Southwest Emerging Technology Symposium 2016 - College of [PDF]

The Department of Mechanical Engineering at The University of Texas at El Paso Proudly Presents

Southwest Emerging Technology Symposium 2016

Saturday, April 9th, 2016 The University of Texas at El Paso 500 W. University Ave. El Paso, Texas 79968



Southwest Emerging Technology Symposium 2016 April 9th, 2016 Wyndham Airport Hotel 2027 Airway Blvd. El Paso, TX 79925

Department of Mechanical Engineering College of Engineering The University of Texas at El Paso 500 W. University Ave. El Paso, Texas 79968 Phone: (915) 747-5450 Fax: (915) 747-5019 [email protected]

Web Link: http://engineering.utep.edu/sets/

FOREWARD Welcome to the Southwest Emerging Technologies Symposium sponsored by Shell. The purpose of the symposium is to encourage communication among the engineers and scientists in and around the El Paso area’s universities and industries. The following individuals and organizations are acknowledged for their assistance with the symposium. Conference Chair:

Dr. Ahsan Choudhuri, The University of Texas El Paso

Technical Chairs:

Dr. Yirong Lin, The University of Texas El Paso Dr. Pavana Prabhakar, The University of Texas El Paso Dr. Norman Love, The University of Texas El Paso

Logistics Committee:

Gloria Salas, The University of Texas El Paso Linda Luna, The University of Texas El Paso

Hosted by:

College of Engineering Department of Mechanical Engineering & MIRO Center for Space Exploration and Technology Research (cSETR) The University of Texas at El Paso

Session Chairs:

Dr. Samia Afrin, The University of Texas at El Paso Mr. David Espalin, The University of Texas at El Paso Dr. Louis Everett, The University of Texas at El Paso Mr. Michael Everett, The University of Texas at El Paso Dr. Angel Flores, The University of Texas at El Paso Dr. Cesar Garcia, Lockheed Martin Aeronautics Dr. Deidra Hodges, The University of Texas at El Paso Dr. Arifur Khan, The University of Texas at El Paso Dr. Chunqiang Li, The University of Texas at El Paso Dr. Chris Navarro, Blue Origin Dr. David Nemir, TXL Group, Inc. Mr. Luis Ochoa, The University of Texas at El Paso Mr. Abraham Trujillo, The University of Texas at El Paso Dr. Jorge Villalobos, Shell WindEnergy Services Inc. Dr. Jianguo Wu, The University of Texas at El Paso

Keynote Speakers:

John Applewhite Chief, Propulsion Systems Branch NASA-Johnson Space Center Charles Chase Sr. Program Manager Lockheed Martin Skunk Works Nick Gonzales Director, Systems Engineering Space Systems Company Lockheed Martin

ii

Southwest Emerging Technology Symposium THE UNIVERSITY OF TEXAS AT EL PASO APRIL 8th – 9th, 2016 April 8th RECEPTION/PREREGISTRATION

The University of Texas at El Paso

Engineering/CCSB Courtyard

6:30 am

REGISTRATION

Wyndham El Paso Airport

Room: Main Lobby

7:30 am

BREAKFAST

Wyndham El Paso Airport

Room: Rosewood

8:00 am

OPENING NOTES

Dr. Yirong Lin Assistant Professor Department of Mechanical Engineering The University of Texas at El Paso

Room: Rosewood

8:05 am

INTRODUCTION

Dr. Ahsan Choudhuri Professor and Chair Department of Mechanical Engineering The University of Texas at El Paso

Room: Rosewood

8:10 am

WELCOME

Dr. Howard C. Daudistel Interim Provost The University of Texas at El Paso

Room: Rosewood

8:30 am

KEYNOTE SPEAKER – I INTRODUCTION Dr. Norman Love

Room: Rosewood

8:30 am

KEYNOTE PRESENTATION – I

Room: Rosewood

4:00 pm

April 9th

TITLE: Integrated LO2/CH4 Technologies and Systems Solutions Relevant to Human Mars Exploration John Applewhite Chief, Propulsion Systems Branch Propulsion & Power Division, NASA JSC

iii

9:50 am

10:00 am

PARALLEL TECHNICAL SESSIONS I Session 1-A

Additive Manufacturing I

Room: Walnut

Session 1-B

Emerging Technologies I

Room: Satinwood

Session 1-C

Space Systems

Room: Sandalwood

Session 1-D Energy and Sustainability I

Room: Orchid

Session 1-E

Room: Poplar

Aerospace and Energy Materials I

SPECIAL SESSION: HIGH SCHOOL INFO SESSION

Room: A/V Theater

Moderator: Dr. Ryan Wicker 10:50 am

BREAK AND POSTER SESSION

Room: Main Lobby

11:10 am

PARALLEL TECHNICAL SESSIONS II Session 2-A

Additive Manufacturing II

Room: Walnut

Session 2-B

Emerging Technologies II

Room: Satinwood

Session 2-C

Propulsion and Energy Technologies I

Room: Sandalwood

Session 2-D

Energy and Sustainability II

Room: Orchid

Session 2-E

Aerospace and Energy Materials II

Room: Poplar

12:10 pm

LUNCH

Room: Rosewood

12:20 pm

KEYNOTE SPEAKER – II INTRODUCTION Dr. Ryan Wicker

Room: Rosewood

12:20 pm

KEYNOTE PRESENTATION – II

Room: Rosewood

TITLE: Interactions Charles Chase Sr. Program Manager Revolutionary Technology Programs at Lockheed Martin Skunk Works 1:50 pm

3:30 pm

PARALLEL TECHNICAL SESSIONS III Session 3-A

Additive Manufacturing III

Room: Walnut

Session 3-B

Emerging Technologies III

Room: Satinwood

Session 3-C

ITAR 1: Propulsion and Energy Technologies II

Room: Sandalwood

Session 3-D

Energy and Sustainability III

Room: Orchid

Session 3-E

Aerospace and Energy Materials III

Room: Poplar

BREAK AND POSTER SESSION

Room: Main Lobby

iv

3:50 pm

PARALLEL TECHNICAL SESSIONS IV Session 4-A

Propulsion and Energy Technologies III

Room: Walnut

Session 4-B

Emerging Technologies IV

Room: Satinwood

Session 4-C

ITAR 2: Propulsion and Energy Technologies IV

Room: Sandalwood

Session 4-D

Energy and Sustainability IV

Room: Orchid

Session 4-E

Aerospace and Energy Materials IV

Room: Poplar

5:30 pm

DINNER

Room: Rosewood

5:40 pm

KEYNOTE SPEAKER - III INTRODUCTION

Room: Rosewood

Dr. Pavana Prabhakar 5:40 pm

KEYNOTE PRESENTATION – III

Room: Rosewood

TITLE: Lockheed Martin Space Systems Company Nick Gonzales Director, Systems Engineering Space Systems Company Lockheed Martin 7:00 pm

ADJOURN

v

ABOUT THE SPEAKER John Applewhite John Applewhite is currently the Chief of the Propulsion Systems Branch in the Propulsion & Power Division (PPD) at the NASA Johnson Space Center (JSC) in Houston. A graduate of Texas A&M University with a BS in Mechanical Engineering, he began his career at NASA in 1989 and now leads a group responsibility for the development and oversight of spacecraft propulsion systems for human spaceflight including systems on the Orion, Commercial Crew, and International Space Station programs. The Propulsion Systems Branch at JSC is also actively involved in technology development and demonstration efforts relevant to human exploration missions with a focus on integrated liquid oxygen (LO2)/methane (CH4) propulsion and integrated spacecraft systems. In his 27 year career at NASA, John has worked primarily in the field of propulsion having previously served as a Subsystem Manager for the Space Shuttle Orbiter Reaction Control System, a System Manager for the ISS propulsion elements and visiting vehicles, a NASA Systems Engineer for the Orbiter Main Propulsion System, and as the Division Chief Engineer with responsibility over all PPD Space Shuttle Orbiter systems including the main propulsion orbital maneuvering and reaction control, and auxiliary power systems. Outside of propulsion, John also served as a Project Lead Engineer for the Orion Crew Module vehicle and as a branch chief in the Vehicle Systems Engineering & Integration Branch.

vi

ABOUT THE SPEAKER Charles Chase Charles Chase manages the Revolutionary Technology Programs organization for the Lockheed Martin Skunk Works. The organization’s charter is to create, mature, and transition a broad range of disruptive technologies with significant Lockheed Martin system impact. Technology focus areas include: power and propulsion systems; RF and optical meta-materials; cognitive and physiological human performance monitoring; new sensor modalities; adaptive and bio-inspired structures; and plasma flow control devices. He recently won the Lockheed Martin Innovate the Future contest with “Power from Low Speed Wind”, and was a semifinalist in the Buckminster Fuller Challenge with “Bio-Inspired Morphing” to eliminate design stagnation. He presented his organization’s work on compact fusion power and propulsion at the 2013 Google “Solve for ” event. He was a low observable engineer on the F-117A production program and has also worked for the Lockheed Martin Space Systems Company as an electromagnetics group lead. He is a co-founder of CBH Technologies, a start-up developing next generation lighting technology and holds numerous patents.

vii

ABOUT THE SPEAKER Nick Gonzales Nick Gonzales serves as Director of Systems Engineering at Lockheed Martin (LM) Space Systems Company (SSC). He is responsible for approximately 1500 engineers and accountable for systems engineering and technology strategy spanning all products, from sensor systems and missile defense to human spaceflight and deep-space missions. Previously, Mr. Gonzales performed the duties of Director and Chief Engineer for Lockheed Martin Space Systems Advanced Interceptor Programs. In this role, he was responsible for guiding the design of Lockheed Martin’s offering for the next generation interceptor kill vehicle. The purpose of this kill vehicle was to enhance homeland defense against missile attack. When the Government decided they would assemble the best industry had to offer into a cohesive design, Mr. Gonzales served as the LM representative to that industry team. Prior to embarking on that journey, Mr. Gonzales dedicated a large portion of his career to developing and fielding the THAAD interceptor. He started this effort in the systems test group where he set up the launch control room for implementation of countdown go/no-go operations and managed the $90M development of the production test equipment used to ATP the missile rounds. He served as the manager of systems integration when the focus was on getting the interceptor qualified for flight. Mr. Gonzales transitioned to CSE during the flight test program, focusing on flight test mission success and on resolving issues discovered during flight operations. Finally, he served as the Director and CE of the THAAD interceptor, driving hard to obtain the approval of the United States Government to release the interceptor to the war fighters in order to defend our troops, friends, and allies in their theater of operation. The THAAD Missile Program employed approximately 200-700 people, dependent on the particular point in the life cycle, and continues to generate 400 million in sales to both US and Foreign Militaries for Lockheed Martin each year. Under his leadership, the THAAD missile team has provided the Missile Defense Agency (MDA) and the Nation with a missile system that has demonstrated groundbreaking capability and reliably achieving a perfect mission success record. Mr. Gonzales has actively participated in and led the development efforts associated with Missile Systems, Missile Software, Missile Structure/Propulsion Elements and all Test Equipment. Mr. Gonzales has over 28 years of successful hands on engineering experience and has held leadership positions in Engineering and Program Management. He received his bachelors of science in mechanical engineering from San Jose State University with a minor in electronics. He is a recipient of the prestigious Lockheed Martin NOVA award and has received various other recognition awards from the corporation. viii

PARALLEL TECHNICAL SESSION I Session 1-A: Additive Manufacturing I Session Chair: David Espalin, UTEP 9:50 am

1A-1

10:10 am

1A-2

10:30 am

1A-3

Session 1-B: Emerging Technologies I Session Chair: Cesar Garcia, LMC 9:50 am

1B-1

10:10 am

1B-2

10:30 am

1B-3

Session 1-C: Space Systems Session Chair: Louis Everett, UTEP 9:50 am

1C-1

10:10 am

1C-2

10:30 am

1C-3

Room: Walnut Wyndham El Paso Airport Design and Development of the Material Handling Components for the Multi3D System S. Ambriz, D. Espalin, J. Coronel, M. Perez, R. Wicker, UTEP Additive Manufacturing for Bonded Composite Joints R. Garcia, E. Acuna, P. Prabhakar, UTEP Smart Parts Fabrication using Electron Beam Melting Additive Manufacturing Technology M. S. Hossain, J. A. Gonzalez, R. Martinez, J. Mireles, Y. Lin A. Choudhuri, R. B. Wicker , UTEP Room: Satinwood Wyndham El Paso Airport Spatially Variant Periodic Structures in Electromagnetics (Presentation Only) Raymond Rumpf, UTEP Numerical Simulation of Ultrasonic Wave Propagation in Fiber-Enhanced Dielectric Nanocomposites for Quality Inspection J. Wu, Y. Lin, B. Tseng, UTEP Location of a Maximum Deflection Point with Fiber Bragg Gratings (FBG) in Polarization Maintaining (PM) Optical Fiber J. Quintana, UTEP Room: Sandalwood Wyndham El Paso Airport Daedalus: LOX/LCH4 Suborbital Testbed Design J.Adams, J. Trillo, A. Johnson, A. Choudhuri, UTEP Trajectory Simulation of a Terrier improved Orion Launch Vehicle D. Camacho, J. Holt, M. Everett, A.l Flores-Abad, UTEP Dynamic Modelling of a Free-Floating Space Manipulator J. Yepez, J. De la Torre, A. Flores-Abad, UTEP

Session 1-D: Energy and Sustainability I

Room: Orchid Wyndham El Paso Airport

Session Chair: Jorge Villalobos, Shell



9:50 am

1D-1

10:10 am

1D-2

10:30 am

1D-3

Bidding in Wholesale Energy Market M. M. P. Chowdhury, J. Juarez, C. Kiekintveld, UTEP Use of Life Cycle Sustinability Assessment for Energy Applications B.A. Benedict, UTEP Transactive Energy Systems E. Galvan, P. Mandal, UTEP



ix

Session 1-E: Aerospace and Energy Materials I Session Chair: David Nemir, TXL Group



9:50 am

1E-1

10:10 am

1E-2

10:30 am

1E-3

Room: Poplar Wyndham El Paso Airport Room Temperature Processed Cuscn Hole Transportation Layers for the use in Perovskite Based Solar Cells J. Galindo, M. Martinez, S. Shahriar, V. Castenada, D. Kava, C. Sana, D. Hodges, UTEP Synthesis of Cu2ZNSNSXSEX-4 Thin Films Absorber Layers for Low-Cost, High Efficiency Thin Film Solar Cells C. Sana, S. Shahriar, J. Galindo, D. Kava, V. Castaneda, M. Martinez, E. Castro, L. Echegoyen, D. Hodges, UTEP Crysatllization of Poly(3-Hexylthiophene) on Carbon Derivatives for Organophotovoltaic Application A. Mishra, V.S.A.Challa, K.C.Nune1, R.D.K Misra, D.Hodges, UTEP



x

SPECIAL SESSION Special Session High School Info Session Moderator: Ryan Wicker, UTEP

Room: AV Theater Wyndham El Paso Airport Dr. Ryan Wicker Professor Mr. and Mrs. McIntosh Muchison I Endowed Chaired Professor Director, W.M. Keck Ceter for 3D Innovation UTEP John Applewhite Chief, Propulsion Systems Branch NASA-Johnson Space Center Nick Gonzales Director, Systems Engineering Space Systems Company Lockheed Martin Michael Everett Staff Engineer UTEP Bowie High School and Hornedo Middle School Stellar Students in Orbit (SSiO)

10:00 am

10:15 am

10:30 am

10:45 am 11:30 am



xi

PARALLEL TECHNICAL SESSION II Session 2-A: Additive Manufacturing II Session Chair: Luis Ochoa, UTEP 11:10 am

2A-1

11:30 am

2A-2

11:50 am

2A-3

Session 2-B: Emerging Technologies II Session Chair: Abraham Trujillo, UTEP 11:10 am

2B-1

11:30 am

2B-2

11:50 am

2B-3

Session 2-C: Propulsion and Energy Technologies I Session Chair: Chris Navarro, Blue Origin 11:10 am

2C-1

11:30 am

2C-2

11:50 am

2C-3

Session 2-D: Energy Sustainability II Session Chair: Jorge Villalobos, Shell



11:10 am

2D-1

11:30 am

2D-2

11:50 am

2D-3

Room: Walnut Wyndham El Paso Airport Developing a Controller for a Copper Wire Embedding Tool J. F. Motta, D. Espalin, R. Wicker, UTEP A Novel Approach to Estimate Corrosion Effect in 3D Printed Bio Material’s using Taguchi Method S. D. Kilari, N. Kim, B. Tseng, UTEP Economic Analysis Between Powder Bed-Based Additive Manufacturing Technologies J. A. Gonzalez, J. Mireles, Y. Lin, R. B. Wicker, UTEP Room: Satinwood Wyndham El Paso Airport Damage Tolerance and Assessment of Unidirectional Carbon Fiber Composites M. Flores, D. Mollenhauer, UTEP Design and Testing of Hybrid Composite Materials for Cryogenic Fuel Tanks R. Avila , Md. Islam , P. Prabhakar, UTEP Stress Corrosion Cracking Susceptibility of Aluminum Foils for Aerospace Applications E. Garcia, C. M. Stewart, UTEP Room: Sandalwood Wyndham El Paso Airport Design of an Optically Accessible High Intensity Turbulence Combustion System A.Acosta-Zamora, A. de la Torre, A. Choudhuri, UTEP Combustion Synthesis of Zirconium Diboride and Hafnium Diboride: Thermodynamic Analysis S. Cordova, E. Shafirovich, UTEP Flame Front Imaging Techniques on a Backward Facing Step Stabilized Flame A. de la Torre, A. Acosta-Zamora, A. Choudhuri, UTEP Room: Orchid Wyndham El Paso Airport Impact of Intermittent Renewable Energy Sources on Power System Analysis L. A. Gutierrez, P. Mandal , UTEP Converting Methane Waste to Valuable Materials R.R.Chianelli, MRTI Comparison of Two Idp Technologies in Detecting and Preventing Cyber-Attacks on Microgrid Communication Networks G. K. Chalamasetty, P. Mandal, and B. Tseng, UTEP



xii

Session 2-E: Aerospace and Energy Materials II Session Chair: Arifur Khan, UTEP



11:10 am

2E-1

11:30 am

2E-2

11:50 am

2E-3

Room: Poplar Wyndham El Paso Airport Zinc Oxide Thin Film Preparation by Single Solution Deposition for Perovskite Solar Cells M. F. Martinez, S. Shahriar, D. Kava, C. Sana, V. Castaneda, J. Galindo, D. R. Hodges, UTEP Pyroelectric Energy Harvesting with High Curie Temperature Material LiNbO3 J. Silva, H. Karim, MD R. H. Sarker, S. Shahriar, M. A. I. Shuvo, D. Delfin, D. Hodges, N. Love, Y. Lin , UTEP Thermal Stability and Oxygen Sensor Characteristic of Ga2O3 Based High Temperature Oxygen Sensors E. J. Rubio, S. Manandhar, C. V. Ramana, UTEP



xiii

PARALLEL TECHNICAL SESSION III Session 3-A: Additive Manufacturing III Session Chair: Cesar Garcia, LMC 1:50 pm

3A-1

2:10 pm

3A-2

2:30 pm

3A-3

2:50 pm

3A-4

3:10 pm

3A-5

Session 3-B: Emerging Technologies III Session Chair: Chunqiang Li, UTEP



1:50 pm

3B-1

2:10 pm

3B-2

2:30 pm

3B-3

2:50 pm

3B-4

3:10 pm

3B-5

Room: Walnut Wyndham El Paso Airport Unified Software for Multi-Functional G-Code: a Method for Implementing Multi Technology Additive Manufacturing E. Aguilera, D. Espalin, E. MacDonald, R. Wicker, UTEP Failure Analysis of Electron Beam Melted Ti-6Al- 4V Tensile Specimen M. S. Haque, E. Arrieta, J. Mireles, C. Carrasco, C. M. Stewart, R. B. Wicker, UTEP Multi3D System: Advanced Manufacturing with Material Handling Robotics J. L. Coronel Jr., S. Ambriz, C. Kim, D. Espalin, R. B. Wicker, UTEP Fabrication and Modeling of Smart Parts using Electron Beam Melting Additive Manufacturing Technology R. Martinez, M. S. Hossain, J. A. Gonzalez, M. A. I. Shuvo, J. Mireles, A. Choudhuri, R. B. Wicker, Y. Lin, UTEP Development of the Thermal Wire Embedding Apparatus for FDM-Printed Parts D. Marquez, UTEP Room: Satinwood Wyndham El Paso Airport Development of High-Temperature Digital Image Correlation Method C. Ramirez, C. M. Stewart, UTEP Optical Second Harmonic Generation Imaging for Ferroelectric Materials Studies Y. Ding, C. Diaz-Moreno, A. Paez, Y.Wang, J. A. López, C. Li, UTEP Metamaterial based Passive Wireless Temperature Sensor for Temperatures up to 500ºC H. Karim, D. Delfin, L. A. Chavez, L. Delfin, J. Avila, C. Rodriguez, R.C. Rumpf, Y. Lin, & A. Choudhuri, UTEP Compressive Properties of Mock Polymer bonded Explosive using Digital Image Correlation C. A. Catzin, C. M. Stewart, UTEP High Temperature Measurement using Lithium Niobate Ceramic Material Md R. H. Sarker, H. Karim, R. Martinez, J. Silva, N.Love, Y. Lin, UTEP



xiv

Session 3-C: ITAR 1 Propulsion and Energy Technologies II Session Chair: Chris Navarro, Blue Origin 1:50 pm

3C-1

2:10 pm

3C-2

2:30 pm

3C-3

2:50 pm

3C-4

3:10 pm

3C-5

Session 3-D: Energy and Sustainabiity III Session Chair: Angel Flores, UTEP 1:50 pm

3D-1

2:10 pm

3D-2

2:30 pm

3D-3

2:50 pm

3D-4

3:10 pm

3D-5

Session 3-E: Aerospace and Energy Matrials III Session Chair: Arifur Khan, UTEP



1:50 pm

3E-1

2:10 pm

3E-2

2:30 pm

3E-3

Room: Sandalwood Wyndham El Paso Airport Decomposition of HAN- and ADN-Based Monopropellants R. Ferguson, E. Shafirovich, UTEP Linear Burning Rate Measurements of Hydroxylammonium Nitrate J. Stahl, E. L. Peterse, Texas A&M University Flame Studies of LMP-103S Decomposition Products K. Hogge, E. Flores, D. Camacho, A. Choudhuri, N. Love, UTEP Development of a Micro-Burner for Hybrid Rocket Combustion Studies J. C. Thomas, J. Stahl, and E. L. Petersen,Texas A&M University High Test Peroxide Thruster Development J. Mona Mejia, A. Choudhuri , UTEP Room: Orchid Wyndham El Paso Airport Analysis of Aerodynamics for Shell Eco Marathon Vehicle using Computational Fluid Dynamics C. Mata, J. F. Chessa, UTEP Study of Learning Effectiveness of Project- based Learning Method on Team-based and Individual-based Projects H. Kim, A. Akundi, Y. Lin, T. Tseng, UTEP Study of Photolysis Rate Coefficients to Improve Air Quality Models for the El Paso-Juarez Airshed S. Mahmud, P. Wangchuk, R. Fitzgerald, W. Stockwell, D. Lu, UTEP Scenario Planning Applications for Energy Issues B.A. Benedict, UTEP Non-Simultanous DG and Capacitor Banks Allocation in Distribution Networks Based on Economic Evaluation S. Sajjadi, P. Mandal, and B. Tseng, UTEP Room: Poplar Wyndham El Paso Airport Temperature Sensing on Woven CFRP by Piezoelectric Particles R. Martinez, E. Tarango, Y. Lin, UTEP Impact Response of Woven Composites A.Castellanos, S. Md Shariful, P. Prabhakar, UTEP Mechanical Properties of Hot Mix Asphalt Materials at Room Temperature for use in Aerospace Landing Applications J. G. Reyes, C. M. Stewart, UTEP



xv

2:50 pm

3E-4

3:10 pm

3E-5

An Element Activation Approach for Modeling Electron Beam Sintering of Titanium C. Beas, J. Chessa, UTEP Artic Exposure Studies of Vinyl Foams for Sandwich Composites C. D. Garcia , P. Prabhakar, UTEP

xvi

PARALLEL TECHNICAL SESSION IV Session 4-A: Propulsion and Energy Technologies III Session Chair: Abraham Trujillo, UTEP 3:50 pm

4A-1

4:10 pm

4A-2

4:30 pm

4A-3

4:50 pm

4A-4

5:10 pm

4A-5

Session 4-B: Emerging Technologies IV Session Chair: Samia Afrin, UTEP 3:50 pm

4B-1

4:10 pm

4B-2

4:30 pm

4B-3

Room: Walnut Wyndham El Paso Airport 500 lbf Liquid Methane - Liquid Oxygen Throttleable Rocket Engine Design J. Carroll, J. Candelaria, J. Trillo, C. Haynes, UTEP Design and Development of a 1 Mw Oxy-Fuel Mhd Combustor B. Lovich, M.J. Hernandez, L. Cabrera, N. Love, UTEP Design and Test of Regenerative Cooling Channels and Injector for a 2000 lbf LOX/LCH4 Engine A. Sandoval, I. Lopez, L. Bugarin, J. Adams, S. Soto, A. Choudhuri, UTEP Conceptual Design and Simulation of a Directly Heated Oxyfuel Supercritical Combustor A.Badhan, A S M A. Chowdhury, D. I. Aguilar, N. D. Love, A. R. Choudhuri, UTEP The Design of a LOX/LCH4 Reaction Control System Engine A. Johnson, D. Ott, P. Nunez, R. Ponce, A. Choudhuri, UTEP Room: Satinwood Wyndham El Airport Free Edge Effect in Multidirectional Laminates M.S. Islam, P. Prabhakar, UTEP Numerical Study of High-Temperature Sco2 Volumetric Receiver for Concentrating Solar Power System. A. Schiaffino, V. Kumar, UTEP

Design Optimization of Sandwich Core and Manufacture through Additive Manufacturing M. Tauhiduzzaman, Md. S. Islam, P. Prabhakar, UTEP



4:50 pm

4B-4

5:10 pm

4B-5

Next Generation Computing Framework for Exascale Simulations A.Chattopadhyay, VMK. Kotteda, V. Kumar, UTEP 2-D Computational Model of a Coaxial Swirl Fuel Injectior J. Aboud, B. Lovich, O. Vidana, L. Cabrera, M.J. Hernandez, N. Love, A. Choudhuri, UTEP



xvii

Session 4-C: ITAR 2 Propulsion and Energy Technologies IV Session Chair: Michael Everett, UTEP 3:50 pm

4C-1

4:10 pm

4C-2

4:30 pm

4C-3

4:50 pm

4C-4

5:10 pm

4C-5

Session 4-D: Energy and Sustainability IV Session Chair: Jianguo Wu, UTEP 3:50 pm

4D-1

4:10 pm

4D-2

4:30 pm

4D-3

4:50 pm

4D-4

5:10 pm

4D-5

Session 4-E: Aerospace and Energy Materials IV Session Chair: Deidra Hodges, UTEP 3:50 pm

4E-1

4:10 pm

4E-2

Room: Sandalwood Wyndham El Paso Airport Laboratory-scale Burning of Composite Solid Propellant using in-situ Synthesized Iron Oxide A. R. Demko, C. Dillier, T. Sammet, K. Grossman, S. Seal, E. L. Petersen, Texas A&M University Delivery System for Bunker J. M. Mejia, J. Holt, N. Love, A. Choudhuri, UTEP Burning Rates of Ap/Htpb-based Solid Rocket Propellants Containing Graphene C. A. M. Dillier, A. R. Demko, T. Sammet, K. Grossman, S. Seal, E. L. Petersen, Texas A&M University Characterization of AF-M315E Propellant for In-Space Applications T.Belcher, J. Valenzuela, N. Love , A. Choudhuri, UTEP Modern Scanning Electron Microscopy in the Study of Solid Propellant Combustion: Surface Structure and Elemental Identification via Eds G. R. Morrow, A. R. Demko, E. L. Petersen, Texas A&M University Room: Orchid Wyndham El Paso Airport Effect of Water and Heat Transport on Three- Phase Transient Behavior of a PEFC A. Nandy, C. Y. Wang, Northern New Mexico College An Analysis of a Photovoltaic Solar System Project in an Institution of Higher Education D. N. De Hoyos, B. Tseng, A.Olivarez, UTEP Hall Effect Measurement Data V. Vidal, UTEP Hough Transform based Automatic Segmentation of Nanofibers from SEM Images Z. Hu, B. Tseng, Y. Lin, J. Wu, UTEP Janus: LOX/LCH4 Robotic Lander Testbed I. Lopez, L. Bugarin, R. Ponce, A. Choudhuri, UTEP Room: Poplar Wyndham El Paso Airport Optimization of Aluminum-Doped Zinc Oxide Thin Films via Variances in Annealing Temperature V. Castaneda, S. Shahriar, C. Sana, M. Martinez, D. Kava, J.Galindo, and D.Hodges, UTEP Study of Tungsten-Yttrium based Coatings for Nuclear Applications G. Martínez, J. Chessa, S. Shutthanandan, T. Tevuthasan, M. Lerche, C.V. Ramana, UTEP

xviii

4:30 pm

4E-3

4:50 pm

4E-4

5:10 pm

4E-5

Characterization and Analysis of Structural Properties, Crystallography and Surface Potential of Perovskite Thin Films S. Shahriar, C. Sana, V. Castaneda, M. Martinez, D. Hodges, UTEP Structural, Dielectric, and Piezoelectric Characterization of Lead-Free Calcium and Cerium Modified BaTiO3 J. A. Duran, C. Orozco, C. V. Ramana, UTEP Physical and Optical Properties of Czts Spin Coating and Doctor Blade Processing D. Kava, S. Shahriar, M. Martinez, C. Sana, J. Galindo, V. Castaneda, D. R. Hodges, UTEP

xix

SETS2016-01 The Southwest Emerging Technology Symposium 2016

DESIGN AND DEVELOPMENT OF THE MATERIAL HANDLING COMPONENTS FOR THE MULTI3D SYSTEM S. Ambriz1,2*, D. Espalin1,3, J. Coronel1,2, M. Perez1, R. Wicker1,2 W.M. Keck Center for 3D Innovation, The University of Texas at El Paso, TX 79968, USA; 2 Deparment of Mechanical Engineering, UTEP, TX 79968, USA; 3 Department of Metallurgical, Materials and Biomedical Engineering, UTEP, TX 79968, USA; 1

* Corresponding author ([email protected])

Keywords: additive manufacturing, Multi3D system, material handling, 3D printed electronics ABSTRACT The final product functionality of parts produced by Additive Manufacturing (AM) can, in part, be improved by the inclusion of multi-material capabilities. The “Multi3D Manufacturing System” that is under development at The University of Texas at El Paso uses material extrusion printing, solid conductor wire embedding, component placement, machine vision and micromaching to enable multi-functional product fabrication. The Multi3D system was designed to transport a workpiece between manufacturing stations via a six-axis robot with a custom designed end-effector, portable build platform (PBP), and a controlled temperature environment or chamber that travels to each manufacturing station. The heated travel envelope (HTE) was included to mitigate thermal shrinkage (and eventually warping) that occurs when a thermoplastic is subjected to a decrease in temperature within a short time frame. Discussed in this work is the design and construction of the HTE and PBP as well as their performance in terms of maintaining the required temperature environment and locating registration within FDM printers. Ultimately, the Multi3D system will be utilized for Aerospace applications to manufacture components for Unmanned Aerial Vehicles (UAVs) and satellites, but other applications where disparate materials are required can be envisioned. 1

Introduction

Additive Manufacturing (AM), otherwise known as 3D printing, is a production process that implements a layer-by-layer approach to fabricate complex 3D objects from CAD models. Research in AM has produced novel processes which have been implemented in various fields such as automotive, biomedical, aerospace, and more. Recently, the fabrication of functional “end-use” products has been a popular trend in the AM field [2]. End-use entails parts that are multi-functional after printing, with more than a structural or aesthetic function. The W.M. Keck Center for 3D Innovation has conducted innovative practices in the past which involve the production of AM parts with embedded electronics. The objective of this research project in general was to further mature these practices by directly minimizing human interaction via the development of the “Multi3D Manufacturing System.” Specifically, the goal was to develop a manufacturing process containing Fused Deposition Modeling (FDM) printers, micro machining, pick and place, wire embedding, machine vision, and a six-axis robot. The Keck Center has conducted research in past for creating a similar system with legacy FDM systems and a pneumatic slide but issues with registration were encountered [1]. The concept of the Multi3D system looked to overcome the limitations of the previous version. As a result, the use of a six axis robot to handle and transport the products was hypothesized to resolve these

P.1

registration issues, by reducing the amount of modifications required to the printers. The capabilities of the Multi3D system (Figure 1) will enable the design and manufacture of novel aerospace components such as UAVs and small satellites. Multiple FDM technologies will enable multimaterial capabilities, which can lead to other various applications where disparate materials are required. 2

Material Handling Components

six-axis robot

FDM

portable build platform

FDM

CNC router

Fig. 1. Photograph of Multi3D Manufacturing System

In order to achieve the overarching goal of the Multi3D system, a methodology was conceived for material handling with three specific components which can appropriately handle and transport the parts developed. The portable build platform (PBP), the heated travel envelope (HTE), and a custom end-effector for the six-axis robot were designed and developed. Figure 2 shows the prototype CAD models of these three components. The purpose of the PBP was to be a platform for which the parts can be produced on, similar to the platform (Platen) normally found in these Stratasys FDM Fortus 400mc printers. The PBP has the ability to grasp a sacrificial build sheet via vacuum similar to the Platen. The PBP differs from the Platen in the aspect that it is removable and portable via the MH50 robot and is able to locate (dock) into each technology of the system. The HTE is a forced convection oven with an actuating curtain door. The purpose of the HTE was to encompass the build area, during the transport of the work piece, in a heated environment to closely replicate the temperatures of the FDM’s heated envelope (enclosure). Maintaining the heated environment around these components will mitigate poor filament bond strength, which ultimately effects mechanical properties in FDM printed parts [3]. The custom endeffector was another component that heated travel was conceived which mounts to the end envelope of the MH50 robot. The end-effector was custom designed to grasp and support the PBP and HTE during transport from station to station of the six-axis robot Multi3D system. The end-effector design used a fork-lift type of beams for support, a pneumatic parallel gripper portable build for grasping and restraining these platform end-effector components, and a port for supplying vacuum to the platform during Fig. 2. CAD model of material handling components transport. 3

Registration Methodology

The designs of the material handling components were evaluated with several experiments which shaped and optimized their designs for the required application. After the designs of

P.2

these components, the required modifications of the different technologies and procedures of the system were all solidified, optimal registration between the FDM printers was required. A calibration print was designed with indicators which work to identify registration in the x-y plane between the printers build area. A calibration print is split so certain sets of indicators are printed on the first FDM and another set are printed on the second FDM. These calibration prints are analyzed and measured optically to find the x, y, and angle offsets between the printers. The calibration prints were reiterated 14 times to work towards the closest achievable registration between the two FDMs. Slight physical modifications to the printers was done to adjust for recorded offsets between each print. The x-offset and y-offset was compensated for first and once the offsets were minimized, the angle offset was compensated for. Figure 3 shows a comparison between calibration print # 11 and #14 y-offset data obtained, where slope difference indicates angle offset and y-intercept difference indicates y-offset. Calibration Print #11 , y-offset

0.05

y = 0.0066x - 0.0184

0.03

y = 0.0037x - 0.0009

0.02 0.01

FDM2 FDM1

0 -0.01

0

2

4

6

8

10

0.15 y = 0.0208x - 0.0007

0.1 0.05

FDM2 FDM1

0

-0.05

-0.02

y = 0.021x - 0.0079

0.2 Y-position (in)

Y-position (in)

0.04

Calibration Print #14 , y-offset

0.25

0

X-position (in)

2

4

6

8

10

X-position (in)

Fig. 3. Comparison of graphs of data obtained from calibration prints #11 and #14

5

Results & Conclusions

The final values of offset recorded in the final calibration print were 0.0027” in the x, 0.0072” in the y, and a .02 degree difference in angle. With the registration between the two FDMs optimized, a demonstration experiment was designed for development of a multi-color part (Figure 4). A total of 82 transfers was conducted between the two printers, each printing 2 layers at a time prior to transfer. This experiment demonstrated the repeatability of the material handling and system in general in the development of a complex part. Future work with this system will entail the development of multi-functional products which will demonstrate capabilities of wire embedding, foil application, and pick and place of components.

Fig. 4. Multi-color part developed by Multi3D system

References [1] D. Espalin, J. Ramirez, F. Medina, and R. Wicker "Multi-material, multi-technology FDM: exploring build process variations." Rapid Prototyping Journal, Vol. 20, Issue 3, pp 236-244, 2014. [2] N. Guo and L. C. Ming "Additive manufacturing: technology, applications and research needs." Frontiers of Mechanical Engineering, Vol. 8, Issue 3, pp 215-243, 2013 [3] Q. Sun, G.M. Rizvi, C.T. Bellehumeur, and P. Gu "Effect of processing conditions on the bonding quality of FDM polymer filaments." Rapid Prototyping Journal, Vol 14, Issue 2, pp 72-80, 2008

P.3

SETS2016-02 The Southwest Emerging Technology Symposium 2016

ADDITIVE MANUFACTURING FOR BONDED COMPOSITE JOINTS 1

R. Garcia1, E. Acuna1, P. Prabhakar1* Department of Mechanical Engineering, University of Texas at El Paso, El Paso, TX 79902, USA; * R. Garcia ([email protected]) * E. Acuna ([email protected]) * P. Prabhakar ([email protected])

Keywords: adhesive, joint, shear ABSTRACT With increasing applications of fiber reinforced polymer matrix composites (FRPCs) in aircrafts, navy structures and automobiles, joining technology to fabricate large components has become a priority for structural engineers. Conventional materials such as steel or aluminum are joined using fasteners and/or bolted joints, which is not favorable for FRPCs as drilling or cutting of the composite may damage fibers causing an adverse effect on the structural integrity. Thus, adhesively bonded joints are becoming a viable option for joining FRPCs. Advantages of bonded joints over traditional mechanical fasteners are lower structural weight and improved damage tolerance. Despite these advantages, bonded joints in primary load-bearing applications often result in overdesign due to the inclusion of mechanical fasteners for additional safety. This is due to the lack of confident adhesive joint materials and designs for composite joining technology. Mechanics based designs for bonded joints are necessary to facilitate efficient use of composites for lightweight applications. This paper presents the use of additive manufacturing to improve the mechanical behavior of bonded joints by redistributing the stresses within the bond area. The effort was split into the following main steps: 1) Conduct experimental study of pure epoxy adhesive joints; 2) Analyze the bonded joint systems computationally to determine the stress distributions; 3) Modify the bond design computationally to redistribute stresses; 4) Enable the optimum designs at the bond region accurately using polymer additive manufacturing; 5) Conduct SLS tests on the new designs to check whether an improvement in the apparent shear strength was achieved by this method.

Fig.1. Dimensions of joints fabricated.

Fig.2. Three modes of failure in bonded substrates.

P.4

Fig.3. Failure types in lap shear tests.

Fig.4. Top view of adhesive failure.

Fig.5. Top view of cohesive failure.

Tab.1. Tensile results per joint.

Fig.6. Un-deformed finite element model.

Fig.7. Deformed finite element model.

1

Lap Shear Study Overview

1.1 Joint Preparation The type of adhesive joint being examined in this report is referred to as a lap joint. The standard used all-throughout the study is ASTM D5868-01[1]. Two carbon fiber substrates were bonded together with LOCTITE Hysol E-120HP epoxy. A total of 5 joints were made and tested. All of the joints were cured at room temperature for 24 hours. The dimension of the joints is shown in Figure 1. The bond line thickness of 0.76 mm was reached by using

P.5

ADDITIVE MANUFACTURING FOR BONDED COMPOSITE JOINTS

glass microspheres, which provided a bond line control of 700-800 microns. The mass ratio between adhesive and microspheres was 10:1. The microspheres were mixed thoroughly prior to being applied on to the substrates. Once the epoxy-microspheres mixture was applied to the 6.45 cm2 overlap, pressure was exerted surrounding the overlap area using c-clamps. It is important to note that this study accounts only for mode II shear failure in bonded substrates, as seen in Figure 2. 1.2 Experimental Procedure According to ASTM D5868-01[1]: (1) Initial grip separation is 75 mm with 25.4 mm minimum of each sample end held in the test grips, (2) The specimen loading rate is 13 mm/min. Note that a loading rate of 0.5 is an important difference compared to other ASTM standards, (3) Prepare a minimum of five lap shear samples in each case and test. 1.3 Digital Image Correlation (DIC) Universal spray paint was applied to one of the edge surfaces per joint. White paint was thoroughly sprayed first. Then, black paint was faintly sprayed to create optically visible dots in the already white surface. Ultimately, the tensile test will be continuously photographed with time intervals using a high quality camera, and then processed through a special software that will determine the separation between dots – deriving in displacement and strain of the joint. 1.4 Computational Model For the computational model, first an entire bonded joint specimen was modeled (see Figures 6 and 7) to determine the local displacement field at the top and bottom edge of the uniform epoxy bond region. This was followed by modeling the bond region with different designs incorporated with the displacement loading applied on the boundaries. The stress distribution was determined for each case, and the ones with uniform stress distribution compared to the pure epoxy case were chosen and printed using polymer additive manufacturing technique. 2

Results

2.1 Discussion The final results per joint were obtained (see Table 1). The ASTM standard required for the publication of individual peak load values, KPa, failure type, and averages by maximum and minimum values. Failure type of the lap joints was decided upon looking at Figure 3. The only two failure modes observed with the five joints tested was adhesive and cohesive. The adhesive failures can be clearly observed in Figure 4. Note that the glass microspheres are visible in Figure 4. The only cohesive failure can be observed in Figure 5. If a conclusion can be made from these five samples, it is that adhesive failure is more common in lap joints than cohesive failure. Therefore, the peak loads are usually going to be less in these types of joints, although five samples are not enough to actually state this as a fact. References [1] ASTM D5868-01 “Standard Test Method for Lap Shear Adhesion for Fiber Reinforced Plastic (FRP) Bonding”. ASTM International, West Conshohocken, PA, 2014.

3 P.6

SETS2016-03 The Southwest Emerging Technology Symposium 2016

SMART PARTS FABRICATION USING ELECTRON BEAM MELTING ADDITIVE MANUFACTURING TECHNOLOGY

1

Mohammad S. Hossain1,2*, Jose A. Gonzalez1,3, Ricardo Martinez Hernandez2, Jorge Mireles1,2 ,Yirong lin2, Ahsan Choudhuri2, and Ryan B. Wicker1,2 W.M. keck Center for 3D Innovation, The University of Texas at El Paso (UTEP), TX 79968, USA; 2 Department of Mechanical Engineering, UTEP, TX 79968, USA; 3 Department of Metallurgical, Materials and Biomedical Engineering, UTEP, TX 79968, USA; * Corresponding author ([email protected])

Keywords: Smart parts, additive manufacturing, electron beam melting, Ti-6Al-4V, piezoceramic ABSTRACT Energy system components with embedded piezoceramic sensors, or smart parts, can be beneficial in obtaining real time performance feedback during operation. Additive manufacturing (AM) technology has been identified as an enabling technology to achieve smart parts where the layerwise process allows the non-intrusive placement of a sensor in any specific location while not altering the part’s design and/or functionality. The embedded sensor within a desired location can allow the real time feedback information from critical areas of smart parts that are exposed directly to high temperatures and pressures. The research herein focuses on the manufacturing of smart parts through the development of a ‘stop and go’ technique. The part’s sensor functionality was demonstrated and approximately 3V was detected with a maximum pressure not exceeding 40 MPa. The research shows the viability of fabricating smart parts using electron beam melting (EBM) AM technology, or a class of powder bed fusion. The fabrication of metallic parts using EBM demonstrates the possibility of fabricating complex smart parts that can revolutionize energy, aerospace, and biomedical industries for applications like pressure tubes, swirlers, air-fuel premixers, and turbine blades. 1 Introduction Embedded sensor within metallic part have tremendous potential to improve in-situ monitoring of critical parameters such as temperature, pressure, and structural health monitoring across diverse applications ranging from turbine blades to combustors. Traditional placement of sensors such as thermocouples onto energy system components (e.g., pressure tubes and turbine blades) typically leaves the sensors on the outer surface of a part, which leaves the sensor unprotected to harsh environments and can lead to nonfunctionality of the sensor that can affect sensor life. Additionally, the sensor placement also involves the use of paste or through an assembly process where a cavity (not present in part’s design) is machined to house the sensor. The addition of a post-process technique for sensor placement can thus interfere with part performance. For example, the aerodynamic performance of a turbine blade can be affected using such cavities. Moreover, the sensor placed in the outer surface may not convey the actual condition or result of the inner surface. These technical challenges result in a need to find alternative advanced manufacturing solutions to produce smart parts.

P.7

A prominent advanced manufacturing technology is electron beam melting (EBM), which is a class of powder bed fusion additive manufacturing (AM) technology that enables fabrication of metallic parts from precursor powder using an electron beam. The EBM technology is well recognized in the field of biomedical, automotive, and aerospace industries for enabling the fabrication of complex shaped metallic parts [1]. EBM technology fabricates parts using metal powder deposited in a layer-by-layer fashion, which enables the opportunity to embed sensors within a desired location without altering the part’s design and/or functionality. The EBM technology was used in this research to demonstrate the creation of smart parts. As a proof of concept, a cylindrical shaped part was fabricated containing piezoceramic material embedded within the part’s body. The process consisted of pausing part fabrication to allow the piezoceramic material to be embedded within a designed cavity. The piezoceramic material can serve as a pressure sensor in the presence of a dynamic load where the piezoelectric effect is measured and related to an applied force [2]. The resulting object is a metallic part with an embedded sensor, or a smart part. Previously, different approaches were taken to embed sensors using AM technology. Li et al. [3] showed the embedding process of thermo-mechanical sensors within a metallic structure using shape deposition modeling technology. The research showed the delamination between the embedded layers and underlying layers due to thermal stress during the fabrication process [3]. Aguilera et al. [4] demonstrated the fabrication of an electromechanical device, or a motor using material extrusion technology. Similar to this work, the process involved pausing the material extrusion process to insert components within the part’s body [4]. To the knowledge of the author’s, a demonstration of embedding sensors through a powder bed fusion process has not been previously demonstrated. The high operating temperature (>7000 C) of Ti-6Al-4V (commonly used powder) using EBM limits the selection of sensors that can be placed during the fabrication process. The work presented here shows the feasibility of embedding sensors within metallic parts using EBM technology that can help enable the fabrication of next- generation smart parts. This manuscript focuses on design, fabrication, material characterization, and sensing capabilities of smart parts enabled by AM, which can be applied toward engineering applications ranging from aerospace, energy, and biomedical sectors. 2 Materials and Methodology 2.1 Materials Ti-6Al-4V powder from Arcam AB (Mölndal, Sweden) was used to fabricate the metal parts used in this research. Fig. 1(a) shows the necessary components used to fabricate the smart parts and Fig. 1(b) shows the assembled view of Fig. 1. Smart parts (a) different components and (b) assembled view the smart parts. The sensor assembly contains piezoceramic material, electrodes, and insulating materials of lead ziconate titanate (PZT), tungsten (W), and alumina, respectively.

P.8

SMART PARTS FABRICATION USING ELECTRON BEAM MELTING ADDITIVE MANUFACTURING TECHNOLOGY

2.2 Stop and Go process A process was developed in this research to achieve the insertion of piezoelectric material within a metallic body at a designed location, which involved pausing the part fabrication process. Here, the sensor was inserted in a pre-designed cavity within the part. Fig. 2 shows the steps involved with the stop and go process, which starts with Fig. 2. Stop and go process a designed cavity within a solid object. The designed cavity is fabricated and the sensor components (i.e. piezoceramic, wires, and sensor housing) are manually inserted (sequence shown in Fig. 1(a)) within the cavity (step 1-2). To restart fabrication, a mask plate is machined (step 3) and the assembled part is placed within the mask plate (step 4). The fabrication continues through standard operating guidelines, completing the fabrication (step 5), and obtaining the final part (step 6). 3 Results 3.1 Smart Parts Fig. 3 shows the fabricated smart parts using the aforementioned stop and go process. The fabricated smart part was ̴ 24 mm in diameter and 28 mm in height. An initial assessment of the electrical components was performed to ensure sensing capability, which involved Fig.3. Smart Parts evaluating the open circuit condition between the electrodes touching the piezoceramic sensor. A SPERRY DM-4400A digital multi-meter (SPERRY Instruments, Menomonee Falls, WI) was used to perform the open circuit test. 3.2 Material Characterization Fig. 4 shows the microstructure of the bottom and top part demonstrating the interfacial bonding of a representative sample. Lamellar like microstructure with alternating α and β phases were observed in the micrographs. The average α grain growth was found to be ̴2.4μm lath thickness within the bottom section, while the top section showed α grain growth averaging ̴1.3μm lath thickness. The heat gradient of the fabrication process was the likely contributor to the microstructural variations observed here [5]. 3.3 Force sensing A compression-compression dynamic loading condition at a 10 Hz frequency was applied to the smart parts. Fig. 5 shows the piezoelectric voltage response in correspondence to the applied force exerted on the smart parts. The results show a good agreement between

3 P.9

the applied force and the electric response, thus supporting the research aim to show the potential of new advanced manufacturing methods to produce smart parts.

Fig.4. Microstructure characterization of smart parts

Fig.5.Voltage response obtained from smart parts in accordance of the applied force

4 Conclusion The work presented here demonstrated the feasibility of fabricating smart parts using powder bed fusion AM technology. The development of a method where a sensor is embedded into the structure can enable a longer sensor life during use under harsh environment. The stop and go process developed here enables the sensor material to be placed within a metallic part at a designed location within the part’s body without altering the part’s overall design and/or functionality. The methods described herein can be used to fabricate smart parts for diverse engineering applications in the field of aerospace, biomedical, and aerospace industries. Acknowledgements This research was supported by the U.S. Department of Energy (DOE) (award No. DEFE0012321). Findings, opinions, conclusions, or suggestions herein are those of the authors, and do not necessarily reflect the views of the DOE. References

[1] I. Gibson et al. “Additive Manufacturing Technologies.” NY: Springer, 2010 [2] A.A. Vives, “Piezoelectric Transducers and Applications”, 2nd ed., Springer, NewYork, 2008 [3] X. Li, A. et al. “Shape deposition manufacturing of smart metallic structures with embedded sensors”, Proceedings of the SPIE’s 7th Annual Int. Symp. on Smart Structures and Materials (2000) 160–171 [4] E. Aguilera, et al., “3D printing of electro mechanical systems”, in: Proceedings of 2013 Annual International Solid Freeform Fabrication Symposium, Austin, TX, 2013 [5] L.E. Murr, et al., “Microstructure and mechanical properties of electron beam-rapid manufactured Ti–6Al– 4Vbiomedical prototypes compared to wrought Ti–6Al–4V”, Mater. Charact. 60(2009) 96–105.

P.10

SETS2016-04 The Southwest Emerging Technology Symposium 2016

NUMERICAL SIMULATION OF ULTRASONIC WAVE PROPAGATION IN FIBER-ENHANCED DIELECTRIC NANOCOMPOSITES FOR QUALITY INSPECTION J. Wu1*, Y. Lin2, B. Tseng1 1 Department of Industrial, Manufacturing and Systems Engineering, University of Texas at El Paso, El Paso, TX 79968, USA; 2 Department of Mechanical Engineering, University of Texas at El Paso, TX 79968, USA; * Corresponding author ([email protected])

Keywords: dielectric nanocomposites; wave propagation; ultrasonic attenuation

ABSTRACT The dielectric nanocomposites where ceramic fibers are mixed in polymer matrix could achieve significantly higher energy density when used in dielectric capacitors. The alignment of these fibers in the polymer matrix have significant influence on the final dielectric properties. This paper investigates the feasibility of utilizing ultrasonic testing to inspect the aligning quality of ceramic fibers through microstructure modeling and wave propagation simulation. The results indicate that the ultrasonic testing has great potential as an NDE tool to inspect the quality of dielectric nanocomposites. 1

Introduction

With an increasing demand on energy storing capabilities, more and more research scientists are focusing on the development of capacitors of high energy density [1]. Dielectric capacitors are most commonly used for their excellent dielectric properties, easy accessibility and low cost. The traditional dielectric materials include polymer and ceramics. Compared with polymer-based capacitors, ceramic-based capacitors have higher relative dielectric permittivity. However, their high dielectric permittivity is at the cost of lower break-down strength, which limits their energy density and performance. To mitigate this issue, many researchers are developing nanocomposites where ceramic fiber materials with higher relative dielectric permittivity are dispersed into polymer matrix that have higher breakdown strength [2, 3] . These nanocomposites could improve both the dielectric permittivity and break-down strength of capacitors, thus achieving a higher energy density. It is found that the alignment of ceramic fibers could significantly influence the dielectric properties [4]. Well-aligned high aspect ratio fibers with the same orientation could increase both permittivity and breakdown strength. The alignment of fibers in polymer matrix has been achieved by many methods, such as extrusion methods [5], and uniaxial stretching assembly [2]. To facilitate a scale-up production, a fast-yet-effective quality inspection technique is critically important to ensure the quality of fiber alignment. The standard method is through microscopic images, which are time-consuming and costly to obtain. Ultrasonic testing is one of the most popular nondestructive evaluation techniques and has been widely used in thickness measurement, flaw detection and microstructure characterization [6, 7]. This paper investigates

P.11

the feasibility of using ultrasonic nondestructive testing method to inspect the alignment of lead zirconate titanate (PZT) in polyvinylindene fluoride (PVDF) matrix through microstructure modelling and ultrasonic wave propagation approach. The rest of this paper is organized as follows. Section 2 introduces the microstructure modeling and wave propagation approach. Section 3 presents the simulation results and conclusion. 2

Microstructure Modeling and Wave Propagation Simulation

2.1 Microstructure Modeling The microstructures of size 0.8mm × 0.8mm are generated for wave simulation. In the microstructure modeling, the fiber alignment is evaluated by the variance σ2 of its orientation angle 𝜃, the angle between the wave propagation direction and the fibers. For example, for the perfect alignment, 𝜃 = 90° or θ~𝑁(90, σ2 = 0); for well-alignment, 𝜃~𝑁(90, 𝜎 2 ) where σ2 is a small value; if there is no specific orientation, the 𝜃 of all fibers are totally random with 𝜃~Uniform(0°, 360°) . The PZT fiber length and width are set to be 50 𝜇𝑚 and 4 𝜇𝑚 respectively. The PZT fiber are uniformly distributed in the PVDF matrix following complete spatial randomness (CSR). Each pixel of the generated microstructural image is set to be Δ𝑥 = 1 𝜇𝑚, which is the spatial step size in wave propagation simulation. The volume fraction of the fibers is set to be about 5%. Three simulated nanocomposites are illustrated in Fig. 1. (1)

(2)

(3)

Fig. 1. Three generated nanocomposites of different fiber alignment: (1) perfect alignment; (2) 𝜃~𝑁(90,102 ); (3) totally random.

2.2 Wave Propagation Simulation Table 1. Acoustic parameters of PZT and PVDF used in the simulation [8]. Density (g/mm3)

Normal velocity(m/s)

Shear velocity(m/s)

PZT

0.0076

4410

2630

PVDF

0.00178

2200

775

The elastodynamic finite integration technique [9] is used to simulate the wave propagation process. The transducer with size of 0.6 mm and central frequency of 20MHz is attached to the

P.12

3

left side of the nanocomposites to generate ultrasonic pulses. The boundary condition is set to be absorptive on the left, top and bottom side, and reflective on the right side. The time step is set to be Δ𝑡 = 1.6 × 10−4 𝜇𝑠. The acoustic parameters of PZT and PVDF is shown in Table 1. The snapshots of the wave propagation process for 𝜃 = 90° are illustrated in Fig. 2.

Fig. 2. Snapshots of wave propagation process at 𝑡 = 0.16 𝜇𝑠 (left) and 𝑡 = 0.32 𝜇𝑠 (right)

2.3 Ultrasonic Attenuation Calculation Ultrasonic attenuation is one of the most commonly used ultrasonic parameters for material characterization. It refers to the decaying rate of the acoustic wave as it propagates through materials. It contains rich information about the microstructures and can be used to infer the fill alignment. The ultrasonic attenuation (dB/mm) is calculated using the spectral ratio analysis technique [6] as: 1

𝑆 (𝑓)

𝐴(𝑓) = 2𝑑 ln |𝑆1 (𝑓)| 2

(1)

where 𝐴 is the attenuation, 𝑑 is the thickness of the specimen, 𝑠1 and 𝑠2 are the frequency spectrum of the initial pulse and the first echo respectively. For each type of microstructures, simulation is repeated 40 times to account for the randomness. 3

Results and Conclusion

Fig. 3 shows the ultrasonic attenuation as a function frequency under four fiber orientation conditions. For the fourth case, 𝜃 = 0°, which means the ultrasonic wave propagates along the direction parallel to the perfect aligned fibers. From the attenuation curves we can see that when the ultrasonic wave propagates along the path perpendicular to perfectly aligned fibers, the ultrasonic energy has the lowest loss. In contrast, when the wave path is parallel to these perfectly aligned fibers, the wave energy reach the highest level. The energy losses of the other two cases are between these two mentioned ones. The energy loss for the uniform case is higher than that when 𝜃~𝑁(90, 102 ) . Based on these curves, we found that the more fibers perpendicular to the wave path, the lower the energy scattering loss. The more fibers parallel to the wave path, the higher the energy scattering loss. This finding could be used to develop ultrasonic testing based quality evaluation techniques to evaluate the fiber aligning quality in dielectric nanocomposites manufacturing. In our future work, experimental study will be conducted to evaluate the ultrasonic testing method.

3 P.13

Fig. 3. Ultrasonic attenuation as a function of frequency under four fiber orientation conditions.

4 [1] [2] [3] [4] [5] [6] [7] [8] [9]

References Y. Cao, P. C. Irwin, and K. Younsi, "The future of nanodielectrics in the electrical power industry," Dielectrics and Electrical Insulation, IEEE Transactions on, vol. 11, pp. 797-807, 2004. H. Tang, Y. Lin, and H. A. Sodano, "Enhanced energy storage in nanocomposite capacitors through aligned PZT nanowires by uniaxial strain assembly," Advanced Energy Materials, vol. 2, pp. 469-476, 2012. M. Rajib, R. Martinez, M. Shuvo, H. Karim, D. Delfin, S. Afrin, G. Rodriguez, R. Chintalapalle, and Y. Lin, "Enhanced Energy Storage of Dielectric Nanocomposites at Elevated Temperatures," International Journal of Applied Ceramic Technology, vol. 13, pp. 125-132, 2016. V. Tomer and C. Randall, "High field dielectric properties of anisotropic polymer-ceramic composites," Journal of Applied Physics, vol. 104, 2008. L. Chen, Y. Hong, X. Chen, Q. Wu, Q. Huang, and X. Luo, "Preparation and properties of polymer matrix piezoelectric composites containing aligned BaTiO3 whiskers," Journal of materials science, vol. 39, pp. 2997-3001, 2004. J. Wu, S. Zhou, and X. Li, "Ultrasonic Attenuation Based Inspection Method for Scale-up Production of A206–Al2O3 Metal Matrix Nanocomposites," Journal of Manufacturing Science and Engineering, vol. 137, p. 011013, 2015. Y. Liu, J. Wu, S. Zhou, and X. Li, "Microstructure Modeling and Ultrasonic Wave Propagation Simulation of A206–Al2O3 Metal Matrix Nanocomposites for Quality Inspection," Journal of Manufacturing Science and Engineering, vol. 138, p. 031008, 2016. J. E. Mark, Physical properties of polymers handbook: Springer, 1996. P. Fellinger, R. Marklein, K. Langenberg, and S. Klaholz, "Numerical modeling of elastic wave propagation and scattering with EFIT—elastodynamic finite integration technique," Wave motion, vol. 21, pp. 47-66, 1995.

P.14

SETS2016-05 The Southwest Emerging Technology Symposium 2016 LOCATION OF A MAXIMUM DEFLECTION POINT WITH FIBER BRAGG GRATINGS (FBG) IN POLARIZATION MAINTAINING (PM) OPTICAL FIBER Joel Quintana1* 1 Department of Electrical and Computer Engineering, UTEP, El Paso, TX 79968,USA; * Corresponding author ([email protected])

Keywords: Bragg, sensor, location, deflection, polarization

ABSTRACT The recent prevalence of composite laminated or fully composite, reusable rocket structures, it is of critical need to evaluate the structural health of each rocket before, after and possibility during each launch. FBG sensors have proven to be excellent devices for structural health monitoring (SHM) and are quickly becoming economically advantageous over electronic, conventional strain gauges [1]. FBG sensors written in PM fiber offer additional dimensions of strain measurement per sensing element, reducing the number of sensing units per area. We build a strain profile of a flat composite plate with concentrated load acting at the center. The FBG PM sensors are placed at random on the surface of the plate. We then calculate spectral shifts caused by the applied axial and traversal strains on the sensors and compare to a predetermined optical baseline. From the optical signal, the shift deltas in wavelength are parsed to correspond to axial and transversal strains. This exercise validates a method of optical detection and shift calculation for multi-axis sensors as an automated, integrated system. Of particular interest is defining the minimum number of sensors required to detect a critical flaw. In this spirit, this paper defines the theoretical minimum number of the PM FBG sensing elements needed to detect a point of maximum deflection on the panel. The point of maximum deflection has been located on a Cartesian plane by formulating orthogonal vectors from the strain tensor of each sensing element at the its location. A group of vector intersects with the minimum Euclidian distance from each other correspond to said point. 1 The Strain Sensor Fiber Bragg gratings in an optical fiber core are primarily defined by, !! = 2!! Λ where the reflected Bragg wavelength λB is the product of the grating effective refractive index ne and the period of refractive index modulation Λ. The overall fractional shift !! , of the Bragg !! wavelength due to the three-dimensional strain (!! , !! , !! ) and temperature (Δ!) are described by[2] equations (1) and (2). These equations assume that the fiber is optically, thermally and mechanically isotropic. Pockel’s constants are !!! , !!" (strain-optic coefficients), ! is the ! thermal expansion coefficient, and !! is the thermal-optic coefficient. When a traversal !" strain is present, birefringence occurs in the fiber core, and the Bragg wavelength exhibits a peak separation as the polarization modes align themselves to the two different effective refractive indices. The Bragg wavelength splits into two peaks representing each of the orthogonal polarization axes. This effect only occurs in the area that is experiencing the strain. The spectral peak separation however, is automatically present in a PM FBG sensor.

P.15

This is due to stress applying parts integrated in the PM fiber cladding. The major propagation axes are commonly designated as the “Fast” and Slow” axes, this designation has ∆! been implemented in the effective index of refraction and wavelength terms !! , !! , ! ! , !∆!! . !"

∆!! !!" ∆!! !!"

= 1− = 1−

!!! ! ! !!

!

!!" !! − !!" !! −

!!! ! ! !!

!

!!! !! − !!" !! −

!!! ! ! !!

!

!!" !! + !!! !! +

! !!! !!! !" ! !!! !!! !"

+ +

!!! ! ! !!

!

!"

(!!! + 2!!" )! Δ!

(1)

(!!! + 2!!" )! !"

(2)

The major propagation axes are commonly designated as the “Fast” and Slow” axes, this designation has been implemented in the effective index of refraction and wavelength terms ∆!! ∆!! !! , !! , , . It’s important to note that the slow and fast axes designate the polarization ! ! !"

!"

directions with respect to the fiber stress applying parts, and not the arbitrary spacial x and y axes, though one can certainly choose to align the slow axis to the x axis or y axis if desirable [4] , and for this work, it is assumed the slow axis lays parallel to the material surface.

Fig.1. Polarization maintaining fiber Bragg grating sensor and it’s polarization axes.

2

Host Material and Strain Evaluation

The host composite material is mathematically characterized by the deflection solution of a circularly clamped plate with concentrated point load acting at the center. The deflection at any point of the plate at a distance r from the center is [5] w=

Pr 2 r P log + (a2 − r 2 ) a 16π D 8π D .

The deflection w is a function of the point load, P, the clamp radius, a, and material properties D which includes Poisson's ratio of .3 and Young’s Modulus of 70 GPa. These values correspond to a homogeneous 0/90° Carbon Tape, pre-preg composite material. N The deflection w is circularly symmetric changes and yields strain profiles in the radius and theta

P.16

(3)

LOCATION OF A MAXIMUM DEFLECTION POINT WITH FIBER BRAGG GRATINGS (FBG) IN POLARIZATION MAINTAINING (PM) OPTICAL FIBER

directions as shown in figure 3. We assume sensor placement on the surface of the plate. Out of plane stress and shear strains all equal 0.

Fig.2. Deflection of circularly clamped composite plate with point load in center and sensor locations

Fig.3. Strain profile

3

Optical Response and Deflection Location

The optical response of each sensor to the external strains at their corresponding locations on the plate was calculated and applied to the untrained baseline optical signal. This is shown in figure 4. It is assumed that the strain is uniform throughout the ~5mm sensor. The shifts in wavelength are parsed [6] into component strain vectors pointing outward from each sensor.

⎛ ε ⎞ ⎛ −ε a ⎜ a ⎟, ⎜ ⎜ εs ⎟ ⎜ εs ⎝ ⎠⎝

⎞⎛ ε ⎟, ⎜ a ⎟ ⎜ −ε s ⎠⎝

⎞ ⎛ −ε a ⎟ and ⎜ ⎟ ⎜ −ε s ⎠ ⎝

⎞ ⎟ ⎟ ⎠

(4)

The strain composite vector has 4 directions in the 2 dimensional strain plane (x,y). The four vectors are superimposed on the deflection plane, one of which is pointing in the direction of increasing deflection. A minimum of 2 sensors is needed to find the point of max deflection,

3 P.17

by strain vector intersections. The sensors however must not share a location with a vector of another sensor.

Fig.4. Optical wavelength unstrained vs. strained sensor

Fig.5. Strain vector lines on deflection plane

References [1] Udd, et al. "Multidimensional strain field measurements using fiber optic grating sensors," Proc. SPIE 3986, p.254 (2000). [2] Kersey, et al. "Fiber grating sensors." Journal of Lightwave Technology 15(8), 1442-1463 (1997). [3] Bosia, et al. "Characterization of the response of fibre Bragg grating sensors subjected to a two-dimensional strain field," Smart materials and Structures 12(6), 925 (2003). [4] Botero-Cadavid.,et al. "Spectral Properties of Locally Pressed Fiber Bragg Gratings Written in Polarization Maintaining Fibers," Journal of Lightwave Technology 28(9), 1291-1297 (2010). [5] Timoshenko, Stephen P., and Sergius Woinowsky-Krieger. Theory of plates and shells. McGraw-hill, 1959. [6] Joel Quintana, Virgilio Gonzalez, "Detection and calculation of reflected spectral shifts in fiber-Bragg gratings (FBG) in polarization maintaining optical fiber", Proceedings of SPIE Vol. 9061, 90614D (2014).

P.18

SETS2016-06 The Southwest Emerging Technology Symposium 2016

DAEDALUS: LOX/LCH4 SUBORBITAL TESTBED DESIGN 1

J.Adams1, J. Trillo1, A. Johnson1, A. Choudhuri1* Mechanical Engineering, University of Texas at El Paso, El Paso, TX 79968, USA; * Corresponding author ([email protected])

Keywords: Daedalus, Suborbital, Vehicle, Methane, Propulsion ABSTRACT LOX/LCH4 propulsion technology has had limited application in real world scenarios. However, due to it wide spread availability in the solar system this propellant combination will allow for greater exploration of space. DAEDALUS is a suborbital vehicle testbed that will be used in order to assess a propulsion system that uses only LOX/LCH4. This vehicle will be divided into three prototypes each assessing different portion of the spacecraft ultimately culminating in a full flight vehicle to be launched from Wallops Flight Test Facility, VA. During this flight the vehicle will follow a predetermined flight profile and transmit data back to a ground control unit for further analysis. From these tests the viability of purely LOX/LCH4 propulsion systems will be determined for future work in propulsion technology. 1

Project Background

1.1 Importance of LOX/LCH4 systems A major limiting factor of deep space exploration is the amount of propellant needed in order to complete a mission. This is due to the need to carry the propellant from Earth throughout the duration of the mission. If propellant can be extracted from the environment of space bodies, it would allow for longer missions and smaller spacecraft which would lower the cost of missions and make space travel more feasible. As of now methane has been found on numerous planets, asteroids, and other bodies spanning from Mars to Uranus1,2, these bodies could be used in order to provide the propellants Liquid Oxygen (LOX) / Liquid Methane (LCH4). Continuing research in this area is vital in order to ensure the availability of the technology needed to utilize the naturally occurring methane as a fuel source for interplanetary travel. 1.2 Mission Overview DAEDALUS is a suborbital testbed to assess the viability of LOX/LCH4 in vacuum and low gravity. This is a key step in the advancement of LOX/LCH4 technology, since these propulsion systems have had limited use in these conditions3. In order to show this propellant combination viability, DAEDALUS will be launched from Earth as the payload of a Terrier Mk 12 sounding rocket. Upon release from the sounding rocket DAEDALUS will conduct a flight maneuver involving a series of CROME 500 pound main engine (ME) starts and PENCIL reaction control engine (RCE) firings to demonstrate the restartability of the engines and its effect on the propulsion system as a whole, as well as the ability to use LOX/LCH4 to control the craft as it moves through the flight maneuver. The project will also allow for the

P.19

DAEDALUS

testing of the technology used in order to create the engines which will likely include additive manufactured components.

2

Project Requirements

2.1 Flight Profile In order to assess DADALUS’ LOX/LCH4 technology in a low gravity & atmosphere conditions the vehicle will conduct a flight maneuver as seen in Figure 1. DAEDALUS is planned to be launched from a Terrier Mk 12 sounding rocket over the Atlantic Ocean from Wallops Flight Test Facility. The Terrier will take the DAEDALUS system to an altitude of 90 miles above sea level. At this point the system will be released from the Terrier, the data acquisition will activate and the system will begin its flight maneuver. In preparation for the flight maneuver the Reaction Control System (RCS) will fire the RCEs in order to orientate the system perpendicular to Earth’s surface. Once the vehicle is orientated the main engine and RCS will go through a series of five engine starts and reorientations. Each firing will be at a different thrust level so as to demonstrate the engines ability to throttle and show that the engines can be started and restarted in a vacuum. Once the flight maneuver has been completed DAEDALUS will then reenter the atmosphere and deploy parachutes allowing the vehicle to splash down into the ocean to be recovered for analysis.

Figure 1. Flight profile of DAEDALUS.

2.2 General Vehicle Dimensions

2 P.20

DAEDALUS

The Terrier sounding rocket has a maximum allowable weight, diameter and length. The vehicle must be contained within a diameter of 15.75 inches and an overall length of 200 inches as well as maximum weight of 500 pounds4. Due to these constraints the vehicle’s subsystems are restricted to a stacked orientation as seen in Figure 2. The system most affected by these constraints are the propellant delivery systems and tanks. Although ideally the tanks would be spherical, to fit in the envelope the tanks must be cylindrical. Included in the system is a high pressure helium tank that works as a regulated pressure source to the thrusters, thus providing constant propellant pressure at injection.

Figure 2. DAEDALUS vehicle in envelope required by Terrier Mk 12.

2.3 Operating Conditions and Operating Criteria In order for the system to operate most efficiently the engines require propellant delivered to the injector at specific conditions. The LOX and LCH4 must be in the liquid state and a pressure 350 psia. This requirement is derived from the needs of the engines which require a minimum of 235 psia for the main engine and 250 psia for the RCS. With these requirements met the main engine is capable of throttling from 125 pounds to 500 pounds which will be used as the baseline for the flight maneuver. The requirements on the fluid also apply additional requirements to the tanks. The tanks must be able to withstand the fluid temperatures (on the range if -120°F), the pressure inside the tank (approximately 350 psia), and ensure that the propellant remains in both liquid state and near the outlets of the tanks to avoid interruption of the flow. 3 Prototype Approach 3.1 Nature and Purpose of the Prototypes DAEDALUS will be broken into three prototypes D-1, D-2, and D-3, as seen in Figure 3. Each will build upon the last, culminating in the D-3 which is the flight article. D-1 is a static test bed allowing for the testing of the engines and propellant delivery systems. After

3 P.21

DAEDALUS

completing D-1, D-2 will integrate the flight hardware and control mechanisms, such as the electronics and autonomous control systems necessary to orientate the vehicle. The testing of D-2 will also allow for analysis of all other flight hardware, such as the tanks and data acquisition system to ensure all necessary systems are operational and able to transmit to a ground control system. Once this test is complete the hardware will then be incorporated into the D-3 flight vehicle to be launched from Wallops Flight Test Facility.

Figure 3. Prototype breakdown and descriptions

3.2 Ongoing work for D-1 Prototype D-1 will measure thrust coming from each engine using three load cells on each engine, as well as the pressure drop through the lines in the setup. These features have been included in the Piping and Instrumentation Diagram (Figure 4) which has been the focus the research up to this point. Also included in this is a partial equipment list (Figure 5) which will be used in order to ensure all equipment is correctly purchased and installed in the test setup. After these are completed the next step will be to design the static test bed that the D-1 prototype will operate on.

4 P.22

DAEDALUS

Figure 4. Piping and instrumentation Diagram for D-1. Figure 5. Partial Equipment list for D-1.

5 P.23

DAEDALUS

References [1] V. Formisano, et all. “Detection of Methane in the Atmosphere of Mars”, Science 03 Dec 2004: Vol. 306, Issue 5702, pp. 1758-1761 DOI: 10.1126/science.1101732 [2] B. Lutz, et all. “Laboratory Band Strengths of Methane and their Application to the Atmospheres of Jupiter, Saturn, Uranus, Neptune, and Titan” The Astronomical Journal 15 Jan 1976, Vol. 203 pp. 541-551 [3] M.Klem et all. “Liquid Oxygen/Liquid Methane Propulsion and Cryogenic Advanced Development” IAC-11-C4.1.5 [4] NASA Goddard Space Flight Center Wallops Flight Facility, “NASA Sounding Rockets User Handbook” July 2015. [5] D. Huzel and D. Huang. “Modern Engineering for Design of Liquid-Propellant Rocket Engines” Vol. 147, 1992

6 P.24

SETS2016-07 The Southwest Emerging Technology Symposium 2016

TRAJECTORY SIMULATION OF A TERRIER IMPROVED ORION LAUNCH VEHICLE 1

D. Camacho, 2J. Holt, 3M. Everett, 4Dr.Angel Flores Abad NASA MIRO Center for Space Exploration and Technology Research (MIRO cSETR) Department of Mechanical Engineering University of Texas at El Paso El Paso, Texas, 79968-0521 [email protected] Keywords: Trajectory, Simulation, Daedalus, Matlab, STK ABSTRACT Before a rocket is launched to a suborbital flight, it is desired to have a mathematical model and simulation environment to aid the analysis and prediction of the actual performance. This paper presents the determination and simulating of the ballistic trajectory of the Daedalus upper stage as an introduction to an ongoing body of work. Daedalus will be launched as the third stage to the Terrier Improved Orion Launch Vehicle in a suborbital flight. The main objective of Daedalus is to both validate research and to test thrusters developed within the Center for Space Exploration and Technology Research (cSETR) at the University of Texas at El Paso, within their actual design environment. The kinematics and dynamics of the vehicle are derived to obtain the trajectory considering the documented performance of the Terrier Improved Orion Launch Vehicle. The resulting equations are firstly introduced in Matlab to obtain numerical insights of the trajectory and secondly in AGI Software to have a more real visualization of the launching location, altitude, impact range and so on. The simulation trajectory is validated by comparison with another model previously developed by NASA using the same flight proven Terrier-Improved Launching vehicle. 1 Introduction 1.1 Project Background Daedalus is a suborbital upper stage rocket that will be used to evaluate thrusters developed by the cSETR. It will be launched as the third stage to a Terrier-Orion sounding rocket. For the spacecraft to be successful, the Guidance, Navigation and Control (GNC) system must be able to track Daedalus’ location, as well as communicate with Daedalus and control Daedalus’ maneuvers, such as when the thrusters will fire. During the mission, Daedalus will test: pulse firing of the Pencil reaction control engines (RCEs), pulse firing of the CROME 500 pound main engine (ME), and the ability to precisely control the spacecraft’s attitude and motions. 2 Mission Approach 2.1 Approach and Flight Profile Before maneuvers for Daedalus can be planned, the general trajectory—including time of flight—must be known. Simulations of the trajectory of Daedalus will be achieved through the use of MATLAB as well as Systems Tool Kit (STK), which is a physics-based software package that allows complex analysis to be performed on air and space systems. NASA’s Mission 1

Graduate Research Student, Mechanical Engineering, The University of Texas at El Paso, 79968
 Research Student, Mechanical Engineering, The University of Texas at El Paso, 79968
 3 Research Engineer, Center for Space Exploration Technology research (cSETR), The University of Texas at El Paso, 79968
 4 Corresponding Author, Research Assistant Professor, Mechanical Engineering, The University of Texas at El Paso, 79968 P.25 2 Undergraduate

The Southwest Emerging Technology Symposium 2016 Planning Lab at Wallops Flight Facility has been using STK to simulate the trajectory, performance, and communications systems of the Terrier-Orion sounding rocket, as well as other rockets and vehicles. Figure 1 below shows a simulation conducted by NASA depicting the terrier-Orion mission [4].

Figure 1: Simulated trajectory of the Terrier-Orion produced by NASA’s Mission Planning Lab. Illustrated in red first stage burn, second stage burn in blue, yellow are the sections in which it will coast.

For Daedalus’ mission, the trajectory will be similar: the first stage, a Terrier Mk 12 booster, will fire for approximately 5.2 seconds. The stage will then separate, with the rocket coasting for 9.8 seconds before the second stage, an Improved Orion booster, ignites. This will fire for approximately 25.4 seconds, with the rocket burning out at an altitude of approximately 30 to 40 km. Daedalus will then separate and begin its mission, firing its RCS thrusters to bring Daedalus to the correct attitude before the apogee of 135 km is reached. Daedalus will perform 5 firings at different thrust levels of the 500 lbs main engine and RCS engines to demonstrate the engines’ ability to throttle, as well as demonstrate that they can be started and restarted in a vacuum. After these maneuvers have been completed, Daedalus will fall back to Earth and re-enter the atmosphere. 2.2 Simulating a Simple Trajectory with MATLAB and STK From Terrier-Orion performance tables provided by NASA (Figure: 2), payload weight and launch angle can be used to estimate an apogee altitude and impact.

P.26

The Southwest Emerging Technology Symposium 2016

Figure 2: Terrier MK12-Improved Orion Predicted Vehicle Performance

A MATLAB program was designed to take user input of payload weight, launch angle, launch location (latitude, longitude, altitude), and launch azimuth. Initially, the shape of the parabolic trajectory was calculated using the kinematics equations for constant acceleration. Below, equation 1, is used to describe the motion of an object. Where t is the time, x is final position, 𝑥0 initial position, 𝑣0 initial velocity, and 𝑎𝑥 the acceleration [2]. This makes assumptions such as the force of gravity doesn’t change with altitude, that there is no air resistance, and ignores the portion of the trajectory at launch while the engines are burning. 1

𝑥 − 𝑥0 = 𝑣0 𝑡 + 2 𝑎𝑥 𝑡 2

(1)

To calculate the latitude and longitude for each data point of the trajectory, great circle equations were used, which estimate the earth as a perfect sphere. These equations take an input of initial latitude and longitude, heading (launch azimuth), and distance traveled, and give an output of ending latitude and longitude. In equation 2 the variable represents 𝜑2 is the final latitude, 𝜑1 initial latitude, 𝜃 is the azimuth angle, 𝛿 distance traveled, and lastly in equation 3 𝜆1 is initial longitude and 𝜆2 final longitude [3]. Because the earth is not really spherical in shape, great circle equations do give some error, although it is typically small, around 0.3%. 𝜑2 = asin(𝑠𝑖𝑛𝜑1 ∙ 𝑐𝑜𝑠𝛿 + 𝑐𝑜𝑠𝜑1 ∙ 𝑠𝑖𝑛𝛿 ∙ 𝑐𝑜𝑠𝜃)

(2)

𝜆2 = 𝜆1 + 𝑎𝑡𝑎𝑛2(𝑠𝑖𝑛𝜃 ∙ 𝑠𝑖𝑛𝛿 ∙ 𝑐𝑜𝑠𝜑1 , 𝑐𝑜𝑠𝛿 − 𝑠𝑖𝑛𝜑1 ∙ 𝑠𝑖𝑛𝜑2 )

(3)

Using this information, an Ephemeris file is generated which can be opened in STK. The Ephemeris format consists of a table of time, latitude, longitude, and altitude points. STK then interpolates the data points to generate a trajectory, which is plotted on the Earth.

P.27

The Southwest Emerging Technology Symposium 2016 To provide a more accurate simulation, it was decided to take into account the change in the force of gravity with altitude. For this task, kinematic equations could not be used in the vertical direction, so a new equation was derived by using the equation for the force of gravity, and Newton’s law (F=m𝑥̈ ). This equation (5) gives the time for an object to fall a set distance. The full derivation of this formula can be read in the references (4) [1]. From (4) it is stated that the force F between two bodies is equal to the product of the gravitational constant G, the mass of the first object 𝑚1 , and the mass of the second object 𝑚2 all divided by the distance between the center of one body to the center of the second body r. As the second body falls to the first body, the acceleration (𝑥̈ ) of 𝑚2 is negative relative to𝑚1 . Then in (5) t is the time, 𝑟𝑖 is the initial distance of separation, 𝑟𝑓 final center-to-center separation distance [1]. 𝐹= 𝑡=

𝐺𝑚1 𝑚2 𝑟2

1 √2𝐺(𝑚1 +𝑚2

(4)

= 𝑚1 𝑥̈ 1 = −𝑚2 𝑥̈ 2 3/2

(√𝑟𝑖 𝑟𝑓 (𝑟𝑖 − 𝑟𝑓 ) + 𝑟𝑖 )

𝑟

𝑐𝑜𝑠 −1 √ 𝑟𝑓 ) 𝑖

(5)

Using the new formula, the shape of the trajectory did not change, however the timing of the trajectory changed slightly, with time above 100 km increasing from 1 s (750 lbs payload, 85 degree launch angle) to 12 s (200 lbs payload, 85 degree launch angle)—the higher the altitude achieved, the larger the change in flight time. Figure 3 shows the trajectories plotted in STK, with the red spacecraft having a constant acceleration from gravity, while the yellow spacecraft has gravity change with altitude [4].

Figure 3: Simulation obtained from STK software representing the trajectory difference when gravity is varied with altitude (yellow) and when it is constant (red)

From STK, tables can then be generated of simulation time, latitude, longitude, altitude, and velocity in the x, y, and z directions.

P.28

The Southwest Emerging Technology Symposium 2016 3

Future Work

In the future, a more in-depth simulation is to be created utilizing extended features of STK, such as STK Pro, and STK’s Astrogator plugin. Once more performance specifications of the Terrier Mk 12 and Improved Orion boosters are known, a more accurate trajectory can be generated. Further, these extended features will allow Daedalus’ maneuvers to be simulated, such as attitude and change in trajectory. STK will also be able to calculate line-of-sight between Daedalus, the launch site, and radar and communications stations. 4

Conclusion

The simulations for the trajectory of the Terroir Improved Orion created on Matlab and STK have demonstrated a trajectory path which can be structured by values such as payload, launch angle, launch location, as well as launch azimuth. The simulations allow there to be a reliable reference for a future launch with a predicted flight path and location once decent of space vehicle indicates. 5

References

[1]

ErikE (http://physics.stackexchange.com/users/1483/erike), Don't heavier objects actually fall faster because they exert their own gravity?, URL (version: 2015-05-25): http://physics.stackexchange.com/q/3534 (Retrieved April 31, 2016) [2]

Young, H and Freedman, R, “University Physics with Modern Physics,” 13th ed., San Francisco, 2012 [3]

Veness, C. (2010 January).Calculate distance,bearing and more between Latitude/Longitude points. http://www.movable-type.co.uk/scripts/latlong.html (Retrieved April 31, 2016) [4]

Wallops Flight Facility Goddard Space Flight Center .Mission Planning Lab: Capability Catalog. http://sites.wff.nasa.gov/mpl/w_terrierorion.html (Retrieved April 31, 2016) [5]

Wallops Flight Facility Sounding Rockets Program Office. Terrier-Improved Orion. Retrieved from http://sites.wff.nasa.gov/code810/vehicles.html. (Retrieved March 31, 2016).

P.29

SETS2016-08 The Southwest Emerging Technology Symposium 2016

DYNAMIC MODELLING OF A FREE-FLOATING SPACE MANIPULATOR 1

J. Yepez1, J. De la Torre Rayas1, A. Flores-Abad1* Mechanical Engineering Department, University of Texas at El Paso, El Paso, TX 79968, USA; * Corresponding author ([email protected])

Keywords: space robot, dynamics, modelling. ABSTRACT Space manipulators are expected to play an important role for satellite servicing and space debris removal. In order to advance in this topic, the dynamic modelling of this type of robots becomes of paramount importance to analyze several aspects of the kinematics and dynamics, and to design and validate control algorithms. As oppose to the robots that operate on ground, space manipulators have a free-floating base, which causes the base satellite to move in response to the manipulators motions. This fact complicates the kinematics and dynamics modelling. In this paper the dynamic modelling of a space robot is presented. The Equations are obtained using conservation of momentum and energy. The obtained model is validated by means of a comparison with Simulink/SimMechanics.

1

Introduction

Space manipulators have been successfully used for many applications such as maneuvering astronauts, berthing and deploying large space structures, constructing and maintaining the International Space Station (ISS), exploring and sample-collecting, satellite on-orbit servicing (technology demos only), etc. All of these manipulation activities dealt with cooperative payloads or target objects and thus, the existing robotics technologies can handle them quite well, though many improvements such as the operational efficiency and dexterity may still be done. However, if a manipulator is expected to perform more challenging and riskier tasks, such as to capture an unknown object, like a piece of space debris, or a non-cooperative object, such as a tumbling satellite, the currently available space robotics technologies are still far from being ready. To make these challenging tasks practical, many enabling technologies have to be further advanced. An important stage of this technology development consists on having the dynamic model of the space robot. Several Universities and Research Centers have developed their dynamics models to simulate and analyze algorithms for space manipulators, i.e [1-3]. In this work, the dynamic model of a space manipulator is developed as an early stage to develop and test control algorithms. 2

Dynamic modelling of the space manipulator

2.1. Kinematics

It is assumed that the robotic system of the servicing spacecraft including both the spacecraft and the manipulator consists of n+1 rigid bodies connected by n single-DOF

P.30

(degree of freedom) joints, as shown in Fig. 1. Body 0 (or Link 0) is the servicing spacecraft which is also the base of the robot. Body i is the ith link of the manipulator. Joint 0 of the system has 6 DOF, which connects the inertial frame to the servicing spacecraft, and joint i is assumed to have only one DOF which articulates links i-1 and i. Unless otherwise specified, all the vectors are assumed to be expressed in the inertial frame F0 which is originated at the mass center C0 of the spacecraft. Moreover, F1 is a frame attached to the spacecraft with its origin at Joint 1 and Fe is the end-effector frame rigidly attached to the end-effector. In the introduced methodology the end-effector position and velocity must be known in order to determine if the capturing point has been reached. The position of the end-effector is defined as n

re

r0 a0

, i 1,2, , n

ai

(1)

i 1

where r0 R3 and re R3 are the positions of the mass center C0 of the spacecraft and the position of the end-effector of the manipulator with respect to the inertial frame F0.; ai R3 is the vector connecting the two connecting joints of the ith link of the manipulator. Link n-1

Link n

Fn

FE

an

an-1

Fn-1

On-1

Link 2

a2 F2 Link 1

Servicing System CM

FI a1 Fc Base spacecraft CM

F1

F0

a0

Base Spacecraft Link 0

C0

Fig. 1 Multibody dynamic system of a servicing spacecraft with an onboard manipulator.

By measuring the attitude and linear displacement of the base spacecraft, plus the robot’s joint angles, the end effector position re can be known. Differentiating (1) with respect to time, the fallowing relationship between the end-effector linear velocity and joint velocity is obtained ve

v 0 ω0

re r0

n

zi

re ρi

i

(2)

i 1

where v 0 R3 and ω0 R3 are the linear and angular velocity vectors of the spacecraft, respectively; v e R3 is the linear velocity vector of the end-effector of the manipulator; z i R3 is the unit vector of the ith joint axis; ρi R3 is the position vector as defined in Fig.1; and i is the joint angle of the ith joint. Besides, the end-effector angular velocity and the joint velocities are related by

P.31

DYNAMIC MODELLING OF A FREE-FLOATING SPACE MANIPULATOR

n

ωe

ω0

zi

(3)

i

i 1

2.2 Dynamics Assuming that there are no external forces acting on the servicing system, the momentum of the system will be conserved. The dynamics equations of the space robot including the servicing spacecraft can be derived in terms of its joint variables as follows [4]: Hθ Cθ τ, (4) n where θ R is the generalized joint coordinates and 1 T θ Hθ θ 2

Cθ

(5)

Rn ,

By following the method described in [4] we obtained the following components for the H and C matrices: M 11

m0

M 12

0

m1

m2

M 13

a1 * m2 * sin(

M 14

a1 * m2 * sin(

M 15

(a2 / 2) * m2 * sin(

M 21

) (a1 / 2) * m1 * sin(

0

1

0

(a1 / 2) * m1 * sin( 1) 0

1

2

) (a0 * m1 * sin( 0 )) / 2 ( a0 * m2 * sin( 0 )) / 2 (a2 / 2) * m2 * sin(

0

1

0

(a2 / 2) * m2 * sin( 1)

0

0

1

1

2

2

)

2)

1

)

0

M 22

m0

M 23

a1 * m2 * cos(

m1

m2

M 24

a1 * m2 * cos(

M 25 M 31

(a2 / 2) * m2 * cos( (a2 / 2) * m2 *sin( 0

) (a1 / 2) * m1 * cos(

0

1

0

(a1 / 2) * m1 * cos( 1) ) 2 ) a1 * m2 *sin(

0

1

1

) (a0 * m1 * cos( 0 )) / 2 ( a0 * m2 * cos( 0 )) / 2 ( a2 / 2) * m2 * cos(

0

1

0

(a2 / 2) * m2 * cos( 1)

0

1

0

)

2)

2

0

1

) (a0 * m2 *sin( 0 )) / 2 ( a0 * m1 *sin( 0 )) / 2 (a1 / 2) * m1 *sin(

0

1

)

M 32 M 23 M 33 I 0 I1 I 2 (a0 2 * m1 ) / 4 (a0 2 * m2 ) / 4 a12 * m2 (a1 / 2) 2 * m1 (a2 / 2) 2 * m2 a0 *(a2 / 2) * m2 *cos( 2 1

2

2

M 34

I1 I 2 a * m2 (a1 / 2) * m1 (a2 / 2) * m2 (a0 (a2 / 2) * m2 *cos(

M 35

I 2 (a2 / 2)2 * m2 (a0 (a2 / 2) * m2 *cos(

M 41

M 14

M 42

M 24

M 43

M 34 a12 * m2

(a1 / 2) 2 * m1

M 44

I1

I2

M 45

I2

(a2 / 2) 2 * m2

M 51

M 15

M 52

M 25

M 53

M 35

M 54

M 45

M 55

I2

C11

0

C12

0

C13

(a1 / 2) * 1* m1 cos(

2

1

2

2

) a0 * a1 * m2 *cos( 1 ) a0 *(a1 / 2) * m1 *cos( 1 ) 2a1 *(a2 / 2) * m2 *cos( 2 )

)) / 2 (a0 * a1 * m2 *cos( 1 )) / 2 ( a0 * (a1 / 2) * m1 *cos( 1 )) / 2 2a1 *(a2 / 2) * m2 *cos( 2 )

)) / 2 a1 *(a2 / 2) * m2 *cos( 2 )

(a2 / 2) 2 * m2

2a1 * (a2 / 2) * m2 * cos( 2 )

a1 * (a2 / 2) * m2 * cos( 2 )

(a2 / 2) 2 * m2

a1 * m2 * cos( C14

1

1

(a1 / 2) *

0

0

0

1

1

) a1 * 1* m2 *cos(

)

* m1 * cos(

( (a2 / 2) * m2 * cos(

0

1

) (a2 / 2) * 2 * m2 *cos(

(a1 / 2) * m1 * cos( 0 0

1

)

a1 * 1

0

2

0

* m2 * cos(

)

1

0

)

0

1

2

) (a2 / 2) * 1* m2 *cos(

(a0 * m1 * cos( 1

)

(a2 / 2) *

a1 * m2 * cos(

0

1

2

)

0

)) / 2

* m2 * cos(

0

0

1

2

) ( (a2 / 2) * m2 *cos(

(a0 * m2 * cos( 1

2

)

( a2 / 2) *

( a1 / 2) * m1 * cos(

0

1

0

0

0

1

)) / 2) *

* m2 * cos(

)) *

2

0

0

1

1

3 P.32

)

2

)

(a2 / 2) *

C15 C21

0

C22

0

C23

1

* m2 * cos(

(a1 / 2) * 1 * m1 *sin(

( (a2 / 2) * m2 *sin(

C24

0

0

1

1

(a1 / 2) * 0 * m1 *sin(

2

(a2 / 2) *

C25

C31

0

C32

0

C33

)

2

(a2 / 2) *

) a1 * 1* m2 *sin(

1

0

0

1

1

2

* m2 * sin(

0

)

* m2 * cos(

0

1

) (a2 / 2) * 2 * m2 *sin(

) (a1 / 2) * m1 *sin(

) a1 * 0 * m2 *sin(

0

1

0

0

1

) (a2 / 2) *

0

2

1

2

)

(a2 / 2) *

0

0

)

(a2 / 2) * m2 * cos(

1

2

0

a1 * (a2 / 2) *

0.5 * a0 * (a2 / 2) *

0

2

* m2 * sin(

* m2 * sin(

2

)

1

0.5 * a0 * (a1 / 2) * 2

)

0

0 1

* m2 * sin(

C41

0

C42

0

0 * m2 * sin(

1

* m1 * sin( 1 )

( (a0 * (a2 / 2) * m2 * sin(

a1 * (a2 / 2) * 1 * m2 *sin( 2 ) a1 * (a2 / 2) *

0.5 * a0 * (a2 / 2) *

0

2)

0

1

1

) (a2 / 2) * 1* m2 *sin(

)

2

0

1

2

2

* m2 * sin(

2

)

C44

a1 * (a2 / 2) *

2

* m2 * sin(

2

)

C45

a1 * (a2 / 2) *

0

* m2 * sin(

2

)

* m2 * sin(

2

)

((a0 * ( a2 / 2) * m2 * sin(

* m2 * sin(

2

)

a1 * ( a2 / 2) * m2 * sin(

0

C52

0

C53

a1 * (a2 / 2) *

1

C54

a1 * ( a2 / 2) *

0

C55

0

0

)

((a0 * (a2 / 2) * m2 * sin(

a1 * (a2 / 2) *

1

1

* m2 * sin(

2

2

2

)) / 2

)

( a0 * a1 * m2 * sin( 1 )) / 2

)*

2 )) / 2

1

0

1

2

1

2

2

2

)*

2

) 0

1

)) *

2

)

1

1

2

0

1

)*

) 0.5* a0 *(a2 / 2) * 1* m2 *sin(

0.5 * a0 * (a2 / 2) *

2

* m2 * sin(

1

(a0 * (a1 / 2) * m1 * sin( 1 )) / 2 *

2

)) *

2

1

2 2

1

2

(a0 * (a1 / 2) * m1 * sin( 1 )) / 2) *

2

a1 * ( a2 / 2) * m2 * sin(

2

)) *

0

1

3. Simulation Results To validate our equations, we modelled our system in Simmechanics (See Fig. 2) and compare the performance of the equations and the Simmechanics model.

P.33

2

)

( a0 * a1 * m2 * sin( 1 )) / 2

)) / 2

1

0

a1 * ( a2 / 2) * m2 * sin(

a1 * (a2 / 2) * m2 * sin(

1

0

)*

( a2 / 2) * m2 * sin(

* m2 * sin( 1 )

* m2 *sin( 2 ) 0.5* a0 * (a2 / 2) * 1* m2 *sin(

a1 * (a2 / 2) *

C51

)) / 2

( (a0 * (a2 / 2) * m2 * sin(

C43

2

) (a2 / 2) * 0 * m2 *sin(

a1 * m2 * sin(

0.5 * a0 * a1 *

2

1

) (a0 * m1 *sin( 0 )) / 2 (a0 * m2 *sin( 0 )) / 2) *

* m2 *sin(

(a1 / 2) * m1 * sin(

1

2

a1 *(a2 / 2) * 2 * m2 *sin( 2 ) 0.5* a0 *(a1 / 2) * 1* m1 *sin( 1 ) 0.5* a0 * a1 * 1* m2 *sin( 1 ) 0.5* a0 *(a2 / 2) * 2 * m2 *sin(

C34

C35

1

1

) a1 * m2 *sin(

0

( (a2 / 2) * m2 * sin(

0

0

)

)

DYNAMIC MODELLING OF A FREE-FLOATING SPACE MANIPULATOR

Fig. 2 The Simmechanis model

The performance of the equations and the Simmechanics model are the same, actually, overlapped. Figure 3 shows the torques input to the robot, as well as the resultant robot’s joint angles. While, Figure 4 depicts the motion of the base satellite as a result of the manipulator’s displacements.

a) b) Fig. 3. a) random joint torques and b) obtained joint angles of the robot

5 P.34

a) b) Fig. 4. a) linear and b) angular displacement of the base satellite. 4. Conclusions In this paper the dynamic model for a 2-DOF space robot was developed and the equations obtained were validated through a comparison with a simmechanics model. This dynamic model will be important in the analysis and control of space manipulators. The simulations show the linear and angular displacements of the base satellite as a result of robot’s motion. Future work includes experimental validation of the model, increasing of the DOF’s and also its application in the development of control algorithms.

References [1] N. Uyama, D. Hirano, H. Nakanishi, K. Nagaoka & K. Yoshida. “Impedance-based contact control of a free-flying space robot with respect to coefficient of restitution”. IEEE/SICE International Symposium on System Integration, pp. 1196-1201, 2011. [2] R. Lampariello, S. Abiko, S., & G. Hirzinger. “Dynamics modeling of structure-varying kinematic chains for free-flying robots,” IEEE International Conference on Robotics and Automation, pp. 1207-1212, 2008. [3] W. Xu, D. Meng, Y. Chen, H. Qian, & Y. Xu. “Dynamics modeling and analysis of a flexible-base space robot for capturing large flexible spacecraft.” Multibody System Dynamics, Vol. 32., No. 3, pp. 357-401, 2014. [4] Y. Xu and T. Kanade, Eds. Space robotics: Dynamics and Control, Kluwer Academic Publishers, 1993.

P.35

SETS2016-09 The Southwest Emerging Technology Symposium 2016

BIDDING IN WHOLESALE ENERGY MARKET 1

Moinul Morshed Porag Chowdhury1, Jesus Juarez1, Dr. Christopher Kiekintveld1* Computer Science Department, University of Texas at El Paso, El Paso, TX 79968, USA; *

[email protected]

Keywords: Smart Grid, Multi Agent System, Machine Learning, Periodic Double Auction Market, Price Prediction.

ABSTRACT The Power TAC simulation emphasizes the strategic problems that broker agents face in managing the economics of a smart grid. The brokers must make trades in multiple markets and to be successful, brokers must make many good predictions about future supply, demand, and prices in the wholesale and tariff markets. In this paper, we investigate the feasibility of using learning strategies to improve the performance of our SPOT agent. Specifically, we investigate the use of decision trees and neural networks to predict the clearing price in the wholesale market and the use of the existing best price prediction to bid in the wholesale market. Preliminary experimental results show that the strategy holds promise, though a more thorough investigation is necessary for SPOT to become more competitive. 1

Introduction

The traditional energy grid lacks several important features such as effective use of pricing and demand response of energy, customer participation, and proper distribution management for variable-output renewable energy sources etc [1]. The smart grid has the potential to address many of these issues by providing a more intelligent energy infrastructure [2]. Researchers rely on rich simulations such as the Power Trading Agent Competition (Power TAC) [1] to explore the characteristics of future smart grids. In the Power TAC smart grid simulation, brokers participate in several markets including the wholesale market, the tariff market, and the load balancing market to purchase energy and sell it to customers. This game was designed as a scenario for the annual Trading Agent Competition, a research competition with over a decade of history [3]. The wholesale market attempts to simulate existing energy markets such as the European or North American wholesale energy markets [4]. It is a “day ahead market” where the energy is a perishable good and it allows brokers to buy and sell quantities of energy for future delivery. Market structures like this exist across many different types of perishable goods, so finding effective, robust, automated bidding strategies for these markets is an important research challenge. 2

Background: Power TAC Wholesale Market

The wholesale market functions as a short-term spot market for buying and selling energy commitments in specific timeslots, where each timeslot represents a simulated hour. At any point in the simulation, agents can participate in auctions to trade energy for the next 24 hours, so there are always 24 active auctions. The simulation clears the bids by matching buy and sell orders, and determines the clearing price for each auction every day. If the min ask

P.36

price has a higher value than the max bid price, the market does not clear. The main problem we consider here is learning to predict the clearing prices of these auctions, which can be used by the agent to implement an effective bidding strategy. 2.1 Learning in the Wholesale Market We have experimented with three different machine learning methods to predict clearing prices in the wholesale market: 1) REPTree (decision tree)[5] 2) Linear Regression and 3) Multilayer Perceptron (a type of neural network). We have also investigated a variety of different features for training the predictors. These include 8 price features that capture information about the recent trading history, such as the clearing price for the previous hour and the prices for the equivalent time slot in the previous day and week. We also include the weather forecast and time of day because the energy production of renewable energy producers (e.g., solar) depends on these factors. The number of participants in the game is included because the amount of competition affects the market clearing price. Finally, we include a moving average of the prices as a convenient way to capture an aggregate price history; this is the predictor used in the baseline version of the SPOT agent. We use the Sample broker to extract training data from the simulation. To generate training data we use simulations with a variety of agent binaries from previous tournaments, as well as a variety of different bootstrap initialization files. We train our models using, and evaluate their ability to predict market clearing prices based on the mean absolute prediction error only for auctions that clear (we do not include auctions that do not clear in the error calculations). 2.2 Prediction Accuracy Comparisons One of the most significant factors we discovered that influences the accuracy of the models is how we handle auctions that do not clear. In many cases, an auction will have no clearing price due to a spread between the bid and ask prices, which results in the simulation returning null values for these prices. This causes significant problems with the price features we use, as well as the final error calculations. To improve this we calculate an estimated clearing price for auctions that do not clear by taking the average of lowest ask price and the highest bid price.

Fig.1. Effect of clearing price estimation.

Fig.2. Prediction models error rates by no. of games.

Figure 1 shows the prediction errors during the course of a single simulation for two different REPTree models trained on 20 games, one with estimated clearing prices and the

P.37

BIDDING IN WHOLESALE ENERGY MARKET

other without. We also include the errors for a simple moving average price predictor as a baseline for comparison. Each data point shows the average error for all auctions in a window of five timeslots. The data show that both REPTree models outperform the moving average predictor, but the version with estimated clearing prices is dramatically better, and produces much more consistent predictions throughout the entire game. Figure 2 shows decision tree slowly improves according to the number of games where other models such as linear regression, default multilayer perceptron do not show this trend and estimated clearing price models are much better than models without estimated prices. 2.3 Bidding in the Periodic Double Auction Market We have used REPTree [5] with est. price predictor in our new SPOT agent to bid in the wholesale market. Our bidding strategy is to find out the auctions where the clearing price is going to be low than a threshold margin and then try to buy higher volume of needed energy from that specific hour ahead auction. Figure 3 shows our new SPOT agent is able to buy high volume of its needed energy when the average clearing price is lowest.

Fig.3. Comparison of Energy Volume with Clearing Price.

3. Conclusion The preliminary results in this paper show that the application of learning strategies to broker agent within Power TAC has immediate benefits in the wholesale market. However, a more comprehensive study is needed to better harness the strength of these learning approaches. Therefore, future work includes learning good bidding strategies in the wholesale market by taking into account the predicted clearing prices as well as empirically evaluating the coupling effects of the learning strategies between the wholesale and tariff markets. References [1] Ketter, W., Peters, M., & Collins, J. (2013, November). Autonomous agents in future energy markets: the 2012 power trading agent competition. InBNAIC 2013: Proceedings of the 25th Benelux Conference on Artificial Intelligence, Delft, The Netherlands, November 7-8, 2013. Delft University of Technology (TU Delft); under the auspices of the Benelux Association for Artificial Intelligence (BNVKI) and the Dutch Research School for Information and Knowledge Systems (SIKS). [2] Ketter, W., Collins, J., & Reddy, P. (2013). Power TAC: A competitive economic simulation of the smart grid. Energy Economics, 39, 262-270. [3] Wellman, M. P., Greenwald, A., Stone, P., & Wurman, P. R. (2003). The 2001 trading agent competition. Electronic Markets, 13(1), 4-12. [4] Ketter, W., Collins, J., Reddy, P. P., Flath, C. M., & de Weerdt, M. (2012). The Power Trading Agent Competition. SSRN Working Paper Series. [5] Elomaa, T., & Kaariainen, M. (2001). An analysis of reduced error pruning.Journal of Artificial Intelligence Research.

3 P.38

SETS2016-10 The Southwest Emerging Technology Symposium 2016 USE OF LIFE CYCLE SUSTAINABILITY ASSESSMENT FOR ENERGY APPLICATIONS 1

B.A. Benedict1 Mechanical Engineering, University of Texas at El Paso, El Paso, TX 79968, USA; * Corresponding author ([email protected])

Keywords: sustainability, life cycle assessment, energy

ABSTRACT Pragmatic and reliable methods for assessing sustainability remain difficult for many organizations. Further, understanding the three elements of sustainability over the full life cycle of products and processes is essential. In some cases, understanding environmental issues is the easiest area. However, economic and social issues are less well understood. Life Cycle Sustainability Analysis is a framework for reviewing all three areas and enabling not only full coverage but understanding balancing and interactions between the elements. This paper reviews the three elements of sustainability, life cycle assessment, and analysis and evaluation of the three elements over a life cycle. These frameworks will be described so as to facilitate development of ways to help decision makers present proposals and illustrate results. Means are presented to enable illustration of findings to both expert and non-expert audiences. Specifically, uses of the life cycle sustainability triangle and the life cycle sustainability dashboard will presented. Examples will be presented for a comparison of solar PV panels. The paper is intended to help practitioners better understand linkages between the three elements of sustainability and ways to analyze them. Introduction Lifecycle sustainability assessment [1] is a relatively new concept, although it is clearly informed by previous efforts. The issue is to try to understand the behavior of environmental issues, cost, and social issues over the life of a project, product [2], [3], or policy. Key elements of each of these will be described. Definitions and approaches Sustainability and its three elements - environment, economics, and social aspects- will be reviewed, as well as the basic factors and approaches of life cycle assessment. Means of evaluating life cycle environmental issues, life cycle costs, and social life cycle effects will be noted. A single snapshot of what is happening at a point in time is not likely to be adequate for describing sustainability, nor is anything that ignores one of the three pillars of focus. Therefore, integration of these three factors is critical to full understanding. Since there is the probability of uncertainty over the life cycle being considered, there should be ways to consider these uncertainties. One powerful tool which may have value is scenario planning.

P.39

Discussion of life cycle assessment will begin by drawing boundaries on the analysis and illustrate the comprehensive evaluation of any and all changes due to the proposed changes. Caution is needed when reviewing life cycle analyses by others to assure that they have included everything. For example, a few years ago, a study by a trade group for the ethanol industry omitted the need to develop new corn fields to replace those whose production was taken for ethanol production. Suggestions will be made for types of items that may be considered for the three sustainability elements. Since social issues are less well understood, a list of such possible considerations will be presented. Further, it will be emphasized that each case requires a separate analysis. This discussion indicates that another use of this analysis is the comparison of alternatives to seek optimum conditions. It also allows for changing estimates of uncertainty to attempt to "bracket" likely performance. The issue of uncertainty is especially evident today as seen in global pricing of oil and its impact.

Displaying Results It is important to be able to sort through the work and data to provide results in such a way that decision-makers can see more quickly where they need to focus attention. The two types of such displays presented herein are the triangle view and the dashboard. The triangle view lays out zero to one hundred percent for each of the three sustainability elements along the three legs of the triangle. This enables consideration of the effect of "weighting" the three elements. Of course, many decisions have previously been made on the basis of only one or two of the elements. Further, it can be seen that some robust solutions may encompass larger regions of the triangle. The dashboard view uses a commonly employed dashboard approach, coupled with color coding to indicate results from very good (typically bright green) to very bad (typically bright red), with shadings between the extremes. This enables a quick view of the overall performance, with alternatives easily compared. Also, results within each of the three elements - environmental life cycle impacts, life cycle costs, and social life cycle assessments - can be seen to help better understand the contributions to overall performance and identify points of redesign and potential investment of resources.

Examples of use within energy sector Examples of areas in which approaches have been demonstrated in the energy sector include the following:

P.40

Solar PV panels Alternative passenger vehicles Hydrogen production from Biomass Building materials Biofuel supply chains The analysis of PV panels [4] will be included as a means of illustrating the process and results, as well as examples of ways to visualize the results. Thoughts for future use Given the centrality of sustainability concerns to existing and future energy development, distribution, and use, it seems logical to see increasing use of this tool. In fact, for the development of robust plans and strategies, such approaches are essential. References [1] M. Finkbeiner, E.M. Schau, A. Lehmann, and M. Traverso, Toward Life Cycle Sustainability Analysis, Sustainability, 2010, 2, 3309-3322, accessed through [2] UNEP, Guidelines for Social Life Cycle Assessment of Products, 2009, accessed through http://www.unep.org/pdf/DTIE_PDFS/DTIx1164xPA-guidelines_sLCA.pdf [3] UNEP, Towards a Life Cycle Sustainability Assessment for Products, 2012, 86 pp The International Journal of Life Cycle Assessment, September 2012, Volume 17, Issue 8, pp 1068-1079 [4] M. Traverso, F. Asdrubali, A. Francis, and M. Finkbeiner, Towards life cycle sustainability assessment: an implementation to photovoltaic modules,

3 P.41

SETS2016-11 The Southwest Emerging Technology Symposium 2016

TRANSACTIVE ENERGY SYSTEMS Eric Galvan and Paras Mandal* Power & Renewable Energy Systems (PRES) Lab, Department of Electrical and Computer Engineering, University of Texas at El Paso, El Paso, TX 79968, USA * Corresponding author ([email protected])

Keywords: Distributed generation, prosumer, smart grid, transactive control. ABSTRACT This paper presents a new approach of transactive energy (TE) for a smart grid community constituted by prosumer community groups (PCG) and consumer groups. To validate the approach, it is tested on a 6-bus system. Our results indicate that the transactive control mechanism can perform adequately in maximizing the profits and minimizing the electricity bill for PCGs. Furthermore, peak load can be reduced with significant increment in asset utilization by utility. 1 Introduction The purpose of the smart grid community (SGC) is to improve the power system efficiency, reliability, resiliency, security, and sustainability. A new approach that is being considered to facilitate the interoperability of smart grid devices and systems is the “Prosumer” architecture. Creating an economically motivated entity, prosumer architecture, allows the consumer to (i) produce, and store power, (ii) own or operate a small power grid, (iii) transport electricity by using electric vehicles (EV), and (iv) make economically optimal decisions regarding its energy utilization [1]. However, to fulfill the potential of the prosumer concept, several requirements need to be addressed, e.g., comprehensive framework for interaction among distributed entities within the grid, a systematic design, and robust and efficient control/optimization algorithms [1]. Prosumer community group (PCG) has emerged as a virtual aggregation of prosumers that might share or be from different locations, but with similar energy behaviors. One of the key benefits of PCG approach is that consumers can directly buy the energy from surrounding PCG, which will deliver the power they commit to, avoiding interrupting the main utility grid. This approach minimizes energy losses as the energy travels a shorter distance to reach the consumers while providing a higher selling price for the prosumers. Another benefit is the PCG can fulfil the energy demand of its own members, while operating off the grid. These types of objectives can motivate the community to save energy, which is encouraging for isolated locations that might have energy limitations. Due to the way the PCG members are selected and categorized in clusters, utilities can contract with these PCGs, making the process of managing prosumers much more efficient. In addition, since each PCG has similar energy sharing behaviors, disagreements among the members are minimized, leading to more stable groups in the long term, which also translates to a sustainable energy sharing process. Under these circumstances, managing prosumers connected to the utility grid has become highly important within the SGC and requires a new approach, i.e., transactive energy (TE). Under the TE approach the volumes of data amassed by the SGC are put to work, integrating both

P.42

utility-owned and third-party-owned resources, including power generation, ancillary services, load management services, and others, in order to secure the lowest-cost electricity in real time [2]. In other words, a TE system utilizes transactive controls (TC) through the smart grid infrastructure to send signals back and forth between utilities, grid operators and individual assets in the grid system, communicating the real-time flow and cost of power. In [3] , a summary of highlights of a five-year Pacific Northwest Smart Grid Demonstration project is presented, where the project’s transactive system operated for approximately two years and demonstrated the potential of TCs. 2 Proposed Transactive Control Strategy Fig. 1 illustrates the application of TC which is a single, integrated, and smart grid incentive signaling approach that combines multiple objectives and constraints (economic and operational) using uniform transactive incentive signals (TIS) and transactive feedback signals (TFS). These signals will propagate through an information network, which is the TC system embedded in the electrical network. The TIS represents the actual delivered cost of electric energy ($/MWh) at a specific system location (e.g., at a transactive node), whereas TFS is a representation of the net electric load (MW) at a specific system location (e.g., at a transactive node). Both signals include the current value and a forecast of future values as forward-looking signals; forecasts spans are day-ahead and updated every hour. Regarding the transactive node, its role is to respond to system conditions as represented by incoming TIS and TFS through decisions about the behavior of local assets, e.g., status and other local information, which are needed from node-owners to calculate the TIS and Fig. 1 Proposed transactive control. TFS. 3 Numerical Results and Analysis The TC is applied to a 6-bus test system that contains 3 consumer groups (CG) and 2 PCGs, (see Fig. 2). The TIS is sent to the Prosumer Group Controller (PGCO) (Transactive node) and from there it is sent to each PCG. The PCGs are considered to have photovoltaics (PV), energy storage systems (ESSs), and loads. Two cases are evaluated with two scenarios in each case. The forecasted price data was obtained from [4] for IESO Ontario market; the forecasted days are: 10/14/2010 (Case-1) and 12/7/2010 (Case-2). In each of the cases, the PCGs receive the TIS, and based on the signal, the PGCO determines the hours during which the ESS will be charged or discharged. However, the amount of energy that can be stored or sold depends on the local energy balance at each PCG, i.e., at any time period, the PCG first executes an energy balance within it and then determines the available energy. Once these calculations are done the PGCO sends the information back to the utility as a TFS, this information allows the utility to plan accordingly. Also, for each case, a comparison of a similar simulation is carried out where the major difference is that the PCGs do not receive

P.43

the TIS nor do they return a TFS, allowing us to demonstrate the benefits of the TIS and TFS signals.

1

To#utility#grid

Table 1 presents a summary of the savings (due to energy not PCG1 CG1 2 consumed) and profits (due to selling energy) for both cases. A 3 detailed analysis in Table 1 shows that in Case 1, the total PCG2 CG2 savings and profits are the same with and without TIS. This is 4 due to a coincidence between a high price and the conventional 5 discharge. This may not be always the case and the objective of CG3 maximizing the profit of the PCGs may not be true all the time 6 under the conventional approach. This can be seen in Case 2 Fig. 2. 6-bus test system. where the conventional discharge did not match the highest price, and in contrary, by using the TIS, maximum profits were achieved. Thus, the fundamental objectives of TE are met, i.e., under the TE approach every participant of the market obtains a benefit. In this demonstration, the PCGs obtain maximum profits and minimize their electricity bill. The utility obtains a reduction in peak demand (and a TFS) that enables a more efficient planning and operation of the grid by reducing the uncertainty in the operation of these PCGs. Table 1. Total energy savings and profits in the considered cases. Scenario 2: without TIS

Scenario 1: with TIS PCG1

PCG2

PCG1

PCG2

Savings ($)

Profits ($)

Savings ($)

Profits ($)

Savings ($)

Profits ($)

Savings ($)

Profits ($)

Case 1

15.67

0.96

20.54

1.84

15.67

0.96

20.56

1.84

Case 2

17.17

1.17

22.45

2.07

16.00

0.78

21.04

1.63

4 Conclusion A TC approach was developed and applied to a 6-bus test system to validate the effectiveness of the control. The results indicated that the TC helps the PCGs to maximize their profits and minimize their electricity bill, as well as, reduce peak load and provide a feedback signal for the utility. The feedback signal enables the utility it to carry out a more efficient planning and operation strategy. References [1] A. J. D. Rathnayaka, V.M. Potdar, T. Dillon, O. Hussain, and S. Kuruppu, “Goal-Oriented Prosumer Community Groups for The Smart Grid,” IEEE Technology and Society Magazine, pp. 42-48, 2014. [2] A. K. Bejestani, A. Annaswamy, and T. Samad, “A Hierarchical Transactive Control Architecture for Renewables Integration in Smart Grids: Analytical Modeling and Stability,” IEEE Transactions on Smart Grid, Vol. 5, No. 4, pp. 2054-2065, Jul. 2014. [3] Pacific Northwest Smartgrid Demonstration Project, “Technology performance report highlights,” 2015. [4] P. Mandal, A. Haque, J. Meng, A. K. Srivastava, and R. Martinez, “A Novel Hybrid Approach Using Wavelet, Firefly Algorithm, and Fuzzy ARTMAP for Day-Ahead Electricity Price Forecasting,” IEEE Transactions on Power Systems, Vol. 28, No. 2, May 2013.

P.44

SETS2016-12 The Southwest Emerging Technology Symposium 2016

Room Temperature Processed CuSCN Hole Transportation Layers for the Use in Perovskite Based Solar Cells Jose Galindo1, Manuel Martinez1, Shaimum Shahriar1, Vanessa Castenada1, Donato Kava1, Cheik Sana1, and Deidra Hodges1 1Electrical and Computer Engineering, The University of Texas at El Paso, El Paso, TX 79968, U.S.A.

ABSTRACT Development of fabrication techniques to produce cost efficient transparent CuSCN hole transportation layers for the rising perovskite-based solar cells. The CuSCN films were deposited by spin coating techniques using the WS650 spin processor from Laurell Technologies. CuSCN is a viable alternative to Spiro-OMeTAD due to its high optical transparency and p-type conductivity. CuSCN has been shown to protect the perovskite from photo-thermal instability. Various characterization instruments will be used for the CuSCN thin films, such as: Cary 5000 UV-Vis spectrophotometer, Hall Effect Measuring System, and 4 point probe. Achieving a high quality material, while having a simple and cost efficient fabrication method, is the primary focus of this study. INTRODUCTION Hybrid organic-inorganic perovskite solar cells have increased in efficiencies of over 20% [1]. Various perovskite device architectures have been employed to achieve high efficiencies. The most common architectures are those that put the perovskite layer in between a mesoporous or solid state metal-oxide n-type and an p-type layer such as 2,2,7,7-tetrakis(N,N-dip-methoxyphenylamine)-9,9- spirobifluo-rene (Spiro-OMeTAD) which have reached efficiencies of over 15% [2]. Many of these high efficiency architectures use organic p-type layers to conduct holes. The organic p-type layers like Spiro-OMeTAD and poly(3,4ethylenedioxythiophene):polystyrenesulfonate (PEDOT:PSS) give efficiencies, but are not chemically stable like perovskite [3]. The main concern when evaluating perovskite solar cells for the eventual commercialization is its stability for long-term use [4][5][6]. Inorganic p-type layers such as Copper Thiocyanate (CuSCN) and NiO have been explored due to their protective properties as capping layers for the perovskite film [7][8]. CuSCN is a p-type material that is highly transparent that has been shown to work in other transparent devices not just including photovoltaic applications [9]. EXPERIMENT The CuSCN solution was a clear liquid which was processed using the received CuSCN (Sigma Aldrich 99%) and dissolved in dipropyl sulfide (Sigma Aldrich 97%) using a concentration of 15 mg mL-1 [10]. The solution was then left to stir at room temperature for 6

P.45

hours. Soda lime glass (SLG) 1 inch samples were used as substrates. The substrates were cleaned using ultrasonic cleaning methods and dried using nitrogen. This allowed for the best coverage of films to be deposited onto the substrates. The perovskite solution was made by stirring a 1:1 molar ratio of PbI2 and CH3NH3I [11]. CH3NH3I was made by reacting methylamine and hydroiodic acid and then stirred at 0°C. The resulting precipitate was recovered and then collected to be dried at 60°C in a vacuum oven [12]. The perovskite solution was synthesized by dissolving PbI2 and CH3NH3I in N,Ndimethylacetamide (DMA, >99% Sigma) and stirred for over 10 hours. A 0.2uL amount of the solution was then spin coated at high rpm using the WS650 spin processor (Laurell Technologies). The samples were then dried in air at 65°C using the Corning PC420D hot plate. The CuSCN layers were first deposited onto the SLG substrates after cleaning by spin coating at 4000 rpm for 30 seconds using the WS650 spin processor (Laurell Technologies) after being heated to 60°C for 5 minutes to allow the solvent to evaporate from the film. The samples were measured using the Signatone Pro-4 4 point probe station to measure sheet resistance and resistivity, the Ecopia HMS-300 Hall Effect measuring system was used to measure mobility and the average hall coefficient, and used the Cary 5000 UV-Vis-NIR spectrophotometer for optical properties. The samples were visually observed to note any degradation in the perovskite films.

DISCUSSION Optical Properties The optical characterization for the samples were obtained using wavelengths from 800 nm to 200 nm. The average transmission data is show in Figure 1.From Figure 1, it can be observed that the CuSCN samples have high transmission percentages as shown in K. Zhao et al’s research [10]. The data shows the CuSCN samples have high transmission rates for the wavelengths 300-800 averaging around 84%. This can be attributed to its high transparent layers that are produces after spin coating. These high transmission highlight the CuSCN’s optical properties in regard to their transparent electronic applications [9].

P.46

Figure 1. Average transmittance data for CuSCN samples.

Electrical Properties The electrical properties of the CuSCN samples were characterized by Hall Effect measurements and four point probe. Table 1 shows the data received from the hall effect measurement system.

Table 1. This table shows the results obtained by hall effect measuring system. Sample Number CuSCN 1 CuSCN 2 CuSCN 3

Mobility (cm2 / Vs) 1.592x101 1.657x101 1.453101

Average Hall Coefficient (m2/C) 4.637x106 4.976x106 4.032x106

Table 2 shows the measurements obtained by four point probe. CuSCN was measured on top of perovskite, and on top of SLG. A measurement of the perovskite sample was also conducted as a reference. Table 2. This table shows the parameters obtained from the four point probe. Material Resistivity ( ) Sheet Resistivity ( ) CuSCN on top of 3.237x104 6.474x108 Perovskite1 CuSCN on top of 2.299x103 2.299x108 2 SLG Perovskite Sample3 1.501x104 3.002x108 1.

-7

Measured with a current of 2.017x10 mA Measured with a current of 2.017x10-7 mA 3. Measured with a current of 4.117 x10-7 mA 2.

P.47

Resistance (Ω) 1.383x108 7.432x107 8.829x107

CONCLUSIONS CuSCN was characterized by multiple techniques, and was shown to help the perovskite layer in terms of thermal and photo stability. The room temperature and fast solution process of the CuSCN layers can be used in flexible solar cells. CuSCN has shown comparable electrical results with ZnO and as an alternative to low-cost solution processed p-type layers. Future work will include the fabrication of an entire solar cell using CuSCN as the p-type layer. ACKNOWLEDGMENTS This work was supported by the University of Texas at El Paso (UTEP) College of Engineering, the Electrical and Computer Engineering Department, and UTEP-Partnerships for Research and Education in Materials (PREM)—NSF Grant DMR-1205302. REFERENCES [1]

“Best Research-Cell Efficiencies.” [Online]. Available: http://www.nrel.gov/ncpv/images/efficiency_chart.jpg. [Accessed: 01-Jan-2016].

[2]

J. Burschka, N. Pellet, S.-J. Moon, R. Humphry-Baker, P. Gao, M. K. Nazeeruddin, and M. Grätzel, “Sequential deposition as a route to high-performance perovskite-sensitized solar cells.,” Nature, vol. 499, no. 7458, pp. 316–9, 2013.

[3]

C.-H. Chiang, Z.-L. Tseng, and C.-G. Wu, “Planar heterojunction perovskite/PC71BM solar cells with enhanced open-circuit voltage via a (2/1)-step spin-coating process,” J. Mater. Chem. A, vol. 2, p. Ahead of Print, 2014.

[4]

M. Jorgensen, K. Norrman, and F. C. Krebs, “Stability/degradation of polymer solar cells,” Sol. Energy Mater. Sol. Cells, vol. 92, no. 7, pp. 686–714, 2008.

[5]

M. D. Irwin, D. B. Buchholz, A. W. Hains, R. P. H. Chang, and T. J. Marks, “p-Type semiconducting nickel oxide as an efficiency-enhancing anode interfacial layer in polymer bulk-heterojunction solar cells,” Proc. Natl. Acad. Sci., vol. 105, no. 8, pp. 2783–2787, 2008.

[6]

K. Norrman, M. V. Madsen, S. a. Gevorgyan, and F. C. Krebs, “Degradation Patterns in Water and Oxygen of an Inverted Polymer Solar Cell,” J. Am. Chem. Soc., vol. 132, no. 17, pp. 16883–16892, 2010.

[7]

A. S. Subbiah, A. Halder, S. Ghosh, N. Mahuli, G. Hodes, and S. K. Sarkar, “Inorganic hole conducting layers for perovskite-based solar cells,” J. Phys. Chem. Lett., vol. 5, no. 10, pp. 1748–1753, 2014.

[8]

S. Ye, W. Sun, Y. Li, W. Yan, H. Peng, Z. Bian, Z. Liu, and C. Huang, “CuSCN-based inverted planar perovskite solar cell with an average PCE of 15.6%,” Nano Lett., p. 150504172654007, 2015.

[9]

P. Pattanasattayavong, N. Yaacobi-Gross, K. Zhao, G. O. N. Ndjawa, J. Li, F. Yan, B. C. O’Regan, A. Amassian, and T. D. Anthopoulos, “Hole-transporting transistors and circuits based on the transparent inorganic semiconductor copper(I) thiocyanate (CuSCN) processed from solution at room temperature,” Adv. Mater., vol. 25, no. 10, pp. 1504–1509, 2013.

P.48

[10]

K. Zhao, R. Munir, B. Yan, Y. Yang, T.-S. Kim, A. Amassian, Solution-processed inorganic copper(I) thiocyanate (CuSCN) hole transporting layers for efficient p–i–n perovskite solar cells, J. Mater. Chem. A. 3 (2015) 20554–20559. doi:10.1039/C5TA04028K.

[11]

J. Burschka, N. Pellet, S.-J. Moon, R. Humphry-Baker, P. Gao, M.K. Nazeeruddin, et al., Sequential deposition as a route to high-performance perovskite-sensitized solar cells., Nature. 499 (2013) 316–9. doi:10.1038/nature12340.

[12]

M. Liu, M.B. Johnston, H.J. Snaith, Efficient planar heterojunction perovskite solar cells by vapour deposition., Nature. 501 (2013) 395–8. doi:10.1038/nature12509.

P.49

SETS2016-13 The Southwest Emerging Technology Symposium 2016

SYNTHESIS OF CU2ZNSNSXSEX-4 THIN FILMS ABSORBER LAYERS FOR LOW-COST, HIGH EFFICIENCY THIN FILM SOLAR CELLS Cheik Sana1, Shaimum Shahriar1, Jose Galindo1, Donato Kava1, Vanessa Castaneda1, Manuel Martinez1, Edison Castro2, Luis Echegoyen2 and Deidra Hodges1 1

Department1Department of Electrical and Computer Engineering, 2 Department of Chemistry * Corresponding author ([email protected])

Keywords: CZTS, CZTSSe, solar cells, photovoltaic, thin film ABSTRACT A non-vacuum, solution based method to deposit Cu2ZnSnSSe4 was investigated. In this approach, a precursor solution of CZTSSe was formed by reacting metal sources copper (II) acetate monohydrate, zinc (II) acetate dehydrate, tin (II) chloride dehydrate and elemental powders of sulfur and selenium powders in a solution of 2-metoxyethanol. The slurry was then spincoated followed by annealing at different temperatures. Optical, structural and electronic characterization of thin films were performed using scanning electron microscope (SEM), thermogravimetric analysis (TGA), X-ray diffractometer (XRD), Raman spectrometer, and UV-Vis spectrophotometer. X-Ray diffractograms show different shifts of the kesterite/stannite (112) peak, which indicates the presence of CZTSSe. The three major peaks of the (112), (220), and (312) planes had respective 2θ in the vicinity of 28°, 47.5° and 56°, characteristics of CZTSSe. The shift of the peaks depends on the ratios of S/Se in the synthetized material, which depends on the annealing parameters. The lattice constants decrease linearly with increasing contents of S in the precursor solution. Raman spectroscopy confirmed the formation of both quaternary CZTS and CZTSe. 1

Introduction

Prominent thin-film photovoltaic (PV) solar cells such as Cu(In,Ga)Se2 and CdTe have attracted a lot of attention and are being commercialized because of their respective power conversion efficiency (PCE) over 21.7% [1] and 21.5% [2] in laboratory conditions. CIGSbased devices are currently the most efficient thin film solar cells on the market. However, the scarcity of indium and gallium in the earth’s crust are a limiting factor in mass production of CIGS solar cells. It is therefore imperative to synthetize new materials for solar cells that could compensate for the current limitations. Recently, earth-abundant kesterite Cu2ZnSn(S,Se)4 (CZTSSe) absorbing materials have been getting significant attention in the thin film solar cells due to their suitable band gaps, high absorption coefficient, and low material cost[3]. Interestingly, the highest PCE, currently 12.6%[4], has been obtained through a hydrazine-based solution method. This high efficiency demonstrates the potential of low-cost CZTSSe thin film solar cells. These methods have two major drawbacks. The first one is the use of hydrazine. Hydrazine is a highly toxic and dangerously unstable compound, which must be handled and stored with special care. The second drawback of these methods is the requirements of either post selenization or sulfurization. As a result, a novel method to synthetize CZTSSe has been investigated. This

P.50

method would eliminate both the use of hydrazine and the selenization/sulfurization phase and could lead the path to a low cost, relatively safe and easily scalable method of fabricating device quality CZTSSe thin films. CZTSSe has abundant elements (Cu, Zn, Sn, S), which are non-toxic and low toxicity element (Se) [5]. It has a tunable optical band gap of 1~1.5 eV and a large optical absorption coefficient of over 104 cm-1 in visible wavelength region for photovoltaic application [6, 7]. These optical characteristics can yield a theoretical PCE of 25 to 30% according to Shockley-Queisser theory [4, 8]. There are two main approaches to synthetize CZTSSe, specifically vacuum based processes and non-vacuum solution based processes [9-13]. The vacuum based processes include RF magnetron sputtering, thermal evaporation, pulsed laser deposition, electron beam evaporation and the non-vacuum processes include spray pyrolysis, photochemical deposition, sol-gel method, spin-coating method and electrodeposition. The non-vacuum solution processes are considered the low cost route for fabricating solar cells [6]. In this paper, CZTSSe is synthesized by mixing copper (II) acetate monohydrate, zinc (II) acetate dehydrate tin (II) chloride dehydrate, sulfur and selenium powders in 2-methoxyethanol with added stabilizer. Thin films are deposited by spin-coating in a non-vacuum environment and annealed in a Rapid Thermal Processing (RTP) furnace. We investigate the influence of annealing temperature on the characteristics of CZTSSe thin films. The structural, morphological, and optical properties have been studied by X-ray diffraction, Raman spectroscopy, Scanning Electron Microscopy, transmission and surface profiling. 2

Experimental

In this study, film preparation began with the preparation of CZTSSe precursors, based on our previous work with CZTS preparation [14]. The CZTSSe precursor solution was obtained by mixing of 4.375 x 10-2 mol of copper (II) acetate monohydrate, 2.875x10-2 mol of zinc (II) acetate dehydrate, 2.875 x 10-2 mol of tin (II) chloride dehydrate, 2.875x10-2 mol of sulfur and 2.875x10-2 mol of selenium in 50.0 mL of 2-methoxyethanol. As a stabilizer, 5.0 mL of monoethanolamine (MEA) was added to the solution. Excess selenium was added to the solution to compensate for the Se evaporated during the annealing phase. The CZTSSe precursor solution was stirred and heated at 45°C for 1 hour to dissolve metal sources, prior to spin-coating. Soda lime glass (SLG) was used as substrates. SLG substrates were cleaned with methanol and acetone, rinsed with deionized water, and dried with nitrogen (N2), to remove hydrocarbons and other contaminants, prior to the film deposition. The solution was deposited on SLG substrates by spin coating at 4,000 rpm for 30s followed by 15 min sintering in air at 350°C on a hot plate. This Spincoating/sintering process was repeated 7 additional times resulting in a total of 8 layers of CZTSSe. After the eighth layer, samples were annealed in a RTP furnace, varying the temperature from 400°C to 550°C, followed by varying annealing time from 15 to 60 minutes. The phase and crystalline structures of the CZTSSe films were characterized by X-ray diffraction (XRD) and Raman spectroscopy. XRD was operated in the 2θ range from 10 to 80° on a Bruker D8 XRD using a step size of 0.02° and a step time of 0.3 s. The crystal structure of the CZTSSe thin films was analyzed using Bruker’s EVA software. The Raman spectra were recorded with a Thermo Scientific DXR SmartRaman spectrometer using a laser

P.51

wavelength of 532nm. The Raman measurements were performed at room temperature. Thermogravimetric analysis (TGA) was done using a Mettler-Toledo TGA/DSC 1 STARe. The morphology and size of the film was measured using Hitachi S-4800 Scanning Electron Microscope. KLA Tencor Alpha Step D 120 was used to analyze the surface profile of the films. Cary 5000 UV-Vis spectrophotometer wavescans in the range of 200 nm to 1100 nm was used to analyze the optical properties including the band gap energy, transmission and absorption of the samples. .

3

Results and discussion

CZTSSe was prepared by reacting copper (II) acetate monohydrate, zinc (II) acetate dehydrate, tin (II) chloride dehydrate, sulfur and selenium in the solvent 2-methoxyethanol. Dissolution of the metal powders and formation of a clear and homogenous precursor solution of CZTSSe can be observed after magnetic stirring of the solution on the hot plate at 45°C. Additional selenium powder was added to the solution prior to spin-coating. Multiple layers were spincoated in order to avoid formation of cracks and pin hole defects on the CZTSSe precursor films. Reported here are the results for the samples annealed under vacuum for 45 minutes at temperatures varying from 450°C to 525°C with a 25°C increment. 3.1 Structural Properties Fig. 1.a shows the X-Ray Diffraction pattern of the samples annealed at different temperatures. There is no pdf data specifically for CZTSSe. The 2θ values for the major peaks located on the following planes (112), (220), (312), are shifted towards lower angle values respective to CZTS (JCPDS Card no. 26-0575) and upper angle values respective to CZTSe (JCPDS Card no. 52-868). As the temperature increases, the peaks are shifted towards higher 2θ values. This could be explained by less Se-replaced CZTSSe and close to phase pure CZTS. As the temperature increases, the samples exhibit more intense and sharper peaks, corresponding to higher crystallinity and larger grains. This confirms the dependency of annealing conditions on the increase of the crystallinity and the size of the crystals. The average crystallite size of the nanoparticles can be calculated from XRD data using Scherrer’s equation [15] " = $λ/(β)*+θ) (1) where D represents the crystallite size, λ = 1.54 Å is the X-ray wavelength (for Cu Kα line), β (in rad) is the peak width at half-maximum, θ is the angle of reflection, and K = 0.9 is the shape factor. The CZTS grains in fig. 1.b have an average size of varying from 10.1nm to 13.8 nm and the CZTSe grains in the annealed sample in fig. 1.b have an average size varying from 6.7 nm to 10.52 nm. The crystal size of the CZTS grains in Fig. 1.b increases as temperature increases while the CZTSe crystal size of fig. 1.b decreases as temperature increases. This can be attributed to the loss of selenium at higher temperature, which gets the CZTSSe samples closer to pure CZTS phase. A higher content of sulfur will decrease the grain size because the covalent radius of sulfur (1.02 Å) is smaller than that of selenium (1.17 Å) [11]. The location of the major peaks suggests that CZTSSe was synthetized. However, XRD patterns of some binary selenides, such as CuSe, CuS, Cu2Se, Cu2S, ZnSe, ZnSe or ternary selenides such as Cu2SnSe3, or Cu2SnSe3 are very analogous to that of CZTSe, making XRD pattern only

3 P.52

insufficient to demonstrate that pure CZTSSe thin film was fabricated [16]. In addition to XRD, Raman spectroscopy was performed to evaluate the secondary phases. Fig. 1.c shows the Raman spectra of the precursor film and the annealed film. All the spectra of the studied CZTSSe exhibited peaks of some secondary phases CZTS. The dominant phase corresponds to CZTS with main vibration located at 338 cm-1. In addition to these peaks, a large peak was observed at between 475 cm-1 and 490 cm-1, corresponding respectively to the undesirable phases of binary CuxS and CuxSe. A weak peak corresponding to ZnSe is detected. However this small amount of ZnSe has no significant effect on the efficiency of kesterite solar cells [17]. Unidentified peaks analogous to other binary or ternary chalcogenides were observed. This corresponds to impurities that will be addressed in our future works. Fig. 1.d presents the thermogravimetric analysis (TGA) curve for the mixed CZTSSe precursor. The TGA sample was prepared from the CZTSSe precursor ink by preheating at 100 °C for 60 min to remove the solvent and low-boiling-point molecules. The initial weight loss from 25°C to 220°C could be attributed to the removal of the remaining solvents, and the sulfur species. The second stage of weight loss, from 220°C to 350°C could be attributed mainly to the loss of Se and Sn. The third weigh loss is attributed to the loss of Zn elements. Finally, the remainder of the material will be mostly Cu, the element with the highest melting temperature. a

b

CZTS

CZTSe

c

d

Fig. 1. (a) XRD profiles of CZTSSe films annealed at temperatures of 450°C, 475°C, 500°C and 525°C; (b)CZTS and CZTSe crystallite size and FWHM of thin films; (c) Raman spectroscopy of CZTSSe films annealed at temperatures of 450°C, 475°C and 500°C; (d) Thermogravimetric analysis of CZTSSe precursor powder

P.53

The elemental composition of the CZTSSe thin films was investigated by energy dispersive X-Ray analysis and the spectra are shown in Fig. 2. The results confirm that the sample contains Cu, Zn, Sn, S and Se. However, the stoichiometry differed from what is reported in the literature. This could be explained by the heavy loss of S and Se during the annealing phase. In order to achieve high-performance CZTSSe solar cells, we need to optimize not only the purity of the crystal phase but also the Cu-poor and Zn-rich composition of CZTSSe thin film. 3.2 Morphological Properties Zhou and al. reported that the morphology and crystal size of the absorber layer are crucial for the performance of the kesterite solar cells [18]. Fig. 3. shows the top-view SEM images of the precursor and the annealed films. Large and densely packed grains with an average size ranging from 1µm to 5µm were observed on the annealed samples. It is clear that the annealing process resulted in larger CZTSSe grain size and more dense thin films. It was also observed that the peripheral grains were larger than the grains towards the center of the substrate. Even though multiple layers of the slurry were spincoated, fracture defects were observed on the surface of the films under SEM. The films have an average thickness of 2.9 µm which is desirable for thin film application. The film’s thickness was measured with KLA Tencor Alpha Step D 120. 3.3 Optical Properties The optical properties of the CZTSe thin films were analyzed using Cary UV-VIS optical transmission spectroscopy measurements and Scanning electron microscopy at room temperature. The optical absorption coefficient (α) was determined from the measured transmittance (T) using the formula. α = A hυ - E g 1/ 2 (2)

(

)

Intensity a.u.

where α is the absorption coefficient, A is a constant, Eg is the energy band gap and hν is the incident photon energy [19]. Fig. 4 shows the plot of (αhν)2 as a function of incident photon 450°C

475°C

500°C

525°C

Energy in KeV

Fig.2. EDAX spectrum of CZTSSe films

Fig.3. SEM images of CZTSSe films annealed at temperatures of 450°C, 475°C, 5 P.54

energy for CZTSSe thin films annealed for 45 minutes at temperatures of 450°C, 475°C, 500°C and 525°C. The estimated direct optical band gap were obtained by extrapolation of the (αhν)2 versus hv plot at α=0. The bandgap energy increases as the temperature increases. It goes from 1.18 eV at the lowest temperature (450°C) to 1.43 eV. This can be explained by the fact that bandgap energy increases as selenium content decreases, which yield to a bandgap closer to pure phase CZTS at the higher temperature. The estimated optical bandgap of CZTSSe thin film on SLG substrate is consistent with the reported values in the literature 1~1.5 eV [20-22]. 4

Conclusion

The CZTSSe thin films were successfully synthetized and deposited using low cost spin coating method. As the annealing temperature increases, there is a greater loss of selenium in the CZTSSe thin films which causes an increase in bandgap energy. The measured bandgap varies from 1.18 to 1.45 eV. The sample annealed at 475°C exhibited the most desirable characteristics. CZTSSe solar cells were successfully fabricated and the results will be published in our future papers. Acknowledgments This work was supported by The University of Texas at El Paso (UTEP) School of Engineering, the National Science Foundation, NSF NNIN at Microelectronics Research Center (MRC) at the University of Texas at Austin and UTEP-UCSB Partnership in Research and Education in Materials (PREM, GRANT # C1205302).

References 1.

Jackson, P., Hariskos, D., Wuerz, R., Kiowski, O., Bauer, A., Friedlmeier, T. M., and Powalla, M., 2015, “Properties of Cu(In,Ga)Se2 solar cells with new record efficiencies up to 21.7%,” Phys. status solidi – Rapid Res. Lett., 9(1), pp. 28–31. 2. Friedlmeier, T. M., Jackson, P., Bauer, A., Hariskos, D., Kiowski, O., Wuerz, R., and Powalla, M., 2015, “Improved Photocurrent in Cu ( In , Ga ) Se2 Solar Cells :,” pp. 7–9. 3. Yang, Y., et al., Solution-processed highly efficient Cu2ZnSnSe4 thin film solar cells by dissolution of elemental Cu, Zn, Sn, and Se powders. ACS Appl Mater Interfaces, 2015. 7(1): p. 460-4. 4. Wang, W., et al., Device Characteristics of CZTSSe Thin-Film Solar Cells with 12.6% Efficiency. Advanced Energy Materials, 2014. 4(7): p. n/a-n/a. 5. Ilari, G.M., et al., solar cell absorbers spin-coated from amine-containing ether solutions. Solar Energy Materials and Solar Cells, 2012. 104: p. 125-130. 6. Ennaoui, A., et al., Cu2ZnSnS4 thin film solar cells from electroplated precursors: Novel low-cost perspective. Thin Solid Films, 2009. 517(7): p. 2511-2514. 7. Ito, K. and T. Nakazawa, Electrical and Optical Properties of Stannite-Type Quaternary Semiconductor Thin Films. Jpn. J. Appl. Phys., 1988. 27. 8. Shockley, W. and H.J. Queisser, Detailed Balance Limit of Efficiency of p-n Junction Solar Cells. Journal of Applied Physics, 1961. 32(3): p. 510. 9. Mitzi, D.B., et al., The path towards a high-performance solution-processed kesterite solar cell. Solar Energy Materials and Solar Cells, 2011. 95(6): p. 1421-1436. 10. Ahmed, E., et al., The influence of annealing processes on the structural, compositional and electro-optical properties of CuIn0. 75Ga0. 25Se2 thin films. J. Mater. Sci.: Mater. El., 1996. 7(3): p. 213-219.

P.55

11. Gao, F., et al., Structural and Optical Properties of In-Free Cu$_{2}$ZnSn(S,Se)$_{4}$ Solar Cell Materials. Japanese Journal of Applied Physics, 2012. 51: p. 10NC29. 12. Liu, T.-H., et al., Metallurgical Mechanism and Optical Properties of CuSnZnSSe Powders Using a 2-Step Sintering Process. Journal of Nanomaterials, 2014. 2014: p. 1-8. 13. Brammertz, G., et al., Electrical characterization of Cu2ZnSnSe4 solar cells from selenization of sputtered metal layers. Thin Solid Films, 2013. 535: p. 348-352. 14. Hodges, D., et al. Development of CZTS thin films by non-vacuum, liquid-based techniques for efficient, low-cost CZTS solar cells. in Photovoltaic Specialists Conference (PVSC), 2013 IEEE 39th. 2013. 15. Jin, C., P. Ramasamy, and J. Kim, Facile hot-injection synthesis of stoichiometric Cu2ZnSnSe4 nanocrystals using bis(triethylsilyl) selenide. Dalton Trans, 2014. 43(25): p. 9481-5. 16. Tian, Q., et al., A robust and low-cost strategy to prepare Cu2ZnSnS4precursor solution and its application in Cu2ZnSn(S,Se)4solar cells. RSC Adv., 2015. 5(6): p. 4184-4190. 17. Hsu, W.-C., et al., Growth mechanisms of co-evaporated kesterite: a comparison of Cu-rich and Zn-rich composition paths. Progress in Photovoltaics: Research and Applications, 2014. 22(1): p. 35-43. 18. Zhou, H., et al., CZTS nanocrystals: a promising approach for next generation thin film photovoltaics. Energy & Environmental Science, 2013. 6(10): p. 2822. 19. Weber, A., et al., In-situ XRD on formation reactions of Cu2ZnSnS4 thin films. Phys. Status Solidi C, 2009. 6(5): p. 1245-1248. 20. Woo, K., et al., Band-gap-graded Cu2ZnSn(S1-x,Se(x))4 solar cells fabricated by an ethanol-based, particulate precursor ink route. Sci Rep, 2013. 3: p. 3069. 21. Salomé, P.M.P., et al., Growth and characterization of Cu2ZnSn(S,Se)4 thin films for solar cells. Solar Energy Materials and Solar Cells, 2012. 101: p. 147-153. 22. Wang, C.-L. and A. Manthiram, Low-Cost CZTSSe Solar Cells Fabricated with Low Band Gap CZTSe Nanocrystals, Environmentally Friendly Binder, and Nonvacuum Processes. ACS Sustainable Chemistry & Engineering, 2014. 2(4): p. 561-568.

7 P.56

SETS2016-14 The Southwest Emerging Technology Symposium 2016

CRYSATLLIZATION OF POLY(3-HEXYLTHIOPHENE) ON CARBON DERIVATIVES FOR ORGANOPHOTOVOLTAIC APPLICATION A. Mishra1, V.S.A.Challa1, K.C.Nune1, R.D.K Misra1*, D.Hodges2 1 Department of Metallurgical, Materials and Biomedical Engineering 2 Department of Electrical and Computer Engineering , University of Texas, El Paso, TX 79968, USA; * Corresponding author ([email protected])

Keywords: poly(3-hexylthiophene), organophotovoltaic, carbon nano tubes, graphene ABSTRACT Conductive polymers such as poly(3-hexylthiophene) (P3HT) has become attractive candidate to bring forth new architecture of organophotovoltaic (OPV) devices. Supramolecular structures involving P3HT and carbon derivatives such as carbon nano tubes (CNTs) and graphene were synthesized by new solution based nucleation approach. Morphology of nucleated nanofibers of the P3HT on CNTs and graphene was characterized and observed high density transcrystalline nanofibers of different dimensions. Nanofibers of the P3HT on CNT have ~ 2-3 nm thickness and 800-1000 nm length moreover in case of P3HT and graphene structure, nanofibers have ~ 8-10 nm thickness and 100 nm in length. These epitaxially grown nanoscale structure have also expressed optical and photoluminescence characteristics which was studied by UV-vis spectroscopy technique. This study indicates accelerated crystallization of P3HT in the presence of carbon derivatives and also suggests mutual strong π-π interaction which is a fundamental indication of charge transport. Thus, explained supramolecular structure has intense potential to establish a connection between effective efficiency and flexible processing for OPV devices. 1 Introduction Over 25 years ago, synthesis of regioregular P3HT fascinated OPV research field because of its excellent properties such as macromolecular self-assembly into polycrystalline, ordered domain, nanowire morphology in differing dimensions and high charge mobility.[1] These inherent qualities of regioregular P3HT make an appropriate candidate for active layer material for OPV cells.[2] Architecture of active layer defines OPV cell performance where we consider strong π-π interactions between thiophene rings and high degree of crystalline packing. Carbon derivatives such as graphene and CNTs are also having inherent high electronic and optoelectronic properties.[3, 4, 5] Composite of crystallized P3HT nanowires on CNTs and graphene can be used to enhance the performance of active layer in which their individual excellent optoelectronic properties can deliver effective charge mobility in OPV devices.[6] π conjugated polymers (such as P3HT) have tendency to grow as thin nano wires during crystallization which provide high ordered one dimensional (1D) charge carrier path.[7,8] Crystallization of P3HT is induced in the presence of CNTs because of the contact area of CNT to crystallize along the tube axis and evolve as transcrystalline nanofibrils.[9] Since graphene also has same graphitic domain like CNTs, it is expected that P3HT can interact with graphene and will crystallize as vertical nanowires on a 2 dimensional (2D)

P.57

lattice of graphene.[10] Through this crystallization, we can control the dispensability of graphene in organic solvent which is a major issue on graphene based nano devices. In conclusion, the described supramolecular structure of P3HT nanowires with CNTs and graphene has all the potential to improve performance of the OPV cell. 2 Experimentation and sample preparation 2.1 Synthesis of Poly (3-hexylthiophene) and CNT hybrid structure We obtained single wall carbon nanotube from global nanotech, Mumbai. Poly(3hexylthiophene) (P3HT) of molecular weight 50–70 kDa (96% regioregularity), was obtained from Acros Chemicals, USA, and decahydronaphthalene (decalin) was obtained from Sigma-Aldrich, USA. 0.1 mg CNT was dissolved in 10 ml of decalin and dispersed using Ultrasonication for 3 hr followed by 20 minutes of probe sonication. Simultaneously, P3HT dissolved in decalin (1mg/ml). P3HT heated to 1500C for 1hr. After proper dispersion of CNT, we added dropwise CNT into P3HT solution and then we quenched this solution to 800C and kept that solution for 8 hrs. After crystallization, we filtered the purple colour solution using filter paper of 0.45 µm size. 2.2 Synthesis of Poly (3-hexylthiophene) Graphene) hybrid structure We obtained graphene, decahydronaphthalene (decalin) and low molecular weight (15-45 kDa) Poly(3hexylthiophene) (P3HT) from Sigma-Aldrich, USA. Graphene was dissolved in decalin (0.1mg/ml) and dispersed properly for 3 hrs. of ultrasonication followed by 20 minutes probe sonication. We also dissolved P3HT in decalin (1mg/ml) and heated to 1500C for 1 hr. After complete dispersion of graphene, we added dropwise graphene into P3HT solution then quenched to 800C for 8 hrs. We filtered resulting purple color solution using filter paper of 0.45 µm size. 3 Observation CNTs and graphene have high tendency to agglomerate which challenges the homogeneous crystallization of P3HT onto the surface of these carbon derivatives. To solve this issue, we dispersed P3HT and mentioned carbon derivatives in the same dispersive media. Fig.1 (a) indicates the crystallized long nanofibrils of P3HT on CNT in a periodic manner. These fibres are nucleated in a transcrystalline fashion along the tube axis which proves that CNT provides large contact area and orientation for growth of P3HT nanowires. Fig.1 (b) illustrates a schematic of P3HT nanofibrils on graphene 2D lattice. Complete analysis of nanocomposite of P3HT and graphene nanoplatelets is under study. We’ve observed the resulting purple color solution during our experimentation which is a prime indication of polymer crystallization on graphene surface.

P.58

a. 1:10

Fig.1 (a) TEM micrograph of crystallized Poly(3hexylthiophene) (P3HT) nanofibrils on CNT surface in a periodic manner[9], (b) Schematic diagram of P3HT crystallization on graphene 2D lattice of graphene.

4 References [1] M. Jeffries, R.D.McCullough “The Conjugated Polymers: Theory, Synthesis, Properties, and Characterization”. 3rd edition, CRC Press (Taylor &Francis Group), 2006. [2] H. Sirringhaus, P. J. Brown, R. H. Friend, M. M. Nielsen, K. Bechgaard, B. M. W. Langeveld-Voss, A. J. H. Spiering, R. A. J. Janssen, E. W. Meijer, P. Herwig and D. M. de Leeuw, “Two-dimensional charge transport in self-organized, high-mobility conjugated polymers”. Nature, 401, 685, 1999. [3] A. K. Feldman, M. L. Steigerwald, X. Guo and C. Nuckolls, “Molecular electronic devices based on singlewalled carbon nanotube electrodes”. Acc. Chem. Res., 2008, 41, 1731, 2008. [4] A. Javey, J. Guo, Q. Wang, M. Lundstrom and H. Dai, “Ballistic carbon nanotube field-effect transistors”. Nature, 424, 654, 2003. [5] KI Bolotin, KJ Sikes, Z Jiang, M Klima, G Fudenberg, J Hone, P Kim, H.L.Stormer, “Ultrahigh electron mobility in suspended graphene”. Solid State Communication, 146,351-355, 2008. [6] J. J. M. Halls, K. Pichler, R. H. Friend, S. C. Moratti , A. B. Holmes, “Exciton diffusion and dissociation in a poly(p‐phenylenevinylene)/C60 heterojunction photovoltaic cell”. Appl. Phys. Lett., 68, 3120, 1996. [7] A. P. H. J. Schenning, E. W. Meijer, “Supramolecular electronics; nanowires from self-assembled πconjugated systems”. Chem. Communication, 26, 3245, 2005 [8] M. Brinkmann, F. Chandezon, R. B. Pansu and C. Julien- Rabant, “Epitaxial Growth of Highly Oriented Fibers of Semiconducting Polymers with a Shish-Kebab-Like Superstructure”. Advanced. Functional Materials, 9, 2759, 2009. [9] R.D.K Misra, D. Depan, V.S.A.Challa, J.S.Shah, “Supramolecular structures fabricated through the epitaxial growth of semiconducting poly(3-hexylthiophene) on carbon nanotubes as building blocks of nanoscale electronics”. Phys. Chem. Chem. Phys., 16, 19122—19129, 2014. [10] A. Chunder, J. Liu, L. Zhai, “Reduced Graphene Oxide/Poly- (3-hexylthiophene) Supramolecular Composites”. Macromolecular Rapid Communication, 31, 380–384, 2010.

3 P.59

SETS2016-15 The Southwest Emerging Technology Symposium 2016 DEVELOPING A CONTROLLER FOR A COPPER WIRE EMBEDDING TOOL J. F. Motta, D. Espalin, R. Wicker W.M. Keck Center for 3D Innovation, The University of Texas at El Paso(UTEP), El Paso, TX 79968, USA

Keywords: additive manufacturing, wire embedding ABSTRACT 3D printing electronics is a great solution for rapid prototyping. Nevertheless, 3D printing electronics involves having a dependable conductor such as copper to unite all electrical components such as resistors, capacitors, diodes, etc. Presently, conductive inks are used to connect electrical components, but inks have several defects. A better approach is using solid copper wire as a conductor having greater conductivity and rated for higher currents when compared to inks. A copper wire embedder is a tool that when attached to a robotic arm or CNC (computer numerical control), copper wire is introduced into a plastic surface and embedded on the surface of the plastic. In order to embed wire successfully all controls of the wire embedding tool should coordinate with minimum delay. This research was focused on developing a controller unit for the wire embedding tool and was created in order to directly communicate with the robot arm or CNC. The controller unit was a success with minor modifications in order to establish better placement of copper wire. 1. Introduction Additive manufacturing is a process which consist of building parts layer by layer that helps reduce lead time when prototyping mechanical and esthetical [1]. 3D printing electronics helps with the prototyping phase of the engineering design process by rapidly creating the desired part and iterating without spending time outsourcing PCB boards. However, the production of a 3D printed electromechanical functional prototype is not easy. Currently there are no commercially available printers that can create these type of specialized parts. The only 3D printer that is able to print electronics is Voxel8 which uses conductive ink in order to connect electrical components. Conductive inks seem like a great solution to printing 3D electronics, but inks have high resistance leading to problems such as overheating and voltage drop, reducing in reliability and performance. The objective of this work is to develop a copper wire embedder control unit that will obtain signals from a robotic motion system and convert the signals to commands such as driving and cutting wire. A control unit utilizing existing modules, such as a motor controllers and relays, was developed to receive signals of 5V input from a robotic system. 2. Materials and Methods A wire embedding tool head was developed and mounted on a robot arm. Once the printed part was completed copper wire was embedded to the surface of the plastic part. In order to achieve a satisfying embedded wire, all electrical components within the controller need to respond quickly without any time delay. A cutting mechanism using a solenoid actuator and two DC motors must

P.60

The Southwest Emerging Technology Symposium 2016 coordinate with exact timing once an output from the robotic arm or CNC is activated. A keypad was programmed into the Arduino in order to accept inputs from a user for adjusting motor speed and the ability to control the wire embedding tool without any robotic arm. Feedback from a liquid crystal display helps the user know what command was selected. Figure 1. shows the completed assembled wire embedding tool that attaches to the robotic arm along with the placement of the electric component. A 24V power supply is connected to a relay module which controls the solenoid actuator that is used to cut the embedded copper wire. One motor is Figure 1. Wire Embedding Tool used to drive wire into the heater blocks in order to heat the copper wire and embed it on the surface of the plastic component. The other motor is used to hold the spool and whenever the driver motor pulls the wire cable the spool motor rotates to reduce tension. Two Arduinos were connected via I2C communication to expand the number of inputs and outputs. The Arduino program contains “if”, “else if”, and “while” functions that enable a command whenever an input is generated. For example, if the robot arm sends a signal of 5V to the Arduino, the Arduino will execute the signal as a command to cut the copper wire. The controller unit consists of two Arduino Uno micro controllers as shown in Figure 2. Arduino is an open source microcontroller that lets the user easily program inputs and outputs [2] Arduino microcontroller was preferred over other microcontrollers because of its popularity and vast variety of examples that help with programming. Figure 2. shows the wire diagram of the completed control unit for the wire embedder. The unit consists of a 9 pin D-Sub connector which is used as an input to communicate with the robot arm using a 5V signal. For the outputs to the wire embedding tool a 15 pin D-Sub connector was used to quickly connect and disconnect the controller unit. The spool and the driving motor are driven via pulse width modulation (PWM).

15 Pin Outputs 1. Left Fan (+) 2. Left Fan (–) 3. Right Fan (+) 4. Right Fan (-) 5. Knife actuator (+) 6. Knife actuator (-) 7. Motor A (-) 8. Motor A (+) 9. Motor B (+) 10. Motor B (-) 11. Cartridge Heater R(+) 12. Cartridge Heater R(-) 13. Cartridge Heater L(+) 14. Cartridge Heater L(-) 15. Open

Figure 2. Controller Unit Schematic P.61

The Southwest Emerging Technology Symposium 2016

3. Results Figure 3 shows the completed control unit for the wire embedding tool head. This unit is capable of changing the temperature of the heater block and read the current temperature. The user is able to input commands to change motor speeds as well as control the wire embedder tool to manually drive the motors and cutter. The controller unit is able to disconnect quickly from both the tool and the robotic arm.

Figure 3. Control Unit for Copper Wire Embedding Tool

4. Conclusions Utilizing copper wire as a conductor in a plastic surface decreases the resistance and increases conductivity when compared to conductive inks. The control unit responded without any time delay to the signals sent by the robot arm to embed copper wire on a plastic surface. It is believed that accuracy is a main concern when embedding copper wire, therefore a precise motor with a motion control will help obtain better results when embedding copper wire. References [1] Gibson, I., Rosen, D.W., and Stucker, B. “Additive Manufacturing Technologies.” New York: Springer, 2010. [2] “What is Arduino?” Web. 16 Mar. 2016.

P.62

SETS2016-16 The Southwest Emerging Technology Symposium 2016

A NOVEL APPROACH TO ESTIMATE CORROSION EFFECT IN 3D PRINTED BIO MATERIAL'S USING TAGUCHI METHOD Sai Dhiresh Kilari1, Nam-Soo Kim1, Tzu-Liang (Bill) Tseng1, Department of Industrial, Manufacturing and Systems Engineering, 2 Department of Metallurgical, Material and Bio Engineering University of Texas at El Paso, TX 79902, USA 1

*Corresponding author ([email protected])

Keywords: Bio Material, Taguchi Method, Polylactic acid, Polycaprolactone, pH ABSTRACT In the present work, the Taguchi method, from the design of experiment (DOE) has been used to optimize the dip testing method to test the weight loss effect of bio degradable materials such as PLA (Polylactic acid) , PCL (Polycaprolactone) and mixture of PLA and PCL. The optimum conditions providing the lowest corrosion rate potential were estimated and the optimum conditions were to be found out. Under what conditions, there is less corrosion is to be found out. As a result of Taguchi analysis in this study, the factor with the most influencing parameter on the corrosion resistance is to be known to show most significant effect. The percentage contributions of pH (measure of how acidic/basic water is), sample and temperature to the corrosion rate are to be seen respectively. Consequently, the Taguchi method was found to be the best promising technique to obtain the optimum conditions for such studies. Moreover, the experimental results obtained confirm the adequacy and effectiveness of this current approach. 1

Introduction

The use of Bio materials in both medical and research fields have been used for about fifty years and they have experienced steady and strong growth in both bio chemistry and material science [1] , The main objectives of the research are, to calculate the corrosion in terms of weight change of the bio materials using immersion testing method and to estimate the best optimized parameters to control corrosion behavior on bio materials using Taguchi method. Bio materials are the materials used to make devices to replace a part or a function of a living body in a reliable and safe manner. Weight loss effect on polymers are often hard to discover, the material may look normal but can in fact be embrittled and have lost its mechanical strength or some of its chemical properties such as its roughness, strength etc. Bio degradable polymeric materials have wide applications; therefore, there are many factors that can lead to degradation or corrosion in these materials. Because the lifetime of a polymeric material cannot be accurately foreseen in a specific corrosive atmosphere, it is necessary to clearly understand the compositions and reaction mechanisms of polymeric materials. Hence, dip testing is done and Taguchi method is used to determine the statistical data to prove the best bio material which has less weight loss effect or degradation rate compared to other materials. The best way to improve or estimate the quality of bio parts is by using a statistical model approach such as Taguchi method 2

Methodology

2.1 Immersion Test The Bio materials used in the process are PLA 100%, PCL 100% and a mixture of PCL 50% and PLA 50% these are made to react with the chemical pH of values ranging pH 1, 3 and 5. In order

P.63

to obtain the pH value, the distilled water is added with certain amounts of acidic or basic solution in order to get the required pH value and the pH meter is used to test the solution. Once the required pH values are obtained, the liquid solution is used for the test of the materials .All acidic solutions have pH values less than 7 and the basic solutions have their pH values greater than 7. The parts are dipped into the solution for 7 days and dried out. The weights of the samples are noted down before and after dipping them into the solution. 2.2 Taguchi Method A statistical model is developed using Taguchi method to improve the quality of material by estimating the corrosion behavior and likewise to determine specific material mixture and parameters required. As we have three variables with three parameters we choose a L9 method [3] . The three variables used with three parameters (Independent Variables) 1. Temperature (25, 50, 75 in degrees Celsius) 2. Samples are (100% PCL, 100% PLA, 50-50% PCL PLA) 3. pH value (1, 3, 5) And the response variable is weight change Determining the number of factor levels

Selecting the Orthogonal array

Determining Signal to Noise (s/n) ratio

Analysis of variance

Obtain the optimum conditions

Fig.1. Conceptual Framework for Taguchi Method

3. Results and Discussion As we see in Fig.2. the data is entered into the L9 Taguchi design matrix S no 1 2

pH 1 1

Bio Sample PCL 100 PCL-PLA 50-50

Temperature 25 50

Weight Change 0.019 0.023

3 4 5 6 7 8

1 3 3 3 5 5

PLA 100 PCL 100 PCL-PLA 50-50 PLA 100 PCL 100 PCL-PLA 50-50

75 50 75 25 75 25

0.030 0.023 0.030 0.005 0.030 0.003

9

5

PLA 100

50

0.006

Fig.2. Design and data for the Taguchi design matrix

As seen in Fig.3. The plot between data means and mean of means show the effect of response data. As we choose lower the better or smaller the better pH 5 has much less effect on the bio sample, so as at 25 degree Celsius temperature it shows less effect. The Bio sample PLA 100% pure has good weight loss resistance compared to other PCL and PLAPCL composition.

P.64

Fig.3. Main effect Plot for means versus Mean of Means Plot

Fig.4. Analysis of variance (ANOVA) from Minitab

As seen in Fig 4 both pH and Temperature are significant at 95% confidence interval from general linear model. The higher the R2 value, the better the model fits. The results of ANOVA indicate that R2 of 99.85% means the model is best fit. Further work can be done to identify signal to noise ratio (s/n ratio) by obtaining more results at different levels and making the values more precise to achieve a better regression model.

References [1] Ratner.BD, Bryant.SJ. “Biomaterials: where we have been and where we are going”. Annu Rev, 2004. [2] Ali.Fathima, Sabirneeza.Abdul Rahiman, Subhashini.Sethumanickam,"Corrosion inhibition, adsorption and thermodynamic properties of poly (vinyl alcohol-cysteine) in molar HCL”.Arabian Journal of Chemistry. [3] Selden PH., "Sales Process Engineering, A Personal Workshop". ASQ Quality Press, (1997). [4] Sung-HoonAhn, Caroline S. Lee, WoobyokJeong,” Development of translucent FDM parts by postprocessing”. Rapid Prototyping Journal, Vol. 10, Iss: 4, pp. 218 – 224, 2004. [5] T.Boothby. and C.Bakis. “Durability of externally bonded fiber-reinforced polymer (FRP) composite systems. Strengthening and Rehabilitation of Civil Infrastructures Using Fibre-Reinforced Polymer (FRP) Composites”.pp.292-322.,2004.

P.65

SETS2016-17 The Southwest Emerging Technology Symposium 2016 Economic analysis between powder bed-based additive manufacturing technologies J. A. Gonzalez1 2*, J. Mireles1,3, Y. Lin3, R. B. Wicker1,3 1

2

W.M. Keck Center for 3D Innovation, University of Texas - El Paso, 79968, USA Department of Metallurgical, Materials and Biomedical Engineering, University of Texas - El Paso, 79968, USA 3 Department of Mechanical Engineering, University of Texas – El Paso, TX 79968, USA * Corresponding author ([email protected]) Keywords: additive manufacturing, electron-beam melting, binder jetting, selective laser melting

ABSTRACT With the growing advancements in additive manufacturing (AM), an increasing number of companies are moving toward the utilization of AM for the fabrication of metal components that are re-engineered for weight savings and improved performance. While growth has been evident across all AM platforms, the technologies able to produce metallic parts have been the ones to receive the most widespread interest. Although a diverse set of platforms exist for metallic part fabrication, the ones that will be evaluated in this paper are powder bed-based technologies available today; which include (1) electron beam melting, (2) selective laser melting, and (3) binder jetting. The analysis described here will compare the energy use and cost efficiency across powder bed platforms to better understand the resources required for direct metal manufacturing using powder bed-based technology. The three technologies have been evaluated in the required equipment, energy consumption during fabrication, as well as cost of post-fabrication steps needed to achieve a final product. Introduction Additive manufacturing (AM) is an advanced manufacturing process that creates components in a layerwise fashion through the selective addition and fusion of material, as opposed to the removal of material that is done by traditional means, such as using a machining process. Although growth has been evident across all AM platforms, the technologies able to produce metallic parts have been the ones to receive the most widespread interest. As further refinements are made to AM technologies, manufacturing of once machined parts will be replaced by products that are additively manufactured, and as a result, have enhanced performance and weight reduction enabled by engineering intricacies previously impossible to achieve through traditional means. Richard D’Aveni (2013) states that “Indeed, the rise of 3D printing and additive manufacturing will replace the competitive dynamics of traditional economies-of-scale production with an economies-of-one production model, at least for some industries and products” which can indicate that more companies will look to incorporate AM technology in a production environment with hopes of enhancing economic dynamics for products of choice. A noteworthy example is GE’s initiative to implement powder bed fusion technology for production of next-generation fuel nozzles. It is expected that an increased number of companies will take a similar approach and that powder bed-based technologies will be at the forefront of this growth. An in depth economic study is needed to gain a better understanding of the resources needed to implement powder bedbased technology for part production. The three main technologies in powder bed AM that utilize metal precursor powder as a means for part production are electron beam melting (EBM), selective 1 P.66

Southwest Emerging Technology Symposium 2016

laser melting (SLM), and binder jetting. This paper focuses on comparing the three powder bedbased technologies in terms of cost efficiency and energy consumption. Methodology Electron-beam melting EBM is a well-developed powder bed fusion technology for part fabrication where metal parts are created using a high-energy electron beam that is selectively directed onto powder precursor material, which is deposited onto a substrate in a layerwise fashion. Metal powder precursor material is melted using an electron beam at roughly 30mA in a vacuum chamber of up to 10-4Torr [1]. Although EBM can theoretically process any conductive material, the standard alloys that have been commercialized for EBM include: Ti-6Al-4V, Ti-6Al-4V ELI, Titanium grade 2, Inconel 718, and cobalt-chrome ASTM F75. EBM can be described as a three-step process involving: 1) powder is spread using a raking mechanism (for Ti-6Al-4V the layer thickness ranges from 50-70µm) [1]; 2) the electron beam strikes the powder with high beam power and high beam speed to achieve sintering; 3) the powder is melted using a lowered beam speed and power to achieve a fully dense layer; 4) the substrate is lowered one layer thickness and the three steps repeat for every layer until part fabrication is complete. There are several advantages to using EBM for the fabrication of metal parts [1], with the most prominent being the high temperature environment which significantly reduces residual stresses and achieves mechanical properties directly post-fabrication that are comparable or better than properties of wrought or cast parts. The most widespread adoption of EBM technology has been biomedical implants where more than 30,000 EBM-fabricated implants have been used for patients requiring hip replacement. The Arcam A2x, provided by the single supplier of EBM technology, was evaluated along with necessary peripherals that are required to achieve a final product (e.g. powder recovery system, chiller, powder vacuum, etc.). Selective laser melting Selective laser melting (SLM) is similar to EBM, where both technologies are considered a class of powder bed fusion technology that selectively processes precursor powder in a layerwise fashion using a high-energy source. However, the similarities stop there, where unlike EBM; SLM uses the energy of a laser beam to melt powder particles (typically 10-45 m in size compared to powders used in EBM that are typically 50-150 m). The finer powder utilized for fabrication with SLM results in parts with improved surface finish. The process takes place in an inert gas environment, which results in parts with high residual stresses that require a post-process heat treatment to achieve a final product [3]. SLM products are being used for applications ranging from dental prostheses to turbine blades and have been applied in a production environment by companies like GE to produce next-generation fuel nozzles. A SLM Solutions 280HL system was considered in this study as well as necessary peripherals that are required to achieve a final product (e.g. powder removal system, band saw, chiller, etc.). Binder Jetting The binder jetting process is very different from powder bed fusion, where instead of a high-energy beam, it uses a piezo-driven device to selectively deposit a binder substance onto a layer of precursor powder particles. The process involves the use of precursor powder material that is added to a powder bed in a layer-by-layer fashion using a stationary roller. The precursor powder is selectively bonded using a piezo-driven device that deposits a binding substance as 2 P.67

Southwest Emerging Technology Symposium 2016

directed by instructions from a computer-aided design (CAD) model. After the binder is deposited, the powder bed moves below a heater that helps cure the deposited binder. Once the heating cycle is complete, the part is prepared for a new layer and the process continues until part fabrication is complete. The part is then transferred to an oven, which helps to fully cure the part by burning off the binder used during the fabrication process. The resulting product is a rigid green body that requires careful handling. If an end-use part is needed that can mechanically compete with parts produced via powder bed fusion, the green body needs to be sintered using the appropriate sintering profile for the material being used [2]. Applications of binder jetting range from sand molds and cores for casting purposes if a green body is produced using sand, to filter and prosthetics is the necessary post-processing of sintering is achieved. For this study, the ExOne M-Flex machine was used for evaluation as well as any necessary peripherals that are required to achieve a final product (e.g. oven, powder removal system, furnace). Results The three technologies being compared here have different process chains to achieve a final product, which resulted in different equipment requirements. Figure 1, shows that the initial purchase of the manufacturing system and its corresponding peripheral equipment, where binder jetting was identified as being the least expensive (at around $448,000USD) when compared to SLM ($700,000) and EBM ($1,066,000). In order to compare energy consumption, the same part geometry was considered for part fabrication across the three platforms. For a single part build with an area of 1,114.27 cm the amount of energy and powder on plotted in Figure 2. A large difference in energy consumed was evident, with binder jetting prevailing as the low cost option . Printer Costs

Printer Model

Arcam A2

EX One M-Flex

SLM 280HL

$0

$200

$400

$600

$800

$1,000

$1,200

Price (Thousands) Base Price + Additional Components

Base Price

Figure 1. Powder bed based AM technologies comparison in cost of equipment of fabrication of components.

3 P.68

Southwest Emerging Technology Symposium 2016

Figure 2. Energy and power consumption for powder bed based AM technologies Conclusion Based on the initial finding and analysis done, binder jetting is a low cost alternative for direct metal AM. Further analysis needs to be conducted to compare the resulting mechanical properties between technologies. The analysis is still ongoing and requires a full analysis of energy and power consumption across all peripheral hardware. References [1] E. L. Murr, M. S. Gaytan, D. A. Ramirez, E. Martinez, J. Hernandez, K. N. Amato, P. W. Shindo, F. R. Medina and R. B. Wicker , "Metal Fabrication by Additive Manufacturing Using Laser and Electron Beam Melting Technologies," Journal of Materials Science & Technology, vol. 28, no. 1, pp. 1-14, 2012. [2] S. M. Gaytan, M. A. Cadena, H. Karim, D. Delfin, Y. Lin, D. Espalin, E. MacDonald and R. B. Wicker, "Fabrication of barium titanate by binfer jetting additive manufacturing technology," Ceramics International, no. 41, pp. 6610-6619, 2015. [3] S. Breman, W. Meiners and A. Diatlov, "Selective Laser Melting A manufacturing technology for the future?," Rapid Manufacturing, no. 2, pp. 33-38, 2012.

4 P.69

SETS2016-18 The Southwest Emerging Technology Symposium 2016

DAMAGE TOLERANCE AND ASSESSMENT OF UNIDIRECTIONAL CARBON FIBER COMPOSITES M. Flores1, D. Mollenhauer2 Civil Engineering Department, University of Texas – El Paso, El Paso, TX 79968, USA; 2 Manufacturing and Materials Directorate, Air Force Research Lab, WPAFB, Ohio, 45435 1

Keywords: composites damage tolerance NDE modeling

ABSTRACT Composites are beginning to grow in structural applications throughout industry. However, certification and damage tolerance is a growing concern in many aerospace and marine applications. Damage tolerance design requires that composite structures have the capability to withstand its design limit load even with the presence of damage. However, determining what type of damage is critical to the structure’s integrity has been limited to delamination and matrix cracks. Non-destructive techniques (NDT) will be utilized to quantify the amount of damage caused by impact experiments. A three-dimensional reconstruction of the images obtained from NDT (X-Ray CT, C-Scan) will be used to pre-impose morphology, delamination, and transverse matrix cracks into a discrete damage model. The objective of this study is to create a high fidelity model to analyze the evolution of damage of post impacted composites under compression. 1

Introduction

The process of integrating experimentation, computational analysis, and nondestructive evaluation is a challenging task. Individually, each discipline comes with sets of assumptions and limitations requiring extensive amounts of research. Performing an experiment heavily depends on the boundary conditions that were applied. However, what is an appropriate boundary condition? Computational modeling provides insight into the boundary conditions to captured the physics of the experiment. However, the underlying boundary conditions are only as good as what the model can predict to an experimental outcome. Nondestructive evaluation provides visualization and understanding of the failure mechanism of the problem. However, each nondestructive technique has limitations on what it can capture in the form of damage. An ill posed problem could arise if a wrong assumption about the failure were applied to the model. Damage tolerance design requires that composites have the capability to maintain load with the presence of damage. Computational models that could predict the damage growth of impacted composites have not been studied to any microscale, mesoscale and component level. The purpose is this study is to integrate nondestructive evaluation into modeling efforts to predict the behavior of failure for impacted composite laminates while under compressive loading. 2

Discrete Damage Modeling

The Discrete Damage Modeling (DDM) approach proposed herein for modeling networks for multiple parallel transverse matrix cracks within individual plies of a laminate and delaminations between plies couples a regularized Mesh Independent Crack (MIC) modeling

P.70

technique [1,2], the step function used in traditional x-FEM approaches to construct local enrichment for a crack discontinuity is replaced with a continuous function that is approximately by the same shape functions as those used for initial displacement approximation. The surface of each crack is replaced with a gradient zone (a volume where the gradient of the approximate step function is nonzero), and the surface fracture energy is replaced with the cohesive energy in the gradient zone. A simulation begins without any initial matrix cracks. As the load increases, matrix cracks are inserted according to a failure criterion. In the present paper, the LaRC03 in-plane stressbased failure criterion [3], as well as the 3D LaRC04 criterion [4], are used. The criterion is evaluated at each integration point of a discretized element and, if the criterion is exceeded, a matrix crack oriented in the fiber direction is added. LaRC04 predicts the crack angle with respect to ply surface, which is 90° under tensile normal stresses whereas in shear and compression dominated loading states can significantly differ from 90°. The crack is inserted using the displacement enrichment necessary to model the displacement jump. The magnitude of the jump is initially zero and is controlled by an interface cohesive law developed by Turon [5]. The same cohesive law is used at the ply interfaces to represent potential delamination surfaces. Fiber failure uses a Progressive Fiber Failure (PFF) methodology based on the uniform degradation of the element stiffness when the fiber failure mode is detected [6,7]. The PFF method continuously degrades stiffness in a given integration point as a function of normal strain in the fiber direction. The degradation begins after a threshold value corresponding to the tensile strength of unidirectional coupons is reached. A Newton-Raphson procedure is applied to find the equilibrium solution at each load step of the implicit incremental solution combining softening mechanisms from cohesive interface matrix damage models and stiffness degradation from fiber failure models. 3

Integrating NDE Into Modeling

Damage from an impacted composite can come in the form of permanent deformation, transverse matrix cracks, delamination and fiber breakage. The process starts by incorporating realistic damage into a discretized damage model for a composite laminate flat plate. The morphology of the deformation was obtained from profilometry and superimposed onto the discretized elements. Filters and smoothing were done to the line profiles to ensure the discretized elements were geometrically stable. The discretized nodes conformed to a fitted linear interpolation of the impacted and back surface line profiles. Delamination was obtained through analyzing immersion ultrasound and X-ray CT images. The benefit of using immersion ultrasound is that it is a relatively quicker technique and provides quantitative information on the delamination. However, it does not provide the spatial representation necessary to identify which interface has the correct delamination pattern. Speckle noise from X-ray CT images makes it hard to differentiate where delamination is for individual interfaces. Therefore, a combined approach was used to extract delamination patterns for each interface. The interfacial patterns were reconstructed into a 3-dimensional STereoLithographic (STL) image. The STL image selected all the nodes and elements within the damaged region and assigned them weak interfacial mechanical properties. The location, angle and orientation of transverse matrix cracks were obtained using x-ray CT. The cracks were prioritized based on it relative location to delamination.

P.71

4

Conclusion

The process of pre-imposing damage into a discrete damage model has been explored. Figure 1 shows how profilometry was used to artificially pre-impose morphology into the model. Figure 2 shows a comparison of the delamination pattern of the simulation (3D) and immersion ultrasound. Figure 3 illustrates how information for the matrix cracks was obtained from of an X-ray CT image and incorporated into the simulation. The technique shows remarkable potential to identify critical damage and its effect on the structure under loading. Nondestructive evaluation was used to assess the amount of damage in a composite structure. The next step is if it can accurately predict how damage growth and ultimate failure. Identifying and modeling critical damage could improve the cost of certifying composite materials. The methodology isn’t limited to composite materials, but it could be used in any application from manufacturing, micromechanics, and infrastructure. Impacted Surface Line Profiles Panel−A 0.04 0.02 0

Height (mm)

−0.02 −0.04 −0.06

Average

−0.08 −0.1 −0.12 −0.14 −40

−30

−20

−10

0 x−axis (mm)

10

20

30

40

b) Simulated pre-imposed morphology Fig.1. Pre-imposing morphology into a simulation using optical profilometry

a) Optical profilometry of morphology

a) Immersion Ultrasound b) Simulation Fig.2. Pre-imposing delamination into a simulation using x-ray CT and ultrasound (3D)

3 P.72

b) Simulation

a) X-ray CT

Fig.3. Pre-imposing matrix cracks into a simulation using x-ray CT images (3D)

References [1] M.J. Swinderman, E.I. Iarve, R.A. Brockman, “Strength prediction in open hole composite laminates by using discrete damage modeling,” AIAA Journal, 51(4), 2012, p936-945 [2] M.J. Swinderman, E.I. Iarve, R.A. Brockman, “Strength prediction in open hole composite laminates by using discrete damage modeling,” AIAA Journal, 51(4), 2012, p936-945 [3] C.G. Davila, P.P. Camanho, C.A. Rose, “Failure criteria for FRP laminates,” Journal of Composite Materials, Vol. 39, No.4 (2005), p323-45 [4] S.T. Pinho, C.G. Davila, P.P. Camanho, L. Iannuci, P. Robinson, “Failure modes and criteria for FRP under in-plane or three-dimensional stress states including shear nonlinearity,” Hampton, VA February 2005, NASA/TM-2005-213530 [5] A. Turon, P.P. Camanho, J. Costa, C.G. Davila, “A damage model for the simulation of delamination in advanced composites under variable-mode loading,” Mechanics of Materials, Vol. 38, No. 11, Nov. 2006, p1072-89. [6] P.P. Camanho, P. Maimi, C.G. Davila, “Prediction of size effects in notched laminates using continuum damage mechanics,” Composite Science and Technology, 2007;67, p2715-2727 [7] P. Maimi, P.P. Camanho, J.A. Mayugo, C.G. Davila, “A continuum damage model for composite laminates: Part I – constitutive model,” Mechanics of Materials, Vol. 39, No.10, 2007, p897-908

P.73

SETS2016-19 The Southwest Emerging Technology Symposium 2016 DESIGN AND TESTING OF HYBRID COMPOSITE MATERIALS FOR CRYOGENIC FUEL TANKS R. Avila1, Md. Islam1, P. Prabhakar1* 1 Department of Mechanical Engineering, University of Texas, El Paso, TX 79968, USA; * Corresponding author ([email protected])

Keywords: Textile Composites, Propellant Tank, Cryogenics, Gradient Temperature, Damage ABSTRACT Cryogenic fuel tanks carry and store cryogenic propellants (oxygen, methane, hydrogen) at subfreezing temperatures in order to generate highly combustible liquids. This tank is exposed to an extremely cold temperature in its interior and to ambient temperature on its external surface creating a large temperature gradient across the wall thickness. Hybrid textile composites with carbon-Kevlar® fabric are explored as means to reduce the thermal gradient effect and enhance the material performance under cryogenic conditions. Previous studies indicated that carbon-Kevlar® textile composites are suitable materials for cryogenic temperatures (77 K). The pristine mechanical properties of carbon composites changed a maximum of 3-4% after cryogenic exposure, while ≈17% for Kevlar® composites. Computational models of hybrid carbon-Kevlar® composites were subjected to cryogenic temperature for extended periods to study the thermal gradient and optimize the design of layups. Six combinations were selected that resulted in low interface stresses and fewer number of peak stresses through the composite thickness. Subsequently, analysis of cryogenic exposure to one surface of hybrid composites was performed computationally to simulate the composite wall containing the liquid fuel. Computational and experimental work was conducted to determine the optimum composite wall designs. An ABS plastic insulating holder was designed and 3D printed to hold the specimens such that only one surface is exposed to LN2. Eight composite layups were exposed to LN2 using the holder to study their response to thermal gradient cryogenic exposure. Based on the results obtained computationally and experimentally, optimum hybrid layups of composites for cryogenic exposure were determined. 1 Introduction Composite materials have higher strength to weight ratio which makes it suitable for aerospace applications where weight is very important. Eventually, there will be a need for reusable space launch vehicles with the capability to complete at least 100 missions. These new vehicles could use composite cryogenic fuel tanks that could lead to 40 % reduction in the weight of the cryogenic fuel tank compared to metallic fuel tank and a 14 % reduction in the overall vehicle weight.”[1], [2]” NASA's Composite Overwrapped Pressure Vessel (COPV) (Fig. 1a) is a well-established cryogenic tank with metallic liner and continuous fiber/matrix system wrapped around it (Fig. 1b). The main purpose of liner in cryogenic tanks is to prevent or minimize the permeation of cryogenic liquid through the walls of the external structure as well as minimize thermal stresses.

P.74

Fig. 1 (a) NASA's Composite Cryogenic Fuel Tank, (b) Schematic of Fiber-Reinforced Laminate Overwrap and (c) Schematic of Textile Composite Pressure Vessel

2 Methodology A preliminary study by Islam et al “[3]”.demonstrated that woven carbon and Kevlar® fiber composites have the potential to be used as cryogenic tank materials. In this study, the focus was on investigating the hybrid design of carbon and Kevlar® fabric for minimizing possible interface stresses that cause delamination and in-plane stresses. Towards that end, a 24 layer composite was considered for analysis with 5 different layer percentages and 10 different layer combinations for each layer percentage, except for the 100 percent combinations. The computational modeling framework was divided into the following two steps: (1) Determine the temperature distribution in the model; (2) Determine the stress distribution in the model. That is, a heat transfer analysis was conducted first on a layered model to determine the temperature distribution upon cryogenic exposure; followed by a stress analysis with temperature distribution as an input to explore its influence on the stress distribution in the layers and at their interfaces.

5

Conclusion

Fig 2. 3D model with dimensions.

3 Experimental Procedure

Table 1. Mechanical properties of carbon and Kevlar fiber lamina.

Composite plates were manufactured using Vacuum Assisted Resin Transfer Molding (VARTM) process. Specimens cut from these plates were then exposed to cryogenic environment followed by short beam shear (SBS) tests at room temperature to determine its

P.75

DESIGN AND TESTING OF HYBRID COMPOSITE MATERIALS FOR CRYOGENIC FUEL TANKS

inter laminar shear stresses (ILSS). Dry fabrics of woven (plain weave) carbon fiber and Kevlar® fiber tows were used as reinforcement with Epon 862/ EPIKURE 9553 hardener as the matrix material. Composite specimens were exposed to cryogenic environment by submerging them in a container filled with LN2 at a temperature of 77 K (-196oC). Short beam shear (SBS) tests were conducted next to measure the inter laminar shear strength (ILSS) of the composites. Twelve samples of each type (four room temperature, four after cryogenic exposure, and four gradient exposure) were tested. 4 Results and Discussion Combination

C24 KCKCKC14KCKCK C3K3C12K3C3 C6K12C6 C4K6C4K6C4 K9C6K9 CKCKCK14CKCKC K24

Condition

# of Carbon Layers

Pristine Cryogenic Gradient Exposure Pristine Cryogenic Gradient Exposure Pristine Cryogenic Gradient Exposure Pristine Cryogenic Gradient Exposure Pristine Cryogenic Gradient Exposure Pristine Cryogenic Gradient Exposure Pristine Cryogenic Gradient Exposure Pristine Cryogenic Gradient Exposure

24 18 18 12 12 6 6 0

ILSS, MPa

39.08 ± 3.5 39.13 ± 3.3 44.37 ± 0.8 34.46 ± 2.0 33.01 ± 2.4 34.73 ± 1.7 30.67 ± 0.9 29.65 ± 1.4 31.24 ±1.3 30.90 ± 1.3 30.02 ± 1.7 27.70 ± 1.1 20.72 ± 1.4 22.75 ± 1.7 21.60 ± 1.3 23.47 ± 1.0 22.13 ± 0.7 22.52 ± 1.3 23.13 ± 0.8 22.25 ± 0.8 23.39 ± 0.2 18.88 ± 2.5 15.71 ± 2.2 18.66 ± 0.5

% Reduction

-1.3% -13.5 % 4.2 % -0.8 % 3.3% -1.9% 2.9% 10.4% -9.8% -4.2% 5.7% 4% 3.8% -1.1% 16.8% 1.2%

Table 2. ILSS and % reduction for the various composite combinations.

Fig 3. ILSS distribution for the various composite combinations.

3 P.76

In this Fig 3, 8 different hybridized composite layups were compared and divided into one 100 % carbon, two 75% carbon- 25% Kevlar, two 50% carbon-50% Kevlar, two 25% carbon75% Kevlar, one 100% Kevlar. As the volume percentage of Carbon Fiber in the composite decreases the Inter-laminar shear strength of the material decreases as well assuming a linear behavior overall the ILSS will decrease as the volume percentage of Carbon decreases or the Kevlar volume increases. The optimum hybrid 24 composite layup for cryogenic applications should have a volume fiber percentage of 75% Carbon- 25% Kevlar; and Kevlar as starting and ending fiber in the arrangement in order to mitigate the most any effects that temperature might have in the ILSS when the composite is subjected to cryogenic temperatures. 5 Conclusions and Future Work Woven carbon and Kevlar ® fiber reinforced hybrid composites exposed to cryogenic environment were computationally and experimentally investigated in this paper. Transient heat transfer analysis followed by mechanical stress analysis were conducted on different material percentages and layer stacking of the same. Six combinations from the thirty combinations computationally investigated for cryogenic and gradient exposure were chosen for experimental testing. Among the hybrid combinations tested the combination KCKCKC14KCKCK has shown the least reduction of ILSS. Further investigation is required to determine the material response of the composite to repeated cryogenic exposure. References [1]Pavlick, M. M., Johnson, W. S., Jensen, B., and Weiser, E., Evaluation of mechanical properties of advanced polymers for composite cryotank applications," Composites Part A: Applied Science and Manufacturing, Vol. 40, No. 4, 2009, pp. 359-367. [2]Kessler, S. S., Matuszeski, T., and McManus, H., Cryocycling and Mechanical Testing of CFRP for X-33 Liquid H2 Fuel Tank Structure, Ph.D. thesis, Technology Laboratory for Advanced Composites, Department of Aeronautics and Astronautics, Massachusetts Institute of Technology, 1998. [3]Islam, M. S., Melendez-Soto, E., Castellanos, A. G., and Prabhakar, P., \Investigation of Woven Composites as Potential Cryogenic Tank Materials," Cryogenics, 2015.

P.77

SETS2016-20 The Southwest Emerging Technology Symposium 2016

STRESS CORROSION CRACKING SUSCEPTIBILITY OF ALUMINUM FOILS FOR AEROSPACE APPLICATIONS

E. Garcia1, C. M. Stewart1* 1 Department of Mechanical Engineering, The University of Texas at El Paso, El Paso, TX 79968, USA; * Corresponding author: Calvin Maurice Stewart ([email protected])

Keywords: Stress corrosion cracking, Mechanical properties, Aluminum Foil ABSTRACT Stress-corrosion cracking (SCC) is a term used usually to describe the service failures of engineering materials that are caused by slow, environmentally induced crack propagation. In this experimental study we describe and examine the susceptibility of generic aluminum alloy foil to stress corrosion cracking under 3.5% w.t NaCl solution. Mechanical properties of the aluminum specimens were investigated using slow strain rate tests of 0.001 mm/min while inside an environmental chamber at a flow rate of 150 ml/min. Smooth foil samples were subjected to monotonic tensile tests. Digital microscope camera was used to observe and perform analysis on the corroded specimen surface. A comparison of stress, strain, and time results of fracture between air and 3.5% NaCl solution at room temperature were calculated. These results showed that, this generic aluminum foil demonstrated corrosion exposure and high stress corrosion cracking susceptibility. 1 Introduction Cracks propagations produced by stress corrosion cracking are the result of combined and synergistic interaction of mechanical stress and corrosion reactions [1]. Recently, there has been an interest in aluminum alloys by many industrial areas as an environment-friend material reducing environment pollution. Al alloys have been widely used in aerospace industry because of their high specific strength -to-density ratio. To deal with challenges such as mass air transportation, the improved damage tolerant properties are needed such as fatigue performance, fracture toughness, and stress corrosion cracking (SCC) resistance [2]. SCC has been drawing attention of researches mainly because it limits the applications of these aluminum alloys significantly due to being susceptible to this behavior. With the rapid development of modern technology, foil metals have found applications in a variety of areas. The mechanical behavior of these materials may be different from that of bulk materials due to size effects. Numerous amount of studies have been conducted to develop corrosion technologies such as cathodic polarization and grain refinement, and other methods of corrosion protection. Marine structures are always subjected to both corrosive environment as well as mechanical loading. The purpose of this experiment is to describe and examine the susceptibility of generic aluminum alloy foil to stress corrosion cracking under 3.5% w.t NaCl solution. 2 Experimental Procedures Methodology There has been different techniques to assess stress corrosion cracking (SCC) susceptibility of materials, however to provide a reasonable and fair estimate of SCC susceptibility, the slow strain rate test (SSRT) was conducted [2-3]. The specimens were made out of commercial generic aluminum foil. The design of these specimens were bone samples and their dimensions were as

P.78

The Southwest Emerging Technology Symposium 2016 follows: gauge length 15 mm, its width 6.5 mm, and thickness 0.024 mm. Mechanical properties of the aluminum specimens were investigated using slow strain rate tests of 0.001 mm/min while inside and environmental chamber of 3.5% NaCl solution at a flow rate of 150 mm/min. Smooth samples were cut and subjected to monotonic tensile tests until fracture in ambient air and under the solution. The SCC susceptibility was calculated by Iσ, Iδ and It, which were defined as the loss of tensile strength, elongation, and time respectively, to failure in air and 3.5% NaCl solution environments with SSRT, as shown below: I

(

a

c

) 100%

(1)

a

I

(

a

c

) 100%

(2)

a

t a tc (3) ) 100% ta where 𝜎𝑎 , 𝛿𝑎 , and 𝑡𝑎 are tensile strength, elongation, and time to failure in air, respectively; and 𝜎𝑐 , 𝛿𝑐 , and 𝑡𝑐 are in solution. The larger the relative loss the higher SCC susceptibility [3]. It

(

Procedure The Al alloy specimens were cut out of the aluminum foil sheet using a 3D-printed polylactic acid guide specimen. Afterwards, each specimen was measure to verify the dimensions and adjust the stress and strain calculations accordingly. The specimens were carefully mounted into the chamber and gripped without any tension to avoid false force readings and start the tests stress-free. The chamber was then filled up at a flow rate of 100 ml/min to reduce any risk of bending the foil or causing any undesired stretch in the foil that could spoil the positioning. Data was acquired from the software provided by the WinTest Digital Control System, where load, displacement, time, and gear flow rate outputs are displayed. An in-line load cell sensed the load and the actuator displacement is sensed using a linear variable displacement transducer (LVDT). Experiments were setup and carried out at 0.001 mm/min for a duration of around 15 hours. 3 Results and Discussion In Fig 1 the stress vs strain curves are shown for the monotonic tension tests of three Aluminum samples in air. The first cracking failures occurred between strain of 0.045 mm/mm and 0.050 mm/mm and they reached an ultimate tensile strength (UTS) between 63 MPa and 66 MPa. These samples experienced a notch failure at this point and the crack propagated until another crack appeared on the opposite edge of the specimen, this behavior was present in all of the Al samples. There wasn’t a brittle fracture as it is observed in how the curve didn’t go immediately to zero and instead it was stepping down due to cracks until the complete detach of the specimen. Fig 2 shows the stress vs strain curves of two Al samples tested in 3.5% NaCl solution. There was a dramatically drop in the ultimate tensile strength and the strain at fracture. The UTS of sample 1 and sample 2 were 10 Mpa and 18 MPa respectively. Fig 3 represent the surface of the corroded aluminum and the salt crystals formed on it. Oxidation process occurred rapidly on this aluminum material as seen in Fig 4 where the failure crack happened and even deteriorated the surface of the sample. From Eq. (1-3), stress corrosion cracking susceptibility was evaluated by the relative losses of tensile strength Iσ, elongation Iδ, and time to failure It.

P.79

The Southwest Emerging Technology Symposium 2016 Table 1. Tensile properties and stress corrosion cracking susceptibility of aluminum foils by SSRT.

Sample No. 1 2

In air 𝜎𝑓 /Mpa

𝛿𝑓 /mm

66.14 66.01

0.678 0.610

Fig 1. Relation curve of Al in air

In 3.5% NaCl Solution 𝑡𝑎 /hr

11.3 10.0

𝜎𝑓 /Mpa

10.48 18.48

𝛿𝑓 /mm

0.105 0.313

𝑡𝑎 /hr

1.7 5.2

Relative Loss Iσ/%

85.44 71.99

Iδ/%

It/%

84.51 48.68

84.8 47.92

Fig 2. Relation curve of Al in 3.5% NaCl

Fig 3. Salt crystals formed on corroded Al at 470x Fig 4. Edges of corroded Al at 470x

4 Conclusion The purpose of this experimental study was to analyze the stress corrosion cracking susceptibility of aluminum foil. Monotonic tension tests under slow strain rate were performed and mechanical properties were evaluated and demonstrated that aluminum foil corroded quickly under flow of 3.5% NaCl solution. Concluded from data table and plots there was a dramatically decrease in stress, elongation, and time failure properties after solution tests. References

[1] R.H. Jones. “Stress-Corrosion Cracking”. 1st Edition, ASM International, 1992. [2] G.M. Ugiansky. J.H. Payer “Stress corrosion cracking – The slow strain-rate technique”. ASTM Committee G-1on Corrosion of Metals in cooperation with the National Association of Corrosion Engineers TPC Committee T-3E on Stress Corrosion Cracking of Metallic Materials, Toronto, ON, ASTM Special Technical Publication 665, 1979. [3] H. SHE, W. CHU, D. SHU, J. WANG, B. SUN0 “Effects of silicon content on microstructure and stress corrosion cracking resistance of 7075 aluminum alloy”. Transactions of Nonferrous Metals Society of China, Vol. 24, pp 23072313, 2014.

P.80

SETS2016-21 The Southwest Emerging Technology Symposium 2016

DESIGN OF AN OPTICALLY ACCESSIBLE HIGH INTENSITY TURBULENCE COMBUSTION SYSTEM 1

A. Acosta-Zamora1, A. de la Torre1, A. Choudhuri1* Department of Mechanical Engineering, The University of Texas at El Paso , El Paso, TX 79968, USA; Keywords: combustor, turbulence, optical diagnostics, design

ABSTRACT This paper focuses on the design of a high intensity turbulent combustion system. The experimental setup consists of a combustor, delivery lines, and lasers diagnostics systems. A backward-facing step flame stabilization as well as optical accessibility features for flow diagnostics drove the combustor design. The combustor was designed for compressible flow (M = 0.3), and a maximum operating pressure of 6 bar. A grid, or perforated plate, is used as the turbulence generator for the experiment. Optical access is provided via the use of quartz windows on three sides of the square combustion chamber. Air and methane delivery lines feed the gases to a premixing chamber before ignition at the combustor. 1

Introduction

The study of high intensity turbulent combustion has been considered an important part of research in recent years. The focus of the project is to study the structure of premixed flames on compressible and high intensity turbulent flow. The use of a backward-facing step for flame stabilization was used effectively by Takahashi and Schmoll [1]. This paper seeks to provide experimental information on turbulent flame structures, specifically the thickened flame regime. Information on the different flame structures and flame regimes are illustrated in literature [2]. The following requirements were established when designing the combustor: System

Requirement

Combustion chamber

Optically accessible Grid turbulence generator Changeable step size (variable dim.) Max Pressure: 6 bar ( 87 psi)

Reactants

Air / CH4

Air

Velocity Mach ~0.3

CH4

Flow rate: Variable (based on equivalent ratio)

Inlet pressures

100 - 600 kPa (14 - 87 psi)

P.81

2

System Description

2.1 Combustion chamber and Laser Systems The combustor section consists of several components; the combustion chamber, an entrance and exhaust regions, a mixing section, a backward facing step, and a grid. The combustion chamber geometry was defined based on existing laboratory flow rate capabilities. The dimensions of the chamber satisfy compressibility conditions of M =0.3. Fluid flow simulations and literature background defined the geometry of the backward facing step. A backward facing step is a sudden increase in flow area which introduces a recirculation zone in the flow field. This feature has been used for flame stabilization in combustion experiments. ANSYS computational simulations were used to analyze the structural integrity; thermal deformation, principal stress and strain components, which further defined the geometry of the chamber design. In order to avoid velocity fluctuations in the flow, the entrance section was designed based on a fifth degree polynomial curve found experimentally [3]. Air and methane delivery lines feed the gases to a mixing chamber prior to entering the combustor for ignition. A hydrogen-air pilot flame, fed through a separate subsystem, serves as the ignition source for the premixed air-methane mixture entering the chamber. This pilot flame is located close to the step, at the expected recirculation gas area. This recirculation zone provides a low gas velocity area, making it an ideal ignition point. For ignition, air is set to the desired test flow rate and methane is slowly introduced until target test flow is achieved and a flame exists inside the combustor. The pilot flame remains on until the desired flow rates are reached and flame anchoring is achieved. Upon flame anchoring the pilot is shut off and the flame imaging is conducted. The combustor system is coupled with a particle imaging velocimetry (PIV) and a laser induced fluorescence (LIF) system. The PIV system is used to investigate the flow characteristics in the backward facing step and inside the combustor, while the LIF system is used to conduct flame front studies. Together these systems can provide the required parameters to define what flame regimes are being studied. Both laser are introduced as laser sheets through the top window of the combustor, while the high speed cameras are placed perpendicular to these laser sheets on either lateral side of the combustor for flame visualization. A single computer, separate from the hardware and instrumentation control computer, is used to control these lasers. Once burned, gases are routed to an exhaust system leading to the outside of the laboratory facilities. The exhaust cooling was designed around existing laboratory safety considerations such as exit velocity and temperature. Cold air and water are introduced at different points of the exhaust to dilute the exit mixture and reduce the exhaust temperature to a safe level. The cold air injection and water cooling tower were designed using momentum principles as well as spray atomization. After combustion, the hot exhaust gas out of the chamber is then cooled by direct spray of water inside a cylindrical chamber. The cooling chamber is designed to cool the exhaust gas temperature below 333K. A panoramic view of the entire assembled experimental setup is shown in Figure 1.

P.82

DESIGN OF AN OPTICALLY ACCESSIBLE HIGH INTENSITY TURBULENCE COMBUSTION SYSTEM

3

Results and conclusions

3.1 Stabilized flame An optically accessible combustion system has been designed and developed. Flow turbulent intensities of 15-30% may be achieved inside of the combustor. A maximum flow velocity of M=0.3 may be obtained at the combustor entrance. The combustion system has been coupled with a particle image velocimetry (PIV) and laser induced fluorescence (LIF) laser diagnostics systems. The system has been validated through the use of cold flow as well as reacting flow. Exhaust cooling has been tested under heat loads of up to 200 kW. Current studies focus on flame stabilization at different increasing bulk flow velocities as well as flame laser imaging. Figure 2 shows a sample of an instantaneous image of stabilized reacting flow (Re-13,500) inside of the combustor. References [1] F. Takahashi, and J. W. Schmoll, “Stabilization and Suppression of a Diffusion Flame Behind a Step”, Presentation at the Joint United States Sections/The Combustion Institute Meeting, Washignton D.C., 1999. [2] S. R. Turns, “An Introduction to Combustion: Concepts and Applications”, 2nd Edition, Mc Graw Hill, 2000 [3] A. R. Choudhuri, “Investigation on the Flame Extinction Limit of Fuel Blends”, Combustion and Propulsion Research Laboratory, Mechanical and Industrail Engineering Department, The University of Texas at El Paso

Fig.1. Panoramic view of experimental setup.

3 P.83

! Fig. 2. Anchored reacting flow inside combustor (flow direction left to right)

P.84

SETS2016-22 The Southwest Emerging Technology Symposium 2016

COMBUSTION SYNTHESIS OF ZIRCONIUM DIBORIDE AND HAFNIUM DIBORIDE: THERMODYNAMIC ANALYSIS 1

S. Cordova1, E. Shafirovich1* Department of Mechanical Engineering, The University of Texas at El Paso, El Paso, TX 79968, USA; * Corresponding author ([email protected])

Keywords: Combustion synthesis, Ultra-high temperature ceramics 1

Introduction

The efficiency of coal-fired power plants could be dramatically increased by using a magnetohydrodynamic (MHD) generator in combination with a steam turbine [1]. As MHD electrodes will operate at very high temperatures in a corrosive environment, new materials that withstand these conditions are needed. Ultra-high temperature ceramics based on zirconium diboride (ZrB2) and hafnium diboride (HfB2) are promising for this application because of their high melting points (over 3200 °C), chemical stability, high electrical and thermal conductivities, and high oxidation resistance [2-4]. Recently, nanoscale ZrB2 powder has been fabricated by self-propagating high-temperature synthesis (SHS) from Zr and B with NaCl diluent [5]. The use of NaCl has decreased the combustion temperature and particle size of the formed ZrB2, which improves the properties. The synthesis from elements, however, is expensive because of the high cost of the raw materials (Zr and Hf). It would be more economical to produce these borides from ZrO2, HfO2, and B2O3 using metallothermic reduction with Mg as the reducing agent. This process can be conducted in the SHS mode, i.e., as a self-sustained propagation of the combustion front, but incomplete conversion of ZrO2 and HfO2 to borides is a major problem [6]. The objective of the present paper is to analyze thermodynamically the feasibility of achieving full oxide-to-boride conversion in a magnesiothermic SHS. Specifically, the purpose of these calculations is to evaluate how the addition of NaCl, MgO, and extra Mg to the stoichiometric (ZrO2:B2O3:Mg = 1:1:5 and HfO2:B2O3:Mg = 1:1:5) mixtures affects the conversion degree and the adiabatic flame temperature. 2

Experimental

The adiabatic flame temperature and combustion products were calculated at 1 atm pressure using THERMO software [7]. First, calculations were conducted for ZrO2−B2O3−Mg and HfO2−B2O3−Mg systems. ZrO2/B2O3 and HfO2/B2O3 mole ratios were maintained at 1 in all calculations, while the mole fraction of Mg in the initial mixture was varied from 0 to 1. Next, calculations were conducted for ZrO2−B2O3−Mg−NaCl, HfO2−B2O3−Mg−NaCl, ZrO2−B2O3−Mg−MgO, and HfO2−B2O3−Mg−MgO systems. Here, ZrO2/B2O3/Mg and HfO2/B2O3/Mg ratios were maintained constant at stoichiometry (1:1:5), while the mole fractions of NaCl and MgO were varied from 0 to 1.

P.85

3

Results

100

2500

80

2000

60

1500

40

1000

20

500

0

0 0

0.2 0.4 0.6 0.8 1 Mole Fraction of Mg in Initial Mixture

Adiabatic Flame Temperature, K

Conversion, %

Figure 1 shows the conversion degree of ZrO2 to ZrB2 and the adiabatic flame temperature for ZrO2−B2O3−Mg system. It is seen that the maximum temperature (2380 K) is reached at stoichiometry. However, 100% conversion of ZrO2 to ZrB2 is not achieved until the mole fraction of Mg in the initial mixture increases to approximately 0.77. At this point, the temperature (2130 K) is high enough for SHS.

Fig. 1. Adiabatic flame temperature and conversion of ZrO2 to ZrB2 vs Mg concentration in ZrO2−B2O3−Mg system.

2000

95

1500 90 1000 85

500

80 0

0 0.2 0.4 0.6 0.8 1 Mole Fraction of NaCl in Initial Mixture

100

Conversion, %

2500

Adiabatic Flame Temperature, K

Conversion, %

100

2500 2000

95

1500 90 1000 85

500

80 0

0 0.2 0.4 0.6 0.8 1 Mole Fraction of MgO in Initial Mixture

Fig. 2. Adiabatic flame temperature and conversion of ZrO2 to ZrB2 vs concentration of (left) NaCl in ZrO2−B2O3−Mg−NaCl system and (right) MgO in ZrO2−B2O3−Mg−MgO system.

P.86

Adiabatic Flame Temperature. K

Figure 2 shows how the added NaCl and MgO affect the conversion degree and the adiabatic flame temperature of ZrO2−B2O3−Mg system. The results indicate that 100% conversion of ZrB2 (or HfB2) can be achieved by increasing the mole fraction of NaCl in the initial mixture to approximately 0.37 and MgO to about 0.27. For both systems, the adiabatic flame temperature at these points (1730 K for NaCl and 2350 K for MgO) is sufficiently high for a self-sustained combustion.

The calculations for Hf-based systems have produced virtually the same results. This is explained by similar thermochemical properties of Zr- and Hf-containing compounds. 4

Conclusions

The thermodynamic calculations have shown that the addition of Mg, NaCl, or MgO to the stoichiometric ZrO2/B2O3/Mg and HfO2/B2O3/Mg mixtures yields 100% conversion of ZrO2 and HfO2 to borides, while the adiabatic flame temperature remains sufficiently high for a selfsustained combustion. The use of NaCl and MgO as additives appears to be promising since no gaseous products are generated by combustion, while the excess Mg may lead to an undesired pressure increase in the reaction chamber because of the relatively low boiling point of Mg. 5

Acknowledgements

This material is based upon work supported by the Department of Energy, National Energy Technology Laboratory under Award Number DE-FE0026333. 6

Disclaimer

This report was prepared as an account of work sponsored by an agency of the United States Government. Neither the United States Government nor any agency thereof, nor any of their employees, makes any warranty, express or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately owned rights. Reference herein to any specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise does not necessarily constitute or imply its endorsement, recommendation, or favoring by the United States Government or any agency thereof. The views and opinions of authors expressed herein do not necessarily state or reflect those of the United States Government or any agency thereof. 7

References

[1] N. Kayukawa, “Open-cycle magnetohydrodynamic electrical power generation: a review and future perspectives,” Prog. Energy Combust. Sci. 30 (2004) 33–60. [2] M.L. Bauccio, ASM Engineered Materials Reference Book. ASM International, Materials Park, OH, 1994. [3] K. Upadhyay, J.M. Yang, and W.P. Hoffman, “Materials for ultrahigh temperature structural application,” Am. Ceram. Soc. Bull. 76 (1997) 51–56. [4] W.G. Fahrenholtz, G.E. Hilmas, I.G. Talmy, and J.A. Zaykoski, “Refractory diborides of zirconium and hafnium,” J. Am. Ceram. Soc. 90 (2007) 1347–1364. [5] H.E. Carmurlu, and F. Maglia, “Preparation of nano-size ZrB2 powder by self-propagating high-temperature synthesis,” J. Eur. Ceram. Soc. (2009) 1501–1506. [6] A.K. Khanra, L.C. Pathak, and M.M. Godkhindi, “Double SHS of ZrB2 powder,” J. Mater. Process. Technol. (2008) 386–390. [7] A.A. Shiryaev, “Thermodynamics of SHS processes: Advanced approach,” Int. J. SHS 4 (1995) 351–362.

P.87

SETS2016-23 The Southwest Emerging Technology Symposium 2016

FLAME FRONT IMAGING TECHNIQUES ON A BACKWARD FACING STEP STABILIZED FLAME 1

A. de la Torre1, A. Acosta-Zamora1, A. Choudhuri1* Department of Mechanical Engineering, The University of Texas at El Paso, El Paso, TX 79968, USA; Keywords: chemiluminescence, OH LIF, combustion, step combustor

ABSTRACT The work presented focuses on a study of flame front imaging techniques on a backward facing step stabilized flame. A backward facing step combustor coupled with a Planar Laser Induced Fluorescence (PLIF) system was used to conduct these studies. This system allowed for the comparison of chemiluminescence and OH LIF imaging techniques on a stabilized flame at Re=13,500 for flame front visualization. 1

Introduction

1.1 Backward Facing Step Combustor Combustion devices such as gas turbines, boilers, furnaces, etc. are designed with the goal of improving their efficiency, and regulate pollutant emission to meet government-imposed standards. Both of these parameters strongly rely on the better understanding of the chemical and physical processes in the flame, which, in most of the cases, are in a turbulent fashion. Premixed combustion has emerged as a leading technology to reduce pollutant emissions and increase efficiencies of combustion devices, making a detailed investigation of the fundamental mechanisms of turbulent premixed flames important [1]. Turbulent flow over a backward facing step is widely used to evaluate the performance of turbulence models. For this study a backward facing step combustor was used as a flame anchoring mechanism. This system takes advantage of the recirculating zone formed due to the sudden expansion. The flame is anchored due to the lower velocity experienced in this region, while the recirculating flow keeps the flame ignited due to the constant intake of fresh air/fuel mixture. 1.2 Laser Induced Fluorescence (LIF) LIF imaging of OH radicals is a popular technique used to conduct flame front studies. This technique is widely used to analyze flame characteristics ranging from laboratory burner flames to gas turbine combustion systems. Fluorescence is produced by the excitation of OH in the flame by a Nd:YAG pump dye laser. This laser system is tunable that allows the operator to scan for the excitation wavelength of the species of interest [2]. The OH radical plays a vital role in the oxidation of hydrocarbon fuels, which takes place at the flame front, making it a suitable flame front tracker [3]. Due to it’s relative higher concentration across

P.88

different operating conditions when compared to other flame front tracking species, like HCO and CH, OH is regarded as a suitable flame front tracking radical. 1.3 Chemiluminescence Chemiluminescence is a simple, inexpensive imaging technique that uses chemical excitation, which is light produced from chemical reactions. The emission of light happens when producing unstable products, that later decay to form more stable ones. Research has shown that chemiluminescence of flame species such as CH, C2, OH and CO2 can be used as a qualitative and/or quantitative combustion diagnostic tool [4]. Unlike the LIF technique, this technique uses only the chemical excitation light emitted by the chemically excited OH. 2

System Description

2.1 Combustion chamber and Laser Systems A combustion chamber assembly houses the reacting flow during experimental procedures. This chamber has optical access through quartz windows located on both lateral sides as well as the topside of the combustor. These windows allow the introduction of laser sheets into the combustor and enable visualization of the reacting flow inside the chamber through the coupling and synchronized high-speed cameras and laser systems. The chamber also houses the step, which introduces a sudden area expansion in the flow path resulting in a recirculation zone that serves as a flame anchoring mechanism. 2.2 Planar Laser Induced Fluorescence System A planar laser induced fluorescence (PLIF) system, consisting of a Nd:YAG pumped Rhodamine 6G dye laser, is integrated with the combustor to generate a 283 nm wavelength laser beam that is later routed to enter as a 50x25x1 mm laser sheet through the top window of the combustor to excite the OH radicals that serve as a flame front marker. Perpendicular to the laser sheet is a high-speed camera coupled with an intensifier unit, and a spectral filter attached to the unit to capture the flame structure. The integrated system is shown in Fig. 1. 3

Methodology

To reduce the noise generated from background flame light when exploring the OH fluorescence, the intensifier, high-speed camera, and laser system have to be synchronized. OH excitation was achieved using the frequency doubled output of a Nd:YAG pumped dye laser tuned to around 283 nm. The fluorescence was then captured using a high-speed camera (>3000 fps at full resolution, 500,000 max fps) coupled with an image intensifier unit. Because the OH radical relaxes from the excited state induced by the laser in a short amount of time, the gate time for the intensifier has to be set at 200 ns. This will suppress the light from the flame while detecting the short lived fluorescent pulse generated by the OH radical. 4

Results

Fig. 2 presents the results of the OH radicals being excited using LIF and the chemiluminescence light emission captured from a CH4/Air flame at equivalence ratio 1

P.89

FLAME FRONT IMAGING TECHNIQUES ON A BACKWARD FACINGS STEP STABILIZED FLAME

flowing from left to right at a Reynolds number of 13,500. Because the OH radicals are short lived during the combustion process an intensifier gate time of 100 - 200 ns is used. The LIF signal provides a 50x25x1 mm volume out of the flame, it is able to generate a sharper image and define the flame front structures much more clearly, while the chemiluminescence image provides more of a field of view image, and requires a longer exposure time to generate a signal. This causes the flame structures to be lost and appear to be blurred out. 5

Conclusions

When comparing OH LIF and chemiluminescence imaging to capture the flame front structures, it is quickly seen that due to the good temporal and spatial resolution, the LIF technique is preferred when conducting studies that require a detailed characterization. The chemiluminescence imaging technique is a viable technique where it is technically difficult, or too expensive to apply the LIF technique such as in flame studies in internal combustion devises, but it is still an appropriate technique to approximate the location of the flame front. References [1] S. Pfadler, F. Beyrau, and A Leipertz “Flame front detection and characterization using conditioned particle image velocimetry (CPIV)” Optics Express, Vol. 15, No. 23, pp15444-15456, 2007. [2] C. Dreyer, S. Spuler, and M. Linne “Calibration of laser induced fluorescence of the OH radical by cavity ringdown spectroscopy in premixed atmospheric pressure flames” Commercial Applications of Combustion in Space, Colorado School of Mines. [3] Malmqvist, Elin. "Thermometry using OH laser-induced fluorescence excitation spectra: A feasibility study." MA thesis. Lund University, 2013. [4] B. Giggins, M.Q. et. al. “An experimental study on the effect of pressure and strain rate on CH chemiluminescence of premixed fuel-lean methane/air flames”. Fuel, Vol. 80, No. 11, pp 1583-1591

Fig.1. Panoramic view of experimental setup and LIF system integration with combustor.

Fig.2. LIF (left) Chemiluminescence (right).

3 P.90

SETS2016-24 The Southwest Emerging Technology Symposium 2016

IMPACT OF INTERMITTENT RENEWABLE ENERGY SOURCES ON POWER SYSTEM ANALYSIS Luis A. Gutierrez and Paras Mandal* Power & Renewable Energy Systems (PRES) Lab, Department of Electrical and Computer Engineering, University of Texas at El Paso, El Paso, TX 79968, USA *Corresponding author ([email protected])

Keywords: Load flow, photovoltaic, voltage profile, wind ABSTRACT This paper is intended to investigate the impact of renewable energy sources (RES) on a power system network by performing load flow analysis. The most common renewable energy technologies include photovoltaics and wind turbines. 1 Introduction During the last years, the integration of renewable energy sources (RES), mainly solar and wind, into power systems has become more important due to global warming. This is why “green” energy sources are expected to play a very important role in the near future. This paper makes use of a 9-bus system using PowerWorld simulator. It focuses on the effects that the integration of renewable energy sources (RES) can have on the load flow analysis of the power network. A load flow study is a steady state analysis that determines the voltages, currents, real, and reactive power flows in a system under certain load conditions. However, this paper will mainly analyze the impact on active power, reactive power, and voltage profile at each bus. This paper considers four cases for load flow analysis: (i) Case-1: base scenario without considering RES, (ii) Case-2: integration of photovoltaic (PV), (iii) integration of wind turbine, and (iv) integration of both PV and wind. The 9-bus includes a slack-generator bus, five loads, two generators, and a transformer. 2 Proposed Approach and Results

P.91

35" 30" 25" 20" 15" 10" 5" 0"

Power&Output&(MW)&

As mentioned in previous section, four cases are considered to analyze the impact of RES integration. The four cases are simulated with the time step simulation (TSS) included in the PowerWorld software. The TSS tool allows the user to specify operating conditions and obtain power flow solutions for a set of points in time. It provides the tools needed to analyze the operation of a power system hour by hour [3]. Fig. 1 shows the daily power output of the PV power and the wind power. The four cases are described as follows.

WIND" PV"

1" 3" 5" 7" 9" 11"13"15"17"19"21"23" Time&(Hour)& Fig. 1. Wind and PV power output [1], [2].

2.1 Case- 1: Initial condition without considering wind and PV The goal of the first case is to set the initial conditions for the load flow analysis without considering RES. 2.2 Case-2: Integration of PV source to bus-7 The integration of a PV power source to bus-7 has a slight significant impact on the power system. As it can be seen in Table 1, the impact of PV integration is such that it reduces the active power (from 467 MW to 463.46 MW) and reactive power (from 511 MVAR to 508.70 MVAR) at the slack bus. The maximum power generation by the PV source for a particular day is 4.68 MW at hour-14. Even with this small amount of power generation by the PV source, it shows a significant impact on the voltage profile at bus 7, i.e., from 0.52 to 0.53 per-unit. Fig. 2 shows a sample of load flow simulation. Table 1 – Load flow analysis results and comparison with or without RES. CASE

P (MW)

Q (MVAR)

BUS

BUS

V (P.U.) BUS

SLACK

5

7

SLACK

5

7

SLACK

5

7

INITIAL

467

x

x

511

x

x

1.00

0.37

0.52

PV

463.46

x

4.68

508.78

x

0

1.00

0.37

0.53

WIND

453.87

18.17

x

497.8

0

x

1.00

0.39

0.53

WIND/PV

451

18.17

4.68

496

0

0

1.00

0.39

0.54

Fig. 2. Load flow analysis using 9-bus system.

P.92

The Southwest Emerging Technology Symposium 2016

2.3 Case- 3: Integration of wind source to bus-5 In Case-3, the wind source is connected to the bus-5. From Fig. 1, it can be observed that the power output of the wind turbine (close to 29 MW at hour-24) is considerably larger than that of a PV source. Comparing with the PV source, the integration of wind power to the grid caused a more desirable impact on the load flow analysis. The active power output from the slack generator has a drop of 13.13 MW. Identically, the reactive power required from the slack also exhibits a decrement of 13.2 MVAR. Advantageously, the voltage profile at bus 5 increased from 0.37 to 0.39 per-unit. 2.4 Case-4: Integration of both PV and wind sources It is expected that the most beneficial impacts on the load flow analysis can be observed when the two different RES are integrated to the power system. Case-4 considers the integration of wind and PV sources to buses 5 and 7, respectively. The slack bus generator experienced a meaningful reduction on its required power generation. It was able to cut the active power output by 16 MW and the reactive power generation lowered from 511 to 496 MVAR. The voltage profile at bus-5 did not exhibit an increment compared to Case 3. However, the voltage profile at bus 7 experienced another desirable increase to reach the value of 0.54 per unit. Table 1 summarizes the results for each case. In this paper, all four cases were simulated considering the peak PV power generation of 4.68 MW (observed at hour-14 of a particular day) and average of wind power at hour-14 and hour-15, i.e., 18.17 MW. Note that the maximum wind power generation was obtained at hour-24 (28.95 MW), however, in our simulated cases, we considered the power generation from the RES for around hour-14 in which PV power is maximum and wind power is about 63% of the total generation. 3 Conclusion This research demonstrated the positive effects that the integration of renewable energy sources has on load flow analysis. The addition of two different renewable energy power sources can help reduce the power output (active and reactive) from the slack bus generator. This can reduce the demand of fossil fuel power generators, which in turn can help reduce the carbon dioxide emissions. It also produces a significant positive effect on the voltage profile. References [1] NRELwebsite, http://www.nrel.gov/ [2] B. Lu and M. Shahidehpour, “Short-term scheduling of battery in a grid-connected PV/battery system,” IEEE Trans. on Power Systems, Vol. 20, No. 2, pp. 1053-1061, May 2005. [3] PowerWorld simulator website, http://www.powerworld.com/

P.93

SETS2016-25 The Southwest Emerging Technology Symposium 2016

CONVERTING METHANE WASTE TO VALUABLE MATERIALS R.R.Chianelli* Materials Research and Technology Institute (MRTI), University of Texas El Paso (UTEP), El Paso, TX 79968, USA; * Corresponding author ([email protected])

Keywords: Methane, Petroleum Flares, Landfill Biology, and Photocatalysis ABSTRACT The United States, Canada, and Mexico are now being called the “New Middle East” because of the huge production of oil and gas, from both conventional and shale sources.[1] This new hydrocarbon production produces two waste materials in large quantities: flare natural gas and tars. The gas is typically burned away in flares because it is cheaper to burn it than to liquefy and sell it. A recent NASA study revealed that space stations flying over Siberia reported thousand of flaring oil wells from conventional oil wells, burning as much energy as the East Coast of the United States uses for travel every day[2]. The same phenomenon (Figure 1) has been reported in the Western states of the United States[3]. Flare gas is typically comprised of approximately 50% methane and 50% CO2. Catalytic reforming of methane with CO2 is possible using catalytic materials producing valuable hydrocarbons that are liquid at room temperature. However, this reaction occurs at temperatures near 700oC. Thus, the cost of putting this process in the petroleum fields is too high. “Flare quenching” methods are now being developed using novel catalytic materials that are producing the reforming using solar power to convert the waste to useful products[4]. In addition, refineries that product heavy crudes produce large quantities of “heavy bottoms” that are used to produce “road tars.” These heavy bottoms contain molecules called asphaltenes. Under moderate conditions, novel catalytic materials can destroy the asphaltenes and remove sulfur and nitrogen. Interestingly, novel materials can be generated during this process. A solar cell made with asphaltenes has been reported[5]. Several more interesting materials made from the “waste asphaltenes” are anticipated to be reported in the future. Thus, there are opportunities to use this so-called waste from oil and gas production using novel catalysts; however, extensive materials research is needed to take advantage of these opportunities.

1

M.A. O’Grady, “The North American Gusher,” The Wall Street Journal (December 9, 2012). R. R. Chianelli, "Don't Waste the Waste", Materials Research Bulletin, Energy Quarterly, Volume 40 November 2015. 3 S. Tomlinson, “What a waste!,” The Daily Mail (2013), available at http://www.dailymail.co.uk/news/article2269517. 4. R. R. Chianelli, B. Torres, “Photochemical Processes and Compositions for Methane Reforming Using Transition Metal Chalcogenide Photocatalysts,” US Patent 0239469 (September 19, 2013). 5 . R. R. Chianelli, K. Castillo, V. Gupta, A.M. Qudah, B. Torres, R.E. Abujnah, “Asphaltene Based Photovoltaic Devices,” US Patent: 12/833,488 (March 5, 2013). 2

P.94

CONVERTING METHANE WASTE TO VALUABLE MATERIALS

Figure 1, Petroleum Flares in the United States: Reference 3

P.95

SETS2016-26 The Southwest Emerging Technology Symposium 2016

COMPARISON OF TWO IDP TECHNOLOGIES IN DETECTING AND PREVENTING CYBER-ATTACKS ON MICROGRID COMMUNICATION NETWORKS Goutham K Chalamasetty1, Paras Mandal1*, and Tzu-Liang (Bill) Tseng2 1 Power & Renewable Energy Systems (PRES) Lab, Department of Electrical and Computer Engineering, 2 Department of Industrial, Manufacturing, and systems Engineering, The University of Texas at El Paso, El Paso, TX, 79968, USA *Corresponding author ([email protected])

Keywords: Cyber security, IDP, MANET, residential microgrid, SCADA.

ABSTRACT This paper aims to compare two intrusion detection and prevention (IDP) technologies in defending cyber-attacks on our proposed supervisory control and data acquisition (SCADA) network. The proposed SCADA network collects the data of power consumption from smart meters in houses and electric vehicles, and sends that data to base station. In this paper, we applied two IDP technologies for providing cyber security to our proposed SCADA network. Results compare the effectiveness of these technologies in preventing cyber-attacks. 1

Introduction

In this smart grid era, one of the major issues is to meet power demand and supply in a secure manner. There is a huge research going in developing technologies to transform the traditional power grid to smart grid, which helps to solve the issue. Supervisory control and data acquisition (SCADA) system is the one which controls the power system operations such as generation, distribution, and transmission. With the recent advancements in smart grid technologies, it is possible to inject power from distributed energy resources (DER) into the grid in order to meet power demand and supply. This made the role of SCADA more challenging as the complexity in the system increases with the integration of DER. The responsibilities of SCADA is to monitor the remote substations such as generation plants, distribution plants, and transmission plants by collecting and analyzing the data from substations. Furthermore, SCADA sends control commands to the substations based on the analyzed information. This paper presents a SCADA network for residential microgrid communication and contributes to compare the effectiveness of two different intrusion detection and prevention (IDP) technologies: (i) monitoring, detecting, and rehabilitation (MDR) approach and (ii) secure knowledge algorithm with anomaly detection technology, which are applied to protect the proposed SCADA network from cyber-attacks. 2

Problem Statement and Solution

The SCADA system communication architecture is a combination of local area networks (LAN) and wide area networks (WAN). There are many vulnerabilities in the current SCADA system

P.96

such as unencrypted data transmission, lack of authentications, lack of proper firewalls, and network design and configuration vulnerabilities. These vulnerabilities benefit cyber attackers to perform the cyber-attacks. In addition to the aforementioned problems, the other issue could be due to a natural disaster where the entire communication infrastructure is destroyed, and the operator has lost the access to control the system. The integration of advanced communication technologies (ACT) such as such as wireless sensor network (WSN), internet, and mobile ad hoc networks (MANET) into the existing SCADA system will overcome most of the aforementioned vulnerabilities [1]. In addition, an efficient integration of ACT brings more reliability, flexibility, redundancy, automation, mobility, accessibility, and security to the SCADA system. This paper presents a SCADA network using MANET for residential microgrid communications. 3

Proposed SCADA Architecture and Applied IDP Technologies

Some of the properties of MANET are: (i) cost effectiveness, (ii) consumes less power, and (iii) no infrastructure is needed for communication. The objective of MANET network (see Fig. 1) is to collect the information from smart meters in houses and electric vehicles (EV) and send that data to control center. This network also connects mobile system operators into the network for handling emergency situations. The dynamic nature of MANET benefits the Fig. 1. Proposed SCADA architecture. cyber attackers to attack the network. In this paper, we applied MDR approach and a secure knowledge algorithm [2], [3] to compare the effectiveness of these two IDP technologies for detecting and preventing denial of service (DoS) attacks. Both of these IDP technologies help to detect and prevent DoS attacks such as malicious nodes and black hole attacks on the network. These IDP technologies were individually applied to the proposed architecture in our previous work [4], [5]. 4

Simulations and Results

The proposed SCADA architecture (see Fig. 1) is developed in network simulator version 2 (NS 2) with 40 nodes in which advanced on demand distance vector algorithm (AODV) is used as routing protocol, k-means clustering algorithm is used for nodes clustering, and network traffic is TCP. The network is simulated for two cases, Case I: MDR approach is applied to the network, when there are four DoS attacked nodes in the network. Case II: a secure knowledge algorithm with anomaly detection is applied to the network, when there are four DoS

P.97

Fig. 2. Case I vs Case II.

attacked nodes in the network. Three major network parameters are considered to test the effectiveness of these IDP technologies, i.e., (i) Packet delivery ratio (PDR), which is the ratio of number of packets sent to the number of packets received, (ii) Network throughput, which is the

Fig. 4. Case I vs Case II.

Fig. 3. Case I vs Case II.

rate at which the data is successfully delivered, and (iii) Delay, time taken for sending packets from source to destination. Case I is represented as red line and Case II as represented in blue line. Figs. 2-4 show the performance of the attacked network when IDP technologies are applied. Case II is more effective for the parameters network throughput and delay (see Figs. 3 and 4), whereas PDR is slightly greater in Case I (see Fig. 2). Test results demonstrated that the secure knowledge algorithm with anomaly detection is a slightly more effective than the MDR approach. 5

Conclusion

This paper presented a SCADA network for residential microgrid communications where two IDP technologies (MDR and secure knowledge algorithm with anomaly detection) were applied. Furthermore, we compared the effectiveness of both the IDP technologies in detecting and preventing DoS attacks on the proposed SCADA network. Future work will focus on exploring more cyber-attacks and develop a new and more efficient IDP technology for the proposed SCADA network. References [1] N.R. Kumar, P. Mohanapriya, and M. Kalaiselvi, “Development of an attack-resistant and secure SCADA system using WSN, MANET, and Internet”. International Journal of Advanced Computer Research Vol. 4, No. 2, pp. 627, 2014. [2] A. Alsumayt, and J. Haggerty, “Using trust based method to detect DoS Attack in MANETs”. PGNet: The convergence of Networking, Broadcasting, and Telecommunications, UK, 2014. [3] Siddiqua, K. Sridevi, and A.A.K. Mohammed, “Preventing black hole attacks in MANETs using secure knowledge algorithm.” In SPACES, International Conference, pp. 421-425. IEEE, 2015. [4] G.K. Chalamasetty, P. Mandal, and B. Tseng, “Secure SCADA communication network for detecting and preventing cyber-attacks on power systems”. In Power System Conference (PSC), Clemson University, IEEE, 2016. [5] G.K. Chalamasetty, P. Mandal, and B. Tseng, “SCADA framework incorporating MANET and IDP for cyber security of residential microgrid communication network”. Smart Grid and Renewable Energy, 2016. (In press).

P.98

SETS2016-27 The Southwest Emerging Technology Symposium 2016 ZINC OXIDE THIN FILM PREPARATION BY SINGLE SOLUTION DEPOSITION FOR PEROVSKITE SOLAR CELLS Manuel F. Martinez1, Shaimum Shahriar1, Donato Kava1, Cheik Sana1, Vanessa Castaneda1, Jose Galindo1, and Deidra R. Hodges1. 1

Electrical and Computer Engineering, The University of Texas at El Paso, El Paso, TX 79968, U.S.A.

ABSTRACT Zinc oxide thin films were prepared via the sol-gel spin-coating method with the use of Laurell’s WS-650MZ-23B spin processor. The film’s annealing parameters were varied to study their impact on the final film morphology and electrical properties. Characterization of the structural properties of the samples was carried on a Bruker D8 Discover X-ray diffractometer (XRD). Electrical characterization was obtained with the use of a Signatone Pro-4 four point probe. Optical characterization of the samples was carried on a Varian Cary 5000 UV-Vis-NIR Spectrophotometer. Samples annealed under a cover are observed to have a higher transmission percentage on the visible light range while having a very small crystallite size and small relative resistivity. Samples annealed under standard atmospheric conditions show a much larger crystallite size and resistivity, and correlated to it, a smaller transmission percentage. Samples annealed under vacuum prove to have a much more reduced optical, electrical, and structural properties when compared to the rest of the samples. INTRODUCTION Zinc oxide thin films are considered as promising candidates for perovskite thin-film perovskite solar cells. Zinc oxide nanocrystals have been used before in perovskite-based solar cells as an electron-selective contact (electron transport layer) that requires no sintering [1] either by depositing them into a mesoporous titanium dioxide (TiO2) scaffold [1][2], or by using them as a stand-alone layer for electron transport [1][3][4]. Both deposition methods have shown efficiencies in solar cells that range from 10.2% to 12.8% [1][3][4]. This material has been used in solar cells because of its wide band gap of 3.37 eV at room temperature and also because it exhibits high exciton binding energy (60 meV), which ensures an efficient excitonic emission and recombination up to room temperature [5]. This material also normally has a charge carrier concentration of 1016 to 1017 cm-3 [6]. Previous studies of the optoelectronic properties of this material were made by preparing samples with a wide variety of techniques such as spray pyrolysis, chemical vapor deposition (CVD), and RF sputtering. However, these techniques require the use of sophisticated equipment and setups that drive the costs of fabrication up [5]. For this reason, recent publications explore the fabrication of ZnO films prepared by the sol-gel method as an inexpensive, simpler, alternative to the use of the material in photovoltaics [7]. A tight control of the electrical and optical properties of zinc oxide thin-films can be achieved by the control of the change in dopant used and the oxygen adsorption by the film [8]. Furthermore, as Ladanov reported, a control of the growth parameters such as pressure, interacting gasses, and temperature of the substrate provide a high repeatability of tailored ZnO films [9]. P.99

In this work, the effects of annealing parameters—such as medium and temperature—on the electrical, optical, and structural characteristics of ZnO thin-films is studied with the intent on later using an optimized ZnO layer composition on perovskite solar cells. EXPERIMENT Zinc acetate dihydrate (Zn(CH3COO)2·2H2O) (99%, Fisher Scientific), was dissolved in a 2-methoxyethanol (99%, Acros)-Mono Ethanolamine (MEA) (≥98%, Sigma) solution at room temperature [10]. The molar ratio of MEA to zinc acetate dihydrate was maintained at 1:1 and the concentration of Zn(CH3COO)2 ·H2O in the solution was 0.75 mol/L. The final homogeneous solution was left stirring on a Corning PC420D hot plate at 120 rpm and kept at a constant temperature of 60 °C with the aid of a temperature controller for 2 hours. After 2 hours, the solution was then cooled to room temperature and left stirring at 120 rpm for 48 hours. ZnO films were prepared on soda lime glass (SLG) substrates by repeated spin-coating using a Laurell Technologies WS-650MZ-23B spin processor. Before the deposition took place, the substrates were heated to 200 °C (Corning PC420D hot plate) for 10 minutes. The zinc oxide solution was deposited on top of the hot SLG substrate with the aid of a plastic syringe and then spun at high revolutions per minute (RPM); this process was repeated for each substrate. After each layer deposition, the substrate was dried at 200 °C for 10 minutes to allow the solvent (2methoxyethanol) to evaporate from the film. Annealing of the substrates was done under 3 different operating conditions to study the effect on the final film characteristics. The first set of samples were annealed at 550 °C for 2 hours (Corning PC420D hot plate) in air. The second set of samples were annealed at 550 °C for 2 hours (Corning PC420D hot plate) with a cover (isolated from air). The final set of samples were annealed on a Rapid Thermal Processing (RTP) furnace (MTI OTD-1200X) under 25 mTorr of vacuum at 200 °C for 30 minutes. The prepared films were characterized via X-ray diffraction (XRD) (Bruker D8 Discover), UV-Vis-NIR spectroscopy (Varian Cary 5000), and four point resistivity probe (Signatone Pro-4).

DISCUSSION Electrical Properties The electrical properties of the final samples were characterized via four point resistivity probe system (Signatone Pro-4), the results for these two measurements are shown in Table 1. Table 1. This table shows the parameters obtained from the four point probe. Medium Resistivity (𝛺 · 𝑐𝑚) Sheet Resistivity (𝛺 · 𝑠𝑞) Resistance (Ω) 1 4 8 Atmosphere 4.208x10 2.630x10 7.903x107 Covered2 3.440x104 2.150x108 6.203x107 Vacuum3 3.822x104 2.388x108 6.448x107 1. -7 Measured with a current of 4.323x10 mA 2. Measured with a current of 6.176x10-7 mA 3. Measured with a current of 3.678x10-7 mA

P.100

The results shown on Table 1 describe the average of obtained parameters for each sample set from a four point probe (Signatone Pro-4). The highest resistivity obtained was observed in the samples annealed under standard atmosphere—this behavior can be explained from the crystallite size of the samples (as described later in the Structural Properties section of this work)—followed by samples annealed under vacuum and finally by samples annealed covered. The electrical properties obtained in this work are directly related to the morphological characteristics that the finalized films describe. In the following section, these properties are explored and explained for the different sample sets. Structural Properties XRD data was obtained from a Bruker D8 Discover X-ray diffractometer with Cu Kα radiation (λ=1.5406 Å) and is shown in Figure 1. Samples covered and samples annealed under standard atmospheric conditions crystallized in the hexagonal structure with a = 3.24982 Å and c = 5.20661 Å (JCPDS 00-036-1451); however, samples annealed covered showed an average crystallite size of 21.7 Å while those annealed under standard atmospheric conditions present an average crystallite size of 531.06 Å as determined by the Debye-Scherrer formula [5]: 𝐷 = 0.9𝜆/(𝐹𝑊𝐻𝑀 cos 𝜃 ) Where λ is the applied X-ray wavelength, FWHM is the full width half maximum, and θ is the diffraction peak angle. XRD data for samples annealed under 25 mTorr of vacuum showed that although some crystallization is present in the sample, it has a crystallite size smaller than those of the samples annealed covered.

Figure 1. XRD Data for prepared samples.

P.101

From Figure 1, it can be observed that both samples annealed covered and samples annealed under standard atmospheric conditions show a preferential crystal growth on the (101) plane, followed by the (002) and (100) planes. This preferential crystal growth is in accordance to previously presented XRD data for similarly prepared ZnO thin-films [11]. Optical Properties Optical characterization for the samples was obtained with the aid of a Cary 5000 UVVis-NIR spectrophotometer with a wavelength setting of 2220 nm to 300 nm. The average data obtained within sample sets is presented in Figures 2 and 3. From Figure 2 it can be observed that the average transmittance of samples annealed covered was far greater than the rest of the samples over the ultraviolet and visible range. Followed by this, samples annealed under standard atmospheric conditions and samples annealed under vacuum described a much more reduced average transmittance. In the visible range, samples annealed under a cover describe a transparent behavior as it shows an average transmittance percentage of 60% while a much larger transmittance is observed for the infrared spectrum; samples annealed under standard pressure and atmosphere exhibit a much more reduced transmittance over the visible light range but are characterized by a more uniform decay; finally, samples annealed under vacuum display a similar behavior to those annealed under standard pressure and atmosphere, they seem to be much less transparent than those annealed covered from external factors. It is also worth noting that absorbance data for all of the samples show a very sharp peak at the 390 nm wavelength, whose energy (~3.17 eV) closely corresponds to the intrinsic bandgap of ZnO (~3.37 eV) [12].

Figure 2. Average transmittance data for finished samples.

P.102

Figure 3. Average absorbance data for finished samples. CONCLUSIONS The effect of varying annealing parameters on the electrical, optical, and structural properties of spin-coated ZnO thin films were studied. Results show that samples annealed covered and under standard atmospheric conditions describe the hexagonal crystal structure with preferential growth on the (101) plane while those annealed under vacuum show little to no crystallization as obtained from XRD data. Higher crystallite size is obtained in samples annealed under standard atmospheric conditions at the expense of a higher resistivity and a lower transmission percentage in the visible range; on the other hand, covered samples have a smaller crystallite size and higher transmission percentages. In all characterization methods described, samples annealed under vacuum describe least preferred electrical, optical, and structural characteristics when compared to the rest of the sample sets. ACKNOWLEDGMENTS This work was supported by the University of Texas at El Paso (UTEP) College of Engineering, the Electrical and Computer Engineering Department, and UTEP-Partnerships for Research and Education in Materials (PREM)—NSF Grant DMR-1205302. REFERENCES [1] [2] [3]

D. Liu and T. L. Kelly, “Perovskite solar cells with a planar heterojunction structure prepared using room-temperature solution processing techniques,” Nat. Photonics, vol. 8, no. 2, pp. 133–138, 2013. W. J. E. Beek, M. M. Wienk, M. Kemerink, X. Yang, and R. A. J. Janssen, “Hybrid zinc oxide conjugated polymer bulk heterojunction solar cells,” J.Phys.Chem.B, vol. 109, pp. 9505–9516, 2005. H.-S. Kim, C.-R. Lee, J.-H. Im, K.-B. Lee, T. Moehl, A. Marchioro, S.-J. Moon, R.

P.103

Humphry-Baker, J.-H. Yum, J. E. Moser, M. Grätzel, and N.-G. Park, “Lead Iodide Perovskite Sensitized All-Solid-State Submicron Thin Film Mesoscopic Solar Cell with Efficiency Exceeding 9%,” Sci. Rep., vol. 2, pp. 1–7, 2012. [4] J. M. Ball, M. M. Lee, A. Hey, and H. J. Snaith, “Low-temperature processed mesosuperstructured to thin-film perovskite solar cells,” Energy Environ. Sci., vol. 6, no. 6, p. 1739, 2013. [5] S. A. Kamaruddin, K.-Y. Chan, H.-K. Yow, M. Zainizan Sahdan, H. Saim, and D. Knipp, “Zinc oxide films prepared by sol–gel spin coating technique,” Appl. Phys. A, vol. 104, no. 1, pp. 263–268, 2011. [6] X. D. Li, T. P. Chen, P. Liu, Y. Liu, and K. C. Leong, “Effects of free electrons and quantum confinement in ultrathin ZnO films: a comparison between undoped and Aldoped ZnO,” Opt. Express, vol. 21, no. 12, pp. 14131–14138, 2013. [7] M. Ohyama, “Sol – Gel Preparation of Transparent and Conductive Aluminum-Doped Zinc Oxide Films with Highly Preferential Crystal Orientation,” J. Am. Ceram. Soc., vol. 32, pp. 1622–1632, 1998. [8] P. Nunes, A. Malik, B. Fernandes, E. Fortunato, P. Vilarinho, and R. Martins, “Influence of the doping and annealing atmosphere on zinc oxide thin films deposited by spray pyrolysisn zinc oxide thin films deposited by spray pyrolysis,” Vacuum, vol. 52, no. 1–2, pp. 45–49, 1999. [9] M. Ladanov, “ZnO Nanostructures : Growth , Characterization and Applications,” no. January, 2012. [10] H. Li, J. Wang, H. Liu, C. Yang, H. Xu, X. Li, and H. Cui, “Sol - Gel preparation of transparent zinc oxide films with highly preferential crystal orientation,” Vacuum, vol. 77, no. 1, pp. 57–62, 2004. [11] T. Sahoo, M. Kim, M. H. Lee, L. W. Jang, J. W. Jeon, J. S. Kwak, I. Y. Ko, and I. H. Lee, “Nanocrystalline ZnO thin films by spin coating-pyrolysis method,” J. Alloys Compd., vol. 491, no. 1–2, pp. 308–313, 2010. [12] X. Zhao, J. Y. Lee, C. R. Kim, J. Heo, C. M. Shin, J. Y. Leem, H. Ryu, J. H. Chang, H. C. Lee, W. G. Jung, C. S. Son, B. C. Shin, W. J. Lee, S. T. Tan, J. Zhao, and X. Sun, “Dependence of the properties of hydrothermally grown ZnO on precursor concentration,” Phys. E Low-Dimensional Syst. Nanostructures, vol. 41, no. 8, pp. 1423–1426, 2009.

P.104

SETS2016-28 The Southwest Emerging Technology Symposium 2016 PYROELECTRIC ENERGY HARVESTING WITH HIGH CURIE TEMPERATURE MATERIAL LiNbO3 Jorge Silva1, Hasanul Karim1, MD Rashedul Hasan Sarker1, Saimum Shahriar2, Mohammad Arif Ishtiaque Shuvo1, Diego Delfin1, Deidra Hodges2, Norman Love1, and Yirong Lin1 1

Department of Mechanical Engineering, University of Texas at El Paso 2 Department of Electrical Engineering, University of Texas at El Paso

Keywords: Energy Harvesting, Pyroelectrictricity, Lithium Niobate (LNB) ABSTRACT Interest in energy harvesting has increased since there is a mass of energy wasted in form of heat. Energy harvesting is a potential solution for energizing future generation sensors and energy storage devices. PZT (Lead Zirconate Titanate) is the most widely material used for ongoing research for piezoelectric and pyroelectric, which it has good piezoelectric and pyroelectric properties. However an alternative solution is desired since Lead on PZT is harmful to human health and to the environment. In this paper, Lithium Niobate (LNB) is chosen to be investigated since it contains no Lead. LNB has lower pyroelectric properties than PZT but still better than other alternative lead free materials such as ZnO. Also, LNB has been selected because its high curie point of 1142 °C, which makes it suitable for high temperature applications where other pyroelectric materials are not suitable. Therefore, a single crystal LNB has been investigated as a source of energy harvesting under alternative heating and cooling environment. A commercial 0.2 F supercapacitor was used as the energy storage device. 1. Introduction The increasing demand for self-powered electronics has enthralled an increase in the number of research into Energy harvesting devices. With the development of wireless technology, sensors and other wireless electronics now can be placed at almost anywhere. Numerous investigations are being carried out to obtain potential energy harvesting solutions such as piezoelectric energy harvesting and solar energy harvesting [1]. However, the harvested power densities from these natural sources are in general low and it is necessary to carry out further investigations to obtain alternative sources and increase efficiency of the conversion process. Pyroelectricity is the change of spontaneous polarization in some anisotropic materials, which can be either single crystal or poly-crystalline aggregates due to a change in temperature [2]. The change in polarization produces a flow of charges to and from the material’s surface. The property has already proved to be useful in sensing application. However, the output voltage and current from the reported energy harvesters are still very low. In this paper, we report LiNbO3 as a lead free alternative for pyroelectric energy harvesting. A thermoelectric combined with a polarity switching circuit was used to produce an alternative heating and cooling scenario. The current generated as a function of temperature change is recorded and the harvesting capability of the material is demonstrated by charging a commercial 0.2F supercapacitor.

P.105

PYROELECTRIC ENERGY HARVESTING WITH HIGH CURIE TEMPERATURE MATERIAL LiNbO3

2. Experimental setup To quantify the energy harvesting capability of LNB a controlled environment is desired with a controlled and alternative heating and cooling cycle. In order to achieve this, the LNB wafer was placed on a thermoelectric cooler. Thermoelectric material creates a temperature difference when a voltage is applied to it. When a voltage is applied to the thermoelectric cooler one surface of the thermoelectric block starts heating up while other surface starts cooling down. If the polarity of the applied voltage is changed, the effect is reversed and the surface that was heating up now cools down. To achieve an alternating heating and cooling cycle, a DC power supply was connected to the thermoelectric cooler through a polarity reversing circuit with two timers to control the start and stop of the reversing. Two leads were taken out from the two surfaces of the LNB wafer and were connected to the energy harvesting circuit containing a supercapacitor that was charged from the current generated from the LNB because of the pyroelectric effect. The voltage of the supercapacitor was recorded using a data acquisition system and Labview. The schematic of the setup is shown in figure 1.

Fig.1. Schematic of the pyroelectric energy harvesting 3. Results and discussion To quantify the current generated by the LNB wafer, the leads from the sample were connected to a picoammeter and the output from the picoammeter was connected to the data acquisition system. Figure 2(a) shows the current generated from the LNB wafer and the rate of temperature change that occurred is shown in Figure 2(b). For a peak value of dT/dt of 0.32 °C/s, 380 nA current was generated. The current density observed here for LNB is 8.3 nA/cm2, which is about 20 times lower than the current density of PZT nanogenerator as reported in [3]. However, the pyroelectric properties of LNB are significantly higher from that of the reported value of other lead free alternatives such as ZnO or PVDF. The pyroelectric current coefficient p of bulk ZnO has been reported to be 0.94 nC/cm2K and for the nanowire it was observed in the range of 1.2~1.5 nC/cm2K [3]. For the LNB wafers tested here, the average pyroelectric coefficient observed was 5-8 nC/cm2K, which is in good agreement with literature [2], which is 4~5 times higher than the reported values for ZnO. For the peak current, the generated open circuit voltage was ~2.6 V, resulting in a peak power of 988 nW.

P.106

PYROELECTRIC ENERGY HARVESTING WITH HIGH CURIE TEMPERATURE MATERIAL LiNbO3

Fig. 2. (a) LNB current generated and (b) rate of change of temperature with time

A 0.2F supercapacitor was charged with a LNB wafer and the resulting voltage data with time is represented in figure 3. Figure 3(a) shows how voltage increased rapidly in the beginning and then slowed down until it reached a constant value approximately 4.5 mV after 9hr. On figure 3(b) charging rate of .6 mV/hr can be seen for the first hour. The low value of voltage and the slow charging rate is due to the significant power drop in rectifying circuit as well as the low current generated by the pyroelectric effect.

Fig.3. Charging of a 0.2F supercapacitor with a LNB wafer (a) Charging for 11.2 hours (b) charging profile for the first hour.

4. Conclusion LiNbO3 as pyroelectric material was investigated for pyroelectric energy harvesting. The density of LNB was found to be around 8 nA/cm2 which is 20 times lower than PZT. However, this small current can be utilized by energy harvesting from wasted energy from environment. Energy harvester was capable of charging a commercial .2F supercapacitor. Increasing the surface area or stacking multiple samples and connect them in parallel can generate more current to achieve desired amount of energy. References [1] [2] [3]

Grätzel, M., Solar energy conversion by dye-sensitized photovoltaic cells. Inorganic chemistry, 2005. 44(20): p. 6841-6851. Sarker, M.R.H., et al., Temperature measurements using a lithium niobate (LiNbO 3) pyroelectric ceramic. Measurement, 2015. 75: p. 104-110. Yang, Y., et al., Pyroelectric nanogenerators for harvesting thermoelectric energy. Nano letters, 2012. 12(6): p. 2833-2838.

3 P.107

SETS2016-29 The Southwest Emerging Technology Symposium 2016

THERMAL STABILITY AND OXYGEN SENSOR CHARACTERISTIC OF GA2O3 BASED HIGH TEMPERATURE OXYGEN SENSORS Ernesto J. Rubio1, Sandeep Manandhar1, Ramana Chintalapalle1* 1 Department of Mechanical Engineering, University of Texas at El Paso, El Paso, TX 79968, USA; * C.V. Ramana ([email protected]) Keywords: Thermal Stability, Oxygen Sensors, Sensitivity, Time Response ABSTRACT Critical understanding and control over high temperature energy processes has become one target for research and development materials and technologies, especially in the area of sensors and actuators where temperature limitations and selectivity issues has not been solve so far. In this particular work, the focus towards novel materials for oxygen sensing based on Gallium Oxide (Ga2O3) and the effect of different dopants. Ga2O3 has proven to detect the presence of oxygen at high temperatures, but the properties and response can be tune by modifying fabrication method. In this case Ga2O3 based thin films deposited via RFsputtering will be analyzed, and the effect on the oxygen sensing properties will be discussed, as well as the effect of tungsten doping into its oxygen response. Furthermore, the thermal stability of the films will be evaluated by comparing the chemical-structural properties of the material before and after heat treatment, and the results will be discussed. 1. Introduction The look towards maximizing the efficiency of energy system while reducing the amount of hazardous emission towards the environment, has lead researcher and scientist interact with different technologies and components which are capable of producing a significant improvement on the energy generation processes. Among the technologies investigated, sensor and actuators have shown capabilities to significantly enhance the efficiency of energy systems, specially the combustion dependent. Furthermore, according to data provided by DOE by accurately detecting the byproducts of the coal fire power plants at temperatures >700 °C, the efficiency could be increased by 1% which could translate into $300 million/year of savings [1]. Metal oxide gas sensors have the capability of perform at the required temperatures depending on the material selection. Gallium oxide is one the materials which its properties (thermal, electrical) can withstand elevated temperatures and effectively sense oxygen quantities, which is one of the key metric for combustion completion. Gallium oxide is a wide band gap semiconductor with five polymorphs: α, β, γ, δ, and ε, where β is the most stable monoclinic phase which is thermally stable up to ~1800 °C which allows this material to operate as a high temperature oxygen sensor without any problem [2]. Nevertheless, some feature of the oxygen response for Ga2O3 can be significantly improved by the introduction of metallic dopants, in particular the response time of the sensor is been a great challenge, whereas the sensitivity, selectivity, and stability of this material need to be preserved [3].

P.108

2. Results and Discussion Figure 1.shows the XPS spectra before and after the films were exposed to high temperature heat treatment (700 °C). The effect of annealing on W- Ga2O3 sputter films on the structural and chemical aspects of the materials was presented, but the mechanism of the pore formation and W-diffusion was not discusses. These phenomena can enable via defect chemistry of the films. As the results from XPS it was determined that W+6 and Ga+3 were the main specimens inside the films, where the ionic size of them are 0.060 nm and 0.062 nm, respectively. Ga2O3 and WO3 show polymorphism, but monoclinic phase is thermodynamically prefer for both oxides, which allows W ions to go substitutional on Ga2O3 matrix. The defect chemistry of the films is presented in the following equation using KrögerVink notation:

The equation shows that the charge equilibria for W-doped films when W acts as a substitutional defect, is achieve by inducing Ga3+ vacancies on the Ga2O3 matrix. This process open paths for W ions to flow to the lower energy state, which in this case is located in the bulk portion of the film. This migration of W ions from the surface to the bulk joint with the re-evaporation of W-oxide specimens will regenerate the pore formations in the films, and the different on coefficient of thermal expansion, will generate the pore formation on the surface of the films. High level of porosity will affect our results from the XPS, more in particular the already discussed O1s peaks which shows broadening due to metal-hydroxil family, which due to pore formations an moister trapping will show higher interactions.

Figure 1. XPS spectra of W-doped Ga2O3 films before heat treatment

The oxygen sensor characteristics of the Ga2O3 based oxygen sensors, their time response and sensitivity is presented in Figure 2 along as the activation energy of the films which is directly proportional to the change in resistance on the films. The basic operation mechanism of the MOS sensors is the following

P.109

The time response of un-doped Ga2O3 was calculated to 90% of the time that the system took to complete the conductivity change when oxygen is present, and the value obtained was 67 seconds. Furthermore, in the case of W-doped films, the time response was reduce to 43 seconds, the sensitivity of the films was not significantly affected by the presence of tungsten.

Figure 2. Oxygen sensor response for W-doped Ga2O3 films

3. Conclusion W-doped Ga2O3 films showed evidence of W-diffusion when the material is exposed to a high temperature heat treatment. Furthermore, the oxygen sensor performance of the films was evaluated were the time response of the films was significantly improve. Similar the appropriate W-content for maximizing the sensing capacities of the films was obtained. References [1]

R. R. Romanosky and S. M. Maley, “Harsh environment sensor development for advanced energy systems,” vol. 8725, p. 87250H, 2013.

[2]

M. Fleischer and H. Meixner, “Gallium Oxide Thin Films: a New Material for Hightemperature,” Sensors Actuators B, vol. 4, no. 1991, pp. 437–441, 2000.

[3]

Y. Li, A. Trinchi, W. Wlodarski, K. Galatsis, and K. Kalantar-Zadeh, “Investigation of the oxygen gas sensing performance of Ga2O3 thin films with different dopants,” Sensors Actuators, B Chem., vol. 93, no. 1–3, pp. 431–434, 2003.

3 P.110

SETS2016-30 The Southwest Emerging Technology Symposium 2016 UNIFIED SOFTWARE FOR MULTI-FUNCTIONAL G-CODE: A METHOD FOR IMPLEMENTING MULTI TECHNOLOGY ADDITIVE MANUFACTURING E. Aguilera, D. Espalin, E. MacDonald, R. Wicker W.M. Keck Center, University of Texas at El Paso, El Paso, TX 79968, USA; [email protected] Keywords: Additive Manufacturing, 3D Printing, Electronics, Automation, Software Abstract Additive manufacturing (AM) started over thirty years ago and with it a manufacturing revolution that moves industrial production into the personal home. With recent interest shifting into multifunctional parts fabricated through AM technologies, unified systems are being developed. Merging different manufacturing technologies into one single machine is a challenge but undergoing research has shown promise in the development of multi-technology systems. Concurrent work is being done in the software and automation aspect of multi-technology systems. This paper explores the challenges and approaches to developing software that interfaces and processes multifunctional CADs and creates files for direct use in multi-technology AM machines. 1 Introduction In recent years, Additive Manufacturing (AM) has become a real contender in the manufacturing industry. Commonly referred as 3D printing, AM is a process which takes a Computer Aided Design (CAD) and fabricates the part in a layer by layer fashion. AM takes advantage of the freedom of design and its quick turnaround time [1]. Recently, it has been a remarkable achievement to manufacture end use parts in a layer by layer fashion and ongoing research has helped shape the process into commercially viable solutions. With these advancements in AM technology, the interest has spurred new ideas and ambition to fabricate multifunctional parts using AM. Until now most multifunctional parts being fabricated consist of a process interrupt approach that uses multiple machines and processes [2]. Alongside hardware and process effort to achieve multifunctional parts, software is also being developed to simplify and aid the user in creating multifunctional parts without human intervention. 2 Interface and Assimilation 2.1 Background Additive manufacturing has recently increased in popularity and although the CAD industry is willing and are supporting the new market, it is slow and frequently changing without a clear path. One of the first hurdles to understand when automating and developing multi-functional g-code is data extraction. Most if not all of AM files are exported in Standard Tessellation Language (STL) which includes vertex data that describes triangles that compose to define a part. Much of the information contain in a CAD is lost when converted into STL format, the industry has evolved and is pushing the Additive Manufacturing Format (AMF) file extension. AMF is written in EXtensible Markup Language (XML), allowing the file to evolve with the technology by making it simple to add or remove information in an organized manner. Although the file type (AMF) allows more information to be saved, the extra information that is needed at this moment is unknown. There has been a wide acceptance of AMF by slicing software that prepares CAD files

P.111

to be processed by the AM machine. Clearly, there is a strong foundation for support of multifunctional g-code in the future. 2.2 Strategy There is a lack of support in extracting additional information from CAD files that include different processes to construct a multi-functional part. Despite there being a multitude of CAD software, SolidWorks is currently being used for its wide acceptance in industry and availability to students. Though SolidWorks does not support data extraction for multi-functional parts, it does support macros and has a built in Application Program Interface (API). The support of an API allows the development of macros that can directly interface with the part that is being designed in SolidWorks. Having a macro grants access to most of the additional information needed to compose the multi-functional g-code. Defining the type of information and structure is dependent on the process that is being accomplish. Adding electronics to AM parts is one way to make multifunctional parts, groundwork for automatic wire embedding [3] is the first process tackled. The first step in defining a circuit in CAD is understating how the machine interprets the process. Simple sketch lines can be used to describe circuit paths and connections. Basic electronics can be hand drawn line by line, yet it is not an effective way of doing intricate electronics. Electrical engineering programs used for designing Printed Circuit Boards (PCB) are very effective at automatically routing complex paths. The paths then are imported into SolidWorks to be placed and aligned into its respective part. Making adding complex electronics simple and straight forward as seen in Fig 1 and Fig 2.

Fig.1. PCB design to DXF layers that can be exported into SolidWorks

Furthermore, every layer that includes a circuit is compromise of a sketch with a name that signifies it is a circuit and a numeric value to show how many there are (e.g., ”circuit1” or ”circuit2” and so on until all the circuits are included). This naming convention allows the macro to find the circuits and their respective heights and export them separately one at a time as DXF files defining the circuit at that level. A unified text file that has a list of all the circuits with their corresponding heights can then be created. Similar approaches can be applied with other processes to extract data and create multi-functional g-code. 3 Integration Much like in the hardware development of interfacing multiple existing technologies into one unified system, multiple software processes already developed are being merged together. For this

P.112

reason, the basic slicing software used is CURA ( an open source AM slicing software that accepts both STL and AMF files). The software, apart from being available to everyone, supports post process plugins from inside its own Graphical User Interface (GUI) making it simple to run and implement post processing for multi-functional parts. Once the exported STL or AMF file is opened in CURA, the part can be oriented and placed in the proper place at the build plate. All that is needed is a simple click of a plugin to run the post process program that uses all of the additional information extracted from the CAD to compose a multi-functional g-code file. Extracting data through the macro and not into a single file like AMF results in much of the reference information lost that would normally be natively kept. Therefore the plugin developed at this stage does more than just merge multiple processes. The plugin has to find every part and circuit (in this case) to orient them and super impose them, as the user has the ability to change the position and rotation of the part arbitrarily. In the end, a design from CAD can be manufactured without any human intervention. Fig 2 shows a test piece (wire embedding is not complete) in which a simple drawing was done in CAD and subsequently processed with it creating machine code that automatically drew the design within the build process, creating a multi-processed part from a single CAD.

Fig. 2. Image shows a CAD file compared to a AM part all in one single process no human intervention

4 Conclusion A strategy for implementing software that helps create multi-functional g-code files to be used in multi-technology AM machines was implemented. With the large industry use of CAD, but slow support of multi-functional parts in them, the need for a software package to extract and implement g-code base AM with multi-functionality is at an all-time high. Identifying the information needed and the processing was a vital step into making a cohesive strategy. References [1] E. MacDonald, R. Salas, D. Espalin, M. Perez, E. Aguilera, D. Muse, and R.B. Wicker “3D Printing for the Rapid Prototyping of Structural Electronics”. IEEE Access, Vol.2, 2014 [2] E. Aguilera, J. Ramos, D. Espalin, F. Cedillos, D Muse, R. Wicker, and E. MacDonald “3D Printing of Electro Mechanical Systems”. 24th International SFF Symposium, Austin, TX, pp 950961, 2013 [3] D. Espalin, D. W. Muse, F. Medina, E. MacDonald, and R. B. Wicker, ‘‘3D Printing multifunctionality: Structures with electronics,’’ Int. J. Adv. Manuf. Technol., Mar. 2014.

P.113

SETS2016-31 The Southwest Emerging Technology Symposium 2016

FAILURE ANALYSIS OF ELECTRON BEAM MELTED TI-6AL4V TENSILE SPECIMEN M. S. Haque1*, E. Arrieta2, J. Mireles3, C. Carrasco2, C. M. Stewart1, R. B. Wicker1 1 Department of Mechanical Engineering, 2 Department of Civil Engineering, 3 W. M. Keck Center for 3D Innovation University of T, El Paso, TX 79968, USA; *([email protected])

Keywords: Electron Beam Melt, Additive manufacturing, Fractography ABSTRACT Growing demand of complex shape components to achieve more economic and efficient power generation can be met by additive manufacturing and rapid prototyping. In this study, electron beam melting, EBM is used to produce Ti-6Al-4V cylindrical bars at zero degree manufacturing orientation. Threaded tensile specimens are machined from them. Stress-strain curves are plotted and discussed. Tensile strength of approximately 1.1 GPa is registered with approximately 10% of elongation. Fractography using optical microscopy is used to analyze the fracture surface. The Ti-6Al-4V samples exhibit trans-granular brittle fracture. 1

Introduction

Titanium alloys are widely used in medical, power generation, automotive, aerospace industries due to its light weight (4.4 gm/cm3), high mechanical strength, corrosion, and temperature resistance[1,2]. Importance of Additive manufacturing has opened the scope to overcome some manufacturing difficulties by its freedom of design, customization, and minimum waste [2]. Given the multiple variables in the process, mechanical behavior of electron beam manufactured Ti-6AL-4V is not completely understood yet. The EBM uses an electron gun to focus electron beam across the building parts guided by a .stl file from a CAD design[1]. In this study, Failure analysis of EB manufactured Ti-6Al-4V tensile samples are conducted. 2

Specimen Fabrication

Three cylindrical rods were fabricated from a CAD file with nominal dimensions of 75mm* 15mm, with a zero degree build angle (layers perpendicular to its cross section). ARCAM Ti6Al-4V prealloyed powder size ranging from 40µ-100µ is used in an ARCAM A2 system with layering thickness set at 0.3mm and melting parameters of 15mA current and 4530 mm/sec beam speed. Specimens where commercially machined at low speeds and high feed rates in a CNC lathe and sized as possible according to ASTM E8-Standard Test Methods for Tension Testing of Metallic Materials threaded specimens, resulting in an average gage length of 400mm and 5.96mm diameter. 3

Tension Test

With the intention of exploring the strength and ductility of metals fabricated by Additive Manufacturing, uniaxial tension tests are conducted, according to the ASTM E8 standard, as a widely accepted method for comparison of metals. Tensile test are performed at room

P.114

Failure Analysis of Electron Beam Melted Ti-6Al-4V Tensile Specimen

1400 1200

Stress,

(MPa)

1000 800 Specimen 1 Specimen 2 Specimen 3

600 400 200

Ti-6Al-4V tensile specimens 0 0.00

0.02

0.04

0.06

Strain,

0.08

0.10

0.12

mm)

Fig.1. Stress-Strain curve for Ti-6AL-4V tensile specimen

Fig. 2. Fractography of Ti-6AL-4V specimens, (a) Fracture surface of specimen 2, (b) specimen 3, (c) Voids and (d) secondary cracks in specimen 3.

P.115

Failure Analysis of Electron Beam Melted Ti-6Al-4V Tensile Specimen

temperature of 25.5°C in an Instron 5969 universal testing machine with an installed load cell of 50k and axial extensometer on the sample. Strain rate of 0.003mm/mm/min was use, as the alloy is expected to be strain-rate sensitive. 4

Results

It is evident from fig. 1 that two of the samples exhibit an ultimate tensile strength, UTS of 1.1 GPa. Specimen 3 failed earlier than the other two at 0.9 GPa, and exhibits less ductility. The first two specimens have strain up to 0.1 mm/mm while the third sample has strain of 0.08 mm/mm. All of the specimens exhibit linear deformation up to elastic limit and after that the stress-strain curve became almost plateau till fracture. Thus the higher elastic limit is observed is 1.0 GPa and the lower elastic limit is observed as 0.8 GPa. Further it is observed that the linear elastic line of the third sample is steeper than the other two lines suggesting that the third specimen is more brittle than the other two. All three specimen exhibits trans-granular brittle fracture, fig. 2. The fractured surface of specimen 2, figure, 2 (a) shows some ductile fracture while the specimen 3 figure 2 (b) exhibits profound fracture. Presence of voids on fractured surface is higher in specimen 3 than specimen 2. Table. 1. Comparison of mechanical properties with standards Properties Yield Strength Ultimate Tensile Strength Elongation

Requirement by ASTM-1108 (cast), GPa 0.758 0.860 >8%

Requirement by ASTM-1472 (wrought), GPa 0.860 0.930 >10%

EBM Ti-6Al-4V, GPa 0.90 1.1 >10%

5 Discussion The presence of void inside the material may be due to manufacturing defects which act as a nucleus of crack propagation and load concentration leading to fracture. Thus less ductility and tensile strength is observed for the third specimen. It is observed that the obtained mechanical properties are higher than the ASTM standards as listed in table 1. However, EBM are not free of manufacturing defects that may lead to poor mechanical properties. Acknowledgement The cylindrical EBM specimens are manufactured by Israel Segura. And the work is partially funded by NSF grant 1405526. 6

References

[1] L. E. Murr, E. V. Esquivel, S. A. Quinones, S. M. Gaytan, M. I. Lopez, E. Y. Martinez, S. W. Stafford, “Microstructures and mechanical properties of electron beam-rapid manufactured Ti–6Al–4V biomedical prototypes compared to wrought Ti–6Al–4V,” Materials characterization, Vol. 60(2), pp. 96-105, 2009. [2] L. Ladani, and L. Roy "Mechanical behavior of Ti-6Al-4V manufactured by electron beam additive fabrication." ASME 2013 International Manufacturing Science and Engineering Conference, collocated with the 41st North American Manufacturing Research Conference. American Society of Mechanical Engineers, 2013.

3 P.116

SETS2016-32 The Southwest Emerging Technology Symposium 2016

MULTI3D SYSTEM: ADVANCED MANUFACTURING WITH MATERIAL HANDLING ROBOTICS J. L. Coronel Jr.1, 2*, S. Ambriz1, 2, C. Kim1, 2, D. Espalin1, 3, R. B. Wicker1, 2 1

W.M. Keck Center for 3D Innovation, The University of Texas at El Paso, El Paso, TX 79968, USA; 2 Department of Mechanical Engineering, UTEP, El Paso, TX 79968, USA; 3 Department of Metallurgical and Materials Engineering, UTEP, El Paso, TX 79968, USA * Corresponding author ([email protected])

Keywords: additive manufacturing, robotics, fused deposition modeling, material handler ABSTRACT The emergence of additive manufacturing (AM), has allowed for the fabrication of complex parts through computer aided design (CAD) models. The Multi3D System aims to incorporate one type of AM called Fused Deposition Modeling (FDM) where polymers are melted and extruded in the shape of the pre-designed part. The system incorporates two FDM printers, a computer numerical control (CNC) router, and a Yaskawa Motoman MH50 robot arm. Via communication with LabVIEW software, the transfer of a portable build platform with the material handling robot arm, allows for the printing of material, and machining with the CNC, all within the same build. The design space is discussed, as well as the testing of the repeatability of the platform placement with the robot. The results highlight that the layer offset for the modified FDM printer for the Multi3D System, is within ±51µm (±0.002”). The offsets are comparable to a factory standard Fortus 400mc printer without modifications. 1

Introduction

Known to most as 3D printing, Additive Manufacturing (AM) is the process in which an object is built by the layer-by-layer deposition of material [1]. The use of computer-aided design (CAD) models allows for the printing of material to incorporate complex geometries. The Multi3D System employs the use of AM technology, via two FDM printers. It also incorporates a CNC Router for subtractive manufacturing, and a Yaskawa Motoman MH50 robot for material handling. Differing from traditional manufacturing methods, the Multi3D System allows for the fabrication of printed electronics through its use of varying manufacturing stations. The following sections will highlight the design space of the system, and the testing used to assess any error introduced by the material handling robot. 2 Material handling robot The material handling robot is tasked with transferring the workpiece from station to station depending on the job requirements. The design space of the Multi3D System was determined by consideration of the payload, 50kg (110.3 lbs), and the horizontal reach, 2,061mm (81.1”) of the robot [2]. The final component placement for the system can be seen on Fig.1. This was determined after considering the control theory of the robot arm with the DenavitHartenberg (DH) parameters [3], [4]. With the DH parameters, a CAD model of the system was made, and through the use of MATLAB software, the path of the robot’s motion was plotted. The simulation served to validate collision avoidance. Once installed, modifications were made to the printers to interact with the build platform. The robot’s motion path

P.117

Fig.1. Multi3D System: manufacturing stations and material handling robot

consisted of six positions on which it entered the FDM2 printer, gripped the platform, lifted and removed it, then exited. The process was reversed for platform placement. Testing was performed to determine the change in accuracy between an unmodified Fortus 400mc printer, and a customized printer that was meant to interact with the material handler and the payload. 3.0 Methodology In order to validate the use of the material handler, modifications were made to printer FDM2, to allow it to interact with the portable build platform. FDM1 was unmodified, retaining its factory settings. A three step stair was designed to allow to measure variations between machines. The base of the stair step was 50.8 x 50.8 mm (2” x 2”) with a height of 19.05 mm (0.75”). Four pauses were strategically inserted to allow for user interaction. For FDM1 the print was simply allowed to resume after the pause. For FDM2, the robot was used to remove the build platform completely from the printer, return the platform to its position, and resume printing. The printed steps can be seen on Fig.2. Data from the two stair steps was captured with an OGP SmartScope Flash 250, through dimensional measurements of the layer-to-layer offset. The data gathered for the 75 layers of each stair step was then plotted, and comparisons were made to assess the impact of the material handling robot in the accuracy of the printers. 4.0 Results The graphical representation of the data is shown on Fig.3. The distance between the edge of each layer and a reference centerline was used. The centerline was

Fig.2. Printed stair step test parts

P.118

generated by connecting the edges of layers 1 and 75. Interpreting the data, we find some difference between the prints. A tolerance of +51μm (+0.002”) and -25μm (-0.001”) was found for the unmodified FDM machine (FDM1), while the modified machine has a tolerance of ±51μm (±0.002”). In terms of percent error, there was a 33% increase in range of layer position relative to the centerline when the portable platform was moved after four pause operations. Considering the layer thickness of the material that is printed by the Fortus 400mc to be 254μm (0.01”), the offset seen can be attributed to other parameters, such as extrusion speed of the material, or temperature difference. Although some error was expected from FDM2, it was comparable to the error found in the factory set, unmodified FDM1.

Fig.3. Centerline offsets between layers on paused stair step tests

5.0 Conclusion Through simulations, a design space was created in which the robot was able to function as the material handler and successfully transfer the build platform throughout the system. A stair step test was conducted to assess the accuracy of the material handling process. Results showed that the offset found in the print conducted on FDM2, where the robot moved the platform, were comparable to the offsets of the unmodified FDM1 printer. Validation of the use of the material handler allowed for a shift in focus on the future work of the Multi3D System, where printed electronics are the desired goal. References [1] Berman, B. “3-D printing: The new industrial revolution”. Business Horizons, Vol. 55, pp 155-162, 2012. [2] Motoman “Yaskawa Motoman Robotics” https://www.motoman.com/datasheets/mh50_mh50-35.pdf, 2011. [3] Denavit, J., Hartenberg, R. “A Kinematic Notation for Lower-Pair Mechanisms Based on Matrices”. Journal of Applied Mechanics, Vol. 22, pp 215-221, 1995. [4] Lewis, F. “Robot Manipulator Control Theory and Practice”. New York: Marcel Deckker, 2004.

P.119

SETS2016-33 The Southwest Emerging Technology Symposium 2016

FABRICATION AND MODELING OF SMART PARTS USING ELECTRON BEAM MELTING ADDITIVE MANUFACTURING TECHNOLOGY Ricardo Martinez1, Mohammad Shojib Hossain1,2, Jose A. Gonzalez2,3, Mohammad Arif Ishtiaque Shuvo1, Jorge Mireles1,2, Ahsan Choudhuri1, Ryan B. Wicker1,2, and Yirong Lin1 1

Department of Mechanical Engineering, The University of Texas at El Paso, TX 79968, USA 2 W.M. Keck Center for 3D Innovation, The University of Texas at El Paso, TX 79968, USA 3 Department of Metallurgical and Materials Engineering, The University of Texas at El Paso, TX 79968, USA Key words: Additive manufacturing, piezoelectric, simulation Abstract Energy system components with embedded sensors, a.k.a. smart parts, can be a pathway in obtaining real time performance feedback and in situ monitoring during operation. The fabrication of smart parts using additive manufacturing (AM) technology allows the flexibility of embedding a sensor within the structure while not negotiating with the required shape and functionality. This paper focuses on the modeling of smart parts that were fabricated using electron beam melting (EBM) additive manufacturing (AM) technology. The simulation of a compression test, following the characteristics of the experimental part is done in ABAQUS. The experimental sensor responses showed good agreement with the applied force in four different frequency conditions, i.e., 10 Hz, 15 Hz, 20 Hz, and 25 Hz. The simulation follows same conditions of loading and frequency and results are summarized in voltage-force ratios. This research shows the advantage of the modeling which will allow accurate characterization of smart parts with complex geometry and different loading situations. 1. Introduction Sensing technologies in energy system application operating at high temperature and high pressure are in great demand, particularly in aerospace, automotive and energy industries [1]. The inclusion of intelligent components would meet requirements for decreased maintenance costs, enhanced structural safety, optimizing efficiency and improved capabilities. In aerospace applications, propulsion systems require continuous monitoring under harsh environments to optimize control and operating parameters [2]. Piezoelectric based sensors have shown to exhibit advantages over other sensing approaches with simple structures, fast response times and ease of integration [3]. This study is an initiative of embedding PZT in metallic parts for energy system components using EBM system. The embedded sensor can be non-intrusive by the parts body from the outside harsh environment. As a proof of concept, the work presented here concentrates on modeling cylindrical shape smart parts. The comparison of the sensing demonstration obtained from the smart parts are reported and it will serve as an initial part of the characterization “smart parts” for high temperature, high pressure applications. The modeling includes the dynamic analysis using ABAQUS simulation of the concept cylinder, as well as the comparison with experimental results from fabricated parts.

P.120

The Southwest Emerging Technology Symposium 2016 2. Fabrication The fabrication process includes the multiple step fabrication using EBM compared to the traditional single build to insert a piezoelectric ceramic material within the part’s body. The access of piezoelectric ceramic material at a certain position was possible due to the layer wise fabrication process using EBM. In this experiment, Arcam EBM S12 (Arcam AB, Mölndal, Sweden) was used to fabricate the smart parts. By gaining access to a non-finished product, a piezoelectric ceramic material was placed in the desired position. Figure 1 shows fabricated smart parts using the “stop and go” process. The resistance between electrodes was measured using a SPERRY DM4400A digital multimeter (SPERRY INSTRUMENTS, Menomonee Falls, WI) to ensure the open loop condition. The open loop condition ensures the smart part’s sensor capability. (b)

(a)

(c)

Figure 1 Smart parts, (a) assembled part, (b) actual fabricated part, and (c) sensor housing leg broken off for sensing purpose. A re-poling of the piezoelectric material is necessary for PZT due to exceeding the Curie temperature. The re-poling of PZT was performed by applying a constant electric field using a power supply, while maintaining it at high temperature to reduce the overall time. It was observed that an electric field as low as 1.3 kV/mm can almost fully recover the piezoelectric coefficient (d33) at room temperature, or by other combinations of heat and time during re-poling process [4]. In this experiment, the smart part was heated at 150 ºC, and electric field of 1.4 kV/mm was applied for 10 hours. Compression - compression testing set up for evaluating sensor capabilities The compression-compression test was performed to evaluate the force sensing capability of the smart part. A MTS Landmark servo-hydraulic test system (Eden Prairie, MN) was used to perform the compression-compression testing. The smart part was compressed at about 0.25 mm, which can be referred as initial position. Then, the part was oscillated at ±0.1 mm from that point at the frequency of 10 Hz, 15 Hz, 20 Hz, and 25 Hz. A DAQ system (NI PCI-6221) was used to record the voltage response. A flat cylindrical fixture was used for the compression-compression

P.121

The Southwest Emerging Technology Symposium 2016 testing. A load concentrator was used to ensure the load being applied at the center of the smart parts. The electrodes were connected with the DAQ system using a clamp to record the voltage response (V). Four different frequencies (10 Hz, 15 Hz, 20 Hz, and 25 Hz) were used to demonstrate the sensing capability across various dynamic loadings. 2. Modeling in ABAQUS The cylinder was modeled following the schematic from Error! Reference source not found.. The metallic parts –first fabrication, insert and second fabrication– were modeled with 10node quadratic tetrahedron (C3D10), global size 3.2mm, properties from Ti6Al4V grade 5 annealed [5] for density 4430 kg/m3, and isotropic properties E=113.8 GPa, v=0.342. Ceramic housing, male and female parts, were modeled using C3D10, global size 1mm, properties from alumina (Al2O3), density 3000 kg/m3, and isotropic properties E=300 GPa, v=0.21 [5, 6]. Electrodes, top and bottom, were modeled using a 20-node quadratic brick. reduced integration (C3D20R), global size 1.2 mm, properties from tungsten, density 19270 kg/m3, and isotropic properties E= 400 GPa, v=0.28 [5]. The piezoelectric component, modeled using a 20-node quadratic piezoelectric brick, reduced integration (C3D20RE), global size 0.8 mm The analysis consisted in a general piezoelectric step, assembling the parts with frictionless contact. A preload and a sinusoidal load of amplitude following the frequencies of 10 Hz, 15 Hz, 20 Hz, and 25 Hz were used for comparison purposes with experimental results The magnitude of the forces were taken from the experimental measurements of force. A schematic of the cylinder is shown in Figure 2. Applied load

Ti64

Alumina housing

Roller boundary condition

Figure 2 Schematic of cylinder modeling in ABAQUS for piezoelectric simulation response under different frequencies of sinusoidal loading. Alumina housing protects piezoelectric material (PZT-5A) and tungsten electrodes.

P.122

The Southwest Emerging Technology Symposium 2016 4. Results and Discussion Force Sensing The sensitivity (DV/DF) between voltage and force at these four frequencies was calculated and shown in Error! Reference source not found.. The sensitivity result provides a comparison to characterize the integrity of the piezoelectric inside the smart part, along with. The sensitivity of a cycle was defined as the ratio of the range of the voltage by the range of the force. The overall sensitivity is the average using all the cycles during the test. The average sensitivities are 0.53, 0.43, 0.44, and 0.42 (V/kN) for these four frequencies 10 Hz, 15 Hz, 20 Hz, and 25 Hz, respectively. The result showed a decreasing trend of sensitivity with the increase of frequency except for the sensitivity response at 15 Hz. The sampling rate at true maximum and minimum force response was not sufficient to provide the accuracy needed in higher frequency loading. Nevertheless, all sensing voltage across different frequencies showed good agreement with the frequencies of the applied force, indicating excellent force/pressure sensing capability of the fabricated smart part. 5. Conclusion The work presented in this paper showed the modeling for a sensor with embedded piezoelectric ceramic material. The simulation presented here is the first step for the overall characterization of smart parts that are built using a stop-and-go process in an EBM manufacturing process. This manufacturing technique (EBM) allow placement of sensors at desired positions, and inaccessible areas within part’s volume. The sensitivity response obtained from this experiment shows the feasibility of employing the smart part in practical applications. However, the simulation showed the need to adapt the frequency dependency on the piezoelectric coupling. Further investigation on the simulation, such as implementing thermal forces and pyroelectricity is being done to estimate the overall behavior of the sensor under different configurations and loading mechanisms. 7. References 1.

2. 3. 4.

5. 6.

Hunter, G.W., et al. Development and application of high-temperature sensors and electronics for propulsion applications. in Defense and Security Symposium. 2006. International Society for Optics and Photonics. Yang, J., A harsh environment wireless pressure sensing solution utilizing high temperature electronics. Sensors, 2013. 13(3): p. 2719-2734. Tressler, J.F., S. Alkoy, and R.E. Newnham, Piezoelectric sensors and sensor materials. Journal of Electroceramics, 1998. 2(4): p. 257-272. Prewitt, A., D., Effects of the Poling Process on Dielectric, Piezoelectric, and Ferroelectric Properties of Lad Ziconate. 2012: Proquest Dissertations and Theses, Thesis (Ph.D), University of Plorida. MatWeb, L., Material Property Data. MatWeb,[Online]. Available: http://www. matweb. com, 2014. Corp., C. 3000°F Rescor 960 Alumina. High Temperature Adhesives and Epoxies, Ceramics, Insulation, Epoxies and Epoxy 2008 11-23-2015; Available from: www.cotronics.com.

P.123

SETS2016-34 The Southwest Emerging Technology Symposium 2016

DEVELOPMENT OF THE THERMAL WIRE EMBEDDING APPARATUS FOR FDM-PRINTED PARTS 1

D. Marquez1* College of Engineering, W. M. Keck Center for 3D Innovations, El Paso, TX 79968, USA; * Daniel A. Marquez ([email protected])

Keywords: Thermal Wire Embedding, Direct Wire Embedding, Additive Manufacturing, Fused Deposition Modeling, 3D Printed Electronics ABSTRACT Current research has focused on enhancing and utilizing the fabrication freedom of 3D printing. Parts produced using Fused Deposition Modeling (FDM) have been generally used for concept modeling and structural applications. This renders the current FDM technology to manufacture limited utilization of products. Incorporating an electronic element to the layerwise FDM-built part will add multi-functionality to the part. The purpose of the current work was to develop an apparatus and method to embed a wire pattern to be used as interconnections between electronic components, an electromagnetic device, a heating element, a heat dissipation element, and/ or use a wire pattern as a mechanical reinforcement on the interlayers and/or on the exterior of an FDM printed part. 1

Introduction

The fabrication of parts through the deposition of a material using a print head, nozzle, or another printer technology is known as 3D printing. Fused Deposition Modeling (FDM) is the most popular kind of Additive Manufacturing (AM) technology, which uses a material extrusion process utilized to make thermoplastic parts through heated extrusion and deposition of materials layer by layer. [2] These 3D printers can fabricate complex geometrical parts that may not be achieved with traditional means of fabrication such as subtractive manufacturing (i.e., milling, lathing) or casting. Although this is a great alternative for such complex geometrical parts, there are still several attributes required to be able to produce end-use products. Recently there have been many efforts focused into providing more functionality to FDM-printed parts on top of them being used for structural applications. A commercially available 3D electronics printer has been released, self-proclaiming it to be the first of its kind, the printer consists of a FDM based printer with an integrated micro dispensing technology. [5] The 3D electronics printer allows electronic components embedded within the part’s layers to be interconnected with the use of conductive silver ink. The use of conductive inks has shown its limits to only operating for low power applications. Even though much work has focused on improving on the conductivity of the inks, the limitations are due to the high resistance found in conductive inks and new materials are required to improve on the limitations. [4] Substituting conductive inks for a more conductive material allows to integrate high powered electronic components into FDM parts. Copper is traditionally used in integrated circuits, printed circuit boards, and electrical wiring due to its very high electrical conductivity. [3]

P.124

Utilizing copper wire on FDM-printed parts gives the ability to incorporate higher powered electronic capabilities. In this paper, we introduce the thermal wire embedding apparatus. This direct wire embedding tool is used to submerge a wire on the interlayer(s) and/or on the exterior surfaces of an FDM-printed part. The embedded wire can be used as a mechanical reinforcement, analogous to carbon fiber used as a reinforcement in polymers, and as interconnections for electronic components, an electromagnetic device, a heating element, or a heat dissipating element. 2

Design and embedding method of the thermal wire embedding tool

The thermal wire embedding apparatus is a fully automated tool used to embed wire onto FDM-printed parts. The thermal wire embedding tool is driven by direct heat conduction, controlled by a temperature controller. The heat conducts to the wire intended to be embedded as well as the surface of the part which is being embedded into. In order to drive the wire, the tool utilizes bidirectional driving motors to control wire feeding. The tool also features a wire cutting mechanism consisting of an electric push solenoid and off-the-shelve blades. The following figure shows the components installed on the thermal wire embedding tool.

Fig.1. A CAD model (left) and front view (right) of the thermal wire embedding tool.

Using G-code, a motion control system is required to install the apparatus and carry out the toolpaths. The method developed with this tool is illustrated in figure 2. The method is as follows, once the thermal wire embedding head approaches the plastic surface, the orifice is to submerge into the thermoplastic surface then dwell in order to melt or soften the plastic. The motors are then signaled to operate, which will allow the wire to displace from the tip’s orifice so that the wire is submerged further into the softened plastic. After the wire is driven further into the substrate, the thermal wire embedding head then moves away from the surface of the plastic at the same speed that the motor is still driving. Once the head has moved away from the surface and halted the motion and wire driver, a blast of cool air from the convective cooling hose is required in order to help solidify the softened plastic and encapsulate the wire. This process creates the initial stake of the wire which will provide the strength required in the solidified plastic with wire to help mitigate the wire from detaching as the thermal wire embedding tool advances from the starting point. The heated tip will then

P.125

DEVELOPMENT OF THE THERMAL WIRE EMBEDDING APPARATUS FOR FDM-PRINTED PARTS

be brought back to the plastic surface, with the feed motors engaged simultaneously as the thermal wire embedding head moves, and immediately traverse as it comes into contact with the surface away from the initial point and continue along its desired toolpath. If a corner or the trace is terminating, a similar step as the initial stake of a trace is required. The final point, with stake and provided cooling, will conclude with actuating the cutting mechanism in order to terminate the trace.

Fig.2. Illustration of wire embedded on a surface.

3 Results The thermal wire embedding tool was used to embed 28 gauge (ϕ=0.0126”, 0.321mm) Nickel Chromium (Ni Cr) wire in order to analyze the mechanical property modifications. Using the ASTM D 638 standard test method for tensile properties of plastics, the tests consisted of tensile testing Z-built Polycarbonate (PC) dog bones. A control set of tensile specimens were not embedded with wire while another set of PC specimens were embedded with the same 28 gauge Ni-Cr wire pattern. The average results of the PC Z-built dog bones with embedded 28 gauge Ni-Cr wire demonstrate an increase in Young’s Modulus of 45%, and increase in UTS of 28%, and an increase in elongation of 11%. The average results of the tensile tests are shown in figure 3.

Fig.3. Tensile testing results.

4 Conclusion/Future Work The thermal wire embedding tool has shown that the ability to embed solid wire can enhance the functionality of a FDM-printed part. The wire can be used as a mechanical reinforcement and as interconnections for electronic components. Future work consists of implementing a CAD pattern-to-gcode software and further improving on the design and performance of the thermal wire embedding apparatus.

3 P.126

References [1] ASTM D638-14, Standard Test Method for Tensile Properties of Plastics, ASTM International, West Conshohocken, PA, 2014, www.astm.org [2] ASTM F2792-12a, Standard Terminology for Additive Manufacturing Technologies, (Withdrawn 2015), ASTM International, West Conshohocken, PA, 2012, www.astm.org [3] "Copper, Cu; Cold-Worked." Online Materials Information Resource – MatWeb [4] D. Roberson, "A novel method for the curing of metal particle loaded conductive inks and pastes" (January 1, 2012). ETD Collection for University of Texas, El Paso. Paper AAI3512749. http://digitalcommons.utep.edu/dissertations/AAI3512749 [5] “Voxel8: 3D Electronics Printing.” http://www.voxel8.co/

P.127

SETS2016-35 The Southwest Emerging Technology Symposium 2016

DEVELOPMENT OF HIGH-TEMPERATURE DIGITAL IMAGE CORRELATION METHOD Christopher Ramirez1, Calvin M. Stewart, Ph.D1 1The University of Texas at El Paso Keywords: high-temperature, digital image correlation, 304 stainless steel, nickel superalloy ABSTRACT The purpose of this study is to develop a digital image correlation (DIC) technique capable of acquiring non-contact strain measurements at temperatures up to 1200°C. At high temperatures, DIC is a challenge because metallic materials glow incandescently in the form of thermal radiation. Thus, a high-temperature DIC technique is required. This study proposes utilizing a split-tube furnace to heat tensile specimens, which are illuminated with white light and ultimately monochromatic blue LED spotlight. A bandpass filter allows only the monochromatic light to be detected by the CCD sensors of the DIC cameras, thus reducing the observed incandescence. To validate this DIC technique, monotonic tensile tests are conducted at room temperature and 600°C on placeholder specimens of 304 Stainless Steel (304SS) and at room temperature for 3D printed compact tensions specimens. Direct strain measurements are obtained with a high-temperature extensometer and will be compared with the non-contact strain measurements obtained with the proposed DIC method. Ultimately, full-field thermal deformation of the specimens will be determined and the acquired images will be compared against reference DIC images illuminated with white light at room temperature. 1

Introduction Measuring the deformation of materials is a crucial element of material characterization that allows engineers to obtain a deeper understanding of the mechanical properties of materials. Digital image correlation (DIC) is an optical technique which tracks miniscule deformations the surface of an object by comparing images taken consecutively during mechanical testing. For our purposes, DIC is especially useful in characterizing materials subjected to extreme environments because it does not require direct contact to determine the strain characteristics of a material. However, a limitation of DIC is observed when the temperature of a specimen surpasses 500°C. In this temperature regime, most materials begin to “glow” incandescently as a result of thermal radiation. That is, there is enough electromagnetic energy generated by the thermal motion of particles within the material to be seen with the naked eye. Because DIC relies heavily on visually tracking the surface of a specimen, any observed incandescence can substantially hinder strain measurements. Thus, if a material is to be tested with DIC at high temperatures, it is necessary to develop methods to counteract its incandescent thermal radiation. The purpose of this research project is to develop a high-temperature DIC technique capable of tensile testing metal alloys at temperature up to 1200°C. Nickel-based superalloys are of particular importance because they exhibit a remarkable combination of hightemperature strength, toughness, and resistance to degradation in corrosive or oxidizing environments[1]. Because of their superior properties, Nickel-based superalloys are widely used in the aerospace field as materials in gas-turbine engines; thus, fully understanding the

P.128

deformation and failure characteristics of these materials is crucial to developing predictive models and making stronger materials. The proposed DIC method in this study will validated in two ways. First, hightemperature monotonic tensile tests will be conducted for 304SS round specimens and surface strain will be measured directly with an extensometer. These strain measurements will be compared against surface strain measurements taken with the proposed DIC method at high temperatures under the same monotonic tensile test parameters. Second, reference DIC images will be taken at room temperature with white light for compact tension (CT) specimens. These images will be compared with DIC images taken with monochromatic blue light and the two sets of images will be compared; the two sets of images should have comparable surface strain measurements if the method is valid. 2

Methodology

2.1 Materials Common materials used for turbine engines are isotropic Nickel-based superalloys such as: Hastelloy X, RA333, Nimonic 263, HS-188, GTD-222, and IN617.[2] However, before conducting full-scale DIC and tensile tests on Nickel superalloys, this study aims to conduct high temperature DIC tests on two placeholder specimen: 304 Stainless Steel (304SS) round specimens and 3D-printed compact tension specimens. The material 304SS is an austenitic FeNi-Cr stainless steel that has been used extensively in the power generation and pressure vessel and piping industry. It has been prepared to meet ASTM standards A276 and A479[3,4]. The material 304SS was selected because its mechanical behavior is well-documented; by comparing the tensile properties we acquire against data from literature, we determine if the proposed DIC method is valid. The round tensile specimens are manufactured according to ASTM standard E8[5] with a gauge length of 1.0 inch and diameter of 0.25 The CT specimen is 3D printed with a Makerbot Replicator 3D Desktop printer using PLA filament. The CT specimen is chosen to act as a placeholder because a mock tensile test is necessary to determine if the experimental setup is capable of performing CT fracture toughness tests. Fracture toughness is a measure of how well a material can resist fracture in the presence of a sharp crack while the material is under linear-elastic stress. The fracture toughness tests conform to ASTM Standard E399[6]. The dimensions of the specimen are shown in Figure 1.

Fig. 1 Dimensions of the CT specimen, conforming to ASTM standard E399.

P.129

2.2 Methods The tensile tests are conducted using an Instron 5969 universal test machine equipment with a 50kN ±100N load cell. Attached to the grips of this load frame are two pull rods (16” 3/16” long each) onto which the round specimens will be threaded and the CT specimens will be mounted with Clevis pins. The long rods are used to keep the Instron grips and load cell cool. An ATS Model 3210 split-tube furnace is used to heat the specimens with three vertically aligned and individually controlled heating zones capable of reaching a maximum working temperature of 1200°C. The furnace is equipped with a 1-inch diameter window that provides an unobstructed view into the interior of the furnace. The furnace is mounted onto the frame of the Instron 5969 machine. An Epsilon Model 3448 high temperature extensometer is used to directly measure displacement of the specimens (shown in Figure 2), while indirect strain is measured by the VIC software from Correlated Solutions. The extensometer is capable of operating at temperatures up to 1200°C and is covered with ceramic sleeves to prevent small disturbances caused by heat turbulence. A Type-K thermocouple is directly welded to the 304SS specimen to measure surface temperature and ensure accuracy of oven temperature. The Instron frame is centrally controlled by a desktop computer running the Bluehill Software (version 3.54) control system. Load, displacement, extension, and temperature are controlled and data is stored. The specimens will be illuminated with a monochromatic blue LED spotlight manufactured by Mightex with a peak wavelength of 455 nanometers. A bandpass filter manufactured by MidOpt with a peak transmission of 470 nanometers will be used in conjunction with this light source and attached to the camera lens of a Grasshopper digital camera. The combination of the monochromatic blue light and bandpass filter will allow only a narrow band of blue light to pass through to the CCD sensors of the camera, thus effectively eliminating the incandescent glow of the heated specimen.

Fig. 2 Experimental setup of 304SS round specimen in oven. A thin thermocouple wire is welded on the left side of the specimen and the free-standing extensometer is attached to the right via ceramic cords under tension.

To characterize the behavior of 304SS, monotonic tensile tests are conducted at room temperature and 600°C. The monotonic tensile tests subject the specimens to a constant strain rate of 0.001 1/s until the specimen ruptures. Monotonic tensile tests are used to determine fundamental tensile properties (modulus of elasticity, ultimate tensile strength, and yield

P.130

strength). To determine an appropriate strain rate for the fracture toughness tests for the CT specimen, the work of Gollerthan et. al[7] is consulted. This group tensile tested CT Nickelbased materials, albeit not at high temperatures. From their work, a strain rate of 0.05 1/min is chosen. The mock CT tensile test was conducted in white light because the monochromatic blue light was not available at the time. Regardless, it is valuable to know if the DIC system is capable of acquiring accurate strain measurements using the furnace setup. 3

Results The results of the monotonic tensile tests for 304SS is shown in Table 1. A stress vs. strain graph for all these tests is presented in Figure 3. These measurements agree well with the given values for this material[8,9]. Images acquired with DIC for the 3D-printed CT specimen are shown in Figure 4. Elastic modulus

Yield strength, 0.2%

Ultimate tensile

(GPa)

offset (MPa)

strength (MPa)

22

1.085

396

662

600

1.045

199

345

Temperature (°C)

Table 1. Mechanical properties of 304SS obtained from monotonic tensile test at a strain rate of 0.001 1/s.

Fig 3. Stress vs. Strain curves of monotonic tensile tests for 304SS at room temperature and 600°C. 11E-4 10E-4 9.0E-4 8.0E-4 7.0E-4 3.0E-4 1.0E-4 0 -2.0E-4 -4.0E-4

Fig. 3 Linear strain measurements (εyy) of CT specimen acquired with DIC at room temperature and with white light.

P.131

4

Conclusion The tensile properties acquired for the 304SS specimens agree well with literature, indicating the monotonic tensile test setup is valid. Therefore, acquiring tensile properties for Nickel superalloys should be accurate as well. For the mock CT tensile test, the DIC images acquired indicate the setup is capable of providing accurate strain measurements. A continuation of this study will involve acquiring DIC images for both the CT specimens and the 304SS round specimen with the monochromatic blue light. If this setup is valid, then strain measurements acquired with the monochromatic blue light should be comparable with what we currently see with the white light setup. References [1] Pollock, T., and Tin, S. “Nickle-based superalloys for advanced turbine engines: chemistry microstructure and properties”, Journal of Propulsion and Power, Vol 22, 2010 [2] Schilke, P. W., Foster, A. D., Pepe, J. J., and Beltran, A. M., 1992, “Advanced Materials Propel Progress in Land-based Gas Turbines ,” Advanced Materials and Processes, Vol 141, No 4, pp. 22-30. [3] American Society of Testing and Methods, "ASTM Standard A276/A276M Specification for Stainless Steel Bars and Shapes," ASTM International, West Conshohocken, PA, 2015. [4] American Society of Testing and Methods, "ASTM Standard A479/A479M Specification for Stainless Steel Bars and Shapes for Use in Boilers and Other Pressure Vessels," ASTM International, West Conshohocken, PA, 2016. [5] American Society of Testing and Methods, "ASTM Standard E8/E8M Test methods for tension testing of metallic materials," ASTM International, West Conshohocken, PA, 2015. [6] American Society of Testing and Methods, "ASTM Standard E399 Standard test method for linear-elastic plane-strain toughness of metallic materials”, ASTM International, West Conshohocken, PA, 2015. [7] Gollerthan, S., Herberg, D., Baruj, A., "Compact tension testing of martensitic/pseudoplastic NiTi shape memory alloys," Materials Science and Engineering: A, Vol. 481-482, 2008, pp. 156 159. [8] American Society of Testing and Methods, “ASTM Standard A276 Standard specification for steel bars and shapes”, ASTM International, West Conshohocken, PA, 2016. [9] Simmons, W.F.,Van Echo, J.A. “Report on the elevated temperature properties of stainless steels.” American Society for Testing and Materials, Philadelphia, PA, 1965

P.132

SETS2016-36 The Southwest Emerging Technology Symposium 2016

OPTICAL SECOND HARMONIC GENERATION IMAGING FOR FERROELECTRIC MATERIALS STUDIES Yu Ding, Carlos Diaz-Moreno, Aurelio Paez, Yongdong Wang, Jorge A. López, Chunqiang Li* Department of Physics, University of Texas at El Paso, El Paso, TX 79968, USA; * Corresponding author ([email protected])

Keywords: Second harmonic generation imaging, ferroelectric domain, lithium niobate

ABSTRACT Optical Second Harmonic Generation (SHG) imaging microscopy is developed as an alternative to fluorescence imaging to observe unique material properties. The polarizationdependent SHG responses of ferroelectric materials are studied. From the measured polar response of LiNbO3 ceramics materials, we are able to find the orientation of the ferroelectric domain. The SHG responses of such materials under different polarizations, including linearly and circularly polarized excitations, are examined. Fine structures of the material are revealed by studying the SHG intensity profiles. Scanning electron microscope (SEM) images are taken as a comparison to SHG images. 1

Introduction

1.1 Introduction to SHG The optical response of a medium to an applied electromagnetic field can be expressed in terms of its polarization density P(t). In a linear dielectric medium, the polarization can be written linearly as P(t) = 𝜒𝜒(1) E (t)

(1)

Here χ (1) is the linear electric susceptibility, and E(t) is the applied electric field. In a nonlinear material, the polarization density can be expressed as Taylor expansion P(t) = P(1) (t)+ P(2) (t)+…+ P(n) (t) = 𝜒𝜒(1) E(t)+ 𝜒𝜒(2) E2(t)+…+ 𝜒𝜒(n) En(t)

(2)

By replacing E(t) with E(eiωt +c.c.), we can find that a polarization field with double frequency 2ω is generated. In this process, the nonlinear response of the material to the external electric field of the excitation light leads to radiations at doubled frequency. This process is thus called frequency doubling or second harmonic generation. Second harmonic generation is a nonlinear optical process only generated in media that are lack of inversion symmetry [1]. 1.2 Introduction to SHG imaging SHG imaging has been an established tool for visualization of materials such as nanoparticles, nanowires and biological samples [2-4]. Unlike the fluorescence signals from

P.133

samples, SHG signal is coherent and stable within a broad range of excitation laser wavelength. Since SHG does not involve any real-state transition, the problems with fluorescence microscopy such as photobleaching and blinking are circumvented. Furthermore, SHG can only be generated in materials that are without centrosymmetry. Therefore, SHG imaging provides high contrast in materials that have high polarization and noncentrosymmetric structures. As a result, the polarization-dependent measurement of the SHG signal can be used to study the material structures in a fast and nondestructive way [5]. 1.3 Ferroelectric effect Ferroelectric effect is a property of certain materials that have a spontaneous electric polarization that can be reversed by the application of an external electric field [6]. Like SHG, ferroelectric effect is observed in symmetry-breaking media. As a result, SHG is a powerful tool of detecting material structures, phase transitions, lattice rotations and electromagnetic coupling in multiferroic materials in a fast and noninvasive way [7,8]. Lithium niobate (LiNbO3) is a ferroelectric oxide crystal with an ABO3 type that has been attracting lots of attentions to its pyroelectric, piezoelectric, electro-optical properties. It has many applications such as birefringence in acoustic transducers, linear acoustic relays, beam deflectors, phase conjugators, dielectric wave guides, memory storage and holographic data processing [6]. 2

Experimental setup

The scanning SHG microscope is based on a femtosecond Ti:Sapphire laser source (SpectraPhysics, Mai-Tai HP). Its pulse duration is about 100 fs width at a repetition rate of 80 MHz. The wavelength of the laser is tunable from 690 nm to 1040 nm, with the maximum power up to 2.5 W. Here we use wavelength at 900 nm to generate the SHG at 450 nm, which corresponds to blue color in the spectrum. The schematic illustration of the experiment setup is shown in Fig. 1.

Fig. 1 Experimental setup for SHG imaging

The fundamental laser beam is tightly focused by a 60× NA 1.2 water-immersion objective lens (Olympus LUMPlanFLN). The generated SHG signal is collected by the same objective lens in an epi-detection geometry. The signal is separated from noise by a dichroic mirror and a 20 nm narrow band pass filter centered at 450 nm, and then is detected by a photomultiplier tube (PMT).

P.134

3

Results and discussion

LiNbO3 was prepared by mechanical milling using lithium carbonate (Li2CO3) and niobium oxide (Nb2O5) as precursors. The milling was performed for 8 hours and it was done in order to reduce the particle size. The particles were bound with PVA and were pressed at 105 kg/cm2 and then sintered at 1000ºC for 5 hours. a

c

b

Fig. 2 SHG image of LiNbO3 at horizontal (a), and vertical polarization (b). SEM image of the same sample (c).

Fig. 2(a) shows the SHG image at horizontal linear polarization (0°). Fig. 2(b) shows the SHG image when the laser polarization is rotated by 90°. At 0° polarization we observe the darkest SHG signal at region A and the brightest at region B. On the other hand, at 90° polarization we observe the brightest SHG signal at region A and the darkest at region B. It clearly shows that region A and region B have perpendicular ferroelectric domains. Furthermore, stripe pattern is also revealed in these regions. The observed SHG intensity profile verifies the existence of fine structures of the material. A scanning electron microscope image is taken for a comparison to SHG images as shown in Fig. 2(c). In summary, we examine the SHG response of the materials under different excitation polarizations. Domain structures are revealed by the SHG images. SEM images are taken as a comparison to SHG images. The SEM image shows that the sample of LiNbO3 was sintered mostly in its crystalline phase. In future experiment we propose a lower temperature sintering for better densification. Future works also include applying the SHG imaging to different materials. Lanthanum doped LiNbO3 is one of the materials we are testing and it shows promising results. Acknowledgement This work was supported by NSF Grant #1429708. We also acknowledge funding from NSF Grant #1205302 for supplies. CDM was supported by Consejo Nacional de Ciencia y Tecnologia-México through Postdoctoral Abroad Program, Solicitation #250381. References [1] [2] [3] [4] [5] [6] [7] [8]

Robert W. Boyd. “Nonlinear optics”. 3rd edition, Academic Press, 2008. Antonio Capretti, et al. Phys. Rev. B 89, 125414, 2014. Godofredo Bautista, et al. Nano Letters, 1564, 2015. R. M. Williams, W. R. Zipfel, and W. W. Webb, Biophys. J. 88, 1377-1386, 2005. Manfred Fiebig, Victor V. Pavlov, and Roman V. Pisarev. J. Opt. Soc. Am. B 22, 96-118, 2005. Volk, T., Wöhlecke, M. “Lithium Niobate”, 1st edition. Springer-Verlag: Berlin, Germany, 2008 M. Fiebig, Th. Lottermoser, D. Fröhlich, A. V. Goltsev, R. V. Pisarev. Nature (London) 419, 818, 2002. Choi, K. J., et al. Science 306, 1005-1009, 2004.

3 P.135

SETS2016-37 The Southwest Emerging Technology Symposium 2016

METAMATERIAL BASED PASSIVE WIRELESS TEMPERATURE SENSOR FOR TEMPERATURES UP TO 500°C

H. Karim1, D. Delfin1, L. A. Chavez1, L. Delfin1, J. Avila2, C. Rodriguez2, R.C. Rumpf2, Y. Lin1*, & A. Choudhuri1 1 Department of Mechanical Engineering, The University of Texas at El Paso, TX 79968, USA; 2 Department of Electrical Engineering, The University of Texas at El Paso, TX 79968, USA; * Corresponding author ([email protected])

Keywords: keywords list (no more than 5) ABSTRACT The recent development of metamaterials has inspired substantial amount of research in many fields. The temperature dependent resonance properties of metamaterial make it suitable for use as a temperature sensor. There is an increasing need of temperature sensors suitable for sustaining at high temperature and harsh environments. However, active sensors is not a suitable solution for that as they are limited by the energy storage devices needed to power them. In this paper we propose a passive wireless temperature sensor suitable for harsh environments. The sensor is a combination of metal CRR (Closed Ring Resonator) arrays with a dielectric ceramic material. The metals are embedded inside the dielectric to protect them from harsh environments. The resonance frequency of the structure depends on the temperature and enables its functionality as a temperature sensor. 1

Introduction

Metamaterials are man-made materials, which can display properties that are otherwise absent in nature. The materials are usually arranged periodically to duplicate the structure of an atom. Depending on the shape, size, orientation, and arrangement, metamaterials can show different exclusive properties such as negative refractive index [1], cloaking [2], and reverse Doppler effect [3]. Wireless passive sensors are getting more and more attentions in the industries. Optical based wireless sensors were developed but the accuracy of these sensors was not satisfactory. Metamaterials were introduced to remove these limitations. Ekmekci et al. demonstrated the feasibility of different types of SRR structures for different types of sensors. They suggested broadside-coupled SRR structure for temperature, humidity and concentration sensor application. The objective of this paper is to propose a metamaterial-based temperature sensor that is able to work in harsh environments such as combustion chambers and will be cheap and easy to replace. The proposed sensor has two closed metal ring resonators embedded in a dielectric material matrix, which separates the two resonators as depicted in Figure 1(a). The dielectric matrix surrounding the metal rings helps protect the metals from harsh and corrosive environments. The whole structure acts as an LC resonance circuit. The equivalent circuit is also suggested in Figure 1(b). The resonance frequency of the structure can be expressed as:

P.136

1

(1) 2π LC where, f is resonance frequency. The simplest expression for capacitance can be given as: A (2) C = ε 0εr d where, ε0 is the relative permittivity of vacuum, εr is the relative permittivity of the dielectric matrix, A is the area of one ring resonator and d is the distance between the two resonators. The detailed modeling and parameter evaluation of the sensor can be found elsewhere.

f=

(a)! !!

!!

!!

Side view !

(b)! !!

Metal rings!

!! !!

Fig.1. (a) Proposed model of the temperature sensor (b) Equivalent circuit 2 Fabrication Barium Titanate (BaTiO3) was selected as the dielectric material and Polyvinyl alcohol (PVA) was chosen as the binder because of their linear dielectric properties with temperature. However, the significant loss values of BaTiO3 creates a problem as a significant amount of energy is absorbed by the material. To reduce the loss of BaTiO3, Boron Nitride, which also has high melting point was mixed with the BaTiO3 and PVA system. Using the Bruggeman principle of mixing Samples of different thickness and sizes as well as with different configurations of Cu washers were fabricated and tested to find out the optimal performance. Fabricated sample are shown in Figure 2. (e)

(f

Figure 2. Die punch fabricated samples, (a) diameter=1 inch, side view (b) diameter=1 inch, top view, (c) diameter=2.75 inch, side view, and (d), diameter=2.75 inch, top view; (e) and (f) represent a set of four 2 in/side samples in top and front view, respectively. 3 Testing and results Free space measurements were performed on the fabricated samples in the cSETR facility by utilizing a set of Gaussian beam antennas, a vector network analyzer and two co-axial cables,

P.137

as seen in Figure 3. The frequency range used was 7 – 17 GHz. The transmission spectrum of the sample was obtained by sending an electromagnetic wave from the interrogating antenna and after material-wave interaction, received by the second antenna. The testing results can be observed in Figure 3. The resonance peak is observed at 14.33 GHz. Metamaterial sensor−Transmission−Ungated 5 0

Transmission S1,2 (dB)

−5 −10 −15 −20 −25 −30 −35 −40 8

9

10

11

12

Freq (GHz)

13

14

15

16

17

Fig 3. Room temperature response of the Metamaterial temperature sensor 4 Conclusion In this paper the concept of a passive wireless temperature sensor using metamaterial has been discussed. A cheap method of fabrication using traditional die-punch was used. A mixture of Boron Nitride with BaTiO3 proved to be useful to reduce the loss of the system and provide with sufficient response. The room temperature characterization of the sensor using a pair of Gaussian beam antenna shows a defined resonance peak at 14.33 GHz. The future work will involve testing at different temperature and measuring the sensitivity. References [1] Smith, D. R., J. B. Pendry, and M. C. K. Wiltshire. "Metamaterials and negative refractive index." Science 305.5685 (2004): 788-792. [2] Cai, Wenshan, et al. "Optical cloaking with metamaterials." Nature photonics 1.4 (2007): 224-227. [3] Lee, Sam Hyeon, et al. "Reverse Doppler effect of sound." arXiv preprint arXiv:0901.2772 (2009).

3 P.138

SETS2016-38 The Southwest Emerging Technology Symposium 2016

COMPRESSIVE PROPERTIES OF MOCK POLYMER BONDED EXPLOSIVE USING DIGITAL IMAGE CORRELATION C. A. Catzin1, C. M. Stewart1* 1

Department of Mechanical Engineering, University of Texas at El Paso, El Paso, TX 79968, USA; * Calvin M. Stewart ([email protected])

Keywords: Mock polymer bonded explosive, solid rocket propellant, Uniaxial Compression, Digital Image Correlation ABSTRACT The mechanical behavior of a mock polymer bonded explosive (PBX) called “Miner mock” was experimentally studied using three-dimensional digital image correlation (3D-DIC) method. Uniaxial compression tests were performed to study the mechanical response of the Miner mock under compressive loading conditions at room temperature. The displacement and strain fields on the surface of the specimen were recorded using a 3D DIC system. Based on the contour plots produced, key features of each loading conditions were identified and recorded. 1 Introduction The concern for safety in explosives handling, storage, and transportation has led to the development of Insensitive Munitions (IM). One of the most successful IM is the polymer bonded explosives (PBXs). Polymer bonded explosives (PBXs) are a complex class of particulate composite materials that are formed by two constituent materials: micron size energetic crystals in a polymer binder material [1]. The content of energetic crystals to polymer binder material differs greatly. The energetic crystals typically compromises about 50 to 98% of the total mass of the composite, depending on the desired explosive output [2]. PBXs possess a wide variety of applications ranging from rocket propellants to the main explosive charge in traditional ammunitions and weapons systems. The origin of PBXs resides in the aerospace industry. The development of PBX binders systems was derived from solid rocket propellant technology. PBXs are an exceptional option for solid rocket propellant, however there exist many gaps in the understanding of their behavior. This study will characterize the compressive properties of a mock PBXs known as Miner mock. A mock PBX has the capability to reproduce the mechanical behaviour of a PBX closely without having any issues related to an accidental explosion [3]. This feature allows the safe study of PBXs which are crucial to fully understand their unique mechanical properties and provide reliable data for constituent models development and enhance its application as a reliable solid rocket propellant. 2 Materials and Test Methods 2.1 Material The mock PBX material known as Miner mock has a formulation containing approximately 50 wt% micron size soda lime glass beads in a bimodal size distribution and 50 wt% high impact polystyrene polymer binder. The Miner mock specimens were manufacture using a hot pneumatic press, typical laboratory mixing equipment, and a silicone release agent. The manufacturing process consisted in mixing the constituent materials, the agglomeration effect of micron(or smaller) particles was exploited during this stage, followed by the application of constant heat and pressure (beyond the melting point of the HIPS) to fully coat the soda lime glass beads. The output of manufacturing was a right circular cylinder of 1.5 inches in

P.139

Compressive Properties of Mock Polymer Bonded Explosive using Digital Image Correlation diameter and 1.85 inches in length. Testing specimens were machined from the asmanufactured cylinders to meet testing specifications. 2.2 Equipment All testing was conducted on an INSTRON 5969 Universal Testing Machine at a temperature of 25°C. The data captured from the INSTRON machine during the tests included time, load, and load head displacement. The three-dimensional digital image correlation (3D DIC) equipment used in this study is a Correlated Solutions VIC-3D DIC system which performs in-situ measurement of the surface displacement and strain fields of specimen during mechanical testing. Prior to correlation, the specimen is primed with a randomly applied speckle pattern that will act as reference points. The speckle pattern was applied using flat/matte general purpose spray paint. A white layer of paint is deposited as a background and topped with black speckles to generate a high contrast speckle pattern. The 3D DIC software, VIC 3D, tracks the displacement of reference points, compares the displacements to a physical reference, and calculates strain using continuum mechanics. A 3 mm calibration pad was used to calibrate VIC 3D for each specimen. 2.3 Testing Method The uniaxial compression test configuration consists of two 6 inches diameter compression platens mounted to the Instron frame as illustrated in Fig.1. The uniaxial compression specimen did not required extensive machining other than polishing the contact surfaces on a PoliMet 1000 using a 200 silicon carbide grid to ensure proper contact. It is important to avoid misalignment by placing the specimen at the center of the compression platens right under the loading axis as indicated in the schematic. Friction can develop at the contact surface leading to a multiaxial state of stress and undesired failure mode. The adaptation of PTFE (poly-tetrafluoro-ethylene) lubricant or tape to reduce friction during testing has been proven successful [4]. The PTFE provides a smooth low coefficient of friction surface between the compression platen and the specimen. Compression testing procedures act in accordance with ASTM D695-10. The cylindrical specimens were compressed at a displacement controlled test rate of 0.14 mm/ min. Platen

D L

Fig.1. Quasistatic Uniaxial Compression Test Configuration with a specimen diameter (D) and a length (L) 3. Results and Analysis The stress-strain curve for uniaxial compression is depicted in Fig.2. The elastic region of the miner mock was determine to end at 0.042 of strain. The elastic modulus was calculated to be 462 MPa. Comparing the elastic modulus obtained in this study to literature, the miner mock behaves similar to the X0242 PBX simulant [5]. The yield point was calculated using the 0.2 % offset. The corresponding value of strain at the yield point is 0.044 and 27.5 MPa of strain and stress, respectively. The plastic region seems to be quite extensive as it starts evolving from 0.044 of strain to 0.081 of strain to the rupture point. The compressive strength is determined to 29 MPa. The strain value corresponding to the compressive strength is 0.06 of 2

P.140

Compressive Properties of Mock Polymer Bonded Explosive using Digital Image Correlation strain. Total fracture occurred at 0.09 of strain. Is possible to perceive a softening effect Stress Strainby the air voids within the specimen around 0.01 of strain. This softening effect is vscause which allow the displacement with a low load.

Stress, MPa

30

20

10

0 0.00

0.02

0.04

0.06

0.08

0.10

0.12

Strain, mm/mm

Fig.2. Stress vs strain curve of uniaxial compression specimen The deformation and fracture behavior of the miner mock specimen under compressive load is depicted in Fig.3. The strain in the x and y direction are shown prior to fracture. The specimen is suffering high deformation along the longitudinal and transverse axis indicated by the stress concentrations marked by the red spots. Fig.3 B indicate that the specimen will fail by shear as stress concentrations localize at the red spots. The specimen failed as predicted. B A

Fig.3. Strain contours provided by VIC 3D (A) Strain in the x direction (B) strain in the y direction. References [1] Cooper, P. W., and Kurowski, S. R., 1996, “Introduction to the Technology of Explosives”. New York: Wiley-VCH. [2] Cady, C. M., Liu, C., Rae, P.J., and Lovato, M. L., 2009, “Thermal and Loading Dynamics of Energetic Materials,” Proc. of the SEM Annual Conference, Albuquerque, NM. [3] Siviour, C. R., Laity, P. R., Proud, W. G., Field, J. E., Porter, D., Church, P. D., Gould, P., and HuntingdonThresher, W., 2008, “High strain rate properties of a polymer-bonded sugar: their dependence on applied and internal constraints,” Proc. R. Soc. A Math. Phys. Eng. Sci., 464(2093), pp. 1229–1255. [4] Jerabek, M., Major, Z., and Lang, R. W., 2010,"Uniaxial compression testing of polymeric materials." Polymer testing, 29(3), pp. 302-309. [5] Cady, C. M., Gray, G. T., Blumenthal, W. R., Peterson, P. D., and Idar, D. J.,1998, “High and Low strain rate compression properties of several energetic material composites as a function of Strain Rate and temperature,” C. Of, S. Energetic, M. Composites, vol. 836. 3

P.141

SETS2016-39 The Southwest Emerging Technology Symposium 2016

HIGH TEMPERATURE MEASUREMENT USING LITHIUM NIOBATE CERAMIC MATERIAL Md Rashedul H Sarker1, Hasanul Karim1, Ricardo Martinez1, Jorge Silva1, Norman Love1*, Yirong Lin1* 1 Department, Affiliation, El Paso, TX 79968, USA; 1 Department of Mechanical Engineering, UTEP, El Paso, TX-79968, USA; * Norman Love ([email protected]); Yirong Lin ([email protected])

Keywords: Pyroelectricity, Sensor, Lithium Niobate, Temperature Measurement ABSTRACT Temperature monitoring for energy generation systems play an important role for control of overall safety and efficiency. Continuous monitoring of real time temperature can lead to enhanced efficiency in these systems. A 1 cm x 1 cm sample of LiNbO3 ceramic with 0.2 cm thickness was prepared as a sensor. LiNbO3 has high Curie temperature (1210 °C) and is applicable for high temperature application as a sensor material. The LiNbO3 sensor and a Ktype thermocouple were placed inside a tube furnace to sense the temperature. Different temperature setting conditions were applied to the sensor including slow heating rate, high heating rate, and steady state conditions for prolonged time period to validate readability and repeatability of the sensor. Temperatures were calculated using current generated from the sensor upon heating or cooling. The calculated temperature from the sensor was compared with the temperature measured by the K-type thermocouple. A range of deviation from 2 % to 11 % was found between the temperature measured by LiNbO3 and thermocouple. 1. Introduction This paper shows the measurement of temperature from room temperature to 450 °C using the pyroelectric properties of a lithium niobate (LiNbO3) ceramic. Lithium niobate is a pyroelectric material which has strong temperature variation dependent spontaneous polarization and loses its pyroelectric property when it is heated above its Curie temperature [1-2]. Pyroelectric materials have been used many applications including thermometer for low temperature applications, monitoring indoor objects, tracking motion direction and distance based on a pyroelectric infrared sensor thermometer [3-5]. Among different types of pyroelectric material, lithium niobate has higher Curie temperature (1210 °C) than PZT and PVDF, which makes this material promising for high temperature measurement applications. Pyroelectricity is a phenomenon of certain materials showing temperature dependent spontaneous polarization. Both increasing and decreasing of temperature cause the changes in energy of atoms, crystal structure, and material polarity [6-7]. Generated current (I) through a homogenous pyroelectric material with temperature T at any time (t) is presented in Eq.1: I

dQ dt

pA

dT dt

(1)

Equation 1 can be rewritten into Eq.2 to calculate the temperature of pyroelectric material at any specific time. tf

Tf

1 Idt Ti pA ti

P.142

(2)

HIGH TEMPERATURE MEASUREMENT USING LITHIUM NIOBATE CERAMIC MATERIAL

2. Experimental Setup For the high temperature tests from room temperature to 450 °C, the LiNbO3 sensor was placed inside of a tube furnace (MTI Corporation, OTF-1200 X-S-UL), Fig.1 (a). The temperature of the tube (5 cm diameter) furnace was set with the help of PID automatic control. The temperature can be set for 30 different programmable steps and reaches a maximum 0.33 ˚C/s heating rate. Experimental tests were done by setting the tube furnace temperature from room temperature to 500 ˚C with two different heating rates of 0.25 ˚C/s and 0.33 ˚C/s. A K-type thermocouple rod was also placed near the LiNbO3 sensor inside the tube furnace to compare with temperature sensing reading from LiNbO3. The two electrical leads were connected to a picoammeter (KEITHLEY 6485), Fig.1 (b), to measure the current output from the LiNbO3.

Fig.1. (a) Tube furnace (b) Picoammeter.

3. Results and Discussion For temperature measurements done at the high temperature range (up to 450 °C), the LiNbO3 sensor was prepared using a high temperature resistive coating, adhesive and ceramic sleeve. The LiNbO3 sensor was then placed inside of a tube furnace, Fig.1 (a), along with a thermocouple rod at room temperature condition with both ends of the tube open to the atmosphere. Later, the tube furnace was set at different operating conditions from room temperature to 300 °C, 400 °C, and 500 °C respectively. At every operating condition, the thermocouple rod temperature readings were less than the set temperature of the furnace because the thermocouple rod was not at the same place as the furnace mounted thermocouple. As the thermocouple rod was placed close to the LiNbO3 sensor, the temperature measured by the thermocouple rod was taken into consideration to compare with the temperature sensing reading from LiNbO3 sensor. The current generated from LiNbO3 sensor at different operating conditions was used to calculate the temperature of the sensor using Eq.2. Figures 2 (a & b) and 3 (a) show the comparison between temperature measured by the LiNbO3 sensor (dotted line) and the thermocouple rod (solid line) at different set conditions of the tube furnace.

Fig.2. Temperature measured by the LiNbO3 sensor and the thermocouple with a furnace set temperature of (a) 300 °C and (b) 400 °C

P.143

HIGH TEMPERATURE MEASUREMENT USING LITHIUM NIOBATE CERAMIC MATERIAL

Fig.3. (a) Temperature measured by the LiNbO3 sensor and the thermocouple with a furnace set temperature of (a) 300 °C and (b) Thermocouple and LiNbO3 sensor temperature comparison at steady state conditions for using the tube furnace.

4. Conclusions A LiNbO3 material was used as a sensor to measure temperature up to 450 °C. The sensor was placed inside of a tube furnace with different heating rate from The sensor can measure temperature at both low and high temperature ranges from 22.5 °C to 450 °C. The sensor can measure the temperatures up to 450 °C with a maximum 11 % deviation from the temperatures measured by the thermocouple. Overall, the study in this paper shows a methodology for measuring temperature using the pyroelectric properties of a LiNbO3 ceramic material. Hence, it is possible to measure high temperatures using LiNbO3, providing the pyroelectric coefficient profile with respect to temperature is known. This study is thought to be of interest for developing high temperature sensors for various industrial and commercial applications.

References [1] Whatmore, R. W. "Pyroelectric devices and materials." Reports on progress in physics 49, no. 12 (1986): 1335. [2] Srinivasan, M. R. "Pyroelectric materials." Bulletin of Materials Science 6, no. 2 (1984): 317-325.A. Green, B. Red and C. Blue “The title of the journal paper”. Journal Name, Vol. 1, No. 1, pp 1-11, 2006. [3] Lang, Sidney B., Steven A. Shaw, Lynn H. Rice, and K. D. Timmerhaus. "Pyroelectric thermometer for use at low temperatures." Review of Scientific Instruments 40, no. 2 (1969): 274-284. [4] Tsai, C. F., and Ming-Shing Young. "Pyroelectric infrared sensor-based thermometer for monitoring indoor objects." Review of Scientific Instruments74, no. 12 (2003): 5267-5273. [5] Zappi, Piero, Elisabetta Farella, and Luca Benini. "Tracking motion direction and distance with pyroelectric IR sensors." Sensors Journal, IEEE 10, no. 9 (2010): 1486-1494. [6] Smith, Brian, and Cristina Amon. "Simultaneous electrothermal test method for pyroelectric microsensors." Journal of Electronic Packaging 129, no. 4 (2007): 504-511. [7] Batra, A.k., Aggarwal M.D. Pyroelectric Materials: Infrared Detectors, Particle Accelerators and Energy Harvesters, SPIE Press, Bellingham, Washington, USA, 2013

3 P.144

SETS2016-40 The Southwest Emerging Technology Symposium 2016 Analysis of Aerodynamics for Shell Eco Marathon vehicle sing Computational Fluid Dynamics C. Mata, J Chessa Department of Mechanical Engineering University of Texas at El Paso El Paso, Texas, 79968-0521 Keywords: Aerodynamics, Road Vehicle, Shell Eco Marathon, Computational Fluid Dynamics, CFD Abstract

A road vehicles aerodynamics can be one of the most influential aspects of its performance. With the increased importance on fuel efficiency in recent years, new road vehicles are being developed smaller in size, with smaller displacement engines, as well as with improved aerodynamics. The aerodynamics of a vehicle can have a significant effect on its fuel efficiency, as well as other important aspects of the vehicles performance such as the top speed, acceleration, and handling. A study focusing on analyzing aerodynamic effects due to vehicle geometries such as wheels covered by the vehicles body in comparison to open wheels outside the body. Where the effects of the wheels show considerable difference in drag force. The effects of vehicle distance from the road surface is also studied, where the distance from the surface shows minimal differences in drag and considerable differences in negative lift force. 1 Introduction 1.1 Wheels in vs Wheels out When considering a design with wheels in vs wheels out, many advantages and disadvantages need to be considered. While the design with wheels out may have a lower frontal surface area, there is an increase of drag due to the geometry of the wheels. The total drag of a moving body is increased drastically when wheels are added. According to W. Hucho, the wheels on a vehicle can amount to half of the vehicles total drag [1]. This can be a substantial source of drag in a vehicle designed for maximum fuel efficiency, which might have a streamlined body with a relatively low drag coefficient. 1.2 Wheels Out A wheel’s rotation will cause drops in the drag coefficient. This is due to a difference in the pressure within the wake region, between a stationary wheel and a rotating one. While a wheel sits on a surface, the rotation causes air to be pushed out from the stagnation point between the ground and the bottom front of the wheel. This fluid being forced out allows for improved flow towards the wake region [2]. The effects of drag caused by wheels can be very complex, and requires a large amount of study. Even though the behavior of flow over a wheel is beyond the

P.145

focus of this study, this can be a good topic for another study. However, due to these effects it is important to compare the resulting drag forces of both wheels in and wheels out. 1.3 Wheels In While the effects of fluid flow over an open wheel seems to make a more favorable argument towards the use of wheels covered by the vehicle’s fairing, there are other negative effects involving wheels in, which may be more significant than those caused by open wheels. The most evident difference is the large increase of frontal surface area. While it is possible to maintain a streamline design with a low drag coefficient, the technicalities of allowing enough room for the wheels within the body cause a large increase in frontal projected area leading in an increase in the drag force acting on the vehicle. 1.4 Vehicle Height While the behavior of air flow over a vehicles body has been well studied and understood, the flow of air underneath has more room for further study. When looking at past studies of flow below the vehicles body, most are simplified. For example, assuming the lower section of a car to be completely smooth [1]. When looking at flow below a vehicles body, there is a larger focus on the effects of lift force as opposed to drag force. [4] While the effects of drag force might be minimal relative to the increases or decreases of lift force, the study will focus on the effects of drag relative to the vehicle height. 2 Methodology and Approach 2.1 Simulation The simulation process is begun by using the defined parameters of the calculation for the vehicle model. The vehicle model used a basic design with only the necessary detail in order to avoid any unnecessary complications further in the simulation process. The wheels used in the model are simplified as well but are designed very closely to the general dimensions of the physical wheels utilized in the vehicle. 2.2 Validation In order to validate results from the simulations, there first must be simulations done for a geometry with a known value for drag, using the same parameters. The geometry used in this validation process is a sphere, which has known values for drag and lift. The simulation was first ran using the 2.286m sphere, and the initial boundaries of the wind tunnel where then adjusted to fit the sphere comfortably. This is done in order to minimize the effects due to the distance of the boundaries. 2.3 Pre-Processing The problem set up is done with the Virtual Wind Tunnel Software. The Simulation is prepared in two different ways, one involving stationary wheels and another involving rotating wheels. In both cases the ground is simulated moving at the same velocity as the fluid inlet, where the velocity of the incoming air is set to 15mph or 6.706 m/s. The boundaries of the wind tunnel are

P.146

set in compliance to the validation, where the vehicle is set 10m away from the inlet, the width is 8m, height is 5m, and the length is set to 40m. 2.4 Solving With the use of Acusolve the simulation is solved, using a high power computing (HPC) cluster. Using the linux command: acuRun -pb vwtAnalysis -dir ACUSIM.DIR -inp vwtAnalysis.inp -np 48 -nt 12 -do all -lsf Which will run the solver using 48 processing cores, at 12 threads per computing node. Acusolve is set up to solve the Navier-Stokes equation, while using the Spalart-Allmaras turbulence model. The convergence tolerance normally used by the solver is set to 0.001. 3 Results 3.1 Wheels In vs Wheels Out As expected the values for a vehicle with rotating wheels results in a lower drag coefficient than that of one with stationary wheels. With the resulting values of each simulation plotted against the element size it is possible to see the convergence of the simulation with the increased refinement.

Figure 1 resulting Drag Coefficient vs the element size The Resulting Drag coefficient for stationary wheels is Cd= 0.396, with an element size of 0.4 m and an element count of 1.9 million elements. This value converges as the element size is further decreased. The value for Cd at an element size of 0.2 m and an element count of 3.7 million elements, results in Cd= 0.334. This value begins to converge as the mesh size is further decreased by half. With an element size of 0.1m the resulting drag coefficient is Cd= 0.318.

P.147

3.2 Vehicle Height When looking at the effects of ground clearance, it is clear how there is a considerable drop in the Lift coefficient. This is due to the drops in pressure under the body caused by the Venturi nozzle effects underneath the vehicle.

Ground Clearance vs Lift Coefficient 0.18 0.16

Lift Coefficeint

0.14 0.12 0.1 0.08 0.06

0.04 0.02 0 0.00

1.00

2.00

3.00

4.00

5.00

Ground Clearance (in)

Figure 2 changes in lift coefficient against ground clearance However the effects of drag increase as the ground clearance drops. Although the changes in drag coefficient are less significant in relation the differences in lift coefficient for the same decrease in ground clearance.

Figure 3 Change of Drag coefficient vs ground clearance

P.148

4

Conclusion

The results show how open wheels can highly affect the drag coefficient of a vehicle in a negative manner. This is mostly due to the geometry of the wheels which create a large wake behind the wheel causing substantial increases in drag. It is also noted that the decrease in ground clearance will increase the drag coefficient while reducing lift due to the increase of flow under the vehicle which in turn will cause a decrease in pressure.

References [1] Hucho, Wolf-Heinrich, 4th ed. Aerodynamics of road vehicles: from fluid mechanics to vehicle engineering. Elsevier, 1998. [2] Cogotti, A., “Aerodynamic Characteristics for Car Wheels,” int. Journal of vehicle Design, SP 3, 1983, pp 173-196 [3] Mercker, E., Bernebrg, H., “On the Simulation of Road Driving of a Passenger Car in a Wind Tunnel Using a Moving Belt and Rotating Wheels.,” 3rd Int. Conf. Innovation and Reliability, Florence, April 8-10, 1992 [4] Wright, P.G., “The Influence of Aerodynamics on the Design of Formula One Racing Cars,” Impact of Aerodynamics on Vehicle Design, Int. J. of Vehicle Design, SP3, 1983, pp 158-172

P.149

SETS2016-41 The Southwest Emerging Technology Symposium 2016

STUDY OF LEARNING EFFECTIVENESS OF PROJECTBASED LEARNING METHOD ON TEAM-BASED AND INDIVIDUAL-BASED PROJECTS H. Kim1, A. Akundi2, Y. Lin1, T. Tseng2* 1 Department of Mechanical Engineering, El Paso, TX 79968, USA; 2 Department of Industrial and Systems Engineering, El Paso, TX 79968, USA; * Corresponding author ([email protected])

Keywords: engineering education, learning effectiveness, project-based learning, team-based project, individual-based project ABSTRACT Education of engineering has been increasingly studied in recent decade as the dramatic changes and challenges of modern industry have been experienced. One of the critical challenges considered in engineering education is to find solutions on complicated analytical issues by using methods such as self-learning. Projected-based learning has become emerged solution recently in academic areas to meet high demands on rapid shifting engineering industry by promoting students in both analytical and inter-personal abilities. Although team projects in the Project-based learning method have advantages such as promoting intercommunication, inter-personal skillset, and critical thinking, drawbacks can be observed when it comes to the equal learning and workload on each individual. This project evaluates the effectiveness of project-based learning, using component design problem based on constraints and expectations and by comparing the final component design, the project report, and evaluation consequences from group and individual projects in two different classes. In addition, the potential influence and performance when working as a team or individually will be evaluated based on written report and quiz from each individual. 1

Introduction

A negative impact of traditional lecture-based approach is the fact that information is carried out by the instructor to student and results in selective learning part of the material so that eventually the overall learning outcome is not uniformly distributed due to the passive role of student [1]. In order to improve this issue, such a flipped classroom and problem-oriented project-based learning were proposed to increase student’s active dedication and information acquisition in engineering education institutions. Flipped classroom methods have overall positive results on participating students while it is not widely implemented in college level. As opposed, project-based learning has been widely implemented and proved by the development and assessment regarding the learning outcome. In the project based-learning method, solving problems asking the fundamental principles and related information allows students attaining the enhanced knowledge acquisition and practice critical thinking [2]. Adderley et al. defined the project based-learning method by using the following concepts [3]: 1) A solution to a problem much be involved in the project; 2) Initiative is needed by the student/group of students, as well as a variety of educational activities; 3) An end product such as a thesis, report, or model is common; 4) Projects are performed for a considerable

P.150

length of time; 5) Professors and teaching assistants perform advisory roles. These concepts plainly summarize core of project-based learning and clear roles of both instructor and students in this method. In recent, engineering education has shifted expectation to practice future engineering professionals in near future as the complexity and interdisciplinary approach for the solutions are growing in industries. Moreover, this expectation is expanded to engineering graduates on not only solving the technical problems based on their major knowledges but also requiring the interdisciplinary skills such as management, communication and collaboration [4, 5]. Based on this trend, even many undergraduate engineering institutions have started to adopt this project-based learning method which can cultivate students with capabilities such as teamwork, communication, problem solving, and life-long learning skills, and have flexibility that can be easily embraced in traditional lecturebased learning. However, there are still disadvantages and it could bring: 1) unequal distribution of work when working as a team, 2) insufficient participation of individual students in team member or project, and 3) project development via dishonest means. These drawbacks result in negative effects on the student learning effectiveness. This research paper compares student performance in a PBL for a junior level in mechanical engineering course, Spring 2015 and Fall classes of Mechanical Design, where shaft design and optimization was combined to curriculum and executed as a group and individually, respectively. 2

Implementation of Project-based learning

The course implemented the project-based learning and assessed in team and individually is a Mechanical Design, 3 hours junior level course in mechanical engineering department at The University of Texas at El Paso. The goal of this course is to design and analyze mechanical components such as gears and shafts by implementing the theoretical concepts learned from prerequisites courses such as Mechanics of Materials and Statics. This class is composed of traditional lecture-based learning (lecture, quizzes, and exams) with the project-based learning method. Each different two projects were implemented in each class in Spring and Fall 2015. The first class is consisted of designing a seatbelt buckle using Computer Aided Design (CAD) and Finite Element Analysis (FEA) to analyze stresses distribution of the component and locate potential points of failure. Second project is carried out by student groups where objective was to design mechanical shaft with a design factor ≥ 3, at the same time maintaining the greatest ratio of Max and Min Von Mises stresses possible. The only constraint on this second project was 1 meter length of shaft but left to student on selection of parameters such as materials, radial dimensions, and overall design. The example of this project’s is shown in Fig. 1 with four different stages. For the assessment of project effectiveness, a class of Sprint 2015 were designed that students work on project as team (5 students) and individually in a class of Fall 2015. In terms of evaluation on these projects, report students turned in and contribution of an oral presentation at the end were evaluated for student performance for the project and development of their soft skills. In addition, a quiz covered with project contents were employed to evaluate the understanding of the materials. Eventually, student opinion of PBL method was recorded based on the anonymous survey and analytical results on student’s performance is discussed in the following section.

P.151

STUDY OF LEARNING EFFECTIVENSS OF PROJECT-BASED LEARNING METHOD ON TEAM-BASED AND INDIVIDUAL-BASED PROJECTS

a)

b)

c)

d)

Fig.1. Example of four different design stages in second project regarding shaft design: a) initial design stage, b) addition of forces and constraints for FEA, c) stress distribution in component, d) shaft deformation under applied load.

3

Evaluation of Effectiveness on Group-based and individual PBL

Performed two same projects in Spring and Fall 2015 were given two weeks to complete and review and Q&A session. Quiz was employed during the project implementation, and after two weeks, project report was submitted from each team or individually with their outcome results. Fig. 2 shows schematic diagram for project evaluation.

Individual Work - 2 weeks/ one student. Project - Shaft Design and Analysis using CAD and FEA.

Project Evaluation - Project report and quiz Team Work - 2 weeks / five student.

Fig. 2. Schematic diagram for setting group and individual projects

The report was graded based on design requirement such as total weight and Max and Min stress and quiz includes mainly related knowledge to project not only within but also outside of the project. The detail evaluation consequence is shown in Table 1. From the table, it was observed that both individual and team project have similar results (comparable results on total weight, Max/Min stress, and blue print) on report evaluation. However, on quiz evaluation, it is found that students who worked on individual project attained better score in knowledge directly related to project. This is assumed that every student is required to understand all required subtasks to complete the project so that their knowledges are comprehensive and deep to the project contents. However, it is found that students worked as a team only were assigned to one specific task each so that they are not required to understand all aspect of knowledge of the project. On the other side, students worked as a team attained higher score in knowledge indirectly related to project. It is assumed that their

3 P.152

knowledges have grown broadly with participation of group discussion thus this inspired learning from each other. Table 1: Evaluation results of Spring 2015 and Fall 2015 projects Team Project Individual Project Evaluation Tools Spring 2015 Fall 2015 Report (Group vs. individual) Total Weight Max. Stress Min. Stress Blue Print Average Quiz (both individual) Knowledge directly related to project Knowledge indirectly related to project Average

4

91 85 90 86 88

85 89 90 95 90

82 96 89

92 83 87

Conclusion

Projected-based learning has become emerged solution recently in academic areas to meet high demands on rapid shifting engineering industry by promoting students in both analytical and inter-personal abilities. However, drawbacks can be observed when it comes to the equal learning and workload on each individual. This project evaluates the effectiveness of projectbased learning, using component design problem based on constraints and expectations and by comparing the final component design, the project report, and evaluation consequences from group and individual projects in two different classes. It is found that based on the results, engineering-based skill and aspects are all similar however in engineering-based knowledge, individual project resulted in better outcome in knowledge directly related to the project because they are required to learn and understand all aspects regarding the project. Team project produced better outcomes in knowledge indirectly related to the project because the group discussion promoted learning broadly from each individual in team. References [1] Tseng, T-. L., Akundi, A., Love, N. “Instructional Setting on Student Learning Effectiveness Using Flipped Classroom in an Engineering Laboratory”. 122th ASEE Annual Conference & Exposition, 2015. [2] Savage, R. N., Chen, K. C., Vanasupa, L. “Integrating Project-based Learning Throughout the Undergraduate Engineering Curriculum”. Journal of STEM Education: Innovations and Research, Vol 8, No 3, pp15-27, 2007. [3] Adderley, K. et al. “Project Methods in Higher Education”. SHRE working party on teaching methods. Techniques group. Society for Research in Higher Education. No 24, pp. 93, 1975. [4] Lehman, M., Christensen, P., Du, X., Thrane, M. “Problem-oriented and Project-based Learning (POPBL) as an Innovative Learning Strategy for Sustainable Development in Engineering Education”. European Journal of Engineering Education. Vol 33, No 3, pp. 283 – 295. 2008. [5] Hadium H., Esche, S. K. “Enhancing the Engineering Curriculum Through Project-based Learning” ASEE/IEEE Frontiers in Education Conference, Vol 2, pp. F3F-1 – 6, 2002.

P.153

SETS2016-42 The Southwest Emerging Technology Symposium 2016

STUDY OF PHOTOLYSIS RATE COEFFICIENTS TO IMPROVE AIR QUALITY MODELS FOR THE EL PASO-JUAREZ AIRSHED Suhail Mahmud1, Pema wangchuk1, Rosa Fitzgerald2, William Stockwell 3, Duanjun Lu4 1 Computational Science Department, El Paso, TX 79968, USA; 2 Physics Department, El Paso, TX 79968, USA; 3 Chemistry Department, Howard University, Washington D.C 4 Physics and Meteorology Department, Jackson State University, MS * Suhail Mahmud ([email protected])

Keywords: Atmospheric, Photolysis, Ozone, Air quality, Simulation

Keywords: Atmosphere,ABSTRACT Photolysis, Ozone. Computation, El Paso The main objective of this work is to measure hemi-spherically integrated spectrally resolved solar photon flux between the wavelengths of 300 and 700 nm (actinic flux), and use the measured actinic flux to improve air quality simulations. Photolysis is the main driver of ozone production and this factor defines the significance of this research work. The actinic flux has been measured during the summer of 2015 in the El paso-Juarez Airshed, at the UTEP location to calculate photolysis rate coefficients for nitrogen dioxide (NO2), ozone (O3) and formaldehyde (HCHO). The improved photolysis rate coefficients have been integrated into a photochemical air quality model (CAMx), and simulations for a selected modeling summer 2015 ozone episode have been performed in an attempt to improve on air quality forecasting. In addition, photolysis rate coefficient has inter-compared with the Tropospheric Ultraviolet visible model (TUV) and the experimental value results in an effort to better understand the photolysis rate at different wavelengths present in the air shed during the chosen study episode. Scientific Background The formation of photochemical air pollution, including ozone and particulate matter (PM), depends on the photolysis of NO 2, O3, HONO, HCHO, aldehydes and ketones and other compounds. For example, the rate of ozone formation is controlled through the photolysis of NO2. A compound’s photolysis frequency, J, is determined by the product of the spherically integrated photon flux (actinic flux), I , the compound’s absorption cross sections, , and its quantum yields, , all integrated over the range of available wavelengths. [1]

P.154

J= I

d

(1)

The measured actinic flux was used to derive the photolysis rate parameters for O3, NO 2, and both photolysis reaction of HCHO. NO2 + h + O(3P) (2) 1 O3 + h O D + O2 (3) HCHO +hv H2 + CO molecular – reaction (4) HCHO +hv(+O3) 2HO2 + CO radical – reaction (5)

Experimental data For the experimental data, an Actinic Flux spectrometer has been used, which is located on the roof of the physical science building at UTEP. This is a site that is representative of the El Paso region. Because El Paso is in a desert ecosystem with no major changes in the terrain during the study period, the albedo will be constant.

Fig.1. Optical sensor installed on the roof of the Physical Sciences Building.

It consists of a hemispherical radiation collection head and a monolithic monochromator with a 512 pixel diode array detector with a spatial resolution of 2.1 nm. The spectrometers will have an extremely fast response time that is able to provide the actinic flux required to determine the photolysis rate coefficient for the photolysis of ozone to produce O1D in 200 ms (or less). Comparison between experimental value and the TUV model value Due to the lack of clouds, two days were selected for simulations. The selected days were July 1 and July 2, 2015.A radiative transfer model, based on the 2-stream Delta Eddington Method, was used to try to simulate the ozone photolysis rate values at the surface that match the experimental ozone photolysis rate values at the surface which were obtained with the

P.155

spectrometer. The ozone column, the albedo, etc were strategically chosen to ac hieve this. This guaranteed a continuous distribution of the ozone photolysis values throughout the entire vertical column in the atmosphere. The inter-comparison is observed on following figures. As observed the surface modeled j-values for ozone are in close agreement to the surface experimental j-values.

Fig 2. Comparison between experimental and model J values

Air Quality Simulations In this project, the version of CAMx V6.1 was used for the air quality simulations. The CAMx model requires a meteorological model to produce meteorological fields and an emissions processing system In this project, the emissions are processed with the Sparse Matrix Operator Kernel Emissions .The SMOKE model is used to convert the source-level emissions (county total emissions) reported on a yearly basis to model-ready emissions which are spatially resolved, hourly and aggregated into model species. [2] In this project, the map projection is the Lambert Conformal, centered at the city of El Paso, TX as observed in the domain configuration shown below.

Fig 3. CAMx Domain Configuration

3 P.156

Three cases were selected for simulation for July1 and 2: Base case (with photolysis j -values used in models), P10 (increasing the photolysis j-values by 10% for all species) and ozone column (using the photolysis j-modeled values to match the instrument’s ozone j-values at the surface, but for ozone only).

Result

Fig 3. Inter-comparison of all 3 cases against TCEQ data for July 1, 2

The above figure shows the time series inter-comparison of all 3 cases against the Texas Commission for Environmental Quality, TCEQ, data for July 1, 2 for the CAMS12 monitoring station, located in the UTEP campus. The Base case and the ozone column case are superimposed on each other and appear as a single line under the scale used. The model simulation for the Base case under-predicts ozone but all cases follow faithfully the overall trend of the experimental ozone results. The case that performed best was the P10 case, as observed in above figure. P10 case shows promising results and demonstrates that optimizing the photolysis rate coefficients will improve the accuracy of air quality simulations and forecasting capability.

References [1] Kim, D., Loughner, C. P., Wetzel, M. A., Goliff, W. S., & Stockwell, W. R. (2007). A comparison of photolysis rate parameters estimated from measured and simulated actinic flux for wintertime conditions at Storm Peak Laboratory, Colorado. Journal of atmospheric chemistry, 57(1), 59-71. [2] Loughner, C. P., Lary, D. J., Sparling, L. C., Cohen, R. C., DeCola, P., & Stockwell, W. R. (2007). A method to determine the spatial resolution required to observe air quality from space. Geoscience and Remote Sensing, IEEE Transactions on, 45(5), 1308-1314.

P.157

SETS2016-43 The Southwest Emerging Technology Symposium 2016

SCENARIO PLANNING APPLICATIONS FOR ENERGY ISSUES 1

B.A. Benedict1 Mechanical Engineering, University of Texas at El Paso, El Paso, TX 79968, USA; * Corresponding author ([email protected])

Keywords: planning, scenarios, energy, uncertainty

ABSTRACT Scenario planning was first used effectively by Royal Dutch Shell approximately 40 years ago. The company recognized that efforts to predict exactly the future are unlikely to be very successful. The premise of scenario planning is that organizations look at possible future trends and project several possible futures (or scenarios). The intent is to project enough such scenarios, even unlikely ones, that they “bracket” possible futures. This enables one to assess the ability of their policy, process, or design to perform positively within any of the scenarios and thus represent a truly robust choice. This paper briefly describes some examples of use of scenario planning within the energy sector, as well as some unusual factors that may influence the outcomes. Introduction Scenario planning arose from a recognition that we could not predict the future with any accuracy at all. Its first use was by Royal Dutch Shell [1] over forty years ago, and it has been widely and effectively employed in many arenas to look at a possible range of actions and the resultant impact on policies, designs, and the like. The intention is to anticipate future possibilities in such a way as to develop the most robust plans, designs, and policies leading to sustainable solutions. [2],[3] Rationale for use of scenario planning There are many historical examples of the failure of traditional predictions. Of more current interest is the severe drop of oil prices. Venezuela, which holds the world's largest petroleum reserves is noteworthy because of their excessive reliance on income from oil exports. Numerous predictions of oil prices to 2050 ranged from $95 to almost $150. If Venezuela had engaged in truly introspective scenario planning, perhaps they would have taken some steps in other strategic directions. For example, low oil prices could encourage development of a broader industrial base. In the field of sustainable development, including energy, there are numerous areas of uncertainty to be considered. These can include, as a minimum, the following: changing environmental regulation; technology improvements; variation in energy prices; impacts of climate changes; and political changes.

P.158

Scenario planning process There are several ways to craft scenarios. One methodology [4] follows these steps: Identifying the right issue(s) Brainstorming critical events Describing the trajectories of critical events Determining combinations of critical events Developing scenarios Comparing strategies and recommendations Reviewing examples The process is most useful when everything can be included. However, often organizations either feel so comfortable or confident in existing procedures that the possibility of change is really not considered. The failure of the Monitor Group, a prominent firm providing scenario planning services, brings this home. The firm did not apply these principles to their own organization. This failure has led to suggestions that it may be best to include the current situation (default belief system) [5] as a scenario. If people are not prepared (or allowed) to be truly introspective, scenario planning use will not truly inform the future directions of that organization. The number of constraints that are possible in any given situation may be excessive and the complexity of evaluation of any policies escalates quickly as constraints are added. Therefore, most studies to which they are exposed have chosen a limited number of constraints for the study. For example, a recent excellent report [6] studied thirty-five proposed energy policies for the US. The measures of success were only two: changes in gasoline prices and reductions in greenhouse gas emissions. Even with only these constraints, a great deal can be learned about the possible effects of a particular policy. While this was not in itself a scenario study, it could easily have been part of one. Examples of use within energy sector Examples to be discussed included two organizations who have projected scenarios to 2050, specifically Royal Dutch Shell [7] and the World Energy Council [8]. Further, reference will be made to other countries or organizations employing this powerful tool. Related items such as disruptions (including disruptive technologies), black swans, wild cards or surprises like the Fukushima incident. Some specific topics changing rapidly may be factors in scenario planning. Examples include the current debates between home solar and the electric power industry and energy storage uncertainty in a rapidly evolving field.

P.159

Thoughts for future use The paper is intended to encourage practitioners to employ scenario planning to develop more robust policies, designs, and operating policies in the energy arena. Use of this powerful tool can lead to more robust and resilient sustainable development and energy development and use.

References [1] M. Lindgren and H. Bandhold, Scenario Planning: The Link between Future and Strategy. New York: Palgrave McMillan, 2003. [2] A. Wilkinson and R. Kupers, “Living in the futures,” Harvard Business Review. Accessed through https://hbr.org/2013/05/living-in-the-futures/ar/1 [3] D. Niles, Forbes, The secret of successful scenario planning, August 3, 2009, Accessed through http://www.forbes.com/2009/08/03/scenario-planning-advice-leadership-managing-planning.html [4] K. Miesing and R. van Ness, Scenario planning – exercise, Organization Management Journal, Teaching and Learning, Vol. 4, No. 2, 2007, pp. 148-167. [5] A. Hutchinson, The Monitor Group: A failure of scenario planning, Spend Matters, November 13, 2012 Accessed through http://spendmatters.com/2012/11/13/monitor-group-a-failure-of-scenario-planning/ [6] A.J. Krupnick, I.W.H. Perry, M. Walls, T. Knowles, and K. Hayes, “Toward a new national energy policy: assessing the options,” National Energy Policy Institute and Resources for the Future, 186 pp., November, 2010, [7] Royal Dutch Shell, Shell energy strategies to 2050. Accessed through http://www.shell.com/content/dam/shell/static/public/downloads/brochures/corporate-pkg/scenarios/shellenergy-scenarios2050.pdf [8] World Energy Council, “World energy scenarios: composing energy futures to 2050”. Accessed through http://www.worldenergy.org/publications/2013/world-energy-scenarios-composing-energy-futures-to-2050/

3 P.160

SETS2016-44 The Southwest Emerging Technology Symposium 2016

NON-SIMULTANOUS DG AND CAPACITOR BANKS ALLOCATION IN DISTRIBUTION NETWORKS BASED ON ECONOMIC EVALUATION 1

Sayyid Mohssen Sajjadi1, Paras Mandal1*, and Tzu-Liang (Bill) Tseng2 Power & Renewable Energy Systems (PRES) Lab, Department of Electrical and Computer Engineering 2 Department of Industrial, Manufacturing and Systems Engineering University of Texas at El Paso, El Paso, TX 79968, USA *Corresponding author ([email protected])

Keywords: Active and reactive power losses, capacitor placement, DG allocation. ABSTRACT This paper presents a multi-objective optimization approach using Genetic Algorithm (GA) to determine the optimal size and location of distributed generation (DG) units, such as wind turbine and PV panels, and parallel capacitor banks non-simultaneously considering multistage load models in distribution systems. The proposed multi-objective function considers technical issues such as voltage profile, line loading, and active and reactive power losses. In non-simultaneous allocation, it is assumed that first DGs are optimally allocated followed by capacitor banks. The proposed algorithm is tested on 9-bus systems. 1

Introduction

With the recent advancement in technology in the context of electric power distribution systems, appropriate allocation of distributed generations (DGs) along with parallel capacitor banks can provide the most economical and reliable approach to supply electricity to the customers. An efficient penetration of DGs and capacitor banks is an effective mechanism for power loss reduction and voltage support [1]. Deficiency of reactive power in distribution systems affects the voltage profile and system losses. Capacitor banks in the distribution systems provide reactive power and can significantly improve the voltage profile as well as reduce peak load losses if they are installed in the optimal locations with suitable sizes [2]. In this paper, a multi-objective optimization approach is proposed to determine nonsimultaneous placement and sizing of DGs and capacitor banks in radial distribution networks with multi-stage load models. The multi-objective function includes active/reactive power losses and voltage profile. Genetic algorithm (GA) is used to determine the optimal solution for a test case of 9-bus systems. 2 Mathematical Model and Objective Function The main objective of this paper is to determine the optimal location and size of DGs and capacitor banks to be installed in the distribution systems in order to reduce the active and reactive power losses and minimize the intake power from the grid. Thus, the proposed multiobjective function can be expressed as follows. • Active power demand reduction from transmission line and active power loss reduction

P.161

∆!!"# = !!"#!"#,! − !!"#!",! = !

!!"#,! + !!"##!"#,! − (!!"#,! + !!"##!",! − !!" !!! !"!,!"

!!" !!!

!!" !!!

!!" !!! !"!,!" )

= !!"##!"#,! − !!"##!",! +

(1)! where !!"#!"#,! !(!!"#!",! )!is the purchased active power from transmission line in j load level without (with) DG sources, !!"#,!!!! is customer demand, !!"##!"#,! !(!!"##!",! ) is distribution network losses in jth load level without (with) DG sources installed in the distribution network. !!" !!is the number of DG units, !!" is the capacity of DG units, and !"!,!" is the generated power by ith DG source in jth load level (MW). Equation (1) can be formulated as !!" !!" ∆!!"# = ∆!!"##,! + !!! (2) !!! !"!,!" th

!

where!!∆!!"##,! is the amount of loss reduction based on the presence of DGs. Therefore, the annual benefits of active power reduction would be !"#!""#!$ = !!!! !!"!,! ×∆!!"# ×!! (3) !

where!!!"!,! is the energy market price in jth load level ($/MWh) and n is the number of load level. InfR and intR are inflation rate and interest rate, respectively. • Reactive power loss reduction The annual benefit of the reactive power loss reduction based on capacitor banks and DGs can be calculated as !"#!""#!$ = !! (!! − !! ) (4) where !! , !! , and !! are the reactive power losses before and after installation of equipment (kVar), and the worth of reactive power ($/kVar), respectively. 3 Optimization Technique for Optimal Allocation of Components GA has been widely applied in most optimization problems as a prominent meta-heuristic algorithm. It is a search algorithm based on the mechanism of natural genetics to obtain the best optimal solution. In GA, a candidate solution for a specific problem is called a chromosome, and consists of a linear list of genes. Each individual represents a possible solution to the problem. A population consists of a finite number of individuals. Each individual is decided by an evaluation mechanism to obtain its fitness value. Based on this fitness value and undergoing genetic operators, a new population is generated iteratively with each successive population referred to as a generation. GA uses three basic stages (initial population, crossover operator, and mutation operator) to manipulate the genetic composition of a population. Details of GA and its associated steps are available in [3]. 4 Numerical Studies and Results Analysis For optimal placement and sizing of DGs and capacitor banks, a software module using MATLAB has been developed. A number of tests are performed on a 9-bus test system, and data are collected from [4] and [5]. Table 1 presents the benefits of different objectives for nonsimultaneous placement of DGs and capacitor banks using GA. Optimal location and size of DGs and capacitor banks are shown in Table 2. Voltage profile before and after placement of

P.162

components are shown in Fig.1. Comparison of the results shows that voltage profile has been improved after inclusion of components. Table 1. Simulation results for non-simultaneous placement of DGs and capacitors. Network condition Benefit of reactive power loss reduction ($) Benefit of active power demand reduction from transmission line and reduction losses ($)

Non-simultaneous placement 840681 8643146

Table 2. Optimal location and size of DGs and capacitors. Network condition Capacity of DGs

Capacity of Capacitors

Non-simultaneous placement 2 MW bus 5 3 MW bus 8 2 MVar bus 9 2.5 MVar bus 5 3 MVar bus 4 3.5 MVar bus 8

Fig. 1. Illustration of improvement in voltage profile at different buses.

5

Conclusion

This paper proposed a multi-objective optimization approach using GA to determine the optimal size and location of DGs and capacitor banks. Test results demonstrated that voltage profile can be significantly improved after non-simultaneous and optimal placement of the components (DGs and capacitor banks). In addition, significant reduction of active and reactive power losses, in terms of benefits ($), were observed. References [1] T. Gozel and M.H. Hocaoglu, “An analytical method for the sizing and sitting of distributed generators in radial systems,” Int. J Electric Power Syst. Res, Vol. 79, pp. 912–918, 2009. [2] T.L. Huang, Y.T. Hsiao, C.H. Chang, J.A. Jiang, “Optimal placement of capacitors in distribution systems using an immune multi-objective algorithm,” Int. J Electric Power Energy Syst. Vol. 30, pp. 184–92, 2008. [3] W. S. Tan, M. Y. Hassan, and M. S. Majid, “Multi population genetic algorithm for allocation and sizing of distributed generation,” In Proc. 2013 IEEE Power Engineering and Optimization Conference (PEDCO) Melaka, Malaysia. [4] S.M. Sajjadi, M.-R. Haghifam, and J. Salehi, “Simultaneous placement of distributed generation and capacitors in distribution networks considering voltage stability index,” Int. J Electric Power Energy Syst. Vol. 46, pp. 366-375, 2013. [5] N. Khalesi, N. Rezaei, and M.-R. Haghifam “DG allocation with application of dynamic programming for loss reduction and reliability improvement” Int. J Electric Power Energy Syst. Vol. 33, pp. 288-295, 2011.

P.163

SETS2016-45 The Southwest Emerging Technology Symposium 2016

TEMPERATURE SENSING ON WOVEN CFRP BY PIEZOELECTRIC PARTICLES 1

Ricardo Martinez, Emilio Tarango*, Yirong Lin Mechanical Engineering, El Paso, TX 79968, USA; * Emilio Tarango [email protected]

Keywords: Composites, Pyroelectric, Carbon Fiber, Multi-Functional Materials

ABSTRACT The temperature sensing capabilities and structural properties of fiber carbon reinforced polymers with embedded piezoelectric particles of lead zirconate titanate (PZT-5A) are studied using three point bending testing and temperature controlled furnace. The composites are fabricated using autoclave molding forming a six-layer composite using carbon and glass woven fabric. The pyroelectric response is estimated and analyzed under different heat loadings. The study demonstrate that the piezoelectric inclusions allow sensing applications as well as power harvesting capabilities by compromising by 11% the maximum strength of pure composites.

1

Introduction

In the last two decades, the use of composite materials in structural applications ranging from aircraft and space structures to automotive and biomedical as well as ballistic armor applications has been growing interest [1]. This polymer based composites are being used due to their higher stiffness and strength per unit weight in comparison with aluminum and titanium alloys. Additionally, the environmental energy harvesting to power low-energy consumption systems is becoming a growing topic for the large deployment of wireless sensor networks and increase of integration and functional density of electronics. Among the existing mechanisms for energy harvesting methods include vibrations (piezoelectric, electromagnetic induction, electrostatic method), thermoelectric materials (Seebecks effect), pyroelectrics (thermal energy variation), and photovoltaic effect (solar cells) [3]. Among these mechanisms, pyroelectricity has revealed useful in sensing applications for a long and useful source of energy. Although pyroelectrics have been displayed low efficiencies, studies have shown that pyroelectric cells can produce currents in the order of 10-7 A, and charges in the order of 10-5 C for temperature fluctuations from 300 K to 360K [2]. The inclusion of pyroelectrics in woven fabric composites would allow low-energy harvesting in different applications such as the construction, aircraft, and automobile industry. This energy would allow the development of more intelligent systems, since lowpower sensors would be implemented in areas were maintenance and location are not costeffective for long periods of time.

P.164

Temperature Sensing on woven CFRP by piezoelectric particle

2

Fabrication Process

The fabrication process for the autoclave molding commences with the preparation of 10” x 10” sample squares sizes of the woven glass/carbon fibers. The fibers must be visually inspected to verify that they are clear of any other material or objects to ensure the most accurate results. Each composite is created with nine (9) layers with six (6) being carbon fiber and three (3) glass fiber. The total weight of these nine layers is measured on the scale shown on the left image inside figure (1). The measured weight is then utilized to find the quantity of lead zirconate titanate (PZT-5A). The system 2000 epoxy resin is created and utilized as the main component of the composite matrix. The other component utilized in the matrix is a high performance hardener called the system 2020. The mix ratio between the epoxy and hardener is 4:1 by volume and these consolidate the matrix. The measured PZT quantity is then combined with the matrix, during this process the PZT amount varies with 0%, 5%, 10%, 15% and 20% by weight. The combined matrix epoxy is then stirred until completely mixed. The mixture is then brushed on to each layer and compacted with two aluminum plates. This is then placed in a vacuum sealed bag to degas the epoxy and compress the composite. The vacuum compressing the plates forces the excess matrix to exit by all four sides, this ensures that the matrix is uniform through the complete composite. The vacuum bag is left to cure for 24 hours and afterwards the separation of the plates is done with regular lab tools. The composites are then cut for several tests that will be discussed later on in the methodology section.

Fig.1. (Left) Scale to measure the weight of the fiber layers,(Middle) Epoxy and hardener, (Right)Vacuum sealed bad

3

Methodology

These multifunctional composites will be characterized in their mechanical and electrical properties. From the mechanical side, elastic and strength assessment will be done by three-point bending experiments (ASTM D790), tensile testing (ASTM D3039), dispersion analysis, and volume fraction characterization. Figure 2 shows the schematic of the mechanical testing. The electrical assessment will be done by subjecting the composites to different loading and thermal conditions and studying the variation of the response of the piezoelectric and pyroelectric effect.

P.165

Fig.2.Types of test the composites will be subjected to. (Top Left)Pyroelectric temperature estimation, (Top Right) Poling process, (Bottom left) 3 Point bending, (Bottom right) Tensile test

4

Preliminary Results and Discussion

Three-point bending experiments showed an overall decrease on the mechanical properties such as bending stiffness, bending strength and toughness. This behavior was expected due to the increase of crack nucleation in the interface between the epoxy/fiber caused by the ceramic inclusions. In Figure 3 the decrease of slope and maximum stress can be observed between the 0% and 10% comparison. Additionally, the pyroelectric experiment showed the expected trend of current generation. During heating the charge builds up by generating negative current, and when cooling occurs, the charge decrease by result of positive current and change of polarization. This phenomenon will be studied by testing three different levels of current, as well as the different compositions of PZT wt% in the composites.

Fig.3.Pyroelectric experiment on 10% composite (left). Three point bending test comparison between 0% and 10% PZT on carbon/glass fiber reinforced composite.

6

References

[1] Cesim Atas and Onur Sayman. An overall view on impact response of woven fabric composite plates. Composite Structures, 82(3):336 345, 2008. [2] A Cuadras, M Gasulla, and Vittorio Ferrari. Thermal energy harvesting through pyroelectricity. Sensors and Actuators A: Physical, 158(1):132 139, 2010. [3] Gael Sebald, Daniel Guyomar, and Amen Agbossou. On thermoelectric and pyroelectric energy harvesting. Smart Materials and Structures, 18(12):125006, 2009.

3 P.166

SETS2016-46 The Southwest Emerging Technology Symposium 2016

IMPACT RESPONSE OF WOVEN COMPOSITES 1

A. Castellanos1, S. Md Shariful1, P. Prabhakar* Mechanical Engineering, The University of Texas at El Paso, El Paso, TX 79968, USA; * Corresponding author ([email protected])

Keywords: Textile composites, Delamination, Impact, Arctic temperature

ABSTRACT Impact tests is investigated for woven composite laminates under arctic conditions. A total of 36 specimens were tested: 18 for room temperature and 18 for arctic conditions. Impact tests were conducted using a guided drop-weight test for energies ranging from 2-35 Joules. The force-time, energy-time, and displacement-time plots were compared for room and arctic temperatures. The damage parameter was calculated for all the specimens. During impact the major damages were fiber breakage, delamination and matrix cracking at the back surface, below the nearly undamaged zone, which were attributed to the bending stresses. 1

Introduction

Sandwich structures have gained popularity for the use in Navy and Marine Corps. Sandwich structures consist of a pair of face sheets and a lightweight core which are joint by an adhesive as it can be seen in Fig. 1. The face sheets are usually made of composite materials and the core by polymer foams. The advantages of these structures are their low weight, high stiffness, high strength, fatigue resistance and absorption capabilities [1]. Despite its several advantages, a major disadvantage of sandwich structures are their low resistance to impact damage due to the layered nature of the face sheets (composite materials) as well as weak adhesion between the face sheets and core. Composite materials consist of layers of woven carbon fibers that are reinforced in polymer matrix. Although the mechanical properties along the fibers are strong and stiff, the region between the woven carbon fiber layers of the composite is a resin-rich region, which is known as interlaminar region as it can be seen in Fig. 2. This region is very susceptible to damage and can result in a premature failure of the composite.

Fig. 1. Sandwich structure composition.

Fig. 2. Composite materials.

P.167

The life of a composite depends on its response to different failure mechanisms, such as delamination or interlaminar fracture, matrix cracking, matrix-fiber debonding, fiber breaking or fiber pullout [2,3]. During manufacturing and service, composite structures are subjected to unplanned impact loadings. Most of the impacts on composites are in the transverse direction, that is, in the through thickness direction of a composite. Low velocity impacts can cause great damage to the composite with typical failure modes like, matrix cracking, delamination and fiber failure [4] . The most common failure mode is delamination. Although impact behavior of composites have been widely studied, seldom work has been reported on impact behavior in arctic conditions. The dynamic response of composite depends on factors such as temperature, moisture and ultraviolet radiation. Composites can become more ductile or brittle depending on the temperature that they are subjected [5]. Researchers have reported that during impact at low temperature more damage is presented in the specimens and the damage area increases as the impact energy increases and the temperature decreases. They have also found that the absorbed energy is dependent on the temperature. In this study, the low velocity impact behavior of carbon fiber-vinylester resin composites at arctic temperatures (-60 oC) will be performed and they will be compared with the results at room temperature. Further, repeated impact tests are being conducted to establish durability and tolerance of these woven composites at arctic Temperatures. The damage due to impact will be evaluated in terms of visual damage of the impacted specimens and by calculating the degree of damage. 2

Manufacturing

Laminates of 16 layers were manufactured according to the ASTM Standard D7136/D7136M in a 12 by 12 in. aluminum mold by the vacuum assisted resin transfer molding (VARTM) process. Laminates were fabricated by placing layers of woven carbon fiber in an aluminum mold with flow media, breather and nylon ply, as it can be seen in Fig. 3. This was followed by enclosing the mold in a vacuum bag and drawing it into vacuum in order to aid infiltration of the resin (vinyl ester resin). The specimens for arctic and room temperatures were manufactured together as one laminate in order to ensure that the curing conditions were identical as it can be seen in Fig. 4. A total of 6 laminates were manufactured. From each laminate 6 samples were obtained (3 for room temperature and 3 for arctic temperature).

Fig. 4. Composite materials.

Fig. 3. VARTM configuration for the laminate.

P.168

IMPACT RESPONSE OF WOVEN COMPOSITES

3

Experiments

Drop-weight impact tests were performed on the specimens with a hemispherical impactor (striker). Each laminate was clamped between two fixtures with a circular testing area as it can be seen in Fig. 5. The impact load was concentrated at the center of the specimen in the out-of-plane direction. The specimens were tested for the following energies: 2, 5, 10, 20, 25, 30, and 35 Joules at room and at arctic temperatures. The specimens were 100 mm long x 100 mm wide with an average thickness of 4.02±0.023 mm. The samples were impacted until perforation or until the degree of damage for the samples was greater than 0.8.

Fig. 5. Impact testing set up.

4

Results and discussion

Impact responses were evaluated in terms of visual damage of the impacted specimens and by calculating the degree of damage. Force-time, energy-time and displacement-time responses were obtained from each test. A total of 3 samples for each energy were tested. Fig. 6. shows the energy-time plot when the specimen was impacted with 20 J. Table 1 shows the damage degree calculated. When the damage degree was lower than 0.5, the main failure mode on the laminate was delamination. When the damage degree was equal or greater than 0.5 the main failure mode was fiber breakage at the bottom of the laminate. Perforation was present when damage parameter exceeds the 0.9.

Table 1. Damage parameter for 20J Fig. 6. Energy-time plot for 20 J

3 P.169

5

Conclusion

Impact tests were performed on laminates with 16 layers at room temperature. The failure mode in these laminates were fiber breakage. When the damage parameter was equal or exceeded 0.5, the damage to the laminate can be seen to the naked eye. As the energy increased, more damage was present in the laminate. 6 Future work Tests will be performed with the same energies as room temperature in arctic conditions. References [1] Bortz, Daniel R., César Merino, and Ignacio Martin-Gullon. "Mechanical Characterization of Hierarchical Carbon Fiber/nanofiber Composite Laminates."Composites Part A: Applied Science and Manufacturing, Vol. 42 No.11, 1584-591. 2011. [2] Czabaj, Michael W., and James G. Ratcliffe. "Comparison of Intralaminar and Interlaminar Mode I Fracture Toughnesses of a Unidirectional IM7/8552 Carbon/epoxy Composite. "Composites Science and Technology, Vol. 89, 15-23. 2013. [3] Szekrényes, András. "Prestressed Fracture Specimen for Delamination Testing of Composites." Int J Fract International Journal of Fracture, Vol. 139, No. 2, 213-37. 2006. [4] C. Bouvet et al., “Low Velocity Impact Modelling in Laminate Composite Panels with Discrete Interface Elements,” International Journal of Solids and structures, Vol. 46: 2809-2821, 2009. [5] Icten, N. "Repeated Impact Behavior of Glass/epoxy Laminates."

SETS2016-47 The Southwest Emerging Technology Symposium 2016

Mechanical Properties of Hot Mix Asphalt Materials at Room Temperature for Use in Aerospace Landing Applications Jesus G. Reyes1 and Calvin M. Stewart1 1 Department of Mechanical Engineering, The University of Texas at El Paso Key Words: Hot mix asphalt, Fracture energy, Fracture mechanics, etc.

Abstract The quality and thus mechanical properties of pavement roads used at airports and space-stations has an enormous impact on the safety of the flights as take-off and landing are generally the most complicated steps to complete a flight. The main objective of this work was to investigate experimentally mechanical properties of a super pave dense graded (SP-D) hot mix asphalt (HMA) using two widely accepted fracture parameters test methods for granular materials: the semi-circular bend (SCB) test and the disc-compact tension test (DCT). The fracture parameters obtained following these procedures are key properties of materials requiring strong mechanical properties due to their applications, such as fracture energy and fracture toughness. The SCB test yielded a lower fracture energy value with a higher COV, while the DCT test results showed the opposite pattern. Digital image correlation was utilized during the test and proved to be useful in predicting the crack path in the samples.

1. Introduction In the aerospace industry, an overlooked area of scientific interest not related to the transportation vehicles in themselves, is the quality and information available on the land strip at airports and space stations where the aircraft and spacecraft takes off and lands. Analyzing the mechanical properties comprehensively therefore, is an area of great importance for researchers and engineers. Recent years have seen a trend appear of researchers in the pavement fields turning to fracture mechanics to better predict the performance of pavement under certain stress conditions [1]. A very important quality of a material is fracture toughness, as it refers to the ability of a material to resist failure with a crack present. Two of the most accepted testing methods to determine this quality in HMA materials are the semi-circular bend (SCB) test and the disc-compact tension test (DCT); the former yields two fracture parameters, fracture energy, Gf and fracture toughness, KIC, while the latter produces the single Gf value to describe the material’s resistance to failure with the presence of a crack in it.

2. Material A dense-grade superpave (SP-D) hot mix asphalt (HMA) was investigated under this study. The HMA mix contains reclaimed asphalt pavement (RAP) and recycled asphalt shingles (RAS). Any HMA asphalt material is usually composed of three constituent materials: the asphalt acting as the binder, the finer grains, and the larger aggregates. For this study, ten SCB specimens with a notch depth of 24 ± 1.5 mm were tested to obtain the fracture energy and toughness. Eight specimens with a diameter of 150 ± 10 mm, thickness of 50±5 mm, two holes of 25±10 mm across a notch with a length and width of 72.5 ± 2.5 mm and 1± .1 mm, respectively of the same HMA material were manufactured and machined. 1 P.171

3. Testing Procedure The DCT test and SCB are both widely accepted by researchers for determining the fracture parameters of asphalt-aggregate mixtures testing; however, the DCT test provides a greater fracture area to occur and is cited in literature as yielding lower COV’s than SCB. The main advantages of SCB is that it is more material-efficient, geometry is easier to machine, and the setup provides flexibility to investigate Mode I, Mode II, and combined Mode failures by the type of notch that is machined. For the determination of fracture energy, the AASHTO TP105-13 standard method of test was employed [2]. The schematic is shown in Figure 1a. The fracture energy, Gf is a fracture mechanics concept that represents the amount of energy required to create a unit surface area of a crack. The fracture energy is obtained by dividing the work of fracture (the total area under the load vs. load line displacement response curve) by the ligament area (the product of the ligament length and thickness). The average fracture toughness, KIC is then computed using this fracture energy value and considering other geometric factors of the specimen. For the disc-shaped compact tension geometry, the ASTM D73-13 [3] method was followed, as shown in Figure 1b. All tests were performed at room temperature.

Figure 1. Schematic of (a) semi-circular bending test and (b) schematic of disc-compact tension test.

4. Results and Analysis With the SCB approach the average fracture energy, Gf was 0.595 kJ-m-2 with a standard deviation and coefficient of variation (COV) of 0.160 kJ-m-2 and 27% respectively, while using DCT the average Gf was .873 kJ-m-2 with a deviation of .103 kJ-m-2 and a COV of 11.9%. The observation is that these specimens presented fraction parameters for both testing methods within the range of each other. The average fracture toughness, KIC was 0.287 MPa-m0.5 with a standard deviation and COV of 0.037 MPa-m0.5 and 13% of COV was also calculated with the SCB method. The results are summarized below in Table 1. Disk Compact Tension Area, Peak Peak (lbs-in) load (lbs) Load, N Average STD Deviation COV, %

Semi-circular Bending Gf, in.lbf/in^2

29.43

133.57

594.16

4.984

3.49 11.85

16.49 12.35

73.36 54.92

0.591 11.85

Gf, kJ/m^2 0.873 Average Std 0.103 Deviatio 11.85 COV, %

Work of Fracture Max Fracture Energy, Load, N ,J kJ/m^2

Notch Depth, mm.

Fracture Toughnes s, KIC

1400

1.57

0.595

24

0.287

178

0.43

0.16

1

0.037

13%

27%

27%

3%

13%

Table 1. Results summary for the semi-circular bending and disk-shaped compact tension tests. 2 P.172

The following images below, in Figure 2, were obtained from the VIC-3D Digital Image Correlation system for each SCB and DCT test. Figure 2a shows the magnitudes of the crack opening strain, εxx, of the SCB test. Figure 2b on the right side, depicts the magnitude of the crack opening strain, εyy, occurring in DCT test right before the crack began to propagate. These images show that the DIC technique can be useful in predicting the crack path of the sample during a fracture parameters test. εxx (10-2) 0.24

0.0131 0.1335

0.0024 0.027

-0.079

(a)

(b) Figure 2. Crack opening strain εxx of SCB test (a) and opening strain εyy of DCT specimen (b). -0.186

5. Conclusion The SCB and DCT tests proved to be adecuate tests to perform to identify fracture energy parameters of HMA asphalt materials. While the SCB standards permit the calculation of the fracture toughness value, KIC, it did yield a higher COV in the comparable value of fracture energy with the DCT standard by more than double (27% to 11.9%), as cited in literature. These parameters can be utilized to label and directly compare HMA materials based on their mechanical properties. Digital Image Correlation is a useful tool for both these tests in obtaining a full-field contour of the strains in the material and predicting the main crack path that fails the specimen.

6. References [1] Bashin, A., Masad, E., Kutay, M. E., Buttlar, W., Kim, Y. R., Marasteanu, M., ... & Carvalho, R. L. (2012). Applications of advanced models to understand behavior and performance of asphalt mixtures. Transportation Research E-Circular, (E-C161). [2] AASHTO TP105-13 (2013) Standard Method Of Test For Determining The Fracture Energy Of Asphalt Mixtures Using The Semicircular Bend Geometry (SCB), American Association Of State Highway Transportation Officials, DC. [3] ASTM D7313-13 (2013) Standard Test Method For Determining Fracture Energy Of Asphalt Aggregate Mixtures Using The Disk-Shaped Geometry.

3 P.173

SETS2016-48 The Southwest Emerging Technology Symposium 2016 An Element Activation Approach for Modeling Electron Beam Sintering of Titanium C. Beas COURI Research Assistant Mechanical Engineering University of Texas at El Paso clbeas @miners.utep.edu

J. Chessa Associate Professor of Mechanical Engineering University of Texas at El Paso [email protected] Keywords: Additive Manufacturing, Electron Beam Melting, Laser Engineering Net Shaping ABSTRACT Additive Manufacturing (AM) allows for rapid generation of complex parts from computer control. Processes such as Electron Beam Melting (EBM), Direct Metal Laser Sintering (DMLS), Laser Engineering Net Shaping (LENS) and Binder Jetting, employ metals such as aluminum, titanium, and steel. Mechanical properties are innately tied to the sintering mechanics associated with the AM processes. A study of the local heating and resulting residual stress formation for a model typical of an EBM process is presented. This method employs an element activation/deletion approach to reproduce the sintering densification of the titanium Ti6Al4V powder in the commercial finite element program Abaqus. The model includes include specifying temperature dependent material properties for the titanium alloy Ti6Al4V, and making use of the DFLUX user defined to model the moving path of the heat source (laser). The effect of the heat flux and heat flux path on the resulting residual stresses is considered. METHODOLOGY AND APPROACH The development of our stress model implements the use of the Abaqus CAE software, Hypermesh software, and the Fortran90 language. Abaqus CAE was chosen due to it’s ease in manipulating factors relevant to our study including specified temperature dependent material properties, element activation/ deletion and a DFLUX user defined subroutine to define the path of the applied surface heat flux. To postprocess the analysis we will use Abaqus View which visually represents residual stresses present after the AM simulation is completed. The DFLUX subroutine was formulated to specify a cubic spline heat distribution and define the path. The path used in this study resembles the AM process, the source moves through rows of elements when a row is completed it moves up one column and continues to navigate through the new row of elements following this pattern until every element has been passed through.

P.174

Employing a refined mesh, a moving heat source, specified heat distribution, temperature dependent material properties, and element activation/ deletion narrows in on the physical factors that affect the residual stresses following the AM process.

RESULTS The figures below show our model using a user defined DFLUX subroutine to model the moving heat flux and temperature dependent material properties.

Fig. 1 A moving heat source propagating through a Ti6Al4V wall

Fig. 2 A validation model showing a Ti6Al4V that was heated to it’s melting temperature of 1660K and cooled back down to 0K exhibiting residual stresses.

Our models will continue to be in development by combining all aspects and employing step in which each element will be activated once the moving heat source has propagated over it.

P.175

CONCLUSION As the development of this this residual stress model continues to be in progress, the validation model and previous studies assure that residual stresses will be present at the end of the AM simulation. Then we will be able to quantify how much these stresses affect the mechanical component as a whole. Following this study we will be able to expand our models into more complicated geometries as well as experiment with the paths and magnitudes of the heat flux to analyze which practices will result in the least amount of residual stresses. Each different practice will then be confirmed by physical failure tests. This project is part of a larger effort to develop coupled thermal-mechanical computational models in which we can emulate the AM process and quantify the residual stresses and microstructural formations from powder to fully dense components. REFERENCES [1] W. (Langley R. C. Seufzer, “Addiative Manufacturing Modeling and Simulation,” NASA/TM-2014218245, 2014. [2] K. Puebla, “Effect of Melt Scan Rate on Microstructure and Macrostructure for Electron Beam Melting of Ti-6Al-4V,” Mater. Sci. Appl., vol. 03, no. 05, pp. 259–264, 2012. [3] S. Kolossov, E. Boillat, R. Glardon, P. Fischer, and M. Locher, “3D FE simulation for temperature evolution in the selective laser sintering process,” Int. J. Mach. Tools Manuf., vol. 44, no. 2–3, pp. 117–123, Feb. 2004. [4] I.A Roberts, C.J Wang, R.Esterlein, M. Stanford, D.J. Mynors, 2010, “A three-dimensional finite element analysis of the temperature field during laser melting of metal powders in additive layer manufacturing”, International Journal of Machine Tools & Manufacture 49. Pp 916-923 [5] M. Akbari, D. Sinton, M. Bahrami, “Moving Geat sources in a half space: Effect of source geometry” ASME 2009 Heat Transfer Summer Conference, 2009.

P.176

SETS2016-49 The Southwest Emerging Technology Symposium 2016

ARTIC EXPOSURE STUDIES OF VINYL FOAMS FOR SANDWICH COMPOSITES

1

C. D. Garcia1*, P. Prabhakar1* Department of Mechanical Engineering, University of Texas at El Paso, El Paso, TX 79968, USA; *

Corresponding author ([email protected])

Keywords: vinyl foam, sea water, diffusivity, mechanical properties ABSTRACT Sea water and cold temperatures appear to have an adverse effect on naval materials resulting in the degradation of their mechanical properties. In this paper, the effects of sea water absorption on the mechanical properties of vinyl foams are discussed. For the study, moisture absorption was periodically measured until weight gain equilibrium or saturation was reached for samples submerged in sea water and deionized water. Diffusivities and saturation values were obtained from moisture uptake curves. It was observed that that the moisture content was higher for the vinyl foam samples submerged in deionized water compared to the samples submerged in synthetic sea water. Diffusivities were 9.227E-06 mm2/sec and 1.390E-05mm2/sec for deionized water and sea water conditioning, respectively. Flexural and compression tests were then conducted on saturated samples to compare their response against non-exposed samples. Experimental findings showed degradation in the flexural modulus and the compressive modulus for saturated wet samples. This occurrence can be observed in both tests with more prominent reduction in samples submerged in deionized water. Such a reduction is attributed to the degradation caused by the water, both DIW and sea water, in the form of surface damage to specimens.. 1

Introduction

With an increased interest in arctic exploration, it has become critical to understand the effects of extreme conditions on naval structures. A major concern for composite materials being used in marine applications is their behavior when exposed to sea water. Due to the corrosion that materials may experience under sea water and cold temperatures, it is important to determine the extent of degradation experienced by naval materials by comparing their mechanical properties. Sandwich composites have been predominantly used for marine applications due to their lightweight, enhanced performance and affordability. Some applications for sandwich composites materials may include but not limited to the use in the ship hulls. Earlier researchers such as Avilés and Aguilar-Montero [1] and Siriruk, Penumadu and Sharma [2] have investigated the effects of moisture absorption on sandwich foam materials. But, seldom work has been reported that discusses the effects of arctic sea water conditioning on foam materials. With the overarching goal of improving the performance of sandwich composites in arctic sea water conditions, the focus of the work presented here is on vinyl foams that are used as core materials in sandwich composites. The objective of this paper is to investigate the effects of sea water absorption in arctic environment on the mechanical properties of vinyl foams in order to develop methods to improve their durability and reliability.

1 P.177

ARTIC EXPOSURE STUDIES OF VINYL FOAMS FOR SANDWICH COMPOSITES

2

Experiments

2.1 Specimens Vinyl ester foam material was chosen for this study due to its potential naval application for sandwich composites. The dimensions of the coupons were determined based on the test the specimens were going to be subjected along with recommendations from ASTM standards C272 and ASTM D5229 for moisture absorption. Specimens were subjected to three-point bending test and compression test after saturation due to water immersion. Two groups of specimen sizes were used. For Group A, the specimen size was set to be 127x12.7x3.2 (mm) to comply with the flexural ASTM D790 testing standard dimensions. For Group B, a specimen size of 75x75x13 (mm) was chosen to comply with the ASTM C272 recommended dimensions. 2.2 Water Uptake: Measurements and Calculations For this study, specimens were submerged in a water bath of deionized water and synthetic sea water. A total of twelve specimens (6 for Group A and 6 for Group B) were submerged for each liquid type. The weight gained by the material was recorded periodically to determine the moisture content in the material. In accordance with ASTM C272, the weighing of specimens was continued until the mass gained for the last 48 hours was less than 2% of the entire mass gained in the previous intervals, which is an indication of saturation. To determine the moisture absorbed by the material at any time t, M(t) was determined using 𝑤(𝑡)−𝑤 𝑀(𝑡) = 𝑤 0 (1) 0

where w(t) is the weight at the given time t and wo is the initial weight. Coefficient of diffusion, or Diffusivity, of the material was obtained by plotting the moisture content against exposure time1/2 experimental curve in combination with the following equation: ℎ 𝑑𝑀(𝑡) 𝐷 = 𝜋(4𝑀 )2 ( 𝑑 𝑡 )2 (2) 𝑠𝑎𝑡

√

where, h is the thickness of the sample, Msat is the saturation value, and

linear region of the moisture content vs √t curve.

𝑑𝑀(𝑡) 𝑑 √𝑡

is the slope of the

2.3 Flexural and Compression Tests To determine the flexural properties, specimens of Group A were tested using ADMET eXpert 5600 Series UTM by following ASTM D790. A crosshead motion rate of 1.365 mm/min was employed to evaluate the flexural strength and flexural modulus of elasticity. Flexural strength and modulus of elasticity were calculated using 3𝑃𝐿

𝜎𝑓𝑚 = 2𝑏𝑑2

&

𝐿3 𝑚

𝐸𝐵 = 4𝑏𝑑3

(3-4)

where σfm is the stress in the outer fibers at the midpoint, P is the load at a given point, L is the support span, b is the width of the beam tested, d is the depth of the beam tested, EB is the flexural modulus and m is the slope of the tangent to the initial linear region of the load-

2 P.178

ARTIC EXPOSURE STUDIES OF VINYL FOAMS FOR SANDWICH COMPOSITES

deflection response. Following ASTM C365 for flatwise compressive properties, specimens of Group B were tested using Instron 5969 Tabletop Universal Testing Systems. The head displacement rate used for the test was 0.1 mm/min. The compressive modulus was calculated using 𝑓𝑐

(𝑃

−𝑃0.001 )𝑡 0.003 −𝛿0.001 )𝐴

𝐸𝑧 = (𝛿 0.003

(5)

𝑓𝑐

were 𝐸𝑧 is the core flatwise compressive chord modulus, P0.0 0x is the applied force at a given

deflection, t is the specimen mean thickness, δ0.00x is the recorded deflection value, and A is the cross-sectional area. 3 Results and Discussion

As it can be seen in Fig. 1 & 2, the moisture content for sea water specimens was considerably lower compared to deionized water. Also, the coefficients for diffusion were 9.227E-06 mm2/sec and 1.390E-05mm2/sec for deionized water and sea water conditioning, respectively. This was expected due to the fact that synthetic seawater has a higher density than deionized water. Moisture Content Increase with Time

Moisture Content Increase with Time

250

180 160

200

140 120

M(t)%

M(t)%

150 100 80

100 60 40

50

Specimen 5 Data fitted curve 0

0

5

10

15

20

25

30

35

40

45

Specimen 3 Data fitted curve

20 0

50

0

5

10

15

t1/2 (h1/2)

20

25

30

35

40

45

t1/2 (h1/2)

Figures 1 & 2. Moisture Content for Deionized Water (left) and Sea Water (right) Specimens

For the flexural tests, wet samples showed a degradation of 10.8% to 13.9% in their flexural modulus compared to dry samples. Also, a decrease in flexural strength is observed of 8.75% between dry and wet conditioning. For Compression tests, wet samples decreased between 18.4 % to 18.8% for its compressive modulus with respect to dry samples. The compressive stress at 2% deflection showed a slightly higher value for dry samples. Table 1 summarizes the average experimental values obtained for each testing. Degradation in the mechanical properties of samples conditioned to deionized and synthetic sea water can be attributed to the surface deterioration of the specimens. Table 1. Specimen Mechanical Properties Compressive Stress at 2% Deflection (MPa) 0.094

Compressive Modulus of Elasticity (MPa)

0.847±0.025

Flexural Modulus (MPa) 23.791±1.337

Deionized Water

0.773 ±0.027

20.483±1.123

0.093±0.003

10.536±0.338

Sea Water

0.773±0.048

21.212±0.588

0.090±0.009

11.048±0.499

Conditioning

Flexural Strength (MPa)

Dry

12.983

3 P.179

ARTIC EXPOSURE STUDIES OF VINYL FOAMS FOR SANDWICH COMPOSITES

References [1] Avilés, F. and Aguilar-Montero, M. (2010), “Moisture absorption in foam-cored composite sandwich structures”. Polym Compos, 31: 714–722. doi: 10.1002/pc.20872. [2] A. Siriruk, D. Penumadu and A. Sharma “Effects of Seawater and Low Temperatures on Polymeric Foam Core Material”. Experimental Mechanics (2012) 52:25–36. DOI 10.1007/s11340-011-9564-2 [3] Gabriele Tagliavia, Maurizio Porfiri, Nikhil Gupta, “Influence of moisture absorption on flexural properties of syntactic foams”, Composites Part B: Engineering, Volume 43, Issue 2, March 2012, Pages 115-123, ISSN 1359-8368.

4 P.180

SETS2016-50 The Southwest Emerging Technology Symposium 2016

500 LBF LIQUID METHANE - LIQUID OXYGEN THROTTLEABLE ROCKET ENGINE DESIGN J. Carroll, J. Candelaria, J. Trillo, C. Haynes Center for Space Exploration and Technology Research (cSETR), University of Texas at El Paso (UTEP), El Paso, TX 79902; * Jonathan Candelaria ([email protected]) *Joseph Carroll ([email protected]) Keywords: Rocket, Throttleable, Liquid Methane, Liquid Oxygen ABSTRACT The CROME project is to design, build, and test a throttleable (4:1) 500 lbf rocket engine using liquid oxygen and liquid methane as propellant. With NASA being on a mission to eventually send humans to Mars, the decision to use methane as a propellant was spurred by future space missions using In-Situ Resource Utilization (ISRU) technology to create methane from the carbon dioxide in the Martian atmosphere. By developing the next generation of rocket engines using liquid methane, this project will contribute to the future of space exploration. The main objective and motivation for developing this engine is to fire it in space and gather performance data in low orbit. The engine will be the main propulsion system used in the Daedalus suborbital demonstrator. The project is expected to take 3 years and is funded under a NASA grant awarded to the University of Texas at El Paso (UTEP) Center for Space Exploration and Technology Research (cSETR). The engine’s highest rated thrust will occur at a nominal chamber pressure of 235 psi. The engine must be capable of being tested at El Paso ambient pressure, therefore, two different nozzles are required. One nozzle will have an area expansion ratio of 2.7:1 (optimum expansion for testing at El Paso elevation) and the other an expansion ratio of 40:1 for space vacuum pressure. These rocket characteristics were calculated using an iterative set of spreadsheets, and additional software (Rocket Propulsion Analysis - RPA) to confirm our findings. Additionally, a system schematic was created to aid in the development of the rocket engine. A combustion chamber and nozzle has been modelled based on the design requirements. The injector components which include the face, manifolds, acoustic instability baffles and interface flange, have been designed. A prototype model has been 3D printed. Currently, a Preliminary Design Review (PDR) is to be held to finalize our injector design in preparation for a Critical Design Review (CDR). Following the CDR, the manufacturing of our design will begin, and the water testing phase follows. Following the water testing of the engine, the first firing test at sea level atmospheric conditions either at the test facility being developed by UTEP or White Sands Missile Range, NM. 1

Methodology and Approach

The project is to design, build, and test a throttleable 500 lbf rocket engine using liquid oxygen and liquid methane propellant. Further exploring Mars is NASA’s current mission One of the reasons that using liquid methane is of interest to us and to NASA is because the carbon dioxide in Mars’ atmosphere can be converted into methane. This method of producing methane from the Martian atmosphere is called In-Situ Resource Utilization (ISRU). This

P.181

500 LBF LIQUID METHANE - LIQUID OXYGEN THROTTLEABLE ROCKET ENGINE DESIGN

engine will serve as the main propulsion system for the Daedalus suborbital test bed, and will be used to gather space flight performance data. Identifying the requirements of the rocket engine, both mission and derived, is a crucial first step for development of the engine. Approximately an entire semester was allocated to developing a comprehensive requirements document. The requirements document is subject to change due to design issues that arise during development. The document includes both the mission and derived requirements, explained further in sections 1.1 and 1.2. 1.1 Mission Requirements The project is driven by the mission. The mission requirements are identified by our customer, cSETR Director and Department Chairman, Dr. Ahsan R. Choudhuri. The mission requirements are as follows. The propellant combination for the rocket engine must be liquid oxygen and liquid methane. The maximum thrust of the engine will be 500 lbf. The rocket engine must be throttleable 4:1, from 500 lbf down to 125 lbf. 1.2 Derived Requirements With the mission requirements set, an additional set of derived requirements was needed to meet the mission requirements. This additional set of requirements was attained by the analysis of the engine performance calculations and parameters. A system diagram of the main components of our engine is shown Fig. 1. The diagram shows how all the requirements are related, and how the derived requirements must be achieved to satisfy one or more of the main requirements. The component requirement hierarchy are shown in Fig. 1. The key components of the engine are the thrust chamber, injector, igniter, structures, delivery system, and instrumentation.

Fig. 1 - System diagram of 500 lbf LOX/LCH4 rocket engine key components. 2

Design Phase

After determining the requirements for the engine, the design stage of development can begin. During the design phase, concepts and ideas are developed for approval by reviewers and staff. This is done to meet the requirements most effectively. The details designed during this phase are key to meeting the requirements. Component selection, geometry, design considerations, are all taken into account during this phase of development. 2.1 Propellant Feed System 2

P.182

500 LBF LIQUID METHANE - LIQUID OXYGEN THROTTLEABLE ROCKET ENGINE DESIGN

The propellant feed system includes all the elements that control the fuel and oxidizer to flow from the tanks to the main combustion chamber. There are many types of feed systems, but we selected the pressure-fed type. This is the simplest feed system type because it does not require any pumps or turbines, but relies solely on the pressure of the upstream propellant tanks. The pressure fed system can be used because of the low chamber pressure. The most important part of the feed system is the valves. This engine will use proportional flow valves, providing throttleability through the valves. It is a requirement that our engine can be throttle between 125-500 pounds of thrust, which correlates to minimum and maximum propellant flow rate. When the valves are partially opened, the flow rates can vary and thus throttle the engine. Other key components to the feed system required for testing are flow meters, pressure transducers, and thermocouples. These components make it possible to measure flow rates, pressure, and temperature in the system. They will be required for testing, but will not be necessary for the flight system. 2.2 Combustion Chamber & Nozzle The combustion chamber is a critical area in a rocket engine. The liquid oxidizer and fuel are atomized and mixed together in the chamber, followed by an ignition of the propellant. The ignition starts the combustion process, and the resulting high temperature, high pressure gases are accelerated through the throat and nozzle to create a force, giving the vehicle momentum. The law of conservation of momentum allows the thrust force to be described by the thrust equation Eq. (1). 𝐹 = 𝑚̇𝑣𝑒 + 𝐴𝑒 (𝑃𝑒 − 𝑃𝑎 )

(1)

Where 𝑚̇ is the propellant mass flow rate, ve is the exit velocity of the exhaust gases, Ae is the area of the nozzle exit, and Pe and Pa are pressures of the exhaust gases and the ambient atmosphere at the exit respectively [1]. When designing the combustion chamber, the goal is complete mixing and combustion, which is can be achieved by proper injector design and chamber dimensions. The standard for quantifying the chambers dimensions is the characteristic length, L*, which is described in Eq. (2), where vc is the volume of the chamber, and At is the area of the throat. 𝐿∗ = 𝑣𝑐⁄𝐴

𝑡

(2)

The actual value for L* is known once the combustion chamber is built, but the value required for maximum performance can only be determined from experimental testing. An initial value for L*, 26 in., was selected based on values that were used in the past by our contacts at Johnson Space Center.

3

P.183

500 LBF LIQUID METHANE - LIQUID OXYGEN THROTTLEABLE ROCKET ENGINE DESIGN

Fig. 2 – Rocket engine injector, chamber, and expansion nozzle assembly 2.3 Injectors The design of the injector is the more complex aspect of rocket engine design. There are different types of injectors and considerations such as performance, acoustic instabilities, film cooling, dribble volume, and manifold design must be designed. Because we are working with cryogenic propellant, the system must be designed to prevent the fluid from boiling into a gas, and ensure that a liquid phase is injected into the chamber. The liquid phase of both propellant types will yield the best performance and combustion. If boiling occurs, the engine will fail to operate correctly. Boiling of the gas will lead to a rapid increase in pressure both at the lines and in the combustion chamber. The injector must have excellent atomization and complete mixing of the propellants must be achieved. The development of our injector is backed by research and previous professional experiences. 2.3.1 Impinging Injector The injector is responsible for injecting, atomizing, and mixing the two propellants in the combustion chamber. It also directs the film cooling (liquid methane) to the combustion chamber walls. There are several different injector types that have been proven to be effective, but our propellant combination and engine size from the mission requirements influenced us to design an unlike impinging injector [2]. Unlike impinging will allow for adequate mixing at point of impingement, as opposed to “like-impinging”, which depends on the propellants mixing further in the combustion chamber and require a significant amount of injection pairs for proper interaction mixing from impinging fans. One of the main requirements for an injector is the total area of injection, which is the sum of all the areas of the injection holes. The injection area required for maximum performance is determined from the required injection velocity and propellant flow rates. In an ideal scenario, an impinging injector would have as many holes as possible, and the holes would be as small as possible. However, since there is a minimum practical manufacturing tolerance for each orifice diameter, the number of injection elements was determined by the minimum allowable orifice diameter and the total area of injection.

4

P.184

500 LBF LIQUID METHANE - LIQUID OXYGEN THROTTLEABLE ROCKET ENGINE DESIGN

Fig. 3 – Impinging injector manifold assembly 2.3.2 Pintle Injector The main advantage of using a pintle injector in a rocket engine is throttleability [3]. The pintle injector, with a complex actuation system allows you to vary the area of injection, thus increasing or decreasing the mass flow rate. A higher flow rate, increases chamber pressure and thus produces a higher thrust. Different designs have been developed in the past and the injector continues to be studied. The main disadvantage of pintle injectors is the poor efficiency. Pintle injectors do not perform as well as other injector designs, such as the impinging plate injector.

Fig. 4 – Pintle Injector Schematic Initially, an actuated variable area pintle injector design and set of requirements was created. The analysis for the area of injection range and sleeve displacement for a pintle post of ¼” diameter proved complex. The injection area range proved too small, and the tolerancing required would be a manufacturing challenge. If future research teams decide to continue this research, the analysis and programs created are documented. The proposed design and complexity of actuation led to the decision to prototype and potentially test a fixed area pintle injector. The fixed area pintle injector is currently under development. This rocket engine concept will be throttled through the valves similarly to the impinging injector. 2.4 Cooling Method The material selected for the combustion chamber and nozzle is Inconel 718. Inconel is a corrosion resistant, high strength material able to withstand operating temperatures of 5

P.185

500 LBF LIQUID METHANE - LIQUID OXYGEN THROTTLEABLE ROCKET ENGINE DESIGN

between -423 - 2300 F [4]. The gases during combustion will reach between 5000 – 6000 F. The wall temperature needs to be maintained by cooling. For both the impinging and pintle injector, film cooling is being utilized. Ports on the injector face near the combustion chamber wall will provide 30% film cooling to keep the wall temperature within operational temperature. During testing, this 30% will be reduced until the amount of fuel is minimized. 3

Status

The team is currently in the design phase of rocket engine development. The details of the impinging injector, such as tolerances, angle of injection per doublet, and overall net injection momentum have been finalized. A 3D printed model of our rocket engine has been created to identify any design or manufacturing concerns. 3.1 Future Work A preliminary design review is to be completed within a few weeks in order to have our injector design evaluated. After this review is completed, another PDR may be necessary to finalize the design of the injector. The team is in the process of working with both local and other manufacturers to get manufacturability feedback and a cost quote. This will affirm our design is feasible from a manufacturing standpoint. The CDR will contain detailed specifications of the design such as tolerances, weld locations, component drawings, interfacing methods, list of purchased components, assembly exploded view of parts, etc. This set of reviews is for the injector alone. An additional set of reviews, in the same sequence must be done for other components of the engine design. 4

References [1] Brown, Charles D. Spacecraft Propulsion. Ed. J. S. Przemieniecki. Washington D.C.: AIAA, 1996. [2] National Aeronautics and Space Administration. Liquid Rocket Engine Injectors. Cleveland: NASA, 1976. Monograph. [3] Gordon A. Dressler, J. Martin Bauer. TRW Pintle Engine Heritage and Performance Characteristics. Conference Paper. Redondo Beach, CA: AIAA, 2000. [4] Special Metals. "Inconel Alloy 718." n.d. Special Metals. pdf. 1 February 2016. .

6

P.186

SETS2016-51 The Southwest Emerging Technology Symposium 2016

DESIGN AND DEVELOPMENT OF A 1 MW OXY-FUEL MHD COMBUSTOR 1

B. Lovich1, M.J. Hernandez, L. Cabrera N. Love* Mechanical Engineering, UTEP, El Paso, TX 79968, USA; * [email protected]

Keywords: Direct power extraction, oxy-fuel, cooling channels

ABSTRACT This paper displays the techniques used for the design of a rocket engine at the 1 MW heat input scale. To purpose of this research is to aid in the investigation of alternative methods of power generation, specifically the MHD direct power extraction. In partnership with the DoE, this research group will provide the initial step of fuel and oxidizer injection, as well as combustion and acceleration of the gas. This combustor is then intended to be utilized by the DoE for direct power extraction. This paper goes over software employed, sources utilized, and equations used for the development of this combustor. This is an up to date breakdown of the methods used for both the 1 MW oxygen and methane combustor as well as its smaller predecessor. 1 Introduction of Direct power extractions Power development is one of the most crucial aspects of modern life as we know it and development of cleaner energy has been of concern recently. Such efforts to limit pollution have been made clean by such organizations such as the United Nations Framework Convention on Climate Change or the (UNFCCC). Only months ago did they pass the Paris Agreement [4] which set high standards on emissions. Green energy remains a large global concern making greener technology more and more sought after. The use of pure oxygen and methane (Oxy-fuel) in a MHD combustor as a stand-alone alternative or retrofitted additions to coal driven turbine generators allows for carbon capture capabilities as well as higher fuel efficiencies. 1.1 Why MHD While Open cycle MHD as use in power generation has been around for quite some time, advancements in limiting technologies have allowed for a promising return. One of the limitations was the strength of the magnetic field required for the power extraction process, however due to the discovery of niobium-titanium alloys for superconducting magnets’ winding in the 1960’s, higher Tesla magnetic fields could be induced [2]. Other previous limitations included computational power, and high temperature resistant materials both of which advanced significantly in recent years. The use of pure oxygen in combustion is cleaner and produces higher flame temperatures than air oxidizer counterparts, which allows for higher Carnot efficiencies. MHD Direct power extraction involves no mechanical aspect of the energy conversion process, which is necessary as any mechanical components would be melted away by the higher flame temperatures of 3000 K or more.

P.187

Fig. 1. Showing the various Carnot Efficiencies of power generation systems.

This process of chemical to electrical energy by combustion and acceleration through a nozzle gives the Oxy-fuel MHD combustors the higher Carnot efficiency capability shown above in Figure 1[3] The combustor discussed later is to be used for MHD direct power extraction however many of the intricacies and other components of these systems will not be discussed. 1.2 Scaling Aspects 1MW energy input was found and decided upon from the lack of research in such sizing with respect to power rating. One of the questions that remains to be answered with regards to open cycle MHD direct power extraction systems is the effects of scaling. The other aspect not well known is the effect on power generation when using the pure oxygen and methane vs the previously used fuel and oxidizer combinations, such as JP4 with oxygen enriched air. Helping answering the question of scaling my lead to much cheaper, much quicker iterations of the direct power extraction systems which could lead to rapid acceleration of the technology. UTEP has already developed and will soon begin testing of an oxy-fuel combustor, this initial combustor has a power rating of ~70 kW. The first combustor is rather small in comparison to the 1MW combustor however the process of designing this smaller design lead to the development of the 1MW combustor in a relatively short manner. The 1MW design has kept the same parameters has the small design whenever possible to give insight into this scaling problem.

P.188

2

Design Calculations

2.1 Parameters considered Parameter name

Symbol

Value

Equivalence ratio

Ф

1.1

Units

CH4 mass flow rate

0.018

kg/s

O2 mass flow rate

0.065

kg/s

Throat diameter

15.4

mm

101.6

mm

1

MW

Chamber Length

L*

Energy Input

2.2 Combustion Chamber The flow rate of methane and oxygen were chosen to have a total heat input of 1 MW, which was the starting point of this design. This was done by finding the higher heating value of methane and then finding the amount of methane needed to supply 1 MW from this value. This mass flow of methane combined with the mass flow of the oxygen, found from the equivalence ratio taken from the previous combustor, could be used to find the minimum throat diameter of the new combustor. CEA was used from this point to find the properties of the fluid given the inputs known. The inputs included the equivalence ratio, the various chamber pressures, methane as fuel, oxygen as an oxidizer, and expansion ratio. Many of the parameters taken from the previous design to help provide insight in the scaling effects include the chamber pressures, the fuel, oxidizer, and the contraction and expansion ratios. From this, assuming full combustion, the constant pressure specific heats, ratio of specific heats, and the Prandtl number at the chamber, throat and exit could be found. These were then used to find the overall heat transfer from the gas side to help in the design of the combustion chamber. The minimum wall thickness has been found to be 0.8 mm for required heat transfer rates to the coolant. This was found using the Bartz’ equation for heat transfer coefficient for the gas side and the equation for combined stress from the Modern Design book [1]. To do this the thickness was varied from 0.5 mm to 10mm in 0.5 mm increments and for each case the heattransfer coefficient was calculated and inputted in the combined stress equation to find the resultant stress on the wall. These were compared to the yield strength of Inconel 718 at the expected wall temperature for various chamber pressures. This was done as the chamber pressure will vary during testing. The limitation of chamber pressure was found for each and 0.8 mm had the range of chamber pressures neccesary for testing similarily to the previous oxy-fuel combustor as shown in Figure 2.

3 P.189

Fig. 2. The various wall thickness vs compressive stress to show the operating pressures of each set.

This figure displays the range of pressures possible for a given wall thickness as calculated from the process above. For 0.5 mm the pressure can theoretically be above 10, however manufacturing of a 0.5 mm thickness would be near impossible. The next step for this combustor will be the water cooling, as this combustor is expected to run for 5 minutes or more. To run for the duration intended the combustion chamber wall must be cooled, and water was chosen as the coolant and cooling channels will be incorporated on the outer wall of the combustion chamber. A rough estimate for required water flow rate has be conducted however this will not be discussed. This is the current state of the design of the 1 MW combustor. 3.1 Summary The new 1 MW combustor is still in its early stages however due to large contributions from the previous MHD direct power extraction team and other helpful sources, large leaps could be made in the design. The consideration of the combined thermal and pressure differential stress proved to be one of the driving factors in this design. The unique purpose of power generation shows how total heat input was used as a starting point for this rocket engine design. The purpose of this paper is to give a brief summary of the importance and techniques used for the design of this rocket engine. This is to show the parameters considered and details of the combustor, as this may help for further development of this technology.

P.190

3.0 References [1] D. Huzel, D. Huang “Modern Engineering for design of liquid-propellent rocket engines” . Volume 147, Washington DC: American Institute of Aeronautics and Astronautics. [2] P. J. Lee, B. Strauss “Nb-Ti - from beginnings to perfection”. Leiden: CRC Press. 2011. [3] T. Ochs, R. Woodside, D. Oryschchyn, and L. Kolczynski “Improvements in Exergetic Efficiency in High-Temperature Oxyfuel”. National Energy Technology LaboratoryIn-house Research. 2014. [4] UNFCCC. Conference of the Parties (COP). “Adoption of the Paris Agreement. Proposal by the President. Paris: United Nations Office at Geneva”. Geneva, 2015.

5 P.191

SETS2016-52 The Southwest Emerging Technology Symposium 2016

DESIGN AND TEST OF REGENERATIVE COOLING CHANNELS AND INJECTOR FOR A 2000 LBF LOX/LCH4 ENGINE A. Sandoval1, I. Lopez1, L. Bugarin1, J. Adams1, S. Soto1, A. Choudhuri1* 1 Mechanical Engineering, The University of Texas at El Paso, El Paso, TX 79968, USA; * Director of Center for Space Exploration and Technology Research (cSETR)

Keywords: engine, injector, methane, regenerative, cooling ABSTRACT This paper discusses the current design work for a regenerative heat exchanger and injector for a 2000 lbf liquid oxygen / liquid methane thruster. The regenerative heat exchanger uses internal rectangular channels in the combustion chamber wall that flow liquid methane from the engine’s nozzle exit to the injector. Implementing this cooling method assures that the methane is not wasted to avoid performance losses, and instead used to increase the energy content of the propellant prior to injection [5]. A shear co-axial and a pintle injector are being designed in parallel to demonstrate a throttleable engine with higher performance using proportional flow control valves and to be able to throttle the engine’s thrust level. The methodology and analysis for design of these components are shown, as well as a description of the testing that will be required for validating the design principles of the components prior to engine manufacturing. 1

Introduction

In rocket engine design, significant design issues include the high temperatures and instabilities associated with combustion. Particular attention is concentrated on the injector and cooling mechanism, as they both have a direct influence on rocket engine performance. Stability is a function of injector design, combustion chamber dimensions, and combustion kinetics. A good injector design has the ability to inject and atomize propellants that burn efficiently and deliver stable combustion. These parameters are dominated by the injection area and the pressure drop across the injector face. Consequently, a design was begun with those two parameters as design drivers for two injectors: a pintle injector and a shear co-axial injector. Although the pintle injector historically has shown good throttleability and combustion stability, it normally has a performance penalty [8]. Meanwhile, a shear co-axial injector should deliver higher performance at the expense of employing an alternative throttle method. However, since the propellant combination for this injector is usually of gas-liquid, it might prove useful should the regenerative fuel gasify before entering the injector [2]. A water test experiment is designed to measure the pressure drop across the injector, as well as visually characterize the spray characteristics for both injectors. Rocket combustion temperatures require an active cooling method to prevent melting of the chamber wall material hence, the use of cooling techniques is required to address this issue. Regenerative cooling is a preferred cooling method in rocket engines because it increases its overall efficiency [5]. Thus, a design was begun for a regenerative heat exchanger. This design uses rectangular channels to flow the coolant (liquid methane) from the nozzle exit of the thrust chamber to an exit near the injector face. However, because the coolant must not vaporize as it

P.192

The Southwest Emerging Technology Symposium 2016 absorbs the heat from the combustion chamber wall, a high heat flux experiment will being run to measure the heat transfer coefficient of methane inside of the desired channel geometry at the expected heat flux. 2

Engine Requirements

The following table contains the 2000 lbf engine requirements considered for the design of the engine components in question (i.e. injector and regenerative heat exchanger). Requirement

Value

Thrust Max Burn Time Min Isp MR Chamber Pressure Inlet operational propellant condition

Injected Propellant Conditions Cooling Method Engine Chamber Material Nozzle

2000 lbf max 4:1 Throttleability 40 s (steady state burn) 230 s (steady state) 2.7 262.8 – 81.1 psia LOX T max: -300 °F LOX P min: 312.8 psia LCH4 T max: -260 °F LCH4 P min: 362.8 psia LOX T: -280 to -300°F LOX P min: 262.8 psia LCH4 T: -260 to -175 °F LCH4 P min: 262.8 psia Fuel Regenerative ; or ≤30 % Fuel Film Cooling Inconel 718 (Alternative: Inconel 625) Expansion ratio = 4 (at sea level)

Table 1: Tabulated engine requirements

3

Injector Design

With the project requirements established, it was possible to create preliminary geometries to design both a pintle and shear co-axial injector, as illustrated in Fig. 1 & 2 respectively (The orange component represents the igniter that is to be employed in the engine). The major areas of interest in the injector design were the pressure drop across the injector and the area of injection. Moreover, the total momentum ratio and the momentum flux ratio were quantified for a pintle and shear co-ax injector, respectively.

Fig.2. Shear co-axial injector design

Fig.1. Pintle injector design

P.193

The Southwest Emerging Technology Symposium 2016 3.1 Pressure Drop and Area of Injection Adequate propellant pressure drop (ΔP) across the injector is of significant importance because it causes the increase in injection velocity necessary for propellant atomization and prevents low frequency instabilities [1][2]. From literature, a reasonable ΔP should be no less than 20 percent of chamber pressure [1][2]. Since the maximum chamber working pressure was to be 250 psig (262.8 psia), a 50 psi ΔP was chosen. In order to create the pressure drop, it is necessary to size the area of injection that will deliver that ΔP. This was done by employing Equation 1 [2]. 𝐴 = 𝑊√

2.238𝐾 𝜌Δ𝑃

(1)

Where A is the area of injection, W is the propellant weight flow rate, K is the minor loss coefficient, ρ is the propellant density, and ΔP is the pressure drop. Employing this formula, it was determined that the total area of injection for both LOX and LCH4 should be 0.247 in2 and 0.126 in2, respectively. Utilizing the injection areas and the flow rates allows for the calculations of the nominal injection velocities; 55.3 ft/s for LOX and 108.3 ft/s for LCH4. 3.2 Total Momentum Ratio – Pintle Injector Design An important design factor in pintle injectors is the total momentum ratio (TMR). The TMR is defined as the ratio between the momentum of the radial jet and the momentum of the axial jet, and it is attained by employing Equation 2 [4]. 𝑇𝑀𝑅 =

(𝑚𝑣)𝑟 (𝑚𝑣)𝑎

(2)

Where m is the mass flow rate, v is the jet velocity, and subscripts r and a stand for radial and axial, respectively. The fuel (LCH4) was selected to be the radial propellant to keep the oxidizer closer to the wall and avoid localized thermal impingement/combustion near the injector faceplate due to a reduced spray angle [1]. Using nominal flow rates, the TMR = 0.609, which is within conventional TMR values. Collision between propellant jets creates a conical spray; the resulting spray cone half angle depends on the TMR. To estimate the spray half angle, θ, a direct sum of the momentum vectors of both the radial and axial jets was conducted to find the resulting vector and the corresponding half angle. This results in the following relationship (Equation. 3) [1]. 𝜃 = 𝑡𝑎𝑛−1 (𝑇𝑀𝑅)

(3)

3.3 Momentum Flux Ratio – Shear Co-Axial Injector Design On the other hand, an important design factor in shear co-axial injectors is the momentum flux ratio. The momentum flux ratio is defined as the ratio between the momentum per unit area of the radial orifice and the momentum per unit area of the axial orifice, and it is attained by employing Equation 4 [4]. 𝑀𝐹𝑅 =

𝑚𝑣 ) 𝐴 𝑟 𝑚𝑣 ( ) 𝐴 𝑎

(

(4)

Where m is the mass flow rate, v is the velocity at the orifice, A is the area of the orifice, and subscripts r and a stand for radial and axial, respectively. Calculations for this injector design were done for a liquid-liquid propellant configuration, as required by the engine requirements. However, because shear injectors usually have a gas-liquid propellant configuration, the oxidizer (LOX) was selected to be the radial propellant in the case of regen fuel gasification.

P.194

The Southwest Emerging Technology Symposium 2016 The liquid is placed in the central port to allow the outer gas to shear the liquid and atomize the propellants. Consequently, the MFR=1.64, which is within conventional MFR values. It is important to note the MFR is a value used for comparison with previous shear co-axial designs to yield an idea on the quality and validity of the design. 4

Regenerative Cooling

A preliminary regenerative heat exchanger concept was designed using the engine requirements and employing a reasonable diameter assumption. 4.1 Channel Dimensions The cooling channel dimensions for the heat exchanger design are dependent on the properties of methane at the different sections of the engine. Different geometries were obtained through an iterative analysis to check whether the methane would remain sub-cooled or change phase through the combustion chamber channels. A channel diameter was first selected for analysis, ensuring that the channels physically fit at the throat of the engine (the smallest section of the engine). Assuming that the heat generated from combustion is absorbed by the engine combustion chamber, a heat balance analysis lead to Equation 5 [7], a function of the number of channels (N) with respect to Reynolds number, Prandtl number, and the heat transfer coefficient for the cooling channel. 4𝑚̇ 0.8 𝜇𝐶𝑝 0.4

𝑁 = [0.023 ( 𝜋𝜇 )

(

𝑘

)

Δ𝑇𝑘

(2𝑞

𝑜𝑢𝑡

−5

)] 𝐷

𝑑4

(5)

Next, defining the minimum thickness permitted between channels at the engine throat (the smallest cross section of the engine and highest heat flux) allows one to substitute Equation 6 [2] . Both these two combined allow a solution for the number of channels required to dissipate the heat at the throat for the channel size selected. 𝑁=

𝜋[𝐷𝑡 +0.8(𝑑+0.04)] 𝑑+0.04

(6)

The next step in this analysis was to use the combustion thermal properties and the adiabatic flame temperature (obtained from NASA CEA) to calculate the bulk temperature of the methane through the cooling channels. This was done at different sections of the engine contour to verify if the methane had remained subcooled. By using the channel size yielded by Equations 5 and 6, an iteration was conducted until methane theoretically proved to be fully liquefied at all points of the engine contour. The current design stands at a total of 49 channels of 0.14’’ hydraulic diameter. These parameters however, are subject to change because the heat transfer associated with the coolant side of the regenerative system is greatly influenced by the coolant properties at the exposed high heat fluxes during combustion. Therefore, an experiment is required to verify that the heat transfer analysis is valid and provide confidence in the function of the heat exchanger design. 5

Future Work

In order to verify the design principles and performance of both the injector and regenerative cooling channels, a series of experiments will be conducted on these individual components. 5.1 Injector – Water Test Set Up A series of water flow tests will be conducted to quantify the pressure drop across the injectors. The objectives for these tests are to measure and validate design flow resistance (ΔP vs. flow

P.195

The Southwest Emerging Technology Symposium 2016 rate) [1], and the visual characteristics of the spray pattern and angle. Furthermore, the mixture ratio and TMR will be varied to observe its effect on the spray angle. Moreover, the water test setup built to measure pressure drop across injectors (shown in Figure 3), will be designed for mobility should future tests require the use of an external facility. 5.2 Regenerative Cooling Channels – High Heat Flux Test Facility A high heat flux test facility was previously assembled and used with the purpose of testing the heat transfer properties of methane at conditions similar to a rocket engine combustion chamber (Fig. 4, right) [ 7]. Experiments will be conducted with this setup to measure the heat transfer characteristics of sub-critical methane flowing through different cooling channels. The channels will be heated using a copper conduction-based thermal concentrator. The left side of Fig. 4 shows a schematic of the high heat flux set up; the bottom block is the thermal concentrator, the middle block an aluminum block for heat flux readings, and the top block the channel sample. The methane will be liquefied with a condensing unit that is part of the setup.

Fig.3. Water test set-up

6

Fig.4. Regenerative cooling channel test set-up

Conclusion

A regenerative heat exchanger and injector were designed for a 2000 lbf liquid oxygen / liquid methane thruster. Experiments on the individual components will follow in order to verify their design principles and performance prior to engine manufacturing. 7

References

[1] I. Lopez, A. Johnson, A. Patel, R. Ponce, M. Lopez, E. Flores, G. Martinez, A. Choudhuri, “Design and water flow testing of a lox/lh4 pintle injector”, 5th Southwest Energy Science and Engineering Symposium, El Paso, TX, 2015. [2] D. K. Huzel, D. H. Huang, “Mondern egineering for design of liquid-propellant rocket engines”, Rev Sub ed., American Institute of Aeronautics and Astronautics, 1992. [3] K. Davis, E. Fortner, M. Heard, H. McCallum, H. Putzke, “Experimental and computational investigation of a dual-bell nozzle”, Worcester Polytechnic Institute, April 23, 2014. [4] V. Yang, M. Habiballah, J. Hulka, M. Popp, “Liquid rocket thrust chambers aspects of modeling, analysis, and design”, Rev Sub ed. American Institute of Aeronautics and Astronautics, 2004. [5] M. E. Boysan, “Analysis of regenerative cooling in liquid propellant rocket engines”, Middle East Technical University, December 2008. [6] C. H. Brown, “Spacecraft propulsion”, AIAA Education Series, American Institute of Aeronautics and Astronautics, 1996. [7] A. G. Trujillo, “An experimental investigation of liquid methane convection and boiling in rocket engine cooling channels”, The University of Texas at El Paso, 2014. [8] G. A. Dressler, J. M. Bauer, “TRW pintle engine heritage and performance characteristics,” AIAA 2000-3871, pp. 1-22, 2000.

P.196

SETS2016-53 The Southwest Emerging Technology Symposium 2016

CONCEPTUAL DESIGN AND SIMULATION OF A DIRECTLY HEATED OXYFUEL SUPERCRITICAL COMBUSTOR A. Badhan1, A S M Arifur Chowdhury1, Daniela I Aguilar1, N. D. Love * and Ahsan R. Choudhuri* 1 Department of Mechanical Engineering, The University of Texas at El Paso, El Paso, TX 79968, USA; * N. D. Love ([email protected])

Keywords: Conceptual layout, Directly heated oxyfuel combustor, Supercritical combustion, Pintle injector.

ABSTRACT The objective of this project is to develop high pressure oxygen and fuel (oxy-fuel) combustion systems that can be integrated into turbines designed for directly heated supercritical CO2 (SCO2) based power cycles. The Directly heated supercritical oxy-fuel gas turbines have the potential to become an essential addition to current power generation systems. They provide a higher thermal efficiency and more effective carbon capture techniques than the existing gas turbines. Due to the higher density of the working fluids the turbomachinery size can be reduced significantly, which in turn will minimize operational cost. However, to achieve supercritical working fluid at the turbine inlet the combustion needs to be conducted under enormous amounts of pressure; these values are is about 10 times those found in the present gas turbines. Current paper projected a conceptual design of directly heated oxyfuel supercritical combustor that can be used for supercritical phase power generation system. For the combustion chamber, bell shape geometry is proposed instead of conventional cylindrical shape geometry to ensure better mixing of the combustion products.

1

Introduction

The motivation of this project is to develop a conceptual design of a directly heated oxyfuel supercritical combustor that can produce supercritical working fluid to drive a turbine to produce 300 MW net power. The usage of indirectly heated SCO2 as working fluid in a power cycle has been growing in recent years on solar thermal and nuclear power plants applications [1] . The directly heated supercritical combustor can be an influential addition to the supercritical phase power generation system since it demonstrates higher thermal efficiency and provides option to recapture more than 90% of exhaust CO2[2,3]. Nevertheless, the design of a directly heated supercritical combustor needs to be developed. The proposed design adopted the conventional rocket engine concept. Four separate powerheads will be used to inject fuel and oxidizer mixer to combustion chamber. Powerheads are designed to fit for both Pintle, and Modified Shear Co-axial injectors. Furthermore, proposed combustion chamber has bell shape geometry which maximize the mixing of supercritical combustion product from four separate powerheads.

P.197

2

Objective

The realization of the combustion process is required to design a directly heated oxyfuel combustor. Despite a linear increase in gas turbine combustion pressures over the last few decades, the current operating pressure range (~ 30 atm) is significantly below the > 100 atm chamber pressure needed for directly heated supercritical power cycles. In contrast rocket combustion chambers are often designed to operate in well excess 100 atm. Thus, they may offer proven legacy technologies for the development of future supercritical oxyfuel combustors. The overarching goal of this project is to design a supercritical combustor based on a Liquid Oxygen (LOX)/Methane Rocket Engine. The proposed conceptual design of directly heated oxyfuel combustor on this paper is focusing on operating envelop determinations, system design. Modeling of the proposed combustor will be based on two different types of injectors: Pintle, and Modified Shear Co-axial injectors. 3 Methodology 3.1 Combustor Fig. 1(a) shows the LOX/Methane, rocket engine derived, O2-CO2/Natural Gas supercritical combustor. Each engine is a complete combustion module and connected to a transition piece to form the entire supercritical combustor unit. The schematic in Fig. 1(a) shows a four-module configuration, however, three- and five module configurations will be considered as well for scaling analysis

(a)

(b)

Fig.1. (a) 9kN LOX/Methane rocket engine derived O2/CO2/Natural Gas combustor, (b) Cut away of system showing power-head, combustors, and transition sections

This particular combustor design has two primary advantages: (i) direct implementations of LOX/Methane engine technologies and (ii) modularity and compatibility with current power turbine layouts. Each combustion module is included a power-head, a combustor-body, and a

P.198

CONCEPTUAL DESIGN AND SIMULATION OF A DIRECTLY HEATED OXYFUEL SUPERCRITICAL COMBUSTOR

transition piece to mesh with the combustor annulus (Fig. 1 (b)). The power-head consists of injector elements, valves, and a torch igniter assembly (Fig. 2 (a)). The proposed combustor is designed to use two types of LOX/Methane injectors: (i) Pintle hole and (ii) modified shear-coaxial [4]. The transition piece is designed to translate the flow from each combustor module to the primary annulus flow. The design shown in Fig. 1 is preliminary in nature, and has not been scaled or optimized. 3.2 Powerhead Design Fig. 2 (a, b, c) shows the powerheads and modified injectors where accommodations are made to introduce CO2 into the combustor. Another component which is critical to the proposed combustor development is the torch igniter assembly. UTEP cSETR has two LOX/Methane igniter technologies: (i) a coaxial swirl based igniter and (ii) LOX/Methane reaction control system (RCS) derived igniters [4, 5]. These igniters are high TRL and have been tested with a wide range of propellant conditions.

(b)

(a)

(c) Fig. 2. (a) Power-head showing injector, valves, and igniter systems, (b) modified shear coaxial injector and (c) modified pintle injector

Therefore, no major modifications is needed for their integration in the proposed combustor. Nonetheless, a limited number of validations tests are needs to be tested at their extended operability in proposed supercritical combustor conditions.

3 P.199

Based on the proposed combustor design, mixing and combustion of reactants is modeled in the next phase. ANSYS FLUENT version 15.0 is employed for combustor modelling. To enhance proper mixing of the reactants, Species transport model with realizable k-epsilon turbulence model was employed. Mass flow inlet and pressure outlet boundary condition was used in this simulation. Figure 3(a) and (b) show the pressure contour and velocity vectors of the proposed supercritical combustor. As shown, the pressure found is between 30 MPA inside the combustion chamber and as fluid proceeds towards the outlet, it creates eddies due to flow fluctuations.

(a)

(b)

Fig. 3. (a) Pressure contour of the proposed supercritical combustor, (b) Velocity vectors of the supercritical combustor.

4 Conclusion Due to high thermal efficiency and efficient carbon capture technique, directly heated supercritical combustor shows a promising prospect in power generation system. Nevertheless, conceptual design of the combustor needs to be developed for further proceedings on research arena. Current paper proposed a conceptual design of directly heated oxyfuel supercritical combustor that can be adopted for supercritical phase power generation system. References [1] Nassar, A., Moroz, L., Burlaka, M., Pagur, P., & Govoruschenko, Y. “Designing supercritical CO2 power plants using an integrated design system”. ASME 2014 Gas Turbine India Conference, New Delhi, India, pp. V001T08A004- V001T08A004, 2014. [2] Wall, T. F. “Combustion processes for carbon capture”, Proceedings of the combustion institute, 31(1), 3147, 2007. [3] NETL, Technology Development for supercritical carbon dioxide (SCO2) based power cycles, retrieved from http://www.netl.doe.gov/research/coal/energy-systems/turbines/supercritical-co2-power-cycles, 2014. [4] Acosta-Zamora, A. and Choudhuri, A., “Development of propellant feed, thrust measurement, and automation control systems for testing LOX/LCH4 reaction control thruster”, 49th AIAA/ASME/SAE/ASEE Joint Propulsion Conference & Exhibit, AIAA 2013-3839. (ITAR Restricted Paper), July 2013. [5] Flores, J., Sanchez, L., Dorado, V. and Choudhuri, A, “Experimental studies of uni-element shear coaxial injector for LOX/LCH4 propulsion research, 49th AIAA/ASME/SAE/ASEE Joint Propulsion Conference & Exhibit, AIAA 2013-3851, July 2013.

P.200

SETS2016-54 The Southwest Emerging Technology Symposium 2016

THE DESIGN OF A LOX/LCH4 REACTION CONTROL SYSTEM ENGINE 1

A. Johnson1, D. Ott1, P. Nunez1, R. Ponce1, Ahsan Choudhuri1* Mechanical Engineering, University of Texas at El Paso (UTEP), El Paso, TX 79968, USA * Director of Center for Space Exploration Technology Research (cSETR)

ABSTRACT The research conducted within this report consists of the design and development of a LOX/LCH4 reaction control system (RCS) engine. The development of this engine is part of a continued partnership between NASA Johnson Space Center (JSC) and cSETR. The engine design was based on previous engine developments which took place both at JSC and cSETR; the main design requirement for the engine was to deliver 5-7 lbf of thrust over steady state operation; the engine will be used in the RCS of the two vehicles being developed at cSETR. The design process along with testing that will take place are highlighted within this report. 1

Introduction

Future space missions are geared towards manned missions to Mars; these missions will require a substantial amount of fuel to deliver a large payload to make a manned mission possible [1]. The reduction in propellant weight for a return trip makes the mission cheaper and allows more space for mission critical equipment. Available in-situ methane resources in space make liquid methane an ideal propellant for space travel. Liquid methane is also less toxic and safer to handle than traditional propellants used in the past, making and propellant involved incidents less troublesome [2]. For the reasons discussed, the fuel and oxidizer for this rocket engine will be liquid methane and liquid oxygen, respectively. A reaction control engine (RCE) was developed to be used within an RCS for two vehicles. The vehicles are Janus & Daedalus, a robotic lander and a suborbital vehicle, respectively. The main requirements set for the engine are to deliver 5-7 lbf of thrust and be able to operate at steady state conditions. Future tests will be conducted to determine whether the engine meets the following performance requirements: reliable ignition, thrust repeatability, and safe steady state operation. 2

Engine Design Parameter Propellant Combination Thrust Mixture Ratio Chamber Pressure Specific Impulse

Description LOX/LCH4 5-7 lbf (nominal) 1.54 (nominal) 100 psia (nominal) 182 s sea level (Janus) 255s vacuum, expansion ratio of 40 (Daedalus)

Table 1. Design requirements for RCS thruster

Table 1 shows the primary requirements for the RCS engine. The design process of the RCE was an iteration of two major components: chamber volume and injection orifice geometry. The engine injector design was based on previous work done by both NASA JSC engineers and UTEP students. For injection 1 P.201

SOUTHWEST EMERGING TECHNOLGY SYMPOSIUM design it is desirable to turn as much of the available pressure loss as possible into injection velocity for best atomization and mixing [3]. In order to obtain proper dimensions for the orifice size, the equations below were applied [3] 2.238𝐾

𝐴𝑖𝑛𝑗 = √

[𝑖𝑛2 ]

𝜌∆𝑃

(1)

Equation 1 gives the total area of the orifices, where K is the head loss coefficient, ρ is the density, and ΔP is the pressure drop across the injector. The value of K was set to 1.7 due to the sharp edge geometry of the orifice [3]. Knowing the area is the sum of all the areas of the injector orifices, the equation below can then be applied to find the diameter of individual injector orifices 3.627𝐾𝑤 2

𝑑𝑜𝑟𝑖𝑓𝑖𝑐𝑒 = (

𝜌∆𝑃𝑁 2

0.25

)

[𝑖𝑛]

(2)

The variables within equation 2 are N and w, where N is the total number of orifices in the injector and w is the weight flow rate of the propellant.

Figure 1. Full RCE assembly (left) and cross section view of the chamber (right)

Figure 2. Injection geometries for both LCH4 (left) and LOX (right)

In the above figures 1 and 2, the final assembly along with the cross sectional view of the chamber and injector can be seen. The engine uses a modified spark igniter with an extended electrode tip for ignition; the ignition method has not been modified from the previous RCS engines that were developed at cSETR [4]. As seen in figure 2 (right), the liquid oxygen (LOX) injector geometry is a set of impinging injector elements. They are set to impinge right after injection since the igniter (which is placed in the center of the chamber) prevents good propellant atomization between all LOX jet streams. There are two main feed line injection ports for the liquid methane (LCH4): one for combustion and the other for film cooling. These were made independent to allow for a variation in percentage of cooling during hot fire testing. Figure 1 (right) shows how film cooling is achieved by placing an internal jacket within the chamber that allows the direction of the LCH4 along the chamber wall. The combustion injection pattern for methane (shown in figure 2, left) is a transverse stream to impinge and shear against the oxygen flow. After propellant injection and mixing, the spark from the igniter electrode arcs along the wall of the chamber 2 P.202

SOUTHWEST EMERGING TECHNOLGY SYMPOSIUM to start combustion. Furthermore, this engine has a slightly under-expanded nozzle for ambient pressure conditions in El Paso.

3

Engine Testing

3.1

Cold Flow Testing

The primary test objectives of cold flow testing are: determine pressure drop across the injector, the mass flow rates, and the discharge coefficient across the injector. These values will be compared to predicted data to see whether the injector parameters match the intended design principles. 3.2

Hot Fire Testing

Hot fire tests will be conducted to determine the following: steady state thrust, specific impulse vs mixture ratio, minimum film cooling percentage, minimum stay time for combustion, and ignition reliability. The film cooling percentage, mixture ratio, and the chamber length will be varied during testing. 4

Conclusion

A reaction control engine was designed with the purpose of being used within the systems of two vehicles. JSC provided cSETR with an initial pencil thruster design for a RCE. Over several design iterations the current RCE prototype is going to be tested in two phases: cold flow and hot fire tests. The cold flow test seeks to determine the relation between pressure drop and flow rate data to be able to determine the desired mixture ratio when firing the engine. Once the flow parameters are quantified, the hot fire tests will test the actual thrust and specific impulse of the RCE while varying several factors such as film cooling percentage, mixture ratio, and chamber length. The next step for this project will be to start generating the interface between the vehicles and the RCE to observe how they work together in a system. 5

References

[1] “NASA's Journey to Mars”. (n.d.). Retrieved January http://www.nasa.gov/content/nasas-journey-to-mars/#.VQDUFvnF_3Q.

1,

2015,

from

[2] S. J. Schneider, J. W. John, and J. G. Zoeckler “Design, Fabrication, and Test of a LOX/LCH 4 Igniter at NASA”. Cleveland, Ohio, Glenn Research Center, 2007 [3] D. K. Huzel and D. H. Huang “Modern engineering for design of liquid-propellant rocket engines”. Washington, DC, AIAA, 1992. [4] J. L. Mena “Performance Evaluation of a LOX-LCH4 Reaction Control System Thruster”. 2014

3 P.203

SETS2016-55 The Southwest Emerging Technology Symposium 2016

FREE EDGE EFFECT IN MULTIDIRECTIONAL LAMINATES 1

M.S. Islam1, P. Prabhakar1* Department of Mechanical Engineering, The University of Texas at El Paso, El Paso, TX 79968, USA; * Corresponding author ([email protected])

Keywords: Free edge, Interface, Laminate, Finite element analysis (FEA)

ABSTRACT In this paper, a generalized plane strain model capable of predicting interface stresses is developed. The analysis of free edge stresses of composite laminates subjected to uniform axial strain by using the generalized plane strain model developed with help of commercially available FEA software Abaqus. The results are compared with previously published results and 3D analysis to validate the model. This model can predict the interlaminar stresses accurately and can be used as a computational tool to predict the interlaminar stresses of different laminate under mechanical loading. 1

Introduction

Localized high interlaminar stresses occur at the free edges of laminated composite materials because of the property mismatch between plies, which is known as the free edge effect. Accurate prediction of stress distribution near the free edge is very important because edge effect is the main cause of delamination or transverse cracking. Classical single layer theories are not able to calculate these out of plane stresses on the free edges. Therefore, various approaches such as finite difference [1], 3D finite element [2], closed form analytical approach [3] , boundary layer theories and layer-wise theories have been used to calculate the interlaminar stresses near the free edges. In this paper, a variational formulation of the Pipes and Pagano [1] is presented. This generalized plane strain formulation allows to determine accurately the stress distribution near the free edges with the help of numerical model. Abaqus (a commercially available software) is used for numerical modeling in this paper. 2

Analytical Modeling

Consider a laminate of length 2L, width 2h and lamina thickness of h as shown in Fig.1. (a). Uniform axial strain is being applied at the edges of and along the x1 direction. L L Cross-section of the laminate is shown in Fig.1. (b). It is assumed that the stress components are independent of x1, so the displacement field U is given by,

P.204

(a)

(b)

Fig.1. (a) 3D laminate and (b) Cross-section of a 3D laminate.

~ U 1 ( x1 , x 2 , x3 ) U 1 ( x 2 , x3 ) ~ U 2 ( x1 , x 2 , x3 ) U 2 ( x 2 , x3 ) ~ U 3 ( x1 , x 2 , x3 ) U 3 ( x 2 , x3 )

x

11 1

(1)

Here, 11 is the applied strain in the laminate along x1 direction. The displacement field U and the associated stress field (U ) are solutions to the following set of equations: ij

0,

xj aijkh

ij

ij

kh

ij

ij

[

ij

xj

xi

[

ij

2b

Fi on

nj ]

Fi on

nj ]

Vi dx1 dx2 dx3

(2a)

) within

,

11

0,

and

(2b)

,

0 on the interfaces

nj ]

Introducing a trial field V defined by (1) with ij

Uj

0 on

nj ]

,

1,2,3 within

(U ), i, j

1 Ui ( 2 xj

(U )

[U i ] 0, [ [

1,2,3 within

i

(2c) k

0

L

L

,

(2d) (2e) (2f) (2g)

0 leads to, Vi with i=1,2,3.

(3)

Applying divergence theorem, Vi dx1dx2 dx3 FiVi dx2 dx3 FiVi dx2 dx3 0 (4) ij L L xj Reminding that Vi does not depend on x1 and taking a section (A-A in Fig.1. (a))in the (x2, x3) plane, finally we obtain,

P.205

FREE EDGE EFFCT IN MULTIDIRECTIONAL LAMINATES

~ U k Vi Vi ai k dx2dx3 dx2dx3 , Vi 11 ai 11 S S x x x Applying divergence theorem on the right hand side of (5), ~ p N U k Vi ai k dx2 dx3 aip 11n pVi ds 11 S Sp x x p 1

(5)

(6)

Where S p is the frontier of the pth layer, n p its outer normal and aip 11 is the corresponding elasticity coefficient. The above formulation can be implemented in several ways using finite element method. The method used here is to modify a thin slice of a 3D model to behave like a generalized 2D model. The model is restricted from any expansion in the x1 direction using multi point constraints which satisfies the requirement that the displacement fields are independent of the x1 direction. 3 Numerical results In this section, the developed finite element model was used to study [+10/-10]s angle ply laminate. The following material properties were used: E 11=97.6 GPa, E22=E33=8.0 GPa, 0.37 and 23 0.37 . Fig.2. shows the comparison G12=G13=3.1 GPa, G23=2.7 GPa, 12 13 of 13 at the +10/-10 interface with Martin et al. [4] and it is found that, the present model is in good agreement with Martin et al. [4].

Fig.2. Distribution of interlaminar shear stress along the +10/-10 interface of [+10/-10]s laminate.

The model was then used for [45/-45/90/0]s and [-45/45/90/0]s laminate to determine the critical interfaces. Fig.3. shows the interface strains for this two types of laminates at different interfaces. These interface strains are compared with 3D model as well and it has found that the results are in good agreement with 3D model. 4 Conclusions In this study, free edge laminates subjected to uniform axial strain is studied. This generalized plane strain model is capable of predicting the interface stresses accurately compared to 3D model. The full 3D model is computationally more expensive.

3 P.206

(a)

(b)

(c)

(d)

(e)

(f)

Fig.3. Distribution of normal and shear strains along (a) 45/-45, (b) -45/90, (c) 90/0 interface of [+45/45/90/0]s laminate and (d) -45/45, (e) 45/90 and (f) 90/0 interface of [-45/45/90/0]s laminate.

References [1] R.B. Pipes and N.J. Pagano “Interlaminar stresses in composite laminates under uniform axial extension”. Journal of Composite Materials, Vol.4, No. 4, pp 538-548, 1970. [2] A.S.D. Wang and F.W. Crossman “Some new results on the edge effect in symmetric composite laminates”. Journal of Composite Materials, Vol.11, No. 1, pp 92-106, 1977. [3] W. Becker “Closed-form solution for the free-edge effect in cross-ply laminates”. Composite Structures, Vol. 26, No. 1-2, pp 39-45, 1993. [4] E. Martin, D. Leguillon and N. Carrere “A twofold strength and toughness criterion for the onset of free-edge shear delamination in angle-ply laminates”. Journal of Solids and Structures, Vol. 47, No. 9, pp 1297-1305, 2010.

P.207

SETS2016-56 The Southwest Emerging Technology Symposium 2016 NUMERICAL STUDY OF HIGH-TEMPERATURE SCO2 VOLUMETRIC RECEIVER FOR CONCENTRATING SOLAR POWER SYSTEM. 1

A. Schiaffino*1, V. Kumar1 Mechanical Engineering, The University of Texas at El Paso, El Paso, TX 79968, USA; * Corresponding author ([email protected])

Keywords: renewable, solar, energy, receiver, CFD

Are you planning on attend the student paper poster: (Yes) ABSTRACT We are living in world with constant increase in energy consumption. Our primary source of providing this huge amount of energy is mostly fossil fuels. Earth has limited amount of these resources and with vast amount of energy demand, these resources are slowly becoming scarce. Every year, energy demand has been increasing throughout the world. According to the renewable energy data book from 2013 [1], United States overall energy consumption grew to 97.3 quadrillion BTU in 2013, a 2.4% increase from 2012. Energy consumption from coal and renewables grew slightly, while consumption from petroleum and natural gas fell slightly. Given that the sun is a highly renewable source of energy, there have been several techniques employed to generate energy by using solar radiation. Devices such as photovoltaic cells are used to generate electricity through an electrochemical reaction, on which the cell absorbs photons coming from solar light and releases electrons. These devices, however, are not useful to generate electricity during night. Another device used to absorb energy from the sun are the molten salt receivers, however, the molten salts used to transport the heat to the plant are not stable beyond 600ºC. [2] A volumetric solar receiver is a device that collects the energy present in the solar light and transfers it to a working fluid. The advantages of using concentrated solar power (CSP) devices is that they can be used during the night, and they achieve high performing efficiencies, since a heliostat field generates a beam of light that is absorbed by the receiver on a very specific part of its geometry. The purpose of this investigation is to develop the idea of a volumetric solar receiver compatible with supercritical carbon dioxide at 20 MPa and 820 ºC, which are the conditions present at most turbines’ inlets. The proposed design consists of a structure, which contains a porous mass on its center, designed to transfer the heat caught from the solar radiation and transfer it to the working fluid in an uniform way. MATERIAL SELECTION Given the temperature and pressure conditions at which the receiver must operate, a proper material needs to be selected in order to increase the endurance of the receiver and ensure its safe operation

P.208

Nickel based alloys were selected as a viable option for the manufacturing of the structure of the receiver, specially Haynes 230, this alloy is a nickel-chromium-tungsten-molybdenum alloy that combines excellent high temperature strength, outstanding resistance to oxidizing environments up to 2100°F (1149°C) for prolonged exposures, and excellent long-term thermal stability. Other attractive features include lower thermal expansion characteristics than most high-temperature alloys. Haynes 230 alloy combines properties which make it ideally suited for a wide variety of component applications in the aerospace and power industries, making it another material option that could be used to fabricate the structure of the receiver. STRUCTURAL DESIGN A design iteration for a structure was designed. The first thickness was calculated by Equation 1, which is included in the ASME Pressure Piping B31.1 Code design equation for pressurized tubes and pipes [3]. P D Equation 1 : Pressure vessel design equation t 2(( S E ) ( P y))

This equation was used to estimate the wall thickness for this design iteration, where t is the minimum thickness required excluding manufacturing tolerance and allowances for corrosion, P is the working pressure, D is the external diame S is the maximum allowable stress at working temperature, E is the joint efficiency factor, and y is the temperature coefficient. For Nickel-alloys y=0.7 at temperatures above 650 C , while for seamless tubes E=1. The S value given to the formula is to be 585 MPa, which is the ultimate strength at 800 C of the Haynes 230. The operating pressure term P is 20 MPa. After performing the calculations, it was decided that the thickness of operating at 800 and 20 MPa will be of 7 millimeters. Porosity Modeling An accurate representation of a porous media was accomplished by using Equation 2. This model was employed, since the flow through the porosity was assumed to be laminar, and fully developed. Equation 2: Ergun equation for fully developed laminar flow through a packed bed Dp 2 150(1

3

)2

The value given for Dp is for 10 micrometers, the porosity has a 75% of void volume. After performing the calculations, it was shown that the permeability for this porous mass was. Also, knowing that the viscous resistance can be obtained by using Equation 3. 1 Equation 3: Viscous Resistance Formulation RV Fluid Structure Interaction Modeling In order to properly simulate and represent the different physical phenomena related with this project, a fluid-structure interaction solver is being developed in order to understand the interaction between the different components of this model, which are flow through a porous media, heat transfer through a cylinder and the corresponding effects that these boundary conditions generate on a structural model.

P.209

This solver has been tested with other problems involving fluid structure interaction, and the results have been promising. One of the problems solved by this algorithm consists of the dampening of the oscillation of a beam by the air in the surroundings. A pressure is applied to the beam, consequently causing a deflection on the beam. Given that the beam is displacing air, this displacement is being feed into Fluent. Once the effects of this displacement are calculated by Fluent, a pressure variant is feed by Fluent to Structural, and so forth. Figure 1 and Figure 2 describe the oscillation of the beam during the simulation. The movement of air can be appreciated on both Figures, and the oscillation of the beam is affected by the air flow surrounding it. Conclusion and future work This paper describes the design of volumetric solar receiver that is going to be operated at 20 MPa of pressure, and within a range of 800 ºC -1000 ºC. The material selected for this design is Haynes 230, material selected for its durability and low deformation under high temperatures, a preliminary thickness of 7 mm was calculated by using the ASME pressure vessel design code. An important aspect of this design process is the utilization of computer modeling to assist with the creation of this prototype. A model for the porous media acting inside the receiver has been accomplished, and a fluid structure interaction solver is being developed in order to replicate in a precise and accurate way the different physical phenomena acting on the receiver.

Figure 1: Oscillation of a beam being damped by air References [1] NREL, “2013 Renewable Energy Data Book.” [Online]. Available: http://www.nrel.gov/docs/fy15osti/62580.pdf. [Accessed: October-2015]. [2] D. Kearney, U. Herrmann, P. Nava, B. Kelly, R. Mahoney, J. Pacheco, R. Cable, N. Potrovitza, D. Blake, and H. Price, “Assessment of a Molten Salt Heat Transfer Fluid in a Parabolic Trough Solar Field,” J. Sol. Energy Eng., vol. 125, no. 2, p. 170, May 2003. [3] American Society of Mechanical Engineers, «ASME Boiler and Pressure Vessel Code,» New York, 2013.

3 P.210

SETS2016-57 The Southwest Emerging Technology Symposium 2016 DESIGN OPTIMIZATION OF SANDWICH CORE AND MANUFACTURE THROUGH ADDITIVE MANUFACTURING M. Tauhiduzzaman1, Md. S. Islam2, P. Prabhakar2* Department of Computational Science, El Paso, TX 79968, USA; 2 Department of Mechanical Engineering, El Paso, TX 79968, USA; * Corresponding author ([email protected])

1

Keywords: Sandwich structures; Additive manufacturing; Shape optimization

ABSTRACT This research presents the finite element modeling of multifunctional sandwich composites to optimize the mechanical characteristics of core structures considering three-dimensional isotropic patterns. Short beam shear test was conducted for each structure that allows us to compare the load vs. displacement response with the model. Peak loads and deformations were recorded to compare the flexural properties. Fused deposition modeling method is employed to build the 3D printed structures. To obtain the new design of the sandwich cores with optimum stiffness and reduced weight shape optimization task is performed by ABAQUS. Stress and weight are the design variables to carry out the optimization method. This optimization method deals with the coordinates of surface nodes; eventually, it creates a new design of the cores that demonstrates versatile performance. Finally, based on the output of the optimization procedure new STL files are imported in the additive manufacturing machine to produce the optimized structure. 1

Introduction

The emergence of sandwich composites introduced a considerable progress in the aerospace, naval, and different engineering industries due to their multifunctional features. Sandwich panels are consisting of the core material, lightweight but exhibits high flexural stiffness and buckling strength, is covered by two facesheets which are stiff skins. Sandwich construction is based on the concept of cellular materials. Manmade cellular solids that used extensively at recent times because of its low weight, high stiffness, and durability. The performance of the cellular structures has spurred the designers to tailor an artificial structure that mimics the design of cellular solids [1]. Typically, stochastic and periodic architecture is being featured by cellular materials that can be classified as open cell and closed cell [2]. In this paper, honeycomb, prismatic, and lattice truss cores are considered for shape optimization technique. The skins of the honeycomb sandwich structure act like I beam because it can carry the load where the top face is in tension, and the bottom face is in compression. The aim of this paper is to address a design optimization procedure for four different sandwich structures that can be fabricated by polymer additive manufacturing process. The additive process where materials are combined layer by layer is used to make the plastic prototypes where computer controlled lasers are used as the energy source. The scope of this paper is as follows, four 3D geometry of different sandwich composites are developed following the geometrical interactions. A finite element modeling is employed to predict the load versus displacement response of each structure. After establishing the material properties, shape

1 P.211

DESIGN OPTIMIZATION OF SANDWICH CORE AND MANUFACTURE THROUGH ADDITIVE MANUFACTURING

optimization is conducted for all of the structures. Finally, four optimized sandwich structures, fabricated using the 3D printing, are utilized to perform an experimental study to validate the modeling and predicts the failure mechanisms. 2

Sandwich Construction & Experimentation

The core of the sandwich structures contains a regular pattern on a plane where the periodicity exists. Core materials feature three main periodic topologies which are prismatic, honeycomb, and lattice truss core. Printing procedure includes some steps initiated by creating CAD modeling, followed by STL conversion, subsequently transferring to the additive manufacturing machine, and finally build the structures by using PLA.

(a)

(b)

(c)

(d)

Fig. 1. Initial core design of a) Prismatic triangular core b) Lattice truss core c) Square honeycomb core d) Hexagonal honeycomb core

One of the challenging factors for modeling is acquiring the perfect material property, such as Young modulus, that has been incorporated to generate optimized shape by mimicking the short beam shear experiments. Experiments and modeling are formed following the ASTM D2344 standard that let us know about the testing procedure, dimension of the structures as well.

2 P.212

DESIGN OPTIMIZATION OF SANDWICH CORE AND MANUFACTURE THROUGH ADDITIVE MANUFACTURING

4

Result and Discussion

Shape optimization method considers stress and weight as a design variable, and design area refers to the core structures. In this process, nodes around the design area are displaced in order achieve the objective and thus new shape is obtained. Table 1, demonstrates the peak loads and deformations of the initial and optimized structures. Failure analysis of four different types of structures delivers the idea about general failure mechanism of the sandwich structures.

Structure

Peak load (lbf)

Deformation Failure mode (inch)

Prismatic triangular structure (Initial)

1132.55

0.14

Facesheet fracture

Prismatic triangular structure (Optimized) Lattice truss core (Initial)

282.61

0.97

Core shear

149.36

0.68

Indentation

Lattice truss core (Optimized)

255.24

0.84

Indentation

Square honeycomb (Initial)

440.12

0.23

Core shear

Square honeycomb (Optimized)

544.74

0.23

Core shear

Hexagonal honeycomb (Initial)

1099.30

412.94

0.14

0.23

Hexagonal honeycomb (Optimized)

Buckling Interlaminar shear

Table 1. Comparison of initial and optimized sandwich structures

5

Conclusion

In the present study, 3D sandwich structures are manufactured to verify the performance of the entire procedure at the same time it demonstrates a good comparison of different types of sandwich composites. According to the shape optimization, the facesheets are not taken into account in the design area, so it supplements more stiffness to the structures. In this research work, a novel process is developed to verify the application of polymer additive manufacturing process by using PLA (Polylactic acid) material. Therefore, the future work will focus on the fabrication process by using ABS (Acrylonitrile Butadiene Styrene). 6

References

[1] D. J. Sypeck, and H.N.G. Wadley “Cellular metal truss core sandwich structures”. Advanced Engineering Materials, Vol. 4, No. 10, pp 759-764, 2002. [2] T. George, and H. N. G. Wadley “Mechanical response of carbon fiber composite sandwich panels with pyramidal truss cores”. Composites part a: applied science and manufacturing, Vol. 47, pp 31-40, 2013.

3 P.213

SETS2016-58 The Southwest Emerging Technology Symposium 2016

NEXT GENERATION COMPUTING FRAMEWORK FOR EXASCALE SIMULATIONS Ashesh Chattopadhyay1*, VMK.Kotteda2 ,Vinod Kumar2 Computational Science Program, University of Texas at El Paso, El Paso, TX 79968, USA 2 Department of Mechanical Engineering, University of Texas at El Paso, El Paso, TX 79968, USA 1

[email protected]

Keywords: Trilinos, MFIX, Next Generation Computing, exascale, integrated framework

Keywords:ABSTRACT keywords list (no more than 5) With an emerging trend in solving large scale multi-physics and engineering problems through scientific computing, high performance computing architectures have provided a citadel for such applications. In this regard previous literatures have provided us with a plethora of libraries, software and toolkits that enable scientific computing on massively parallel architectures. In this work we talk about a developing language independent framework that integrates two of such powerful libraries and heads way for real time exascale simulations of engineering problems especially in the areas of multiphase flow. One of the libraries used in this work is Trilinos which is an object oriented framework composed of a number of different packages that amongst other things provide capabilities for implementing large scale massively parallel linear and non-linear solvers while the other is MFIX , which has been widely used as a large scale multiphase solver. While the code design in MFIX is based on FORTRAN, Trilinos has been designed in an objected oriented C++ 11 framework. In this paper we describe the methodology for integrating these two separate libraries in a language independent framework, thus exploiting capabilities of the finest and most impressive traits of both these packages in solving complex multi-physics problems laying way for exascale computing abstractions.

1

Introduction

In general, parallel and distributed systems have found their applications in large scale engineering problems since some time now [1]. Real time Next-Generation computing has also seen some visibility in current literature [2] .Trilinos as a tool has been recently developed with all its integrated packages and had been used previously [3] to report case study on low mach fluid and solve cascade failure [4] problems. Sufficient emphasis has been put into developing parallel algorithms and software to handle large engineering problems [5] and the necessary architecture required to handle such computing [6]. At the same time MIFX has grown in popularity for providing multiphase solvers specific to certain engineering application especially in fluidized bed problems with inherent concurrency and parallelism build into their toolkit. Previous work have used MFIX [7, 8] to build and solve numerical models of certain multiphase flow problems. This work has tried to use and compare Trilinos large scale linear solvers with MFIX solvers and build a language independent framework [9]

P.214

to integrate both these libraries in order to exploit the massive parallelism and speed-up provided by both. 2

Comparison of Trilinos and MFIX

2.1 Trilinos Trilinos provides an object oriented framework using C++ 11 with a two-level software abstraction [10]. The higher level of abstraction is a common object oriented framework for programming while the second level contains a number of scientific computing packages enabling inherent parallelism and exascale computations. In this work we have explored the capabilities of a few specific packages including Epetra package, that provides a library of constructors for building and operating on large sparse and dense matrices in parallel, The AztecOO library that implements a GMRES[11] iterative solver integrated with IF PACK which provides capabilities for preconditioning. The inherent parallelism provided by the Epetra_Map class has been used to build coefficient matrices corresponding to a billion difference equations. We have tested the Trilinos frame work to solve a transient heat conduction problem with a million elements problem size and have later on compared the solver used for this particular problem with the MFIX solver. Figure 1 shows the speedup for the AztecOO GMRES based solver with a 10 million elements problem size. 2.2 MFIX MFIX is a multiphase solver that has been in the industry for some time now, and has become a standard for implementing multiphase flow problems such as a fluidized bed. MFIX, dedicated to multiphase flow has capabilities to solver large problems though iterative solvers implemented in FORTRAN. At the same time it has provided parallelization through various techniques such as domain-decomposition etc. Figure1 shows a Tri-diagonal matrix with 10 million elements problem size solved on multiple processors with MFIX and Trilinos.

Figure1. Speed up graphs for Trlinos AztecOO and MFIX linear solver on UTEP Research Cloud.

P.215

NEXT GENERATION COMPUTING FRAMEWORK FOR EXASCALE SIMULATIONS

3

Integration of MFIX and Trilinos in a Language independent framework

Given that fact the MFIX code design was in FORTRAN which has no object oriented capabilities while Trilinos was designed in an object oriented C++ 11 framework, their integration has no trivial solution. In order to do that in a language independent framework we would need to communicate via an object that can access both the FORTRAN and the C++11 frameworks[9]. A C object was chosen to access data from the FORTRAN objects and pass them into a C++ function which would in turn communicate with the Trilinos package. It has to be kept in mind that such information passing is not trivial since such information would contain multi-dimensional arrays of millions of elements which would require extensive memory management for successful implementation. However with such integration we can obtain better performance by using MFIX capabilities of setting up such large scale problems with minimum effort and Trilinos capabilities for providing extremely high speed solvers. Figure2 shows a schematic of the software abstraction for this integration.

Figure2. Schematic depicting the software abstraction of the integration 4 Conclusions Such integration as has been developed in this work lays foundations of exascale simulation especially in the areas of multiphase flow. Moreover it provides application writers with a language independent framework to solve large scale multiphase problems with minimal optimization effort on parallelization owing to the fine tuning of these individual libraries.It opens windows for Next-Generation computing frameworks for complex multi-physics problems.

3 P.216

5

References

[1] H.Power et al. “High Performance Computing in Engineering Applications:Applications to Partial Differential Equations” WIT Press 1995. [2] J.A.Stankovic “Real-Time Computing System-The Next Generation” Technical Report 1988 ACM [3] P.Lin et al. “ Towards Extreme Scale Simuation with Next Generation Trilinos:A Low Mach Fluid Application Case Study” . Parallel & Distributed Processing Symposium Workshops (IPDPSW), 2014 IEEE International, Phoenix, Arizona. [4] Christopher Palmer et al. “Developing a dynamic model of cascading failure for high performance computing using trilinos”. Proceedings of the first international workshop on High performance computing, networking and analytics for the power grid, New York, 2011. [5] D. Gaston et al. “Parallel multiphysics algorithms and software for computational nuclear engineering” Journal of Physics :Conference Series, Volume 180, Number 1. [6] David E. Bernholdt et al. “ A Component Architecture for High Performance Scientific Computing” International Journal for High Performance Computing Applications, vol-20,2006. [7] T. Mackeen et al. “ Simulation and experimental validation of a freely bubbling bed of FCC catalyst” Power Technology, Elsevier 2003. [8] N. Xie et al. “ Simulations of multiphase reactive flows in fluidized bed using situ adaptive tabulation” Combustion Theory and Modeling Taylor & Francis-2004. [9] Aytekin et al. “Modernizing A Legacy Open Source CFD Code By Leveraging Scalable Parallel Preconditioners and Linear Equations Solvers” Proceedings of the 27 th International Conference on Parallel Computational Fluid Dynamics 2015, Montreal, Canada. [10] M.Heroux et al. “ An overview of Trilinos” Sandia National Labs 2003. [11] Y. Saad et al “ A Generalized Minimum Residual Algorithm” Siam Journal on Scientific Computing 1986.

P.217

SETS2016-59 The Southwest Emerging Technology Symposium 2016

2-D COMPUTATIONAL MODEL OF A COAXIAL SWIRL FUEL INJECTOR J. Aboud, B. Lovich, O. Vidana, L. Cabrera, M.J. Hernandez, N. Love1 A. Choudhuri1 1 Mechanical Engineering Department, UTEP, El Paso, TX 79968, USA; Keywords: MHD, CFD simulation, Swirl injector, gas-gas ABSTRACT A direct power extraction system is relatively complicated and one of the crucial portions of the system is the fuel injector. Computational analysis was used because building and performing experimental testing on an injector design is time consuming and expensive. Computational analysis can be done relatively easily and provides adequate results for performance of the actual model. The purpose of this paper is to present a 2-D computational model of a coaxial swirl fuel injector that used in a current project of direct power extraction system. For this reason, the details and parameters used for the development of the 2-D computational model used will be discussed in this paper. The fuel injector mentioned is that of the MHD direct power extraction project which uses methane as the fuel. 1. Introduction: Traditional power generation relies on using hydraulic, thermal and nuclear resources. In all these systems the potential, or thermal energy, is first converted into mechanical energy and which is then transferred to electrical energy [1]. Over the years, these traditional power generations have been an incredible source of reliability, and efficiency. The main problem with this progression lies in the fact that conversion processes must always have an associated efficiency. Since this conversion efficiency will never be 100% the implication is that some energy will be lost in the process. One possible solution is Magnetohydrodynamic (MHD) power generation, which works by converting the enthalpy from a high temperature working fluid into electrical power using Faraday’s Law of electromagnetic induction.[1] Due to the fact that MHD generators have no mechanically moving parts, the direct conversion of thermal energy to electrical energy is further simplified.[2] MHD power generation systems’ efficiency is usually described in terms of enthalpy extraction ratio, which is the ratio of power output over thermal input. Extraction ratios can range from 10 – 30%.As recently as 2004, open-cycle generators experimentally demonstrated an 1115% enthalpy extraction ratio based on the type of electrodes the system has. [3] Typically, MHD power generator apparatus design consists of a pressure-fed, fuel and oxidizer delivery system, combustion chamber with a converging-diverging De Laval nozzle structure, a magnet, and electrodes. [4] A critical component that governs flames in a combustor is the injector manifold. The fuel and oxidizer system is ideally designed to insure fully mix products before the combustion reaction takes place. Therefore the fuel and oxidizer system play a major role in the system by driving the flame temperature and exhaust products velocity.

P.218

The Southwest Emerging Technology Symposium 2016

2. Fuel injectors: Since the 1990’s, the gas-gas injector technology has been preferred as a key technology of the Full Flow Stage Combustion (FFSC) cycle engine. In addition, gas-gas injectors started to be widely used in small engines because the gas-gas injection combustion process contains straightforward mixing and reaction processes without atomization and vaporization. [5] Recently the computational model for complex systems has increased, due to the impressive result. The computational model could be done in shorter time and less coast comparing with traditional ways. Furthermore, the computational model usually is a mid-solution between the theoretical and experimental, also it gives an estimate about how the result would look in the experiment. this particular study focuses on investigating coaxial swirl gas-gas injector geometries using computational fluid dynamic model. The fuel injection is delivered through tangential orifices in this design. Fig. 1 shows a representative fuel injector configuration of interest. The manifold responsible for creating a momentum vector that travels in a centripetal direction. This momentum vector is created due to the offset implemented in each injection point. The four tangential ports shown in Fig. 1 represent the fuel injectors. The oxidizer is introduced through the center port and shears the tangentially flowing fuel to create enhanced turbulence mixing behaviors before combustion. 3. Geometry: The requirements of this design are accomplish to be well-distribution of the fuel in the manifold. This includes investigating the velocity distribution among all tangential inlets, uniform swirl velocities, and reduction of any low pressure regions in the combustor. Also, length of the ports was constrained by manufacturing techniques. While optimizing the injection system, many variables in the manifold geometry were manipulated using ANSYS Fluent. The results from the different parameters improved flow in the injector. The dimensions of the final iteration are: The width of the manifold ring is 6.1 mm; the inner diameter is 19 mm while the outer diameter is 25.1. Every orifice is 1.59 mm in diameter with an offset distance of 4.21 mm from the center. The injection plane is introduced at 115-degrees in respect to the x-axis. The primary injection pipe’s diameter is 3.05 mm.

P.219

Fig. 1. Swirl coaxial injector geometry

The Southwest Emerging Technology Symposium 2016

4. CFD simulation The fuel injector system mesh was created using the ANSYS Workbench meshing tool and modified by the ANSYS Fluent adapt region tool. The mesh contained a combination of triangular and quadrilateral elements with inflation layers added on the injector walls as shown in Fig. 2. The mesh originally contained 909 nodes with 788 elements but after the adaption the number increased to 12,470 nodes with 25,255 elements. The mesh has a minimum orthogonal quality of 0.111, maximum orthogonal quality of 0.999, minimum aspect ratio of 1.0035, and maximum aspect ratio of 23.4.

Fig. 2. Swirl coaxial injector mesh

5. Setup and Boundary conditions: Boundary conditions for this model can be seen in Table-1 If not specified, values were left at their default setting. Table 1. Setup and Boundary Conditions Section

Input

General

Time - Transient Gravitational acceleration - 9.81 m/s

Models

Multiphase – Volume of Fluid Viscous – Standard k-epsilon Methane Density – 0.6654 Cp – Piecewise-polynomial

Materials

Primary – Air Secondary – Methane

Phases

Mixture inlet: Velocity - 200 m/s Hydraulic diameter – 0.00305 m Methane inlet: Volume fraction – 1 Mixture outlet: Pressure - 689 KPa Hydraulic diameter – 0.001588 m

Boundary conditions

P.220

The Southwest Emerging Technology Symposium 2016

6. Result: 6.1. Pressure: Figure. 3. Shows the static pressure distribution inside the fuel injector. It can be seen that the pressure is similarly distributed among the four ports for this given dimension. With pressure at the port of approximately 660 KPa and a pressure drop of 50 KPa. The pressure drop is the significant parameter in the fuel injector because it determines the pressure required for the fuel line to achieve the testing Fig. 3. Fuel injector static pressure contours shown in KPa pressure in the combustion chamber.

6.2. Velocity Figure. 4. presents the velocity in the fuel 1 4 injector. The contour shows the distribution of the fuel velocity among the ports. It can be seen that the minimum port velocity of 15 m/s and maximum port velocity of 195 m/s. This diversity in the velocity among the port is due 2 3 to position of the port relative to the inlet. As it is shown the first and second ports have the lowest velocity as they are located near the Fig. 4. Fuel injector velocity contours shown in m/s inlet, and the third and fourth have the highest velocity because they located further downstream. The significant of the velocity is the determination of the mass flow rate in the ports which determines the total fuel mass flow rate in the combustion chamber.

6.3. Volume fraction Figures 5-A, B, C, D, and E show the fuel flow at 0.1, 0.5, 1, 2 and 4.45 ms respectively. The volume of fraction diagrams show how the flow would propagate through the injector manifold, filling each injector from the first and fourth port first, and then the second and finally the third one. The time necessary to fill the system with fuel is determined to be 4.45 ms. The final remnants of the air exits the fuel manifold at 4.5 ms. As it is shown the volume fraction diagram is an important tool as it provided insight into fluid flow at any given time.

P.221

The Southwest Emerging Technology Symposium 2016

Fig. 9-A. Injector volume fraction contour of methane at 0.1 ms

Fig. 9-B. Injector volume fraction contour of methane at 0.5 ms

Fig. 9-C. Injector volume fraction contour of methane at 1 ms

Fig. 9-D. Injector volume fraction contour of methane at 2 ms

Fig. 9-E. Injector volume fraction contour of methane at 4.45 ms

7. Discussion: Originally the L/D was set to 0.75 and was steadily increased by set increments starting from 1. However, it was observed that increasing the length to diameter ratio (L/D) of the ports proved beneficial to the flow characteristics mentioned previously and resulted in better distributed flows within the injectors both with respect to velocity and pressure. On the other hand, it would also increase pressure drop, which leads to higher pressure in the line to achieve the desired chamber pressure. As a result, The L/D is determined to be of 4.18. Also the location of the inlet was determined to be in between the first and the fourth ports because being closer to any port would vastly disrupt velocity and pressure distribution, and showing bias towards the closest port which cause the fuel injector to perform inadequately.

P.222

The Southwest Emerging Technology Symposium 2016

8. Conclusion The purpose of this paper was to present a 2-D computational model of a coaxial swirl fuel injector that is used in a current project of a Direct Power Extraction system. The requirement was to investigate the velocity distribution, pressure drop, and filling time. The geometry was selected after many iterations to look as shown in fig.1. The mesh was generated with ANSYS Workbench meshing tool and improved with adapt region tool; with a minimum orthogonal quality of 0.111, and minimum aspect ratio of 1.0035. For the setup and boundary condition, everything was kept as default except for the sections listed in table 1. In the result section, contours displayed the distribution of the pressure, velocity, and volume fraction respectively in figures 3 through 5-E. The contours indicated the pressure drop of 50 KPa, 15 and 200 m/s as minimum and maximize velocity at the ports, and 4.5 ms was the time needed to fill up the fuel manifold. This 2-D computational model result will be used in the fuel manifold for the MHD direct power extraction project. In addition, this model will serve as a base for the 3-D computational model for the same project, and furthermore a coaxial fuel injector will be investigated using the knowledge gained from this model. 9. Acknowledgments The research is supported by the US Department of Energy, under award DE-FE-0024062 (Project Manager Jason Hissam). However, any opinions, findings, conclusions, or recommendations expressed herein are those of the authors and do not necessarily reflect the views of the Department of Energy. 10. [1] [2] [3]

[4]

[5]

References Ajith Krishnan, R., & Jinshah, B. S. (2013). Magnetohydrodynamic Power Generation. International Journal of Scientific and Research Publications, 3(6), 1-11. Kayukawa, N. (2004). Open-cycle magnetohydrodynamic electrical power generation: a review and future perspectives. Progress in Energy and Combustion Science, 30(1), 33-60. Rosero, J., Enriquez, G., Aboud, J., Lovich, B., Gamboa, C., and Love, N., “Review of MHD Power Generation Systems,” 5th Southwest Energy Science and Engineering Symposium, El Paso, Texas, April 2015. Hernandez, M., Cabrera, L., Vidaña, O., Chaidez, M., Love, N., and Choudhuri, A., “A Systematic Framework for the Design of an Open-Cycle Direct Power Extraction Combustion Chamber,” SciTech 2016, AIAA, San Diego, CA, January 4 – 8, 2016 Xiaowei, W., Guobiao, C., Yushan, G., & Ping, J. (2009). Large flow rate shear-coaxial gasgas injector.

P.223

SETS2016-60 The Southwest Emerging Technology Symposium 2016

EFFECT OF WATER AND HEAT TRANSPORT ON THREEPHASE TRANSIENT BEHAVIOR OF A PEFC 1

A. Nandy1*, C. Y. Wang2 College of Engineering and Technology, Northern New Mexico College, Espanola, NM 87532, USA; 2 Mechanical Engineering, Pennsylvania State University, University Park, PA 16802, USA * Corresponding author ([email protected])

ABSTRACT The dynamic three-phase behavior of a Polymer Electrolyte Fuel Cell (PEFC) immediately after successful startup from subzero temperature is not well understood yet, although it plays the pivotal role on the post-cold start PEFC performance. A transient, one-dimensional, threephase PEFC model is developed to capture complex and coupled water and heat transport mechanisms which govern the dynamic cell performance starting from subzero temperature towards normal operation. Ice formation during cold start operation, ice melting at freezing point and liquid water transport process are captured and are correlated with the cell performance. The effect of load rise after successful cold start, resulting electrode flooding by liquid water and cell performance loss are also elucidated. 1

Mathematical Model

The mathematical model for this study is presented in detail in author’s previous work [1]. This model has been derived and extended based on the previous cold-start models [2], [3], [4]. The anode gas diffusion layer (AGDL), anode catalyst layer (ACL), ionomeric membrane, cathode catalyst layer (CCL), and cathode gas diffusion layer (CGDL) are included in the computational domain. Only for the cell thermal mass calculation the bipolar plates are included. Variations only in the through-plane direction are captured in this 1-D model, and the in-plane and the along-the-channel variations are not accounted for. Spatially uniform cell temperature is assumed and a lumped thermal model is used for the energy conservation. 1.1 Cell Voltage The cell voltage is obtained by subtracting the activation losses and ohmic losses from the equilibrium potential as U cell U 0 IR (1) a c In the above Eq. (1), the expression for the cathode over-potential ( c) is obtained following Tafel kinetics, whereas the anode over-potential is negligibly small due to fast Hydrogen oxidation reaction kinetics. Also, the Ohmic loss is considered only due to protonic resistance in the membrane and Cathode CL.

1.2 Heat Balance- Lumped Thermal Model The total heat generation (J/m2/s) during post-cold start could be calculated as

Qtotal

Qrev

Qirrev

(2)

Q phase

where, Q phase is the heat generation term due to phase transitions, which consists of different source and sink terms depending on the cell temperature (cold-start vs. post-cold-start state) Heat losses to the ambient can be formulated in the same way as is done in the cold start model [4]

P.224

1 (3) mC p out T mC p in T0 h T T0 A where, h and T0 denote the convective heat transfer coefficient and ambient temperature, respectively. The overall heat balance for the cell can be formulated as: Qloss

mC p

cell

T

T0

t 0

(4)

Qtotal Qloss dt

1.3 Water Transport Model Water Content in MEA – In membrane electrode assembly, the generalized water content (λ)-explicit transport equation in 1D form [3] is used and solved in conjunction with appropriate boundary conditions. Water Balance in Cathode CL - CCL flooding and resulting Oxygen transport and cell performance is dictated by a delicate balance in cathode CL. A water balance with respect to the CCL control volume is performed by assuming uniform water content and homogeneous phase transition in CCL and is formulated as n pro

nEOD

nCCL

nBD

nc, outflow nice

nLiqAccu

nLiqTran

(5)

0

n pro (water production) and n EOD (electroosmotic dragged water from ACL) are designated

with a positive sign as they bring water in the cathode CL. nCCL (CCL water uptake),

n BD (back diffusion due to water content gradient), nc , outflow (water vapor outflow with exhaust gas), and n LiqTran (liquid water transport to CGDL) are considered with a negative sign. At subzero temperatures, ice is formed and nice is considered to be negative. When the cell temperature reaches 0oC, ice starts to melt and nice is considered as positive. n LiqAccu is the rate of water accumulation in CCL. If successful self startup is achieved, ice begins to melt at 0oC and liquid water starts to form. Hence, two sub-models, ice melting model and liquid water transport model to CGDL are also established in order to perform the water balance calculation in the PEFC during post-cold start [1]. 2. Physical Parameters and Operating Conditions The physical and geometrical parameters are given in Table 1. A stand-alone FORTRAN code is developed based on the model formulation which is capable of capturing complex physical phenomena occurring in PEFC cold start and subsequent post-cold start process. The GDL is discretized into 30 numerical elements and the membrane is discretized into 10 numerical elements, and are found to be sufficient to resolve the through-plane variations. Only one numerical element is used in the CL as uniform water content in the CL is assumed. Table 1: Summary of Physical Parameters and Operating Conditions Geometric Parameters/ Physical Properties GDL thickness CL thickness Membrane thickness Dry membrane density, ρmem Ionomer fraction in CL Porosity of CL/GDL

300 µm 10 µm 30 µm 1980 kg/m3 0.2 0.5

Operating Conditions Initial membrane water content Initial ice fraction Startup temperature Startup current density Switch to current density at 50C Heat Transfer coefficient between cell and surrounding

P.225

7 0.4 -10oC 100 mA/cm2 500 mA/cm2 40 W/m2K

3. Results and Discussions As shown in Figure 1-a, the cell temperature increases gradually due to heat generation (heat generation and removal rates as a function of time are shown in Figure 1-d) and a successful self startup is achieved. Due to ice melting at constant temperature, a small plateau in the temperature plot at 00C at around 53 s is observed. As the total heat generation rate increases due to load rise (from 100 mA/cm2 to 500 mA/cm2) at 50C, the cell temperature also rises much faster. The variation of ice fraction and liquid saturation in the CCL with respect to time are displayed in Figure 1-b. There is no ice precipitation during the first 29 s of cold start, and the water produced is mainly absorbed by the membrane and CCL to increase their water content. In this stage, the cell voltage experiences an increase due to lowered protonic resistance resulting from the hydration of the membrane (see Figure 1-a). Once the ionomers of CCL get saturated with water, ice starts to precipitate and the maximum ice fraction in CCL reaches a value of 0.47 when the cell temperature reaches freezing point (as shown in Figure 1-b). Liquid water starts to form at the freezing point due to ice melting and is being transported from CCL to CGDL resulting in a liquid saturation value of 0.40 after completion of ice melting (Figure 1-b). Following ice melting, the transport of liquid water causes a reduction in the liquid saturation in CCL, and eventually a relatively steady state value is reached (~0.25). Both Oxygen blockage and ECA reduction effects are reduced due to lowered liquid saturation level and hence the cell voltage experiences a rise till the current density is increased at 50C. The cell temperature continues to increase after completion of ice melting, which also contributes to the increase in cell voltage. Figure 1-c shows different water fluxes in and out of the CCL. In this figure, a positive flux indicates water accumulation and a negative flux represents water removal from the CCL. As shown in Figure 1-c, at the initial phase of cold start, the water removal rate by the outflow gas is negligibly small but it increases significantly at a higher cell temperature as the vapor transport rate increases considerably with cell temperature rise. The CCL ionomers water uptake rate goes to zero in about 29 s, upon water saturation (note at this instant ice starts to precipitate). Before load rise, membrane water uptake increases at a fast rate at the initial stage, and then almost levels off, followed by a decrease in the rate. This decrease in membrane water uptake rate prior to load rise is caused by the decreasing driving force (as the difference in water content value between CCL and Membrane decreases). The product water generation rate and electroosmotic drag (EOD) increases instantaneously with load rise, whereas the back diffusion of water into membrane and ACL lags behind, resulting in a net decrease in the membrane and ACL water content. As the load is increased, it is followed by an immediate supply of water from membrane to CCL and after ~90 s, the membrane starts to take water from CCL again (Figure 1-c). This interesting phenomenon occurs as the EOD and the product water (both proportional to the current density) increases instantaneously upon load rise, whereas the back diffusion of water lags behind and it takes some time to catch up. Hence, the flooding situation in CCL becomes more severe upon load increase, because of the immediate water supply from the membrane. This effect contributes to the sharp voltage drop once the load is increased as shown in Figure 1-a. As the liquid formation rate inside CCL is increased upon load rise (to 500mA/cm2), the liquid water transport rate to CGDL also increases, but it lags behind the liquid formation rate, resulting in an increase in liquid saturation in CCL (form a value of ~0.25 to ~0.71; Figure 1-b). This

3 P.226

causes a sharp increase in Oxygen blockage and ECA reduction which in turn causes a sharp drop in cell voltage by 0.15 V (from 0.73 to 0.58 V; Figure 1-a). The voltage dropdown is followed by a gradual recovery due to movement of liquid water into CGDL and increasing cell temperature.

Ice fraction/ Liquid saturation in CCL

0.8 0.7

Ice fraction Liquid saturation

0.6 0.5 0.4 0.3 0.2 0.1 0.0 0

20

40

60

80

100

120

80

100

120

Time (s)

(a)

(b) 3000

Product water

0.06

Outflow water Back diffusion

0.02

CCL water uptake

0.00 -0.02 -0.04 -0.06 0

20

Reaction Heat

2500

Membrane water uptake

0.04

Heat Generation/Removal Rate (W/m2)

Water Generation/Removal Rate (mol/m2 s)

Dragged water

40

60

80

100

Entropic Heat

2000

Joule Heat Heat Loss- outflow gas

1500

Heat Loss- BP

1000 500 0 0

20

40

60

-500

-1000

120

Time (s)

Time (s)

(d)

(c)

Figure 1. a) Cell temperature and voltage variation with time; b) CCL ice fraction and liquid saturation variation with time; c) Variations of water generation and removal rates with time; d) Variations of heat generation and removal rates with time.

References [1] A. Nandy “Modeling Gas Purge and Three-Phase Transients in a Polymer Electrolyte Fuel Cell”. PhD Dissertation, The Pennsylvania State University, 2012. [2] F. Jiang, W. Fang, and C.Y. Wang “Non-isothermal cold start of polymer electrolyte fuel cell”. Electrochimica Acta. Vol. 53, pp 610-627, 2007. [3] L. Mao and C.Y. Wang “Analysis of cold start of polymer electrolyte fuel cell”. J. Electrochem. Society. Vol. 154, pp B139-B146, 2007. [4] A. Nandy, F. Jiang, S. Ge, C. Y. Wang, K. S. Chen “Effect of Cathode Pore Volume on PEM Fuel Cell Cold Start”. J. Electrochem. Society, Vol. 157 (5), pp. B726-736, 2010.

P.227

SETS2016-61 The Southwest Emerging Technology Symposium 2016

AN ANALYSIS OF A PHOTOVOLTAIC SOLAR SYSTEM PROJECT IN AN INSTITUTION OF HIGHER EDUCATION D. N. De Hoyos1, B. Tseng2*, A.Olivarez3 Environmental Science and Engineering, Doctoral Student 2 Department of Industrial, Manufacturing and System Engineering, 3 Department of Teacher Education, College of Education University of Texas at El Paso, El Paso, TX 79968, USA *Corresponding author ([email protected]) Keywords: Solar, Photovoltaic, AASHE 1

ABSTRACT This paper is a review of the impact of photovoltaic solar panels installations on institutions of higher education throughout the United States. It affords universities to utilize this type of alternative energy source as a resource for college campuses. The use of solar photovoltaic (PV) and other renewable sources to meet rising electricity demand by a growing world population has gained traction in recent years. In order to find alternative sources of energy to aid in the reduction of our nation’s dependency on non-renewable fuels, energy sources include the use of solar energy panels. Photovoltaic (PV) sources were predicted to become one of the biggest contributors to electricity generation among all renewable energy generation candidates by 2040. It includes the impact of sustainable procurement, its challenges and barriers faced during implementation of solar panels. In 2009, President Barack Obama implemented the American Recovery and Reinvestment Act (ARRA) or The Recovery Act which was a stimulus package enacted by the 111th United States Congress. [1] The intent of these initiatives was to provide substantial energy savings and reduce dependence on the electrical grid and net metering savings during the peak energy-use hours. Many campuses are installing photovoltaic (PV) solar electric arrays. While rarely as costeffective as energy conservation, PV becomes more cost-effective when conventional electric rates are high and ample incentives are offered by state government or local utilities. 1

Introduction

Some mature, alternative generation methods are wind, power, photovoltaic panels, biogas and fuel cells. Among them photovoltaic panels are the most accepted and most convenient methods that can be used. This technology of sustainable energy was identified as a form of renewable energy utilizing funding allocated by the United State government. “Energy Efficiency is an example of an energy efficiency project on an institution of higher education. From all the projects undertaken by university members, initiatives related to conserve energy on campus are the most preferred. It is because “campus energy costs typically constitute 30% of a university’s total operations and maintenance budget.” [2] As a result, numerous universities have saved many dollars in energy cost. [3] 2

How a Solar Panel Works

The solar panels are made up of photovoltaic (PV) cells, which convert sunlight into direct current (DC) electricity throughout the day. Solar or photovoltaic (photo = light, voltaic =

P.228

voltage or electricity) cells are created from special materials such as Silicon (Si) mixed with other elements, which when exposed to sunlight will generate an electrical current. Basically sunlight is absorbed into the photovoltaic material, which in turn knocks electrons within the material loose. This allows the electrons to flow freely within the material structure, creating an electrical current. [4]

Fig.1. How a solar panel works [5] One of the reasons, so many campuses have adopted PV installations is due to the practicality, the aesthetics and the functionality. With soaring energy costs, many Universities are facing ever spiraling energy bills when it comes to heating and providing electricity to the many rooms within their buildings. Solar energy is one very affordable option for Universities to consider, paying for itself in a matter of a few years and then continuing to slice the energy bills faced by academic establishments. [6] Nearly all Universities own modern buildings that would be suitable candidates for installing solar paneling. If the building has a good sized area of roof space, it has a wasted potential for using this space to generate solar electricity. Installing solar panels also provides a great resource for all environmental elements of academia. [7] 3

Types of Solar Systems

There are three different types of mounting options for solar systems which are rooftop, ground or car port structures. Several types of the most common – roof top are listed below. Rooftop system which is flush mounted, affixed. Flush Mounted which is affixed Ballasted tilt mounted which is typical a flat roof Affixed tilt – mounted which is typically a flat roof) A typical system is comprised of crystalline panels, is either roof or ground mounted, is grid tied with fixed mounting, net metered and distributed generation. At the institution being discussed, it was both roof and canopy mounted. What can differ in systems is system size, location, building height, flush/tilt, and ballasted/affixed. For a commercial property the cost

P.229

range from $4.25 - $5.25/Watt. 1kW of solar PV produces approximately 1200 kWh per year. 4

American Association of Advancement of Sustainability in Higher Education

The Association for the Advancement of Sustainability in Higher Education maintains a database for institutions of higher education documenting Solar PV projects throughout the country. The data entered includes the following: institution, year installed, panel manufacturer, installed cost (USD) , type of funding (USD), match (Institutional) (USD), estimated annual utility savings (USD), capacity (kilowatts), annual production (kWh) and installation type. The highest number of installations (by State) include: Arizona, California, Massachusetts, Texas, Colorado, Ohio, New Jersey Oregon, New York and Wisconsin. These installations were implemented between 1993 and 2015. The top three campus installations were at Arizona State University, The University of Arizona and Colorado State University. The majority of the installations were roof top mount installation types. [8] 5

Life Cycle Assessment of the Electric Generation by Means of Photovoltaic Panels

Life cycle analysis is an assessment process that analyzes a product/system from cradle to grave, that incorporates the raw material extraction to the disposal of the product/system. Figure 2 demonstrate the life cycle stages of a solar photovoltaic process. The life cycle of a solar photovoltaic is divided into three different stages. The life cycle analysis is conducted based on these three stages [9]. The first stage combines the pre-power generation processes. These includes raw material extraction, material production, module manufacture, system installation, and plant construction. The second stage discusses about the system operational stage. The third stage includes that decommissioning and the disposal of the system.

Fig.2. Life Cycle Stages of Solar Photovoltaic Power Generation Process [10] 6

Photovoltaic Cost

The United States still has the highest prices among major PV markets. U.S. installed PV prices are more than double German prices for systems under 100 kW, but also much higher than prices in the UK, Italy and France. U.S. installed prices are even higher than Japanese installed PV system prices for the 10-100 kW range, even though Japan has by far the highest PV module prices of any major market. Prices for PV plants above 5 MW completed in 2013 remained steady at $3.00 per watt. Installed PV system costs varied widely from state to state, depending on segment. [11]

P.230

7

Benefits of Using PV Solar Panels

Benefits for Solar Generating System include: low maintenance requiring occasional cleaning, remote performance monitoring, annual checks for the electrical wiring inverter and panels). The system has a 30 + year life. Solar also utilizes underused property (rooftops, parking lots, etc.) to provide onsite electricity. Solar distributed generation provides localized electricity – no extra burden on transmission infrastructure and it is an extremely important component to help us meet increasing global electricity needs. [12] The power source of the sun is absolutely free and produces no pollution, and it is infinite in nature. 8

Summary

Many campuses are installing photovoltaic (PV) solar electric arrays. While rarely as costeffective as energy conservation, PV becomes more cost-effective when conventional electric rates are high and ample incentives are offered by state government or local utilities. The amount of available sunlight is another important factor. [13] According to a Penn State study focusing on silicon cells a sustainable energy system must fulfill three absolute criteria: It must produce more energy over its lifetime than what is used to produce the system, it must not deplete a natural resource over time and it must not create by-products having a negative impact on society or the environment. [13] Solar panels are inherently sustainable when compared to other technologies due to their low environmental concerns, energy efficiencies and low maintenance costs. To determine their full sustainability, the risks in life cycle analysis must be compared to conventional systems risks; this is a possible point of future research. References [1] Steinbrook, R. (2009). Health care and the American recovery and reinvestment act. New England Journal of Medicine, 360(11), 1057-1060. [2] Ayres, R.U.et al, 1998 [3] Prugh, T., Costanza, R., & Daly, H. E. (2000). The local politics of global sustainability. Island Press. [4] http://redare.com.au/solar/about/solarpanels/ [5] http://www.redarc.com.au/solar/about/solarpanels/ [6] http://www.theecoexperts.co.uk/solar-panels-university-buildings [7] http://www.theecoexperts.co.uk/solar-panels-university-buildings [8] http://www.aashe.org/wiki/cool-campus-how-guide-college-and-university-climate-action-planning/53install-renewable-energ [9] NREL, Oregon.gov [10] http://www.pv-magazine.com/news/detail/beitrag/us-installed-solar-pv-costs-continue-to-fall_100016490/ [11] http://www.slideshare.net/iandiamond/renergyco-solar-pv-for-commercial-properties [12] http://www.aashe.org/wiki/cool-campus-how-guide-college-and-university-climate-action-planning/53install-renewable-energ [13] Pearce, J., & Lau, A. (2002, January). Net energy analysis for sustainable energy production from silicon based solar cells. In ASME Solar 2002: International Solar Energy Conference (pp. 181-186). American Society of Mechanical Engineers.

P.231

SETS2016-62 The Southwest Emerging Technology Symposium 2016

Hall Effect Measurement Data Valerie Vidal, University of Texas at El Paso

I. Introduction The Hall Effect Measurement system is used to measure the details of conduction in semiconductors such as carrier density, resistivity, conductivity, mobility, and bulk concentration. It’s widely used in the electronics industry in materials research and device manufacturing. By using a combination of a magnetic field and current through a sample, this creates electrical current that becomes transverse voltage. A thin Hall probe is placed in the magnetic field in order to measure the transverse voltage. The Lorentz Effect force is the Hall effect. This concept is when an electron moves across an electric field perpendicular to a magnetic field which creates a magnetic force. This force pushes carriers into circular paths in the magnetic field. When the force from the Hall effect balances with the Lorentz force, the transverse voltage is proportional to the magnetic field and current.

II. Approach The goal was to measure the sample’s thickness of the first layer. Comparing conductivity and resistivity would tell us how the thickness affects these factors and how the current is flowing through the sample. Applying silver paste by hand and spin coding were the approach to place the layer on the sample. Silver (Ag) paste was placed on a glass sample and thinned out. Three specific methods were used when applying the substance: Silver paste without shaking the substance, silver paste shaking the substance, and spin coding. When the silver paste (Ag_paste_1) was applied without shaking the solution, the layer was 1288-1595nm thick. When the paste (Ag_paste_2) was shaken, the thickness was 4637-5373nm. Two other samples were spin coded, (Ag_sc) and (Ag_witness) measured at 4890-5607nm and 60 nm, respectively. III. Data Before applying, the paste bottle was only shaken once. Once applied, it was thinned out with a piece of glass until the liquid solution was completely flat on the sample. This sample (Ag_paste_1) appeared very thin compared to the others. The results were not expected due to the fact that conductivity is low and resistivity is high. Even though the thickness was at a thinner range than the others, the data did not prove the silver paste is conductive in this sample.

P.232

(Ag_paste_1)

Bulk Concentration Mobility Resistivity A-C Cross Hall Magneto-Resistance

-3.65E+11 1.85E+01 9.23E+05 -1.36E+07 3.78E+09

Cm^3 Cm^2 / Vs ohms Cm m^2 / C ohms

Sheet Concentration Conductivity Average Hall Coefficient B-D Cross Hall Coefficient Ratio of Vertical/Horizontal

-5.83E+07 1.08E-06 -1.71E+07 -2.06E+07 6.63E-01

/Cm^2 1/ ohms Cm m^2 / C m^2 / C

In the second sample (Ag_paste_2), the paste bottle was shaken multiple times (5-7 shakes). When applied, the paste was thicker than the first sample. This sample was also thinned out with a piece of glass. The results proved that the silver paste is conductive due to the conductivity being high and resistivity is low. (Ag_paste_2)

Bulk Concentration Mobility Resistivity A-C Cross Hall Magneto-Resistance

-4.34E+22 2.36E+00 6.09E-05 -1.31E-03 8.39E-04

Cm^3 Cm^2 / Vs ohms Cm m^2 / C ohms

Sheet Concentration Conductivity Average Hall Coefficient B-D Cross Hall Coefficient Ratio of Vertical/Horizontal

-2.17E+19 1.64E+04 -1.44E-04 1.02E-03 3.63E-01

/Cm^2 1/ ohms Cm m^2 / C m^2 / C

The silver paste on (Ag_sc) was spin coded on this sample. The results show mobility is greater than the (Ag_paste_2) sample but the conductivity is lower. (Ag_sc)

Bulk Concentration Mobility Resistivity A-C Cross Hall Magneto-Resistance

-4.88E+21 1.07E+01 1.20E-04 -8.76E-05 1.48E-03

Cm^3 Cm^2 / Vs ohms Cm m^2 / C ohms

Sheet Concentration Conductivity Average Hall Coefficient B-D Cross Hall Coefficient Ratio of Vertical/Horizontal

-2.54E+18 8.37E+03 -1.28E-03 -2.47E-03 5.80E-01

/Cm^2 1/ ohms Cm m^2 / C m^2 / C

Silver layer (Ag_witness) was also spin coded. This sample had the best results. Conductivity was high, resistivity was low, it had mobility, and the bulk concentration was at E+22. (Ag_witness)

Bulk Concentration Mobility Resistivity A-C Cross Hall Magneto-Resistance

-4.76E+22 1.48E+01 8.82E-06 -1.58E-04 1.70E-03

Cm^3 Cm^2 / Vs ohms Cm m^2 / C ohms

Sheet Concentration Conductivity Average Hall Coefficient B-D Cross Hall Coefficient Ratio of Vertical/Horizontal

P.233

-2.86E+17 1.13E+05 -1.31E-04 -1.05E-04 3.75E-02

/Cm^2 1/ ohms Cm m^2 / C m^2 / C

IV. Conclusion The data collected concluded that silver paste (Ag) has conductivity along with low resistivity. Even though (Ag_paste_1) gave opposite results, the other three samples proved otherwise. (Ag_paste_1) could have been affected by multiple factors due to the results stating the conductivity was low. The best method according to the data is to apply the silver paste by spin coding. The conductivity was high, resistivity was low, and the bulk concentration was at a E+21-E+22. References Dr. Jeff Lindemuth. Hall Mobility Measurement Of Solar Cell Material. Magnetics Technology International. 2012; 36-39.

P.234

SETS2016-63 The Southwest Emerging Technology Symposium 2016

HOUGH TRANSFORM BASED AUTOMATIC SEGMENTATION OF NANOFIBERS FROM SEM IMAGES Z. Hu1, B. Tseng2, Y. Lin3, J. Wu2*, 1 Department of Electrical and Computer Engineering, University of Texas at El Paso, El Paso, TX 79968, USA 2 Department of Industrial, Manufacturing and System Engineering, University of Texas at El Paso, El Paso, TX 79968, USA 3

Department of Mechanical Engineering, University of Texas at El Paso, El Paso, TX 79968 *Corresponding author ([email protected])

Keywords: Hough transform, Nanofiber, Nanocomposites, SEM image

ABSTRACT In this paper, a partitioning Hough transform method is proposed to automatically extract nanofibers from SEM images for nanocomposites morphology analysis. This method first divides the image into partitions with connected fibers and then apply the Hough transform to the skeleton of each partition to avoid the influence of other partitions. Compared with the standard Hough transform, this method is much more robust and has higher segmentation accuracy.

1 Introduction Nanofibers have been widely used as reinforcement to form various nanocomposites with significantly enhanced mechanical and other functional properties [1, 2]. The morphology (e.g., spatial distribution, fiber orientation) of nanofibers in the base material plays a critical role in determining the enhancement of material properties. In structural applications, uniform distribution of nanofibers in terms of both spatial location and orientation is desirable to achieve the best mechanical properties. However, in some other applications, alignment of nanofibers with the same orientation is more preferable. For example, in the dielectric nanocomposites manufacturing used in high performance capacitors, well-aligned nanofibers could increase both permittivity and breakdown strength, and thus increase the energy density of capacitors [3]. Currently, the standard quality inspection technique is morphology analysis of nanofibers imbeded in the base material based on microscopic images, e.g., scanning electron microscope (SEM) images. The morphology analysis is often based on visual inspection of SEM images, which is subjective and lack of quantitative measurement of distribution quality of nanofibers. This paper develops a novel image segmentation method, i.e., partitioning Hough transform method, to automatically identify both the location and orientation of nanofibers for further morphology

P.235

The Southwest Emerging Technology Symposium 2016 analysis. The rest of the paper is organized as follows. Section 2 will present the methodology. The case study and results will be given in Section 3.

2 Methodology 2.1 Introduction of Hough Transform Hough transform was invented by Richard Duda and Peter Hart in 1972, named as “generalized Hough transform” since the 1962 patent of Paul Hough. It is an important segmentation method to detect edge or a certain class of shapes by voting procedure. The classic Hough transform was used to detect straight lines in the image, and then it extend to arbitrary shapes like circle and ellipse. The basic idea of Hough transform is to transfer a point from the physical domain to a curve in its parameters’ domain. A line in x-y plane can be uniquely defined by its distance 𝑟 from the origin and the orientation angle 𝜃 as 𝑐𝑜𝑠 𝜃 + 𝑦 𝑠𝑖𝑛 𝜃 = 𝑟 , as shown in Fig. 1 Therefore, this parameterization maps every line in x-y plane to a point (𝜃, 𝑟) in 𝜃 − 𝑟 domain. For a point 𝑃0 (𝑥0 , 𝑦0 ), all lines passing through it could be expressed as 𝑥0 𝑐𝑜𝑠 𝜃 + 𝑦0 𝑠𝑖𝑛 𝜃 = 𝑟 where 𝑟 is the distance from the origin to the line, and 𝜃 is the angle. Therefore, the point 𝑃0 is mapped from (𝑥0 , 𝑦0 ) to a curve in the 𝜃 − 𝑟 domain. Based on this transform, for points lying on the same line in x-y plane, their corresponding sinusoidal curves in 𝜃 − 𝑟 plane will pass through a common point, which is the line in the 𝜃 − 𝑟 domain.

r 𝜃

x

O

Fig.1. Illustration of Hough transformat

In the Hough transform algorithm, the 𝜃-𝑟 parameter space is divided into accumulator cells to form a two-dimensional matrix. The value of the cell is the number of times a mapping line of points in x-y space intersects that cell. Cells receiving a minimum number of “votes” are assumed to correspond to lines in x-y space. 2.2 Segmentation using Partitioning Hough transform To extract nanofibers, we could use Hough transform to detect lines in SEM images. However, due to the large width, a nanofiber could result in many lines in detection. To overcome this issue, “Skeleton” operation could applied to get the morphological skeleton and then apply Hough transform. However, there is still an issue in the above method. The skeleton of other nanofibers could contribute to the accumulator cell values and may significantly influence the detection accuracy, especially when the density of nanofibers is very large. To overcome this problem, we propose the partitioning Hough transform, where the partitioning is first applied to segment SEM

P.236

The Southwest Emerging Technology Symposium 2016 images into multiple images based on connected components and then apply Hough transform to each segmented image. The Partitioning Hough transform is illustrated in Table 1 as follows:

Table 1. Partitioning Hough transform algorithm for nanofiber segmentation. Convert the SEM image to a binary image Partition the binary image into 𝑛 SEM images with each having one connected component For i=1:n Extract the morphological “skeleton” of image i Perform Hough transform on image i to get the Hough matrix Identify the peaks from the Hough matrix Detect straight lines based on the peaks and Hough matrix End

2.3 Image Simulation Simulated SEM images of nanofiber reinforced nanocomposites are used to evaluate the proposed method. The image size is 2500 × 1900. The orientation of the nanofibers follows uniform distribution. Also, the locations of nanofibers follow complete spatial randomness. The length and width of each fiber follow normal distribution with mean of 4 and 120 respectively. Fig.2 shows a simulated SEM image.

Fig.2. Illustration of simulated SEM image

3 Results and Conclusion Fig.3 shows the results of Hough transform without partitioning and the partitioning Hough method. The straight lines indicate the nanofibers. There are in total 200 nanofibers in the simulated image. For the Hough transform without partitioning, 184 straight fibers are correctly detected (marked with red and yellow “x” on both end), which indicates an accuracy of 92%. For the partitioning Hough transform method, the accuracy is increased to 97% with in total 194 straight lines have been correctly detected. Note that the partitioning step would result in more significant improvement with larger density of nanofibers.

P.237

The Southwest Emerging Technology Symposium 2016 In conclusion, the developed method based on Hough transform is very effective and efficient in the nanofiber segmentation. With this method, nanofibers morphology analysis can be automatically implemented. The improvement and application of this method to real SEM images will be left to our future work.

(a)

(b)

Fig.3. Segmentation results for (a) Hough transform without partitioning and (b) partitioning Hough transform method.

4 Reference [1] [2] [3]

C. Bakis, L. C. Bank, V. Brown, E. Cosenza, J. Davalos, J. Lesko, A. Machida, S. Rizkalla, and T. Triantafillou, "Fiber-reinforced polymer composites for construction-state-of-the-art review," Journal of Composites for Construction, vol. 6, pp. 73-87, 2002. Z.-M. Huang, Y.-Z. Zhang, M. Kotaki, and S. Ramakrishna, "A review on polymer nanofibers by electrospinning and their applications in nanocomposites," Composites science and technology, vol. 63, pp. 2223-2253, 2003. H. Tang, Y. Lin, and H. A. Sodano, "Enhanced energy storage in nanocomposite capacitors through aligned PZT nanowires by uniaxial strain assembly," Advanced Energy Materials, vol. 2, pp. 469476, 2012.

P.238

SETS2016-64 The Southwest Emerging Technology Symposium 2016

JANUS: LOX/LCH4 ROBOTIC LANDER TESTBED 1

I. Lopez1, L. Bugarin1, R. Ponce1, Ahsan Choudhuri1* Mechanical Engineering, University of Texas at El Paso, El Paso, TX 79968, USA * Director of Center for Space Exploration Technology Research ([email protected])

Keywords: Lander, Vehicle, Methane, Propulsion, Janus

ABSTRACT This paper discusses the initial development & requirement definition for Janus, a robotic lander that will operate of a liquid oxygen (LOX) and liquid methane (LCH4) integrated propulsion system. The focus of the project includes the definition of the project, an overview of the vehicle’s derived requirements (including the lander envelope, mission flight profile, and expected propellant requirements and deliverables), as well as an overview of the current developmental approach with the first of three prototype systems. 1 Project Introduction LOX/LCH4 propellant technologies have been of recent interest for several advantages they present over technologies that employ more conventional propellant combinations. These advantages include easier handling and storage versus than other cryogenics (e.g. liquid hydrogen), and it is non-toxic (unlike hydrazine) Furthermore, methane can be potentially obtained on the surface of other planets (e.g. Mars), which is known as in-situ resource utilization (ISRU). ISRU would effectively reduce the propellant mass of launch vehicles allowing for greater payloads on deep space missions [1]. Moreover, with the use of a solid oxide fuel cell (SOFC), the LOX/LCH4 combination could allow for an integrated system that provides all propulsion and power requirements for a vehicle, essentially reducing vehicle complexity by decreasing the propellant requirement to only two constituents. JANUS is a robotic lander vehicle that serves as a methane propulsion technology testbed. The goal of the project is to develope a flight-capable vehicle that incorporates various methane technologies into a fully operational autonomous (i.e. robotic) system. These technologies include the use of a LOX/LCH4 propulsion system and a methane SOFC. Moreover, the project will potentially demonstrate additive manufacturing for vehicle component design and implementation, as well as composite tank and vehicle structures. The vehicle flight mission comprises a non-tethered flight demonstration that successfully conducts a takeoff, midair hover, and soft landing under full autonomous system control. The vehicle is expected to remain in low altitude when hovering, and the propulsion system use a P.239

The Southwest Emerging Technology Symposium 2016 gimbaled & throttleable 500 – 2000 lbf LOX/LCH4 engine (i.e. cSETR CROME-X), as well as 4 (or more) 5 lbf reaction control system (RCS) engines (i.e. cSETR Pencil Thruster). 2 Project Requirements 2.1 Vehicle Description & Dimensions The vehicle size and geometry was first envisioned from a sized down version of NASA’s Project Morpheus vehicle, a LOX/LCH4 planetary lander [2]. Consequently, Janus’ dimensions are based on a scale factor around half from the Morpheus vehicles size. The overall geometry, size, and dimensions are summarized in Fig. 1 and Table 1. The vehicle will comprise of two integrated systems: the propulsion system and power system. The propulsion system consists of several subcomponents, including the main engine, igniter system, RCS, tanks, & propellant manifold. Power integration will include the use of an SOFC to prove the feasibility of methane integration on both systems. Initial vehicle prototypes will likely employ a battery power supply to simplify the system.

Parameter

Description

Height

13 ft

Width

5.5 ft

Tank Diameter

4 ft

Vehicle Dry Mass

900 lbm

Propellant Mass

200 – 300 lbm

Table 1. Janus size and dimensions

Fig. 1. Janus concept shape

The vehicle is a two tank vertical configuration. This option was selected to minimize the number of tanks for operation, given that Janus runs on a bipropellant system (to have a mass-balanced vehicle another alternative would be a four tank horizontal layout). Moreover, stacking components in vertical form gives greater control authority to the vehicle gimbal, which is the vehicle pitch and yaw control mechanism. As a result, the gimbaled main engine will provide dynamic control for vehicle navigation and stability. The RCS will be provide roll control while in flight. The RCS was limited to roll because a reaction control engine (RCE) thrust is much smaller than the main engine (5 lbf vs. 2000 lbf), so a gimbal is required for main vehicle control and translation. The pressurization method for the engines is a blowdown system. In a blowdown system, the tanks are pressurized before usage and the pressure decays as the propellant is used. This has the advantage of simplicity and fewer component required [3][4]. Although the thrust and propellant flow rate is drops with time during vehicle flight, the tanks are sized to allow for acceptable P.240

The Southwest Emerging Technology Symposium 2016 pressure loss during the flight mission. (≈ 30 seconds), giving a blowdown ratio (ratio of initial pressure to final pressure) close to unity. Consequently, the tank size and initial ullage volume mitigate thrust decay during vehicle flight. 2.2 Flight Profile To demonstrate full use of the propulsion system, a flight profile was defined so as to employ all engines at their thrust capabilities whilst maintaining dynamic control of the lander. Consequently, the flight of the vehicle was chosen to be takeoff-hover-landing maneuver, as shown in Fig. 2. The vehicle is not initially expected to perform sideways maneuvers. The initial rise maneuver is expected to last around 10 seconds and take the lander around 20 feet high. Next, it will hover in place and roll 180° for another 10 seconds. Finally, the vehicle descend for a soft landing in another 10 second maneuver, ending a projected 30 second flight. Albeit simple, the flight profile specified will allow operation of the main engine over a range of thrust levels, employ the RCS, and show dynamic control and stability of the lander.

Fig. 2. Janus flight profile description

2.3 Operating Conditions Having defined the size of the vehicle and flight requirements, it became possible to delineate the minimum engine performance requirements, mainly the specific impulse (Isp) and the operating thrust range. As previously mentioned, the main engine to be used is CROME-X, a 2000 lbf engine with a 4:1 thrust ratio, and the Pencil thruster, a 5 lbf RCE, both currently under development at UTEP cSETR. For operation the engines carry a number of fluid condition requirements that have to be delivered by the vehicle. This allowed a preliminary definition of lander tank and manifold requirements. All these are summarized in Table 2.

P.241

The Southwest Emerging Technology Symposium 2016 Parameter Operating Thrust Main Engine Min Isp Total thrust RCS # of thrusters LOX Temperature cSETR Engine requirements from LCH4 Temperature lander Delivery Pressure

Description 900 – 1200 lbf 230 s 10 lbf 4 (minimum) -300°F -260°F ≈ 300 psig

Table 2. Engine requirements and lander deliverables

3 Prototype Approach 3.1 Nature and Purpose of Prototypes In order to asses Janus’ system & component viability and simplify the design process the project will be broken into three major prototypes. The purpose and advantages of this approach are: Purpose Develop multiple vehicles that incorporate gradual improvements to vehicle design with each vehicle iteration Facilitate the design process and focus the short-term goals for vehicle development Advantages Allows evolution and progressive maturity of vehicle design and test hardware Enables formalization of development and testing strategies for future projects Multiple independent vehicles prevent component/system failure setbacks Multitude prototypes and objectives permits easy project transition to new generation of students 3.2 Prototypes Development & Description Table 3 shows the description of each prototype that will be developed under this project. Prototype

J-1

Characteristics Static test bed

J-2

Flight test bed Nonautonomous (Tethered)

J-3

Flight test bed Autonomous (Non-tethered)

Purpose Test propulsion system Demonstrate structural integrity of thrust mount & gimbal Develop flight hardware Create flight envelope Demonstrate vehicle dynamic integrity Show telemetry capability Implement A.M. Components Possibly test composite tank Demonstrate autonomous GNC integration Employ SOFC power supply

Testing Capabilities Propulsion System Integration Data Acquisition Actuation (Gimbal and Valve) Mission duty cycle Develop procedures & protocols Same as J-1 Landing gear reliability Telemetry/Communications Limited GNC Flight Monitoring Instruments Same as J-2 SOFC Power (Non-flight first) Autonomous GNC control

Table 3. Description of purpose and testing capabilities of each vehicle prototype P.242

The Southwest Emerging Technology Symposium 2016 3.3 J-1 Prototype The current work involves the development of the first prototype (J-1), which will be a static test bed of the propulsion system. Having a non-flight test bed will allow simpler validation of fundamental equipment and operations of the lander which will help improve on the flight test bed. Presently a list of equipment and schematics are being created to develop and construct J-1, which will allow testing capabilities as described in Table 3. These include piping & power schematics, a master equipment list, an interface control document, ground support equipment, etc. Fig. 3 shows a P&ID schematic of J-1.

Fig. 3. P&ID of J-1

4

References

[1] “NASA's Journey to Mars”. NASA. Retrieved January 19, 2015, from https://morpheuslander.jsc.nasa.gov/about/ [2] “Project Morpheus”. NASA. Retrieved February 8, 2015, from http://www.nasa.gov/content/nasas-journey-tomars/#.VQDUFvnF_3Q [3] C. D. Brown, “Spacecraft Propulsion”, 3rd edition, American Institute of Aeronautics and Astronautics, 1996. [4] D. K. Huzel, D. H. Huang, “Modern engineering for design of liquid-propellant rocket engines”, Rev Sub ed., American Institute of Aeronautics and Astronautics, 1992.

P.243

SETS2016-65 The Southwest Emerging Technology Symposium 2016

OPTIMIZATION OF ALUMINUM-DOPED ZINC OXIDE THIN FILMS VIA VARIANCES IN ANNEALING TEMPERATURE V. Castaneda1, S. Shahriar1, C. Sana1, M. Martinez1, D. Kava1, J.Galindo1, and D.Hodges1* 1 Department of Electrical and Computer Engineering, The University of Texas at El Paso, El Paso, TX 79968, USA; * Corresponding author ([email protected])

Keywords: thin films, annealing conditions ABSTRACT During the formation of thin film solar cells many layers are combined to make an efficient cell. The optimization conditions of the front contact layer is investigated by the effect of annealing temperature on sputtered aluminum-doped zinc oxide (AZO) thin films. The AZO thin films were prepared by sputtering on soda-lime glass in a vacuum environment. Samples were created; and were annealed in air for 30 minutes at various temperatures (100 °C, 200 °C, 250 °C, 300 °C, 350°C, and 400°C) and a control sample that was not annealed. Data was then collected using various methods that include UV/Vis/NIR spectroscopy, X-ray diffraction, and Hall Effect measurements. Using the optical, structural, and electrical properties of AZO thin films collected, the ideal annealing temperature for an AZO front contact can be determined.

Fig. 1 XRD patterns of AZO films of 100nm thickness annealed under at different temperatures. Table 1- Parameters of the AZO thin films Annealing temperature Room temp. 100⁰C 200⁰C 250⁰C 300⁰C

FWHM of (002) peak 1.460⁰

Crystallite Size (Å)

2θ (deg)

d spacing (Å)

55.60

34.89

2.56938

0.556⁰ 0.399⁰ 0.867⁰ 0.418⁰

146.3 204.0 93.80 194.6

35.09 34.49 34.60 34.95

2.55520 2.59766 2.59035 2.56510

P.244

350⁰C 400⁰C

0.531⁰ 0.607⁰

153.3 134.1

34.90 34.65

2.56868 2.58671

FWHM-full width at half maximum; 2θ- diffraction angle

Fig. 2 Crystallite size and FWHM of AZO thin films annealed at temperatures from 100⁰C to 400⁰C

Fig. 3 AZO films transmittance; y-axis: transmittance %; x-axis wavelength from 350 nm to 800nm.

1

Introduction

Due to their good adhesion to substrates, high conductivity, and high transmittance values, transparent conductive oxide (TCO) films have been widely used recently in applications such as light emitting diodes, flat-panel displays and thin film solar cells. There are various kinds of TCOs such as fluorine-doped tin oxide (FTO), aluminum-doped zinc oxide, and tindoped indium oxide, with the latter being the most commonly used [1]. The disadvantage of ITO is that it is becoming scarce due to its high cost, and toxicity [2]. As an alternative to ITO, AZO has been considered because of its similarities to ITO and the advantages it presents over ITO. These advantages include thermal stability, low cost, and because it is a natural resource it is highly abundant [3].The similarities include properties of good electrical conductivity, and high transmittance. TCOs are generally used as a front contact in thin film solar cells. Research has shown that in order to improve the overall performance of thin film solar cells it is important to optimize the characteristics of the front transparent conductive

P.245

OPTIMIZATION OF ALUMINUM-DOPED ZINC OXIDE THIN FILMS VIA VARIANCES IN ANNEALING TEMPERATURE

oxide [4]. For this reason, this paper focuses on the optimization of AZO thin films that will ultimately be used as a front contact in our thin film solar cells. To optimize our AZO thin films, our study concentrated on understanding the influence of annealing temperature on the optical and electrical properties of AZO. In this study, the AZO samples were deposited onto soda-lime glass substrates using RF sputtering. To compare the difference in the thin film parameters when annealed, one sample was not annealed while the other AZO thin films were annealed in air (Corning PC420D hot plate). The annealed samples were annealed using temperatures ranging from 100°C-400°C. Data was collected using various characterization methods such as X-ray diffraction to evaluate crystallite structure, UV/Vis/NIR spectroscopy to measure transmittance and Hall Effect measurements. 2

Experimental Details

Before deposition the AZO single layer films were cleaned using methanol, acetone and rinsed in deionized water to remove contaminants. The deposition of AZO layers was done on soda-lime glass substrates by RF sputtering. Sputtering was done using a Kurt J.Leske PVD75 system with the power kept at 125W for 45 minutes. Afterwards, AZO films were annealed for 30 minutes in air at 100°C, 200 °C, 250°C, 300 °C, 350°C, and 400°C, respectively. The resulting thickness of the thin films was about 100 nm. Resistivity and conductivity were measured using Hall Effect measurements (Ecopia-HMS3000). The crystallite size and structure were characterized using X-Ray diffraction (Bruker XRD). Also optimal transmittance was measured using a spectrophotometer Varian Cary 5000 175-3300 nm. 3 Results and Discussion To investigate the effects of annealing temperature on the structural properties of AZO films, the samples were annealed for varied temperature in a vacuum atmosphere. The AZO films demonstrated a typical c-axis preferred orientation. The peaks obtained from the samples were consistent with zinc-oxide peaks according to JCPDS 00-036-14513. As can be seen in Fig.1, they all had their highest peak close to the (002) diffraction peak, which is at 34º. For the sample that was annealed at room temperature the angle of diffraction was at 34.89º, at 100ºC annealing temperature it was at 35.09 º and at 200 º C the angle of diffraction was 34.49º. As temperature increased the angle began increasing to 34.60 º at 250 ºC, and at 300ºC the angle became 34.95 º. After 300 ºC it began decreasing to 34.9 º at 350 ºC and 34.65 º at 400ºC. From Table 1 it can be seen that the smallest value of FWHM (0.399 º) was at 200 ºC. As expected the largest grain size was at 200 ºC, since the FWHM is inversely proportional to the grain size. Fig. 2 shows the crystallite size increased with temperature from 100⁰C to 200ºC and decreased from 300ºC to 400ºC. Also the FWHM decreased with temperature from 100ºC to 200ºC and increased from 300ºC to 400ºC. The electrical properties of the samples were investigated using Hall measurements. From the Hall measurements we found that the resistivity of most of the samples fell around the range of 10-1 Ωcm to 10-3 Ωcm. The resistivity decreased from 4.48 x 10-2 Ω cm to 7.8 x 10-3 Ω cm when going from a room temperature annealed sample to a sample annealed at 100 ºC. The conductivity also increased from 1.027 x 101/Ω-cm to 1.477x101/Ω-cm, which is

3 P.246

consistent with the suggestion that annealing improves the electrical properties of the AZO thin films [7]. In Fig. 3 the optimal transmittance spectra of the AZO films annealed under different temperatures, from the wavelength range of 350 nm to 800 nm, is shown. The average transmittance of the AZO films, at the range of 400 nm to 800 nm, was above 80% for all samples except the one annealed at 400ºC. From the transmittance graphs we can see that the majority of the samples fell within the general requirement of thin films to have an average optical transmittance above 80% in the visible wavelength range [7].For the sample annealed at 200⁰C, the highest transmittance (96.66%) was obtained. We saw the lowest transmittance from the AZO film that was not annealed (95.99%). Although we observed different transmittance values for each sample, the transmittance difference was minimal. This indicates that annealing temperature does not have a significant effect on AZO thin film transmittance. 4 Conclusion AZO single layer thin films were sputtered on a soda lime glass substrate using RF sputtering in order to study the effects of annealing temperature. The samples were annealed at temperatures ranging from 100⁰C- 400⁰C. The samples demonstrated a strong c-axis orientation with the strongest peak at (002). Overall the sample annealed at 200⁰C demonstrated the best structural and optical characteristics, with the highest transmittance; as well as the largest grain size. The best electrical properties were achieved from the sample annealed at 100°C. All the samples obtained an average transmittance above 80%, showing that the AZO thin films prepared in our study are suitable to be used as transparent conductive oxides for thin film applications. From our investigation, the optimal annealing temperature parameter appears to be low temperatures from 100°C to 200⁰C. 3. References [1] P. Prepelita, V. Craciun, F. Garoi, A. Staicu, “Effect of annealing treatment on the structural and optical properties of AZO samples”. Applied Surface Science, Vol. 352, pp 23-27, 2015. [2] J.H. Park, J.M. Shin, S.Y. Cha, J.W. Park, S.Y. Jeong, H. K. Pak and C.R. Cho. “Deposition-temperature effects on AZO thin films prepared by rf magnetron sputtering and their physical properties”. Journal of Korean Physical Society, Vol. 49, pp 584-588, 2006. [3] T. Minami, S. Suzuki, T. Miyata, “Transparent conducting impurity-co-doped ZnO:Al thin films prepared by magnetron sputtering”, Thin Solid Films, Vol. 398–399, pp 53-58, 2001. [4] V. Tvarozek, P. Sutta, I. Novotny, Ballo et al., “Preparation of transparent conductive AZO thin films for solar cells". Advanced Semiconductor Devices and Microsystems. ASDAM 2008. International Conference , Vol., No., pp.275-278, 2008.

P.247

OPTIMIZATION OF ALUMINUM-DOPED ZINC OXIDE THIN FILMS VIA VARIANCES IN ANNEALING TEMPERATURE

[5] C. Besleaga, L. Ion, S. Antohe, “AZO thin films synthesized by rf-Magnetron Sputtering: the role of deposition power”, Romanian Reports in Physics, Vol. 66, No.4,pp 993-1001, 2014. [6] Hong-lie SHEN, Hui ZHANG, Lin-feng LU, Feng JIANG, Chao YANG, “Preparation and properties of AZO thin films on different substrates”, Progress in Natural Science: Materials International, Vol. 20, pp 44-48,2010. [7] Y. Lin, C. Chu, H. Wu,J.L. Huang, “Study of Azo thin films under different annealing atmosphere on structural, optical, and electrical properties by rf Magnetron Sputtering”. International MultiConference of Engineers and Computer Scientists 2015, Volume II, 2015. [8] S. A. Kamaruddin, K.-Y. Chan, H.-K. Yow, M. Zainizan Sahdan, H. Saim, and D. Knipp, “Zinc oxide films prepared by sol–gel spin coating technique,” Appl. Phys. A, vol. 104, No. 1, pp. 263–268, 2011.

5 P.248

SETS2016-66 The Southwest Emerging Technology Symposium 2016

STUDY OF TUNGSTEN-YTTRIUM BASED COATINGS FOR NUCLEAR APPLICATIONS Gustavo Martínez1, Jack Chessa1, Shuttha Shutthanandan2, Theva Tevuthasan2, Michael Lerche3 and C.V. Ramana1* 1

.Department of Mechanical Engineering, The University of Texas at El Paso, El Paso, TX, 799682. 2 .Environmental Molecular Sciences Laboratory (EMSL), Richland WA, 080341 3.McClellan Nuclear Research Center, UC Davis, One Shields Avenue, Davis, CA 95616* *Corresponding author ([email protected])

Keywords: Tungsten, Yttrium, Structural

ABSTRACT Material failure is one of the most considerable setbacks needed to be addressed by the materials research community to develop the current magnetic confinement nuclear fusion power reactors concepts that demand materials to serve under extreme conditions. Residual film stress resulting during material sputtering often contributes to material behaviour such as cracking, buckling, or delamination during materials performance. Tungsten (W) films ~300nm, were fabricated using DC current sputtering of a W target The residual stress, structure, morphology relationship with residual stress and mechanical properties as a function of substrate temperature (TS= RT-400 °C) during sputtering. W films were amorphous at TS≤RT, at which point a structural transformation occurs and BCC (α-W) is identified and stabilized for all subsequent growth temperatures. Studies showed that as grain size formation increases the residual stress distribution will reach the maximum (Max=62MPA) and stabilize after 350°C. Although the films go through a phase, the residual stress still continues to follow a parabolic pattern, indicating that the stresses mainly depend on grain organization rather than atomic packing. From Nano-scratch it is found that depth penetration decreases with increasing sputtering temperature. It is also observed that RT shows a complete delamination of the film and the pile-up decreases in height as 300C is reached. The results are presented and discussed. 1

Introduction

Recently it is in the research community the fundamental structure-property evaluation of Tungsten and Tungsten-Yttrium (W-Y) films variable with dependent sputtering pressure PAr. It was found that the sputtering pressure can tailor the phase structure and mechanical properties of the film. A transformation from body centered cubic ( -W) was favored at

P.249

STUDY OF TUNGSTEN-YTTRIUM BASED COATINGS FOR NUCLEAR APPLICATIONS

sputtering pressures greater than 0.63E-2 mbar whereas a metastable A15 phase ( -W) was favored when sputtering pressures of 0.46E-2 or lower where used. In order to achieve an isotropic, crack-resistant material, mechanical W-Y allow a fundamental understanding of mechanical and electrical properties must be investigated. This quarter was devoted to grow comparable W-Y samples with a variable content of Y in the lattice and its electrical and mechanical properties were compared to pure W, W 95 –Y 5 wt. % and W 90 –Y 10 wt. %.It is well known that the mechanical properties of nanocrystalline metals apparently differ from those of the conventional polycrystalline materials [1,5]. In the case of hard metal films, some studies have re- ported that the very high hardness of Ta [6] and Mo[7] films were obtained as 11.6 and 11.8 GPa, respectively, which are much higher than those of bulk Ta and Mo. Tungsten (W) and W-based alloyed films possess many attractive properties, such as high melting temperature, high mechanical strength, and good metal barrier performance [8,9]. Taking account of the fact that bulk W is harder than corresponding Ta and Mo, W films is anticipated to be of great scientific and technological interest. In this work, W films were prepared by magnetron sputtering onto Si (100) substrates. In order to improve the adhesion properties between W films and Si substrates, 30 nm Cr sticking layers was deposited prior to the W deposition for a set of samples. Nano-indentation was used to characterize the hardness and modulus of W films. Surprisingly, the ultrahigh hardness (24.5 GPa, 21.3 GPa) which are significantly higher than that of coarse-grained W in bulk form (3.92 GPa[10]) are obtained for the deposited and annealed W films with the average grain size of about 26.9 and 32.5 nm, respectively. The relationship between hardness enhancement and the microstructure features is discussed. 2

Experimental Methodology

Nano-crystalline tungsten coatings were deposited onto Silicon (100) and Sapphire (C-plane) using DC sputtering technique. The parameters used for this deposition fallow the DEC 2013 growth parameters. The film thickness was varied from 2 min of sputtering time deposition to 10 min in order to investigate crystallographic formation and texture with respect to film thickness. In general one of the many advantages of using physical vapor deposition processes is the ability to manipulate phase and microstructure with not only sputtering time, but also with deposition pressure, depositing power, substrate orientation. Something to consider is phase change due to exposition of the sample to external environments such as substrate interference. A more detailed table describing the film growth condition parameters is described in table 1. 2.1 Microstructure Characterization of the films Depending on the growth conditions, W thin films are usually made up of either the stable a phase (bcc lattice), the metastable B phase (A15 cubic lattice), or a mixture of both phases [12]. As it can be appreciated in figure 1 corresponding to pure W samples with variable deposition pressure, all the peaks in XRD data can be correlated to either bcc or A15 structure. The occurrence of dominant (210) is observed at sample W-Y 1 as highly crystalline with a mixture of (200) phase at PAr =0.46x102 mbar. As the deposition pressure increases to PAr =0.63x102 the dominant peak corresponds to a phase change in

P.250

STUDY OF TUNGSTEN-YTTRIUM BASED COATINGS FOR NUCLEAR APPLICATIONS

(110) and also a mixture with (200) plane. For the samples W-Y 3-6 the phase changes to a dominant (110) with different intensities of the peak. In plot 1B obviously, the effect of Yttrium content in the lattice affects the crystallite size making bigger crystals when compared to pure tungsten. Overall phase change can be considered the same for both when increasing the deposition pressure, the behavior is the same. Therefore it is concluded that the desired phase with less than 5% of Y wt. % can be tailored with the sputtering pressure as the appearance of the metastable β-W phase is related to the lowered deposition flux of W atoms, increased film porosity and correspondingly to the higher probability of oxygen incorporation all this related to sputtering pressure. This same behavior was previously observed by K. Salomon et al in pure W films prepared by magnetron sputtered films. 2.2 Tables and Figures Table 1

Table.1. Table 1 Depicts growth parameters conditions for W films

Figure 1

3. References

[1]A.H. Chokshi, A. Rosen, J. Karch and H. Gleiter. Scripta Mater., 1989, 23, 1679. [2] C.A. Schuh, T.G. Nieh and H. Iwasaki: Acta Mater.,2003, 51(2), 431.

[3] J. Schiotz and K.W. Jacobsen: Science, 2003, 301,1357. [4] H. Gleiter: Acta Mater., 2000, 48(1), 1. [5 ] K.S. Kumar, H.V. Swygenhoven and S. Suresh: Acta

3 P.251

STUDY OF TUNGSTEN-YTTRIUM BASED COATINGS FOR NUCLEAR APPLICATIONS

Mater., 2003, 51(19), 5743. [6]M. Zhang, B. Yang, J. Chu and T.G. Nieh: Scripta Mater., 2006, 54(7), 1227. [7 ] K.B. Yoder, A.A. Elmustafa, J.C. Lin, R.A. Hoffman and D.S. Stone: J. Phys D: Appl. Phys., 2003, 36,884. [8] K.Y. Ahn: Thin Solid Films, 1987, 153(1-3), 469 [9] V.G. Glebovsky, V.Y. Yaschak, V.V. Baranov and E.L. Sackovich: Thin Solid Films, 1995, 257(1), 1. [10] S. Eroglu, H. Ekren and T. Baykara: Scripta Mater.,1997, 38(1), 131. [11]M. Gutierrez, H. li and J. Patton. J. Thin Film Surface resistivity and Materials Engineering, 0-24 (2002)[12]S.M. Rossnagel, I.C.

P.252

SETS2016-67 The Southwest Emerging Technology Symposium 2016

CHARACTERIZATION AND ANALYSIS OF STRUCTURAL PROPERTIES, CRYSTALLOGRAPHY AND SURFACE POTENTIAL OF PEROVSKITE THIN FILMS Shaimum Shahriar1, Cheik Sana1, Vanessa Castaneda1, Manuel Martinez1 and Deidra Hodges1* Edison Castro2 and Luis Echegoyen2 Tahmina Akter2 and Geoffrey Saupe2 Eva Deemer3 and Russell Chianelli2 1Department of Electrical and Computer Engineering, The University of Texas at El Paso, El Paso, TX 79968, USA 2Department of Chemistry, The University of Texas at El Paso, El Paso, TX 79968, USA 3Metallurgical, Materials and Biomedical Engineering, The University of Texas at El Paso, El Paso, TX 79968, USA * Deidra Hodges ([email protected])

Keywords: Solution deposition, Crystallographic structure, Infrared (IR) spectroscopy, Crystalline and Surface potential

ABSTRACT The development of perovskite thin films by spin-coating, deposition techniques have been investigated. The methyl ammonium lead iodide (CH3NH3PbI3) perovskite has a direct band gap of 1.5 eV and a large absorption coefficient around 1×105 cm-1. It was deposited by a non-vacuum liquid-based coating method with 2 steps method by using Laurell Technologies WS650 spin processor. Characterization and Analysis of Structural Properties, Crystallography, Surface Roughness and Surface Potential of Perovskite were performed by using the Thermo Scientific DXR Smart Raman spectrometer, the Perkin Elmer spectrum 100 Fourier Transform Infrared spectrometer (IR spectrometer), Hitachi S-4800 Scanning Electron Microscope (SEM), the Bruker D8 Discover X-ray diffractometer (XRD), the NTMDT Ntegra Atomic force microscopy (AFM), KLA-Tencor D-120 Profilometer and the Scanning Kelvin Probe Microscopy (SKPM). Results were used to determine the fingerprint of each element, the crystal structure, orientation and crystallite size, surface roughness and surface potential of perovskite thin films. Raman peaks of perovskite appear at 54 cm-1, 65 cm-1, 81 cm-1, 103 cm-1 and 249 cm-1 wave number. IR spectroscopy identifies the methyl – CH3, N-H, C-H, C-N and M-X (metal – halide) bonds at 2870 cm-1, 3400 – 3500 cm-1, 1020 – 1220 cm-1 and at 750 cm-1 wave numbers. The XRD diffraction peaks, (110), (202), (220), (312), (224) and (314) were found at 14.077°, 24.47°, 28.354°, 31.768°, 40.48° and 43.09°. The uniformity and compact crystalline nature were observed in the films by SEM. The roughness, thickness, crystalline and dense-grained uniform morphology were exhibited by AFM. The thickness of the layer was ~ 350 nm which were determined by profilometer. The observed variation of surface potential in the 0 - 140 mV in case of perovskite thin film on the conductive substrate might reveal a presence of surface charges.

P.253

1

INTRODUCTION

Many different photovoltaic technologies are being developed for large-scale solar energy conversion [1]. Different materials have been used in nanostructured devices, with the goal of attaining high-efficiency thin film solar cells in such a way has yet to be achieved [2]. The hybrid inorganic– organic solar cell concept is ‘material agnostic’ in that it aims to use the optimum material for each individual function. Any semiconductor material that is easy to process, inexpensive and abundant can be used, with the aim of delivering a high-efficiency solar cell. Organometal halide perovskites have recently emerged as a promising material for high efficiency nanostructured devices [3-6]. A simple planar heterojunction solar cell incorporating vapor-deposited perovskite as the absorbing layer can have solar-to-electrical power conversion efficiencies (PCE) of over 15% (as measured under simulated full sunlight) [7]. With these materials, a PCE of 20% is achieved from single junction structure [8]. This demonstrates that perovskite absorbers can function at the highest efficiencies in simplified device architectures, without the need for complex nanostructures. Perovskite has a direct band gap of 1.60 eV and a large absorption coefficient (1x 105 cm-1) and very high charge carrier mobility [9]. In this study we report on characterization and analysis of perovskite thin films are done by using a 2 step method. The fingerprint of each element is identified by using the Thermo Scientific DXR SmartRaman spectrometer and the Perkin Elmer spectrum 100 Fourier Transform Infrared spectrometer (IR spectrometer). Crystallography were characterized by using the Hitachi S-4800 scanning electron microscope (SEM), the Bruker D8 Discover X-ray diffractometer (XRD) and the NT-MDT Ntegra Atomic force microscopy (AFM) to determine the crystal structure, orientation and crystallite size. The thickness of perovskite thin film was determined by KLA-Tencor Alpha-Step Profilometer. The surface potential profile was analyzed by using the Scanning Kelvin Probe Microscopy (SKPM). 2

THEORY

Soda lime glass (SLG) was used as substrates. SLG substrates were cleaned with acetone, methanol and deionized water rinses to remove hydrocarbons and other contaminants and dried with nitrogen (N2) gas. CH3NH3PbI3 can be fabricated by mixing two compounds (PbI2, CH3NH3I). CH3NH3I was synthesized by mixing methylamine (27.86 ml, 40% in methanol) and hydroiodic acid (30 ml, 57 wt.% in water) at 0⁰C and stirred for 2-3 hours. The precipitate was recovered by evaporation at 50⁰C - 70⁰C for 1 h. The product was washed with diethyl ether three times and then finally dried at 60⁰C in vacuum oven overnight [4]. In 1-2 ml N, N-dimethylformamide (DMF, 99.8%), 462 mg PbI2 was dissolved at 50⁰C - 100⁰C to make 1M PbI2 solution. Twenty microliters PbI2 solution was spin-coated at 3000 rpm for 30s, which was dried at 40⁰C for 3-5 min and 100⁰C - 150⁰C for 5 minutes consecutively. One hundred microliters of 0.063M CH3NH3I solution in 2-propanol was loaded on the PbI2coated substrate for 30s, which was spun at 4000 rpm and then dried at 100⁰C for 5-15 minutes [7,10].

P.254

3

DISCUSSION

3.1 Structural Properties Raman spectroscopy was performed in the determination of a single phase perovskite. Raman measurements with excitation wavelength at 532 nm were performed with the Thermo Scientific DXR SmartRaman spectrometer. The laser excitation power was 2 mW and the measurement was performed at room temperature of the methyl ammonium lead iodide film deposited on soda lime glass shown in fig.1. Low-frequency raman spectra were obtained from 50 cm−1 – 450 cm-1 for this experiment. The vibrational modes of the PbX6 octahedral were observed below 250 cm−1. The Raman spectra below 230 cm−1 involve vibrations of Pb−X bonds consisting of heteropolar ionic/covalent interactions in the inorganic framework. Raman bands due to the Pb−X vibrations of the octahedral structure are expected in the LF range [11]. In table I, among them the symmetric stretching mode 81 cm-1 is expected to be the highest in frequency, followed with the asymmetric stretching mode 103 cm-1 and the asymmetric deformation or asymmetric bending mode 65 cm-1. We had symmetric bending at 54 cm-1 and torsion at 249 cm-1 due to methyl ammonia (MA) vibration.

Fig.1. Raman spectroscopy for 2 step method perovskite single layer deposited on soda lime glass substrate.

Table I. Raman shift vs. assignment Raman shift (cm-1) Assignment 54 δs (X−Pb−X) 65 δas (X−Pb−X) 81 νs (X−Pb−X) 103 νas (Pb−X) 249 τ (MA) Where, δs = symmetric bending, δas = asymmetric bending, νs = symmetric stretching, νas = asymmetric stretching and τ = torsion. Infrared spectroscopy (IR spectroscopy) is the spectroscopy that deals with the infrared region of the electromagnetic spectrum that is light with a longer wavelength and lower frequency than visible light. It covers a range of techniques, mostly based on absorption spectroscopy. As with all spectroscopic techniques, it can be used to identify and study chemicals [12]. We used IR spectroscopy to identify the chemical bond of perovskite. In

3 P.255

perovskite we have methyl –CH3, N-H, C-H, C-N and M-X (Metal – Halide) bonds. In fig.2 we can identify –CH3 at 2870 cm-1, N-H at 3400 – 3500 cm-1, C-N at 1020 – 1220 cm-1 and M-X which is Pb-I bond at 750 cm-1. Characteristic internal vibrational modes of MA appear at high frequencies, just below 1000 cm−1 (C−N stretching), at 1400 −1600 cm−1 (CH3 and NH3 bending) and at around 3000 cm−1 (CH3 and NH3 stretching). The most interesting vibrational modes are the rocking vibration and low-frequency torsion of MA around the C−N axis. The MA rocking mode appears in between 911 and 923 cm−1 and sharpens, increases in intensity.

Fig.2. FTIR spectra response observed in 2 step method of CH3NH3PbI3 film.

The dynamics of the formation of the perovskite were monitored by optical absorption, emission and X-ray diffraction (XRD) spectroscopy. The Bruker D8 Discover XRD diffractometer with Cu Kα radiation (λ=1.5406 Å). All the diffraction peaks, (110), (202), (220), (312), (224) and (314) were assigned at 14.077°, 24.47°, 28.354°, 31.768°, 40.48° and 43.09° which are shown in fig.3. The results shows also the secondary phases PbI2. In table II, the crystallite size is inversely proportional to the full width at half maximum (FWHM) peak width based on the Scherrer’s relation [13-15]. Crystallite sizes of the perovskite thin films can be analyzed by using the crystallite size of CH3NH3I and PbI2. The Scherrer’s relation is used to calculate the crystallite sizes. D = 0.9λ/ cosθ

(1)

Where, D is the diameter of the crystallite, λ is the X-ray wavelength, β is the line broadening at half the maximum intensity (FWHM), after subtracting the instrumental line broadening, in radians and θ is the Bragg angle. Table II. FWHM vs Crystallite size between PbI2 and CH3NH3I FWHM (⁰) 0.266 PbI2 0.168 CH3NH3I

Crystallite size (Å) 305.8 485.5

The interplanar spacing of (110) Miller indices of perovskite is 6.05774 Å.

P.256

Fig.3. XRD profiles for 2 step perovskite thin film.

3.2 Crystallography The morphology of perovskite thin films is important to produce large, densely packed and uniform grains. Producing dense films with few interconnection between large grains is required for high performance PV devices [16]. In fig.4, it had been reported that perovskite had small grains. The uniformity and crystalline nature was observed in the films. They were compacted grains.

b)

a)

Fig.4. SEM surface micrographs for perovskite thin films a) compact grains and b) mixture of small and large grains.

The thickness of the film was reported in fig.5 a) from profilometer. Our optimal goal for thickness is 330 nm but we observed an average around 350 nm. We saw waviness on top of the sample in fig.5 b) and the height was more than 300 nm by using AFM.

5 P.257

Fig.5.a) The thickness of a 2 step perovskite film from profilometry and b) the topology image from AFM of a 5µm × 5µm perovskite thin film.

3.3 Surface Potential Scanning kelvin probe microscopy (SKPM) is employed to examine the surface and Grain boundaries (GBs) in the perovskite film. The SKPM has been used to determine the surface potential difference between GBs and inner grains in a thin film solar cell, which helps to reveal the band bending in the energy band diagram around the GBs. Spatial maps of topography and surface potential are shown in fig.6.a) The film surface potential was reported in fig.6.b) and it was around 0 - 140 mV.

Fig.6.a) SKPM topography of the surface potential of the film prepared by the 2 step deposition method and b) Surface potential vs position curve

4

CONCLUSIONS

Raman peaks of perovskite appear at 54 cm-1, 65 cm-1, 81 cm-1, 103 cm-1 and 249 cm-1 wave number. IR spectroscopy identifies the methyl –CH3, N-H, C-H, C-N and M-X (metal – halide) bonds. The XRD diffraction peaks were found at 14.077°, 24.47°, 28.354°, 31.768°,

P.258

40.48° and 43.09°. The uniformity and compact crystalline nature were observed in the films by SEM. The roughness, thickness, crystalline and dense-grained uniform morphology were exhibited by AFM. The thickness of the layer was ~ 350 nm which were determined by profilometry. The observed variation of surface potential in the 0 - 140 mV in case of perovskite thin film on the conductive substrate might reveal a presence of surface charges. We observed based on all this characterizing results that we have synthesize perovskite thin film successfully and incorporated these thin film into solar cell. Results of these devices will be presented in a separate paper. References [1] M. Graetzel, R. A. Janssen, D. B. Mitzi and E. H. Sargent, Nature 488 (7411), 304-312 (2012). [2] M. A. Green, K. Emery, Y. Hishikawa, W. Warta and E. D. Dunlop, Progress in photovoltaics: research and applications 23 (1), 1-9 (2015). [3] M. M. Lee, J. Teuscher, T. Miyasaka, T. N. Murakami and H. J. Snaith, Science 338 (6107), 643-647 (2012). [4] 4. A. Kojima, K. Teshima, Y. Shirai and T. Miyasaka, Journal of the American Chemical Society 131 (17), 6050-6051 (2009). [5] 5. H.-S. Kim, C.-R. Lee, J.-H. Im, K.-B. Lee, T. Moehl, A. Marchioro, S.-J. Moon, R. Humphry-Baker, J.-H. Yum and J. E. Moser, Scientific reports 2 (2012). [6] 6. J. H. Noh, S. H. Im, J. H. Heo, T. N. Mandal and S. I. Seok, Nano letters 13 (4), 1764-1769 (2013). [7] 7. P. Docampo, J. M. Ball, M. Darwich, G. E. Eperon and H. J. Snaith, Nature communications 4 (2013). [8] 8. S. Collavini, S. F. Völker and J. L. Delgado, Angewandte Chemie International Edition 54 (34), 97579759 (2015). [9] C. C. Stoumpos, C. D. Malliakas and M. G. Kanatzidis, Inorganic chemistry 52 (15), 9019-9038 (2013). [10] T. Minemoto and M. Murata, Journal of Applied Physics 116 (5), 054505 (2014). [11] R. G. Niemann, A. G. Kontos, D. Palles, E. I. Kamitsos, A. Kaltzoglou, F. Brivio, P. Falaras and P. J. Cameron, The Journal of Physical Chemistry C (2016). [12] N. J. Jeon, J. H. Noh, Y. C. Kim, W. S. Yang, S. Ryu and S. I. Seok, Nature materials 13 (9), 897-903 (2014). [13] P. Scherrer, Bestimmung der inneren Struktur und der Größe von Kolloidteilchen mittels Röntgenstrahlen. (Springer, 1912). [14] J. I. Langford and A. Wilson, Journal of Applied Crystallography 11 (2), 102-113 (1978). [15] J. He, L. Sun, K. Zhang, W. Wang, J. Jiang, Y. Chen, P. Yang and J. Chu, Applied Surface Science 264, 133-138 (2013). [16] A. Sharenko and M. F. Toney, Journal of the American Chemical Society (2015).

7 P.259

SETS2016-68 The Southwest Emerging Technology Symposium 2016 STRUCTURAL, DIELECTRIC, AND PIEZOELECTRIC CHARACTERIZATION OF LEAD-FREE CALCIUM AND CERIUM BaTiO3 MODIFIED 1

Juan A. Duran 1, Cristian Orozco 1 and C. V. Ramana1* Department of Mechanical Engineering, University of Texas at El Paso, El Paso, Texas 79968, USA *Corresponding author: [email protected].

Keywords: Barium-Titanate Ceramics, Calcium and Cerium doping, Dielectric Constant leap, Electrical Properties. ABSTRACT Structure, morphology, and regulation of the dielectric properties via close-composition intervals is demonstrated for variable-cerium, constant-calcium co-doped barium titanate (Ba0.80Ca0.20CeyTi1-yO3; x=0.0-0.25; referred to BCCT). The effect of variable Ce-content on the structure and dielectric properties of BCCT is investigated. X-ray diffraction spectra confirms the studied samples are mainly in BT tetragonal phase with a small secondary phase detected as CaTiO3 in BCCT for y = 0.20 and 0.25. However, the lattice parameter reduction was evident with increasing Ce-content. Composition-driven dielectric constant leap (4,0005,500) was observed from intrinsic BCT to BCCT for (y = 0.0-0.04). The temperaturedependent dielectric constant showed a transition temperature, which decreased with progressive addition of Cerium content. The Curie point, Tc, diminishes from 120 to 50 °C for (y = 0.0-0.04) showing a decrease in the ferroelectric to paraelectric state. Hence, the solubility limit for cerium in BCCT ceramics may have been reached. 1 Introduction 1.1 Exploration of potential Semiconductors Polycrystalline electro-ceramics such as Barium Titanate (BaTiO3) have opened a new era of exploration in the semiconductor industry. Interest in lead-free ferroelectric piezoelectrics has captured the attention of researchers over the years due to the on-going need to find a potential replacement of commercial piezoelectric-lead zirconate titanate (PZT) based sensors and actuators, which currently face global restrictions due to its high toxicity lead-content [1]. Barium Titanate ceramics are widely used in the manufacturing process of multilayer capacitors, thermistors, etc., due to its relatively high dielectric constant [1, 2]. It has been demonstrated that the co-doping approach is an efficient method of improved physical and electrical properties for this family of compounds, having the general formula ABO3. Now-adays BaTiO3 has become the basic capacitor material in the semiconductor technology [1, 3]. As part of the ongoing research on lead-free piezolectrics, it is of interest to further explore the different types of cations whose substitution at the Ba and/or Ti sites could enhance the piezoelectric properties, while maintaining its Curie temperature for its practical applications. In our attempt to obtain insulating ceramics to improve the electrical properties of this compound for energy conversion and storage application purposes, we devote this analysis to calcium and cerium doped barium titanate at close composition intervals in the dilute concentration regime. Cerium can exist in two oxidation states, Ce 3+ and Ce 4+. Substitution of Ti 4+ ions with donor dopant tetravalent Ce 4+ is assumed, further discussion will be addressed as structure, dielectric, and piezoelectric characterization will be used to

P.260

The Southwest Emerging Technology Symposium 2016

further confirm the morphology of the material. It is interesting to note that the ionic sizes for Ti 4+ (0.0605 nm for coordination number 6) and Ce 4+ (0.087 nm for coordination number 6). It has been reported that those cations bigger in size than Ti 4+, may improve the piezoelectric properties of barium titanate [1-3]. Moreover, from the point of view of piezoelectricity, it is desirable that the substituting cations does not increase the electrical conductivity of the specimen [3]. 2 Experimental 2.1 Electro-ceramics Synthesis Barium Carbonate (BaCO3), Calcium Carbonate (CaCO3), Titanium Oxide (TiO2), and Cerium Oxide (CeO2) were used in synthesis of BCCT ceramics. Synthesis of the BCCT ceramics has been carried out with Ca content kept constant at 20%, which is suitable for providing temperature stability. Stoichiometric amounts of Ba0.80Ca0.20CeyTi1-yO3 (y=0.00.25) were weighed and ground in ethanol for 2 h in agate mortar and pestle. The mixed powders were then immediately calcined at 1250°C for 10 h for carbonate to burn out. Phase identified powders were made into pellets by uniaxially pressing in a 13 mm die with an applied pressure of 100 MPa for 4 min each. Ceramic thick capacitors were built having 1.02 mm in thickness and 7.46 mm in diameter. The pellets were then sintered at 1480 °C for 6 h in air in a box furnace with heating and cooling rates of 10°C/min. The BCCT ceramics were characterized by X-ray diffraction (D8 Discovery X-ray diffractometer with Cu Kα radiation, λ = 1.54 Å). The dielectric response of the sintered materials was measured using a Hewlett– Packard 4284A impedance gain phase analyzer over the frequency ranging from 100 to 1 MHz and at an oscillation voltage of 1 V. The measurements were performed over a temperature range from 20 to 150 °C using an inbuilt heating system. 3 Results and Discussion 3.1 X-ray diffraction X-ray diffraction (XRD) patterns of BCCT ceramics were analyzed as a function of variable Ce composition. All the main peaks observed are comparable to those of the powder XRD pattern of pure BT tetragonal at room temperature. However, peak due to minor secondary phase corresponding to CaTiO3 was observed in the BCCT with increasing Ce content. The diffraction peaks for the CaTiO3 phase may be due to the preparation of samples CaCO3 and TiO2. Due to Ca doping the peak intensity i.e. (121), (202), and (042) reflections of CaTiO3 phase increases slightly, this also means that because Calcium has been added, the formation of CaTiO3 was induced. It is important to notice that these changes reflect the chemistry of the Ce content on the crystal lattice of BCT ceramics due to the fact that Ca is kept constant throughout the studies. 3.2 Dielectric Properties The variation of dielectric constant of BCCT ceramics with Ce content (y) is presented in Figure 1. The data shown are at different constant frequency values. It is evident from Fig. 1 that the Ce content strongly influences the dielectric constant of the samples. The dielectric

P.261

The Southwest Emerging Technology Symposium 2016

4000

(Ba0.80Ca0.20)TiO3

100Hz 1 kHz 10 kHz 100 kHz 1 MHz

3500 3000 2500 2000 1500 1000 500 20

40

60

80

100 120 140 160

Dielectric Constant ( )

Dielectric Constant ( )

constant increases with increasing y values in BCCT ceramics. The increase in the dielectric constant with the inclusion of Ce could be due to the fact that, with the inclusion of smaller size Ti4+ ions for larger Ce4+ ions, the BCT lattice is distorted giving rise to increase in the atomic polarizability subsequently the dielectric constant. From Fig. 2 (a-c) it is seen that the dielectric constant increases up to a maximum value, ε' max, and then decreases with an increase of temperature for compounds y = 0.0, 0.02, 0.04. The maxima, ε' max, corresponds to a transition temperature where ceramics undergo a structural shift from ferroelectric (tetragonal) to paraelectric (cubic) phase. Such transition temperature is termed Curie point, Tc [1, 2]. Similarly, as Ce content is introduced, the ε' max value tends to shift towards the low – temperature region. This phenomena may be due to the solubility limit of Ce content in BCT ceramics. There is a decrease in the ferroelectric to paraelectric phase transition temperature, Tc from 120 to 50 °C for y = 0.00 to y = 0.04 with an increase in Ce doping concentration. Such decrease further indicates a partial substitution of Ti 4+ ion with Ce 4+ ion [2].

Dielectric Constant ( )

20

(Ba0.80Ca0.20)(Ce0.02Ti0.98)O3

1200 1000 800 600

100Hz 1 kHz 10 kHz 100 kHz 1 MHz

400 200 20

40

60

80

100

120

140

160

(°C)

(°C) 5500 5000 4500 4000 3500 3000 2500 2000 1500 1000

1400

(Ba0.80Ca0.20)(Ce0.04Ti0.96)O3

100Hz 1 kHz 10 kHz 100 kHz 1 MHz

40

60

80

100

120

140

160

(°C) Fig. 1. The variation of the dielectric constant with temperature over the range of 100 Hz – 1 MHz of Ba0.80Ca0.20CeyTi1-yO3 compounds for (a) y=0.00, (b) y=0.02, (c) y=0.04. Fig. 2. The variation of the lattice parameters, a and c, of Ba0.80Ca0.20CeyTi1-yO3 ceramics as a function of Ce concentration y.

References [1] Brajesh, K., Kalyani, A. K., & Ranjan, R. “Ferroelectric instabilities and enhanced piezoelectric response in Ce modified BaTiO3 lead-free ceramics”. Applied Physics Letters, 2015. [2] Brajesh, K., Kalyani, A. K., & Ranjan, R. “Ferroelectric instabilities and enhanced piezoelectric response in Ce modified BaTiO3 lead-free ceramics”. Applied Physics Letters, 2015. [3] Heywang, Walter, Karl Lubitz, and Wolfram Wersing. “Piezoelectricity: evolution and future of a technology”, Vol. 114. Springer Science & Business Media, 2008.

3 P.262

SETS2016-69 The Southwest Emerging Technology Symposium 2016

PHYSICAL AND OPTICAL PROPERTIES OF CZTS SPIN COATING AND DOCTOR BLADE PROCSSING

Donato Kava1, Shaimum Shahriar1, Manuel Martinez1, Cheik Sana1, Jose Galindo1, Vanessa Castaneda1, and Deidra R. Hodges1 Electrical and Computer Engineering Department, University of Texas at El Paso, El Paso, TX 79968, USA; * Corresponding author ([email protected])

Keywords: CZTS Thin films, spin coating, doctor blading, and photovoltaic cells.

ABSTRACT Cu2ZnSnS4 (CZTS) is an emerging photovoltaic (PV) material intended to replace copper indium gallium (di) selenide(CIGS). Processing costs are one aspect in need study for large scale production of CZTS solar cells as a CIGS successor. Spin coating and doctor blading are two low cost processing techniques studied. Characterization and analysis of the thin films were performed using Raman spectroscopy, scanning election microscope (Hitachi S4800), X-ray diffraction (Philipps X’Pert), Proflimeter (Veeco Dektak 150), UV-Vis-NIR Spectrophotometer (Carry 5000) measurements. Device characterization confirms fabricated CZTS is an optimal material alternative to conventional cells. 1

INTRODUCTION

Research in thin film solar cells continues to make substantial progress towards developing not only economically viable, high-efficiency photovoltaic (PV) devices but specifically using CZTS allows for a solar cell that is also environmentally non-toxic. If Earth abundant element thin film solar cells can be developed with equal or better efficiency to CIGS or CdTe, they will likely be the long-term solution of choice for low-cost terawatt scale PV[1]. Besides the issue of abundance, heavy metal cadmium has experienced resistance towards adoption in some countries because of the toxicity[2]. The first CZTS solar cell was constructed from a heterostructure with cadmium tin oxide, yielding an open circuit voltage of 165 mV under AM1.5 illumination (no efficiency reported)[2]. Work on this technology has since improved. For example by Katagiri et al. on CZTS thin films prepared by thermally evaporated elements and binary chalcogenides in high vacuum, have resulted in a solar cell with a conversion efficiency (PCE) of 6.7% after a preferential etching in deionized water (DIW) [3]. However, using vacuum raises costs for large scale manufacturing, thus making spin coating CZTS efficiencies economically viable alternative. Additionally, efficiencies have reached as high as 12.6% efficiency was reported by Wang et al., using a hydrazine-based CZTSe approach[4]. As hydrazine is toxic and highly reactive, handling procedures make large scale manufacturing a problem. Using 2methoxyethanol as a solvent can solve this problem as it has a high boiling point and low reactivity [5]. The most important physical quantities for processing are the final film thickness and uniformity of the layers. [6]. The optimized spin coating speed was reported to be 3000 rpm [7]. Additionally, CZTS layers have shown optimize grain structures when annealed at 550° C resulting in densely packed grains with a few large 1 micron grains [8].

P.263

Since the microstructure of the films is mainly determined by the substrate temperature, the lattice match of the compound, the substrate properties, the growth process direction, and the growth rate and pressure during deposition of the films[8], by optimizing these properties CZTS thin films have been produced using non-vacuum liquid-based spin coating procedures and doctor blade printing. 2

EXPERIMENTAL

Prior to use, substrates are were cleaned with acetone, and deionized water rinses and dried with nitrogen (N2), to remove hydrocarbons and other contaminates. Thin heat resistant tape strips were laid for later profilometer measurements. Thin films were prepared using a CZTS precursor created by mixing Alfa Aesar copper (II) acetate monohydrate, Fisher Scientific zinc (II) acetate dihydrate, and Alfa Aesar tin (II) chloride dihydrate of 4.375 x 10-2, 2.1875 x 10-2 and 2.1875 x 10-2 mol, respectively, in a 50 ml solvent of Acros 2-methoxyethanol. Alfa Aesar sulfur powder was added as the sulfur source at 8.75 x 10-2 mol. 5ml of sigma monothanolamine is added and used as a stabilizer. The solution is stirred and heated at 45° C for 1 hour to dissolve metal sources. The solution was deposited on SLG substrates using spin coating the speed of 3000RPM. The film was then dried in air at 300° C for 15 minutes on a hot plate. This process was repeated 19 additional times and results in 20 layers of coated and dried films. After the fifth layer is complete the samples are to be annealed varying the temperature from 400°C to 600°C. For the doctor blade samples, the p-type layer was deposited by running a glass microscope slide over a high temperature kylpton tape mask. The film is dried in air at 75° C for 20 minutes on a hot plate. The process is repeated for bi-layer deposition. After the layer is complete the samples are to be annealed varying the temperature from 400°C to 600°C. RESULTS AND DISCUSSION

The desired thickness for the CZTS layer is 800nm. This thickness is presumed to give the best electrical properties. The work starts with the optimization of the CZTS layer. . XRD patterns were used to determine crystal structure, orientation, and crystallite size of the CZTS layer. Peaks at (112), (220), and (312) and compared for both dr blade and spin coating.

Fig. 4.

XRD profiles of CZTS spin coated samples

P.264

References [1] [2] [3] . [4] [5] [6] [7] [8]

W. Ki and H. W. Hillhouse, "Earth-Abundant Element Photovoltaics Directly from Soluble Precursors with High Yield Using a Non-Toxic Solvent," Advanced Energy Materials, vol. 1, pp. 732-735, 2011. D. B. Mitzi, O. Gunawan, T. K. Todorov, K. Wang, and S. Guha, "The path towards a highperformance solution-processed kesterite solar cell," Solar Energy Materials and Solar Cells, vol. 95, pp. 1421-1436, 6// 2011. K. Hironori, J. Kazuo, Y. Satoru, K. Tsuyoshi, M. Win Shwe, F. Tatsuo, et al., "Enhanced Conversion Efficiencies of Cu 2 ZnSnS 4 -Based Thin Film Solar Cells by Using Preferential Etching Technique," Applied Physics Express, vol. 1, p. 041201, 2008 W. Wang, M. T. Winkler, O. Gunawan, T. Gokmen, T. K. Todorov, Y. Zhu, et al., "Device Characteristics of CZTSSe Thin-Film Solar Cells with 12.6% Efficiency," Advanced Energy Materials, vol. 4, pp. n/a-n/a, 2014. H.-m. Zhou, D.-q. Yi, Z.-m. Yu, L.-r. Xiao, and J. Li, "Preparation of aluminum doped zinc oxide films and the study of their microstructure, electrical and optical properties," Thin Solid Films, vol. 515, pp. 6909-6914, 6/13/ 2007. C. J. Lawrence, "The mechanics of spin coating of polymer films," Physics of Fluids (1958-1988), vol. 31, pp. 2786-2795, 1988. Y. Yue, G. Jie, T. Prabhakar, and Y. Yanfa, "Effects of Spin Speed on the Properties of Spin-coated Cu2ZnSnS4 Thin Films and Solar Cells Based on DMSO Solution," in Photovoltaic Specialist Conference (PVSC), 2014 IEEE 40th, 2014, pp. 0448-0451. B. J. Deidra Hodges, Toussaint Moseley, Aaron Love, Caleb Burke, Edward Jones, Irina Tyx, Manoj Chaulogain, and Ophelia Johnson, "Development of CZTS Thin Films by Non-vacuum, Liquid-based Techniques for Efficient, Low-cost CZTS Solar Cells," IEEE Transactions on Device and Materials Reliability vol. 13, p. 4, 2013.

3 P.265

Session at a Glance Abbreviation Title Keynote Presentation I - John Applewhite 1-A Additive Manufacturing I 1-B Emerging Technologies I 1-C Space Systems 1-D Energy and Sustainability I 1-E Aerospace and Energy Materials I Special High School Info Session 2-A Additive Manufacturing II 2-B Emerging Technologies II 2-C Propulsion and Energy Technologies I 2-D Energy & Sustainability II 2-E Aerospace and Energy Materials II Keynote Presentation II - Charles Chase 3-A Additive Manufacturing III 3-B Emerging Technologies III 3-C ITAR 1 - Propulsion and Energy Technologies II 3-D Energy and Sustainability III 3-E Aerospace and Energy Materials III 4-A Propulsion and Energy Technologies III 4-B Emerging Technologies IV 4-C Propulsion and Energy Technologies IV 4-D Energy and Sustainability IV 4-E Aerospace and Energy Materials IV Keynote Presentation III - Nick Gonzales

Start Time 8:30 AM 9:50 AM 9:50 AM 9:50 AM 9:50 AM 9:50 AM 10:00 AM 11:10 PM 11:10 PM 11:10 PM 11:10 PM 11:10 PM 12:20 PM 1:50 PM 1:50 PM 1:50 PM 1:50 PM 1:50 PM 3:50 PM 3:50 PM 3:50 PM 3:50 PM 3:50 PM 5:40 PM

End Time 9:50 AM 10:50 AM 10:50 AM 10:50 AM 10:50 AM 10:50 AM 12:00 PM 12:10 PM 12:10 PM 12:10 PM 12:10 PM 12:10 PM 1:30 PM 3:30 PM 3:30 PM 3:30 PM 3:30 PM 3:30 PM 5:30 PM 5:30 PM 5:30 PM 5:30 PM 5:30 PM 7:00 PM

Location Rosewood Walnut Satinwood Sandalwood Orchid Poplar AV Theater Walnut Satinwood Sandalwood Orchid Poplar Rosewood Walnut Satinwood Sandalwood Poplar Orchid Walnut Satinwood Sandalwood Poplar Orchid Rosewood

CONFERENCE ENTRANCE REGISTRATION

SATINWOOD

SESSIONS: 1D-2D 3D-4D

ORCHID

POSTER SESSION

SESSIONS: 1B-2B 3B-4B WALNUT

POPLAR

WILLOW/ROSEWOOD/OAKWOOD Keynote Presentations Breakfast/Lunch/Dinner

SANDALWOOD

SESSIONS: 1E-2E 3E-4E

SESSIONS: 1A-2A 3A-4A

SESSIONS: 1C-2C 3C-4C

BATHROOMS

EXECUTIVE BOARDROOM Symposium Office

A/V THEATRE High School Info Session

WYNDHAM MAIN ENTRANCE

Aboud J, 4B-5 Acuna E, 1A-2 Adams J, 1C-1, 4A-3 Aguilar D, 4A-4 Aguilera E, 3A-1 Akter T, 4E-3 Akundi A, 3D-2 Ambriz S, 3A-3, 1A-1 Arif M, 2E-2 Arrieta E, 3A-2 Avila J, 3B-3 Avila R, 2B-2 Badhan A, 4A-4 Beas C, 3E-4 Belcher T, 4C-4 Benedict B, 3D-4, 1D-2 Bugarin L, 4A-3, 4D-5 Cabrera L, 4B-5 Camacho D, 1C-2, 3C-3 Candelaria J, 4A-1 Carrasco C, 3A-2 Carroll J, 4A-1 Castaneda V, 4E-5, 4E-1, 4E-3, 2E-1, 1E1, 1E-2 Castellanos A, 3E-2 Castro E, 4E-3, 1E-2 Catzin C, 3B-4 Chalamasetty G, 2D-3 Challa V, 1E-3

Author Index Chattopadhyay A, 4B-4 Chavez L, 3B-3 Chessa J, 3D-1, 4E-2, 3E-4 Chianelli R, 2D-2 Chintalapalle R, 4E-2, 4E-4, 2E-3 Choudhuri A, 1A-3, 3A-4, 3B-3, 1C-1, 4A-3, 3C-5, 2C-1, 4A-5, 2C-3, 4C-4, 3C-3, 4A-4, 4C-2, 4B-5, 4D-5 Choudhuri Ar, 4A-4 Chowdhury M, 1D-1 Cordova S, 2C-2 Coronel J, 3A-3, 1A-1 De Hoyos D, 4D-2 De la Torre A, 2C-1, 2C-3 De la Torre J, 1C-3 Deemer E, 4E-3 Delfin D, 2E-2, 3B-3 Delfin L, 3B-3 Demko A, 4C-1, 4C-5, 4C-3 Dillier C, 4C-1, 4C-3 Ding Y, 3B-2 Duran J, 4E-4 Echegoyen L, 4E-3, 1E-2 Espalin D, 2A-1, 3A-1, 3A-3, 1A-1 Everett M, 1C-2 Ferguson R, 3D-1 Fitzgerald R, 3D-3 Flores A, 1C-2, 1C-3 Flores E, 3C-3 Flores M, 2B-1

Galindo J, 4E-5, 4E-1, 2E-1, 1E-1, 1E-2 Galvan E, 1D-3 Garcia C, 3E-5 Garcia E, 2B-3 Garcia R, 1A-2 Gonzalez J, 1A-3, 3A-4 Gonzalez J, 2A-3 Grossman K, 4C-1, 4C-3 Gutierrez L, 2D-1 Haque M, 3A-2 Hasanul K, 2E-2 Haynes C, 4A-1 Hernandez M, 4A-2, 4B-5 Hodges D, 2E-2, 4E-5, 4E-1, 4E-3, 2E-1, 1E-1, 1E-2, 1E-3 Hogge K, 3C-3 Holt J, 1C-2, 4C-2 Hossain M, 1A-3 Hossain M, 3A-4 Hu Z, 4D-4 Islam S, 4B-1, 2B-2, 4B-3 Johnson A, 1C-1, 4A-5 Juarez J, 1D-1 Karim H, 3B-3, 3B-5 Kava D, 4E-5, 4E-1, 2E-1, 1E-1, 1E-2 Kiekintveld C, 1D-1 Kilari S, 2A-2 Kim C, 3A-3 Kim H, 3D-2

Kim N, 2A-2 Kotteda V, 4B-4 Kumar V, 4B-4, 4B-2 Lerche M, 4E-2 Li C, 3B-2 Lin Y, 1A-3, 3A-4, 3D-2, 1B-2, 4D-4, 2E2, 3B-3, 3B-5, 3E-1, 2A-3 Lopez I, 4A-3, 4D5 Lopez J, 3B-2 Love N, 2E-2, 3B-5, 4A-2, 4C-4, 3C-3, 4A-4, 4C-2, 4B-5 Lovich B, 4A-2 Lu D, 3D-3 Mahmud S, 3D-3 Manandhar S, 2E-3 Mandal P, 2D-1, 1D-3, 2D-3, 3D-5 Marquez D, 3A-5 Martinez G, 4E-2 Martinez M, 4E-5, 4E-1, 4E-3, 2E-1, 1E1, 1E-2 Martinez R, 1A-3, 3A-4, 3B-5, 3E-1 Mata C, 3D-1 McDonald E, 3A-1 Mejia J, 3C-5, 4C-2 Mireles J, 1A-3, 3A-4, 3A-2, 2A-3 Mishra A, 1E-3 Misra R, 1E-3 Mollenhauer D, 2B-1 Moreno D, 3B-2 Morrow G, 4C-5

Motta J, 2A-1 Nandy A, 4D-1 Nunel K, 1E-3 Nunez P, 4A-5 Olivarez A, 4D-2 Orozco C, 4E-4 Ott D, 4A-5 Paez A, 3B-2 Perez M, 1A-1 Petersen E, 4C-1, 3C-2, 4C-5, 4C-3, 3C-4 Ponce R, 4A-5, 4D-5 Prabhakar P, 3E-5, 3E-2, 4B-1, 2B-2, 4B3, 1A-2 Quintana J, 1B-3 Ramirez C, 3B-1 Reyes J, 3E-3 Rodriguez C, 3B-3 Rubio E, 2E-3 Rumpf R, 3B-3 Sajjadi S, 3D-5 Sammet T, 4C-1, 4C-3 Sana C, 4E-5, 4E-1, 4E-3, 2E-1, 1E-1, 1E-2 Sandoval A, 4A-3 Sarker R, 2E-2, 3B-5 Saupe G, 4E-3 Schiaffino A, 4B-2 Seal S, 4C-1, 4C-3 Shafirovich E, 2C-2, 3D-1, 3C-1 Shahriar S, 2E-2, 4E-5, 4E-1, 4E-3,

2E-1, 1E-1, 1E-2 Shariful S, 3E-2 Shutthanandan S, 4E-2 Shuvo I, 2E-2 Shuvo M, 3A-4 Silva J, 2E-2, 3B-5 Soto S, 4A-3 Stahl J, 3C-2, 3C-4 Stewart C, 3B-1, 3B-4, 2B-3, 3A-2, 3E-3 Stockwell W, 3D-3 Tarango E, 3E-1 Tauhiduzzaman M, 4B-3 Tevuthasan T, 4E-2 Thomas J, 3C-4 Trillo J, 1C-1, 4A-1 Tseng T, 3D-2, 4D-2, 1B-2, 4D-4, 2D-3, 3D-5, 2A-2 Valenzuela J, 4C-4 Vidal V, 4D-3 Vidana O, 4B-5 Wang C, 4D-1 Wang Y, 3B-2 Wangchuk P, 3D-3 Wicker R, 1A-3, 3A-4, 2A-1, 3A-2, 2A-3, 3A-1, 3A-3, 1A-1 Wu J, 1B-2, 4D-4 Yepez J, 1C-3 Zamora A, 2C-1, 2C-3

Department of Mechanical Engineering ǀ Engineering Annex, Suite 126 The University of Texas at El Paso ǀ 500 W. University Avenue ǀ El Paso, TX 79968-0521 Phone (915) 747 5450 ǀ Fax (915) 747 5019 ǀ me.utep.edu [email protected] ǀ [email protected] ǀ [email protected]