2011 OF THE CROATIAN ACADEMY OF ENGINEERING [PDF]

ANNUAL 2010/2011 OF THE CROATIAN ACADEMY OF ENGINEERING

CROATIAN ACADEMY OF ENGINEERING

Annu. Croat. Acad. Eng.

ISSN 1332-3482

ANNUAL 2010/2011 OF THE CROATIAN ACADEMY OF ENGINEERING Editor-in-Chief Vilko Žiljak

Zagreb, 2012

Published by Croatian Academy of Engineering, 28 Kaþiü St., 10000 Zagreb, Croatia

Editor-in-Chief Prof. Vilko Žiljak, Ph.D., Vice president of the Croatian Academy of Engineering

Editorial Board Prof. Stanko Tonkoviü, Ph.D. Prof. Vilko Žiljak, Ph.D. Prof. emer. Zlatko Kniewald, Ph.D.

ISSN 1332-3482 Annual 2010/2011 of the Croatian Academy of Engineering Annu. Croat. Acad. Eng.

Prepress and Press INTERGRAFIKA, Zagreb

Circullation 300

Table of contents B. Pribiþeviü, D. Medak, A. Ĉapo GEODETIC CONTRIBUTION TO THE GEODYNAMIC RESEARCH OF THE AREA OF THE CITY OF ZAGREB .................................................... 11 Drago Katoviü, Andrea Katoviü, Marija Krnþeviü SPANISH BROOM (SPARTIUM JUNCEUM L.) .............................................. 23 Matanoviü Davorin, Moslavac Bojan, Nediljka Gaurina-Meÿimurec THE WAY TO REDUCE PIPE WEARING WHILE DRILLING ...................... 38 Dinko Mikuliü, Vladimir Koroman FUNDAMENTALS OF MILITARY PRODUCTION DEVELOPMENT IN THE CROATIAN WAR FOR INDEPENDENCE FROM 1991 TO 1993 .......... 46 Darko Stipaniþev INTELLIGENT FOREST FIRE MONITORING SYSTEM – FROM IDEA TO REALIZATION ................................................................................... 58 Igor Petroviü, Davorin Kovaþiü LABORATORY TESTING AND NUMERICAL MODELLING OF MBT WASTE DEFORMABILITY............................................................................... 74 Darko Dujmoviü, Boris Androiü, Ivan Lukaþeviü BEAM-TO-COLUMN JOINT MODELLING TO EC3...................................... 89 Darko Dujmoviü, Boris Androiü, Josip Piskovic MODELLING OF JOINT BEHAVIOUR IN STEEL FRAMES ...................... 109 Juraj Božiþeviü, Marijan Andrašec DEVELOPMENT STRATEGY OF ECOINDUSTRAL PARK RAŠA - EIPR ..................................................................................................... 121 Marijan Bošnjak POSIBLE MODEL OF MATHEMATICAL EVALUATION OF BEHAVIOUR DISORDER THERAPY ............................................................ 136

Mario Žagar, Ivica Crnkoviü, Darko Stipaniþev, Maja Štula, Juraj Feljan, Luka Lednicki, Josip Maras, Ana Petriþiü DICES: DISTRIBUTED COMPONENT-BASED EMBEDDED SOFTWARE SYSTEMS ................................................................................... 154 Martin Žagar, Hrvoje Mlinariü, Josip Knezoviü FRAMEWORK FOR 3D MOTION FIELD ESTIMATION AND RECONSTRUCTION........................................................................................ 168 Gaurina-Meÿimurec Nediljka, Pašic Borivoje, Matanoviü Davorin LABORATORY EVALUATION OF MUD DIFFERENTIAL STICKING TENDENCY AND SPOTTING FLUID EFFECTIVENESS ............................ 184 Ružica ýunko, Sanja Ercegoviü Ražiü USE OF PLASMA TECHNOLOGY FOR MODIFICATION OF TEXTILES ................................................................................................... 199 Žaneta Ugarþiü-Hardi IMPORTANCE OF SALT CONTENT REDUCTION IN BAKERY PRODUCTS...................................................................................... 213

Annual 2010/2011 of the Croatian Academy of Engineering

7

Croatian Academy of Engineering is sponsoring scienti¿c meetings, symposia and conferences organized by the Departments and Centers of the Academy. Important events in the Croatian Academy of Engineering in 2010 and 2011 are: 2010 25th Annual Assembly of the Croatian Academy of Engineering was held on March 27th, 2010 where Awards of the Academy for 2009 were granted and the new Statute of the Academy adopted. In accordance with the provisions of the new Statute of the Academy, a new categorization of membership in the Academy was implemented. The most important innovation is the termination of Associate Members of the Academy and the transition of Associates and Members into the unique status of a Full Member of the Academy. Annual 2009 of the Croatian Academy of Engineering with the new Who is Who in the Croatian Academy of Engineering was published. President of the Republic of Croatia Prof. Ivo Josipoviü, Ph.D. visited the Academy on September 15th, 2010. In November 2010 the Academy submitted its comments on the draft of the Law on Science to the Ministry of Science, Education and Sports of the Republic of Croatia.

8

Annual 2010/2011 of the Croatian Academy of Engineering

2011 Croatian Academy of Engineering became a partner in EU SETA - South East Transport Axis. The holder of the project activities from the part of the Academy is its Center for Traf¿c Engineering and distinguished experts from the Department of Transport of the Academy have taken part in the project. On January 14th, 2011 the Academy organized a conference entitled "Engineering Ethics and the Croatian Economy" at the Faculty of Electrical Engineering and Computing in Zagreb. In the organization of the Croatian Academy of Engineering and the Ministry of Science, Education and Sports of the Republic of Croatia, under the auspices of the President of the Republic of Croatia, Central Celebration of the 300th Anniversary of the Birth of Ruÿer Boškoviü was held at the Vatroslav Lisinski Concert Hall on May 17, 2011. On the same day, after the celebration, 26th Annual Assembly of the Croatian Academy of Engineering was held upon which Academy Awards for 2010 were presented. Co-organized by the Biotechnical Center of the Academy, Faculty of Food Technology and Biotechnology of the University of Zagreb and the Centre for Environment Protection and Development of Sustainable Technologies of the Academy, international symposium "The life and Achievements of Prof. Emer. Vera Johanides " was held on September 28th, 2011, and her memorial bust was unveiled in front of the House of the Academy in 28 Kaþiü St., Zagreb. Bulletin of the Academy in Croatian and English "Tehniþke znanosti / Engineering Power vol. 10 (1) 2011 was published. 2012 Bulletin of the Academy in Croatian and English "Tehniþke znanosti / Engineering Power vol. 10 (special issue) 2011 entirely devoted to Ruÿer Boškoviü and 300th anniversary of his birth was published. In March 2012 Croatian Academy of Engineering signed Agreement on Scienti¿c and Technical Cooperation with the Academy of Medical Sciences of Croatia, Croatian Academy of Legal Sciences and the Academy of Forestry Sciences. Organized by the Scienti¿c Committee for Agriculture and Forestry of the Croatian Academy of Sciences and Arts and Biotechnical Center of the Croatian Academy of Engineering, a scienti¿c conference "Food as the Foundation of Health and Longevity" was held on June 18th 2012.

Annual 2010/2011 of the Croatian Academy of Engineering

9

In September 2012 representatives of the Croatian Academy of Engineering met with the delegation of the Chinese Academy of Engineering (CAE) in Zagreb. CAE representatives were: Prof. Pan Yunhe, Executive Vice President of CAE, Prof. Zheng Xiaoguang and Wang Xiaowen, Ph.D., Deputy Directors of the Department for International Cooperation of CAE, and Ms. Zhang Song, Secretary to Prof. Pan Yunhea. On behalf of the Academy Chinese delegation was received by Prof. V. Žiljak, Ph.D.,Vice President, and members of the Academy N. Periü, Dean of the Faculty of Electrical Engineering and Computing, and Prof. M. Cifrek, Ph.D. The meeting was also attended by a group of scientists from the Faculty of Electrical Engineering and Computing.

At the conference “The Future of Printing” held in September of this year Prof. Anayath Rejendrakumar from India gave a comprehensive overview of the state and development of graphics technology in the world. The lecture was held at the Croatian Chamber of Commerce in the presence of many engineers and members of the Academy.

10

Annual 2010/2011 of the Croatian Academy of Engineering

In November 2012 the Department of Systems and Cybernetics and the Centre for Development Studies and Projects of the Academy started "Talks about the Present and Future of Engineering and Biotechnological Sciences," as an activity of the Academy members who would at their monthly meetings reÀect upon important development issues. Information about the events associated with the Academy could be found on our web site: www.hatz.hr

Annual 2010/2011 of the Croatian Academy of Engineering

11

Geodetic Contribution to the Geodynamic Research of the Area of the City of Zagreb B. Pribičević, D. Medak, and A. Đapo University of Zagreb, Faculty of Geodesy email: {bpribic, dmedak, adapo}@geof.hr

ABSTRACT Paper presents 11 year long interdisciplinary research of geodynamic processes of the area of the City of Zagreb. The Geodynamic GPS-Network of the City of Zagreb represents the longest and the most intensive research effort in the ¿eld of geodynamics in Croatia. Since the establishment of the Network in 1997, several series of precise GPS measurements have been conducted on specially stabilized points of Geodynamical Network of City of Zagreb with purpose of investigation of tectonic movements and related seismic activity of the wider area of the City of Zagreb. The Network has been densi¿ed in 2005 in the most active region of northeastern Mount Medvednica. Since then, several GPS campaigns have been conducted. Processing of observation data was done with scienti¿c software GAMIT/ GLOBK, developed by MIT. From this series of GPS measurements geodetic model of tectonic movements has been created. In the area of interest, independent geological investigations have been conducted through even longer period of time which resulted in geological model of tectonic movements. The correlation coef¿cient between geodetic and geologic model has been calculated and shows high degree of correlation thus giving credibility to both methods of research. Systematic analysis has been conducted over geodetic and geologic results giving as a result unique interdisciplinary model of crust movements over wider Zagreb area. This interdisciplinary interpretation of obtained geodetic movements leads to a new scienti¿c insight about geodynamics of the City of Zagreb area. The results of this scienti¿c research will be used to delineate zones of potential earthquake hazard or tectonically caused landslides. Keywords: geodesy, GPS, geodynamics, GAMIT, tectonics Introduction It has been known for a long time that the wider area of Zagreb is geodynamically active, (Prelogoviü, Cvijanoviü 1981) (Kuk et al. 2000), (Kuk et al. 2000a), (Tomljenoviü, 2002) however, there were no data on the actual size of those movements, nor on their spatial orientation. The ¿rst geodetic research in this direction was made through the implementation of the project the “Basic GPS network of the City of Zagreb” (ýoliü et al. 1999) (Medak, Pribiþeviü 2001). Through the realization of the project “Basic GPS-Network of the City of Zagreb” in 1997, Croatian capital got a modern, geodetic foundation of high accuracy. Network was planned as the basis for investigations of tectonic movements and related seismic activity of the wider area of the City of Zagreb. Basic part of the network consists of 43 specially stabilized geodetic points to meet the speci¿c criteria for geodynamic points. After the second series of GPS measurements in year 2001, the network has become “Geodynamic Network of the City of Zagreb” (Medak, Pribiþeviü 2001), (Medak, Pribiþeviü 2002), (Medak,

12

Annual 2010/2011 of the Croatian Academy of Engineering

Pribiþeviü 2003). The City of Zagreb has recognized the importance of this project and GPS campaigns have been performed in years 2004, 2006, 2007 and in 2008 (Pribiþeviü et al 2004), (Medak et al 2007a). After six series of GPS-measurements in period from 1997 to 2008, the analysis of the results with scienti¿c software GAMIT/GLOBK show signi¿cant movement on GPS points as a result of geodynamic activity in the research area (Medak, Pribiþeviü 2004), (Medak, Pribiþeviü 2006), (Pribiþeviü et al. 2007), (Ĉapo 2005), (Ĉapo 2009). From the analysis results the geodetic model of tectonic movements has been created and scienti¿c comparison with geologic model created on the basis of age-long research. The correlation coef¿cient between geodetic and geologic model has been calculated and shows high degree of correlation thus giving credibility to both methods of research. Systematic analysis has been conducted over geodetic and geologic results giving as a result unique interdisciplinary model of crust movements over wider Zagreb area (Ĉapo et al. 2009), (Ĉapo 2009). Geodynamic Network of the City of Zagreb Underlying network has already in its preliminary design devised a dual role: ¿rst it was the basis for the establishment of homogeneous GPS network of the City of Zagreb, while since 2001 it has became the “Geodynamic Network of the City of Zagreb. Such role was possible primarily due to development of GPS technology and achieving high accuracy, which made possible measurement of geodynamic movements (Blewitt 1993), (Bock et al. 1993), (Segall, Davis 1997). This fact now allows geodesy to actively participate together with other professions and disciplines in the important interdisciplinary research for the City of Zagreb. The aim of such a special geodetic network is determination of actual geodynamic movements over a longer period of time with very high accuracy. This research represents a qualitative contribution to the geodetic profession in a variety of interdisciplinary research, such as monitoring of earthquake prediction indicators in Zagreb and its surroundings. Thus, modern geodesy with its high accuracy enters in an entirely new area of research with purpose of determining new parameters microseismic zoning and landslide monitoring in the area of the City of Zagreb. Today, in this type of research, together with geologists, seismologists, geotechnicists, geophysicists and other scientists, their signi¿cant contribution may be given by the geodesists (Altiner 1999.), (Altiner et al. 2001), (Medak et al. 2002), (Grenercy, 2005), (Pinter et al. 2004). It goes without saying that this segment of the research is of great importance because the City of Zagreb has the highest concentration of population and industry in the Republic of Croatia. Therefore, it is certain that even now the city administration should think about measures to be taken in order to preserve human life and property in case of natural disasters such as earthquakes, landslides and similar. Figures 1-7 show cracks and damages on the objects in the Zagreb area due to tectonic movements.

Annual 2010/2011 of the Croatian Academy of Engineering

13

Figure 1: a, b) Cracks in the wall of the church of St. Marie in Granešina with marked direction of the sliding – size of the cracks is between 5-10 cm (c) Example o the cracks parallel to the orientation of local compression stress above south church window; (d) same crack extended to the road between the church and the rectory (Medak et al 2007b)

Figure 2. Gornja Planina, Podizeljska Street 12 – a) crack on the concrete Àoor parallel to the fault from zone Stubica-Kašina b) In the house yard a concrete curb with a crack of 1.5 cm in size, created in just one year. (Ĉapo 2005)

Figure 3. Example of cracks due to movements in the zone of Zagreb fault - Vidovec – cracks 4 cm wide (Ĉapo 2009)

14

Annual 2010/2011 of the Croatian Academy of Engineering

Figure 4. a) Donja Planina, Planinarska Street 21. Whole house has been cut with the crack of NNW-SSE direction. On the facade a signi¿cant horizontal right shift of 2 cm is notable. The facade was made two years prior; b). Donja Planina 13- south facade – cracks with horizontal right movement with size 3cm (Ĉapo 2009).

Figure 5. During 2006 year in the period of a relatively larger displacement amplitudes measured at points Geodynamic network, in fault zone Stubica-Kašina landslides were activated. Particularly activated was the well known landslide in the Planina Gornja. Figure shows the cracking of the road (Ĉapo 2009).

Figure 6. a) In Gornji Mikuliüi on the route of main fault of the Zagreb zone the landslide has been reactivated in 2008. b) Particularly noticeable are shifts of the ground north of the road, where considerable damage to fences has been recorded (Ĉapo 2009).

Annual 2010/2011 of the Croatian Academy of Engineering

15

Figure 7: In the spring of 2008 a new landslide in Vidovec was created. Landslide is located within the route of the main faults of the Zagreb fault zone. In the overlying wing of the cliffs there is a fault in the relief. Due to tectonic activity there was an escarpment destabilization and the activation of landslides (Ĉapo 2009).

Design and the stabilization of the Geodynamic Network Given the size of the City of Zagreb and the need to encompass the wider area, the Geodynamic network covers an area of app. 700 km2. The main characteristic of this mostly raster formatted network is that the distance between points is approximately seven kilometers in a sparsely populated area, and in urban areas of the City, density is slightly higher. Namely, given the fact that the City of Zagreb is seismically very active, the Network is designed so that the recent structural fabric of the Zagreb area is optimally covered. Network points are placed in relation to fault zones to positions that will show the truest geodynamic movements in the research area. Figure 8 shows the placement of the points in the Zagreb area and positions of main faults and epicenters of largest earthquakes in the area.

Figure 8. Locations of the Geodynamic network points and positions of main faults and epicenters of largest earthquakes in the area (Ĉapo 2005)

16

Annual 2010/2011 of the Croatian Academy of Engineering

Thanks to the signi¿cant increase in accuracy, dependence of geodetic data on time (the fourth dimension) is apparent. That is why special attention has been paid to the stabilization of the points of the geodynamic network. The aim was to ensure the stability of points over a longer period of time and provide accurate reoccupation of GPS antenna. In this way determination of actual geodynamic displacements was enabled with very high accuracy over a longer period (Solariü, 1999), (Schmitt, 1985), (Gerasimenko et al. 2000). Works on the stabilization of points of the Geodynamic network of the City of Zagreb were performed according to the design created by the experts from the University of Zagreb, Faculty of Geodesy, with the interdisciplinary help of engineers, geologists and seismologists. The Geodynamic network consists of more than 40 specially stabilized points on wider area of the City of Zagreb and covers area of about 700km2 (ýoliü et al 1999), (Medak, Pribiþeviü 2002). In total there were 33 pillars built, 32 of which come with the above ground stabilization, and only one with the sub terrain stabilization (1028 King Tomislav Square). The remaining 7 points of the Geodynamic network has a different stabilization, which also meets all the stability requirements that are set before this special network (Medak, Pribiþeviü 2002) (Pribiþeviü et al 2007). As shown on the ¿gure 1, most of the monuments have pilots that go up to 14 meters deep to consolidated ground. On the top of every monument is a steel mark with winding for the special extension thus insuring stability of points through long periods of time and precise GPS antenna reoccupation.

Figure 9. Specially stabilized monument on Geodynamic network of the City of Zagreb

After seven years of intensive research, it became clear that the Geodynamic network must be densi¿ed in the seismotectonically most active area. Densifying is of particular importance for the preliminary assessment of the landslide susceptibility in those urban areas that lie on the clay layers in the northern part of the zone: Šestine Grmošica, Granešina, Mikuliüi, Vugrovec and Kašina. In year 2005, ¿ve new geodynamic points were established in that part of Zagreb (Medak et al. 2007a).

Annual 2010/2011 of the Croatian Academy of Engineering

17

GPS-observations Since 1997 GPS measurements on the Geodynamic network of the City of Zagreb were conducted through GPS measurement campaigns every few years. Each campaign consists of two to three 24 hour sessions. First two campaigns in 1997 and 2001 had tree 24 hour sessions with 15 second observation interval, thus giving 5670 epochs. All later campaigns had 24 hour sessions but with 30 second interval, giving 2880 epochs (Medak et al 2007a) (Ĉapo 2009).

Figure 10. Distribution of points according to the observation sessions (2004, 2006, 2008)

All conducted campaigns had only Trimble GPS receivers and antennas. There were 8 conducted GPS campaigns: 1997, 2001, 2003, 2004, 2005, 2006, 2007 and 2008. Only ¿ve of those were observation of complete Geodynamic network: 1997, 2001, 2004, 2006, and 2008 (41 points). Two campaigns were conducted for observing densi¿cation points 2005, 2007 (11 and 21 points) (Pribiþeviü et al. 2007) (Ĉapo et al. 2009). Table 1 shows all conducted campaigns including number of points and used instruments Table 1: GPS campaigns on Geodynamic network of the City of Zagreb

Campaign

Date

sessions

points

receivers

Zagreb 1997

27.-29.10.1997.

2

43

27

Zagreb 2001

25.-28.06.2001.

3

40

16

Zagreb 2003

22.-23.06.2003.

1

13

13

Zagreb 2004

17.-20.06.2004.

3

39

13

Zagreb 2005

10.-11.09.2005.

1

11

11

Zagreb 2006

22.-25.06.2006.

3

41

13

Zagreb 2007

13.-15.07.2007.

2

21

13

Zagreb 2008

10.-13.06.2008.

3

41

13

18

Annual 2010/2011 of the Croatian Academy of Engineering

Figure 11: Velocity rates and their error ellipses for period 2006-2007 (Ĉapo 2009), (Wessel 2004) Table 2: Statistical representation of absolute values of velocity rates for 2006–2007 in mm/yr.

min. max. avg.

vij

vȜ

vhz mm/yr

vH mm/yr

0,2 10,70 3,14

0,57 19,36 4,86

1,43 19,37 6,43

0,60 50,27 16,25

Velocity rates calculated for the complete period from 1997 to 2008 are much smaller due to different active zones of the Zagreb area from year to year giving thus sometimes opposite velocity directions for the same points. Velocity rates for the period 1997 to 2008 are shown in Figure 12 and Table 3 shows statistical representation of absolute values of velocity rates for period 1997–2008 in mm/yr. Table 3: Statistical representation of absolute values of velocity rates for 1997–2008 in mm/yr.

min. max. avg.

vij

vȜ

0,03 3,93 0,83

0,04 3,42 1,03

vhz mm/yr 0,12 4,34 1,45

vH mm/yr 0,02 17,48 1,97

Annual 2010/2011 of the Croatian Academy of Engineering

19

Figure 13: Geodetic velocity model created using IDW interpolation with respect to faults model

(1)

Table 4: Correlation coef¿cient by Spearman formula

Calculating of correlation coef¿cient by above given formula (1) for Spearman rank correlation we get rs=0.94 with 1% signi¿cance level (or 99% in Gaussian model). The correlation coef¿cient between geodetic and geologic model shows high degree of correlation thus giving credibility to both methods of research (Ĉapo 2009).

20

Annual 2010/2011 of the Croatian Academy of Engineering

CONCLUSION Geodynamic study of the City of Zagreb has started in 1997 and since 2008 there has been seven GPS campaigns for the purpose of determining geodynamic movements on the points of the Geodynamic network of the City of Zagreb. Through the project Geodynamic study of the City of Zagreb, based on independent multiple precise geodetic and geological measurements, original models of tectonic movements on Zagreb area have been created. Geodetic model is based on precise GPS satellite measurements, and geological model is based on the long-term geologic measurements and studies. For this purpose data collected through GPS measurements on seven GPS measurement campaigns in the period since 1997 until 2008. Year has been used. GPS measurements were processed in the scienti¿c GAMIT/GLOBK software, designed at MIT speci¿cally for processing GPS measurements on the geodynamic network. It calculated the geodynamic movements on the points of the Geodynamic network using modern methods of Kalman ¿ltering. The result of the velocity models at the points Geodynamic network were obtained for the period 1997- 2001, 2001-2004, 2004-2006, 2006-2007, 2007-2008, 2006-2007-2008 and cumulatively for the whole period from 1997-2008. The maximum absolute value of displacement for the total solution in the horizontal direction is 4.3mm/yr in vertical direction 17.5mm/yr. The largest absolute values of displacements were obtained during the period 2006-2007 and are amounted to 19.4mm/yr and 50.3mm/yr in the horizontal and vertical direction respectively. It is evident that it represents signi¿cant geodynamic movements. However, it should be noted that the mean value of velocity is 3 mm/yr. Velocity model of geodynamic movements for the whole period of the study has been created using IDW interpolation method with the inclusion of fault models of the Zagreb area. In that way the original geodetic model of the velocity ¿eld for the wider area of the City of Zagreb. Geostatistical analysis has been performed and the correlation coef¿cient has been calculated for the geodetic and geological model of the geodynamic movements, by using Spearman correlation coef¿cient formula. The resulting value of 0.94 with signi¿cance level of 1% (or 99% in Gaussian model) indicates a high degree of correlation between the geodetic and geological model of the geodynamic movements in the subject area. This proves the credibility of eleven year long research independently conducted by geological and geodetic methods. In conclusion, it can be stated that as part of this research a unique interdisciplinary model of crust movements over wider Zagreb area has been created for the ¿rst time. It can be applied to precisely de¿ne the boundaries of seismic micro-zoning of the City of Zagreb. On the other hand, results of the study can help in decision making process about reconstruction and adaptation of important structures (reinforcement of foundations and structural elements) and also they can be used to more accurately de¿ne the zones of landslides caused by tectonic movements. REFERENCES 1. 2.

Altiner, Y. (1999). Analytical Surface Deformation Theory for Detection of the Earth’s Crust Movements. Springer Verlag. Altiner, Y., Marjanoviü-Kavanagh, R., Medak, D., Mediü, Z., Prelogoviü, E., Pribiþeviü, B., et al. (2001). Is Adria a Promontory or does it exist as an Independent Microplate? In J. Sledzinski (Ed.), Proceedings of the EDS G9 Symposium “Geodetic and Geodynamic Programmes of the CEI (Central European Initiative), 25-30 March 2001.

Annual 2010/2011 of the Croatian Academy of Engineering

3.

4. 5.

6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16.

17. 18.

19.

21

Reports on Geodesy (p. 225-229). Nice, France: Warsaw University of Technology. Blewitt G. (1993): Advances in Global Positioning System technology for geodynamics investigations: 1978–1992. In Contrib. Space Geodesy Geodyn.: Technol. Geodyn., Geodyn. Ser., ed. DE Smith, DL Turcotte, 25: 195–213. Washington, DC: Am. Geophys. Union. 213 pp. Bock Y., D., Agnew, Fang, P., Genrich, J., Hager, B., Herring, T., Hudnut, K., et al. (1993). Detection of crustal deformation from the Landers earthquake sequence using continuous geodetic measurements. Nature, 361 , 337-340 ýoliü, K., Prelogoviü, E., Pribiþeviü, B., Švehla, D. (1999). Hrvatski geodinamiþki projekt CRODYN i GPS mreža Grada Zagreba. In A. Bajiü (Ed.), Znanstveni skup Andrija Mohoroviþiü - 140. obljetnica roÿenja: zbornik radova (p. 141-152). Zagreb: Državni hidrometeorološki zavod. Dong, D., Herring, T., & King, R. (1998). Estimating regional deformation from a combination of space and terrestrial geodetic data. Journal of Geodesy, 72 (4), 200214. Ĉapo, A. (2005). Obrada i interpretacija geodetskih mjerenja na geodinamiþkoj mreži Grada Zagreba. magistarski rad. Zagreb: Geodetski fakultet, 12.05. 2005., 174 str. Voditelj: Medak, Damir. Ĉapo, A (2009): Korelacija geodetskog i geološkog modela tektonskih pomaka na primjeru šireg podruþja Grada Zagreba. doktorska disertacija. Zagreb: Geodetski fakultet, 08.05. 2009, 199 str. Voditelj: Pribiþeviü, Boško. Ĉapo, Almin; Pribiþeviü, Boško; Medak, Damir; Prelogoviü, Eduard (2009): Correlation between Geodetic and Geological Models in the Geodynamic Network of the City Of Zagreb. Reports on geodesy. 86 (2009), 1; p 115-122. Grenerczy, G., Sella, G., Stein, S., Kenyeres, A. (2005). Tectonic implications of the GPS velocity ¿eld in the northern Adriatic region. Geophys. Res. Lett., 32 . Gerasimenko, M. D., Shestakov, N. V., Teruyuki, K. (2000). On optimal geodetic network design for fault-mechanics studies (Vol. 52). Earth Planets Space. Herring, T., King, R., McClusky, S. (2006a). Documentation for the MIT Global Kalman ¿lter VLBI and GPS analysis program: GLOBK 10.3. Cambridge. Herring, T., King, R., McClusky, S. (2006b). Documentation for the MIT GPS analysis software: GAMIT 10.3. Cambridge. Kuk, V., Prelogoviü, E., Dragiþeviü, I. (2000). Seismotectonically Active Zones in the Dinarides. Geol. Croatica, 53 (2), 295-303. Kuk, V., Prelogoviü, E., Soviü, L., Kuk, K., Mariü, K. (2000a). Seizmološke i seizmotektonske znaþajke šireg zagrebaþkog podruþja. Graÿevinar, 52 (11), 647-653. Medak, D., Pribiþeviü, B. (2001). Geodynamic GPS-Network of the City of Zagreb - First Results. In The Stephan Mueller topical conference of the European Geophysical Society: Quantitative neotectonic and seismic hazard assessment: new integrated approaches for environmental management. Balatonfüred, Hungary. Medak, D., Pribiþeviü, B. (2002). Geodinamiþka mreža Grada Zagreba. In T. Bašiü (Ed.), Zbornik Geodetskog fakulteta Sveuþilišta u Zagrebu povodom 40. obljetnice samostalnog djelovanja 1962-2002. pp. 145-156. Zagreb. Medak, D., Pribiþeviü, B., Prelogoviü, E. (2002). Determination of the recent structural fabric in the Alps-Dinarides area by combination of geodetic and geologic methods. In M. Brilly (Ed.), Raziskave s podroþja geodezije in geo¿zike. p. 57-64. Ljubljana: Slovensko združenje za geodezijo in geo¿ziko. Medak, D., Pribiþeviü, B. (2003). Processing of Geodynamic GPSNetworks with GAMIT Software. Reports on Geodesy, Warsaw University of Technology, 64 (1), 75-84.

22

Annual 2010/2011 of the Croatian Academy of Engineering

20. Medak, D., Pribiþeviü, B. (2004). Processing of Geodynamic GPS networks in Croatia with GAMIT Software. In N. Pinter, G. Grenerczy, J. Webber, S. Stein, & D. Medak (Eds.), The Adria Microplate: GPS Geodesy, Tectonics and Hazards (Vol. 61, p. 247256). Veszprem, Hungary: Springer. 21. Medak, D., Pribiþeviü, B. (2006): Processing of geodynamic GPS-networks in Croatia with GAMIT software, The Adria Microplate, GPS Geodesy, Tectonics and Hazards / Pinter, Nicholas et al. (eds). 22. Medak, D., Pribiþeviü, B., Ĉapo, A. (2007a). Progušüenje toþaka Geodinamiþke mreže Grada Zagreba u podsljemenskoj zoni. Geodetski list , 61(84)(4), 247-258. 23. Medak, D., Pribiþeviü, B., Prelogoviü, E., Ĉapo, A. (2007b). Primjene geodetskogeodinamiþkih GPS-mjerenja za monitoring tektonski uvjetovanih klizišta. In Simpozij o inženjerskoj geodeziji (p. 229-241). Beli Manastir. 24. Pinter, N, Grenerczy, G., Webber, J., Stein, S., Medak, D. (Eds.) (2004), The Adria Microplate: GPS Geodesy, Tectonics and Hazards .Vol. 61. Veszprem, Hungary: Springer. 25. Prelogoviü, E., Cvijanoviü, D. (1981). Potres u Medvednici 1880. godine. Geol. vjesnik (34), 137-146. 26. Pribiþeviü, B., Medak, D., Prelogoviü, E. (2004). Geodinamika prostora Grada Zagreba. Geodetski list , 58(81)(1), 51-65. 27. Pribiþeviü, B., Medak, D., Prelogoviü, E., Ĉapo, A. (2007). Geodinamika prostora Grada Zagreba. Zagreb: Geodetski fakultet Sveuþilište u Zagrebu. Znanstvena monogra¿ja. 28. Schmitt, G. (1985). Review of network design: Criteria, risk functions, design ordering. In Grafarend & F. Sanco (Eds.), Optimization and design of geodetic network (p. 610). Berlin etc.: Springer. 29. Segall, P., Davis, L. J. (1997): GPS Applications for Geodynamics and Earthquake Studies. Annual Review of Earth and Planetary Sciences. Vol. 25: 301-336. May 1997. 30. Solariü, M. (1999). Suradnja srednje europskih zemalja u geodeziji i geodinamici. In A. Bajiü (Ed.), Znanstveni skup Andrija Mohoroviþiü - 140. obljetnica roÿenja: zbornik radova (p. 165-177). Zagreb: Državni hidrometeorološki zavod. 31. Šošiü, I. (2004). Primijenjena statistika. Zagreb: Školska knjiga, Udžbenici Sveuþilišta u Zagrebu. 32. Tomljenoviü, B. (2002). Strukturne znaþajke Medvednice i Samoborskog gorja. Doktorska disertacija. Rudarsko-geološko-naftni fakultet, Zagreb. 33. Wessel, P., Smith, W. H. F. (2004). The Generic Mapping Tools – Technical Reference and Cookbook (4th ed.).

Annual 2010/2011 of the Croatian Academy of Engineering

23

Spanish Broom (Spartium Junceum L.) Drago Katović, Ph.D. University of Zagreb Faculty of Textile Technology Zagreb, Prilaz baruna Filipovića 28A e-mail: [email protected]

Andrea Katović PhD University of Calabria, Faculty of Engineering Department of Chemical Engineering and Materials Via P.Bucci, Cubo 44 A, Rende (CS), Italy (e-mail: [email protected])

Marija Krnčević B.Sc. of Ethnology and Archeology, Museum of Šibenik, Gradska vrata 3 [email protected])

LA GINESTRA O IL FIORE DEL DESERTO Qui su l’arida schiena Del formidabil monte Sterminator Vesevo, La qual null’altro allegra arbor né ¿ore Tuoi cespi solitari intorno spargi Odorata ginestra, Contenta dei deserti … Giacomo Leopardi

Abstract The Spanish Broom (Spartium Junceum L.) plant is almost a forgotten textile raw material. This paper presents a review of procedures practiced in the past and nowadays for obtaining ¿bers from this plant. It discusses new discoveries about its use as a component of special composite ¿bers, as well as its possible bene¿ts in households of poorer limestone areas. It also reports on come chemical and physical properties of the ¿bers extracted from local plants. Key words: Spanish Broom, maceration, composites, ash, Klason lignin, cellulose, hemcellulose, FTIR, SEM, microwave, composites Introduction Whilst Àax and hemp have mostly been used as textile raw material of cellulosic origin

24

Annual 2010/2011 of the Croatian Academy of Engineering

in plains, in coastal areas of the Mediterranean wild Spanish Broom has been used as raw textile material since ancient times. However, Spanish Broom (Spartium Junceum L.) (Croatian: žuka or brnistra; Italian: ginestra) is almost completely forgotten today, and it is only sporadically mentioned as raw textile material. The habitat of Spanish Broom is the Mediterranean area of south Europe, south-west Asia and north-west Africa. In Italy, in the Mediterranean area of olive groves, it climbs the altitude of 975 meters. In Turkey, Syria and Palestine, it reaches altitudes of 1,700 meters. It is regarded weed in the USA and New Zealand; where there is a tendency to eradicate it in order to save indigenous plants.

Figure 1 Spartium Junceum L.

There are numerous anatomic adjustments to dry soil visible in microscopic structures of Spanish Broom’s vegetative organs. One of them is xerophtytic (xerophytes – types of plants adjusted to dry climate) adjustment of the leaf during its short life span and transformation of its internal structure dominated by the chain of parenchyma. The upper part of the stem took over the function of leaf, whereas sclerenchyma’s ¿bers and conductive elements take over the majority part of its secondary units. An unusual feature of the root’s primary structure is its underdeveloped endoderm, and the secondary structure points to its storage and mechanical roles. [1]. Morphological structures of Spanish Broom, more precisely its longitudinal and cross-section of shoots were recorded with a FE-SEM (Field Emission-Scanning Electron Microscope) of the company TESCAN, with a steamer and unit for EDX analysis (Energy Dispersive X-Ray Analysis). Before the microscopic recording, samples were processed for 180 seconds in a steamer with gold and palladium, using operative voltage of 5-15 kV. The results of images show that Spanish Broom’s shoots have two basic layers: rigid and woody inner layer, porous in the middle part, making the plant light, and an outer layer, skin intertwined with sinewy ¿bers. The cross-section of

Annual 2010/2011 of the Croatian Academy of Engineering

25

Spanish Broom also points to epidermis, parenchyma of the bark with chlorophyll, endoderm, bundles of root ¿bers and bundles of pericyle ¿bers. Head shaped units of root and pericycle ¿bers, connected in an almost continuous ring, are also visible.

Figure 2 SEM micrograph of cross-section of Spanish Broom vermenes (SEM) Hitachi S-3400 N

Cross-section of the woody shoot of Spanish Broom, which give rigidness to the shoot, and enables passage of cell juices alongside the shoot

Figure 3 SEM micrograph of Cross-section of the woody (lignin) part

26

Annual 2010/2011 of the Croatian Academy of Engineering

Figure 4 SEM micrograph of Cross-section of ¿bers of Spanish Broom

The cross-section of technical ¿bers shows their irregular shape, with visible regular elementary circular ¿bers in some places.

Figure 5 Longitudinal micrograph SEM of Spanish Broom technical ¿bers with visible elementary ¿bers (SEM FE/Mira Tescan);

Annual 2010/2011 of the Croatian Academy of Engineering

27

Figure 6 Elementary ¿ber of Spanish Broom, recorded with scanning electron microscope (SEM FE/Mira Tescan);

THE APPLICATION OF SPANISH BROOM IN HISTORY Spanish Broom is a plant that was known even to the ancient Romans. In the past, Greeks, Romans and Carthaginians used it as raw material for the manufacture of ropes, nets, bags, sails. They also used it for covering roofs and even clothing. The Àower of the Spanish Broom is important for the coastal apiculture. Aristotle and Pliny praised the honey produced from Spanish Broom. In general, out of all ancient writers about Spanish Broom, it was Pliny Senior who wrote the most. He called it genista and he writes: Ginestra quoque vinculi,… Àores apibus gratissimi (Spanish Broom is used for tying and making crumples,… and bees often land on its Àowers). However, in several places he mentions a plant called Spartum (Latin form of the Dioskorides’ description of the plant sparton). After a detailed analysis of the application and processing, it is regarded the same plant today, or rather (Spartium Junceum L.). Already in the period of the ancient Rome, Spanish Broom ¿elds (genestium – or ager in later Latin) were cultivated as any other ¿eld, and the plant or sprouts were sown in ploughed up furrows, as the Roman agronomist Columella describes. According to Vergil, hedges of Spanish Broom were planted in addition to willow, hazel, elder and other plants. Pliny wrote that the sowing and planting of Spanish Broom was crucial for peasants. Amongst other things, its branches provide excellent material for tying vine and young trees. The most intensive production of Spanish Broom as raw material for obtaining textile fabric was in Italy during the 1930’s, when annual production of Spanish Broom, due to economic sanctions in Italy (war in Abyssinia) and the proclamation policy of self-suf¿ciency, reached 700,000 tones. It was planted on 300,000 hectares and there were 61 processing factories [2].

28

Annual 2010/2011 of the Croatian Academy of Engineering

However, the industrial processing of the textile raw material completely extinguished in the mid 20th century. There were only several associations of Spanish Broom producers left in Europe that obtained subsidized EU projects with the goal of maintaining and improving old crafts. However, during the past 10 years there has been a growing interest in Spanish Broom as raw material in the production of composite materials. Findings of Liburnian ships, named Serilias, dating from the 2nd century, prove that Spanish Broom was used quite early in our parts of the Adriatic. The special feature of this type of ship is the way parts of its structure were connected by sewing individual elements. All dilemmas surrounding this type of ship building were resolved by ancient writers, who also transposed the name of the vessel constructed using this sewing technique. Ancient writers left us with data about the special type of vessel used by Liburnians and Histrians and their construction techniques. Marcus Verrius Flacus, grammarian from the Augustan period, gathered grammatically interesting, but faintly familiar words in his work De verborum signi¿catu. He also mentions the name serilia for Istrian and Liburnian ships. This record came to us through Sextus Pompeious Festus, who lived in the 200 AD. He writes: According to Verrius, Serilia is the name for Istrian and Liburnian vessels, compressed with Àax and Spanish Broom ropes. Its name thus derives from the Latin word conserere (string together) and contexere (to weave). In his play Niptra, Pacuvius writes: No wooden pin hold the ship’s haul together. It is rather sewn using Àax and Spanish Broom ropes. He used a descriptive expression and invented word for ropes spun from Spanish Broom. [1]. In this area, Spanish Broom was used throughout the Middle Ages, according to several historical records. The oldest record, even though it only mentions Spanish Broom, does point to its processing for textile purposes. This data is recorded in the collection of laws and regulations Statuta et leges civitatis et insulae Curzlae (1214 – 1558). According to chapter XCVIII, it was prohibited to gather on moles of Zavalatica, Prigradica and Blace, since they were probably places where women came to soak Àax and Spanish Broom. Provisions for the protection of useful plants were prescribed quite often. One of the examples is the Statute of Braþ from 1656[3]. The next preserved record on Spanish Broom was written in Glagolitic alphabet and dates from 1674. It describes the area of Šibenik. It is a testament and reads: Ostavlan Matii svekrivi jednu kanicu novu i brnistru, ka j(e) g(o)tova. It is not quite clear what does ¿nished Spanish Broom actually refer to. In the book Viaggio in Dalmazia, written by the famous Italian travel writer Alberto Fortis in 1774, he describes his visit to the island of Murter including the processing and use of Spanish Broom: “The favorite craft of Betinjans, population of the western part of the island, is picking, soaking and weaving of Spanish Broom which they ¿nd on the coast of Istria and islands of Kvarner. They make fabric from it, which they use for making sacks and sometimes shirts and skirts for peasants. There is no doubt this plant could produce far better artifacts if processed in a more re¿ned manner. They use the sea to soak the bundles.” The cultivation of Spanish Broom is unfamiliar today in our parts. However, it was widely spread in the past. The shortage of textile in 1916 motivated Friar Andrija Rajkoviü to use the people’s experience in Spanish Broom processing and use it to a greater extent, even in industry. He explained to authorities at that time the need, bene¿t and possibilities of processing Spanish Broom, sending several samples of ready-made fabric. In the same memo he writes: For the past 40 or 50 years, Spanish Broom has been widely used, but now manufacturing of this type of fabric is completely neglected. He also attached a guide How

Annual 2010/2011 of the Croatian Academy of Engineering

29

to process Spanish broom fabric. [3] During the ‘30s of the past century, foresters widely recommended the cultivation of Spanish Broom (Spartium junceum) in the coastal karst areas, more precisely M. Aniü Ph. D in Forest Journal in 1937 and engineer Stane Benko in his booklet on Spanish Broom, published by the Ministry of Industry and Mining, in Zagreb, 1946. Member of the Institute for Forestry Research Ante Premužiü explains the possibility of using Spanish Broom for economic purposes in his report Systematic cultivation of Spanish broom in our karst areas, published in the Forest Journal. He wrote that diverse biological properties could be put to good use for forestation of poor soil of the coastal limestone area since Spanish Broom grows on clay and limestone soil of neutral, acid and alkali reaction. Positive sides of its cultivation were highly praised, giving these studies a more propaganda character. Alongside the coast of east Adriatic, people used and processed Spanish Broom in various ways, but it was also harvested in Istria and the coastal area of Dubrovnik in the period of Austrian-Hungarian Monarchy, and later between the two World Wars. It was then harvested by the army, tied up in bundles and exported to Italy for manufacturing military suits. There were many attempts of industrial processing in our country. Let us remember the previously mentioned initiative of Rajkoviü, which spread across the entire Adriatic, gaining the status of industrial processing. Immediately after the World War I in 1919, there was an attempt in Omiš to process Spanish Broom in bulk. A factory was also built for its processing. However, the ¿rst attempt failed since Spanish Broom had been soaked in the sea for six months, thereby rotting. There were no further attempts from then on. Only after the end of World War II, Spanish Broom was rediscovered, mostly due to a shortage of textile raw material during the post-war period. The processing actually started two years after the W. W. II (1947). Around two trucks of Spanish Broom, harvested near Rabac and Labin (Istria), were turned over to the district of Pula to send it for processing in the factory in Fažana. It was also decided to build factories in Vodice and Zakuþac near Omiš. There was a second attempt to revive the production, this time using state-of-the art methods, quite near where the ¿rst plant had been constructed in 1919. The process of maceration was chemical, and a device for hot-air drying was used. In the paper factory in Rijeka, a department was turned into a laboratory for testing. In order to save on transport of raw material, someone suggested turning an old ship into a Àoating factory, which would cruise in the Adriatic and process Spanish Broom, quite expensive to harvest. The results of this test production were taken over for further processing by certain spinning mills in Croatia and Slovenia, and so ¿rst test products were actually produced. However, the process stops there. The whole initiative, which had started positively, stopped abruptly, allegedly due to high purchasing price, and so traditional processing remained in the hands of the people. [3]. There were other attempts to process Spanish Broom industrially. In Vodice, between 1946 and 1949, a factory was built for the processing of Spanish Broom. It worked until 1954, when it was shut down due to unpro¿tability. The same factory is scheduled for demolition today. According to former workers of the factory (Tvornica Žuke) in Vodice the technological procedure for obtaining ¿bbers was as follows: branches of Spanish Broom with their shoots were soaked in sodium alkali for several days. After resting in pools with the alkali, ¿bers would separate from the other parts. Bundles of Spanish Broom would be subject to water jets under pressure of four to six atmospheres. The water jets would separate ¿b-

30

Annual 2010/2011 of the Croatian Academy of Engineering

ers from the connective tissue, transferring them to pools for rinsing off the alkali. After air drying, periodic exposure to rain would have a positive effect on the quality. Fibers in the shape of hemp were then mechanically cleansed from any eventual remainders of the connective tissue. They were pressed in bales and sent for further processing in the factory Duga Resa. The resulting technical ¿bers of Spanish Broom were used for producing ropes. The original use of Spanish Broom can be found in Petriþani near Zadar. Freshly cut shoots were chopped and beaten into powder. This powder was sprinkled on a wet mesh, and the ingredients which it would absorb, would protect it against rotting. Spanish Broom shoots were also used for manufacturing smaller-sized coops. [3] Several associations of Spanish Broom producers remain in Europe today. The European Union allocated subsidies to Italy with a goal of preserving and developing old trades. There is also a craft shop which produces sails made of Spanish Broom for the reconstruction of old boats. In Pirovac, the Porart Association organizes traditional harvesting and maceration of Spanish Broom, once practiced in our parts. (Figure 7-9).

Figure 7 Harvesting of Spanish Broom

Figure 8 Tying shoots in bundles - fašine

Annual 2010/2011 of the Croatian Academy of Engineering

31

Figure 9 Maceration of Spanish Broom in the sea

There are several items made from Spanish Broom in our museums, or rather its combination of other ¿bers. The sheet – lancun – from the Island of Pag (Jakišnica), 207x160 cm, was weaved in four threads in approximately 1945. It consists of two parts and it was washed and whitened by damping in dew and sun exposure. It is in good shape, with a few smears.

Figure 10 Lancun (sheet) made of Spanish Broom from Pag (collection of the Museum of Ethnography in Split)

Figure 11 Sheet (collection of the Town Museum of Šibenik)

32

Annual 2010/2011 of the Croatian Academy of Engineering

Cover from the Island of Capri (local name: klašnje), 130x80 cm. It is mostly made of Spanish Broom, whilst old pieces of cloth were used for weft. They were of uniformed size, depending on the width of the weaving loom.

Figure 12 Cover (basically Spanish Broom) from Capri (collection of the Town Museum of Šibenik)

The Faculty of Textile Technology, within the framework of the Textile Science Research Center, analyzed a sample of a fuse from a Roman ceramic oil lamp found in Split. The oil lamp dates from the 3rd or 4th century AD. Considering this item is so rare, we were interested from which material the fuse was made of.

Figure 13 Roman ceramic oil lamp (lucerna)

For the purpose of the fuse analysis, several images were taken with a scanning electron microscope (SEM) of the company Tescan Mira/FE .

Annual 2010/2011 of the Croatian Academy of Engineering

Figure 14 - 20

SEM images of the fuse (variations in zoom)

33

34

Annual 2010/2011 of the Croatian Academy of Engineering

The images show ¿bers covered with ashes, but not completely burned (grey part of the sample). Otherwise, they would not be identi¿able with precision. There are also black clumps of dirt (larger shapeless pieces) For the purpose of analyzing the fuse FTIR (Fourier Transform Infra Red) specters of ¿bers were also recorded, which might have been used in the 3rd or 4th century AD in our parts: cotton, Àax, hemp and Spanish Broom. The resulting curves of the mentioned ¿bers were compared with the spectrum of the fuse sample.

Figure 21 Fuse spectrum

After attempts of separating parts of the fuse (grey part) from the black part, which mostly contains dirt, the difference in FTIR specters was evident, although insuf¿cient for a de¿nite conclusion. The differences, which relate to a total of spikes, double spike at wavelength ca. 2900 cm-1, and at 1700 cm-1 (arc) and 1540 cm-1 (spike), represent the existence of non-cellulose natural ¿bers (wax, pectin, lignin). At ¿rst glance that would mean it is not cotton or Spanish Broom. However, this would be acceptable only if we take into consideration special methods for ¿ber processing, practiced after the 19th century. It was then assumed, based on the color and SEM images, that the fuse was actually a remainder of incomplete burning process. FTIR ¿ber specters were then recorded after burning, and it was concluded it was probably not cotton. THE APPLICATION OF SPANISH BROOM IN FUTURE In the past, natural ¿bers were not taken into account as reinforcements for polymeric materials because of some problems associated with their use. Lack of good interfacial adhesion, low degradation temperature, properties variability depending on the quality of harvest, age and body of the plant, as well poor resistance against moisture make the use of natural ¿ber reinforced composites less attractive than synthetic. However, the production of composites reinforced with synthetic ¿bers and matrices (glass, carbon, aramid) requires a large amount of energy, and they are dif¿cult to recycle or they are completely unrecyclable. Moreover, natural ¿bers in cases of ¿re decrease the content of toxic gases of combustion. [10]. For the past several years there has been an increase in the number of scienti¿c studies

Annual 2010/2011 of the Croatian Academy of Engineering

35

relating to the production of Spanish Broom as raw material for composite materials [68]. The advantage of Spanish Broom over Àax and hemp is that it can grow in the most unfavorable limestone soil; it is resilient to draught; and once planted it can be used during a period of up to twenty years, whilst Àax and hemp demand high quality soil each year. Its cultivation on rocky areas would have several bene¿ts. For the purpose of analyzing the possibility of using Spanish Broom as composite material, inborn samples were harvested in the surroundings of Šibenik. Freshly picked shoots contained up to 35% of moisture, compared with the mass of samples dried in standard conditions. Samples were tied in bundles, soaked in sea water for 21 days and then ¿bers were separated from the remaining woody part, using a mechanical procedure. This traditional procedure of extracting Spanish Broom was compared with several contemporary processes, using NaOH, with subsequent thermal processing at 130 °C and rapid decompression. [9]. Processing would last for approximately six hours, which is signi¿cantly less compared with water processing between 20 and 80 days. According to the latest studies of the Faculty of Textile Technology – Institute for Textile Technology and Ecology – the possibility of microwave processing is currently undergoing research. According to present and promising results, total maceration period was reduced to 10 minutes, with a signi¿cant decrease in the use of applied chemicals.

Figure 22 Energy for production of some ¿bers

Energy for production of Spanish Broom depends on the type of applied maceration and ranges within the framework of other natural ¿bers, according to our research.

Tensile strength (MPa)

Cotton 300-700

Flax 500-900

Hemp 310-750

Spanish broom 750

Figure 23 Some natural ¿bers properties

Spanish Broom ¿bers were extracted from the plant’s branches harvested in Dalmatia (Šibenik area) which contained 35 % of moisture compared with branches dried under

36

Annual 2010/2011 of the Croatian Academy of Engineering

standard conditions. The IR spectra of the ¿ber sample obtained after alkaline extraction was recorded on the ATR-FTIR spectrometer (Spectrum 100, Perkin Elmer), while cotton, Àax and hemp ¿bers were taken as reference samples. From the recorded FT-IR spectra of the ¿ber samples it is visible that all ¿bers have peak characteristic of cellulose although they present some differences (Fig. 24.) The intensity of signal at 1734 cm-1 corresponding to the C=C esther band related to pectin is higher in Àax than in hemp ¿bers while this band is not found in Spanish Broom ¿ber sample. On the contrary, only a weak peak at ca. 1500 cm-1 corresponding to the C=C in-plane aromatic vibrations from lignin can be observed in the case of Spanish Broom ¿bers, providing an ulterior evidence the applied chemical treatment was adequate for the almost complete removal of non-cellulosic compounds from the Broom ¿bers. On the other hand, the two little sharp peaks observed over a broader peak in the area of 2850 to 2950 cm-1, attributed to the CH2 and CH groups of long alkyl chains of waxes, are present in the spectra of Àax and hemp.

Figure 24 FTIR (Fourier Transform Infra Red) spectra of some bast ¿bers compared to the cotton ¿ber

In Fig. 25 the FTIR spectra of Spanish Broom ¿bers obtained after four different extraction procedures are presented. According to the previously highlighted peaks that point the presence of pectin, lignin and/or wax, it can be seen that the procedures B and D resulted in the best removal of the non-cellulosic compounds of the ¿ber. Fiber treatment (B) with sodium hydroxide at 105 °C for 2 h removes completely wax and lignin while a weak peak related to pectin at 1740 cm-1 is still visible. On the other hand, ¿ber treatment (D) with the same alkaline solution but heated in a microwave oven for only 10 minutes (900 W) resulted in greater removal of pectins and maybe not total elimination of waxes. The latter treatment seems very promising, especially from the economical point of view.

Annual 2010/2011 of the Croatian Academy of Engineering

37

Figure 25 FTIR spectrum of ¿bers obtained from Spanish Broom applying four different procedures compared to the cotton ¿ber.

REFERENCES 1.

Kovaþeviü Z., Krnþeviü M., Katoviü A. Katoviü D.: Brnistra – zaboravljena tekstilna sirovina Tekstil 2010 –in press 2. http://www.lammatest.rete.toscana.it/lammatest/documenti/ginestra_manuale.pdf Accessed November 2010 3. Stojanoviü A.: Brnestra (Žuka- Spartium Junceum) (1962) Etnološki Zavod Filozofskog fakelteta Sveuþilišta u Zagrebu, 4, 3-45 4. Katoviü D., Katoviü A., Krnþeviü M.: Spanish Broom - History and Perspective (Spartium Junceum), Journal of Natural Fibers 2011 - in press 5. Nekkaa S., Chebira F., Haddaoui N.(2006): Effect of Fiber Treatment on the Mechanical and Rheological Properties of Polypropylene/Broom Fiber Spartium Junceum Composites Journal of Engineering and Applied Sciences 1(3), 278-283 6. Cerchiara,T., Chidichimo G., Ragusa, M.I., Belsito E.L., Liguorib, A. Arioli A. (2010): Characterization and utilization of Spanish Broom (Spartium junceum L.) seed oil Industrial Crops and Products 31, 423–426 7. Avella M., Casale L., Dell’Erba R., Focher B., Martuscelli E., Marzetti A.M. (1998): Broom Fibers as Reinforcing Materials for Polypropylene-Based Composites J. Appl Polym Sci 68, 1077–1089 8. Gabriele B., Teresa Cerchiara T., Salerno G., Chidichimo G., Vetere M.V. Alampi C, Gallucci M.C., Conidi C., Cassano A.( 2010): A new physical-chemical process for the ef¿cient production of cellulose ¿bers from (Spartium junceum L.) Bioresource Technology 101, 724–729 9. Cerchiara T., Chidichimo G., Gallucci M.C. Vuono D., (2010): Effects of Extraction (Spartium junceum L.) Fibres, Fibres & Textiles in Eastern Europe 2(79) 13-16 10. Cristaldi G. Latteri A. Recca G., Cicala G. (2010) : Woven Fabric Engineering , Composites Based on Natursal Fibre Fabrics, Sciyo, Rijeka, pp.317-342, ISBN 978-953307-194-7

38

Annual 2010/2011 of the Croatian Academy of Engineering

THE WAY TO REDUCE PIPE WEARING WHILE DRILLING Matanović Davorin, Moslavac Bojan, Nediljka Gaurina-Međimurec University of Zagreb Faculty of Mining, Geology and Petroleum Engineering Pierottijeva 6, 10000 Zagreb, Republic of Croatia Tel. +385 1 5535 835; fax: +385 1 48 36 074 E-mail: [email protected]

Abstract While drilling there is an additional load on the coiled tubing due to friction with the well walls/rocks regardless the drilling mode; rotating or sliding. According to recent investigations fatigue is the major factor controlling the service life of coiled tubing. The rate of fatigue damage accumulation is affected by a number of operating factors: mechanical damage, corrosion damage, damage incurred by exceeding the ultimate capacity of the string, and damage due to cycling plastic strain or fatigue. Determination of the actual friction coef¿cient is essential in de¿ning the project technical feasibility. As the part of drilling system, drilling Àuid can impact on the reduction of friction coef¿cient by implementation of lubricants. Selection of lubricants, their compatibility with drilling Àuid and rocks was determined through laboratory testing. The results are shown and elaborated in the paper. Key words: drilling, pipe wearing, lubricants

Introduction Horizontal wells are more expensive to drill than vertical wells so increased production must offset increased drilling costs. In drilling horizontal or highly deviated wells, more serious problems appear than in drilling vertical wells. These problems are: poor hole cleaning, excessive torque and drag, hole ¿lling, pipe sticking, well bore instability, loss of circulation, formation damage, poor cement job, and dif¿culties at logging jobs. From that reason, the principle factors that require extra consideration for choosing a Àuid for drilling a horizontal well are: formation damages, hole cleaning, hole stability and lubricity. Good lubricating properties of the drilling Àuid can improve greatly drilling operation ef¿ciency. When drilling long sections of highly deviated wells, there is considerable contact between the drill string and the borehole wall, generating frictional resistance to movement. The friction between the drill pipe and the wall will increase torque and power required to rotate the drill pipe and the stress on the drill pipe. It could also interfere with running the pipe in and out of the hole. The frictional resistance to movement can be large enough to be the limiting factor in horizontal and extended-reach drilling. Since there are many factors that affect torque and drag, it is sometimes dif¿cult to detect what is causing increased down hole friction. Regardless of the drilling Àuid or lubricant used, down hole friction can be reduced by

Annual 2010/2011 of the Croatian Academy of Engineering

39

conditioning the Àuid to achieve a thin, slick, compressible ¿lter cake, and by practicing good hole-cleaning techniques to minimize the cuttings bed. In order to design the best suitable drill-in Àuid for horizontal well, drilling problems generally encountered in each ¿eld should be studied and laboratory tests planned and conducted. Systematic approach to drilling optimization To optimize a drilling process there is a need to develop methods and processes that will improve drilling ef¿ciency. The method must be systematic, logical and must have quanti¿able approach. The applicable method is one that de¿nes and quanti¿es two sets of parameters for the drilling operation: performance objectives and requirements, as shown in Fig. 1. The ¿gure is taken from reference [1] but is slightly changed to apply for coiled tubing drilling with down hole positive displacement motor.

Fig. 1 Systematic approach inÀuence diagram

To achieve minimal interval cost we have to maximize on-bottom time and effective rate of penetration. First branch determines parameters that are not direct drilling parameters, but depend on well path and used equipment. Second branch determines parameters that are of greatest inÀuence in drilling process. Bit design and used down hole motor combination ef¿ciency according to observed formation characteristics are essential. To obtain optimal performance of bit-motor combination optimal Àow rate must be obtained and proper Àuid used. In deviated or horizontal wells proper determination of weight on bit in any moment or point of the well trajectory is essential. The determination of compressive load in vertical section, build section or horizontal section depends on well trajectory and coiled tubing dimensions. Buckling of coiled tubing and lock-up condition is the next limitation that must be considered. Because of different possible well paths and segment length and slope (Fig. 2) it is necessary to determine forces and stresses in coiled tubing.

40

Annual 2010/2011 of the Croatian Academy of Engineering

Fig 2. Possible well paths and pipe buckling in them

When coiled tubing is run into horizontal section, it is subjected to compressive load due to frictional drag. The axial compressive load is highest at the end of the built section and lowest at the far end of the well. When compressive load exceeds the critical buckling load, the coiled tubing will initiate a sinusoidal buckling. Further increase of the compressive load can result in a helical buckling of the coiled tubing. Helical buckling can also occur in the vertical section when “slacking-off” weight at the surface to push coiled tubing through the well. In horizontal section coiled tubing will buckle when the axial compressive load exceeds critical (sinusoidal) buckling load. When the axial compressive load reaches helical buckling load, helical buckling occurs, and is fully formed. In the build-up section, coiled tubing under axial compression will be pushed against the lower side of the well bore before the onset of sinusoidal or helical buckling. In this section buckling will not easily occur because of distributed lateral force, and the need of larger buckling load in build-up section than in straight well bore. When pushing coiled tubing into horizontal section with no helical buckling present, the axial load simply increases linearly along the tubular. This is due to the friction drag from the tubular weight, without considering any additional frictional force from sinusoidal buckling, since usually it is very small. In vertical section without helical buckling there is theoretically no frictional drag and the axial compressive load decreases linear. If helical buckling occurs in the vertical section due to “slack-off”, the axial load distribution decreases nonlinear:

Annual 2010/2011 of the Croatian Academy of Engineering

41

Maximum axial compressive load that is transmitted to the bottom of vertical section to kick-off point (KOP) by “slacking-off” is [2]: (1) Where: (Fb,max) is maximal axial compressive load at the bottom of vertical section, (E) is Young’s modulus of elasticity, (I) is moment of inertia of tubular, (we)- tubular weight in Àuid, (r) is radial clearance between well bore and tubular, (ȝ) is coef¿cient of friction and (H) is vertical well bore depth. The compressive axial load at the end of the build section (EOC) is: (2) Where: (Feoc) is axial load at the end of build section, (Fkf) is axial load at the kickoff point, (R) is radius of well bore curvature and (e) is natural logarithm base. “Lockup” refers to situation where the bit or packer load can not be increased by “slackingoff” weight at the surface, or the tubulars can not be pushed further in the well bore. And the maximal horizontal reach without helical buckling in horizontal section will be: (3) Where: (Lh) is horizontal section length. It is obvious that coef¿cient of friction is of importance when calculating loads on tubular. Measurements conducted by Johancsik at all [3] have shown that it can vary from 0,2 to 0,4 in real well bore situations. The reduction of friction will allow greater horizontal lengths with smaller pipe wearing. The capability to do so has two main bene¿ts. First, deep, highly deviated wells can be planned to minimize torque and drag by selection of most appropriate well path. Second, more complete knowledge of drill string loading allows choose of drill string components according to systematic approach that considers the extra forces involved. They have showed that tension increment and torsion increment depend on coef¿cient of friction and normal force: (4) (5) Where: (ǻFt) is axial tension acting at lower end of element, ( ) is average inclination angle of element, (Fn) is normal force acting on element, (ǻM) is increase in torsion over length of element, and (rp) is characteristic radius of drill string element. In Eq. 4, the plus sign is for upward motion and the minus sign is for downward motion.

42

Annual 2010/2011 of the Croatian Academy of Engineering

LUBRICANTS Friction reduction can be achieved by using a drilling Àuid with good lubricating qualities such as oil-base Àuids, polymer Àuids, and glycol-derivative systems [4]. Oil-based Àuids often are chosen for extended reach wells because they are highly inhibitive and exhibit good lubrication properties. Unfortunately, their use in environmentally sensitive areas is discouraged or even prohibited [5]. There are also a large number of lubricant additives available; these should be tested for Àuid and formation compatibility as some of them can have detrimental effects. In general, lubricants may be of a mechanical nature (beads, nut plug, etc.), general borehole or ¿lm lubricants, or extreme-pressure (metal-to-metal) lubricants. Natural grease and oils abandoned because of cheaper oxidant stable mineral oils, are now, because of ecological aspect, used again. Rape oil and castor oil are used in production of biodegradable lubricants. Rape oil has also other favorable properties: good lubricity, high viscosity, good adhesion on metal surfaces and very good load bearing. Problems on high and low temperatures and with foaming can be solved by adding additives. When added in water based mud (WBM) such problems are solved with the use of emulsi¿ers, antifoaming agents and for warehouse purposes pour point depressants. Lubricants are designed to reduce torque to increase horsepower at the bit by reducing the coef¿cient of friction. In water based drilling Àuids certain oils, graphite powder, glycols and soaps are used for this purpose. To be economic, a drilling Àuid lubricant must reduce torque and drag, and operating costs by performing the following functions: • reduce possibilities of twist-offs, • shorten trip time, • lessen the conditions for differential sticking, and • lower the amount of energy necessary to run the rig. Depending on their chemical composition and state of dispersion or solubility in the base drilling Àuid, lubricants: • can coat metal surfaces, reducing the adhesion of steel to the ¿lter cake; • can be incorporated into the ¿lter cake and provide better Àuid-loss control (resulting in thinner cakes); and • can be incorporated into the ¿lter cake to reduce the yield stress of the cake. In many Àuid systems, it was found that the most effective additives are those operating by more than one of the above mechanisms.

LABORATORY TESTS AND RESULTS Laboratory tests of the effect of two commercially available lubricants L1 and L2, and new developed lubricant Horma-138 [7] on rheological, ¿ltration and lubricating properties of selected drill-in Àuids were carried out. They are water-dispersible lubricants designed to decrease the coef¿cient of friction in all water-base muds, reducing torque and drag. They also have a unique wettability characteristic which lowers the potential of BHA balling [7]. InÀuence of addition different concentrations (0%, 2%, 3%) of three lubricants on rheological and ¿ltration properties in different drill-in Àuids were carried out according API standard using Fann viscometer and API ¿lter press. The three unweighted, water based

Annual 2010/2011 of the Croatian Academy of Engineering

43

drill-in Àuid systems were tested. Fluids are designated as WBM 1, WBM 2 and WBM 3 (Table 1). Lubricating properties were studied using the Lubricity Tester. BrieÀy a steel test block that simulates the wall of the hole is pressed against the test ring by a torque arm. The torque is measured by the intensity of current that is required to turn the ring at a constant rotation when immersed in the Àuid that is tested. The API procedure (RP 13B) was followed for the tests, in laboratory tests the rotational velocity was 60 min-1 (60 rpm) and the force applied was 16,95 Nm (150 lb/inch). Results are given as a torque value that can vary from 0 to 50. The lower the value, the better the lubricating properties. The coef¿cient of friction is calculated from torque reading using adequate formulas. Table 1. Effect of various lubricants on mud properties and lubricating ability

It is obvious that adding of HORMA-138 lubricant has the greatest inÀuence on plastic viscosity and yield point in comparison to Àuid without lubricants. For all three lubricants there is an signi¿cant reduction of coef¿cient of friction (COF). Fig. 3 is an illustration of the inÀuence of lubricant concentration on COF in selected water based muds.

Annual 2010/2011 of the Croatian Academy of Engineering

44

Fig. 3. Ef¿ciency of selected lubricants in mud WBM 1

The comparison of different lubricants inÀuence on lubricating properties of WBM is shown in Fig. 4. It can be seen that with 3% of HORMA-138 lubricant in WBM 1 and WBM 2 there is a 60% reduction of COF. In WBM 3 because of original excellent lubricating properties there is only 12% of COF reduction.

Fig. 4. Comparison of various lubricants impact on mud lubricating ability

Conclusions All used lubricants have certain but not important impact on rheological and ¿ltration properties of tested Àuids, which is otherwise desirable. By adding 2 to 3% of lubricant in WBM satisfactory reduction of COF is achieved. New developed lubricant HORMA-138 is in range of commercially available lubricants and is applicable for ¿eld testing. Achieved COF reduction when implemented in force and horizontal reach calculations

Annual 2010/2011 of the Croatian Academy of Engineering

45

shows that for the same horizontal reach there is the need for smaller axial loads on pipe, that means smaller amount of stresses imposed on pipe while drilling. Nomenclature E - Young’s modulus of elasticity, Pa e - natural logarithm base, 2,21828 Fb,max - maximal axial compressive load at the bottom of vertical section, N Feoc -axial load at the end of build section, N Fn - axial load acting on element, N Fkf - axial load at the kickoff point, N ǻFt - axial tension acting at lower end of element, N H - vertical well bore depth, m I - moment of inertia of tubulars, m4 Lh - horizontal section length, m ǻM - increase in torsion over length of element, N•m r - radial clearance between well bore and tubulars, m rp - characteristic radius of drill string element, m R - radius of well bore curvature, m we - tubular weight in Àuid, N/m ȝ - friction factor ʌ - Ludolf’s number, 3,14 References 1.

2.

3. 4.

5. 6.

7.

Wolfson L, Mensa-Wilmot G, Coolidge R. ( 1998) Systematic Approach Enhances Drilling Optimization and PDC Bit Performance in North Slope ERD Program, SPE 50557, SPE Annual Technical Conference and Exhibition, New Orleans, 27-30 September, p. 1-12 Wu J, Juvkam-Wold HC. (1993) Drilling and Completing Horizontal Wells With Coiled Tubing, SPE 26336, 68th Annual Conference and Exhibition, Houston, 3-6 October, p. 221-233 Johancsik CA, Friesen DB, Dawson R. (1984) Torque and Drag in Directional WellsPrediction and Measurement, Journal of Petroleum Technology, June, p. 987-992 Argillier J-F, Audibert A, Janssen M, Demoulin A. (1996) Development of a New Non-Polluting Ester Based Lubricant for Water Based Muds: Laboratory and Field Tests Results, SPE 36862, European Petroleum Conference, Milan, Italy, October 2224, p. 401-410 Bol GM. (1986) Effect of Mud Composition on Wear and Friction on Casing and Tool Joints, SPE Drilling Engineering, October, p. 369-376 Evans N, Langlois B, Audibert-Hayet A, Dalmazzone C, Deballe E. High Performance Emulsi¿ers for Synthetic Based Muds, SPE 63101, SPE Annual Technical Conference and Exhibition, Dallas, TX, October 1-4, p. 1-11 Gaurina-Meÿimurec N. (1998) Horizontal Well Drill-In Fluids, The Mining, Geological, Petroleum Engineering Bulletin, vol. 10, Zagreb, Croatia, p. 73-76

Annual 2010/2011 of the Croatian Academy of Engineering

46

FUNDAMENTALS OF MILITARY PRODUCTION DEVELOPMENT IN THE CROATIAN WAR FOR INDEPENDENCE FROM 1991 TO 1993 Dinko Mikulić University of Apllied Sciences Velika Gorica Zagrebačka cesta 5, 10410 Velika Gorica, Croatia e-mail: [email protected] Tel. 01/6230-761, Mob. 099 6222-503

Vladimir Koroman Brodarski institut d.o.o., Av. V. Holjevca 20, 10020 Zagreb, Croatia e-mail: [email protected] Tel. 01/6504-401, Mob. 098 211-324

Summary There were several phases in organizing the military production during the Homeland War from 1991 to 1993. In the ¿rst phase, emphasis was placed on human creativity and top level improvisation in producing armoured vehicles and light arms. In the second stage, the military armament and equipment production development fundamentals were established by forming The Sector for Development, Production and Scienti¿c Research. In the third phase, a systematically organised massive arms and equipment military production and overhaul had begun by forming The Production Department. In this paper, based on an example of armoured vehicles development, the fundamentals of Croatian military production are explained, which ful¿lled the military missions and reconciled the requests, thus serving the purpose of the country's liberation. Key words: armoured vehicles, development and production, Croatian War of Independence (Homeland war) Introduction The production of military equipment for Croatian National Guard (Zbor narodne garde, ZNG) soldiers began in the summer of 1991, a year of war. The equipment was produced manually in small and large plants as well as in workshops and garages. Firstly, trucks were armoured and given as present to the Croatian soldier who ¿ghts for freedom. It was a help to the police in the ¿rst place, then to the Croatian National Guard. Such improvised vehicles gave protection from infantry weapons and enabled the completion of various logistic tasks, from food and ammunition transport to the transport of the wounded. Since Croatia didn't have any armoured vehicles at its disposal, excluding a part of the police personnel carriers, these handcrafted armoured vehicles in the hands of the brave Croatian Homeland War defenders as seen on TV screens, strongly raised people's morale and gave a huge impetus to faster production and other defence ¿nances. Approximately ¿fty ¿rms participated in their production, mostly in Zagreb, then in Split, Slavonski Brod, Rijeka etc.

Annual 2010/2011 of the Croatian Academy of Engineering

47

1 The beginning of organized production First it began in the Croatian Parliament and the Government of the Republic of Croatia. Direct contacts between the Headquarters of the military command and the economy resulted in the admission of company's leaders to the Croatian Parliament, as well as the Defence Minister, lieutenant general Martin Špegelj, and associates. Defence Minister Deputy V. Kikerec and Mr M. Hlad, a Parliament representative, accepted in their of¿ces persons and company requests for development and production of military equipment among which were also the offers for the production of armoured vehicles. Having graduated from the Technical Military Academy from ýrnomerec in Zagreb, volunteers, military-technical experts offered their professional services to the Ministry of Defence leaders. As professional advisers, they took over the coordination of development and production of various equipment. The development coordination at Zagreb-based companies TAZ, Hidroelektra-Mehanizacija, Autodubrava, Autoservis-Borongaj and others soon followed as well as offering professional instructions for vehicle armouring, body armours, small individual weapons, etc. 1.1 The sector for Development, Production and Scientific Research

Systematic development of defence funding began by forming The Sector for Development, Production and Scienti¿c Research. President of the Republic of Croatia Franjo Tuÿman, PhD, issued the Decision on the Appointment of Gojko Šušak as the Defence Minister on 16 September 1991, while Croatian Prime Minister Franjo Greguriü, PhD, passed the Decision on the Appointment of Vladimir Volareviü, PhD, as the Defence Minister Deputy, who took over the organization of military equipment and armaments development and production and also formed The Sector for Development, Production and Scienti¿c Research (temporary department at the Ministry of Defence). The ¿rst staff of the just formed sector comprised military experts, mostly lecturers from the Technical Military Academy from ýrnomerec in Zagreb and experts from the Army Depot Mainentance Bregana, who as volunteers enlisted for Croatian military service, and who were competent for speci¿c ¿eld of military equipment and armaments development. These were the following Croatian army of¿cers: Dinko Mikuliü, MSc, motor vehicles; Zdenko Matijašþiü, MSc, motor vehicles; Vjekoslav Stojkoviü, PhD, motor vehicles; Mijo Vrhovski, PhD, motor vehicles; Simeon Kovaþev, PhD, motor vehicles; Nikola Gambiroža, PhD, explosives; Mirjana Gambiroža Jukiü, PhD, explosives, Milan Prpiü, MSc, explosives, Mladen Pleše, PhD, explosives; Mladen Barkoviü, PhD, electronic devices; Artur Goriþanin, MSc, electronic devices; Aljoša Božikoviü, MSc, radars; Mirko Jukl, MSc, radars; Mihalj Strmeþki, MSc, air defence; Vladimir Lebinac, MSc, electronic devices, Goran Prokopec, MSc, telecommunication; Ivan Pokaz, MSc; Mirko Kukolj, MSc, armaments; Milan Ivanuševiü, MSc, NBC equipment; Josip Petroviü, electronic devices; Josip Vinter, overhaul; Vinko Turk, artillery weapons; Ivan Peica, air defence, Vitoš Bebiü, engineering; Zdravko Paveliü, armaments; Petar Radonjiü, navy. The cooperation was established between The Sector for Development, Production and Scienti¿c Research of the Croatian Ministry of Defence and the Department of Energy of the Croatian Ministry of Industry for the purpose of coordination and unobstructed work of professional advisers of the Ministry of Defence in companies.

48

Annual 2010/2011 of the Croatian Academy of Engineering

1.2 The list of some companies working on the production of armoured vehicles

Autodubrava, Zagreb, Brodosplit, Split, TAZ – Bus Factory Zagreb, Autoservis Borongaj, Zagreb, Hidroelektra-Niskogradnja-Mehanizacija, Zagreb, Industrogradnja, Zagreb, Tehnika, Zagreb, TPK, Zagreb, J.Gredelj, HŽP Zagreb, Jedinstvo, Zagreb, Tehnomehanika, Marija Bistrica, Ĉ.Ĉakoviü Specijalna vozila, Sl.Brod, ýelik/TVIL, Križevci, Torpedo, Rijeka, Rijekaprojekt, and others. Production of some armoured vehicles, Semptember-December, 1991

Key meetings at companies Key meetings of the representatives of The Sector for Development, Production and Scienti¿c Research of the Croatian Ministry of Defence and the Department of Energy of the Croatian Ministry of Industry with representatives of the ¿rms which offer the military production of armoured vehicles, were held as follows: Meeting in the factory Torpedo Rijeka, 1 October 1991 Meeting in the factory Ĉ. Ĉakoviü, Slavonski Brod, 15 October 1991 Meeting in the shipyard Brodosplit, Split, 4 December 1991 The basic manufacturers' questions were about the necessary number of vehicles and the way or their development as well as the possibility of project ¿nancing. In the production of armoured personnel carrier based on the truck chassis, emphasis must be put on chas-

Annual 2010/2011 of the Croatian Academy of Engineering

49

sis standardization while for armouring the usage of steel plates (plates for ship) 8-10 mm thick. As a technical institute Brodarski institut offered the help of its experts for the purpose of superstructure and certain systems development.

Fig. 1. Armoured personnel carriers, Otoþac, Gacka, Krešimir i Domagoj; Industrogradnja, Zagreb, 1991 1.3 A letter to the producers of military equipment and armament, 10 November 1991

In cooperation with the Ministry of Industry and after talks with potential armaments and military equipment manufacturers, Defence Minister Deputy V. Volareviü sent the manufacturers the following letter: The Republic of Croatia The Ministry of Defence Zagreb, 10 November 1991 Class: 213-01/91-01/30 Reg. no.: 512-08-91-1 Armaments and military equipment production Please activate all available capacities for mass production of the assigned – arranged armaments and military equipment assortment for the needs of the armed battle and war operations. All produced quantities with the highest degree of achieved reliability – quality should be delivered to the commanders of the units in your territories, while the necessary documentation should be made about the delivery. Defence Minister Deputy Vladimir Volareviü, PhD

50

Annual 2010/2011 of the Croatian Academy of Engineering

Typical example of Commands’ requests The Republic of Croatia Starigrad-Paklenica Command of the Croatian Army No.: 155-1-9/ Date: 13 November 1991 The Ministry of Defence Fax 041 451-882 Dinko Mikuliü For the purposes of our command, please assign us one armoured vehicle with the following characteristics: - It must be protected against the shooting and cumulative effects - The possibility of mounting the machine-gun - Such vehicle is urgent to us Ivan Cerovac Commander 1.4 Report on motor vehicles production and development, 20 November 1991, to the Defence Minister Deputy

1.

The production of an armoured engineering vehicle Straško is running according to the delivery plan (15 delivered) 2. The production of armoured personnel carriers in Brodosplit ¿rms is ¿nished. The works are ¿nished at Hidroelektra, Tehnika-Bjelovar. 3. Armour steel HPA-10 of various thicknesses is provided for vehicle armouring, placed in the vicinity of TAZ (Tvornica autobusa Zagreb) 4. The production of safety car tires with ¿lling mass, placed and examined in Hidroelektra-Mehanizacija, Žitnjak, Zagreb. 5. The development of mine cleaner with disks, the examination of its effectiveness is under way 6. The project of multi-purpose armoured vehicle is prepared in cooperation with Brodarski Institute Zagreb, on the TAM 150 T11 6x6 chassis 7. Tactical-technical requirements are prepared for development of armoured light vehicle. 8. Preparations for the possibility of cooperation on terrain vehicle are being made 9. The project of shielded medical buses in TAZ is being prepared. 10. The production of military truck on TAM 110 T7 4x 4 chassis is being prepared. After production of ¿rst armoured vehicles, the manufacturers gained a valuable experience, so they asked for new available armoured vehicles, trucks type TAM 260, TAM 170, TAM 190 and similar. However, there wasn't a suf¿cient number of such vehicles at disposal because a huge number of vehicles was mobilized, besides building companies which had the most trucks would be left without work resources. It is estimated that approximately 200 various armoured vehicles are handcrafted and there were over 1000 people working on them, in more than 50 ¿rms and workshops in the Republic of Croatia.

Annual 2010/2011 of the Croatian Academy of Engineering

51

2 Croatian Army requirements and organized production of armoured vehicles Considering the war situation, the Croatian Army required increasingly better military equipment, with better technical protection characteristics, with greater mobility and better possibility of combat activity as well as their supporting logistics, spare parts and intervention team. In this way the Croatian Army wanted to increase tactical and strategic mobility of the Croatian Army at the whole operation zone of the Republic of Croatia. Requirements for better equipment were based upon three important factors: armour protection, mobility and ¿repower. Commanders demand the mobility of ¿repower and this is possible only by means of armoured combat vehicles for it can lead to faster change in conducting combat operations. It was not enough anymore that the soldier could only be transported at the battle¿eld protected by armour, moreover ¿repower from various arms was required as well as supporting technical logistics. Such vehicles from which infantry can lead the battle should be produced and leave the vehicle only in emergency case and reconnaissance. The basis of it can be multi-purpose light combat vehicles on wheels intended for various combat activities in the system of mechanized reconnaissance units, forces for fast activities and general support forces. Thus, armoured vehicles on wheels complement with armoured vehicles on tracks during tactical activity. Means of transport for troop transport, military trucks, ambulance vehicles, towing vehicles for MB-120 mm heavy mortars were required from the Croatian army as a priority, and then requiring better armoured vehicles with the possibility to apply weapons. For the purpose of vehicle standardization for the army multi-purpose vehicles are recommended in order to be used for various tasks. The Sector for Development, Production and Scienti¿c Research intensi¿es development and production for defence, which speeds up the implementation of projects for armoured vehicles and supporting logistics development. During the 1991 - 1993 period the following production of armoured vehicles, trucks and overhaul support was established and realized: - 120 mm Self-propelled mortar, year 1991 - Military truck TK 130 T-7 4x 4 Torpedo, year 1992 - Light Armoured Vehicle (LAV) 4x 4, year 1993 - Army Maintenance Depot, Zagreb, year 1991 2.1 Development and production of 120 mm self-propelled mortars

It is estimated that the Croatian Army has at its disposal a suf¿cient number of terrain vehicles (TAM 150 T11 6x6) which can meet tactical-technical requirements of multi-purpose armoured vehicle production. The project was given to a scienti¿c-research institution – Brodarski Institute Zagreb, which already has experts and experience in the development of complex military objects. It was the ¿rst complex development project of the Ministry of Defence which was initiated with the scienti¿c-research institution. The complete preparation of the project was led by professional adviser in The Sector for Development, Production and Scienti¿c Research, Dinko Mikuliü, MSc, and the head of the Special Projects Department in Brodarski Institute, Vladimir Koroman, MSc. The implementing body – Brodarski Institute obliged to design, construct and deliver the prototype and establish the production of the multi-purpose armoured vehicle with a built-in MB-120 mm mortar and PAM 12.7 mm heavy machine-gun. There are three project phases: 1. The construction of a functional model to full functionality

52

Annual 2010/2011 of the Croatian Academy of Engineering

2. The construction of a prototype based on the approved functional model 3. The organization of production based on the approved prototype Deadline for the completion of the 1st phase works was 10 December 1991. Within the Institute, people specialising in various technical ¿elds worked day and night in the preparation of the ¿rst functional model and prototype. Technical documentation was prepared according to the simultaneous engineering principle (simultaneous phase overlap), regarding the armour design, as well as hydraulic suspension for the impact amortization, mortar ¿tting and ammunition storage. The vehicle is designed for endurance according to possible impacts of the MB-120 mm heavy mortar, produced by “Prvomajska-Alatni strojevi / INAS”- Žitnjak. Several manufacturers participated in the 1st and 2nd development phase: TAZ Zagreb(for armour), Strojogradnja Samobor (for hydraulics), Rade Konþar – special devices and systems (laser cutting of sheet metal), Obrada Rovinj (mortar chassis at the vehicle), Elmech (inert mines – mines for testing) and other manufacturers. The testing of the self-propelled mortar functional model was carried out on 22 November 1991 on Velešec polygon in the vicinity of Velika Gorica. SoMB-120 mm self-propelled mortar model was placed on hydraulic props and levelled. Recording equipment was connected as well. The usage of mortars was carried out by a group of military of¿cers of the General Staff of the Croatian Armed Forces. 9 inert mines were ¿red. The vehicle was checked and there weren’t any deformations on it. With the change of functional model to prototype, a prototype was completely ¿nished and prepared for ¿nal testing on endurance and accuracy of targets shooting, with 99 ¿red inert mines. Further testings on mortar endurance and accuracy showed the correctness of suspension solution and the way of ¿re control system. At the beginning of 1992, the prototype was given to the 108th Brigade ZNG, Slavonski Brod. A self-propelled mortar prototype was used during the Homeland War. According to the 3 rd phase of the contract - the organization of production based on the approved prototype – two more SoMB-120 mm self-propelled mortars were produced in TAZ factory according to the documentation and supervision of works by Brodarski Institut. Note: In time of completion of self-propelled mortar production, Ivan ýermak was appointed as the Defence Minister Deputy for logistics and forms Logistics Sector which includes several departments: Department for Production, which takes over the organization of military production (from The Sector for Development, Production and Scienti¿c Research), Procurement Department, Technical and Trade Department, Construction Department and Telecommunications Department. Colonel Nikola Gambiroža, PhD, was appointed as Head of The Department for Production on 9 January 1992. Both just produced SoMB-120 mm self-propelled mortars were delivered to the Croatian Army on 7 January 1991 before Operation ''Maslenica'' in which self-propelled mortars were successfully used in liberating operation of Zadar hinterland. The contribution of Brodarski Institute to the development of Croatian Army equipment and armaments, among which was also SoMB self-propelled mortar, was shown when Croatian President Franjo Tuÿman and Minister of Defence Gojko Šušak came to visit Brodarski Institute on 31 March 1993.

Annual 2010/2011 of the Croatian Academy of Engineering

53

Fig.2. Functional model of SoMB 120 mm self-propelled mortar with 12.7 mm, Kiseli, Brodarski Institut, Zagreb, 1991

Fig.3. The testing of SoMB 120 mm self-propelled mortar prototype, Velešec, 1992 2.2 Development and production of military trucks TK 130 T-7 4x4 Torpedo

The idea of military truck production based on TAM 110 T7 4x 4 chassis fell on fertile ground in the tractor and construction machines factory Torpedo, Rijeka. After the production of armoured engineering vehicle Straško/HIAV on the universal excavator chassis, the management board decided to make preparations for the domestic production of military trucks in cooperation with the factory TAM Maribor. For the purpose of domestic production share increase, Boženko Vuþkoviü, MSc and Zdenko Novak, MSc, reconstructed the cabin and cargo crate, which was completely produced in Progres ¿rm-Jastrebarsko. Nedjeljko Ški¿ü, MSc, proposed the ¿tting of 130 HP. domestic diesel-engine Torpedo, which would be technically and logistically better than 110 HP engine. Feasibility study justi¿ed the truck domestic production. A new manufacturing department for vehicle production was prepared within the factory. In the production system Director Neven Dediü and Head of Procurement Department Marþelo Jurman worked out a plan for mounting

54

Annual 2010/2011 of the Croatian Academy of Engineering

parts procurement, thus the production in the new manufacturing department could start. A part of the chassis (frame, axle, gear box) must have been imported from the manufacturer TAM Maribor, while everything else was manufactured in Croatia (diesel engine Torpedo, the cabin, cargo crate, tarpaulin, other parts). Thus, vehicle price around 50% of purchase price of the similar imported vehicle was acceptable. A domestic share of approximately 60 % of product price was relevant. Product Quality Regulation was made, while all vehicles before the delivery passed all necessary polygon tests on Šljunþara polygon, under the Grobnik automobile race-track. The vehicles had provided logistical support, a 2-year guarantee and spare parts for the next ten-year-period. It was a successful decision in truck production, according to the guidelines of war industry, the reduction of costs and vehicle standardization in the Croatian Army. First ''0'' series vehicles comprising 20 trucks were delivered at the end of 1992, then series comprising 200 trucks were delivered in 1993-1994 period. It provided the tactical and strategic mobility of the Croatian Army during the Homeland War. In this time Torpedo TK-130 4x 4 military vehicles could be mostly seen on Croatian roads. This means that the production of Torpedo TK-130 4x 4 trucks with the capacity of 3 tons(it can accommodate 12 soldiers with complete equipment) was a very important factor of the Croatian Army capability during the Homeland War. Besides, a series production of military trucks proves that the cooperation with foreign partners on the basis of economical pro¿tability is possible even during the laid embargo. 2.3 Development and production of light armoured vehicles LOV 4x4

On the basis of good experience of ‘’0’’ series military truck production, an initiative was accepted that the manufacturer Torpedo offers the solution of the Light Armoured Vehicle on the same TK 130 T7 4x 4 vehicle chassis. A huge degree of uni¿cation of both products was achieved in this way, as well as the low price of development and production. After that, appropriate tactical-technical requirements for light armoured vehicles (LOV 4x4) were prepared. The weight of 6.5 tons light armoured vehicles will be propelled by an intensi¿ed 150 HP diesel engine, with the maximum vehicle speed 120 km/h. The vehicle crew is made of 2 + 8 soldiers. Two items of ''0'' series LOV OP are developed in 1992. Main engineer Zdenko Novak, MSc, designed a modern form of Armoured Combat Vehicle. The vehicle height was below 2 metres (1860 mm). The front part of the vehicle was well designed, the door construction for fast and easy entering and exiting of the driver, then for fast entering and exiting of the crew. The power of diesel engine was increased to150 HP, so that the vehicle got more available power for terrain mobility. In cooperation with experts from Brodarski Institut, contemporary methods were used to check the ¿rmness of the armour with regard to vibrations. The capacity of LOV OP include 2 tons artillery and 8 soldiers. The main protection of Light Armoured Vehicles is the armour made of special steel which protects the crew from all types 7.62 x 51 AP (NATO protection) ammunition which would be ¿red from the distance of 30 metres as well as the 155 mm shell fragments. 6-8-9 mm thick Armox special steel plates are applied as well as the safety glass. Due to the greater ¿re exposure of the armoured surfaces, the vehicle got the greater armour thickness. On the bottom side of the vehicle an additional two-layered antitank mines protection was made. Trial factory testings of the Light Armoured Vehicles are carried out on Šljunþara polygon, above Rijeka. The representatives of the factory, the Ministry of Defence, the Faculty of Mechanical Engineering and Naval Architecture and Brodarski Institute were present. The vehicle demonstration was very good. It is de¿ned that the braking regulator must be ¿tted on the rear axle, Run Flat tires must be provided and making the opening of the roof doors

Annual 2010/2011 of the Croatian Academy of Engineering

55

and rotation of turret easier. The ¿nal testing of light armoured vehicles as a type LOV IZV 4x4 with the 12.7 mm heavy machine-gun was carried out on polygon of Jastrebarsko barracks. During the testing Chief of General Staff of Armed Forces Anton Tus and representatives of Armoured units of the Croatian Army were present and very satis¿ed with the possibilities of barriers overcoming, especially cross water-trench capability as well as the speed of entering and exiting of the soldiers from the armoured vehicles.

Fig. 4. The testing of LOV 4x4 OP prototype, Jastrebarsko, 13 February 1993.

Fig. 5. Testing of LOV RAK 24/128 mm prototype, Zeþevo, Šibenik, 1993.

56

Annual 2010/2011 of the Croatian Academy of Engineering

After the operational capacity of the armoured vehicle was shown, a serial quantity of different types of vehicles was arranged. The delivery of these vehicles increased the tactical and strategic capability of the Croatian Army and its readiness to reject the enemy forces and end the war. Light Armoured Vehicle was developed in the following types: Armoured Personnel Carrier (LOV OP 4x4), Reconnaissance Vehicle (LOV IZV 4x4), Electronic Warfare Vehicle (LOV ED 4x4), Command Vehicle (LOV Z 4x4), Multiple rocket launcher (LOV RAK 24/128 4x4), Nuclear-Biological Vehicle (LOV ABK 4x4) and Reconnaissance patrol and artillery ¿re control system (LOV UP). 12.7 mm air-defence machine-gun is a part of the standard equipment of all Light Armoured Vehicles. The Reconnaissance Vehicle for electronic warfare uses an additional armaments – 20 mm RT-20 hand gun and 8-tube launcher system Obad (8 RL 60 M93), 60 mm calibre, with range of 8000 m. In order to increase artillery ¿repower or artillery rocket combat a type of RAK 24/128 Light Armoured Vehicle is developed. This vehicle is equipped with the multiple rocket launcher with 28 tubes of 128 mm diameter. The rocket range is 8.550 m, up to 13.500 m. A prototype LOV-2 as an improved type of previous series is developed. It comprised covered winch on the front part, easier vehicle conducting, better brakes, better electronic equipment, telecommunications system and working area. A complex development project of military LOV 4x4 Torpedo linked 4 important factors: construction, technical, conducting and economic. It gave its logistic identity. Logistic engineering requirements were set already in the 1st phase of tactical-technical requirements setting so that the total logistic support for lime time is assessed through the ’’iceberg’’ costs. Production and service price of such vehicles is more acceptable than the price of such imported vehicles. The price of the basic LOV OP 4x4 is 6 times lower than the offered price of basic Piranha 6x6 armoured vehicle, which is the proof of the possible costeffectiveness of the further production. This project of the original LOV design, low contour and domestic production proved the suf¿cient domestic professional knowledge and it showed that the army can rely on the domestic science and production in dealing with the complex military systems in the hardest war conditions. 2.4 Army Maintenance Depot

Staff of The Sector for Development, Production and Scienti¿c Research examined the disposable halls used for manufacture in Jedinstvo company in Jankomir in Zagreb and prepared the necessary documentation on maintenance realization and the overhaul of the Croatian Army military technology. Colonel Mijo Vrhovski, Col. Josip Vinter, Zdenko Matijašþiü, Col. Vjekoslav Stojkoviü, Col. Josip Petroviü and other military of¿cers of the Croatian Army were working on this project. A unit of army maintenance depot support was formed on the basis of this project on 14 October 1991, ¿rst for the overhaul of armoured means. The development of technical support system for the development of armoured vehicles began in this way, then began the formation of the 310th logistic support brigade. The 310th logistic support brigade signi¿cantly contributed to the development and maintenance of military equipment and armaments. Among other things, it maintained the armoured vehicles of different manufacturers and it gave help at the lowlands ¿elds throughout the Republic of Croatia. Army Maintenance Depot, as a part of the Logistics Command, became a unit of the Armed Forces responsible for the implementation of the overhaul highest level (the 3rd level) of armaments means, Croatian Army military equipment and it is also a support to the units of the Croatian Armed Forces which take part in the international peacekeeping missions.

Annual 2010/2011 of the Croatian Academy of Engineering

57

3 Conclusion During the Serbian aggression against the independent Republic of Croatia, people’s motivation and organized directing of equipment and armaments projects enabled the production of improvised combat means. The units of the Croatian Army required these vehicles in the ¿rst 3 years of the Homeland war. The ¿rst quantities of the armoured personnel carriers justi¿ed their purpose. Regarding the direct commanders' demand and soldiers' information from the ¿eld, the Croatian Army was satis¿ed. Around 1000 people from many ¿rms and workshops took part in the handcrafted production of 200 armoured vehicles. Requirements for military equipment caused the need for war industry directing in order to ensure the logistics of Croatian Army equipment in a faster and more rational way. Apart from the protection of people from infantry weapons using the sandwich armour, these vehicles could not provide the necessary mobility of ¿repower. Therefore, the development projects of contemporary armoured vehicles with armaments on chassis of the highly mobile military trucks, SoMB-120 (6x6) self-propelled mortars and Light Armoured Vehicles (LOV 4x4) were initiated. The developed domestic armoured vehicles were used in the three crucial country liberating operations ’’Maslenica ’’, ’’Bljesak ’’ and ’’Oluja ’’. Their usage in the hands of the Croatian Army soldiers, their operation and their production is carried out with regard to military tasks as well as in view of foreign contemporary solutions. A series production of military trucks proves that the cooperation with foreign partners is possible even during the embargo. In the end, there were several phases in organizing the military production during the Homeland War from 1991 to 1993. In the ¿rst phase, emphasis was placed on human creativity and top level improvisation in producing armoured vehicles and light arms. In the second stage, the military armament and equipment production development fundamentals were established by forming The Sector for Development, Production and Scienti¿c Research. In the third phase, a systematically organised massive arms and equipment military production and overhaul had begun by forming The Production Department. In this paper, based on an example of armoured vehicles development, the fundamentals of Croatian military production are explained, which ful¿lled the military missions and reconciled the requests, thus serving the purpose of the country's liberation. 4 References 1. 2. 3. 4. 5. 6. 7. 8.

Ivkoviü Z (2001) Hrvatski oklopnjaci 1991.-1993.: Ministarstvo obrane Republike Hrvatske, Zagreb, 2001. Poster (1996) Ruþno izraÿena oklopna vozila Hrvatske vojske 1991-1992: Galerija Zvonimir, Zagreb, 1996. Gambiroža N (1992) Proizvedeno u Hrvatskoj: Hrvatski vojnik, br. 17/92, Zagreb,1992. Bandula D (1997) Hrvatska vojna industrija: Velebit, 31.12.1997, Zagreb. Mallet J (1998) Croatian tanks and rocket launchers hits of Eurosatory 1998: DAILY, Janes Defence Weekly, Le Bourget, Pariz, 1998. Mikuliü D (1991/1992) Kolegij blok: vlastite bilješke, Zagreb, 1992. Mikuliü D (1991/1992) Oklopljena vozila: slikovni album, Zagreb, 1992. Mikuliü D (1992/1993) Oklopna borbena vozila: slikovni album, Zagreb, 1993.

58

Annual 2010/2011 of the Croatian Academy of Engineering

Intelligent Forest Fire Monitoring System – from idea to realization Darko Stipaničev Department for Modeling and Intelligent Systems & Center for Wildfire Research University of Split Faculty of Electrical Engineering, Mechanical Engineering and Naval Architecture R.Boškovića 32, 21000 Split, Croatia [email protected]

Abstract Wild¿res, including forest ¿res, are natural phenomena that cause signi¿cant economic damage and have quite devastating effect to the environment all over the world. Early ¿re detection on one side and quick and appropriate intervention on the other one, are of vital importance for wild¿re damage minimization. The ¿re season 2003 was quite severe one, particularly in Split and Dalmatia County. Provoked by great damages caused by 2003 wild¿res in autumn 2003 a project was initiated at University of Split Faculty of Electrical Engineering, Mechanical Engineering and Naval Architecture Department for Modeling and Intelligent Systems, having the main goal to ¿nd a way how advanced information – communication technologies (ICT) could be used to improve wild¿re prevention and protection. After three years of research and testing in 2006 the Intelligent Forest Fire Monitoring System called iForestFire was presented. iForestFire belongs to the last generation of wild¿re monitoring and surveillance systems, having a lot of innovative and advanced features. Since 2006 it has been applied in various Croatian national and nature parks, but also the whole Istria region is covered by iForestFire network having 29 monitoring stations and 7 operational centers. The paper describes main features of iForestFire system with emphasize on its development and implementation. Key words: wild¿re, wild¿re detection, smoke detection, intelligent system 1 Introduction Wild¿re or wildland ¿re is any uncontrolled burning of natural vegetation (grass, shrub, forest timber, litter and slash) in the wilderness area. When the vegetation layer is more precisely known, the more speci¿c names could be used like forest ¿re, grass ¿re or bush¿re. Wild¿res represent a constant threat to ecological systems, infrastructure and human lives. According to prognoses, wild¿res, including ¿re clearing in tropical rain forest, will halve the world forest stand by the year 2030. In Europe, up to 10,000 km2 of vegetation is destroyed by ¿res every year, and up to 100,000 km2 in North America and Russia. Wild¿res are responsible for approximately 20% of CO2 emission into the atmosphere (Kührt et al, 2001). Croatia belongs to countries with high wild¿re risk. In summer season, seven coastal counties in Croatia, including in particular the Adriatic islands, are permanently exposed from high to very high ¿re risks. This is due to meteorological conditions, densely spaced conifer forests and a lot of tourists. According to Croatian Forests data (Žaja, 2008) from

Annual 2010/2011 of the Croatian Academy of Engineering

59

1992 to 2007 there were 4.851 wild¿res in Croatia and the burning area was 251.910 ha. Fire seasons 2000 and 2003 were particularly severe with 706 and 532 wild¿res and the total damage caused by wild¿res was huge. For example in 2003 only in Split and Dalmatia County there were 130 wild¿res, the total burned area was 9.700 ha, the direct damage caused by wild¿res (¿re¿ghters interventions and post-¿re terrain recovery) was 16 mil.€, and the indirect damage, taking into account energetic equivalent of lost woody biomass, was assessed to 66 mil.€ (Stipaniþev et al., 2004, Stipaniþev et al. 2007). Wild¿res in 2003 were particularly catastrophic on Split and Dalmatia County islands. For example, on islands Hvar and Braþ the burning area was between 1/4 and 1/3 of total islands area and the small island Biševo near Vis has been totally burnt. Figure 1 shows photos of few 2003 ¿res on Split and Dalmatia County islands photographed by the author of this paper.

Figure 1 – Wild¿res on Dalmatian islands in 2003

Provoked by great damages caused by 2003 wild¿res in autumn 2003, a project was initiated at Department for Modeling and Intelligent Systems University of Split Faculty of Electrical Engineering, Mechanical Engineering and Naval Architecture, having the main goal to ¿nd a way how advanced information – communication technologies (ICT) could be used to improve wild¿re prevention and protection. After detailed survey of research topics and implementation of ICT systems in wild¿re prevention task in other countries subjected to wild¿res, we have focused ourselves primarily on these three topics: - Automatic early wild¿re detection, - Calculation of micro location wild¿re risk index, and - Wild¿re behavior and propagation simulation. The ¿rst topic is the most important one, because the only effective way to minimize damage caused by wild¿res is wild¿re early detection and fast and appropriate reaction, apart from preventive measures. Great efforts are therefore made to achieve early wild¿re detection, which is traditionally based on human wild¿re surveillance, realized by 24 hours observation by human observers located on selected monitoring spots. In Croatia the human wild¿res surveillance is mainly organized by Croatian Forests – the governmental organization responsible for protection and exploitation of forests in state ownership. Human observers are usually equipped only with standard binoculars and communication equipment and their observation area is only the area covered by their sight of view. A rather new, technically more advanced approach to human wild¿re surveillance is video cameras based human wild¿re surveillance and monitoring when remotely controlled video cameras are installed on monitoring spots, and the human observer is located in the observation center. Such system could be used not only for early ¿re detection, but also for

60

Annual 2010/2011 of the Croatian Academy of Engineering

distant video presence on ¿re remote location.

Figure 2 – The difference between human wild¿re surveillance (left) and video camera based human wild¿re surveillance (right)

The video cameras based human wild¿res surveillance has many advantages in comparison to direct human observation from monitoring spots. Few of them are: - A wider area could be covered, because one human observer could monitor few video monitoring ¿eld units. - Cameras are usually equipped with power zoom (optical zoom with 22 x magni¿cation) so the observer could easily inspect suspected areas. - System usually has video storing capabilities, at least for the last couple of days, and that is quite useful for post-¿re analysis. The main limitation of video cameras based human surveillance is that ¿re detection depends entirely on the human observation. The observer is located in more comfortable environment, the observation center, but he (or she) has to carefully watch multiple computer monitors at the same time, so problems like fatigue, boredom and loss of concentration could be encountered. That was the main reason for introduction of various forms of automatic and advanced automatic wild¿re surveillance and monitoring systems and networks. The system described in this paper, named iForestFire® – Intelligent Forest Fire Monitoring System (Croatian name is IPNAS® – Inteligentni Protupožarni NAdzorni Sustav) (iForestFire, 2011), belongs to the last generation of advanced automatic wild¿re surveillance and monitoring systems, having a lot of innovative and advance features. It was entirely developed at University of Split Faculty of Electrical Engineering, Mechanical Engineering and Naval Architecture, so iForestFire is a good example of university – industry cooperation, because today it is a successful commercial product installed in various Croatian national and nature parks and Istria County, but also a product that have signi¿cant export opportunities. In the rest of this paper, a short introduction to automatic wild¿re monitoring and surveillance systems will be given, main features of iForestFire system will be described with emphasis to its advanced and innovative aspects and ¿nally a short discussion about iForestFire development, implementation and commercialization will be added. 2 Automatic wildfire surveillance and monitoring system The research and system development in the area of automatic wild¿re surveillance system was extended in the last couple of years. There are two main types of automatic wild¿re

Annual 2010/2011 of the Croatian Academy of Engineering

61

surveillance systems: satellite systems based on satellite wild¿re monitoring and terrestrial systems based on wild¿re monitoring from ground monitoring stations. In this moment terrestrial systems are more useful because early ¿re detection is not the only one task that contemporary automatic wild¿re surveillance and monitoring system has to ful¿ll. Beside the automatic detection of wild¿res, the distance video presence on ¿re remote location is almost of equal importance and today’s satellite systems could not ful¿ll that task. Satellite systems could be used only for wild¿re detection with limited spatial and time resolution. Therefore the automatic wild¿re surveillance and monitoring ground-based or terrestrial system used as an enhancement of human based wild¿re surveillance are dominant today. In ground-based or terrestrial systems different kinds of ¿re detection sensors could be used: - Video cameras sensitive in visible spectra where wild¿re detection is based on smoke recognition during the day and ¿re Àames recognition during the night. - Infrared (IR) thermal imaging cameras where wild¿re detection is based on detection of heat Àux from the ¿re (Arrue et al, 2000). - Optical spectrometry that identi¿es the spectral characteristics of smoke (Forest Fire Finder, 2010). - Light detection and ranging (LIDAR) systems that measure laser light backscattered by the smoke particles (Utkin et al, 2003) - Radio-Acoustic Sounding System (RASS) for remote temperature measurements and thermal sensing of a particular forest region (Sahin et al, 2009). - Acoustic Volumetric Scanner (VAS) that recognizes the ¿re acoustic emission spectrum as a results of acoustic ¿re sensing (Viegas et al., 2008). - Sensor network based system, where a number of sensor nodes (in most cases wireless sensors) are deployed in forest, measuring different environmental variables used for ¿re detection. There are lot of different approaches, from more or less standard wireless sensor network (Byungrak et al, 2006), application of so called Fiber Optic Sensor Network (FOSN) developed within the EU-FIRE project (Viegas et al, 2008) to exotic proposals where animals have to be used as mobile biological sensors equipped with sensor devices (Sahin, 2008). Each technology has its advantages and disadvantages. Most of them are promising, but they are still in experimental stage, particularly sensor networks, VAS, RASS and LIDAR systems. For example LIDAR - light detection and ranging system is used to carry out chemical detection from great distances and has the potential to be an ef¿cient system for wild¿re detection. However, it requires the lighting of the horizon with a laser beam that causes public health risks, besides not being very feasible from the economic point of view. Because of that, today in commercial use are mostly video based system equipped with cameras sensitive in visible spectra and/or infrared spectra and systems based on optical spectrometry. Optical spectrometry is rather new technology. It is based on a chemical analysis of the atmosphere by an optical spectrometry system. A telescope is coupled with optical sensor connected to a spectrometer unit with optical cable. The system analyzes the way the sunlight is absorbed by the atmosphere. It is quite ef¿cient having the smallest number of false alarms, but of course it has a lot of disadvantages too. The main one is that it scans the space above the tree crowns on horizon, so the smoke has to be higher than the horizon. The second one is dubious night detection when standard video camera is used ¿rst to detect light and then optical spectrometry is used to detect ¿re Àames. Because of that in commercial optical spectrometry based systems, the video cameras in visible spectra are also included. Infrared systems are good choice for wild¿re detection, but their price is still quite high in comparison to video cameras sensitive in visible spectra, they have limited space resolution and they could not be used for distant video presence. Therefore,

62

Annual 2010/2011 of the Croatian Academy of Engineering

contemporary systems for wild¿re detection based on infrared cameras are usually also equipped with cameras sensitive in visible spectra. The conclusion is that almost in all commercially available systems, the camera sensitive in visible spectra is also present. As an old adage said: “The best hunting solution is to kill two rabbits with one shot.” we think that today the most suitable solution for terrestrial automatic wild¿re detection and monitoring systems is to use cameras sensitive in visible spectra, particularly from the price/quality point of view. If you want to build a network, you need a lot of monitoring devices and the price of various advanced ¿re detection systems are much higher than today’s high quality video cameras. Additional feature of today’s video cameras is their dual sensitivity. They are usually color cameras sensitive in visible spectra during the day, and black and white cameras sensitive in near IR spectra during the night, so the detection capabilities, particularly in sunrise/sunset parts of the day are greatly improved. The second reason why visible spectra video cameras based system are the best choice is because their way of detecting wild¿res is the most close to human based wild¿re detection. Human wild¿re observers primarily use his (or her) vision sensor (eyes) to detect wild¿re. Sometimes humans use additional visual enhancement devices like binocular to check suspicious areas, but humans never use other sensors for early wild¿re detection. During our research and development phase we have spoken with a lot various human wild¿re observers and no one of them said that he (or she) is using for example hearing (ears) to detect wild¿res. That was the reason why our efforts from the beginning were to develop ground-based wild¿re monitoring system based on visible cameras sensitive in visible spectra. Off course we were not the only ones who have thought of that way. In various countries that encounter high risk of wild¿res, various terrestrial systems based on cameras sensitive in visible spectra were developed and proposed. In all of them automatic wild¿re detection was based on smoke recognition during the day and Àame recognition during the night. The main disadvantage of those systems is rather high false alarms rate, due to atmospheric conditions (clouds, shadows, dust particles), light reÀections and human activities. Therefore, systems are usually designed as semi-automatic systems, which means that a human operator supervises the automatic wild¿re detection and his (or her) decision is the ¿nal one. After the ¿re alarm is generated and suspicious part of the image is marked, the human operator con¿rms or discards the alarm. The task of a human operator is not to monitor camera displays all the time, like in video cameras based human wild¿re surveillance mentioned in previous section, but only to con¿rm or discard possible ¿re alarms. If the human operator is not sure about a ¿re alarm, he (or she) could switch the system to manual operation and make additional inspections using camera pan, tilt and zoom features. Using such semi-automatic surveillance system, human operator ef¿ciency is highly improved. One operator can manage more video monitoring units but also his (or her) fatigue is greatly reduced. iForestFire - Intelligent Forest Fire Monitoring System belongs to this category. It is an innovative, cloud computing based, semi-automatic wild¿re detection system with quite advance distant video presence capabilities. 3 Main features of iForestFire - Intelligent Forest Fire Monitoring Systems iForestFire is integrated and intelligent video based wild¿re surveillance and monitoring system. Wild¿res are detected in incipient stage using advanced image processing and image analyses methods. Intelligent ¿re recognition algorithms analyze images automatically, trying to ¿nd visual signs of wild¿res, particularly wild¿re smoke during the day and wild¿re Àames during the night. If something suspicious is found, pre-alarm is generated

Annual 2010/2011 of the Croatian Academy of Engineering

63

and appropriate image parts are visibly marked. The operator inspects suspicious image parts and decides is it really the wild¿re or not. The system is capable to work with both types of cameras: video cameras sensitive in visible and near IR spectra and real IR thermal imaging cameras, but video cameras sensitive in visible spectra are preferred. Theoretical background of iForestFire is innovative and newly introduced wild¿re observer network theory based on three-layer sensor network architecture, formal theory of perception and notation of observer (Stipaniþev et al, 2007a; Šeriü et al, 2009). Wild¿re observer, illustrated in Figure 3, is the core element of iForestFire system. It has three horizontal layers: data or sensor layer, information or service layer and knowledge or application layer, vertically interconnected by low-level or data observer working as proprioception unit (syntactic and semantic validation of sensors and sensors data) and two high-level observers, the image ¿re observer and the decision ¿re observer, working as exteroception units (making conclusions based on sensory data).

Figure 3 – Wild¿re observer is organized as three-layer observer network

iForestFire is a cloud computing or Web Information System (WIS) which means that the operator could be located on any location with broadband Internet connection and his (or her) user interface is standard Web browser. The system is based on ¿eld units and a central processing unit. The ¿eld unit includes the day & night, pan/tilt/zoom controlled IP based video camera and IP based mini meteorological stations connected by wired or wireless LAN to a central processing unit where all analysis, calculation, presentation, image and data archiving are done. The system is also an example of Future Generation Communication Environment (FGCA) where all applications and services are focused on users, and “user” in our case is the natural environment, having the main task its own wild¿re protection. For such environment behavior the term environmental intelligence (EI) was introduced (Stipaniþev et al, 2007a; Šeriü et al, 2009).

64

Annual 2010/2011 of the Croatian Academy of Engineering

iForestFire is both integral and intelligent user friendly system. Its organizational structure is shown in Figure 4.

Figure 4 – The structure of iForestFire system

It has three data bases: data warehouse with input images and alarm images, SQL database with meteorological data and temporal information of alarm images and GIS database with all relevant GIS data. iForestFire has ¿ve working modes: - Manual Mode – user-friendly camera manual control by pan, tilt and zoom. - Automatic Mode – automatic ¿re detection based on images captured by video cameras in the visible and near infrared spectra. - Archive Mode – video and meteorological data archive retrieval using various userfriendly procedures. - Simulation Mode – ¿re behavior modeling and ¿re-spread simulation using meteorological data and various GIS layers. - Fire Risk Calculation Mode - micro location ¿re risk index calculation using, not only meteorological data, but sociological parameters connected with forest ¿res too. iForestFire is integral because it is based on three different types of data: - Real time video data. Digital video stream is used in both, automatic and manual system modes. In automatic mode the video stream is a source of images for automatic wild¿re detection and in manual mode the video stream is used for distant video presence and distant video inspection. - Real time meteorological data. The meteorological data is used in the post-processing unit for false alarm reduction, but it is important for wild¿re risk calculation during the monitoring phase and wild¿re spread behavior modeling during the ¿re-¿ghting phase. Main meteorological parameters are measured using high tech IP based ultra sound mini meteorological station (iMeteo, 2011), also developed during iForestFire project. - GIS (Geographical Information System) database. GIS system stores not only informa-

Annual 2010/2011 of the Croatian Academy of Engineering

65

tion on pure geographical data (elevations, road locations, water resources etc.), but also all other relevant forest ¿re information related to a geographic position, like ¿re history, land cover – land use, roads and forest corridors and similar. These data are used for userfriendly camera pan/tilt control but also they are quite useful for ¿re¿ghting management activities. GIS data is essential for Simulation Mode and Fire Risk Calculation Mode. iForestFire is intelligent because it is based on arti¿cial intelligence (AI), computational intelligence (CI) and distributed intelligence (DI) technologies like: - Multi - agent based architecture. The system software organization is based on agent architecture. Intelligent software agents are responsible for sensors integrity testing, image and meteorological data collecting, syntactic and semantic image and data validation, image and data storing, image pre-processing processing and post-processing and pre-alarms and alarms generation. All agents share the same ontology and speak the same agent communication language (ACL) (Šeriü et al, 2009). To demonstrate the system complexity, let us mention that on one server having 5 monitoring locations with 16 preset positions on each video unit, more then 300 agents are working in parallel. - Advanced image processing and analyses algorithms. In its automatic mode, the wild¿re detection is based on various advanced image processing, image analyzing and image understanding algorithms. Various algorithms work in parallel based on advanced motion detection, advanced image segmentation, ¿res smoke dynamic pattern analysis, color-space analysis and texture analysis (Krstiniü et al,2009; Krstiniü et al, 2011). Typical detection result for monitoring station located in Buzet region (Istria) is shown in Figure 5. - Advanced procedures for false alarms reduction. In post-processing analysis, various methods derived from intelligent technologies ¿eld are used to reduce the number of false alarms, as for example advanced image processing techniques (Jakovþeviü et al, 2009), rule-based expert system, data fusion algorithms (Stipaniþev et al, 2007b) and integration of ¿re risk index calculation with automatic adjustment of detection sensitivity (Bugariü et al, 2009). Algorithms have a number of tuning parameters, but our experience was that users adjust them rarely. The poorly adjusted parameters sometimes cause overly false alarm generation. That was the reason why we have introduced the possibility of automatic parameter adjustment based on meteorological data fusion and augmented reality features. Results of ¿re risk index calculation are used to automatically increase or decrease system detection sensitivity on various image regions. Also a powerful QoS (Quality of service) was developed, particularly related to wild¿re observer detection quality evaluation (Jakovþeviü et al, 2010). QoS is used particularly as a tool for further improvements of detection quality. - Augmented reality. The system is geo-referenced, so for every image pixel the corresponding geo-coordinate could be known and vice versa. The augmented reality features, now in experimental phase, based on fusion and integration of GIS information and real time video images are used in both, automatic and manual mode. Two examples of augmented reality use in automatic mode are automatic adjustment of detection sensitivity and determination of smoke location geo-coordinates. In manual mode important GIS information could be shown on video screen like toponyms, coordinates, altitudes, but GIS data are also used in advanced cameras manual control. In system design phase particular attention was given to create a user-friendly system. All iForestFire modules and components could be reached and administrated through dynamic and interactive Web pages, where real time video and meteorological data are shown together with GIS data and user friendly interface for camera pan/tilt/zoom camera control. From the beginning, the ¿re¿ghters were involved in experiments with iForestFire system prototype, so the ¿nal user interface was designed taking into account their advices. Figure 5 shows a typical camera control screen, and a typical ¿re alarm screen.

66

Annual 2010/2011 of the Croatian Academy of Engineering

Figure 5 - iForestFire camera control screen with various manual control modes and typical ¿re alarm screen at location Buzet in Istria. The user can accept or discard the generated alarm.

For right decision about ¿re¿ghting intervention, both the early ¿re detection and appropriate judgment about the potential ¿re danger are important. That is the reason why from the ¿re¿ghters’ point of view, both automatic detection of wild¿res and manual camera control modes are of equal importance. Because of that, we have implemented in iForestFire various user-friendly procedures for cameras manual control like: - Geo-referenced camera map control. The user can control camera pan movement by simple clicking on geo-referenced map. The camera control system is integrated with GIS, so it automatically detects and informs the user is the chosen point visible from the camera location or not. - Geo-referenced one-click multiple cameras map control. In regions where the monitoring cameras network is established (like Istria County) the so-called one-click multiple cameras control was implemented. The user simply click on geo-referenced map and in background visibility of that location is calculated for all cameras in neighborhood, appropriate azimuth and elevation angles are calculated and all cameras are automatically pan and tilt moved to show chosen location (Stipaniþev et al, 2009). - Camera control using panorama image. In left upper corner of Figure 5 the 360o panorama image is shown. By simply clicking on panoramic image camera moves to chosen position by both pan and tilt. - Camera control using preset positions. Camera could by simple click on preset thumbs, moved to preset positions, pre-de¿ned by pan, tilt and zoom. - Virtual pan-tilt-zoom commands and joystick emulation. Virtual commands are shown in upper right part of Figure 5. Simple, self-explaining virtual commands for pan tilt and zoom camera control were implemented, together with joystick software emulation. iForestFire has a powerful archive retrieval methods for input images, generated alarm images and meteorological data. All of them could be easily reached and analyzed using various image and image data retrieval procedures based on advanced Internet technologies. Each monitoring station is equipped with advanced IP based ultra sound mini meteorological station used for measuring the most important meteorological parameters: air temperature, relative humidity, air pressure, wind speed, wind direction and wind gust. Additionally it is possible to measure other meteorological parameters important for wild-

Annual 2010/2011 of the Croatian Academy of Engineering

67

¿re behavior and system performances, as for example isolation, precipitation, moisture, ground temperature or lighting activity. Meteorological data are used in Automatic Mode for false alarms reductions in post-processing procedures, but also in our experimental Fire Risk Calculation Mode and Simulation Mode. Figure 6 shows the system interfaces for micro-location wild¿re risk calculation and simulation of wild¿re propagation.

Figure 6 – Experimental system for micro-location wild¿re risk calculation (left) and simulation of wild¿re propagation (right)

In Fire Risk Calculation and Simulation Mode external meteorological services are also used, for example the results of meteorological simulations performed by simulation model ALADIN-HR. These data are automatically collected from the servers of Meteorological and Hydrological Service of Croatia twice a day. Figure 6 shows ALADIN-HR wind data superimposed on ¿re risk and simulation of ¿re propagation maps. iForestFire is also award wining ICT project. In 2008 iForestFire won ¿rst price (Tesla Golden Egg) on VIDI e-novation award, competition founded by VIDI publication and Rudjer Boskovic institute. It was elected as the most prominent and innovative ICT project in Croatia in 2008. 4 Development and commercialization of iForestFire system Development of Intelligent Forest Fire Monitoring System started in autumn 2003 as a seminar on the postgraduate (Mr.Sc.) study at Faculty of Electrical Engineering, Mechanical Engineering and Naval Architecture, provoked by great damages caused by 2003 wild¿res. The student Damir Krstiniü (who was also the voluntary ¿re¿ghter at island Hvar) enrolled the course Digital Image Processing and Analyses. The results of his seminar entitled “Wild¿re smoke segmentation” were more than promising, so we have decided to apply for technology project, and in 2003 we have received a grant TP-03/0023-09 ''System for early forest ¿re detection based on cameras in visible spectra" supported by the Ministry of Science, Education and Sport of Republic Croatia. The grant funds were enough to start more intensive research, but not enough to ¿nished it, so a part of our research were also funded by ordinary project 023-0232005-2003 "AgISEco Agent-based intelligent environmental monitoring and protection systems", by support of Split and Dalmatia County authorities through a study "Holistic approach to forest ¿re protection in Split and Dalmatia County" and by our own resources. In initial system development four researchers were involved - Darko Stipaniþev as a team leader, Maja Štula as a designer of overall Web based information system, Damir Krstiniü

68

Annual 2010/2011 of the Croatian Academy of Engineering

as a designer of wild¿re detection algorithms and Ljiljana Šeriü as a designer of wild¿re observer agent architecture. From the beginning the main ¿re¿ghters adviser was Tomislav Vuko, the vice commander of Croatian ¿re¿ghters for Adriatic Coast and Islands. Research started in 2003 with wild¿res video materials collecting necessary for detection algorithm development. Controlled ¿res were burned at island Hvar in cooperation with voluntary ¿re¿ghters. Lot of video material was recorded and soon the ¿rst version of offline detection algorithm was developed. Figure 7 shows examples of ¿rst algorithm detections in the typical landscape of island Hvar. Recording and collecting wild¿re images and video sequences is quite important for detection algorithm development and testing so this activity was carried out all the time. Today our database has more than 2.500 images selected and segmented by reference (ground truth) human observer. Figure 8 shows one typical example. On manually segmented image all regions were distinguished, because in the latest version of our detection algorithm region context based wild¿re alarm reduction has been applied so it was important to have manually segmented, not only wild¿re smoke – no smoke regions, but also other regions like sky, water, vegetation, man-made objects etc.

Figure 7 – Examples of detection images in the typical landscape for Croatian coasts and islands (island Hvar 2003)

Figure 8 – Input image and corresponding manually segmented image regions represented by various gray tonalities

Annual 2010/2011 of the Croatian Academy of Engineering

69

After successful laboratory implementation, the real life ¿eld-testing was performed during ¿re seasons 2004 and 2005. Three experimental monitoring stations were installed on Marjan Hill near Split, Vidova gora on island Braþ and FESB faculty building in Split. These ¿eld tests were performed in cooperation with Fire Brigade Administration for the Coast, Protection and Rescue Operations Administration, Ministry of Interiors in Divulje so the main monitoring centers were in Divulje and the second one in our laboratory. Figure 9 shows the experimental system layout and Figure 10 the experimental station on Marjan hill.

Figure 9 – The experimental system layout used for iForestFire prototype testing in 2004 and 2005

After the testing period a lot of improvements in both, hardware and software design were implemented and in 2006 the commercialization phase has started and the ¿rst real life monitoring system was installed in National Park Paklenica supported by Ministry of Culture of Republic Croatia. The system was partially developed as a technological project supported by Ministry of Science, Education and Sport of Republic Croatia and in contract signed with them our obligations were also the system commercialization. Two models were possible – to establish a new spin-off company or to ¿nd a strategic partner. In that moment, the second option was more convenient for us, so we found the strategic partner in company Lama d.o.o. from Split (iForesFire, 2011). Lama was responsible for system promotion, selling, installation and maintenance, but further system improvements, development and research were and remain until today our obligations. The development of system like iForetFire is a never-ending story. The system version is now 2.7 (October 2010) and beside the initial researchers involved in system development from the begineeng, in today’s version signi¿cant contributions were given also by Toni Jakovþeviü (advanced IP based meteorological station, detection algorithm), Marin Bugariü (Web GIS based system features), Josip Maras (component based system architecture), Petar Jerþiü (software developement) and Kaja Radiü (electronic components design and realization).

70

Annual 2010/2011 of the Croatian Academy of Engineering

Figure 10 – The experimental station on Marjan hill

Since 2006 the system was successfully applied in various Croatian national and nature parks (7 monitoring stations and 5 operational centers), but the most advanced is network application in Istria County. Istria system called Istria iForestFire Net (Stipaniþev et al, 2010) has 29 monitoring stations and 7 operational centers, mutually interconnected using encrypted VPN and hardware ¿rewalls developed by our strategic partner – the company Lama. Figure 11 shows system layout and few screen prints of that system. iForestFire also has a quite promising export potentials. In March 2011 two demo units are in installation phase, one in Greece and the other in Portugal. Last but not least it is important to emphasize that iForestFire has initiated a lot of scienti¿c research, too. As a result of research connected with system development three PhD theses were written and defended (D.Krstiniü, Lj.Šeriü and T.Jakovcevic), two PhD theses are in preparation phase and lot of scienti¿c papers were published in journals, books and conference proceedings. A specialized Web portal dedicated to wild¿re observers and smoke recognition has been created and maintained (http://wild¿re.fesb.hr), but also our interest in more general scienti¿c wild¿re research has resulted in establishment of Center for Wild¿re Research (http://cipop.fesb.hr).

Annual 2010/2011 of the Croatian Academy of Engineering

71

Figure 11 – Istria iForestFire Net – advanced wild¿re monitoring network in Istria County

5 Conclusion The only effective way to minimize damages caused by wild¿res is wild¿re early detection and fast and appropriate reaction, apart from preventive measures. Great efforts are therefore made to achieve early forest ¿re detection, which is traditionally based on human surveillance. This paper shows how modern ICT technologies could be used for automatic wild¿re detection and monitoring. It describes the advanced wild¿re surveillance and monitoring system named iForestFire and entirely developed at University of Split Faculty of Electrical Engineering, Mechanical Engineering and Naval Architecture. iForestFire is a practical realization of the observer network theory. Observer network was de¿ned as an advanced sensor network described using formal theory of perception and a notation of the observer. iForestFire is both integral and intelligent system. Integral because it is based on various data types (images, meteorological data, GIS data) and intelligent because it has a lot of features derived from various intelligent technologies (multiagent architecture, advanced image analysis and image recognition algorithms, advanced procedures for false alarm reduction, augmented reality). The system was developed as a technological and scienti¿c research project, but today it is commercial system widely used for advanced wild¿re surveillance and monitoring in various Croatian national and nature parks and Istria region, so it is a good example of successful university – industry cooperation, particularly because its development has provoked a lot of scienti¿c research activities resulting in a number of PhD thesis and published papers in journals, books and conference proceedings.

Annual 2010/2011 of the Croatian Academy of Engineering

72

References 1. 2. 3.

4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19.

Arrue B C, Ollero A, Martinez de Dios J R (2000) An intelligent system for false alarm reduction in infrared forest-¿re detection. IEEE Int. Systems, May/June 2000, 64-72. Bugariü M, Jakovþeviü T, Stipaniþev D (2009) Automatic adjustement of detection parameters in forest ¿re video monitoring system, MIPRO, May 2009, Opatija, pp. 270-275 Byungrak S, Yong-Sork H, and Jung-Gyu K, (2006) A Design and Implementation of Forest-Fires Surveillance System based on Wireless Sensor Networks for South Korea Mountains, IJCSNS Int. Journal of Computer Science and Network Security, V.6 No.9B, pp.124-130 Forest Fire Finder NGNS IS, http://www.ngns-is.com/html/Àorestas/fff_intro_eng. html (Accessed March, 1 2011) iForestFire (2011) Itelligent Forest Fire Monitorig System, http://iforest¿re.fesb.hr (Accessed March, 1 2011) iMeteo (2011) Advance Ultrasound IP based Mini Meteorological Station, http://imeteo.fesb.hr (Accessed March, 1 2011) Jakovþeviü T, Stipaniþev D, Krstiniü D, (2009) False alarm reduction in forest ¿re video monitoring system, MIPRO 2009, May 25-29, Opatija, 264-269 Jakovþeviü T, Šeriü Lj, Stipaniþev D, Krstiniü D, (2010) Wild¿re smoke detection algorithms evaluation, VI Int. Conf. on Forest Fire Research, Coimbra, Nov 15-18 2010. Krstiniü D, Jakovþeviü T, Stipaniþev D, (2009) Histogram-Based Smoke Segmentation in Forest Fire Detection System, Information Technology and Control, Vol.38, No.3, 2009. 237-244 Krstiniü D, Kuzmaniü Skelin A, Slapniþar I (2011) Fast Two-Step Histogram-Based Image Segmentation, IET image processing. 5 (2011), 1; 63-72 Kührt E, Knollenberg J, Mertens V (2001) An automatic early warning system for forest ¿res, Annals of Burns and Fire Disasters, vol. XIV, n. 3, Sept. 2001 Sahin Y, Animals as Mobile Biological Sensors for Forest Fire Detection, Sensors 2007, 7, 3084-3099 Sahin Y G, Turker Ince S, Early Forest Fire Detection Using Radio-Acoustic Sounding System, Sensors 2009, 9, 1485-1498 Stipaniþev D, Hrasnik B (2004) Integral, holistic model of wild¿re prevention in Split and Dalmatia County, Expert Study for Split and Dalmatia County, FESB Split, 2004. (231 pages) (in Croatian) Stipaniþev D, Hrastnik B, Vujþiü R (2007) Holistic Approach to Forest Fire Protection in Split and Dalmatia County of Croatia, Wild¿re 2007 Int.Conference, Sevilla, Spain, May 2007. Stipaniþev D, Bodrožiü Lj, Štula M (2007a) Environmental Intelligence based on Advanced Sensor Networks, Proc.of 14th Int.Conference on Systems, Signals and Image Processing, Maribor, Slovenija, 27-30.6.2007 Stipaniþev D, Bodrožiü Lj, Štula M (2007b) Data Fusion in Observer Networks, Proc. of Second (IEEE) International Workshop on Information Fusion and Dissemination in Wireless Sensor Networks. Pisa, Italia, 08.10.2007, 1-6 Stipaniþev D, Bugariü M, Bodrožiü Lj (2009) Integration of Forest Fire Video Monitoring System and Geographic Information System, Proc. of 51st Int.Symp ELMAR 2009, Zadar, Sept 2009, pp.49-52 Stipaniþev D, Štula M, Krstiniü D, Šeriü Lj, Jakovþeviü T, Bugariü M (2010) Advanced automatic wild¿re surveillance and monitoring network, VI International Conference on Forest Fire Research, Coimbra, Porugal, Nov. 15 – 18, 2010, 053, 15 pages

Annual 2010/2011 of the Croatian Academy of Engineering

73

20. Šeriü Lj, Stipaniþev D, Štula M (2009) Observer network and forest ¿re detection, Information Fusion, DOI: 10.1016/j.inffus.2009.12.003, on-line from Dec 28, 2009. 21. Utkin A, Vilar R, Feasibility of forest-¿re smoke detection using lidar, International Journal of Wildland Fire 2003, 12(2) 159 – 166 22. Viegas X, Pita L P, Nielsen F, Haddad K, Calisti Tassini C, D’Altrui G, Quaranta V, Dimino I, Acoustic and thermal characterization of a forest ¿re event, Proceedings of SPIE, Remote sensing of ¿re, San Diego CA , 2008, vol. 7089, pp. 708904.1708904.12 23. Žaja D (2008) The role of Croatian Forests d.o.o. in ¿ghtings again wild¿res, International Workshop – new Methodes and Approaches to Wild¿res Prevention and Protection, Dec, 8-11 2008., Makarska (in Croatian)

Annual 2010/2011 of the Croatian Academy of Engineering

74

Laboratory testing and numerical modelling of MBT waste deformability Igor Petrović, Ph.D. University of Zagreb, Faculty of Geotechnical Engineering Hallerova aleja 7, Varaždin, Croatia e-mail: [email protected]

Prof. Davorin Kovačić, Ph.D. University of Zagreb, Faculty of Geotechnical Engineering Hallerova aleja 7, Varaždin, Croatia e-mail: [email protected]

Abstract The paper deals with the experimental investigation and numerical simulation of the mechanical behaviour of samples from a “Mechanical Biological Treatment” waste deposit (MBT waste deposit). Laboratory tests with samples of different moisture contents were conducted to determine the basic geotechnical characteristics as well as the stress-strain relations under loading and reloading conditions. When it is required to evaluate the loaddeformation characteristics of specimens containing relatively large-diameter grain particles, like MBT waste for example, the standard oedometric apparatus with cell (sample) diameter of 56 mm is evidently inadequate. Therefore, at the University of Zagreb, an oedometer of 500 mm in internal diameter was designed, fabricated, assembled and evaluated. As only limited experimental results were available, for numerical simulations a simpli¿ed hypoplastic model was used. For the calibration of this model it was suf¿cient to conduct oedometer tests and to use a simple procedure to estimate the angle of friction. The model proved to be satisfactory for modelling the compaction behaviour of MBT waste observed in experiments. Key words: compressibility, hypoplasticity, MBT waste, oedometer, stiffness 1 Introduction The waste management is actually one of the most important issues related to the protection of the environment. Remarkably high amount of money is invested in waste management, including the separation, recycling, transport and permanent land¿lling of municipal solid waste. According to the Waste Management Strategy of the Republic of Croatia (2005), a more intensive process of rehabilitation and improvement of sanitary conditions on a large number of uncontrolled waste dumps started in 1990s. The objective of the Strategy is to establish framework within which Croatia should reduce the amounts of waste it produces, and to be able to manage it in a sustainable manner. Taking the assessment of current situation and a vision of an appropriate waste management system as a starting point, the Strategy sets the objectives and offers suggestions for their gradual achievement by 2025.

Annual 2010/2011 of the Croatian Academy of Engineering

75

An ideal project of waste management in the Republic of Croatia corresponds to the socalled non-land¿ll concept. In order to achieve this project it would be necessary to close the circle starting from avoidance of waste production, reduction of quantities and harmfulness, recycling and recovery (mechanical, biological, energy recovery) to the use of inert residue. The Strategy has set quantitative targets that determine the dynamics of meeting strategic objectives (Table 1). Deadlines are determined with the expected time delays in relation to EU regulations. Therefore the following basic activities are to be performed: - rehabilitation and closure of the existing open dumps, - extension and improvement of those land¿lls which will remain in operation, - construction of new waste management centres in which the inevitable waste should be land¿lled. Table 1. Quantitative targets for land¿lls Targets Regional centres for waste management County centres for waste management “Of¿cial land¿lls” * Percentage of remedied land¿lls (% from the number determined in 2000).

Year 2005.

2010.

2015.

2020.

2025.

0

1-2

2-3

3

4

0

3-7

7-10

10-14

14-21

187

100

50

30

14-21

5

65

75

85

100

*Group “Of¿cial land¿lls“ consists of the following categories de¿ned above: legal land¿lls, land¿lls in the process of legalization, of¿cial land¿lls, and land¿lls by agreement A waste land¿ll has to be regarded as an engineered structure for which long-term safety has to be proved on the basis of an appropriate geotechnical design. This means that the waste body has to be stable against slope failure. Furthermore, the deformations behaviour and stresses imposed on subsoil, liner systems and internal structures have to be assessed. Figure 1 gives schematic illustration of possible failure mechanisms and deformations problems (Jessberger et al., 1995)

76

Annual 2010/2011 of the Croatian Academy of Engineering

Figure 1. Stability and deformation of a land¿ll

The study of the mechanical behaviour of municipal solid waste by monitoring the existing land¿lls and by laboratory testing of waste samples was named “waste mechanics” in the handbook issued by German Geotechnical Society in the chapter: Principles of Waste Mechanics in the Geotechnical Design of Land¿lls (1993). The results of a survey conducted in the period 1992 - 2000 (for both continental and coastal parts of the Croatia) show that the average annual composition of municipal solid waste has a biodegradable portion of 74,5 % (Domanovac et al., 2002). Waste Management Plan of the Republic of Croatia for 2007-2015 (2007) requires that until 2020, the biodegradable portion of municipal solid waste should be reduced to 35 % of the mass portion of biodegradable municipal solid waste produced in 1997. According to Ordinance on the methods and conditions for the land¿ll of waste categories and operational requirements for waste land¿lls (2007), which is based on Council Directive 1999/31/EC of 26 April 1999 on the land¿ll waste, only pre-treated waste is allowed to be dumped in land¿lls. Most of county waste management plans anticipated mechanical biological treatment (MBT) as the optimum waste treatment method. The goal of mechanical biological waste treatment is the recycling of separated fractions (which leads to the reduction of land¿ll volume) on one side, and the biological stabilisation of the waste on the other side. One of the consequences of the waste treatment is the change of physical, biological and chemical properties of the material to be land¿lled, compared with the wastes which have been land¿lled before. It can be concluded that the knowledge of mechanical parameters of MBT waste for the design of land¿lls, which will be built within waste management centres, is necessary. The extended operation period of a land¿ll due to the reduction of land¿ll volume is directly dependent on the deformability properties of MBT waste which are in the focus of the present research. Most commercial geotechnical laboratories have been utilizing some kind of Casagrandetype oedometer (Casagrande, 1936) for routine measurements of deformability parameters of ¿ne-grained soils for decades. Other names for this test are: one-dimensional compression test, the con¿ned compression test and the consolidation test. The apparatus contains an oedometer cell that can be installed in a loading frame so that a predetermined vertical consolidation stress can be applied to the specimen. The soil specimen is restrained laterally by a steel oedometer cell. The top and bottom surface of the specimen are in contact

Annual 2010/2011 of the Croatian Academy of Engineering

77

with porous stone discs, so that the excess pore water pressure induced in the specimen by the applied loading can be dissipated through simultaneous upward and downward drainage during consolidation. The specimen has to be fully submerged in water during the test to eliminate the surface tension of water at the drainage surfaces, so as to allow unrestricted drainage of water. The test is considered to be one dimensional as both specimen deformation and drainage occur in the vertical direction only. Oedometer samples usually have a diameter of 56 mm and the thickness-diameter ratio 1:3 to 1:4. Based on the above mentioned consideration, the manufacturing of large oedometer of 500 mm in internal diameter, suitable for testing MBT waste samples, was initiated within the scienti¿c project “Characterization of municipal solid waste”. The basic outline of the project was given by Petroviü (2008). The manufacturing of large oedometer started in September 2008 and lasted until March 2009 when it was assembled at the Faculty of Geotechnical Engineering. 2 Basic geotechnical characteristics of tested MBT waste material The MBT waste samples used for testing were imported from an MBT waste land¿ll in Austria as in the Republic of Croatia there is no MBT facility yet installed. Naturally there was no experience in laboratory testing of MBT samples so far. Prior to laboratory testing the waste was dried in the open air for two weeks. After drying the equilibrium moisture content was determined according to ASTM D 2974 as in case of organic soil samples. The samples were dried in the oven at the temperature 1050C. The equilibrium water content was 7 %. Figure 2 shows the particle size distribution curve for the air dried MBT waste material according to HRN.U.B1.018. It can be seen that the largest particle diameter does not exceed 30 mm and therefore ful¿ls the requirement that the largest particle diameter should not exceed the 1/5 of the sample height which, for the oedometer used, is 40 mm. Since the ¿ne-grained portion of the waste material is less than 10 %, a particle size analysis using the hydrometer test procedure was not conducted. According to the Uni¿ed Soil Classi¿cation System (USCS), the MBT waste can be classi¿ed as a coarse grained material, as the particle sizes lie in the range between 0.05 and 30 mm. The uniformity coef¿cient and coef¿cient of curvature are Cu = 26 and Cc = 1.5, respectively. Therefore the tested MBT waste is a well-graded material and it can be well compacted.

Figure 2. Particle size distribution curve of MBT waste

78

Annual 2010/2011 of the Croatian Academy of Engineering

The average density of the solid particles, ȡs, was determined according to the ASTM D 854 on the air free samples with maximum particle size of 4.75 mm using a pycnometer of a known volume. The particles larger than 4.75 mm were removed by sieving. The value ȡs = 2147 [kg/m3] was determined. Detailed description of the laboratory testing of basic MBT waste parameters is given by Petroviü et al. (2011b). 3 Design, fabrication and assembly of a large oedometer Since standards for testing compressibility behaviour of MBT waste do not exist it was concluded that a large oedometer must be designed, manufactured, assembled and evaluated in accordance with standards for testing compressibility behaviour of soils. This is quite in line with the fact that waste mechanics had been developed on the basis of soil mechanics principles. The most relevant standards used for de¿ning the technical speci¿cations of large oedoemeter are British (BS 1377: Part 5: 1990) and American (ASTM D 2435) standards for compressibility testing as well as Croatian standard Eurocode 7: Geotechnical design – Part 2 - Ground investigation and testing (HRN EN 1997-2:2008). Technical speci¿cations of new oedometer were de¿ned in collaboration with several mechanical engineering, industrial management and IT companies from Varaždin. The diameter of the new oedometer cell is 500 mm and the height is 200 mm. It means that minimum height-diameter ratio 1:2,5 required by the British standard in order to minimize the effect of the side friction forces, is satis¿ed. Maximum vertical pressure on sample is 2000 kN/m2. The force (pressure) on specimen is transmitted by means of hydraulic loading actuator. Pressure transducer has a minimum accuracy ± 0,05 % of full range output (FRO). Displacement transducer has a minimum accuracy ± 0,05 % FRO and minimum stroke of 9 cm. The lower and upper lids, as well as the pressure plate, are made from structural steel (all protected against corrosion). Compression tests are fully automated (computer-controlled). Figure 3 shows a ¿nal set-up of the large oedometer system while Figures 4, 5, 6 and 7 present selected parts of a large oedometer during the production phase.

Figure 3. Final set-up of the large oedometer system

Annual 2010/2011 of the Croatian Academy of Engineering

79

Figure 4 presents an upper and lower lid of oedometric cell. Maximum deformation of lids at highest load should not exceeds 0,7 mm. Figure 5 presents hydraulic loading actuator while Figure 6 presents oedometric cell ring polished to burnish so that friction between sample and cell ring is reduced to minimum. Figure 7 presents drainage porous plates which are required for dissipation of the excess pore pressure. Drainage porous plates are made from 1 mm thick stainless steel. The clogging of the drainage holes should be prevented with the use of the ¿lter paper or geotextile placed between the specimen and drainage plate.

Figure 4. An upper and lower lid of oedometric cell

Figure 5. Hydraulic loading actuator

Figure 6. Cell ring

80

Annual 2010/2011 of the Croatian Academy of Engineering

Figure 7. Porous drainage plate

As the innovative parts of the large oedometer are still subject of legal protection (patent registration), detailed technical description of the apparatus can not be given before the completion of the legal procedure. 4 Oedometric compression tests and friction angle 4.1 Sample preparation

In the present paper the properties of two MBT waste samples are presented, where Sample A is characterized by an initial moisture content of w = 65 % and Sample B is characterized by an initial moisture content of w = 41 %. In order to obtain a homogenous distribution of moisture content through the whole specimen after mixing with water, the wetted samples were left untouched for the next 24 hours. Then the samples were installed in the oedometer cell in ¿ve layers. On average, each layer was 4 cm thick and pre-compacted to a de¿ned initial void ratio. The basic geotechnical parameters for both samples at the beginning of the test are presented in Table 2 respectively. Herein ȡ denotes the density, w the moisture content, ȡd the dry density, ȡs the average density of solid particles, e the void ratio and S the degree of saturation. Table 2 Basic geotechnical parameters of tested samples

Sample A Sample B

ȡ [kg/m3] 1366 1093

w [%] 65 31

ȡd [kg/m3] 828 834

ȡs [kg/m3] 2147 2147

e0 / 1,59 1,57

S [%] 88 42

4.2 Consolidation procedure

The initial pressure obtained from the pressure plate was 4,7 kN/m2. A sequence of loadings was applied, where the next load step was applied after the completion of consolidation of the preceding loading step. The loading steps are 36, 82 and 180 kN/m2, followed by unloading to 36 kN/m2, then reloading to 365 kN/m2 and ¿nally unloading to 4,7 kN/m2.

Annual 2010/2011 of the Croatian Academy of Engineering

81

4.3 Oedometric test results and their evaluation

Figure 8 presents stress-strain relationship for samples A and B respectively.

Figure 8. Stress-strain relationship for one-dimensional compression

The graphs on Figure 8 can be used in order to determine secant compressibility moduli for usual load increments (Table 3) Table 3. Secant compressibility moduli Load Increment Sample A Sample B

4,7-50 [kN/m2] 566 657

50-100 [kN/m2] 1190 1316

100-200 [kN/m2] 1515 1639

200-360 [kN/m2] 3666 2860

The values from Table 3 are compared with test results published by other researchers (Table 4). Table 4. Compressibility moduli published by other researchers Load

(0) 25-50

50-100

100-200

200-400

400-600

280-420

Fraction

increment/ Ref.

[kN/m2]

[kN/m2]

[kN/m2]

[kN/m2]

[kN/m2]

[kN/m2]

[mm]

Duelmann (2002)

730

1480

2460

4920

/

/

0-30

840

1990

1870

3290

/

/

0-20

800

1000

1800

/

/

/

0-20

1070

1590

1680

2880

/

/

0-40

500

1100

1600

2800

/

/

0-40

940

1490

2440

3030

/

/

0-60

600

1300

2000

2800

/

/

0-60

KuehleWeidemeier (2003)

82

Annual 2010/2011 of the Croatian Academy of Engineering

Bidlingmaier et al. (1999)

355

1670

1947

2904

4514

/

< 60

239

905

1570

2973

4977

/

< 100

Ziehmann (1999)

/

/

/

/

/

5100

/

/

/

/

/

/

7900

< 60

By comparison of the values of compressibility moduli from Table 3 and Table 4 it can be concluded that the values obtained for the load increments 0 – 50 (25-50), 50 – 100 and 100 – 200 kN/m2 ¿t well with the values of compressibility moduli by other researchers. The values obtained for load increment 200 – 360 kN/m2 are somewhat higher than the values published by Bidlingmaier et al. (1999) and by Kuehle-Weidemeier (2003). Yet according to Duellmann (2002) i Ziehmann (1999) it is possible to obtain even higher values of compressibility moduli for approximately same load increment. In addition, Figure 9 presents comparison of measured stress-strain curves with stress-strain curves of MBT waste published by Duelmann (2002) and Kuehle-Weidemeier (2003).

Figure 9 Comparison of measured stress-strain relationships with stress-strain relationships published by other researchers

Figure 9 shows a good agreement, with respect to the relative deformation, of the selected stress-strain relationships with stress-strain relationships published by Duelmann (2002) and Kuehle-Weidemeier (2003). It should be also emphasized that the largest particle size of MBT samples used by Duelmann (2002) and Kuehle-Weidemeier (2003) corresponds well to the largest particle size of the tested MBT waste material. The elasto-plastic behaviour of MBT waste de¿ned by Duelmann (2002) has been also con¿rmed. Detailed description of the oedometric compression tests of MBT waste samples is given by Petroviü et al. (2010) and Petroviü et al. (2011c). 4.4 Friction angle

In order to estimate the critical friction angle ijc, the angle of repose of the dry waste material was determined. The measurements were conducted on two separate samples at the bottom and at the top of the repose in twelve different points. The values measured were as follows: 370, 310, 310, 410, 380, 380, 350, 400, 410, 380, 500, 430. If the highest and lowest values are considered to be unrealistic and are disregarded, then the mean value is 390 with a standard deviation of ± 2.450.

Annual 2010/2011 of the Croatian Academy of Engineering

83

5 Constitutive model In order to model the behaviour under oedometric conditions ¿rst a relation between the void ratio e and the mean effective pressure p for monotonic compression is considered. In particular, the decrease of e with an increase of p is described by the following exponential function:

(1) where the constant e0 denotes the void ratio for p § 0, hoe has the dimension of stress and n is a dimensionless constant. It can be noted that function (1) is similar to the approximation function for isotropic compression proposed by Bauer (1996), i.e.

(2)

While in the function by Bauer the quantities eio, , and hs are de¿ned for the isotropic compression curve starting from the loosest possible state of the grain skeleton, the parameters eo, p, n and hoe in Eq. (1) are related to monotonic oedometric compression. Herein the mean pressure reads p = - (ı11+ ı22+ ı33)/3 = -(1+2K0) ı11/3, where ı11 is the vertical stress, ı22 = ı33 the lateral stress under zero lateral strain and K0 denotes the pressure coef¿cient at rest, i.e. K0= ı22 / ı11= ı33 / ı11. Usually K0 is obtained in experiments by measuring the vertical and horizontal stresses simultaneously. As this was not carried out in the experiments conducted, it is necessary to make estimation for the value of K0. The most established semi-empirical function in engineering practice is the relation proposed by Jaky (1944), where K0 is related to the critical friction angle ijc according to: K0 = 1 - sin ijc. A comprehensive review of experimental data obtained from different soils, however, shows that K0 not only depends on the friction angle but also on the current density and the pressure level. Bauer (1997), for instance, reported a remarkable agreement between experiments carried out under different densities and the corresponding predictions with the hypoplastic model proposed by Bauer (1996) and Gudehus (1996). The framework of hypoplasticity is therefore suitable to get an approximation for K0 and to model the inelastic and non-linear stress strain behaviour under axisymmetric loading and unloading. In particular for oedometric boundary conditions the constitutive equations (3-4) are proposed, which present a simpli¿ed version of the hypoplastic model by Bauer (1996) and Gudehus (1996).

(3) (4)

84

Annual 2010/2011 of the Croatian Academy of Engineering

Assuming a constant grain density, the balance equation of mass leads to a relation between the rate of the void rate, , and the volumetric strain rate, , i.e. equation (5) can be derived: (5) is the time derivative of the Cauchy stress are normalized Herein, quantities, is the strain rate, and is the rate of void ratio. Factor is related to the critical friction angle ijc and for general stress paths also depends on the Lode angle in the deviator plane (Bauer, 2000). For axisymmetric loading paths factor reduces to (Bauer & Herle, 2000):

(6) In Eq. (3) and Eq. (4) the density factor >0 is related to the limit void ratios as outlined in more detail by Bauer (1996) and Gudehus (1996). In particular 1 for a loose material. In the present paper is assumed to be constant. The compression law (1) is embedded in the stiffness factor using the consistency condition for oedometric compression, which leads to the following relation:

(7) Under monotonic oedometric compression the following expression is obtained from Eq. (3) and Eq. (4) for the stress ratio

(8) K0 in Eq. (8) is also called the pressure coef¿cient at rest. It depends on the critical friction angle ijc and the density factor . From Figure 10 it can be seen that Eq. (8) allows a ¿ner adaptation of K0. In particular higher friction angles ijc and lower values of the density factor give signi¿cantly lower values for K0 while Jaky’s relation is independent of the current density. The hypoplastic model proposed for MBT waste material includes ¿ve constants. Taking into account Eq. (8) for K0, a critical friction angle of ijc = 39º and =1.1, the values eo, hoe and n of the approximation function (1) can be calibrated. For Sample A, a comparison of the results obtained from Eq. (1) with the experiments is shown in Figure 11. For the numerical simulations discussed in Section 6 the constants used are presented in Table 5.

Annual 2010/2011 of the Croatian Academy of Engineering

85

Figure 10. Comparison of K0 values obtained with the hypoplastic constitutive model (HM) for different density factor and friction angles ijc with the function by Jaky (1944)

Table 5. Constants used for numerical simulations

Sample A

hs [kN/m2] 3519

n

e0

fd

0.389

1.75

1.1

[0] 39

Figure 11. Comparison of the oedometric compression relation according to Eq. (1) with the experimental results for Sample A

6 Comparison of numerical simulations with experiments In the following the results obtained from numerical simulations of oedometric compression tests are compared with the corresponding experiments by Petroviü (2010). All simulations are conducted under drained conditions. Figure 12 show a good agreement between the results obtained from the numerical simulation (solid curve) and the experimental data (dots). Is should be noticed that with the presented hypoplastic model only loading and unloading can be modelled. For the simulation of loading cycles an extension of the constitutive model is needed as proposed for instance by Bauer & Wu (1992), Niemunis & Herle (1997).

86

Annual 2010/2011 of the Croatian Academy of Engineering

Figure 12. Oedometric compression and extension of Sample A

Detailed description of the numerical modelling of MBT waste deformability is given by Petroviü et al. (2011a). 7 Conclusions A 500 mm internal diameter Casagrande-type oedometer has been designed, manufactured, and assembled at the University of Zagreb. The manufacturing costs were about 17.000,00 EUR. It was a rather demanding task to produce a laboratory device capable to give reliable results with very limited ¿nancial resources. In order to achieve this goal many innovative and unusual technical solutions were applied. Some of the innovative parts of the large oedometer are currently in the process of legal protection (patent registration). The laboratory results obtained with the new device prove that even with limited resources it is still possible to achieve a respectable scienti¿c result as long as scientists keep their minds wide open. The results obtained from the evaluation program indicate that they are comparable to the results published by other authors. The particle size distribution curve con¿rmed that the MBT waste can be classi¿ed as a coarse grained material. A good agreement of stressstrain curves for samples A and B proves that new apparatus can provide repeatability of the oedometer test. It has been veri¿ed that the magnitude of vertical relative deformation is in the range 15 - 25 % for given load range. The elasto-plastic behaviour of MBT waste has been also con¿rmed. Simpli¿ed hypoplastic model proved to be satisfactory for modelling the compaction behaviour of MBT waste observed in experiments A complete presentation of the laboratory testing and numerical modelling of MBT waste deformability together with the recommendations for future research is given by Petroviü (2010). 8 Acknowledgements The ¿nancial support of the Ministry of Science, Education and Sports of the Republic of Croatia for the project "Characterization of municipal solid waste" (160-0831529-3031) is gratefully acknowledged.

Annual 2010/2011 of the Croatian Academy of Engineering

87

References 1. 2. 3. 4. 5.

6. 7. 8.

9.

10. 11. 12. 13. 14. 15. 16. 17. 18.

Bauer E (1996) Calibration of a comprehensive hypoplastic model for granular materials. Soils and Foundations., Vol. 36(1), pp 13-26 Bauer E (1997) The critical state concept in hypoplasticity. 9th International Conference on Computer Methods and Advances in Geomechanics, IACMAG 97, Wuhan, China, Balkema, pp 691–696. Bauer E (2000) Conditions for embedding Casagrande’s critical states into hypoplasticity. Mechanics of Cohesive-Frictional Materials, 5. pp 125-148 Bauer E, Wu W (1992) A hypoplastic model for granular soils under cyclic loading, In: Kolymbas (ed.) Proceedings of the International Workshop on Modern Approaches to Plasticity, Elsevier, pp 225-245 Bidlingmaier W, Scheelhaase T, Maile A (1999) Langzeitverhalten von mechanischbiologisch vorbehandeltem Restmuell auf der Deponie, Abschlußbericht zum Teilvorhaben 3.1 des BMBF-Verbundvorhabens „Mechanisch-biologische Behandlung von zu deponierenden Abfaellen“, Universitaet Gesamthochschule Essen, Fachbereich 10 – Bauwesen, Fachgebiet Abfallwirtschaft, 1999 Casagrande A (1936) The Determination of Preconsolidation Load and its Practical Signi¿cance. Proceedings, 1st International Conference on Soil Mechanics Foundation Engineering, Cambridge, Vol. 3, pp. 60–64. Council Directive 1999/31/EC of 26 April 1999 on the land¿ll waste, Of¿cial Journal of the European Communities, L 182, 16.7.1999. Domanovac T, Orašanin R (2002) Municipal solid waste composition for continental and coastal part of Croatia. In: Milanoviü Z. (ed.) Proceedings of VIIth International Symposium Waste Management Zagreb 2002, mtg-topgraf, Zagreb, pp 61-68 (in Croatian) Duellmann H (2002) Untersuchungen zum Einbau von MBA-Abfaellen auf der Zentraldeponie Hannover, Laboruntersuchungen zum Verdichtungs-, Durchlaessigkeits-, Last-Setzungs- und Scherverhalten. Februar 2002. Im Auftrag des Abfallwirtschaftsbetriebes Hannover Geotechnics of Land¿ll Design and Remedial Works Technical Recommendations – GLR, Second Edition, Ernst & Sohn, Berlin, 1993. Gudehus G (1996) A comprehensive constitutive equation for granular materials. Soils and Foundations, Vol 36(1), pp 1-12 Jaky J (1944) The Coef¿cient of Earth Pressure at Rest. Journal for Society of Hungarian Architects and Engineers, pp 355-358 Jessberger HL, Kockel R (1995) Determination and assessment of the mechanical properties of waste materials. In: Sarsby (ed.) Waste Disposal by Land¿lls – GREEN'93, Balkema, Bolton, pp 313-322 Kuehle-Weidemeier M. (2003) Land¿lling and properties of MBP waste, Proceedings Sardinia 2003, Ninth International Waste Management and Land¿ll Symposium, S. Margherita di Pula, Cagliari, Italy Niemunis A, Herle I (1997) Hypoplastic model for cohesionless soils with elastic strain range. Mechanics of Cohesive-Frictional Materials, Vol. 2(4), pp 279-299 Ordinance on the methods and conditions for the land¿ll of waste, categories and operational requirements for waste land¿lls (2007) Of¿cial Gazette of the Republic of Croatia, No. 117/07 (in Croatian) Petroviü I (2008) Deformability of mechanically and biologically treated municipal solid waste, In: Milanoviü Z. (ed.) Proceedings of Xth International Symposium Waste Management Zagreb 2008, Gospodarstvo i okoliš, Zagreb, pp 329-334 (in Croatian)

88

Annual 2010/2011 of the Croatian Academy of Engineering

19. Petroviü I (2010) Modelling the behaviour of mechanically and biologically treated municipal solid waste. PhD Thesis, University of Zagreb, Faculty of Civil Engineering, Zagreb, Croatia (in Croatian) 20. Petroviü I, Kovaþiü D (2010) Measurement of stiffness modulus of mechanically and biologically treated waste, In: Milanoviü Z. (ed.) Proceedings of XIth International Symposium Waste Management Zagreb 2010, Udruga za gospodarenje otpadom, Zagreb,(in Croatian) 21. Petroviü I, Bauer E (2011a) A simple hypoplastic model for simulating the mechanical behaviour of MBT waste, In: The Second International Symposium on Computational Geomechanics (ComGeo II), Cavtat, Croatia, 27-29 April, 2011 22. Petroviü I, Szavits-Nossan V, Štuhec D (2011b) Laboratory testing of waste after biomechanical treatment. Graÿevinar 63(1), pp 43-53 (in Croatian) 23. Petroviü I, Szavits-Nossan V, Kovaþiü D (2011c) Deformability of mechanically and biologically treated municipal solid waste, Graÿevinar 63(3), pp 255-264 (in Croatian) 24. Waste Management Strategy of the Republic of Croatia (2005) Of¿cial Gazette of the Republic of Croatia, No. 130/05 (in Croatian) 25. Waste Management Plan of the Republic of Croatia for the period 2007-2015 (2007) Of¿cial Gazette of the Republic of Croatia, No. 85/07 (in Croatian) 26. Ziehmann G (1999) Veraenderung des mechanischen Verhaltens durch die mechanische und biologische Vorbehandlung, Deponierung von vorbehandelten Siedlungsabfaellen. Veroeffentlichungen des Zentrums fuer Abfallforschung der Technischen Universitaet Braunschweig, Heft 14, S. 1 – 9

Annual 2010/2011 of the Croatian Academy of Engineering

89

BEAM-TO-COLUMN JOINT MODELLING TO EC3 Darko Dujmović University of Zagreb, Faculty of Civil Engineering; Croatian Academy of Engineering, Zagreb, Croatia

Boris Androić IA Projektiranje Structural Engineering L.t.d.; Croatian Academy of Engineering, Zagreb, Croatia

Ivan Lukačević University of Zagreb, Faculty of Civil Engineering, Zagreb, Croatia

ABSTRACT The modelling of beam to column joints is analyzed taking into account real behaviour of such joints as described by the Mj-ij curve. General considerations relating to the new joint design philosophy, as based on Eurocode 3, are given in the initial part of the paper. The method of components, an analytical method for characterization of joints, is then described. Components enabling bolted joint between beams and columns are identi¿ed and analyzed. The use of the described modelling procedure is shown by means of a specially devised computer program CoP (Connection Program). Key words: beam to column joint, modelling, component method, bolted joint, application of the new approach 1 INTRODUCTION Laboratory research and the progress of numeric methods have stimulated the development of a more realistic approach to dividing joints for everyday engineering practice. The component method is an ef¿cient way of solving complex behaviour of steel structure joint. The component method, as well as the mechanical models which describe individual joint types can be used to de¿ne basic joint characteristics. These are: stiffness, bending resistance and rotation capacity. 2 BEAM-TO-COLUMN JOINT MODEL Joint analysis requires the following to be taken into consideration: • material non-linearity (plasticity, strain-hardening), • non-linear contact and slip, • geometrical non-linearity (local instability), • residual stresses, • complex geometrical con¿gurations. For practical purposes, the joint analysis is conducted on simpler models. This is the reason the design procedure known as the components method has been adopted in Eurocode 3 (EC3). The method suits the simpli¿ed mechanical spring and rigid links model very well. The joint is represented by adequate rigid and Àexible components. They represent individual joint parts, Fig. 1.

90

Annual 2010/2011 of the Croatian Academy of Engineering

Fig. 1. Component model of beam-to-column joint

Typical bolted joint components are: • column web panel in shear, • end-plate in bending, • beam Àange in bending, • beam web in tension, • beam Àange and web in compression, • bolts in tension • welds. When using the component method, the ¿rst characteristics to be considered are those of resistance, stiffness and ductility in individual components. These characteristics are combined to reach an estimate of mechanical joint behaviour. A non-linear force-deformation (F-ǻ) curve can be used to show individual joint behaviour, Fig. 2.

Fig. 2. Force-deformation (F-ǻ) curve of ductile component: a) actual behaviour and b) bi-linear approximation

Simpler behaviour idealisations are possible. The simpli¿ed components model of EN 1993-1-8, for instance, combines the joint bending response with the column web panel response to shear. This type of equivalent rotational spring has been shown in Fig. 3.

Annual 2010/2011 of the Croatian Academy of Engineering

91

Fig. 3. Equivalent rotational spring model

Fig. 4. Application of the component method to a welded joint

The application of the component method requires the following steps: • Identi¿cation of the active components from a global list of components (20 different components currently codi¿ed in EN 1993-1-8). • Evaluation of the force deformation response (F-ǻ) of each individual basic component (initial stiffness, design resistance etc.). • Assembly of the active components, using representative mechanical model shown in Fig. 2, for the evaluation of the mechanical characteristics of the whole joint (initial stiffness, design resistance etc. or the whole deformability M-ij curve). These steps are illustrated in Fig. 4. in simple case of a beam-to-column welded joint.

92

Annual 2010/2011 of the Croatian Academy of Engineering

Understanding the mechanical behaviour of different joint components allows the analysis of a large number of different joint con¿gurations. The purpose of the component method is to determine the F-ǻ curve for each individual spring. Only the linear stiffness of each component is required to estimate initial joint stiffness. The ductility estimate, however, requires knowledge of the non-linear F-ǻ response for each component. The following points encompass only the components that are relevant for the beam-tocolumn joint. For the joint ductility estimate, the joint components are divided into three groups, analogous with the cross-section classi¿cation: • components with high ductility, • components with limited ductility, • components with brittle failure. 3 DESIGN RESISTANCE OF COMPONENTS 3.1 Components with high ductility

These components are represented by the F-ǻ curve which in the ¿rst part shows linear elastic behaviour and in the second part allows an increase in deformation with an increase in force. The component deformation capacity is very large. • Column web panel in shear Extensive tests show the column web having a resistance reserve in shear. The following equation gives the design shear resistance of unstiffened column web panel, Vwp,Rd, for single-sided framed joints:

(1) where: fy,wc the yield strength of the column web, Avc the shear area of the column, ȖM0 the partial factor for resistance of cross-sections whatever the class is. In the case of welded cross-sections, the shear column area, Avc, coincides with the web area, whereas in the case of rolled sections it is given by: (2) where: Ac bc tfc twc rc

the total area of the column, the Àange width, the Àange thickness, the web thickness, the root-radius of the web-Àange junction.

Eq. (1) disregards the axial force in the column. Using the Von Mises yield criterion it is possible to determine the reduced resistance value which takes into consideration the axial column force. The 0,9 reduction coef¿cient partially solves this problem, which has also

Annual 2010/2011 of the Croatian Academy of Engineering

93

been accepted into EN 1993-1-8. The resistance of the unstiffened column web in shear for single-sided joints, Vwp,Rd, is:

(3) The shear deformation of the column web, ĭs, in the overall initial rotation of joint is given by:

(4) where: V the shear force on the column web taken as 2ȈFi (Fi denoting the force in each bolt row and i the bolt row), G the shear modulus. Thus the corresponding axial stiffness, Kwp, is: (5) where: z the lever arm between the compressive and the tensile areas, E the modulus of elasticity. Dividing the axial stiffness, Kwp, with the modulus of elasticity E the Eq. (5) is: (5a) where: kwp the stiffness coef¿cient of column web in shear, the designation in EN 1993-1-8 is k1. From Eq. (4) it can be observed that the stiffness of this component depends on the applied shear force on the column web. Given that, in general, internal forces transmitted by the lower and upper column and (for internal nodes with unbalanced moments) left beam may also be present, the applied shear force must also be modi¿ed by a transformation parameter ȕ, EN 1993-1-8, to deal with this effect. For different joint con¿gurations and unbalanced bending moments, the stiffness coef¿cient, kWP, Eq. (5a), must be modi¿ed by the transformation parameter ȕ:

94

Annual 2010/2011 of the Croatian Academy of Engineering

(6) For stiffened column web shear deformation can be disregarded (kwp = ’). It ought to be noted that in slender webs instability becomes relevant, but has so far not been included in EC 3. • End-plate in bending The deformation of this component is evaluated using the simple substitute model, T-stub. It represents the tension joint zone behaviour, and is shown in Fig. 5.

Fig. 5. Substitute model, the T-stub

Three possible failure modes of the equivalent T-stub Àange are assumed, as shown in Fig. 6.

Fig. 6. Failure modes of the T-stub Àange

The following failure modes are possible: Mode 1: End-plate yielding without bolt failure

Annual 2010/2011 of the Croatian Academy of Engineering

95

(7a) Mode 2: Simultaneous yielding of end-plate with bolt failure

(7b) Mode 3: Bolt failure without end-plate yielding (7c) where: m the distance between the bolt centreline and the face of the weld connecting the beam web to the end-plate, see Fig. 7, n the effective distance to the free edge, , but , see Fig. 7, Ft,Rd the design tension resistance of a bolt, ™Ft,Rd the total value of Ft,Rd for all bolts in the T-stub. Design bending resistance of an end-plate depending on failure mode is: (8a) (8b) where: tf fy

the effective width of the end-plate in bending, the thickness of the end-plate, the yield strength of the end-plate.

Dimensions of an equivalent T-stub Àange are shown in Fig. 7. For end-plate bending resistance, of the possible three failure modes, the minimum value is used. The T-stub stiffness coef¿cient analytical equations can be derived from classical beam theory, when the effective width is correctly estimated, so that:

(9)

96

Annual 2010/2011 of the Croatian Academy of Engineering

where: the stiffness coef¿cient of the end-plate in bending, the designation in EN 1993kb,p 1-8 is k5, tp is the thickness of a end-plate.

Fig. 7. Dimensions of an equivalent T-stub Àange

• Column Àange in bending This component behaves similarly to the end-plate in bending, so the equivalent T-stub approach can also be utilized. The exception is the case when the Àange is stiffened. The unstiffened Àange is supposed to have the same levels of ductility and stiffness. The resistance equations dependent on failure modes are given in the following: Mode 1: Column Àange yielding without bolt failure

(10a) Mode 2: Simultaneous yielding of column Àange with bolt failure

(10b) Mode 3: Bolt failure without column Àange yielding (10c) For column Àange bending resistance the value which is used is the minimum out of the three failure modes.

Annual 2010/2011 of the Croatian Academy of Engineering

97

The component stiffness coef¿cient is:

(11) The symbols have the same meaning as in the case of the end-plate, with the exception that the column and not the end-plate Àange characteristics are used. The designation of this stiffness coef¿cient in EN 1993-1-8 is k4. • Beam web in tension For bolted end-plate joints, the design tension resistance of the beam web, Ft,wb,Rd, is:

(12) where: beff,t,wb the effective width of the beam web in tension, equal to the effective length of the equivalent T-stub representing the end-plate in bending, twb the thickness of beam web, fy,wb the yield strength of the beam web. The initial stiffness for this component can be taken to be in¿nite. 3.2 Components with limited ductility

These components are characterized by an F-¨ curve with limit point after which the curve falls. • Column web in compression This component achieves limited ductile behaviour when the curve falls after the maximum resistance has been reached. The resistance for this component can be grouped according to two different types of criteria: - crushing resistance, - buckling resistance. The crushing resistance must take into account the interaction between: - the local stresses that arise from the shear stresses in the panel zone, - the vertical normal stresses due to axial load and bending moment in the column, - the horizontal normal stresses transmitted by the beam Àanges. Using the von Mises yield criterion, the design crushing resistance, Fc,wc,Rd, is:

(13) The effective width of the column web in compression, beff,c,wc, for bolted end-plate joints is:

98

Annual 2010/2011 of the Croatian Academy of Engineering

(14) where: ap the effective thickness of the weld, Fig. 8, s=r for rolled column sections, sp the length obtained by dispersion at 45º through the end-plate.

Fig. 8. Transverse compression on an unstiffened column

Reduction factor, kwc, accounts for the inÀuence of vertical normal stress due to axial force and the bending moment, ıcom,Ed, and is given in the following:

(15)

Reduction factor Ȧ accounts for the shear interaction, and equals:

Annual 2010/2011 of the Croatian Academy of Engineering

99

(16)

with

(17a)

(17b) where: ȕ is the transformation parameter from EN 1993-1-8. The buckling resistance, Fc,wc,Rd, is taken approximately using the Winter equation:

(18) The reduction factor for plate buckling (column web panel), ȡ, is:

(19a) or (19b) The normalised plate slenderness (column web panel),

, is:

(20) The initial deformation of the ĭc component is:

100

Annual 2010/2011 of the Croatian Academy of Engineering

(21) where: N the resultant compressive force, taken as 2ȈFi (Fi denote the force in each bolt row and i the bolt row), Ac the effective web area in compression zone, Ac = twc•beff,c,wc, the depth between column ¿llets, dc hc the beam depth minus beam Àange thickness. So that the initial (axial) stiffness becomes:

(22) Dividing the axial stiffness, Kc,wc, with the modulus of elasticity E the Eq. (22) is: (22a) It must be noted that for the calculation of the stiffness coef¿cient, kc,wc, a reduction of the effective width used for the resistance calculation is adopted (0,7×beff,c,wc). The designation of this stiffness coef¿cient in EN 1993-1-8 is k2. • Column web in tension Excluding instability occurrences, the resistance of this component approaches the column web in compression. The design resistance of an unstiffened column web, Ft,wc,Rd, equals:

(23) where the symbols mean the same as earlier, with t (tension) replacing c (compression). It shall be noted that EN 1993-1-8 disregards the inÀuence of vertical stresses arising from the column. Analogous to the preceding case, the initial deformation of this component ĭw is: (24) where: T the resultant tensile force, taken as 2ȈFi (Fi denoting the force in each bolt row and i the bolt row), At the effective web area in the tensile zone, At = twc×beff,t,wc,

Annual 2010/2011 of the Croatian Academy of Engineering

dc ht

101

the depth between column ¿llets, the distance from the tensile force to the centre of compression.

Initial (axial) stiffness equals:

(25) Dividing the axial stiffness, Kt,wc, with the modulus of elasticity E the Eq. (25) is: (25a) This means that Eq. (25a) provides the stiffness coef¿cient, kt,wc, of the column web in tension. The designation of this stiffness coef¿cient in EN 1993-1-8 is k3. • Beam Àange and beam web in compression This component, the beam Àange and web in compression adjacent to the connection of beam, provides a limitation to the resistance of the joint. It is therefore necessary to determine maximum component resistance, using:

(26) Its initial stiffness is taken as in¿nity. 3.3 Components with brittle failure

These components behave linearly until collapse. Before failure they show very small deformation. • Bolts in tension The bolts show a linear force-deformation (F-¨) response up to failure. The resistance and initial stiffness of each bolt are:

(27) (28) where: As the tensile area of the bolt, fub the ultimate tensile strength of bolts, Lb the sum of the thickness of the connected plates (the column Àange and the endplate), the thickness of the washers and the half thickness of the nut and the bolt head.

102

Annual 2010/2011 of the Croatian Academy of Engineering

• Welds Welds are virtually undeformable (Kw = ’). A rigid-plastic model is adopted as adequate, and the design resistance is:

(29) where: a the effective thickness of the weld, the ultimate tensile strength of the weld, fu ȕw the correlation factor. 4 ASSEMBLY OF THE COMPONENTS Components assembly is the third and the last step of the components method. It consists of assembling components in such a way that the mechanical characteristics of the entire joint can be reached. The relationship between the components characteristics and the joint components is based on “the distribution of internal forces in the joint”. The manner in which the external forces bearing on the joint are distributed to joint components is determined for the given set (analogy with structural member cross-section). Fig. 9 shows the example of a spring model for bolted joint with end-plate. Each component possesses stiffness coef¿cient, ki. The springs are connected serially and/or parallely, and the initial rotational stiffness, Sj,ini, is:

(30)

The bending resistance of the joint, Mj,Rd, equals: (31) where: Ftr,Rd the effective design tension resistance of bolt-row r, hr the distance from bolt-row r to the centre of compression, r the bolt-row number.

Annual 2010/2011 of the Croatian Academy of Engineering

103

Fig. 9. Components represented by springs on bolted beam to column joint

5 PRACTICAL APPLICATION OF THE COMPONENTS METHOD The example shows the new approach to steel beam-to-column joint design adopted in EC3, in practical application. The beam to column joint in Fig. 10 is a real industrial portal frame joint with an inclined beam (Į = 5°), the length of L = 19,5 m, and h = 6,32 high (in the ridge H = 7,20 m). A haunch 1,5 m long supports the beam. The design of this kind of joint is based on the traditional approach and was intuitively held to be rigid. The second step is the design of mechanical joint characteristics (joint No. 1) conducted based on the EC3, Part 1.8, and the values are: • Initial rotational stiffness Sj,ini = 260 863 kNm/rad, • Bending resistance Mj,Rd = 554 kNm. In order to agree with the designer’s expectations, every value has to be higher than: • the inÀexibility limit, based on the rotational stiffness 129 769 kNm/rad, • the maximum bending moment, transferred to the joint in global frames analysis, equals Mj,Ed = 416 kNm. Since both conditions were ful¿lled, the joint design can be considered satisfactory.

104

Annual 2010/2011 of the Croatian Academy of Engineering

Fig. 10. Detail of the joint

The joint stiffness proof based on the EN 1993-1-8 has not been conducted, and the joint was simply taken as inÀexible based on the designer’s experience. According to the EC3 additional efforts are needed to check whether the rigid joint could be considered unnecessary. Before any further discussion, the following needs to be established: • A lack of knowledge necessary to design joints, when the designer’s joint geometry is de¿ned, leads to a systematic application of transversal stiffeners of the column (in both the tension and compression joint zones), together with thick end-plates, • As a result of excessive stiffening the internal force distribution is of a linear kind, which explains why the elastic design approach has traditionally been used for determining bolt forces, In such design procedures, the bolts are weak joint components and the plate approach to design can be considered valid, but it often results in an excessive dimensioning of joints and uneconomical joint framings, since the execution depends on the stiffness level to such an extent. Stiffness even prevents easy assembly at site. Finally, the lack of ductility related to bolt failure is far from satisfactory, when joint ductility is taken into consideration. Part 1.8 of the EC 3 suggests as an alternative a set of rules for design, where the type of failure has not been chosen in advance. The stiffness and resistance characteristics for all participating components have been taken into account in the design. In other words, real joint characteristics are calculated, and the designer is left with the decisions to change the joint geometry as needed, choosing a ductile type of failure, thus making the joint more resistant, stiffer and more ductile. In the case at hand, the joint characteristics exceed requirements (the resistance is higher than required, and the rotational stiffness of 129 769 kNm/rad is enough to secure rigidity in the joint, as was supposed in the frame analysis). Simplifying joint geometry thus decreases the joint execution costs, and those of the entire structure based on that. A more signi¿cant fact here is that in this way we design joints that will achieve the required levels of reliability. Intuitive stiffness forecasts, though made to create a more reliable joint, may have an effect that is just the opposite. The following table shows different structural joint con¿gurations. They are under consideration with the goal of simplifying joint geometry, taking into account the frame analysis requirements.

Annual 2010/2011 of the Croatian Academy of Engineering

105

• Joint No. 1: 6 bolt rows, lengthened end-plate, 6 stiffeners of column web (three pairs). • Joint No. 2: 5 bolt rows, 6 stiffeners of column web. • Joint No. 3: 5 bolt rows, 4 stiffeners of column web (compression and tension zone). • Joint No. 4: 5 bolt rows, 2 stiffeners of column web (compression zone). • Joint No. 5: 5 bolt rows, no stiffeners of column web. • Joint No. 6: 4 bolt rows, 2 stiffeners of column web (compression zone). • Joint No. 7: 3 bolt rows, 2 stiffeners of column web (compression zone). • Joint No. 1, which the designer has suggested, ful¿ls the frame structure analysis requirements, where the stiffness is such that the joint can be considered rigid, and the bending resistance is greater than the bending moment, which was reached through global frame analysis. • Since the joint resistance is signi¿cantly greater than required, it is concluded that the extended end-plate is unnecessary, and the upper bolt row is eliminated, joint No. 2. This decreases joint stiffness, but within rigidity limits. • The third step (joint No. 3) is to reduce the number of stiffeners, by neither eliminating the middle stiffener, which does not decrease joint stiffness nor bending resistance. The middle stiffener is thus ineffective. • Next, the stiffener is eliminated from the tension zone (joint No. 4), and joint bending resistance and stiffness are reduced, although still completing the requirements of global analysis. • Following that, the stiffener would also be eliminated from the compression zone. That sort of joint, however (joint No. 5), does not satisfy requirements, since the resistance is lower than the action effect. • Because the relevant mode of failure is the column web in compression, now a step is taken back to joint No. 4 with stiffeners in the compression zone, with the elimination of 1 row of bolts. That kind of joint (joint No. 6) is entirely satisfactory. • The number of bolts is then reduced to 3, and joint no. 7 is no longer satisfactory as far as bending resistance is concerned (the spacing between the ¿rst and the second row of bolts have been increased from 90 to 120 mm).

106

Annual 2010/2011 of the Croatian Academy of Engineering

Table 1. Joint con¿gurations with corresponding resistance and stiffness values, as well as frame analysis requirement ful¿lment

Joint No. 6 represents the solution. Bending resistance is reduced to an acceptable level, and as far as stiffness goes, the joint still belongs to the rigid joint. For each of these structural solutions, Table 1 gives the resistance and stiffness values, as well as whether the frame analysis requirements have been ful¿lled. This analysis shows very clearly that even for rigid joints, the new approach to joint design based on EC 3 can have serious bene¿ts. This is the result of the clearer de¿nition of the inÀexibility notion, and the usage of advanced design models for joint stiffness and resistance determination. 6 CONCLUSION An entirely different approach that consists in a realistic representation of joints in structure modelling is shown. Joints in a broader sense connect all types of non-monolithic structural elements and are therefore the most important within structural elements and

Annual 2010/2011 of the Croatian Academy of Engineering

107

within the entire structure. In spite of that and owing to the lack of knowledge about their actual characteristics and the way of modelling, joints have been so far neglected and very simpli¿ed assumptions related to forming of models have been adopted. This has resulted in expensive structural solutions that often have very low reliability. This new approach can be successfully applied in every-day engineering practice and further savings can be made due to the decrease in the self weight of structures or reduced costs of execution and assembly. The shown example represents the practical application of the new approach according to EC 3. Further research of joint behaviour should be conducted by use of methods of structural reliability theory. With the collection of statistical data on joint components that would be obtained in laboratory, their stochastic models could be established. By solving the limit state equations the comparison of the reliability level of different types of joints would be make possible. This way optimal solutions can be chosen in line with general reliability requirements according to EN 1990: 2001, Eurocode: Basis of structural design. REFERENCES 1. 2. 3.

4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14.

Zoetemeijer, P. (1974) A design method for the tension side of statically-loaded bolted beam-to-column joints, Heron, pp 1-59 Yee, Y.L.; Melchers, R.E. (1986) Moment-rotation curves for bolted connections, Journal of Structural Engineering 112 3, pp 615-635 Janss, J.; Jaspart, J.P. (1987) Strength and behaviour of in plane weak axis joints and of 3-D joints, In: Bjorhovde, R.; Colson, A.; Zandonini, R. (Editors): Connections in steel structures, Proceedings of the International Workshop on Joints, Elsevier Applied Science, New York Zoetemeijer, P. (1990) Summary of the research on bolted beam-to-column connections, Report 25-6-90-2. Faculty of Civil Engineering, Stevin Laboratory - Steel Structures, Delft University of Technology, Delft European Convention for Constructional Steelwork (1992) Analysis Design of Steel Frames with Semi-rigid Joints, Publication 67, Brussels: ECCS Weinand, K. (1992) SERICON – Databank on Joints in Building Frames, Proceedings of the 1st COST C1 Workshop, Strasbourg, pp 28-30 Shi, Y.J.; Chan, S.L.; Wong, Y.L. (1996) Modelling for moment rotations characteristic for end-plate joints, Journal of Structural Engineering 122 11, pp 1300-1306 Bursi, O. S.; Jaspart, J. P. (1997) Benchmarks for Finite Element Modelling of Bolted Steel Connections; Journal of Constructional Steel Research, Vol. 43, No. 1-3, pp 1742 Cruz, P. J.S; da Silva, L. S.; Rodrigues, D. S; Simões, R. A. D. (1998) Database for the semi-rigid behaviour of beam-to-column connections in seismic regions, Journal of Constructional Steel Research 46 1-3, pp 233-234 Huber, G.; Tschemmernegg, F. (1998) Modelling of steel connections, Journal of Constructional Steel Research 45 2, pp 199-216 Kuhlmann, U.; Davison, J.B.; Kattner, M. (1998) Structural systems and rotation capacity, Proceeding of COST Conference on Control of the Semi-rigid Behaviour of Civil Engineering Structural Connections, Liège, pp 167-176 Weynand, K.; Jaspart, J.-P.; Steenhuis, M. (1998) Economy Studies of Steel Building Frames with Semi-Rigid Joints, Journal of Constructional Steel Research 46 1-3 Kuhlmann U. (1999) InÀuence of axial forces on the component: web under compression, Proceeding of COST-C1 Working Group Meeting, C1/WG2/99-01, Thessaloniki Faella, C.; Piluso, V.; Rizzano, G. (2000) Structural Steel Semirigid Connections, The-

108

15. 16. 17. 18. 19. 20. 21. 22. 23.

Annual 2010/2011 of the Croatian Academy of Engineering

ory, Design and Software, CRC Press LLC, 2000 N.W. Corporate Blvd., Boca Raton, Florida 33431 CoP Software, (2002) Version 2002R03, RWTH Aachen, MSM Liège, ICCS Hoofddorp da Silva, L. S.; Lima, L.; Vellasco, S.; Andrade, S. (2002) Experimental behaviour of end-plate beam-to-column joints under bending and axial force, Database reporting and discussion of results. Report on ECCS-TC10, Meeting in Ljubljana European Committee for Standardization (CEN) (2002) “EN 1990: 2002, Eurocode 0: Basis of structural design“, Final draft, Brussels Dujmoviü, D.; Androiü, B.; Skejiü, D. (2003) Modeliranje prikljuþaka þeliþnih okvirnih konstrukcija, Graÿevinar 55 6, pp 339-348 Dujmoviü, D.; Skejiü, D.; Androiü, B. (2003) Modeliranje prikljuþka nosaþ-stup prema Eurokodu 3, Graÿevinar 55 7, pp 397-405 Girão Coelho, A.M. (2004) Characterization of the ductility of bolted end plate beamto-column steel connections, PhD Thesis. University of Coimbra, Coimbra, Portugal Girão Coelho, A. M.; Bijlaard, F. S. K.; da Silva, L. S. (2004) Experimental assessment of the ductility of extended end plate connections, Engineering Structures 26, pp 1185-1206 European Committee for Standardization (CEN) (2005) EN 1993-1-8:2005, Eurocode 3: Design of steel structures, Part 1.8: Design of joints, May 2005, Brussels Skejiü, D.; Dujmoviü, D.; Androiü, B. (2008) Reliability of the bending resistance of welded beam-to-column joints, Journal of Constructional Steel Research, Vol.64, No.4, pp 388-399

Annual 2010/2011 of the Croatian Academy of Engineering

109

MODELLING OF JOINT BEHAVIOUR IN STEEL FRAMES Darko Dujmovic University of Zagreb, Faculty of Civil Engineering; Croatian Academy of Engineering Zagreb, Croatia.

Boris Androic IA Projektiranje Structural Engineering L.t.d., Croatian Academy of Engineering, Zagreb, Croatia

Josip Piskovic University of Zagreb, Faculty of Civil Engineering, Department for Structural Engineering, Zagreb, Croatia.

ABSTRACT The interaction between steel frames and their joints is explained, and the joint dimensioning method according to EC 3 is presented. Four basic terms related to novel approach to joint dimensioning are described. These terms are: characterization, classi¿cation, modelling and idealisation. An emphasis is placed on the advantages of component method, adopted in EC 3, when compared to the procedure currently used for joint modelling. Economic bases for the use of the new joint dimensioning method are also explained. KEY WORDS: Joints, frame structures, component method, characterization, execution costs 1 INTRODUCTION Traditionally the beam-column joints have been considered as pinned without any resistance and stiffness (simple joints) or as completely rigid with full resistance (continuous joints). This approach is especially convenient for frame structures that are classi¿ed as braced and non-sway, where most beam-column joints do not demand for a moment transfer (pinned joints). Joints that must transfer moments are usually a constituent part of a bracing system. Therefore, they must be completely rigid. In reality, both of these assumptions for simple and continuous joints can be incorrect and uneconomical, and they present only the limit cases of realistic behaviour of joints described by the relationship of the bending moment M and the rotation angle ij. Because of insuf¿cient knowledge about the behaviour of joints in the region between limit cases, procedures have been adopted within the traditional approach to designing which completely neglected this range, and in many cases they provide uneconomical solutions or solutions that are not so reliable. The basic problem was the fact that had not been taken into consideration and that is that the structural properties of joints must be harmonised with structural properties of elements that are being considered. That means that this insuf¿ciency was a particular problem in terms of global structural analysis. By introducing a new engineering discipline called structural modelling, the insuf¿ciencies of the traditional approach had began to be consid-

110

Annual 2010/2011 of the Croatian Academy of Engineering

ered. Numerous laboratory researches and the development of numerical methods encouraged a faster development of the idea regarding a more realistic division of joints for the needs of every-day engineering practice. European standards for structures prescribed a system of joint division that classi¿es joints regarding stiffness (nominally pinned, semi-rigid and rigid), and resistance (nominally pinned, partial-strength and full-strength). Also, joints can be classi¿ed regarding ductility, although the criteria for this classi¿cation have not clearly been prescribed within Eurocode 3 yet. 2 STRUCTURAL BEHAVIOUR OF JOINTS A connection is de¿ned as a set of physical components that mechanically fastens the elements that it connects. It is considered that the connection is concentrated at the place where fastening is realized, for example, where the end of the beam and the column for the beam-column joint around the stronger axis meet. When the connection and the appropriate zone of interaction between the connected structural elements are considered together, the expression joint is used, Figure 1. Depending on the number of elements in the plane that are connected, single-sided, Figure 1.a), and double-sided joints con¿gurations, Figure 1.b), are de¿ned. It is a well-known fact that rotational behaviour of actual joints ranges between two limits: ¿xed and pinned. Figure 2. illustrates such behaviour in an elastic area of a single-sided joint of a column and a beam.

Fig. 1. Joints and connections

Fig. 2. Classi¿cation of joints according to rotational stiffness

Figure 2.a) presents a rigid joint. It is assumed that all elements of the joint are stiff enough. There is no difference between rotations at the ends of the structural elements connected

Annual 2010/2011 of the Croatian Academy of Engineering

111

in the joint. The joint rotation is a node rotation usually used in frame structure analysis methods. A joint that does not have stiffness, Figure 2.b), is nominally pinned. A connected beam behaves like a simple supported beam. For cases in between, the stiffness of a joint is neither zero nor in¿nite. The moment that the joint transfers will have as a consequence the difference of ij between absolute rotations of two connected structural elements, Figure 2.c). In this case this concerns a semi-rigid joint. The simplest way of introducing these notions within global analysis is by means of a rotational spring placed between the ends of the structural elements that are being connected. Rotational stiffness of spring S is the parameter which links the carried-over moment Mj with the appropriate rotation ij, which in turn is the difference between absolute rotations of two connected structural elements. Rotational stiffness is de¿ned as the slope of curve Mj -ij which depends on the properties of the joint. When this rotational stiffness equals zero, or when it is relatively low, the joint is in the class nominally pinned. On the other hand, when rotational stiffness S is in¿nite, or relatively large, the joint is in the class of a rigid joint. In all other in-between cases, the joint belongs to the class of semi-rigid joints.

Fig. 3. Modelling of joints for the case of an elastic global analysis

For semi-rigid joints, loads will cause the bending moment Mj as well as the appropriate rotation ij between the connected structural elements. The moment and the appropriate rotation are linked by a relationship whic depends on the structural properties of the joint. This relationship is illustrated in Figure 3, where, for the purpose of simpli¿cation, it is presumed that the global analysis was performed with linear elastic assumptions. Within the global analysis, the effect that semi-rigid joints have in relation to the effect of rigid joints or nominally pinned ones, is not only the modi¿cation of displacement but also the distribution and size of internal forces and bending moments in the structure. The concept of semi-rigid joints has been introduced into Eurocode 3 for the design of steel structures regarding static loads. 3 JOINT REPRESENTATION 3.1 Introduction

Research regarding joints was focused mostly on the following aspects:

112

Annual 2010/2011 of the Croatian Academy of Engineering

• evaluation of mechanical properties of joints by means of rotational stiffness, moment resistance and rotational capacity, • procedures of analysis and dimensioning of frames including the behaviour of joints. From this approach it had been understood that there is a step in between these ones that is to be taken into consideration, with the goal of consistent integration of the actual response of the joint within the frame analysis, and it is called the joint representation. The joint representation includes four steps as follows: • joint characterisation: determination of stiffness, resistance and ductility of the joint with the help of the Mj -ij curve that describes its behaviour, • classi¿cation of joints: provides boundary conditions for the application of traditional joint modelling types, i.e. pinned or rigid, • joint idealisation: derivation of a simpli¿ed Mj -ij curve so that it is adjusted to the speci¿c procedures of frame global analysis, e.g. the linear idealisation for the elastic analysis, • joint modelling: the manner in which the joint is physically presented in lieu of the frame analysis. 3.2 Joint characterisation

A more exact yet also more expensive way to characterisation of deformability and resistance of joints is laboratory research. This way is basically limited to research activities and is not recommended for every day practice. On the other hand, numerous mathematical models have been developed that can be placed into four major sets as follows: • curve ¿tting, • simpli¿ed analytical models, • mechanical models, • analysis by means of ¿nite elements. The procedure for the characterisation of mechanical properties of joints adopted by Eurocode 3, Part 1.8, is based on the component method. The identi¿cation of various components that constitute the joint (bolts, welds, stiffeners) provides a good picture of the complexity of the joint analysis. The analysis of a joint requires exact consideration of many occurrences: non-linearity of materials (plasticity, strain-hardening), non-linear contact and slip, geometrical non-linearity (local instability), residual stresses and complex geometrical con¿gurations. Although numerical procedures that apply non-linear ¿nite elements can take into consideration all these complexities, they require long-lasting procedures and are very sensitive to the possibility of modelling and analysis. Therefore, for practical reasons, the planned approach must be based on simple models that leave out many variables. The component method precisely ¿ts the simpli¿ed mechanical model consisting of springs and rigid links. In this the joint is simulated by means of the appropriate selection of rigid and Àexible components. These components represent speci¿c parts of the joint that, depending on the type of loading, make an identi¿ed contribution to one or more of its structural properties. Basic components that are included in Eurocode 3 are as follows: 1. Column web panel in shear 2. Column web in transverse compression 3. Column web in transverse tension 4. Column Àange in bending 5. End-plate in bending 6. Flange cleat in bending 7. Beam or column Àange and web in compression

Annual 2010/2011 of the Croatian Academy of Engineering

113

8. Beam web in tension 9. Plate in tension or compression 10. Bolts in tension 11. Bolts in shear 12. Bolts in bearing (on beam Àange, column Àange, end-plate or cleat) 13. Concrete in compression including grout 14. Base plate in bending under compression 15. Base plate in bending under tension 16. Anchor bolts in tension 17. Anchor bolts in shear 18. Anchor bolts in bearing 19. Welds 20. Haunched beam. When using the component method, the characteristics of resistance, stiffness and ductility of basic components are evaluated ¿rst. These characteristics are then combined with the goal of achieving the mechanical characteristics of the entire joint. As a result, the application of the component method is according to that fairly wide, since it can be applied to all steel joints for which the characteristics of constituent components can be identi¿ed. In general, each of these components is characterized by a non-linear relationship between the force and deformation (F-ǻ), although more simple idealisations are also possible. Several models made of spring and rigid links are proposed, which consist of the same basic components. The application of the component method on steel joints requires the following steps: 1. Selection of the relevant (active) components from the general list of components (20 different components have thus been further de¿ned in Part 1.8, EC 3). 2. Evaluation of the force-deformation response of each component. 3. “Assembly” of the active components for the evaluation of the moment-rotation (Mj -ij) joint response, by means of a representative mechanical model. Its application can suit various levels of element “processing”, e.g. a simpli¿ed characterisation of components is possible whenever only the resistance or the initial stiffness of a joint is sought. Familiarization with the mechanical behaviour of various joint components enables the analysis of a large number of various joint con¿gurations with a relatively low number of components. The key element in the component method therefore relates to the characterisation of the F-ǻ relationship for each particular spring (component). For the evaluation of the initial stiffness of a joint only the linear stiffness of each component is necessary, while for the evaluation of ductility it is necessary to know the non-linear F-ǻ relationship of each component. Classi¿cation of joints 3.2.1 Classification according to stiffness

Classi¿cation according to stiffness for rigid, semi-rigid and pinned joints is performed by simply comparing the design stiffness of the joint with two stiffness boundaries, Figure 4. With the goal of simpli¿cation, the stiffness boundaries are set in such a way that they allow a direct comparison with the calculated initial stiffness Sj,ini of the joint, for any type of joint idealisation that is being used in the analysis and is presumed beforehand. The boundaries of classi¿cation according to stiffness are as follows:

114

Annual 2010/2011 of the Croatian Academy of Engineering

• rigid joint (unbraced frames), (braced frames), • semi-rigid joint (unbraced frames), (braced frames), • pinned joint

EI/L denotes the Àexural stiffness of the connected beam.

Fig. 4. Boundaries of classi¿cation according to joint stiffness 3.2.2 Classification according to resistance

Classi¿cation according to resistance consists of comparing the design joint resistance moment, Mj,Rd, with the boundaries of “full-strength” and “pinned”, Figure 5. Boundaries of classi¿cation according to resistance are as follows: • full strength joint • partial-strength joint • pinned joint

Mfull strength denotes the design resistance of the weaker structural element in the connection. It shall be noted that the classi¿cation based on experimental Mj-ij joint characteristics was not taken into consideration because only the design properties were considered.

Annual 2010/2011 of the Croatian Academy of Engineering

115

Fig. 5. Boundaries of classi¿cation of joint stiffness 3.2.3 Classification according to ductility

Adequate experience and appropriate detailing result in the so-called pinned joints that exhibit a suf¿cient rotation capacity. That means they can sustain the rotations imposed on them. In the case of moment resistant joints a notion of ductility class that includes rotation capacity is introduced. This problem has not been dealt with in more detail within Eurocode 3, although extensive research is being done regarding this problem. In order to evaluate the ductility of a joint one can apply an analogy with the classi¿cation of cross-sections according to rotation capacity. Thus, joints can be classi¿ed into three classes concerning ductility, see Figure 6.

Fig. 6. Ductility classes of joints – classi¿cation according to rotation capacity

116

Annual 2010/2011 of the Croatian Academy of Engineering

Class 1 - Able to reach MRd and with a suf¿ciently good rotation capacity to allow plastic design of the frame. Class 2 - Able to reach MRd but with a reduced plastic rotation capacity. A plastic veri¿cation of the section is anyway allowed. Class 3 - Where brittle failure (or instability) limits the moment resistance and does not allow full redistribution of the internal forces in the joint. 3.3 Joint modelling

The following types of joint modelling have traditionally been considered: • for rotational stiffness: rigid or pinned; • for resistance: full-strength or pinned. However, as far as joint rotational stiffness and economical joint dimensioning are concerned, they can be semi-rigid. This provides a new possibility of joint modelling: semirigid/full-strength and semi-rigid/partial-strength. Eurocode 3 takes into consideration these possibilities with three joint modelling types, Table 1: • continuous: covers only the rigid/full-strength case, • semi-continuous: covers the rigid/partial-strength, semi-rigid/full strength and semi-rigid/partial-strength cases. • simple: covers only the pinned case. Table 1. Joint modelling types Stiffness Rigid Semi-rigid Pinned

Full-strength Continuous Semi-continuous -

Resistance Partial-strength Semi-continuous Semi-continuous -

Pinned Simple

The notions continuous, semi-continuous and simple have the following meaning: • continuous: the joint ensures a full rotational continuity between the ele ments it connects; • semi-continuous: the joint ensures only a partial rotational continuity between the elements it connects (connecting elements); • simple: the joint prevents any rotational continuity between the elements it connects. These meanings are linked to the type of the global analysis of the frame. In the case of an elastic global analysis of the frame, only the stiffness properties of the joint are important for joint modelling In the case of an ideal plastic analysis, the main characteristic of the joint is resistance. In all other cases, the properties of stiffness as well as resistance are a guide for the manner in which joints are to be modelled. The above listed possibilities as well as their physical interpretation for analysis are provided in Table 2 and in Figure 7, respectively

Annual 2010/2011 of the Croatian Academy of Engineering

117

Table 2: Joint modelling and frame analysis Type of frame analysis Joint modelling

Elastic

Rigid plastic

Continuous

Rigid

Full-strength

Semi-continuous

Semi-rigid

Partial-strength

Semi-rigid/fullstrength Semi-rigid/partial strength Simple

Pinned

Pinned

Elastic-perfectly plastic and elastoplastic Rigid/full-strength Rigid/partialstrength

Pinned

Fig. 7. Simpli¿ed joint modelling for frame analysis

3.4 Joint idealisation

The Mj-ij curve that describes non-linear behaviour of joints is not convenient for everyday practice. However, it can be idealised without a more signi¿cant loss of exactness. One of the simplest idealisations is the one done by means of the elastic-perfectly plastic relationship, Figure 8. The advantage of such modelling is in the fact that it is similar to the one that is applied for modelling of member cross-sections that are subjected to bending. The Mj,Rd moment that corresponds to the yield plateau is within EC 3 called design moment resistance. It can be considered as a pseudo-plastic moment resistance of the joint. The effects of strain-hardening and the possible membrane effects are neglected, which explains the difference in Figure 8 between the actual M-ij curve and the yield plateau in the idealisation. The value of constant stiffness of joint Sj.ini/Ș has been explained in a number of papers, and practical values are provided in EC 3, Part 1.8. The coef¿cient Ș results from a highly

118

Annual 2010/2011 of the Croatian Academy of Engineering

expressed non-linearity of the Mj-ij joint curves in comparison to the M-ij curves for structural elements.

Fig. 8. Bi-linearisation of the M-ij curves

Actually there are different possibilities of Mj-ij joint curve idealisations. The choice of one of them depends on the type of global analysis of the frame used, and EC 3 lists them according to the following: • elastic idealisation for the elastic analysis, • perfectly plastic idealisation for the perfectly plastic analysis, • non-linear idealisation for the elastic-plastic analysis. 4 CHOICE OF A JOINT MODEL Joints presented as rigid or pinned in an analysis should be dimensioned in such a way that they satisfy the conditions for a classi¿cation of a rigid or pinned joint. The model for a semi-rigid joint can be more or less complex. It can be modelled as a spring for which M-ij relationship is in the range between linear elastic and non-linear, see Figure 9. The elastic global analysis requires that the behaviour of a joint be modelled as linear elastic. The elastic – perfectly plastic analysis requires a bi-linear joint model. So, the type of global analysis directly inÀuences the complexity of the adopted joint model. It should be mentioned that formation of plastic hinges in the joints is allowed in the application of plastic analysis.

Fig. 9. Presentation of different Mj-ij relationships

5 ECONOMIC JUSTIFICATION OF THE NEW APPROACH A modern approach to the behaviour of joints is based on the realistic behaviour of joints as semi-rigid. It is founded on economic justi¿cation of a structure in which the joints have been considered as semi-rigid. To obtain the minimum price of a steel structure two design

Annual 2010/2011 of the Croatian Academy of Engineering

119

strategies are possible regarding the execution of the joint: • Simpli¿cation of joint details which results in reducing the costs of workshop execution. It is important for unbraced frames where joints transfer signi¿cant moments and those are rigid joints. • By reducing cross-sections of elements which results in decreasing the expenses for material. It is signi¿cant for braced frames with nominally pinned joints. Generally speaking both strategies will lead to the application of semi-rigid joints. In the case of rigid joints an economic solution can be achieved if the stiffness of joint is close to the classi¿cation limit of semi-rigid joint. Different economic studies show that there is a possibility of saving with the application of semi-rigid joints and the adequate design of 20-25% in the case of unbraced frames and 5-9% in the case of braced frames. With the assumption that the price of steel frames amounts to about 10% of the total price of a commercial building and about 20% for an industrial building a decrease in total costs can be estimated to 4-5% for unbraced frames. For braced frame systems saving of 1-2% is possible. The costs of material and workshop execution depend on the relative stiffness of the joint, Figure 10.

Fig. 10. Qualitative relation of a steel structure depending on the relative stiffness of the joint

With an increase of the relative stiffness of joints the costs of material is decreased, curve A, whereas the costs of labour grow, curve B. For the total costs that are a sum of the costs of curves A and B, the minimum can be found and therefore the optimum relative stiffness as well. In many cases the value that leads to the optimum design of a structural system by taking into consideration the minimum of the total costs does not lie in the area of nominally pinned nor in the area of semi-rigid joints. Looking at the trends in the past few decades it is obvious that the costs of labour grow compared to the costs of the used material (see curve B*). As a result, see Figure 10, it is clear that a fast development of the optimum stiffness of joints shifts more towards semirigid joints. Therefore, in an attempt to ¿nd an economic solution for steel structures the use of semi-rigid joints will become increasingly more interesting. 6 CONCLUSION The design of structural joints that is proposed in Eurocode 3, is based on the research conducted over the past few decades. Eurocode 3 gives practical guidelines for the characterisation of most of steel joints (beam-to-column joints, beam and column splices, column base plates etc.) of frames made of I or H cross sections, loaded by static load. Further research is focused on extending the guidelines for characterisation of joints as composite

120

Annual 2010/2011 of the Croatian Academy of Engineering

joints (concrete and steel) and joints made of other types of cross sections and other types of load. The new approach, semi-rigid joints – semi-continuous structure, gives more possibilities than the traditional approach in type of framing. This is based on the fact that the joint characteristics in the design have been considered as variables that are chosen to comply with the particular requirements of a speci¿c project. For braced frames, the new approach allows the use of smaller cross sections of beams and a reduction of the costs of the structural frame. For unbraced frames the new approach reduces the complexity of the joint geometry and thus results in major saving in the execution costs. REFERENCES 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15.

Eurocode 3, Design of Steel Structures, Part 1.1: General Rules and Rules for Buildings, European Prestandard ENV 1993-1-1, Brussels: CEN, 1992. Revised Annex J of Eurocode 3. Joints in Building Frames, European Prestandard ENV 1993-1-1: 1992/A2, Brussels: CEN, 1998. Eurocode 3, Design of Steel Structures, Part 1.8: Design of Joints, European Prestandard prEN 1993-1-8, Final Draft, Brussels: CEN, 2001. Maquoi, R.; Chabrolin, B.: Frame Design Including Joint Behaviour, ECSC Report 18563, Luxembourg, Of¿ce for Of¿cial Publications of the European Communities, 1998. Nethercot, D.; Zandonini, R.: Methods of prediction of joint behaviour – Beam to column connections, In: Naranayan, R. (ed.) Structural Connections – Stability and Strength, Elsevier Applied Science, 1989., 23-62 European Convention for Constructional Steelwork, Analysis Design of Steel Frames with Semi-rigid Joints, Publication 67, Brussels: ECCS, 1992. Zoetemeijer, P.: A design method for the tension side of statically-loaded bolted beamto-column joints, Heron 1974; 1-59 Weynand, K.; Jaspart, J.-P.; Steenhuis, M.: The stiffnes model of revised Annex J of Eurocode 3, In: Bjorhovde, R.; Colson, A.; Zandonini, R., (editors): Connections in steel structures, Trento, 1995., 441-452 Huber, G.; Tschemmernegg, F.: Modelling of steel connections, Journal of Constructional Steel Research 45 (1998) 2, 199-216 Weynand, K.; Jaspart, J.-P.; Steenhuis, M.: Economy Studies of Steel Building Frames with Semi-Rigid Joints, Journal of Constructional Steel Research 46:1-3 (1998) Girăo Coelho, A. M.; Bijlaard, F. S. K.; da Silva, L. S.:Experimental assessment of the ductility of extended end plateconnections. Engineering Structures 26 (2004), 11851206. da Silva, L. S.; Lima, L.; Vellasco, S.; Andrade, S.: Experimental behaviour of endplate beam-to-column joints under bending and axial force, Database reporting and discussion of results. Report on ECCS-TC10, Meeting in Ljubljana, April 2002. Faella, C.; Piluso, V.; Rizzano, G.: Structural Steel Semirigid Connections, Theory, Design and Software, CRC Press LLC, 2000 N.W. Corporate Blvd., Boca Raton, Florida 33431, 2000. Cruz, P. J.S; da Silva, L. S.; Rodrigues, D. S; SimĘes, R. A. D.: Database for the semi-rigid behaviour of beam-to-column connections in seismic regions, Journal of Constructional Steel Research 46 (1998) 1-3, 233-234 Robot Millennium Program - Version 17.5: User's Manual, RoboBAT, June 2004.

Annual 2010/2011 of the Croatian Academy of Engineering

121

Development strategy of ECOINDUSTRAL PARK RAŠA – EIPR Prof.dr.sc. Juraj Božičević Elitech Ltd. Trg kralja Tomislava 18 Zagreb 10000, Croatia [email protected], Tel. 00385 1 4922264

Dr.sc. Marijan Andrašec Ecooleum Ltd. Trg kralja Tomislava 18, Zagreb 10000, Croatia [email protected], Tel. 00385 98 404755

Abstract The pith of the strategy of development of the Ecoindustrial Park Raša is presented. The Park is situated in the bay of Raša, unique but improperly used and maintained Croatian fjord. The starting point in conceiving of the Ecoindustrial Park Raša is synergy of service and production activities, successful cooperation of companies aiming to achieve results on the following principle: unity is more successful than the amount of successfully combined activity of singular companies, units of the system. The contribution will be the care about industrial ecology, strengthening of the integration of economic raise and protection of environment as well as industrial symbiosis that will improve the economy and the quality of life in the Community of Raša. Key words: ecoindustrial park, sustainable development, Port Raša, Community of Raša Contents 1.0 2.0 2.1 2.2 3.0 3.1 3.2 3.3 3.4 3.5 3.6 4.0 4.1 4.2 4.3 4.4 4.5

Introduction The Community of Raša Geography Present situation Ecoindustrial Park Raša – EIPR - conceptualization What is an Ecoindustrial Park – EIP? Basics for conceptualization of EIPR Projects in EIPR proposed for realization Ecoindustrial Park and the Community of Raša Stakeholders Knowledge and experience on which venture relies Ecoindustrial Park Raša – EIPR - realization The way of achieving realization of EIPR Economy of the project EIPR Risks Development effects achieved with EIPR Priority tasks

122

5.0 6.0

Annual 2010/2011 of the Croatian Academy of Engineering

Conclusion References

1 Introduction When, during the early 1990s, more precisely, in 1993, Croatian System Society has organized the ¿rst conference called The Port as a Complex System, our guiding idea was the promotion of knowledge about the importance of the modernization of the Croatian ports and the strategic consideration about the role of the ports in the development of the economy. At that time the ¿rst author was, as the Associate Minister for higher education in the Ministry of Education and Sports, participated at the ¿rst Croatian – Netherlands negotiations about the cooperation in education area... so he took the opportunity to visit the Rotterdam port and learn on higher maritime education, as well as on the training equipment at the High Maritime College, with the intention to transfer gained experiences to the Maritime Faculty in Rijeka. He was accompanied by a host, a Dutchman, who was the member of the management board of the Rotterdam port and who presented him numerous data which illustrated the importance of the port and its industrial area for the economy of the Netherlands. The host was especially interested in Croatia’s plans – in which way will Croatia, after has gained independence, use its maritime position, and how Croatian strategic plan of development of economy and the ports will be conceived. We shall quote the private communication of the ¿rst author (3): The Dutchman emphasized the extraordinary geographical position and maritime features of the Port of Rijeka and neighboring ports within its jurisdiction. He had excellent knowledge of Croatian possibilities, and, since I was preparing for the conference „Port as a complex system“ in Rijeka, I presented him ambitiously the vision of the Conference Program board and our idea about ports development in Croatia... He asked me if we really have full understanding of the geo-strategic importance of the Port of Rijeka and its possible competitiveness in relation to the Rotterdam port. When we shaked hands during the parting, he said to me, half-jokingly, something which stunned me: «I will kill myself if the Port of Rijeka takes even 1% of the middleEuropean traf¿c from the Port of Rotterdam.» That was the war time in Croatia. The Port of Rijeka was destination of just a few ships. When I talked about this event to my closest cooperate in the project „Port as the Complex System“, dr. Juraj Maÿariü, he suggested me not to talk about the Dutchman’s statement, because no one will understand... However, he proposed that we talk about this in inner circle as a very serious topic, and that the program of our future conferences should be gradually developed in such way that, in the sphere of the economic decision making, we could develop the awareness about the need for the strategy of the economic development which is leaning upon the maritime orientation. The fourth conference ”Port as a Complex System”“ held in 1999, was dedicated to strategic topics: „Rijeka – the Main Croatian Port and joined Rijeka and Zagreb – Croatian and European economic Focus“. (4, 5) However, since 2000, there was no more interest for the Conferences.... I was introduced to the operations of the Port of Rijeka ¿ve years later, when I started to cooperate with the Community of Raša. The talks with the mayor of the Community encouraged me to think about the sustainable development of the port area of Raša, which is in concession of the Port of Rijeka. Really, about very strategic Task. The mayor of Raša accepted the idea of consulting company Elitech Ltd. to make the development concept of the Community of Raša as socially responsible community, the basis of the wise exploration of its natural resources, revitalisation and further development of

Annual 2010/2011 of the Croatian Academy of Engineering

123

its economy, based upon the idea of eco-industrial society. The following agreement was accepted as the frame work: The target of sustainable development is to facilitate and ful¿ll of basic necessities and quality lives to all the citizens of the Community, having in mind not to risk the quality of life of future generations. If we don’t take care about sustainable development we could face an insecure future. Therefore the sustainable development is our best long-term interest. We started with work in autumn 2005. We had number of meetings with the mayor and the head of the County of Istria, who suggested that the Community of Raša, in cooperation with consulting company Elitech Ltd., should establish the Working group for strategic development of port area and sign an agreement upon it. With the acceptance of Letter of Intent, signed by Community of Raša, County of Istria, Port of Rijeka, Port Authority of Rijeka and Elitech Ltd. on 15th December 2006, Working group started with meetings and discussions in early 2007. As a basis for discussions has been the study „Development of Eco-energy Valley of Raša and Port area Raša based upon the production of biofuel”, made by Naval Research Institute in Zagreb in cooperation with the company Elitech Ltd. and the company Ecooleum Ltd., which joined the Working group. Ecooleum Ltd was established with the purpose to realize conceived projects. The idea of technological platform Ecoenergy Valley was developed and presented to the Croatian Institute for Technology as a possible EU-project in order to ¿nance the project. Aside to eco-energy production, the particular attention was focused upon traf¿c infrastructure, especially in revitalizing of the railway as important support to port activity. In the period from 2008 to 2010 the idea of Eco-industrial society has been developed in ¿nal form, and the basic conception is presented in this paper. In 2008. the Forum for Social Responsible Activity was founded to facilitate communication with local community. The companies Holcim and Murexin - Istarska tvornica vapna joined the Forum, and the consultations were held upon the Sustainable Development of Community of Raša. Initially, the important contribute was made by Croatian Systems Society and their study Ecoproduct areas of Croatia and then with their project Systemic considerations of social responsibility in the living area (2008-9), ¿nanced by Ministry of Environment Protection, Space Planning and Construction. In the case of Ecoindustrial park of Raša it has been conceived the interaction of portrelated, industrial, energy, agricultural, maricultural and touristic activities that rely on restructured and highly functional communal and traf¿c infrastructure – port, road and railway. We would like to emphasize the part of activities in management of eco-industrial park that will be very important for successful achievement of sustainable development and industrial symbiosis, for the reduction of costs, common use of information, services, energy and byproducts as raw material for number of users. From the point of view of the Community of Raša, it was particularly important to motivate enterprising spirit in the development of port activity. It is presumable that the building of the Park will be especially bene¿cial and a big impulse on port activity, and the Port of Rijeka, which is the concessionaire, should be encouraged to show more responsibility on that valuable national resource.

124

Annual 2010/2011 of the Croatian Academy of Engineering

2 The Community of Raša 2.1 Geography

The Community of Raša is situated on south-eastern side of Istrian peninsula. Its surface is 7.954 ha, it contains 23 settlements with 3535 inhabitants and the shore length of 48 km. It is part of Region of Istria. The Community of Raša have got high quality land for agricultural production that is partly owned by private persons, and partly is property of Republic of Croatia and of the Community. It is planted with fodder, vegetable in lowlands, and with grapes, fruits and medicinal cultures and olives on the sunny positions. Forests are mostly of private property, and the rest of it is property of Republic of Croatia and are managed by Croatian Forests. The forests are not important for exploitation but are very important for ecological balance, prevention of erosion, improvement of microclimatic conditions and development of agrotourismus and recreation. Climatic conditions in Community of Raša are characterized with bigger quantity of rains than the west side of Istria. Rainy period is at its peak in October and minimum in July. The most signi¿cant winds are northeastern, especially bura in winter period and summer’s night wind in most part of shore territory, southeastern – especially jugo (sirocco) in spring and autumn, northwestern – especially during summer period. Soils are shallow and stony, and smaller areas of deep soil we can ¿nd in typical karstic forms – short lanes and karstic heights, predominantly red and brown soil. In the area of the river Raša we ¿nd hidromor¿c (gley) soil, and on the edges aluvial-coluvial soils that are salinized on the estuary of river. The river of Raša is the main and the most important water-current in water-supply-system “Raša-Boljunþica“. Its characteristic is the frequent torrents, so at times extremely high water is to be expected. The main water sources are Fonte Gaj and Kokoti. The town of Raša with 1800 inhabitants does not have waste water puri¿cation facility. Across the Community of Raša there is state road D21 Pula-Labin, number of regional and local roads. Numbers of unclassi¿ed roads are under community’s jurisdiction. Railroad Lupoglav-Štalije is used exclusively for transport; it’s not electri¿ed, it’s in group of mountain railways and is quali¿ed for average speed of 60 km/h. Railway station is in Most Raša. There are several ports on the territory of the Community of Raša. The port Bršica is the only one classi¿ed for international sea shipping (state-level port). The port of Koromaþno is special purpose port – industrial port. The ports Sveta Marina, Trget and Tunarica are of local importance (¿shery). The main direction of magistral gas pipeline (for international transport) of national importance Casal Borsetti – Karlovac DN700, transits along eastern shore of Istria and is passing through a part of Community’s territory. 2.2 Present situation

Once industrially developed and supported by mining (coal mines, coal power plant) and by production of machine-tools (Prvomajska factory), the Community of Raša became poor with the closure of coal mine and failure of Prvomajska, leaning nowadays on companies Holcim, ITV Murexin and hardly active Port. The territory of the Port of Raša, the most valuable industrial zone in Istria, has long tradition in energetic sector (coal), and the industrial culture of inhabitants was developed. In

Annual 2010/2011 of the Croatian Academy of Engineering

125

the past it was mostly used to support local industry, and therefore never had signi¿cant value as transit port for Middle-European area, although it was connected by railroad to Trieste and further to North Italy and Middle Europe. The port is used for stone and timber transport. Port installations and operative quay are in dilapidated conditions so the new one should be made, other traf¿c, energy and water infrastructure should be made too and operative and warehousing surface should be widened. The bottom of the bay should be cleaned from the mud. On the location of Bršica a stone pit existed, that was closed in 1978, so this area should be revitalized by removing the part of the hill between microlocations Bršica and Štalije. Agricultural land in river Raša valley and along the channels are poorly used, because of the irresponsible maintenance of the channels, non resolved water overÀow protection and salinization. The territory of Raša is rich with high quality lime stone, so the stone exploitation, as well as in other parts of Istria, is a part of traditional exploitation. On that basis the production of lime (ITV – Istrian Lime Factory) and cement (Holcim) were established, so the further establishment of complementary production with higher grade of material processing would be natural development of this industry. The project of waste water drainage should be replaced with improved project with which should be projected high quality equipment for depuration and obtained technical water for agricultural production on uncultivated agricultural land. The urban plan of Community of Raša in the past did not provide for appropriate development of the Community of Raša, as well as of the port-area. It should be upgraded in the new urban plan. 3 Ecoindustrial Park Raša – EIPR - conceptualization 3.1 What is an Ecoindustrial Park – EIP?

An Ecoindustrial Park is unity of production and service activities that cultivate cooperation and with emphasized care about environment and materials, harmonizing energy, water and materials Àows. Acting together, the members of business units take care of the common pro¿t, which is bigger than amount of single pro¿t that the single company would achieve optimizing proper activity independently. In that way economic success of the companies increases and minimizes its inÀuence on environment. That is a new concept of sustainable development of the local community which protects the economy of local society from undesired inÀuences of globalization. EIP has rich and different possibilities of development including diversity of environment cultivation, infrastructure, single out¿t and resources, common services etc. Since the beginning of 90s, a number of ecoindustrial parks were established all over the world, and their activities proved that public and private sector is interested in such concept of sustainable development. (8, 9, 10, 11, 12) Ecoindustrial parks are developed in three basic forms: Transformation of an existing industrial park, Revitalization of previously used sites (brown¿elds development). New industrial site development.

126

Annual 2010/2011 of the Croatian Academy of Engineering

Often these appear in combination. When existing industrial parks undergo expansion, this can be developed as an eco-industrial park. The form of ElP development brings certain preconditions that affect the whole planning process: Existing industrial parks have the companies for potential synergies, but often the interaction between the companies is not developed, making it dif¿cult to get to know companies and their potential in detail. Park management and local authorities have to invest time in community interaction among the companies. As in developing traditional industrial parks, similar strategies can be used to establish ecoindustrial parks. The most commonly practiced or researched strategies are: Anchor tenant; This involves having a large industrial user, often a power plant, sugar re¿nery or other types of operation with large-scale material Àows, which the industrial park will be developed around. Materials or By-product exchange; This can occur either in an eco-industrial park or in a regional network of businesses, including industrial parks. This involves using one industry’s waste or “byproducts” as another industry’s raw materials. Resource Recovery System; is an expanded concept of the byproduct exchange and includes all forms of material and product recovery in an industrial complex (waste management, by-product exchange, recycling and remanufacturing). Energy connections; involves maximizing energy ef¿ciency through design or rehabilitation, co-generation, and energy cascading. It can also include the utilization of renewable energy. Thematic Park; a number of ElPs focus on co-locating a speci¿c type of industry. Examples of such thematic parks are: Agriculture-based EIP or Agro-ElP; provides support for sustainable farming and food processing and includes several basic types of ¿rms and agencies which may be recruited as tenants. Farms and food processing industries have a high Àow of organic material that can easily be further processed and maximized in its resource value. Chemical EIP; the chemical industry uses co-location of production units and upstream-downstream synergies as their core business. Sector-speci¿c EIP (Textile, Electronics etc.); a number of regions have historical industrial areas with an agglomeration of sector industries, often as industrial clusters, but also as industrial parks. Companies located there usually have similar problems and less potential for synergies, but strong links between companies (associations) can stimulate joint problem solving and innovation concepts. Resource Recovery EIP; companies involved in recycling, reuse, remanufacturing, niche collection of materials, and manufacturing from recovered materials form a cluster with good synergies among them. Environmental Technology Parks/EIP; the concept focuses on the promotion of environmental technology industries. Aside from possible innovation synergies, their colocation does not necessarily provide any bene¿t in industrial symbiosis. Many environmental technology companies produce like ordinary manufacturing businesses. 3.2 Basics for conceptualization of EIPR

Considering the past experiences and the present situation it is natural wish of Community’s management to initiate development on new basis and use extraordinary natural resources, especially port-area, with the special attention on sustainable development approach. The proposed study should bring the strategy of development of Ecoindustrial Park of

Annual 2010/2011 of the Croatian Academy of Engineering

127

Raša as the basis of sustainable development of the Community of Raša. The conditions relevant for conceptualization of EIPR are: there are existing old industrial zone and economically weakly developed port, not connected all subjects in space and do not cooperate, so they should be brought to create new relations through new ideas, local community is interested in forming EIPR, because it should solve its infrastructural and ecological dif¿culties and economical development, present entities could be netted with new projects, location is important also on Croatian level (state port!), and at North Adriatic area is the only free and high quality industrial space. Extraordinary convenient traf¿c settlement: port, road and railway should be natural encouragement for the logistic support of activities in EIPR. With planning and restoring of the railway and in the future with the tunneling of the mountain Uþka, EIPR will have additional value. The activities on settling of the port area, removal of ruins of the abandoned separation of the coal and the formation of initial industrial plateaus will be important initial activities, followed by necessary reconstruction of quay and removal of mud from the bay. One of the ¿rst and important steps in realization of this unique Croatian project is settlement of the neglected port area and forming of landscape and horticulturally shaped industrial area and reconstruction of quay. In order to manage the activities in Park, a social, non-pro¿t Company for environment economy should be established, against the model of public-private partnership, as a common model of the management in eco-industrial parks. Besides institutional co-property of local community, the idea is that local inhabitants should become shareholders, and this should favor acceptance of the project and increase the possibilities for long-term successful and sustainable development of port, transport and industrial activities, as well as provide attractive bene¿ts for internal and foreign investors. The company for management of the Park will take care about the development and building of EIPR, so acting in concordance with the targets de¿ned with this strategy will implement the model of the Park and will select the participating companies, investors and act as coordinator, manage and coordinate closely with the Community of Raša. For the success of the project it is important to clearly de¿ne WHAT and HOW a support of the citizens should give the impulse for general development. We have to establish multidisciplinary team in the company for the management of EIPR, in which there would participate, together with the experts in branches of business, people from the executive authorities. Such expert team should, in accordance with the rules of business, start working on the plan of development beginning with the de¿nition of development targets, available human resources through the ways of its executing, from planning the future development till the preparation of the conditions for its achievement. Large amount of funds should be invested in port space design, but it should be emphasized that this project should, with wisely designed project of environment design and use of obtained material, bring pro¿t to use in further investments in space and infrastructure development of the Park. So the conditions to attract investors should be created gradually, and the interest for the project has been already shown from the investors from Austria, Italy, Sweden, Denmark and Malaysia. It should be taken into account that the Croatian economy is in big dif¿culties: huge public debt, unemployment, the deprivation of quality of life, but also the leak of the current knowledge and skills. In such circumstances the development should be conceived considering existence and fast initialization with new economic initiatives.

128

Annual 2010/2011 of the Croatian Academy of Engineering

Also, it should be kept in mind that the business conditions have changed in last ten years both in Croatia and in the world. There is number of barriers in business development and creative and highly innovative business penetrations will be needed as well as wellconceived marketing, and organized care for maintaining the human resources above all. Elaboration of the program of development should be the beginning. Well-conceived projects should attract investors. It is important to emphasize that projects cannot be accomplished just through the preparation of the area of industrial zone and calling the investors to come and ful¿ll their interest, which is the present Croatian practice. We should act completely different from the way it’s done actually. 3.3 Projects in EIPR proposed for realization Territory of EIPR covers the port of Bršica with port area with particular accentuation on industrial microlocations: Bršica, Štalije and Vlaška, and areas surrounding port area along the river Raša and lime factory, areas along the channels to the industrial location Vlaška and the city of Raša (see picture). In the period from the establishment of Working group and presentation of the basic ideas about settlement of the eco production in port territory, the concept of the Park was gradually built. During the study based on industrial ecology as the basis of the project, different industrial, agricultural and service activities came to the surface, which could be connected in the community on the principles of modern eco industrial parks. Special attention was paid on biofuel production based on imported raw materials, and on the other hand the development of agricultural production on poorly tended land and processing agricultural products, as well as on waste water puri¿cation and reuse.

There were various ideas about energy, production of biogas, wind power station Goli project and other useful and complementary ventures, in order to settle operative port and industrial synergic system. We will present the most important tasks: The building-up of the operative port and industrial surface, which will be cleaned of the ruins from previous industrial period, and then the hill between ex quarry Bršica and Štalije will be removed. For that task the Community of Raša and Port authority of Rijeka should jointly prepare the tende. Stone exploitation as building material with further processing should provide important initial income for development of EIPR. Production of biofuel on the bio-material basis Biodiesel as a modern ecological fuel guarantees good positioning of the Park and connection with other economies in Croatia and Middle Europe. The idea of this clue project is supported by knowledge and experience in Biodiesel production in Croatia. Croatian and foreign experts and business partners were consulted and the whole width was taken in consideration: technical, technological, microeconomics, energetic, political, organisational and markets knowledge relevant for the project realisation, Croatian, European and global situation and trends have been considered. With the increase of bioethanol use, which is mostly imported in Europe from overseas countries, it opens the possibility of logistic activity in that direction. Bioenergy business means creating ¿nal products or semi-products, and could be realized in Štalije location, and means higher (and more pro¿table) phase of technological development in that ¿eld.

Annual 2010/2011 of the Croatian Academy of Engineering

129

Industrial power plant on biofuels Industrial location with big production plants should have proper power plant that brings supplying security, but uses by-products from energy production plant (hot water, steam, energy power..). Considering great delicacy of the relation between electro-energy projects and environment, the particular attention was paid to the power station on biofuel. The basic idea is that it should be exclusively in function for local society: use of biodiesel and bio-oils; „green energy“ use of biogas as a product in cleaning of waste water from Raša and Labin high level of work Àexibility: industrial power plant netting with local heat and electric energy consumers; for example food production (greenhouses) and food processing (dryers, refrigeration plants…) Water management Raša and all Labin territory have a large quantity of water from the coal-mine galleries that could be used as technical water for irrigation in agriculture. In case it will be shown that this water could be supplied continuously in necessary clean level, it could be used as an export product for Adriatic islands and South Italy. Labin and Raša do have a problem of waste water processing, but with the use of modern technology where waste is used to produce biogas, and puri¿ed water for agricultural irrigation, we would create a system that should economically and ecologically friendly bring several bene¿ts: puri¿cation of waste water, produced energy could be used for pumping water for irrigation in agriculture, puri¿ed water could be used for irrigation, ecologically contaminated channels in valley could be cleaned and used for aquacultural production. Processing of mineral raw materials Some of Istria producers of white cement and other products stopped the production or have uncertain future, because of the location or lousy industrial politics. The location of EIPR is by all means favorable for development of complementary production with higher level of processing of raw minerals, and means natural development through cooperation with existing factories. The micro- and nano-technology allow a wide range of products from lime-stone mineral basis and usefull in processing of plastic in food, in metallurgy, in agriculture, in medicine etc. Building development strategy in that segment is very important task in EIPR project. Agricultural production and processing There is a huge potential in vegetable and Àower cultivation through modern technology. To get economically successful project the following is important: production direction and market availability, the use of new cultivating technologies, surface, quantity of production, irrigation water and irrigation system availability, costs of energy for winter period heating and for product processing, for example drying. On the other hand, the production and processing of industrial cultures as hemp might be very pro¿table. The availability of waste heating from energy station might be good condition for economic success of the project. In Raša valley the conditions: water availability, as well as possibility of the building of the processing facilities for agricultural products are very good. The port and railway facilitate transportation on the market. 3.4 Ecoindustrial Park and the Community of Raša

Every industrial park is in interaction with its living space (surroundings) and depends on it from the point of view of natural resources, services and social development and values. It is extremely important that the cooperation and understanding of the Community’s leadership and EIPR management is established since the initial phase of EIPR. The inhabitants

130

Annual 2010/2011 of the Croatian Academy of Engineering

are included from the beginning through the public discussions in the project and its inÀuence on living space, and number of bene¿ts: contributions to the economical and social development, improvement of infrastructure, improvement of life quality are accepted as logical result. For that purpose the consultations are organized, and Forum for socially responsible teritory, which is stimulating discussions about actual problems of sustainable development of the Community, has been established. EIPR will be the impulse for the development of educational institutions, and will stimulate employment and attract working force from the surroundings and therefore stimulate residential building. It leads on strengthening and diversi¿cation of existing infrastructure and services. Social activities will be developed, cultural institutions will be built, and it is to be expected that Raša becomes a friendly place for the start of new businesses, a place with high quality of life. Following are the goals of the Project Long term goals Realisation of highly ef¿cient and economically sustainable and ef¿cient ecoindustrial park of manufacturing and service industries, which would be based on principles of industrial symbiosis and realizing bene¿ts for the inhabitants of Raša and Labinština. Short term goals Enterprise management in the life space and management of EIPR Environment design of EIPR, landscape and horticultural design Renewal of the transport infrastructure Renewal and equipment of port area Trget – Bršica – Štalije Management of storm water and inundation protection Ecologically ef¿cient management with waste water from Raša and Labin and with production of biogas Processing of the mineral raw material Agricultural production and processing of agricultural products Production of biofuel and management of bioenergy Mariculture Water from deserted mine galleries as market product Building of Marina and Wind power station Goli as collaborative projects will support ef¿ciency of EIPR, so they are considered as the additional targets The development of tourism should be designed in harmony with EIPR and leaning on original ideas and innovation as separately important target 3.5 Stakeholders

We will point out initial statement of the major of Raša in the publication Prosperity towards sustainable development (7): A set of synergy related projects based on production of biofuel, management of water and agricultural production, logistic support, commodity business tailored to the needs of development of the Community of Raša, Istria and Croatia is in complete spirit of strategy of EU and Croatia on realisation of sustainable development, energy ef¿ciency and protection of nature with reduction of emission. The program will be realized with emphasis on practice of public – private partnership and is consistent with Croatian strategy of export support, and is its important contribution. Local community, Labin area and Istria are naturally interested for good solutions in activating the port of Raša, for use of industrial areas and for the prosperity of the economy in that space.

Annual 2010/2011 of the Croatian Academy of Engineering

131

Local entities that are leaning on available resources (raw material, water, energy, transport) with this project will gain great developing impulse and support, and location and market advantages that ensure the future prosperity. Republic of Croatia with this project will gain a rounded conception of development of the ports in basin of Rijeka, the solution for neglected port of Raša and impulses for the development of the rail transport route on the eastern part of Uþka-mountain. In Middle European area the interest for port of Raša exists as the port specialized for biomaterial with the possibility for their processing in area settled for industrial plants. The main reason for such thinking is the fact that in other north Adriatic ports there is no possibility to accept additional liquid cargo and to supply big biodiesel plants in middle Europe with vegetable oils, there is a need for sea transport, as well as for supplement of biofuel (biodiesel, ethanol). The interest for the port of Raša has been shown from the entrepreneurs from Italy and several Asiatic countries, especially from Malaysia and India. 3.6 Knowledge and experience on which venture relies

The suggested strategy is based on the synergy of knowledge and experience of their authors, who participated as individuals in number of projects and solved number of engineering and economic development tasks. We should emphasize the knowledge and experience by which the mayor of Raša supported the designing of the strategy, because knowing local society and its wide environment is especially important for strategy development. All this knowledge and experience has been implemented in the idea of this strategy. There exist huge literature dedicated to development and building of ecoindustrial parks, but it can be understood and implemented just with wide experience and understanding of systemic thinking. 4 Ecoindustrial Park Raša – EIPR - realization 4.1 The way of achieving realization of EIPR

Good cooperation of interested institutions, economic entities and management through the activity of Working group, Forum for socially responsible activity and public presentations and discussions, as well as good cooperation with Croatian System Society, form the basis of friendly surroundings for realization of the project EIPR. Everything mentioned above, together with learned experiences, particular knowledge and information, as well as gathered circle of experts, constitutes the intellectual capital without which would it be impossible to initiate realization of the project EIPR on the current cognitive and project level. The idea is for the Working group to continue its activity as a part of expert advisory panel of the company for the economy in life space and EIPR management. The ¿rst development task is appointment of project manager and establishment of the management of Ecoindustrial Park of Raša. This ¿rst in order of priority tasks will realize the Working group for strategy of development in cooperation with the Community of Raša and will create a multidisciplinary body – Expert working group or Expert advisory council of appropriate size and competence. It will be a support to the project manager (leader) and will jointly establish and operate the Park management. Expert working group will continue with technical work, de¿ne methodology of work, and prepare necessary documentation and detailed operational plan. Together with the Forum for socially responsible activity will promote the idea of EIPR and report to the public about

132

Annual 2010/2011 of the Croatian Academy of Engineering

progress of activities. With the progress and raise of EIPR it will be established consortium of companies, members of Park, which will be important support to the activity of Park management. 4.2 Economy of the project EIPR

Generally it can be predicted that potential of the space and suggested projects expressed as an annual income is around 200 million €. In different phases of realization of EIPR the projects need different amount of funds, in relation with actual structure of project realization, and are provided from public and private sources. In initial activities the needs for funds generally are: execution of Preliminary feasibility studies and preliminary engineering studies execution of EIP’s infrastructure and marketing activities for attraction of investors execution of possible pilot-plant and building of objects for initial settlement of entrepreneurs in rent establishment of the institutions that support development of EIP, as business incubators and development centers activities with which local / wider community supports the development of EIP and acts in accordance with projected targets The idea of EIPR could be realized if foreign investors with capital are attracted, that will be driving force of the project. It should also be considered ¿nancing from EU and Croatian funds. 4.3 Risks

The hardest thing is to predict, minimize and solve potential risks that could endanger project. But risk is a part of every day’s life and the causes should be individuated and separated. The suggested project is complex and includes number of subprojects and therefore different risk causes. We are listing the more important possible causes, and in more detailed way the causes will be analyzed during the elaboration of the singular project tasks: the success of the management of the EIPR careful choice of collaborative companies non favorable economic circumstances competition inÀuence Investors are loath to the risks, so it is important to decrease their exposure to the risks through wisely de¿ned conditions of investments. Let’s emphasize that, for example, public private partnership cannot signi¿cantly decrease exposure to risks, but there are several other useful approaches of decreasing risks that should be examined. Thinking the past and the present we cannot think good enough and live the already present future: informatic, biotechnological, nanotechnological that has already begun in highly developed countries greatly progressed and moved away. Is it possible to think and organize overcoming of the technical gap? Therefore positive thinking should be promoted, which is focused to the future, creativeness and permanent innovation, the understanding of our Àexible and creative abilities; it should wisely be open to the most developed world and its permanently changeable life. Let’s point that the old saying which states that mental blockade is the biggest obstacle on progress towards the future is true that is so called „embedded human resistance against anything new“, and the Europeans are the biggest slaves of such resistance and their lag behind Americans and Japanese is partly caused because of it. We are Europeans as well,

Annual 2010/2011 of the Croatian Academy of Engineering

133

and we also carry this defect, not so much when spending is concerned, but true when it concerns work and production. We hardly see the embryo of big changes, and therefore we defer too late. As a useful example we should examine the unsuccessful Lisabon strategy because of which EU Council in 2000 thought EU as „the most dynamic and competitive economy on the world…“ and the majority of targets was not reached till the predicted data, 2010 year. European culture is one of the basic risks to the possible success of our project, and we can list number of others. In the ¿rst line there is the possibility of initial ¿nancing of the venture, understanding of the surroundings, choice of the project management, then understanding of the surrounding community about the need of its achievement. 4.4 Development effects achieved with EIPR

On macro plan - Excellently organized community of producing agricultural and service activities taking care of industrial ecology by principles of ecoindustrial park, dependent on highly developed infrastructure: sea, road, rail and air. On micro plan - Support to the sustainable development of the living space and economy of the Community of Raša, employment and rich social life, work and family life, achieving reputation of the place of possible business success and quality of life. 4.5 Priority tasks

Realization of the whole EIPR project, as well as partial project courses inside of the system starts with the realization of project documentation and forming of singular project courses with all the necessary elements explained that are needed for initiate the realization, as it is used in profession. We are mentioning here the tasks that should be executed in short term period, in order to initiate the project: Preliminary project management with life space EIPR and coordination of activities Conceptual space design planning in EIPR Preliminary project of restoring and furnishing of the Port of Raša Preliminary project of activity and management with waste water from Labin and Raša Preliminary project of Cluster for the processing of the mineral raw material Preliminary project of agricultural production 5 Conclusion The above described idea of sustainable development of the Community of Raša and preparations for development of the Ecoindustrial Park of Raša, as well as description of circumstances in which the development tasks should be accomplished, tell us about a vision of a Park as ¿rst such Croatian venture: with wise management it can become widely recognized and accepted successful business model. During the deliberation of the Strategy of development of port area Raša points of view were de¿ned, that are starting points of sustainable development and with which development projects and activities should harmonize: System of projects thought on the basis of space, natural, infrastructural and technological resources Activating of the port of Raša and port area of Raša and its surroundings, as „anchorage“ of EIPR, that will through specializing operations have advantage to the other ports of Northern Adriatic. Planned activities in port area won’t disturb other development projects in spaces or tour-

134

Annual 2010/2011 of the Croatian Academy of Engineering

istic programs The area of EIPR will be designed by cleaning of the remained ruins from previous industrial periods and other activities (Bršica, Štalije, Vlaška, agricultural land along channel3) and by removing the stones (hill) between the deserted stone mine Bršica and Štalije area operational and industrial space will be formed. The basis is created for the development of strategically important and clean activities, the production that will in good way place Raša in regional environment (transport, bioenergy, food, water). Realization of the exemplary management with water: storm water and inundation protection, using the water from mine galleries from Raša and Labin, and the water from puri¿cation of waste water. It predicts netted system of energy management with emphasis on renewable sources of energy (biodiesel, biogas, wind power stations) Revitalisation and realization of ecological agricultural production in the valley of Raša, considering good climatic conditions, availability of water, energetic and other support of industrial systems Compliance to the rigorous norms of landscape conservation should harmonize with touristic activity and harmony with landscape beauty in nearby area. Realization of EIPR project follows usual procedure and phases of development for such project with predicted tasks and subprojects have to be realized: Company for space management (PGP) and management of EIPR Designing the space of EIPR: functional, landscape and horticultural design Renewal of traf¿c infrastructure Renewal and furnishing of the port-area Bršica-Štalije Management of storm water and inundation protection Ecologically ef¿cient treatment of waste water from Labin and Raša with biogas production Basis for forming of a Cluster for mineral raw material processing and strategy of development in that segment Agricultural production and processing of agricultural products Production of biofuel and management of bio energy Mariculture Water from deserted mine galleries as market product Building of marina and wind power station Goli as collaborative projects that would be supportive to the success of EIPR, so are emphasized as possible additional targets Realization of the whole project EIPR as well as singular units inside the Park will start with design of project documentation. 6 References 1. 2. 3. 4.

Andrašec M. (2009), Tehnologijske platforme, hrvatski razvojni projekti u EU okruženju. In: Božiþeviü J. (eds) Inovacijska kultura i tehnologijski razvoj, CROSS, Zagreb, pp 109-114 Baas L. (2008), Industrial Symbiosis in the Rotterdam Harbour and Industrial Complex, Bus. Strat. Env 17, pp 330-340 Božicevic J. (1993), Private communication, Zagreb Božiþeviü J. (ed) (1999), Luka kao složen sustav, Rijeka – glavna hrvatska luka, HATZ-CROSS, Zagreb

Annual 2010/2011 of the Croatian Academy of Engineering

5.

135

Božiþeviü J. (ed) (1999), Povezani Rijeka i Zagreb hrvatsko i europsko gospodarsko žarište, HATZ, Zagreb 6. Božiþeviü J., Andrašec M. (2009), Opüina Raša – Mudra zajednica, Opüina Raša, Raša 7. Božiþeviü J., Andrašec M. (2009), Napredak prema održivom razvoju, Opüina Raša, Raša 8. Doyle B. et al. (1996), Eco-industrial Parks, TRI, Research Triangle Park, NC, USA 9. Fleig A. K. (2000), Eco industrial Parks, Deutsche Gesellschaft fur Technische Zusamanarbeit GmbH, Eschborn, Deutschland 10. Lowe E. A. (2001), Eco-industrial Park Handbook for Asian Developing Countries, Report to Asian Development Bank, Indigo Development, Hidden Valley Lake, CA, USA 11. Koenig A.W. (ed.) (2005), The Eco-Industrial Park Development, A Guide for Chinese Government Of¿cials and Industrial Park Managers, EU-China Environmental Management Cooperation Programme - Industry Development, Beijing 12. Tarantini M., Paolo A., Dominici A., Peruzzi A., Dell’Isola M. (2007), Guidelines for the Settlement and the Management of the Sustainable Industrial Areas, The Experience of the LIFE SIAM Project, Bologna

136

Annual 2010/2011 of the Croatian Academy of Engineering

POSIBLE MODEL OF MATHEMATICAL EVALUATION OF BEHAVIOUR DISORDER THERAPY Prof. Dr. sc. Marijan Bošnjak, retired; member emeritus of Croatian Academy of Engineering (HATZ), HR-10000 Zagreb, Croatia;

Summary Because of enormous complexity of mental and behavioural disorders and illnesses there is still continuous need for further advances in their treatment. Treatment frequently asks the application of chemotherapeutic agents in addition to other conventional therapeutic actions. The effects of treatments should be adequately evaluated. In this work special emphasis is given to medical treatments when chemotherapeutic agents are applied. Therefore, the major importance was given to chemotherapeutic agent pharmacokinetics. Consequently, the mathematical model composed from more than twenty differential equations was applied to describe pharmacokinetics of chemotherapeutic agent distribution in human body, kinetics of development of positive and negative effects of applied chemotherapeutic treatment, effects on body viability and development of compounds inhibiting the effects of present chemo-therapeutic agent and causing the organism dependence on applied substances. Computer simulation was applied to test mathematical model convenience. Obtained computer simulation data con¿rmed the model adequacy suggesting its application in explaining real cases of combined psychotherapy treatments. Special convenience of proposed mathematical model type showed to be its con¿rmation of necessity for controlling chemotherapeutic substance distribution in bodies of treated patients as a prerequisite for more reliable successful patient treatments with reduced accompanied negative side effects. Key words: Psychotherapy pharmacokinetics. Mathematical modelling. Computer simulation application. 1 Introduction Despite the evident advances in discovering the causes of mental and behavioural disorders and illnesses as well as in increasing the frequency of successful patient disorder and illness treatments, resulting in patient recoveries, there is still continuous need for further advances in this area of medicine, because of enormous complexity of problems asking to be successfully solved. One of important prerequisites for further advances appears to be the application of appropriate approaches to the evaluation of results of medical treatments. Some consequences (positive and/or negative) of applied medical treatment can appear even very fast after its start, the others one can observe much later, whereas it can also happen that the desired effects cannot appear at all. Of course, the best situation can be considered that when desired positive effects appear already during the beginning phase of medical treatment. Since, in addition to direct communications of psychiatrists and employed personnel with patients the psychotherapy includes also different psychophysical activities and frequently

Annual 2010/2011 of the Croatian Academy of Engineering

137

an application of chemotherapeutic agents, the particular effects of all mentioned therapeutic actions should be appropriately evaluated prior to the conclusions referring to the answer on the question whether the applied therapy was adequate and whether it can be considered as successful with respect to patient recovery and his health improvement. This work refers to an attempt of advancing medical treatment by developing the one of possible models for mathematical evaluation of results of applied psychotherapy. Although the present approach to the development of mathematical model does not neglect a decisive role of engaged psychiatrists and other professional persons, the emphasis is given to the role of applied chemotherapeutic agents and their effects on patients subjected to the therapy. The impression based on the real practice of neglecting the relevance for assaying or estimating amounts of chemotherapeutic agents in particular compartments of patient body (mainly caused by ¿nancial reasons) provoked such an orientation to mathematical modelling. 2 Mathematical modelling 2.1 Development of fundamental model

As already long time known >1, 2, 10, 14@, different approaches to the mathematical modelling of biochemical reaction systems can be applied. In the case of biological systems the frequent practice is to describe them by systems of differential equations >110, 12-15@. These can be relatively simple (composed from 2 – 3 differential equations) or complex ones characterised with series of differential equations. Commonly, the simple systems of differential equations can be solved analytically or by applying computer simulation, whereas for complex systems the best choice appears to be the application of computer simulations. As pointed out in recent publication >2A@, the whole system of human organism certainly is so complex that one can consider illusory to expect that one can perfectly describe it by adequate mathematical model. It is also pointed out that the accepted practice became the description only of parts of realty and mainly more or less approximately. Consequently, process events in live organisms can be described by series of appropriate differential equations, as demonstrated with reference to microbial cells and/or microbial biomass >2-6, 8-10, 12-15@ and speci¿cally with reference to mammalian bodies >2, 8, 9, 15@. The data published recently >2@ could be considered to be the most relevant ones in forming the mathematical model of this work. Prior to the beginning of possible mathematical model formation one should select and group the most relevant patient body process events assumed to happen during the applied therapy and to indicate (at least roughly) the most probable therapy consequences. In any case the therapy effects on patient survival and viability maintenance should not be neglected in forming a mathematical model. The emphasis given to the evaluation of applied chemotherapeutic agent effects asks a more detailed consideration of chemotherapeutic agent pharmacokinetics. Data and explanations expressed in the recent work >2A@ suggest a similar approach in this work. Usual practice is to apply chemotherapeutic agents orally, although in some cases a parenteral application can be recommended. When applying chemotherapeutic agents orally one of possibilities is to consider that ventricular-intestinal organism compartment is composed from two sub-compartments. Then one can start mathematical modelling by putting differential equations: dFve/dt =-k1˜Fve

/1/

138

Annual 2010/2011 of the Croatian Academy of Engineering

dFit/dt =k1 ˜Fve –k2 ˜Fit –k3 ˜Fit

/2/

if neglecting chemotherapeutic drug resorption rate value in ventricular part (Such oversimpli¿cation cannot be recommended in every case), and if supposing that some part of drug would be excreted through faeces, i. e. if one can apply the equation dFfc/dt =k2 ˜ Fit

/3/

If, however, a chemotherapeutic agent resorption rate is enough signi¿cant already in ventricular part, then instead of equation /1/ the following equation should be applied: dFve/dt =-k1˜Fve –a˜Fve

/1a/

All organism parts are interconnected adequately and therefore communication between them is well established when one considers normal viable organism. However, with respect to organism survival and communications between organism parts the main role could be assigned to the cardio-vascular system and its connection with respiratory system, if respecting the control role of cerebral-neural system. Since the blood communicates with all organism compartments the rate of changes of chemotherapeutic drug mass in the blood could be expressed by equation dFsng/dt =k3 ˜Fit –(k4+k6+k7+k9+kk) ˜Fsng+k5 ˜Fcr+k8 ˜Fhp+ka ˜Ft+ +ki ˜Fmen

/4/

or by equation dFsng/dt =a ˜ Fve +k3 ˜ Fit –(k4+k6+k7+k9+kk) ˜ Fsng+k5 ˜Fcr+k8 ˜Fhp+ka ˜Ft+ +ki ˜Fmen /4a/ Communication with nephritic-urinary system is mainly irreversible and therefore drug transfer rate towards this system can be presented by applying the equation dFnu/dt =k6 ˜Fsng

/5/

The liver has important role and communicates reversibly with blood. The equation dFhp/dt =k7 ˜Fsng –k8 ˜Fhp

/6/

is considered to be mathematically adequate. For the main part of tissue one can apply the equations: dFt/dt =k9 ˜Fsng –(ka+kd) ˜Ft+ke ˜Ftr

/7/

dFtr/dt =kd ˜Ft –ke ˜Ftr

/8/

if drug transfer to tissue receptor happens reversibly. The hearth communicates reversibly with the blood, as the engine of blood circulation and as the main part of cardio-vascular system. Its reversible communication with the

Annual 2010/2011 of the Croatian Academy of Engineering

139

respiratory system is necessary and is a prerequisite of body survival. The equations dFcr/dt =k4 ˜Fsng –(k5+kb) ˜Fcr+kc ˜Frsp

/9/

and dFrsp/dt =kb ˜Fcr –kc ˜Frsp

/10/

appear to be adequate for application. Cerebral-neural and consequently mental system should be obligatorily considered. The rate of changes of drug amount in this system can be represented by equation dFmen/dt =kk ˜Fsng –ki ˜Fmen

/11/

Presence of chemotherapeutic substance in the cerebral-neural system can induce positive and/or negative effects inÀuencing their intensities. One of possibilities to express these consequences could be by applying the equations dMenp/dt =ku ˜Fmen ˜(1-Menp/Mpm)

/12/

dMenn/dt =kv ˜Fmen ˜(1-Menn/Mnm)

/13/

Equations /12/ and /13/ determine the maximal values which can be attained during medication, if neglecting the probable inhibitory effects of increased Fmen quantities. Since it is much more probable that the inhibitory effects would reÀect rather on positive than negative consequences, it seems to be suf¿cient to modify the equation /12/ with taking into account inhibitory effects of increasing Fmen quantities. The equation, dMenp/dt =ku ˜Fmen ˜(1-f˜Fmen-Menp/Mpm)

/12a/

can be considered adequate one for an explanation of expected consequences. The experience and experts recommendations concerning the application of chemotherapeutic agents in psychotherapy suggest us to be cautious, especially in cases of longterm their application. As mentioned recently >7@, none of known antipsychotics used to treat patients is quite adequate. Those called the typical antipsychotics are regarded as ef¿cacious at ameliorating positive symptoms, but generally ineffective against negative symptoms and with accompanied displaying of extra-pyramidal side effects. Atypical antipsychotics are also ineffective against negative symptoms. Therefore, one should consider possible negative effects of applied chemical substances on organism body of patients subjected to psychotherapy. One of ways of evaluating the consequences is by estimating values of arbitrarily de¿ned body viability, Bv, and body non-viability, Bnv. In present approach it is supposed that differential equations dBv/dt =-ks .Bv ˜Fsng+kr ˜Srv ˜Bnv

/14/

dBnv/dt =ks ˜Bv ˜Fsng–kr ˜Srv ˜Bnv

/15/

dSrv/dt =-kr ˜Srv ˜Bnv +krv ˜(1-Srv)

/16/

140

Annual 2010/2011 of the Croatian Academy of Engineering

can serve adequately to such a purpose, if one considers that live organism by his metabolism supplies continuously the own body with energy and appropriate substances in necessary amounts for survival, and that “revitalization substrate”, Srv, as the notion represents them. The effects of psychotherapy on patient recovery and his health improvement should be evaluated. However, one cannot say that actually reliable and exactly de¿ned methods are applied systematically. Therefore, every attempt in approaching to the problem solution can be useful. One of possibilities seems to be by proposing the mathematical expression in which the recognised effects are considered with respect to the participation of particular factors of inÀuence. No doubt, both the ef¿ciency of psychiatrists and employed professional personnel and effects of applied chemotherapeutic substances should be evaluated on appropriate ways. As the ¿rst step in searching an acceptable mathematical model could be the investigation of application convenience of equations dPs/dt =z˜(1-Ps)

/17/

dUt/dt =w˜Ps+x˜Menp-y˜Menn

/18/

Since the decision on the choice and dosing psychotherapeutic agents essentially depends on psychiatrists, perhaps it is more adequate instead of equation /17/ to apply the equation dPs/dt =z˜(1-Ps)˜(1+Menp)/(1+Menn)

/17a/

The idea to propose equations /17/, /17a/ and /18/ is based on the assumption that engaged psychiatrists and professional personnel are ethically and by their education and experience quite adequate to act ef¿ciently in positive sense. It is also supposed that both positive and negative consequences should be evaluated, primarily without an investigation whether the applied chemotherapeutic agent acts as reaction substance in biochemical processes or as signal substance inducing the biochemical processes which lead to desired patient behaviour and health improvements. Although there is series of other factors playing important roles in people therapies (religion, meteorological factors, environmental conditions in general, etc.), the evaluation of consequences of their inÀuence cannot be recommended for detailed consideration in this work, because of actual insuf¿cient knowledge to de¿ne and estimate exactly the effects of their inÀuence. 2.2 Modelling of the development of organism dependence on applied drugs

Frequently, organism dependence on applied drug can be induced to develop. Then it can be produced even with danger and irreversible consequences. Various mechanisms of induction and development of organism dependence on drugs can be supposed. With respect to mathematical modelling one of possibilities is to suppose that a dependence product, Dp, is produced in proportion to the rate of negative effects, Menn production and that it acts as an inhibitor of positive effects and/or as a drug inactivation factor. If, supposedly, one could consider that induction and development of dependence product formation mainly happen in the cerebral-neural compartment, and if roughly it could be suf¿cient to focus our interest to events taking place in cerebral-neural system, then the following approach could be applied in transforming the discussed mathematical model: Let suppose that the dependence product, Dp inactivates the applied drug to high extent irreversibly, i.e. in accordance to the scheme:

Annual 2010/2011 of the Croatian Academy of Engineering

Dp + Fmen ĺ Fmni

141

/Sh-1/

If the expressed reduction rate of positive effects becomes proportional to Dp amount, then the applicability of following equations can be expected: dDp/dt =b ˜dMenn/dt –c ˜Dp ˜Fmen

/19/

dFmni/dt =c ˜Dp ˜Fmen

/20/

dFmen/dt =kk ˜Fsng –ki ˜Fmen - c ˜Dp ˜Fmen

/21/

2.3 When one can stop to apply chemotherapeutic agent?

If one observes that positive effects do not show a tendency to decrease, then one can recommend a starting of the procedure of decreasing drug doses, or even of stopping a given drug application, especially if with stopping drug application negative effects would start to reduce. Therefore, it appears that instead of equation /12a/ the equation /22/ can be recommended for application, i.e. dMenp/dt =ku ˜Fmen ˜(1-f˜Fmen-Menp/Mpm) –d ˜Dp

/22/

Such a consideration does not ask detailed discussion referring to the kinetics of inactivated drug, Fmni. If both positive and negative effects will spontaneously reduce in proportion to their speci¿c rates, then instead of equation /22/ and equation /13/ one can propose the equations: dMenp/dt =ku ˜Fmen ˜(1-f˜Fmen-Menp/Mpm) –d ˜Dp –g ˜Menp

/23/

dMenn/dt =kv ˜Fmen ˜(1-Menn/Mnm) –h ˜Menn

/24/

Equations /23/ and /24/ suggest the drugs with higher h/g values should be preferred for application under a condition they do not differ with respect to other their properties. To facilitate the understanding of properties of transformed mathematical model some simpli¿cations can be introduced in order to have more convenient ¿nal model. It seems that lumping of the terms Fve and Fit into common Fit, as well as of Fhp, Ftr and Ft into common term Ft could markedly correspond to the desired goal. Then the equations /1/, /1a/, /6/ and /8/ could be neglected, whereas instead of equations /2/, /4/, /4a/ and /7/ the corresponding equations /25/, /26/ and /27/ could be applied: dFit/dt =-k1 .Fit –k2 .Fit

/25/

dFsng/dt =k1 ˜Fit –(k4+k6+k9+kk) ˜Fsng +k5 ˜Fcr +ka ˜Ft +ki˜Fmen

/26/

dFt/dt =k9 ˜Fsng –ka ˜Ft

/27/

2.4 Modelling of the psychotherapeutic substance transformation

The applied psychotherapeutic agent, depending on its chemical structure and its

142

Annual 2010/2011 of the Croatian Academy of Engineering

instantaneous position in organism body, can be and usually is transformed into degradation products, at least partly. Degree and the rate of transformation depend on both substance characteristics and biodegradative capability of particular organism, especially of organism corresponding parts. The formed degradation products can be therapeutically inactive, more or less toxic, of higher or lower therapeutic effects, when compared with original substance, etc. Commonly, signi¿cant or even the main part of applied substance mass is eliminated from organism body as the substance of original structure, while the other part is eliminated as applied substance degradation products. In many cases, the applied substance is primarily converted into the product (metabolite) of similar therapeutic effects as original substance. Of course, there is spectrum of ef¿ciency which could be expressed by formed metabolites. In this work, pharmacokinetics of formed therapeutically active product is considered to demonstrate how some speci¿c situations can be described by corresponding mathematical model. Substance transformation can be considered as to happen in organism body in general, or as to happen mainly in the speci¿c tissue (e.g. in liver), or in body tissue in general. Since cardiovascular system communicate with every particular body compartment, one can consider the substance transformation quickly reÀected to the state in cardiovascular system. Therefore, in the extended mathematical model, instead of equation /26/ the equations dFsng/dt =k1˜Fit-(k4+k6+k9+kk+j)˜Fsng+k5˜Fcr+ka˜Ft+ki˜Fmen

/26a/

dFtsng/dt =j˜Fsng-(k6+kk)˜Ftsng+ka˜Ftt+ki˜Fmt

/28/

should be applied, if neglecting the formed product af¿nity to the hearth and if supposing its transfers concerning other body compartments happen with same speci¿c rates as those referring to transfers of the original substance. In addition, the corresponding new equations should be also included: dFnut/dt =k6˜Ftsng

/29/

dFtt/dt =k9˜Ftsng-ka˜Ftt

/30/

dFmt/dt =kk˜Ftsng-ki˜Fmt

/31/

It is supposed that formed product inÀuences also positive and negative effects of patient behaviour as well as body viability. Therefore, instead of equations /12/, /12a/, /13/, /14/ and /15/ the following equations should be applied: dMenp/dt =ku˜(Fmen+Fmt)˜(1-f˜Fmen-Menp/Mpm)

/32/

dMenn/dt =kv˜(Fmen+Fmt)˜(1-Menn/Mnm)

/33/

dBv/dt =-ks˜Bv˜(Fsng+Ftsng)+kr˜Srv˜Bnv

/34/

dBnv/dt =ks˜Bv˜(Fsng+Ftsng)-kr˜Srv˜Bnv

/35/

Of course, the other possibilities could be supposed. However, there is no need to consider them now.

Annual 2010/2011 of the Croatian Academy of Engineering

143

2.5 Relevance of drug application method and pharmaceutical product form

It is known that different drug application methods are practised. Parenteral application is preferred when fast effects are desired and/or when active substance is sensitive to the action of active factors present in ventricular-intestinal compartment. Active substance stability and pharmaceutical drug form properties play important role in applied drug convenience and therapeutic ef¿ciency. Active substance half-life time is considered to be one of very relevant pharmacokinetics parameters, especially when different active substances should be applied during therapy. In the case of per oral drug application the resorption rate of active substance markedly affects its pharmacokinetics and therefore therapeutic consequences. The fast resorption rate is desired when the fast therapeutic effects should result (e.g. when applying hypnotics). However, for longer therapy periods and for maintenance of more stable active substance concentrations with minimal oscillations of active substance concentrations in corresponding organism compartments the slower active substance resorption rates could be recommended to be realised. Some of computer simulation data demonstrating the effects of different substance resorption rates are shown in Tables 2 – 3 and Figs 3a,b. 3 Computer simulations Computer programme Scientist enables a solving of number of different scienti¿c problems. Based on previous and especially of recent experience >2, 5, 6@ it was applied also in this work. 4 Results and discussion 4.1 Fundamental comments

Chosen examples of computer simulations are presented in Fig. 1 to Fig. 3. To see how the estimated consequences relate to quantities in particular organism body compartment the calculation results are presented in Table 1. Table 1

Effects of some psychotherapeutic parameters (ku, kv, w, x, y, z) on the psychotherapy ef¿ciency (Ut). Ut values refer to the 50th hour of the computer simulation process. Parameter name

ku

kv

w

x

y

z

Ut

Values

0.02 0.022 0.02 0.02 0.02 0.02 0.02 0.02 0.02

0.015 0.025 0.02 0.02 0.02 0.02 0.02 0.01 0.01

0.001 0.001 0.001 0.001 0.001 0.002 0.002 0.002 0.002

0.001 0.001 0.001 0.002 0.001 0.001 0.001 0.001 0.002

0.001 0.001 0.001 0.001 0.002 0.002 0.002 0.002 0.001

0.001 0.001 0.001 0.001 0.001 0.001 0.002 0.002 0.002

0.00492 -0.000525 0.00123 0.0218 -0.01934 -0.01811 -0.01573 0.000668 0.033608

It is known that distribution of pharmacokinetic substances depends on substance properties,

144

Annual 2010/2011 of the Croatian Academy of Engineering

way of its application and its applied amount, af¿nity of organism compartments to accept it and to retain its quantity, particular organism compartment mass and physiological state, whole organism viability and activity, etc. In this work, values of computer simulation parameters were chosen arbitrarily in order to demonstrate acceptability of proposed approach and possible applicability of supposed mathematical model, expecting for both to be conditionally adequate. Chosen computer simulation examples demonstrated in Fig. 1 to Fig. 3 show how parameter values reÀect on distribution of pharmacokinetic substance in treated patient organism. Psychotherapy experts know well that rarely one can expect evident patient health improvement during very short time of treatment. Therefore, one can suggest considering the consequences of longer patient treatment periods. Some insight one can obtain by observing Fig. 1. Psychotherapy pharmacokinetics

0.25

0.00 0

10

20

30

Time (h)

40

0.046

-0.005

50

0.65

0.30

-0.05

-0.40 0

FVE_CALC vs T FIT_CALC vs T FFC_CALC vs T FSNG_CALC vs T FCR_CALC vs T FNU_CALC vs T FHP_CALC vs T FT_CALC vs T FTR_CALC vs T FMEN_CALC vs T FPL_CALC vs T MENP_CALC vs T MENN_CALC vs T BV_CALC vs T BNV_CALC vs T SRV_CALC vs T UT_CALC vs T PS_CALC vs T

a)

0.200

1.00

10

20

30

Time (h)

40

0.138

0.075

Ut,Ps (units)

0.098

Y (mg)

0.50

0.149

Ut,Ps (units)

0.75

10.0 9.5 9.0 8.5 8.0 7.5 7.0 6.5 6.0 5.5 5.0 4.5 4.0 3.5 3.0 2.5 2.0 1.5 1.0 0.5 0.0

Bv,Bnv,Srv,Menp,Menn (nor.units)

0.200

1.00

Bv,Bnv,Srv,Menp,Menn (nor.units)

Y (mg)

Psychotherapy pharmacokinetics 10.0 9.5 9.0 8.5 8.0 7.5 7.0 6.5 6.0 5.5 5.0 4.5 4.0 3.5 3.0 2.5 2.0 1.5 1.0 0.5 0.0

0.012

-0.050

50 FVE_CALC vs T FIT_CALC vs T FFC_CALC vs T FSNG_CALC vs T FCR_CALC vs T FNU_CALC vs T FHP_CALC vs T FT_CALC vs T FTR_CALC vs T FMEN_CALC vs T FPL_CALC vs T MENP_CALC vs T MENN_CALC vs T BV_CALC vs T BNV_CALC vs T SRV_CALC vs T UT_CALC vs T PS_CALC vs T

b) Fig. 1a,b

Simulated psychotherapy course for applied same computer simulation parameters during two consecutive cycles: k1=0.3; k2=0.01; k3=0.8; k4=0.1; k5=0.5; k6=0.02; k7=0.05; k8=0.08; k9=0.1; ka=0.1; kb=0.05; kc=0.03; kd=0.1; ke=0.05; Kk=0.1; ki=0.15; ku=0.02; kv=0.01; Mpm=1.0; Mnm=1.0; ks=0.005; kr=0.1; krv=0.5; w=0.002; x=0.002; y=0.001; z=0.002; f =0.65. Initial value for Fve was the same for both a) and b) cycles; other initial independent variable quantities of the 2nd cycle (b) were those at the 50th hour of the simulated ¿rst (a) cycle; As shown, some relative improvement one can expect if chosen substance can induce or cause an improvement and if it cannot lead to negative consequences of applied chemotherapeutic treatment. Results suggest the giving of more importance to the role of experts whenever the experts can inÀuence the patients by inducing positive consequences at them. Data presented in Table 1 can explain the reasons why one can recommend the estimations and/or assaying of chemotherapeutic quantities in particular patient organism

Annual 2010/2011 of the Croatian Academy of Engineering

145

compartments. Fig. 1a, 1b demonstrate that equation /12a/ can be applied in explaining negative consequences of increasing the amounts of pharmacokinetic substances in cerebral-neural tissues. Evidently, data in Fig. 1a, 1b support the reasons why one should recommend the assaying and/or estimating of quantity distribution of chemotherapeutic agents in particular compartments of human organism body. The organism dependence on applied drugs can be explained if applying the transformed mathematical models. One example of computer simulation is demonstrated in Fig. 2a, 2b. Psycho- therapy pharmacokinetics

0.00 0.0

4.8

9.6

14.4

Time (h)

19.2

0.013

0.000

6.0 5.7 5.4 5.1 4.8 4.5 4.2 3.9 3.6 3.3 3.0 2.7 2.4 2.1 1.8 1.5 1.2 0.9 0.6 0.3 0.0

24.0

1.00

0.75

0.50

0.25

0.00 0.0

FIT_CALC vs T FF_CALC vs T FSG_CALC vs T FC_CALC vs T FTSG_CALC vs T FNU_CALC vs T FT_CALC vs T FM_CALC vs T FR_CALC vs T MEP_CALC vs T MEN_CALC vs T BV_CALC vs T BNV_CALC vs T SRV_CALC vs T U_CALC vs T PS_CALC vs T DP_CALC vs T FMI_CALC vs T FTT_CALC vs T FMT_CALC vs T FUT_CALC vs T

a)

4.8

9.6

14.4

Time (h)

19.2

0.100

0.075

0.050

0.025

Y3-U,Ps,Dp (arb.units)

0.25

0.025

Y (mg)

0.50

0.038

Y3-U,Ps,Dp (arb.units)

0.75

0.050

Mep,Men,Bv,Srv (nor.units)

1.00

Mep,Men,Bv,Srv (nor.units)

Y (mg)

Psycho- therapy pharmacokinetics 6.0 5.7 5.4 5.1 4.8 4.5 4.2 3.9 3.6 3.3 3.0 2.7 2.4 2.1 1.8 1.5 1.2 0.9 0.6 0.3 0.0

0.000

24.0 FIT_CALC vs T FF_CALC vs T FSG_CALC vs T FC_CALC vs T FTSG_CALC vs T FNU_CALC vs T FT_CALC vs T FM_CALC vs T FR_CALC vs T MEP_CALC vs T MEN_CALC vs T BV_CALC vs T BNV_CALC vs T SRV_CALC vs T U_CALC vs T PS_CALC vs T DP_CALC vs T FMI_CALC vs T FTT_CALC vs T FMT_CALC vs T FUT_CALC vs T

b) Fig. 2a,b:

Simulation of active degradation product formation by applied drug transformation. Computer simulation based on transformed extended mathematical model. Applied values of computer simulation parameters: k1=0.8; k2=0.01; k4=0.1; k5=0.5; k6=0.02; k9=0.1; Kk=0.1; ka=0.1; b=0.01; c=0.03; d=0.0001; ki=0.15; ku=0.02; f=0.5; kv=0.01; g=0.0001; h=0.1; Mpm=1.0; Mnm=1.0; ks=0.005; kr=0.1; krv=0.5; w=0.002; x=0.002; y=0.001; z=0.002; New parameter j=0.12 was applied. Initial drug quantity for both a) and b) was the same; the other quantities applied for b) were those ¿nal for a); In practice, patients are frequently classi¿ed as belonging to different groups or types, based on the criteria of debatable reliability and therefore sometimes wrongly or even with danger consequences. Here, perhaps is reasonable to mention the Latina sentence: Quot homines tot sententiae, which refers to all people in general and therefore it certainly can be applicable to patients as well. Therefore, it follows that the best patient classi¿cation criterion seems to be by considering every patient as individual asking speci¿c treatment and strong control of drug quantity distribution during therapy. Data in Tables 2 – 3 and Figs. 3a,b show one of possible ways for therapy improvements or even optimisation.

146

Annual 2010/2011 of the Croatian Academy of Engineering

As demonstrated, slowed resorption leads to bene¿cial consequences. Therefore, there is reason for production of pharmaceutical products of properties which correspond to desired active substance resorption rate. Table 2 Effects of active substance resorption rate (value of resorption rate constant k1) on values of pharmacokinetics parameters at 10th hour after drug application. Parameter name Initial value FIT FF FSG FC FTSG FNU FT FM FR MEP MEN BV BNV SRV U PS DP FMI FTT FMT FUT * **

100.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.99 0.01 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0

Value at 10th hour for given k1 value k1 =0.8/h 0.030354 1.2342 16.699* 3.8810 17.915 6.2270 18.144 14.235 2.1823 -5.8849 0.59911 0.33840 0.66160 0.88626 -0.058650 -0.022766 0.0011259 0.0048652 9.0919 7.7490 2.6062

k1 =0.08/h 40.657 6.5937 15.413** 2.8076 8.5609 2.5503 8.6117 7.2429 0.82074 -1.1067 0.28760 0.57298 0.42702 0.93331 -0.005600 0.014274 0.0016591 0.0012169 3.1410 2.75743 0.84128 maximum attained during the 2nd hour, maximum attained during the 8th hour

Table 3 Effect of active substance repeated per-oral additions on pharmacokinetics quantities. Application of drug form adapted for active substance resorption rate de¿ned with k1 = 0.08/h.

Annual 2010/2011 of the Croatian Academy of Engineering

Quantity name

Initial value

Simulated value Simulated value after 8th hour after 25 mg active substance addition at 8th hour and after 16 hours

FIT FF FSG FC FTSG FNU FT FM FR MEP MEN BV BNV SRV U PS DP FMI FTT FMT FUT

100.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.99 0.01 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0

48.675 5.7028 15.600 2.6998 7.0583 1.9287 7.0789 6.1546 0.58643 -0.54687 0.19988 0.63995 0.36005 0.94798 -0.0018992 0.013684 0.0014112 0.00058767 2.0975 1.8886 0.52811

35.861 9.9043 17.548 3.3293 13.705 4.7860 13.049 10.210 1.5909 -3.2256 0.51099 0.42220 0.57780 0.90219 -0.034447 0.0042820 0.0013068 0.0038032 7.0013 5.7933 2.2175

147

Simulated values after 50 mg active substance additions at 8th and at 16th hour Value after Value after 24th hour 16th hour 48.030 47.716 11.330 16.920 21.448 25.419 4.0042 4.8585 15.469 24.329 5.2682 9.3740 14.819 20.868 11.749 15.583 1.7375 3.0097 -3.7939 -6.4175 0.53363 0.69866 0.38628 0.27897 0.61372 0.72103 0.89739 0.87648 -0.037620 -0.12630 0.0022673 -0.039698 0.0012736 0.00039878 0.0040628 0.0065879 7.5256 14.736 6.2655 11.627 2.3495 5.5524

148

Annual 2010/2011 of the Croatian Academy of Engineering Psycho- therapy pharmacokinetics

0.0 0

2

4

6

Time (h)

8

0.025

0.000

1.00

0.75

0.50

0.25

0.00 0

10 FIT_CALC vs T FF_CALC vs T FSG_CALC vs T FC_CALC vs T FTSG_CALC vs T FNU_CALC vs T FT_CALC vs T FM_CALC vs T FR_CALC vs T MEP_CALC vs T MEN_CALC vs T BV_CALC vs T BNV_CALC vs T SRV_CALC vs T U_CALC vs T PS_CALC vs T DP_CALC vs T FMI_CALC vs T FTT_CALC vs T FMT_CALC vs T FUT_CALC vs T

a)

2

4

6

Time (h)

8

0.100

0.075

0.050

0.025

Y3-U,Ps,Dp (arb.units)

0.3

0.050

100 95 90 85 80 75 70 65 60 55 50 45 40 35 30 25 20 15 10 5 0

Mep,Men,Bv,Srv (nor.units)

0.5

0.075

Y (mg)

0.8

0.100

Y3-U,Ps,Dp (arb.units)

1.0

Mep,Men,Bv,Srv (nor.units)

Y (mg)

Psycho- therapy pharmacokinetics 100 95 90 85 80 75 70 65 60 55 50 45 40 35 30 25 20 15 10 5 0

0.000

10 FIT_CALC vs T FF_CALC vs T FSG_CALC vs T FC_CALC vs T FTSG_CALC vs T FNU_CALC vs T FT_CALC vs T FM_CALC vs T FR_CALC vs T MEP_CALC vs T MEN_CALC vs T BV_CALC vs T BNV_CALC vs T SRV_CALC vs T U_CALC vs T PS_CALC vs T DP_CALC vs T FMI_CALC vs T FTT_CALC vs T FMT_CALC vs T FUT_CALC vs T

b) Fig. 3a,b.

Effect of active substance resorption rate on pharmacokinetics quantities: a) k1=0.8; b) k1=0.08; Concerning a chemotherapeutic substance distribution in organism body one should point out that in the proposed mathematical model the chemotherapeutic substance quantities represent chemotherapeutic substance masses in particular compartments, not their mass concentrations in corresponding compartments. Such an approach to expressing chemotherapeutic substance distribution can be considered as more convenient than that referring to chemotherapeutic substance mass concentrations, since at given time the corresponding substance mass concentration cannot be considered to be constant with reference to every compartment part. 4.2 Speculative comments (conclusions)

Enormous complexity of psychotherapy asks serious comments. Some questions one can put here: Are we capable to recognise, in every case, which part of patient behaviour can be and should be changed, and how to proceed in order to realise desired positive changes?! Are we always sure that chosen chemotherapeutic agent is adequate one?! Whether at all the application of any chemotherapeutic agent can be recommended?! Is it reasonable a neglecting of prayers?! Religious and spiritual factors in general certainly should be taken into account with giving them adequate relevance. Which degree of tranquilisation of particular patient could be considered as optimal one for his health?! How to enhance the endogenous production of tranquilisation factors (endorphin, e.g.) in improving the

Annual 2010/2011 of the Croatian Academy of Engineering

149

dreaming and sleeping in order to reduce the need for an application of hypnotics?! Why not to consider that particular effect of any factor can be quantitatively evaluated, at least approximately?! Is it possible to de¿ne the position of psychiatrists, since the most frequent situation is that they are simultaneously engaged in treating relatively large number of patients?! One of answers on all mentioned questions certainly should be that exact information on chemotherapeutic substance distribution in organism body is of essential importance for achieving of positive results of psychotherapeutic treatment which includes the application of chemotherapeutic substances. Therefore, methods of assaying and estimating of chemotherapeutic substance contents in particular human organism compartments should be available for application whenever one estimates that such information could be helpful. Application of chemotherapeutic substances in general can induce or cause different undesired side effects. These can refer to particular organism body compartments or to the whole organism. Live organism body can be induced, and frequently is capable to degrade and/or transform the used substance. As a result of organism body answer could be the overproduction of some products asking therapeutic action as well. Therefore, one suggests a serious consideration of all these effects and possible their description. The extended mathematical model can enable a more exact expression of particular side effect causes and consequences. The approach similar to that applied for cerebral-neural compartment can be also applied for any considered compartment. Then, particular positive and/or undesired negative effects could be evaluated with reference to the whole organism body or to particular its organs or tissues. As shown recently >2A@ by expressing roughly the decisive role of cardio-vascular and respiratory compartments in affecting the quantity of “revitalisation substrate”, Srv, in this work one can focus onto evaluation of consequences referring to vital functions of liver, or of nephritic-urinary and other organism body compartments. Concerning nephritic-urinary system it should be pointed out that the total amounts of pharmacokinetic agent transferred to it differ from those present in it at given time, because of periodic withdrawals of urine from nephritic-urinary system. Therefore, the amounts of pharmacokinetic substances actually present in the nephritic-urinary system represent the difference between the total amounts transferred into the system and the total amounts withdrawn from the system. Similar consideration as for nephritic-urinary system can also be applied concerning the pharmacokinetic substance quantity present in the faeces of large intestine. The advantages of developed mathematical model application can become especially evident if one tends to estimate consequences of simultaneous application of different drugs in the therapy. Based on the kinetic relationships established for particular substance one can predict the consequences of simultaneous applications of two or more different substances during the planned psychotherapy treatments, if applying computer simulations. These could be performed with reference to every patient. It is important to point out that our possibilities to represent the events in organism body by mathematical models are not limited only on models presented here. Our knowledge referring to other events can also be translated into corresponding mathematical models, and therefore in those convenient for the use in performing desired computer simulations. By the way, it is important to point out that presented mathematical model can simply be transformed into much simpler models by neglecting relevance of those events which should not be obligatory considered for evaluation of consequences of more relevant events. Finally, there are questions which should not be avoided. One of them is how to estimate normalised values of quantities of Bv, Bnv, Srv, Menp, Menn, etc. No doubt, a reliability

150

Annual 2010/2011 of the Croatian Academy of Engineering

of estimations will markedly depend on the experience of psychotherapy experts and on the availability of acceptable evaluation methods. However, the experts certainly can and know de¿ne, at least approximately, the maximal speci¿c quantity values, and compare them with values estimated during the patient observation and treatments. Therefore, there is a certain possibility for at least slight advances in evaluating the ef¿ciency of psychotherapy treatments. Of course, evaluations should be performed in the right time. One cannot recommend e.g. a postponing of Bnv evaluations till a situation when the pathologist opinion would become the most relevant. On the other hand, we should always take into account the fact that the frequency of successful therapy cases shows positive trend. Moreover, relevant pharmaceutical companies (e.g. LILLY USA, LLC. >11@) give enough data for optimal approach to the use of psychotherapeutic agents. Therefore, there is no doubt for a need of controlling drug distribution in human body during therapy. Already long time introduced practice of determination of alcohol concentrations in blood samples of persons causing traf¿c accidents justify the giving adequate relevance to the control of psychotherapeutic drugs distribution in the body of every patient subjected to psychotherapy which includes the drug application. Symbols a kinetic constant, h-1 (T -1) Bv body vitality, dimensionless Bnv body non-vitality, dimensionless b kinetic constant, dimensionless c kinetic constant, h-1 (T -1) Dp dependence product normalised mass, dimensionless d kinetic constant, h-1 (T -1) Fcr pharmacokinetic substance mass in hearth, mg (M) Ffc pharmacokinetic substance mass in faeces, mg (M) Fhp pharmacokinetic substance mass in liver, mg (M) Fmen pharmacokinetic substance mass in neural-brain system, mg (M) Fmni inhibited Fmen, mg (M) Fmt transformed pharmacokinetic substance mass in neural-brain system, mg (M) Fnu pharmacokinetic substance mass in nephritic–urinary system, mg (M) Fnut transformed pharmacokinetic substance mass in nephritic–urinary system, mg (M) Frsp , Fplm pharmacokinetic substance mass in respiratory system, mg (M) Fsng pharmacokinetic substance mass in blood, mg (M) Ftsng transformed pharmacokinetic substance mass in blood, mg (M) Ft pharmacokinetic substance mass in tissue, mg (M) Ftt transformed pharmacokinetic substance mass in tissue, mg (M) Ftr pharmacokinetic substance mass in tissue receptor, mg (M) Fve, Fit pharmacokinetic substance mass in ventricular (Fve)-intestinal (Fit) compartment, mg (M) g kinetic constant, h-1 (T -1) h kinetic constant, h-1 (T -1) j kinetic constant, h-1 (T -1) -1 k1 kinetic constant with reference to Fve, h (T -1) -1 k2 kinetic constant with reference to Fit, h (T-1) -1 k3 kinetic constant with reference to Fit, h (T -1) k4 kinetic constant with reference to Fsng, h-1 (T -1) -1 k5 kinetic constant with reference to Fcr, h (T -1)

Annual 2010/2011 of the Croatian Academy of Engineering

k6 k7 k8 k9 ka kb kc kd ke ki kk kr krv ks ku kv Menn Menp Mnm Mpm Ps Srv Ut w x y z

-1

kinetic constant with reference to Fsng, h kinetic constant with reference to Fsng, h-1 kinetic constant with reference to Fhp, h-1 kinetic constant with reference to Fsng, , h-1 kinetic constant with reference to Ft, h-1 kinetic constant with reference to Fcr, h-1 kinetic constant with reference to Frsp, h-1 kinetic constant with reference to Ft, h-1 kinetic constant with reference to Ftr, h-1 kinetic constant with reference to Fmen,h-1 kinetic constant with reference to Fsng,,h-1 kinetic constant with reference to Bnv and Srv, h-1 kinetic constant with reference to Srv, h-1 kinetic constant with reference to Bv and Fsng, mg-1h-1 kinetic constant with reference to Fmen and Menp, mg-1h-1 kinetic constant with reference to Fmen and Menn, mg-1h-1 normalised quantity of negative effects caused by Fmen, dimensionless normalised quantity of positive effects caused by Fmen, dimensionless maximal value of Menn, dimensionless maximal value of Menp, dimensionless normalised effect of psychiatrist and personnel, dimensionless normalised quantity of revitalisation substrate, dimensionless normalised psychotherapy ef¿ciency, dimensionless kinetic constant with reference to Ps, h-1 kinetic constant with reference to Menp, h-1 kinetic constant with reference to Menn, h-1 kinetic constant with reference to Ps, h-1

151 -1

(T ) (T -1) (T -1) (T -1) (T -1) (T -1) (T -1) (T -1) (T -1) (T -1) (T -1) (T -1) (T -1) (M-1T -1) (M-1T -1) (M-1T -1)

(T -1) (T -1) (T -1) (T -1)

References 1. 2. 3. 4.

5. 6. 7. 8.

Bošnjak M. (1982) Mathematical modelling of biochemical reaction systems (in Croatian) Kem. Ind. 31 545 -559 Bošnjak M. (2009) Introduction to Kinetics of Microbial Processes ( in Croatian: Uvod u kinetiku mikrobnih procesa), Graphis, Zagreb, (2A= pp. 275-306, Appendix) Bošnjak M., Topolovec V., Vrana M. (1978), Growth kinetics of Streptomyces erythreus during erythromycin biosynthesis, J. Appl. Chem. Biotechnol. 28 791- 798 Bošnjak M., Topolovec V., Johanides V. (1979), Growth kinetics and antibiotic synthesis during the repeated fed batch culture of streptomycetes, In: Armiger W. B. (ed.), Computer Applications in Fermentation Technology, Biotechnol. Bioeng. Symp. No. 9 155 -165, Wiley & Sons Bošnjak M., Bago Joksoviü A., Pigac J., Bošnjak Cihlar Ž., Hranueli D., (2006) Applicability of mathematical models in de¿ning the behaviour kinetics distinction among microbial strains, Chem. Biochem. Eng. Q. 20 375 -388 Bošnjak M., Udikoviü Koliü N., Petriü I., Cihlar D., Hršak D., (2010) Integrated approach to mathematical modelling of atrazine degradation in different reaction systems, Food Technol. Biotechnol.48 392-403 Ciof¿ Ch. L., Liu Sh., Wolf M. A., (2010) Recent development in glycine transporter-1 inhibitors, In: J. E. Macor (Ed.), Annual Reports in Medicinal Chemistry,Vol.45, pp. 19-35, Academic Press, Burlington, MA 01803, USA Karba R., Mrhar A., Bremšak F., Kozjak F., (1977) Estimation of pharmacokinetics

152

9.

10.

11. 12. 13. 14. 15.

Annual 2010/2011 of the Croatian Academy of Engineering

parameters of amoxicillin and ampicilline by identi¿cation (in Slovenian: Doloþevanje farmakokinetiþnih parametrov amoksicilina in ampicilina s pomoþjo identi¿kacije), In: Informatica 77, Proc. Yug. Intern. Symp. 6 218, Slov. Soc. Informatika, Ljubljana van der Kleijin E., Guelen P. J. M., van Wijk C., Baars I., (1975) Clinical pharmacokinetics in monitoring chronic medication with anti-epileptic drugs, In: Schneider H., Janz D., Gardner-Thorpe C., Meinardi H., Shervin A. L. (eds.), Clinical Pharmacology of Anti-Epileptic Drugs, Springer, Berlin- Heidelberg –New York, pp. 11 -13 Kossen N. W., (1979) Mathematical modelling of fermentation processes: Scope and limitation, In: Bull A. T., Elwood D. C., Ratledge C. (eds.) Microbial Technology: Current State, Future Prospects, 29th Symp. Soc. Gen. Microb., Cambridge, University Press, Cambridge, , pp. 327 -35 LILLY USA, LLC Internet sites. Li R. C., Nix D. E., Schentag J. J., (2006) Pharmacodynamic modelling of bacterial kinetics: E-Lactam antibiotics against Escherichia coli, J. Pharm. Sciences 83 970 -975 Nikolaou M., Schilling A. N., Vo Giao, Chang Kai-tai, Tam V. H., (2007) Modeling of microbial population responses to time-periodic concentrations of antimicrobial agents, Anuals Biomed. Eng. 35 1458 -1470 Ramkrishna D., (1979) Statistical models of cell populations, In: Ghose T. K., Fiechter A., Blakebrough N. (eds.), Advances in Biochemical Engineering, Vol. 11, Springer, Berlin-Heidelberg-New York, pp. 1 -47 Roepke H. und Riemann J.,(1979) Analog Computer in Chemie und Biologie, Springer, Berlin, 1969

Appendix Spirituality based therapy and its possible relevance with reference to behaviour disorder treatments There are cases of health improvements observed after patients being subjected to spirituality based therapy. Therefore, its relevance should not be neglected regardless whether somebody is religious or not. However, the evaluation of effects based on such a therapy probably asks the speci¿c approach and explanations. Certainly there is no need to explain why the health of those religious was improved, while the cases of those declared as not being religious ask some comments. In this work the effects of signals were not discussed. The source of signals can be of different characters, but possible explanation of their effects could be similar. One of explanation possibilities is to suppose that signals activate some biochemical mechanisms capable to repair the organism normal behaviour. Applied drug dose titration In psychotherapy, depending on particular patient, one or more different drugs can be applied. In cases when applied drug reliably showed adequate positive effects with absence or with minimal negative side effects the estimation and maintenance of optimal drug dosing do not represent problem, if necessary information on drug properties is available. More complicate situation appears when the quite adequate drug for particular patient is not available, or when patient behaviour asks the searching of adequacy of known and available different drugs. Then it can happen that none of available known drugs is quite adequate and that the combination of different available drugs should be applied. If the most

Annual 2010/2011 of the Croatian Academy of Engineering

153

convenient of available drugs is chosen for starting treatment is applied on appropriate way, then it is recommendable to monitor the consequences to decide how to continue, taking into account drug properties referring to drug half-life time and resorption rate as well as character and intensity of observed consequences. The observed side effects could suggest additional application of another drug suitable to suppress observed negative effects. If such a drug can induce side effects reducing the positive effects of applied starting drug, then careful titration become obligate. If this would not be done adequately, then it can happen that due to overdosing instead of health improvement the health aggravation could result. Therefore, in addition to commonly applied trial and error methods in optimizing chemotherapeutic treatment the computer simulation of applied drugs pharmacokinetics and supplying some necessary experimental data could be recommended in order to optimise therapy. Information on the author: Prof. Dr. sc. Marijan Bošnjak, retired; member emeritus of Croatian Academy of Engineering (HATZ), HR-10000 Zagreb, Croatia; E-mail: [email protected] Home address: HR-10000 Zagreb, Slovenska 19, Croatia Tel. (385)-1-3702507

154

Annual 2010/2011 of the Croatian Academy of Engineering

DICES: Distributed Component-based Embedded Software Systems Mario Žagar Faculty of Electrical Engineering and Computing University of Zagreb, Zagreb, Croatia [email protected]

Ivica Crnković Mälardalen Real-Time Research Centre Mälerdalen University, Västerås, Sweden [email protected]

Darko Stipaničev Faculty of Electrical Engineering, Mechanical Engineering and Naval Architecture University of Split, Split, Croatia [email protected]

Maja Štula Faculty of Electrical Engineering, Mechanical Engineering and Naval Architecture University of Split, Split, Croatia [email protected]

Juraj Feljan Mälardalen Real-Time Research Centre Mälerdalen University, Västerås, Sweden [email protected]

Luka Lednicki Faculty of Electrical Engineering and Computing University of Zagreb, Zagreb, Croatia [email protected]

Josip Maras Faculty of Electrical Engineering, Mechanical Engineering and Naval Architecture University of Split, Split, Croatia [email protected]

Ana Petričić Faculty of Electrical Engineering and Computing University of Zagreb, Zagreb, Croatia [email protected]

Abstract This article gives a short overview of the contribution of DICES project. The goal of the project is to advance theories and technologies used in development of distributed embedded systems. Three examples of the contributions are presented: a) reverse engineering of webbased applications, design extraction and extraction of reusable user-interface controls,

Annual 2010/2011 of the Croatian Academy of Engineering

155

b) a fretwork for building systems that use UPnP devices and treat them as components in the same way as software components, and c) PRIDE – development environment for designing, modeling and developing embedded systems, based on ProCom technology. Keywords: Software components, embedded systems, web-based applications, reverse engineering 1 Introduction DICES (Distributed Component-based Embedded Software Systems) is a project funded by UKF (Unity Through Knowledge Fund) with contributions from Faculty of Electrical Engineering and Computing (FER) - University of Zagreb, Faculty of Electrical Engineering, Mechanical Engineering and Naval Architecture (FESB), University of Split, School of Innovation, Design and Engineering - Mälardalen University (MDU) in Sweden, and Ericsson Nikola Tesla. The project was performed during three years, from 2007 to 2011. The DICES project goal is to advance development of distributed embedded software systems with emphasis on software reusability and predictability of software quality. The aim of the project is increasing the software development ef¿ciency and quality by applying service-oriented and component-based approaches. The overall presence of distributed embedded systems in the modern society is a fact. Examples of such systems are telecommunication systems, grid systems, control and information systems of vehicular systems (cars, trains), environmental monitoring systems. Embedded systems development is one of the strategic research areas of EU-FP7 programmes. It is also of signi¿cant importance in Croatia, since many leading companies in Croatia either produce such systems (e.g. Konþar, Ericsson Nikola Tesla) or use such systems (e.g. Pliva, or many small companies). DICES is addressing ef¿cient reusability of software components and prediction of the important properties for embedded systems: resource utilization, and performance, by applying the service-oriented software engineering and component-based software engineering methods and technologies. The main contribution of DICES is provided in two directions: a) analysis of legacy webbased systems, extraction of its design and its automatic modeling, and extraction of user interface controls and packaging them as reusable units, and development of componentbased technology appropriate for design of embedded systems. The reverse engineering, architecture recovery and extraction of reusable UI units is applied on “iForestFire - Intelligent Forest Fire System” system developed at FESB Split, which enables thorough validation of the approach and provides input for further development of this system and possible commercialization of the improved product. Development of component-based technology is done in cooperation with Swedish Strategic Research Centre, PROGRESS that has developed theories and methods for modeling component-based embedded systems. DICES contribution was in development of PRIDE, PROGRESS Integrated Development Environment, a tool modeling component-based embedded systems, simulation and analysis of resource utilization. In this article we give a short overview of these contributions as follows. In section 2 we describe Componentizing Web Information Systems, including phpModeller, a tool for design extraction and its presentation in UML. Section 3 shows a component model that includes both, software and hardware components based on UPnP devices and enables a unique treatment for software and hardware components. Section 4 describes PRIDE.

156

Annual 2010/2011 of the Croatian Academy of Engineering

2 Analysis and Componenitization of Web-based systems Web-based systems are systems that use the Internet as its core infrastructure and are accessible from anywhere in the world via the Web. Often these systems extensively communicate with databases and third-party information sources. In some cases they even communicate with embedded devices such as cameras or meteo-stations in order to provide real-time information about the environment. One of these systems is the iForestFire system developed at the University of Split, Croatia. iForestFire is an intelligent and integrated video based monitoring system for early detection of forest ¿re. The main idea is that forest ¿res are detected in incipient stage using advanced image processing and image analyses methods. The system is based on ¿eld units and central processing unit. The ¿eld units are embedded devices and include day & night, pan/tilt/zoom controlled IP based video cameras and IP based mini meteorological stations connected by wired or wireless LAN to the central processing unit where all analysis, calculation, presentation, image and data archiving is done. In the scope of the DICES project, iForestFire was chosen as a case study application in order to test the ideas of componentization of Web Information Systems that extensively communicate with embedded devices. Since one of the goals of the DICES project was the analysis of legacy web-based systems, extraction of its design and its automatic modeling, extraction of User Interface controls and packaging them as reusable units in this section, we will describe two approaches that we have developed in order to achieve those goals: phpModeler – an approach for reverse engineering of legacy web applications, and Firecrow, an approach for extracting reusable Web User Interface controls. 2.1 PhpModeler – Reverse Engineering Legacy Web Applications

In order to componentize web information systems, ¿rst a correct assessment of the current state is necessary. Web information systems, especially ones that are communicating with complex embedded devices can be hard to understand because a single behavior can be realized by code executed on a large number of nodes (embedded devices, web servers, clients, etc.). Also, there usually exist many inter-dependencies between certain parts of the system that are hard to track and manage manually, and naturally the complexity of these inter-dependencies grows with the size of the web application. In order to tackle this problem and to improve development and maintenance ef¿ciency, these applications need to be modeled on a high level of abstraction. One of the ways to achieve this goal is to use a process called Reverse Engineering (RE). Reverse Engineering (RE) is used for information extraction from source artifacts (primarily source code) and their transformation to easily understandable abstract representations (e.g. standard UML diagrams). Even though there exists a certain number of already available tools for reverse engineering Web Information Systems (WARE [6], WebUml [3], ReWeb [10], etc.), none of them have the following functionalities: • Static analysis of web application source code which as an end result produces UML diagrams that can be used for architecture recovery • Generates dependency models that show inter-dependence between web application elements • Visualizes Web application evolution. In order to tackle this problem we have developed a plugin for the Eclipse IDE (Integrated Development Environment) that facilitates modeling and web application architecture recovery called phpModeler. It has three main features: page modeling, dependency modeling and model comparison.

Annual 2010/2011 of the Croatian Academy of Engineering

157

In order to be able to build page UML models, the information important for those models has to be extracted. phpModeler gathers this information by parsing PHP, HTML, and JavaScript code responsible for the behavior of the web application. Once this information has been collected further analysis techniques can be used to establish dependencies between parts of the system. For every web page entity (JavaScript library, database table, ¿le, PHP library, web page), the analysis ¿nds all other entities dependent on it. For example, for every database table this means locating all web pages that access them and for function libraries all web pages that use those functions. Naturally, all information gathered in the analysis process can be generated to UML diagrams.

Figure 1: Reverse Engineering Web applications with phpModeler.

Web information systems evolve with the addition of new functionalities. This usually means that the existing code is modi¿ed to support the new user requirements. At different steps of the WIS evolution the system is represented with different models, so phpModeler provides a way to track the differences between two existing models. This gives a simpler, and more visual representation of the evolution of the target system. The whole process is shown in Figure 1. We used phpModeler in the reverse engineering process of the iForestFire system in order to recover the architecture and gain better understanding of the system. 2.2 Firecrow – reusing Web User Interface controls

Important part of the web application code is the code that is used to realize the userinterface of the web application. Web application user-interface (UI) is often composed of distinctive UI elements, the so called UI controls. Similar controls are often used in different parts of web applications (and even in different web applications) and facilitating their reuse could lead to faster development. Unfortunately, preparing code for reuse is a slow process which is often not a priority. Often, when developers encounter problems that have already been solved in the past, rather than re-inventing the wheel, or spending time componentizing, they reuse the code that is already functioning in another context [4]. Reuse tasks are complex and prone to errors, primarily because it is dif¿cult to establish the minimum amount of code responsible

158

Annual 2010/2011 of the Croatian Academy of Engineering

for implementing the desired behavior [8]. In the web application domain, reuse is especially dif¿cult: there is no trivial mapping between source code and the page displayed in the browser, responsible code is often scattered between several ¿les and is found in between code that is not important from the reuse perspective. Next, the developer has to locate and download the necessary resources, source code, and adjust for the changes so that the component can be easily integrated into an already existing system. The structure of a web page is de¿ned by HTML code, the presentation by CSS (Cascading Style Sheets) code, and the behavior by JavaScript code. In addition, a web page usually contains various resources such as images or fonts. The interplay of these four basic elements produces the end result displayed in the user’s web browser. Visually and behaviorally a web page can be viewed as a collection of UI controls, where each control is de¿ned by a combination of HTML, CSS, JavaScript and resources (images, videos, fonts, etc.) that are intermixed with code and resources de¿ning other parts of the web page. In order to reuse a web UI control, we have to extract all that is necessary for the control to be visually and functionally autonomous. This means extracting all HTML, CSS, JavaScript and resources that are used in the visual presentation and the desired behavior of the control. The process can be separated into three phases: 1) Interaction recording, 2) Resource extraction, and 3) UI control reuse (Figure 2).

Figure 2: Extracting Web User-Interface controls.

The ¿rst step of the Interaction recording phase is to select the HTML node that de¿nes the chosen UI control. Next, the user performs a series of interactions that represent the behavior of the control. The purpose of this phase is to gather a log of all resources required for replicating visual and behavioral aspects of the control. This is done by logging all executed code, all CSS styles and all resources used in the lifecycle of the control.

Annual 2010/2011 of the Croatian Academy of Engineering

159

When the user chooses to end the recording, the process enters the Resource extraction phase, where code models for all code ¿les (HTML, CSS, and JavaScript) are built. Based on those models and logs gathered during the recording phase, the code necessary for replicating the visuals and the demonstrated behavior is extracted. After the extraction phase is completed, the user can enter the Reuse phase and automatically integrate the extracted control in an already existing web page, either by replacing, or by embedding it inside an existing node. With this, a full cycle is completed: from seeing the potential for reuse, through control extraction, all the way to actual reuse and gaining new functionalities on the target web application. The whole process is currently supported by the Firecrow tool, which is an extension for the Firebug web debugger. Currently, the tool can be used from the Firefox web browser, but it can be ported to any other web browser that provides communication with a JavaScript debugger, and a DOM explorer (e.g. IE, Chrome, Opera). The interaction recording phase is the only part of the process that is browser dependent; building source models, extracting code, downloading resources, merging code and resources are all functionalities that are all encapsulated in a library that can be called from any browser on any operating system. The whole source of the program can be downloaded from [18]. 3 Hardware and Software Components The main purpose of embedded systems is to monitor or control processes in their environment. This interaction of software with the real world requires integration of software components with hardware components such as sensors and actuators. However, coping with these two different parts of embedded systems simultaneously is still a challenge. Current component models for embedded systems mostly focus on software components and rarely try to provide extensive support hardware component [7]. In DICES project we have also addressed this aspect of component-based software development. One of the results of our research is UComp component model and supporting technology [9]. 3.1 UComp – component-based development for hardware and software components

The focus of UComp is on distributed systems whose functionality is implemented using various devices connected to a computer network. These devices may either be physical, i.e. realized using hardware, or virtual, i.e. realized using software applications. Main goal of UComp is to utilise the component-based approach and manage the hardware and software components in a uniform way. Further, our goal is to apply the component-based approach during whole life-cycle of a system, including run-time phase. By having the components available at run-time, systems can be extremely Àexible because any modi¿cations can be done while the system is running and the embedded devices are deployed. In addition, this would allow easier late deployment of new devices (i.e. components) or replacement of existing ones during run-time. 3.1.1 The UComp component model

UComp distinguishes between two types of components: Device components and software components. Device components are used to communicate with hardware or virtual network devices and provide their functionalities in a component-based manner. Functionality of software components is fully implemented in program code, and these components are not associated with any devices. They are used for additional computations in order to avoid the need for “glue code”.

160

Annual 2010/2011 of the Croatian Academy of Engineering

Component interfaces consist of input and output ports. Ports can be considered as component access points, used for a component to exchange data and control (triggering) signals. System execution follows the pipes and ¿lter pattern. Data and triggering signals from output port of one component can be directed to input ports of one or more components. When an input port receives a signal from the output port it is connected to, it becomes active. A component can be con¿gured to start it’s execution when either all or just selected ports are activated. This is done by selecting activation types for ports. There are three activation types for input ports: • Trigger. A component is activated if all input ports with activation type set to trigger are active. • Priority trigger. A component is activated if any of its input port with activation type set to priority trigger is active. • Data. If port’s activation type is set to data, it is only used to receive data, and does not affect the triggering of the component. The graphical representation of output ports and all input port types can be seen in Figure 3. The ¿gure shows an instance of Component A having input ports a (data port), b (trigger port) and c (priority trigger port), and an output port out.

Figure 3: Graphical representation of an UComp component and its input and output ports. 3.1.2 Device Components

Device components are the base for accomplishing the uniform treatment of components of a system, which is not dependant on component realization (hardware or software). They represent hardware (physical) and virtual (realized using software applications) network devices. Device components, together with their input and output ports, are automatically generated by the UComp framework using device descriptions. Automatic generation of devices and their ports eliminates need for specialized drivers or manual con¿guration of such components by the developer of the system. Every device component signals if the actual device is available on the network. This information can be very useful in distributed systems where connection between components is not reliable and some of them can be temporarily unavailable. Device components can represent actions (synchronous request-response communication) or events (asynchronous publish-subscribe messaging) of network devices. Therefore, we have de¿ned two types of device components: action components and event components. Action Components. Action components are designed to wrap around synchronous action invocations or data queries of a device connected to the network. When an action component is triggered, action invocation is performed, using values of its input ports as arguments. When the invocation ¿nishes, results of the invocation are

Annual 2010/2011 of the Croatian Academy of Engineering

161

propagated to output ports of the components. Event Components. Event components allow receiving asynchronous messages from devices. These messages may signal data changes or other events that device may provide. Interfaces of event components have only output ports, which obtain their values from event noti¿cations provided by the network device. 3.1.3 Software Components

Functionality of software components is fully implemented by program code, and they are not bound to any hardware elements. They are used to process the data received from, or sent to, device components or manipulate the execution of components. Their function can vary from very simple (for example addition of two numbers) to complex data processing. Execution Semantics Initially, all components in the system are in an idle state waiting to be activated for execution. Activation can be caused either by the triggering signals received at the input ports of the component, or by its internal events. Action components and most software components are passive, meaning that they execute only when they are triggered by signals received from other components, while event components and some software components are active and thus may start their execution by an internal event. Realisation of UComp Component Model The UComp architecture, shown in Figure 4, is realised as Java application that implements the Universal Plug and Play (UpnP) [16] technology to control devices available on the network, process their data, and relay data between them. The application communicates with devices through a single UPnP control point implemented using CyberLink UPnP stack [11]. Centralized architecture allows us to process data received from a device by UComp application before it is forwarded to other devices, making the system much more Àexible and eliminating the need to change the code of devices to adapt them to the needs of the developed system. Also, run-time modi¿cation of systems is much easier. System’s behaviour can be modi¿ed by simple changes in the interconnection of components (or by changing the components themselves) in the central application. 4 Development environment

Figure 4: The UComp architecture.

162

Annual 2010/2011 of the Croatian Academy of Engineering

To facilitate the development we have created a tool for visual development of UComp systems - UComp Developer. The UComp Developer enables browsing available device and software components using a tree structure, visual representation of components on a development panel, modifying connections between them, setting their properties and the properties of their ports, and starting and stopping the execution of the developed system. Systems developed with this tool are saved or restored from XML ¿les. A screen-shot of UComp Developer is shown in Figure 5.

Figure 5: Screenshot of UComp Developer tool.

5 An overview of the PRIDE tool PROGRESS-IDE (PRIDE) is an integrated development environment used for software development of distributed embedded systems (ES), primarily in the automotive, automation and telecom domains. The grand vision of Progress-IDE is to cover the whole software development cycle of distributed embedded systems, from early system design to deployment and synthesis. The development process of ES requires a strong emphasis on analysis, veri¿cation and validation in order to ensure the necessary quality of the delivered product, therefore PRIDE integrates various analysis tools. Compared to the majority of existing IDEs that focus mainly on the programming aspect, PRIDE uses the notion of component as a ¿rst-class concept allowing the manipulation and modelling of components, with reusability as one of the key concerns and supporting: • Component and system design using the ProCom component model [13], • System analysis (worst-case execution time, model checking of behavioural models and fault propagation and transformation calculus), • Deployment modelling • Code synthesis.

Annual 2010/2011 of the Croatian Academy of Engineering

163

In the following list we give some overall reÀections that apply to Progress-IDE. • Progress-IDE covers the whole system development process, • Progress-IDE is developed as a stand-alone application on top of Eclipse RCP, • Progress-IDE consists of a number of editors: architectural editors, attribute editor, node description (virtual/physical), allocation editor, code generation, various analysis editors, • Progress-IDE is user-friendly by providing a user interface respecting the usability features, partial auto-completion features, default values etc. 5.1 The PRIDE approach

In the recent years ES development has changed signi¿cantly due to the rapid increase of software in these systems and software becoming as complex as in conventional systems. In non-embedded domains, new approaches such as model-based, component-based, and service-oriented development have been proposed to manage software complexity and there is a trend to apply these approaches also in ES development. ES correctness is strongly correlated to speci¿c extra-functional properties (EFPs) such as timing (e.g. execution and response time) or dependability (e.g. reliability and safety) under constrained resources such as memory, energy, or computation speed. This calls for additional domain-speci¿c technologies that provide support not only for functional development but also for analysis and veri¿cation of EFPs. As a possible solution, we have been developing a new component-based approach [5] built around a two-layer component model - ProCom, which addresses the particularity of ES development from big complex functionalities, to small, close to control loop functionalities. This approach requires speci¿c tool support that enables: Ef¿cient system design by using existing components, Seamless integration of different tools to provide the analysis and veri¿cation required for system correctness, and Ef¿cient EFP management of components and systems. PRIDE uses reusable software components as the central development units, and as a means to support and aggregate various analysis and veri¿cation techniques. In difference to similar approaches [1, 2, 17], PRIDE puts emphasis on EFPs during the entire lifecycle - from early speci¿cation to deployment and synthesis. PRIDE has been designed to support four design strategies that are especially important to consider for having an ef¿cient component-based development of ES. Levels of abstraction Using components throughout the whole development process implies that the component concept spans a wide range of abstractions, from a vague and incomplete early speci¿cation, to very “concrete” with a ¿xed speci¿cation, a corresponding implementation and information about their EFPs. This means that components at different levels of abstraction must be able to co-exist within the same model. Component granularity In distributed ES, components span a large variety in size and complexity; the larger components are typically active (i.e. with their own thread of execution) with an asynchronous message passing communication style, whereas the smaller components are responsible for a part of control functionality with a strong synchronization. For an ef¿cient development, a support for handling different types of components must be provided. Component vs. system development The common distinction between component development and system development brings

164

Annual 2010/2011 of the Croatian Academy of Engineering

issues in ES development, where the coupling between the hardware platform and the software is particularly tight. As a consequence, component development needs some knowledge of where the components are to be deployed. This requires support to handle the coupling between components, system and target platform, while still allowing separate development of components and systems. Extra-functional properties The correctness of ES is ascertained based on both the functional and extra-functional aspects. However, many EFPs typically encountered in ES are assessed through different methods during the development lifecycle (from early estimation to precise measurements) and different values may be obtained according to the characteristics of the resources of the platform on which the components are to be deployed. For this reason, an ef¿cient ES development should provide a means for specifying, managing and verifying these properties with respect to the context in which their values are provided. To comply with the design strategies, the following requirements have been identi¿ed as principles that guided the design and development of PRIDE: Allowing to move freely between any development stages, Displaying the consequences of a change in the system or within a component, Supporting the coupling with the hardware platform, and Enabling and enforcing the analysis, validation and veri¿cation steps. In addition, a central requirement relates to the notion of component. Components are the main units of development and seen as rich-design artifacts that exist throughout the whole development lifecycle, from early design stage, in which little information about them exists, to deployment and synthesis stages, in which they are fully implemented. PRIDE views a component as a collection of all the development artefacts (requirements, models, EFPs, documentation, tests, source code, etc.), and enables their manipulation in a uniform way. Driven by the aforementioned principles, several tools have been developed and tightly integrated into PRIDE. Since the PRIDE is built as an Eclipse RCP application, it can be easily extended with addition of new plugins. 5.2 Implementation of PRIDE

The Eclipse platform is chosen as the supporting architecture for PRIDE. In order to reduce the overhead which exists when using Eclipse directly as the main integration environment, it has been decided that the PRIDE will be developed as a stand-alone application built on top of the Eclipse Rich Client Platform (Eclipse RCP). This decision is driven by the will to keep the control over the features present in the environment such as, for example, avoiding the presence of menus not directly related to the particular use of the PRIDE. When using Eclipse RCP, only the features explicitly added to the environment are present. In other words, the layout and function of the environment are fully controlled by the plugin developers. The component architecture design part of the environment was implemented using EMF [14] and GMF [15]. EMF is a modelling framework and code generation facility for building tools based on a structured data model. It offers a graphical editor for describing metamodels. With the aid of this, the ProCom metamodel (model of the ProCom model) was de¿ned. From this graphical description of the metamodel, EMF automatically generates Java classes representing the model, classes used for modifying the model and textual tree editors for the model. GEF then takes this existing application model (i.e. the model generated using EMF) and quickly creates a rich graphical editor. However, this automatically generated editor has rudimentary functionality and needs to be further

Annual 2010/2011 of the Croatian Academy of Engineering

165

manually modi¿ed and tweaked in order to achieve the desired functionality. Software architecture modelling and analysis As already mentioned, PRIDE uses component-based approach which allows components to be developed independently and reused in different contexts. Reusability is one of the main concepts in PRIDE aiming to signi¿cantly shorten development time. The tool makes a distinction between a component type and a component instance. Each reusage of a component creates a component instance of the given component type. By editing the component all of its instances are affected. In PRIDE components are rich design entities encapsulating a collection of development artefacts; requirements, various models (e.g. architectural, behavioural, resource usage etc.), EFPs, documentation, tests and source code. Figure 6 shows a screenshot from PRIDE with main parts highlighted. PRIDE’s modelling part consists of a component explorer and component editors.

Figure 6: A screenshot from PRIDE showing a) the component explorer; b) the component editor; c) the code editor; d) the repository browser; and e) the attribute framework.

Component Explorer Enables browsing the list of the components available in the current development project. In it a component owns a prede¿ned and extensible information structure corresponding to a rich component concept. It also provides a support for component versioning and importing and exporting from a project to a component repository making them available to other projects thus facilitating the component reusability. Component Editors Although the ProCom component model distinguishes between two different types of components (ProSave and ProSys), in the component editors all components are treated in a uniform way. Component editors provide two independent views on a component, external and internal view. The external view manages the component speci¿cation and interface such as the information about the component name, its interface (services and ports) and EFPs. The internal view focus on component’s internal structure i.e. its realization and implementation and it depends in the component realization type. For primitive components, the internal view is linked to the component implementation and

166

Annual 2010/2011 of the Croatian Academy of Engineering

the source code is displayed. For composite components, the internal view corresponds to an interconnection of subcomponent instances and a graphical form is made available allowing to make modi¿cations in this inner structure (addition/deletion of component instances, connectors, change in the connections, etc.). The analysis support in PRIDE is based on two main parts, an attribute framework and an analysis framework. Attribute Framework The purpose of the attribute framework [12] is to provide a uniform and user-friendly structure to seamlessly manage EFPs in a systematic way. The attribute framework enables attaching extra-functional properties to any architectural element such as a speci¿c port, service or the component as a whole. Attributes are de¿ned by attribute types, and include attribute values with metadata and the speci¿cation of the conditions under which the attribute value is valid. One key feature is that the attribute framework allows an attribute to be given additional values during the development without replacing old values. This allows de¿ning of early estimates for EFPs even before the component has been implemented and use it for analysis in early stages of system development. New, userde¿ned attribute types can also be added to the model. Analysis framework The analysis framework provides a common platform for integrating in a consistent way various analysis techniques, ranging from simple constraint checking and attribute derivation (e.g., propagating port type information over connections), to complex external analysis tools. Analysis results can either be presented to the user directly, or stored as component attributes. They are also added to a common analysis result log, allowing the user easy access to earlier analysis results. PRIDE also allows to easily integrate new analysis techniques together with their associated EFPs. 6 Conclusion In this paper we gave an overview of some of the results of DICES project. The project focus was on distributed embedded systems and the overview shows that a large variety of technologies and approaches are needed for an ef¿cient development and maintenance of embedded systems. The project is focused on several phases in the lifecycle: a) reverse engineering and analysis of the existing legacy code that helps in reusability in development of new systems from existing components; b) modelling and designing embedded systems that include both software and hardware components, and ability to perform different types of analysis important for embedded systems (e.g. resources utilization, performance, timing characteristics). The work done shows that the embedded systems of today are complex system, in the ¿rst hand with complex software that is approaching in size and complexity software from other domains. Acknowledgment This work was supported by the Swedish Foundation for Strategic Research via the strategic research centre Progress, Croatian Ministry of science, education and sports via the research project Software engineering in ubiquitous computing, and the Unity Through Knowledge Fund via the Dices project.

Annual 2010/2011 of the Croatian Academy of Engineering

167

References 1. 2. 3. 4. 5. 6.

7. 8. 9.

10. 11. 12.

13.

14. 15. 16. 17. 18.

Akerholm M., Carlson J., Fredriksson J., Hansson H., Hakansson J., Möller A., Pettersson P., Tivoli M. (2007) The SAVE Approach to Component-Based Development of Vehicular Systems, In: Journal of Systems and Software, Elsevier Science, pp. 655-667 Articus Systems, Rubus Software Components, http://www.arcticus-systems.com , Accessed 7 March 2011 Bellettini C., Marchetto A., Trentini A. (2004) WebUml: reverse engineering of web applications, In: Symposium on Applied computing, ACM New York, NY, USA, 16621669 Brandt J., Guo J.P., Lewenstein J., Klemmer S.R. (2008) Opportunistic programming: How rapid ideation and prototyping occur in practice, In: Workshop on End-user software engineering, ACM New York, NY, USA, pp. 1-5 Bureš T., Carlson J., Sentilles S., Vulgarakis A. (2008) A Component Model Family for Vehicular Embedded Systems, In: The Third International Conference on Software Engineering Advances, IEEE Computer Society Washington, DC, USA, pp. 437-444 Di Lucca G., Fasolino A.r., Tramontana P. (2004) Reverse engineering Web applications: the WARE approach, In: Journal of Software Maintenance and Evolution: Research and Practice - Special issue: Web site evolution Volume 16 Issue 1-2, John Wiley & Sons, Inc. New York, NY, USA, pp. 71-101 Feljan J., Lednicki L., Maras J., Petricic A., Crnkoviü I. (2009) Classi¿cation and survey of component models, MRTC report ISSN 1404-3041 ISRN MDH-MRTC-242/2009-1-SE, Mälardalen Real-Time Research Centre, Mälardalen University Holmes R.(2008) Pragmatic Software Reuse, Ph.D. Dissertation, University of Calgary Lednicki, L. and Petricic, A. and Zagar, M. (2009) A Component-Based Technology for Hardware and Software Components, In: Software Engineering and Advanced Applications, 2009. SEAA ‘09. 35th Euromicro Conference on, IEEE Computer Society Washington, DC, USA, pp. 450 -453 Ricca F., Tonella P. (2001) Understanding and Restructuring Web Sites with ReWeb, IEEE Multimedia, Vol 8, Issue 2, pp. 40-51 Satoshi Konno, CyberLink for Java, http://www.cybergarage.org/twiki/bin/view/Main/ CyberLinkForJava , Accessed 7 March 2011 Sentilles S., ŠtČpán P., Carlson J., Crnkoviü I. (2009) Integration of Extra-Functional Properties in Component Models, In: Proceedings of the 12th International Symposium on Component-Based Software Engineering (CBSE ‘09), Springer-Verlag Berlin, Heidelberg, pp. 173-190 Sentilles S., Vulgarakis A., Bureš T., Carlson J., Crnkoviü I. (2008) A Component Model for Control-Intensive Distributed Embedded Systems, In: Proceedings of the 11th International Symposium on Component-Based Software Engineering (CBSE ‘08), Springer-Verlag Berlin, Heidelberg, pp. 310-317 The Eclipse Foundation, Eclipse Modeling Framework Project, http://www.eclipse.org/ emf/ , Accessed 7 March 2011 The Eclipse Fundation, Graphical Modeling Framework, http://http://www.eclipse.org/ gmf , Accessed 7 March 2011 UPnP forum, Universal Plug and Play, http://www.upnp.org , Accessed 7 March 2011 van Ommering R., van der Linden F., Kramer J., Magee J. (2000) The Koala component model for consumer electronics software, In: IEEE Computer, , IEEE Computer Society Washington, DC, USA, pp. 78-85 Maras J., Štula M., Carlson J, Crnkoviü I. (2011). Firecrow web page, http://www.fesb. hr/~jomaras/?id=Firecrow, Accessed 7 March 2011

168

Annual 2010/2011 of the Croatian Academy of Engineering

Framework for 3D Motion Field Estimation and Reconstruction Martin Žagar, Hrvoje Mlinarić, Josip Knezović Department of Control and Computer Engineering, Faculty of Electrical Engineering and Computing, Unska 3, HR-10000 Zagreb, Croatia

Abstract: We propose the framework for the motion estimation of 3D objects based on the motion vectors that form motion ¿elds and the motion reconstruction based on the 3D rotations and the factorization method. The classical two-dimensional motion ¿eld approach is extended to three dimensions, e.g. on the volumetric objects that are moving in time. When dealing with the real world multiple moving objects and the complex scenes, lots of objects are moving with different motions, both in space and in time. In this context, an object can be described as a part of a scene that moves with a coherent motion and the scene can be broken down into a number of regions, each of which can be well approximated by its own motion. The ¿rst part of this paper describes the motion vectors based techniques which are used for the motion estimation, and the second part addresses problems with the reconstruction of the 3D motion and structure. Proposed methods estimate reconstruction from a sparse set of the matched volume features on the 3D neuroimage in NIfTI format. We evaluate these techniques according to the different problems they address. Keywords: 3D motion estimation, differential techniques, matching techniques, motion and structure reconstruction, factorization method. 1 INTRODUCTION The motion estimation is used to eliminate a large amount of temporal and spatial redundancy that exists in sequences of 3D data. Most video encoders perform the motion estimation in video sequences to search for the motion information in sequential video frames that can improve compression. The difference between the current frame and the predicted frame in motion estimation and motion reconstruction (based on previous and/or future frames) is coded and transmitted. The better the prediction is, the smaller the error and hence the transmission bit rates are. If a scene is still, then a good prediction for the particular object in the current frame is the same as object in the previous frame and the error is close to zero. However, when there is a motion in a sequence, an estimation of the predicted position of the object in the current frame can be made with the motion vectors [7], [10]. Many techniques based on the motion vectors that form motion ¿elds can be roughly divided into two major classes: differential techniques and feature-based matching techniques [9]. The differential techniques are based on the spatial and temporal variations of the volume brightness at all voxels, and can be regarded as methods for computing optical Àow. The matching techniques, instead, estimate the disparity of special volume points (features) between frames. Proposed methods are evaluated on 3D neuroimages in NIfTI (Neuroimaging Informatics Technology Initiative) format which is designed to

Annual 2010/2011 of the Croatian Academy of Engineering

169

maximize the usability of neuroimaging computing tools and provide a common resource for researches in magnetic resonance imaging (MRI). It is used for the detailed visualization of internal structure and functions of the human body. The main purpose of determining motion in MRI is to reduce a vast amount of acquired data in coding process. For different types of MRI, different motion techniques should be used for motion estimation and motion reconstruction. Estimation of 3D motion ¿eld is based on the motion vectors and is used to predict and to describe motion between the current and the next frame. With coding only motion vectors instead of whole dataset, higher compression ratio can be achieved. On the other hand, in the motion reconstruction, 3D motion is reconstructed from a sparse set of matched volume features. The paper is organized as follows. Our proposed differential and matching techniques are described in Section 2. We adapted motion vectors based techniques as one of appropriate techniques for motion estimation and factorization method for 3D motion reconstruction. Factorization method is described in Section 3.2, while the beginning of Section 3 presents our 3D edge detection method for segmentation, based on Canny 2D edge detection [11]. The segmentation is necessary when dealing with multiple moving objects to detect and to segment objects that are moving with different motions in space in time. Experimental results of proposed methods and techniques are presented in Section 4 and concluded in Section 5. 2 MOTION VECTORS BASED TECHNIQUES 2.1 Differential Techniques based on Optical Flow

A large number of differential techniques for computing the optical Àow have been proposed [1], [3], [8]. Some of them require the solution of a system of partial differential equations, others the computation of second and higher-order derivatives of the volume brightness and some of them require the least-squares estimates of the parameters characterizing the optical Àow. Methods in the latter class have at least two advantages over those in the ¿rst two: • They are not iterative; therefore they are genuinely local and less biased by possible discontinuities of the motion ¿eld than iterative methods. • They do not involve derivatives of order higher than the ¿rst; therefore, they are less sensitive to noise than methods requiring higher-order derivatives. 2.1.1 Optical flow

Our proposed differential technique is the extension of 2D motion ¿eld estimation ([3], [11]) to 3D. It is based on the least-squares estimates of the parameters characterizing the optical Àow. The basic assumption is that the motion ¿eld is well approximated by a r constant vector ¿eld velocity v within any small region of the volume plane. The optical Àow is the approximation of the motion ¿eld which can be computed from time-varying volume sequences. For each point pi within a small patch Q of size N x N x N, it can be written (∇E )T vr + ∂E = 0 (1) ∂t where E = E(x,y,z,t) is volumetric brightness for voxel at point p(x,y,z) at time t. The spatial and the temporal derivatives of the volumetric brightness are computed at points (voxels)

170

Annual 2010/2011 of the Croatian Academy of Engineering

p1 , p2 , K , p N 3 ∈ Q r . Therefore, the optical Àow can be estimated within Q as the constant vector, v that minimizes the functional

& \ >v @

wE º ª T & ¦ «¬’E v  wt »¼ . pi Q 2

(2)

The solution of the least squares problem can be found by solving the linear system AT Av = AT b [3]. The i-th row of the N3 x 3 matrix A is the spatial gradient evaluated at point pi:

⎤ ⎡∇E ( p1 ) ⎢∇E ( p ) ⎥ 2 ⎥ A=⎢ ⎥ ⎢M ⎥ ⎢ ⎣∇E ( p NxNxN )⎦

(3)

and b is the N3-dimensional vector of partial temporal derivatives of the volumetric brightness, evaluated at p1 , p 2 , K , p N 3 after a sign change:

∂E ⎡ ∂E ( p NxNxN )⎤⎥ b = − ⎢ ( p1 ), K , ∂ t ∂ t ⎣ ⎦

T

.

(4)

The least squares solution of the constrained system can be obtained as −1 r v = (AT A) AT b .

(5)

r is the optical Àow (the estimate of the motion ¿eld) at the centre of patch Q. By ψ [v ] repeating this procedure for all cubic patches, a dense optical Àow is obtained. This algorithm can be summarized as follows. The input is a time-varying sequence of n volumes E1, E2, …, En. Let Q be a sub-volume region of N x N x N voxels. 1. 2. 3.

Filter each image of the sequence along each spatial dimension with a Gaussian ¿lter of standard deviation equal to σ s . Filter each image of the sequence along the temporal dimension with a Gaussian ¿lter of standard deviation σ t . For each voxel of each volume of the sequence: (a) Compute the matrix A and the vector b using (3) and (4) (b) Compute the optical Àow using (5)

Values σ s and σ t are set to 1,5 voxels and 1,5 frames respectively, because of the smoothness. The output is the optical Àow computed in the last step. 2.2 Feature-based Matching Techniques

The second class of methods for estimating the motion ¿eld is formed by the so-called matching techniques, which estimate the motion ¿eld at feature points only. The result is a sparse motion ¿eld [7]. One proposed solution is a two-frame analysis (¿nding feature disparities between consecutive volumetric frames). The robustness of frame-to-frame matching can be improved by motion tracking of a feature across a long volumetric

Annual 2010/2011 of the Croatian Academy of Engineering

171

sequence. We will now describe and adapt two different feature-based matching techniques from 2D to 3D: the constant Àow algorithm and the feature tracking tracking. 2.2.1 Constant Flow Algorithm

Features can be de¿ned as centres of those regions for which the smallest eigenvalue of ATA computed over small, cubic regions, is greater than a threshold. The idea of the matching method is simple; the displacement of such feature points can be computed by iterating the constant Àow algorithm [4]. The input is formed by two frames of a volumetric sequence, I1 and I2 and a set of corresponding feature points in the two frames. Let Q1, Q2, and Q’ be three N x N x N cubic regions. Let d be the unknown displacement between I1 and I2 of a feature point p on which Q1 is centred. Feature point p is centre point of the shape segmented and de¿ned by the edge and corner detection described in Section 3.1. The procedure consists of three steps. First, the uniform displacement of the cube region Q2 is estimated through constant Àow, and added to the current displacement estimate (initially set to 0). Second, the patch Q2 is wrapped according to the estimated Àow. This means that Q2 is displaced according to the estimated Àow, and the resulting patch, Q’, is restored in the voxel grid of frame I2. If the estimated Àow normalized in time equals (vx, vy, vz), the grey value at voxel (i, j, k) of Q’ can be obtained from the values of the voxels of Q2 close to (i – vx, j – vy, k – vz). Third, the ¿rst and second steps are iterated until a stopping criterion is met. For all feature points p: 1. Set d = 0 and centre Q1 on p. 2. Estimate the displacement d0 of p, centre of Q1, and let d = d + d0. Let Q’ be the patch obtained by warping Q1 according to d0. Compute S, the sum of the squared differences between the new patch Q’ and the corresponding patch Q2 in the frame I2. If S > τ , where τ is the threshold, set Q1 = Q’ and go to step 1; otherwise exit. The output is an estimate of d for all feature points. In reality, long volumetric sequences rather than just pairs of frames have to be analyzed. The motion of feature points is expected to be continuous, and therefore predictable, in most cases. The disparities computed between frames Ii-1 and Ii-2, Ii-2 and Ii-3 and so on, can be used to make predictions on the disparities between Ii-1 and Ii, before observing frame Ii. This algorithm gives satisfying results only for continuous, predictable motions which can be computed based on disparities between pairs of frames only. To improve the algorithm results for long volumetric sequences we propose to use the feature tracking. 2.2.2 Feature Tracking

The feature tracking solves the problem of matching features from frame to frame in a long sequence of volumes. Our approach is based on [2] and [5] and extended on 3D. First, it is necessary to formalize the tracking problem. A new frame of the volumetric sequence is acquired and processed at each instant t i = t i 0 + kΔT , where k is a natural number and ΔT = 1 is denoted as a sampling interval. Discrete equally-spaced time instants can be indicated with t i assumed that, for simplicity, ΔT = 1 is small enough to capture the system’s dynamics. The state does not change much between consecutive time instants, and a linear system model is an adequate approximation of the state change within ΔT = 1 , i.e. the motion of feature points from frame to frame is considered to be linear. If only one feature point,

pi = [xi , yi , z i ] is considered, in the frame acquired at instant ti, moving T

172

Annual 2010/2011 of the Croatian Academy of Engineering

[

with vector ¿eld velocity vi = v x ,i , v y ,i , v z ,i with the state vector

[

]

T

, the motion on the cubic plane is described

si = xi , yi , zi , v x ,i , v y ,i , v z ,i

]

T

.

si

A physical system is modelled by a state vector , often called simply the state, and a set s of equations, called the system model. The state is a time-dependent vector, is stated as s (t ) , the components of which are system variables, in a suf¿cient number to capture i the dynamic properties of the system. The system model is a vector equation describing the evolution of the state in time. Assuming a suf¿ciently small sampling interval (and therefore constant feature velocity between the frames), the system model can be written as i

pi = pi −1 + vi −1 + ξ i −1

(6)

vi = vi −1 + η i −1

(7)

where ξ i −1 and η i −1 are zero-mean, random, vector-modelling additive system noise. In terms of the state vector si rewrites si = φi −1 si −1 + wi −1 with

φi −1

⎡1 ⎢0 ⎢ ⎢0 =⎢ ⎢0 ⎢0 ⎢ ⎣⎢0

0 0 1 0 0⎤ 1 0 0 1 0⎥⎥ 0 1 0 0 1⎥ ⎥ 0 0 1 0 0⎥ 0 0 0 1 0⎥ ⎥ 0 0 0 0 1⎦⎥

⎡ξ i −1 ⎤ and wi −1 = ⎢ ⎥ . ⎣η i −1 ⎦

(8)

(9)

The difference between the predicted state and the real state is measured with the state correction via the ¿lter gain factor Kt , known from Kalman ¿ltering. Estimation step, a priori estimation, estimation error variance and correction step are calculated based on [6].

3 3D MOTION RECONSTRUCTION When dealing with the real world multiple moving objects and complex scenes, lots of objects are moving with the different motions. In this context, an object can be described as a part of a scene that moves with a coherent motion and the scene can be broken down into a number of regions, each of which can be well approximated by its own motion. Segmenting objects against ¿xed background, based on the edge detection, as a solution of ¿nding the regions of the object corresponding to the different moving objects is presented in Section 3.1. The 3D motion can be reconstructed from a sparse set of matched features. If the average disparity between consecutive frames is small, the reconstruction can gain in stability and robustness from the time integration of long sequences of frames. Moreover, in this case, motion can be de¿ned as a combination of small 3D rotations around coordinate axes and axis and angle. This method is described in Section 3.2. If, on the contrary, the

Annual 2010/2011 of the Croatian Academy of Engineering

173

average disparity between frames is large, this problem can be dealt with by using a stereolike algorithm, applied to a pair of frames. The factorization method as a typical example of stereo-like algorithm is described in Section 3.3. 3.1 Segmentation based on the Edge Detection

Relax the assumption that the object motion is described by a single general 3D motion to deal with the problem of multiple motions. Identifying moving objects can be seen as a problem of detecting and segmenting objects against a ¿xed background. The goal is to ¿nd the regions of the volume corresponding to the different moving objects. This problem can be thought of as a classi¿cation problem. One has to classify the voxels of each frame within objects that are detected. The Canny edge detector is, at the moment, the most widely used edge detection algorithm in multimedia systems. Constructing a Canny detector requires the formulation of a mathematical model of the edges and the corners. The edges of the intensity volumes can be modelled according to their intensity values and pro¿les. The edge detection operator returns a value for the ¿rst derivative in the horizontal and the vertical direction. The depth dimension (the third dimension) can be achieved by modifying a regular Canny 2D detector [11]. The edge gradient of a 3D volumetric object G can be ∂G ∂G determined by computing the magnitude gradient components G x = ∂x , G y = ∂y ∂G G = z and ∂z for each voxel point p(x,y,z) and can be displayed as an volume which intensity levels are proportional to the magnitude of the local intensity changes. The second step is to estimate the edge strength with

ed ( x, y, z ) = G x2 + G y2 + Gz2

as well as the

orientation of the edge normal.

Figure 1 Orientation of edge normal.

As shown in Figure 1, the orientation of the edge normal is speci¿ed by angles ș and ϕ

174

Annual 2010/2011 of the Croatian Academy of Engineering

that can be computed by equations Gy tan θ = Gx

(10)

and

tan ϕ =

Gz

(Gx )2 + (G y )2

.

(11)

Weighted summations of the voxel intensities in local neighbourhoods can be listed as a numerical array. Intensity gradients which are large are more likely to correspond to edges than if they are small. Making the assumption that the important edges should be along continuous curves in the volume allows us to follow a faint section of a given line and to discard a few noisy pixels that do not constitute a line but have produced large gradients. This marks out the edges we can be fairly sure are genuine. Starting from these, using the directional information derived earlier, edges can be traced through the frames. The output of gradient based edge detection is a binary volume indicating where the edges are in order to decide whether an edge has been found. From complementary output from the edge tracing step, the binary edge map obtained in this way can also be treated as a set of edge curves. 3.2 3D Rotations

This section addresses 3D rotations as a base for reconstruction of 3D motions. In most cases it is useful to represent a rotation by means of more natural parameterization. Two parameterizations are proposed: rotations around the coordinate axes and axis and angle. Rotations around the coordinate axes can be expressed as the result of three consecutive rotations around the coordinate axes, e1, e2 and e3, by angles Į, ȕ and Ȗ respectively. The angles are then the three free parameters of linear transformation represented by a 3x3 matrix R and each rotation is expressed as a rotation matrix, Rj, rotating vectors around ej, that is,

0 ⎡1 ⎢ R1 (α ) = 0 cos α ⎢ ⎢⎣0 sin α ⎡ cos β R2 ( β ) = ⎢ 0 ⎢ ⎣⎢− sin β ⎡cos γ R3 (γ ) = ⎢ sin γ ⎢ ⎢⎣ 0

0

⎤ − sin α ⎥ ⎥ cos α ⎥⎦

(12)

0 sin β ⎤ 1 0 ⎥ ⎥ 0 cos β ⎦⎥

(13)

− sin γ cos γ 0

0⎤ 0⎥ ⎥ 1⎥⎦

.

(14)

Annual 2010/2011 of the Croatian Academy of Engineering

175

The matrix R describing the overall rotation is the product of the Rj (the order of overall multiplications matters; different sequences give different results with the same triplet of angles): R = R1 R2 R3 = cos β cos γ ⎡ = ⎢ sin α sin β cos γ + cos α sin γ ⎢ ⎢⎣ − cos α sin β cos γ + sin α sin γ

− cos β sin γ − sin α sin β sin γ + cos α cos γ cos α sin β sin γ + sin α cos γ

sin β ⎤ − sin α cos β ⎥ ⎥ cos α cos γ ⎥⎦

(15)

Second parameterization is as follows. According to Euler’s theorem, any 3D rotation can be described as a rotation by an angle, ș, around an axis identi¿ed by a unit vector r T n = [n1 n2 n3 ] . The corresponding rotation matrix R can then be obtained in terms of ș and the components of n which gives a total of four parameters. The redundancy of this parameterization (four parameters for three degrees of freedom) is eliminated by adding n12 + n22 + n32 [5]. the constraint that n has a unit norm, that is by dividing each ni by The matrix R in terms of ș and n is given by

⎡ n12 n1n2 ⎢ R = I cos θ + (1 − cos θ )⎢n2 n1 n22 ⎢ n3n1 n3n2 ⎣

n1n3 ⎤ ⎡ 0 ⎥ n2 n3 ⎥ + sin θ ⎢ n3 ⎢ n32 ⎥⎦ ⎢⎣− n2

− n3 0 n1

n2 ⎤ − n1 ⎥ ⎥ 0 ⎥⎦

(16)

Conversely, both ș and n can be obtained from the eigenvalues and the eigenvectors of R. The three eigenvalues of R are 1, cos θ + i sin θ and cos θ − i sin θ , where i is the imaginary unit. The unit vector n is proportional to the eigenvector of R, corresponding the eigenvalue 1; the angle ș can be obtained from either of the two complex eigenvalues. 3.3 Factorization Method for Motion Reconstruction

Of the many methods proposed in the literature for the case of large average disparity between frames, the most used is the factorization method, which is simple to implement and gives very good (and numerically stable) results for objects viewed from rather large distances [8]. In this paper it is adopted to three dimensions. The necessary assumption is that the position of n volume points corresponding to the feature points p1, p2, ..., pn, has been tracked in N frames, with N > 3. This is equivalent to acquiring the entire sequence before starting any processing. Tracking method is described in Section 2.2.2. If

pij = [xij , yij , zij ]

T

denote the j-th voxel (j = 1, ..., n) at the i-th frame (i = 1, ..., N),

and think of the xij, yij and zij as entries of three N x n matrices, X, Y and Z, respectively, the measurement matrix 3N x n can be formed as ⎡X ⎤ W = ⎢Y ⎥ . ⎢ ⎥ ⎢⎣ Z ⎥⎦

(17)

176

Annual 2010/2011 of the Croatian Academy of Engineering

For use in the rank theorem, the mean of the entries on the same row is subtracted from each xij, yij and zij: ~ (18) xij = xij − xi ~ (19) yij = yij − yi ~ (20) zij = zij − zi where

xi =

1 n ∑ xij n j =1

(21)

yi =

1 n ∑ yij n j =1

(22)

zi =

1 n ∑ zij n j =1

(23)

are the coordinates of pi , the centroid of the voxels in the i-th frame. Furthermore, ~ xij , ~ yij ~ ~ ~ and ~ zij can be denoted as entries of three N x n matrices, X , Y and Z , and form the 3N x n matrix W, called the registered measurement matrix:

~ ⎡X ⎤ ~ ⎢ ~⎥ W = ⎢Y ⎥ . ⎢ Z~ ⎥ ⎣ ⎦

(24)

The factorization method is based on the proof of a simple but fundamental result that the registered measurement matrix without noise has at most the rank 3 [8]. The proof is based ~ on the decomposition (factorization of W ) into the product of a 3N x 3 matrix, R, and a 3 x n matrix, C. R describes the frame-to-frame rotation with respect to the points pi. C describes the structure of points (coordinates).

Figure 2 Factorization method.

Consider all quantities expressed in an object-centred reference frame with the origin in

Annual 2010/2011 of the Croatian Academy of Engineering

177

the centroid of p1, ..., pn (Figure 2), and let ii, ji and ki denote the unit vectors of the i-th frame (at time instant i). It can be seen from Figure 2 that

xij = iiT ( p j − d i )

(25)

yij = jiT ( p j − d i )

(26)

zij = kiT ( p j − d i )

(27)

where di is the vector from the original reference frame to the origin of the i-th frame; moreover as the origin is in the centroid of the points,

1 n ∑ pj = 0 . n j =1

(28)

Now, it can be rewritten

1 ~ xij = iiT ( p j − d i ) − ∑ iiT ( pm − d i ) = iiT p j n m=1 n

1 ~ yij = jiT ( p j − d i ) − ∑ jiT ( pm − d i ) = jiT p j n m=1 n 1 ~ zij = kiT ( p j − d i ) − ∑ kiT ( pm − d i ) = kiT p j n m=1

(29)

n

(30) .

(31)

Therefore, the rotation matrix R size of 3N x 3 can be de¿ned as

⎡ i1T ⎤ ⎢ ⎥ ⎢M ⎥ ⎢ inT ⎥ ⎢ T⎥ ⎢ j1 ⎥ R=⎢ M ⎥ ⎢ T⎥ ⎢ jn ⎥ ⎢k1T ⎥ ⎢ ⎥ ⎢M ⎥ ⎢k nT ⎥ ⎣ ⎦

(32)

and a 3 x n shape matrix C as

C = [ p1 K pn ]

~

so it can be written that W = RC .

(33)

178

Annual 2010/2011 of the Croatian Academy of Engineering

4 EXPERIMENTAL EVALUATION This section gives the experimental details of the implemented methods and algorithms on the four-dimensional data sets (3D objects that change in time) collected from an MRI scanner (i.e. MRI brain volumes). Each dataset is representing the brain motion in an MRI scanner. The framework for the estimation of different 3D motions is represented in Figure 3. Different types of the 3D motion are estimated with the different techniques. For the typical linear motion, such as translation, we propose use of the constant Àow algorithm. Rotations in space in time are approximated with the 3D rotation methods. The optical Àow and the feature tracking are proposed for estimation of a long sequence of data or accelerating motion. In case of large disparity between frames we propose the factorization method. The whole process is semi-automated because a medical expert should know what kind of the input data collected from MRI scanner is analyzed, and based on that, the system will select an appropriate technique for given 3D motion.

Figure 3 Proposed framework for 3D motion.

Three consecutive frames of the original experimental dataset (size of 429 kB) are shown in Figure 4. For motion estimation between two frames we use the constant Àow algorithm (Figure 4). The resulting displacement between the ¿rst and the second frame is computed as

d =

d = [ (i − vx ), ( j − v y ), (k − vz ) ] = [153 ,2,0]

(i − vx ) + ( j − v y ) + (k − vz ) 2

2

2

= 153 ,013

and

. The resulting displacement d is

used to estimate the movement between the second and the third frame.

Figure 4 Example of motion between three consecutive frames. Motion between the second and the third frame is determined by the constant Àow algorithm.

Annual 2010/2011 of the Croatian Academy of Engineering

179

Figure 5 Example of motion between three consecutive frames. Motion between the second and the third frame is determined by the optical Àow.

In Figure 5 the same procedure of motion estimation between the second and the third frame is repeated by using of the optical Àow algorithm. There is a signi¿cant visual difference between the estimated and real position of 3D object. Also, in case of the constant Àow algorithm, we have to code only the displacement d which is signi¿cantly less (for three consecutive frames 151 kB) than coding the whole optical Àow for the each voxel (386 kB).

Figure 6 Motion in long sequence of volumes.

Second example is a long sequence of volumes (shown in Figure 6). In this case the feature tracking is the most appropriate technique for the motion estimation, because the displacement in a long sequence of data is not a constant value. In this example feature tracking is described with

pi = [517 ,38,92]

T

and vi = [− 5,0,1]T , the motion on the

volume plane is described with the state vector

T si = [517 ,38,92,−5,0,1] . Estimated

180

Annual 2010/2011 of the Croatian Academy of Engineering

state vector si +1 , with assumed ΔT = 1 , gives values

T si +1 = [512 ,38,94,−52 ,6,−14 ] .

Estimated state vectors for consecutive 8 frames are stated below:

si +1 = [512 ,38,94,−52 ,6,14 ]

T

si + 2 = [460,44,108 ,−37 ,2,−1]

T

,

,

si + 3 = [423,46,107 ,−11,−12,13 ]

,

si + 5 = [413 ,34,122 ,9,6,−1]

si + 4 = [412 ,34,120 ,1,0,2]

T

T

,

si + 6 = [422 ,40,121,17 ,8,5]

T

T

,

si + 7 = [439,48,126,6,24,−2]

,

si + 8 = [445,72,124 ,0,0,0]

T

T

,

.

To recreate motion on the synthesized 3D-models, it is necessary to ¿nd a mathematical solution to tie the analysis to the synthesis. To extract motion information from the speci¿c features of multiple moving objects, the parameters of the visualization semantics of the system that will synthesize the motion, should be known. To do this we ¿rst separate the objects with methods described in Section 3.1. Results are shown in Figure 7.

Figure 7 Example of segmentation; upper left is the original frame; upper right the segmented frame based on the edge and the corner detection of the outer shape; lower left all edges detected based on the methods described in Section 3.1; lower right the ¿nal segmentation of the brain with the inner cavities.

It is also necessary to relate these parameters to the actions that must be applied to the 3D-model in order to recreate motion. These motion models translate the results into an object-visualization parameters building motion ¿elds of the sub-objects in a 3D rotation

Annual 2010/2011 of the Croatian Academy of Engineering

181

and the factorization method for the motion reconstruction.

Figure 8 3D rotation.

Example shown in Figure 8 represents a rigid motion ¿eld that is simply described by a rotation matrix R R = R1 R2 R3 =

− cos(− 13°)sin(29°) cos(− 13°) cos(29°) sin(− 13°) ⎡ ⎤ ⎢ sin(8°)sin(− 13°) cos(29°) + cos(8°)sin(29°) − sin(8°)sin(− 13°)sin(29°) + cos(8°) cos(29°) − sin(8°) cos(− 13°)⎥ ⎢ ⎥ ⎢⎣− cos(8°)sin(− 13°) cos(29°) + sin(8°)sin(29°) cos(8°)sin(− 13°)sin(29°) + sin(8°) cos(29°) cos(8°) cos(29°) ⎥⎦ ⎡ 0,1478 − 0,1478 − 0,225 ⎤ = ⎢⎢0,8509 0,8813 − 0,1356 ⎥⎥ ⎢⎣0,2623 − 0,0015 0,866 ⎥⎦

Example of the factorization method is shown in Figure 9. Aim is to form a registration measurement matrix. In this example, there is a large average disparity between the current and the reference frame. The registered measurement matrix, the rotation matrix and the shape matrix are ⎡ i1T ⎤ ⎡− 0,27 − 0,76 0,9 ⎤ ⎡− 2,84 ⎤ ⎢ T⎥ ⎢ ⎥ ⎢ − 1,13 ⎥ j ⎢ 1 ⎥ ⎢ − 0,73 0,58 − 0,24⎥ ⎢ ⎥ ⎢k1T ⎥ ⎢ − 0,82 0,24 0,31 ⎥ ⎢ − 2,13 ⎥ ⎢ T⎥ ⎢ ⎥ ⎢ ⎥ ⎡4⎤ ⎢ i 2 ⎥ ⎢ − 0,35 − 0,69 0,75 ⎥ ⎢ − 3,065 ⎥ ⎢3,5⎥ ~ ⎢ ⎢ T ⎥ and , ⎢ ⎥ C = ⎥ R = j 2 = − 0,69 0,72 − 0,31 W = − 0,55 ⎢ ⎥. ⎢ ⎥ ⎢ ⎥ ⎢ ⎥ ⎢⎣ 1 ⎥⎦ ⎢k 2T ⎥ ⎢ − 0,91 0,15 0,24 ⎥ ⎢− 2,875 ⎥ ⎢ ⎥ T ⎢ ⎥ ⎢− 3,515 ⎥ ⎢ i3 ⎥ ⎢− 0,49 − 0,61 0,58 ⎥ ⎢ ⎥ ⎢ j T ⎥ ⎢ − 0,59 0,79 − 0,39 ⎥ ⎢ 0,015 ⎥ ⎢ 3T ⎥ ⎢ ⎢− 3,515 ⎥ 0,12 ⎥⎦ ⎢⎣k 3 ⎥⎦ ⎣ − 0,97 0,07 ⎣ ⎦

182

Annual 2010/2011 of the Croatian Academy of Engineering

Figure 9 Example of factorization method.

~

The factorization of the registered measurement matrix W as the product of R and C suggests a method for the reconstruction of structure and motion from a sequence of tracked voxels. 5 CONCLUSION In this paper we have applied different algorithms for the 3D motion estimation and reconstruction. For the estimation of the constant motion ¿eld between only two consecutive frames, the constant Àow algorithm is satisfying enough, and it is not computational extensive as the differential techniques because it is oriented only on a sparse set of the volume points, denoted as the feature points. For estimating motion in long sequence of 3D objects it is more appropriate to use the feature tracking. The main purpose of this framework for the 3D motion ¿eld estimation in 3D MRI data is to reduce a vast amount of acquired data, in coding process. For different types of MRI, different motion techniques should be used for the motion estimation and the motion reconstruction. By coding only a sparse set of information about the volume motion, higher compression ratios can be achieved. For example, size of the original dataset is 429 kB, and with coding only the reference frame and the motion ¿eld based on the combination of proposed algorithms for different kind of 3D motions, the size can be reduced on only 56 kB which gives the compression ratio of 1 : 8.

Annual 2010/2011 of the Croatian Academy of Engineering

183

6 REFERENCES 1.

Aubert G, Deriche R, Kornprobst P (1999) Computing Optical Flow via Variational Techniques. SIAM Journal of Applied Mathematics, Vol. 60, Issue 1: 156-182 2. Black MJ, Anandan P (1996) The robust estimation of multiple motions: parametric and piecewise-smooth Àow ¿elds. Computer Vision and Image Understanding, Vol. 63, Issue 1: 75-104 3. Fashandi H, Fazel-Rezai R, Pistorius S (2007) Optical Flow and Total Least Squares Solution for Multi-scale Data in an Over-Determined System. Lecture Notes in Computer Science, Advances in Visual Computing, Springer Berlin/Heidelberg, Vol. 4842: 33-42 4. Favaro P, Soatto S (2007) 3-D Shape Estimation and Image Restoration: Exploiting Defocus and Motion Blur. Springer-Verlag, London, pp 240-292 5. Furht B, Greenberg J, Westwater R (1997) Motion Estimation Algorithms for Video Compression. Kluwer Academic Publishers, Boston, pp 120-156 6. Messelodi S, Modena CM, Segata N, Zanin M (2005) A Kalman Filter Based Background Updating Algorithm Robust to Sharp Illumination Changes. Lecture Notes in Computer Science, Vol. 3617: 163–170 7. Su Y, Sun MT, Hsu V (2005) Global motion estimation from coarsely sampled motion vector ¿eld and the applications. Circuits and Systems for Video Technology, IEEE Transactions on, Vol. 15, Issue 2: 232-242 8. Ueshiba T, Tomita F (2000) A factorization method for multiple perspective views via iterative depth estimation. Systems and Computers in Japan, Vol. 31, Issue 13:87-95 9. Yuille AL, Hallinan PW, Cohen DS (1992) Feature extraction from faces using deformable templates. International Journal of Computer Vision, Springer Netherlands, Vol. 8, Issue 2: 99-111 10. Wills J, Agarwal S, Belongie S (2006) A Feature-based Approach for Dense Segmentation and Estimation of Large Disparity Motion. International Journal of Computer Vision, Springer Netherlands, Vol. 68, Issue 2: 125-143 11. Žagar M (2009) 4D Medical Data Compression Architecture. PhD Thesis, University of Zagreb, Zagreb, Croatia

184

Annual 2010/2011 of the Croatian Academy of Engineering

Laboratory Evaluation of Mud Differential Sticking Tendency and Spotting Fluid Effectiveness Gaurina-Međimurec Nediljka University of Zagreb, Faculty of Mining, Geology and Petroleum Engineering Pierottijeva 6, Zagreb, Croatia [email protected], Tel. +385 1 5535 825

Pašic Borivoje University of Zagreb, Faculty of Mining, Geology and Petroleum Engineering Pierottijeva 6, Zagreb, Croatia [email protected], Tel. +385 1 5535 840

Matanović Davorin University of Zagreb, Faculty of Mining, Geology and Petroleum Engineering Pierottijeva 6, Zagreb, Croatia [email protected], Tel. +385 1 5535 835

Abstract Differential-pressure pipe sticking occurs when a portion of the drill string becomes embedded in a ¿lter cake that creates on the wall of a permeable formation during drilling. The dominant force is usually associated with a difference in pressures between the hydrostatic pressure of the mud and the pore pressure in the contact area, though adhesion and cohesion may also contribute some resistance to pipe movement. Some level of sticking occurs routinely in drilling operations and differential-pressure-pipe-sticking problems may not be totally prevented. These events only became problematic if the force required to pull the pipe free exceeds the pipe strength. If sticking does occur, several methods can be used for freeing the stuck pipe but the most common approach is to place a small volume of oil, or special spotting Àuid in a wellbore annulus to free differentially stuck pipe. Spotting Àuid reduces the stuck area and the forces of cohesion between the pipe and cake, and allowing pipe to be pulled free. Differential sticking tendency of different drilling Àuids has been determined in laboratory using sticking tester as well as inÀuence of spotting Àuids on freeing differentially stuck pipe. Results of the testing are presented in the paper. Key words: differential sticking, stuck pipe, spotting Àuid, ¿lter cake, drilling 1 INTRODUCTION Pipe sticking is, for most drilling organizations, the greatest drilling problem worldwide. It results in a signi¿cant amount of non-productive time and ends up as one of the major causes of increased well costs (Reid et al., 2000; Pal et al., 2000). Pipe sticking may result in abandonment of the current hole and force a sidetrack, It is estimated that the cost of stuck pipe in deep oil and gas wells can be approximately 25% of the overall budget. In some areas, events related to differentially stuck pipe can be responsible for as much as 40% of the total well cost (Reid et al, 2000). Causes of stuck pipe can be devided in

Annual 2010/2011 of the Croatian Academy of Engineering

185

two categories (Isambourg, et al., 1999): (a) mechanical (key seating, formation-related wellbore instability, wellbore geometry (deviation and ledges), inadequate hole cleaning, junk in hole or collapsed casing, cement related) and (b) differential pressure (wall sticking). Differential pressure sticking is usually indicated when the drill string cannot be rotated, raised or lowered, but full circulation at normal pressure can be established (BushnellWatson and Panesar, 1991). The force required to pull the pipe free can exceed the strength of the pipe. Usually, even if the stuck condition starts with the possibility of limited pipe rotation or vertical movement, it will degrade to the inability to move the pipe at all. The risk of differentially stuck pipe increases when drilling depleted reservoirs and avoids when drilling underbalanced. Stuck pipe causes and methods for their freeing are shown in Figure 1 (BHI, 1998).

STUCK PIPE CAUSE

MECHANICAL

DIFFERENTIAL STICKING

KEY SEATING

WELLBORE GEOMETRY

INADEQUATE HOLE CLEANING

JUNK OR COLLAPSED CASING

CEMENT RELATED

DRILLSTRING JAMMED

DRILLSTRING JAMMED

HOLE PACKED OFF

DRILLSTRING JAMMED

DRILLSTRING JAMMED

WORK STRING UP IF RIH. WORK STRING DOWN IF P.O.O.H.

WORK STRING DOWN TO ESTABLISH OR IMPROVE CIRCULATION

WORK STRING UP AND DOWN

WORK STRING UP AND DOWN. PUMP ACID IF AVAILABLE

WORK STRING DOWN AND ROTATE

DRILLSTRING JAMMED

SLUMP STRING AND ROTATE. REDUCE FLUID WEIGHT. UTILIZE SPOTTING FLUID

Figure 1. Stuck pipe causes and methods for their freeing (BHI, 1998.)

2 MECHANISMS OF DIFFERENTIAL STICKING Differential sticking by de¿nition is a situation in which the drilling assembly (pipe, drill collars and bottomhole assembly) is stuck in ¿lter cake that was previously deposited on a permeable zone (Figure 2a). A signi¿cant overbalance must also exist and the pipe must be stationary, or almost stationary, to allow the bond between the cake and the pipe to develop.

186

Annual 2010/2011 of the Croatian Academy of Engineering

The pipe is held in the cake by a difference in pressures (ǻP) between the hydrostatic pressure of the mud (Pm) and the formation (pore) pressure (Pf) in the permeable zone. This pressure (ǻP) acts upon the area of the pipe in contact with ¿lter cake and isolated from the pressure of the mud in the hole by new ¿lter cake (Figure 2b). There must be a minimum penetration of the drill pipe before a signi¿cant change of pressure at pipe/cake interface can occur. Therefore if ¿lter cake can be made thin enough, sticking can be avoided. Three differential forces have been used to free the pipe: axial (equivalent to working the pipe up or down), radial (equivalent to pulling the pipe across the wellbore) and torque (rotation of the pipe) (Figure 2c). In practice it is likely to be a combination of an axial force and torque which is used to free the pipe. The force required to free differentially stuck pipe must overcome adhesion and the differential pressure exerted by the mud. It has tendency to increase with time until all water is expelled from the ¿lter cake. Once sticking is established, a signi¿cant force is required to free the pipe, even if the mud overbalance is removed. At best, several hours of rig time can be spent in various freeing operations. In more serious cases, the pipe cannot be freed and the well has to be sidetracked or abondoned. Pullout force needed to free a stuck pipe is equal to: F = ǻP·A·— (Eq. 1) where F - pullout force, N ǻP - differential pressure, Pa A - contacta area, m2 — - coef¿cient of friction (adhesion) between the collars and the cake Value of F is also increased with compressibility and thickness of the ¿lter cake, hole deviation, and diameter of the drill collars. It is decreased with increase in diametar of the hole. The pullout force is time-dependent since contact area (A) and coef¿cient of friction (—) increases with time.

Figure 2. Pipe before stuck (a), stuck pipe (b) and forces to free the pipe (c)

Annual 2010/2011 of the Croatian Academy of Engineering

187

3 DIFFERENTIAL STICKING CAUSES Differential sticking causes could be: (a) relatively high differential pressure and (b) mud cake characteristics (thickness, permeability, lubricity). In the situations when is not possible to reduce the differential pressure by reducing the mud weight the option is to act on the mud cake (Outmans, 1974). In situation when pipe is rotating, dynamic ¿lter cake is formed and drill collar penetrates only a short distance into mud ¿lter cake (Figure 3a) but when pipe is stationary, static mud ¿lter cake is formed and drill collar is pushed into mud ¿lter cake by differential pressure (Figure 3b). In highly deviated wellbores, with pipe being stationary, pressure between mud ¿lter cake and drill collar varies from zero to ǻP (Figure 3c).

Figure 3. Situation in wellbore with pipe rotating (a) and pipe stationary (b and c) 3.1 Methods of reducing the risk of differential sticking

Differential sticking tendencies of mud depend on mud ¿lter cake properties: thickness, shear strenght, and lubricity. These ¿lter cake properties are inÀuenced by a combination of variables such as: mud overbalance, solids content of the mud (both high-gravity and low-gravity solids), mud type (e.g., oil-based, polymer water-based, gel water-based), speci¿c mud composition, and Àuid loss. Early detection of differential pressure sticking risks could be made through observation of torque and drag levels while drilling to detect any sign of deviation from a normal trend for the well. The objective of the stuck pipe avoidance practices is to ensure conditions are maintained at all times that allow pulling force to exceed sticking force (shear strenght) (Dupriest et al, 2010). To mitigate differential pressure sticking events, operators often (Montgomery et al., 2007, Simon et al., 2005): • • • • • •

Minimize the overbalance (by decreasing mud weight), Minimize stationary time (drill string rotates at all times), Minimize drilled length through low pressure formations, Increase drill collar and drill string stabilization, Optimize Àuid properties in attempts to minimize the risk of sticking, and Select a drilling Àuid that will yield smooth ¿lter cake with low coef¿cient of friction.

The addition of certain lubricants to water- and oil- based muds will reduce the risk of differential sticking and, should sticking still occur, reduce the force needed to free the stuck pipe or tool. Lubricants are designed to reduce the coef¿cient of friction of drilling Àuid which decreases torque and drag. Depending on their chemical composition and state of dispersion or solubility in the base mud, lubricants can: (a) coat metal surfaces, reducing

188

Annual 2010/2011 of the Croatian Academy of Engineering

the adhesion of steel to the mud cake, (b) be incorporated into the ¿lter cake and provide better Àuid-loss control (resulting in thinner cakes), and (c) be incorporated into the ¿lter cake to reduce the yield stress of the cake. However, despite the best efforts of operators a differential pressure sticking event may still occur (Ayers et al., 1989). The best cure for differential sticking is to prevent it by use of drill-collar stabilizers and, more important, conscientiouslly shortening the intervals of rest when pipe is opposite permeable formations (Dupriest et al., 2010). 3.2 Methods of freeing the stuck pipe

Methods used to get the pipe free, in addition to pulling and torquing the pipe, include: (a) lowering hydrostatic pressure in the wellbore (by reducing the mud weight; this will reduce the differential pressure; should not be used if well control is a problem), (b) placing a spotting Àuid next to the stuck zone and (c) applying shock force just above the stuck point by mechanical jarring, or (d) all the above. The most common approach, however, to getting pipe free is to pump a chemical spotting Àuid in a wellbore annulus. Before placement of spotting Àuid, the depth (free point) to where the drill string is free and where sticking starts must be determined. This free point can be calculated using measurements taken on the rig Àoor. Knowing the stretch ǻL and the forces applied F1 and F2, Hooke’s law, the length of the drill string from the surface to the free point (Lf) is equal to: (Eq. 2) where • • • • •

E is the Elastic Modulus (Young’s Modulus) of steel (i.e., 200 GPa), A is the cross-sectional area of the pipe body (m2), ǻL is the stretch distance (elastical stretch of the free portion of the drill string (m), F1 is the force to place the entire drill string in tension (N), F2 is a force greater than F1 but less that the force limited by the yield stress of the pipe grade (N).

3.3 Spotting fluid

A spotting Àuid (spot) is a small volume or pill of any substance, oil or water base, that is positioned in the wellbore to achieve a speci¿c purpose. Most service companies provide multiple spotting Àuid options. The purpose of the spotting Àuid is to dissolve or break down the ¿lter cake so the pipe can be freed. Possible mechanisms of spotting Àuid action to help free differentially stuck pipe are (Montgomery et al., 2007): • • • • • •

Breaks the capillary forces that hold the drillstring against the wellbore wall, Penetrates, dehydrates and cracks (breaks up) the mud ¿lter cake, Reduces the contact (stuck) area between pipe and wall, Reduces the forces needed to work the pipe free, Increases drillstring lubricity throughout stuck zone. Allows pipe to be pulled free.

Spotting Àuids need to be in place as quickly as possible (within six hours after pipe becoming stuck). Figure 4 shows free drill collar in the center of hole (a), initial (b) and ¿nal (c) position of drill collar in contact with wall, before spotting Àuid. When pipe sticking

Annual 2010/2011 of the Croatian Academy of Engineering

189

occurs, pupming of spotting Àuid has to be as quickly as possible to minimize continued ¿lter cake buildup which leads to higher contact angle and pull free forces (Fig. 4c). Mud ¿lter cake shrunk after spotting Àuid (Fig. 4d) and drill collar is freed (Fig. 4e).

Figure 4. Position of drill collar in the hole before and after spotting Àuid

Spotting Àuids could be devided into four main categories: (a) water-based spotting Àuids, (b) diesel-based spotting Àuids, (c) synthetic-based spotting Àuids, and (d) acid based spotting Àuid. They could be unweighted or weighted (spotting Àuid, viscosi¿er and weighting material) and consist of detergents, soaps, oils, surfactants and other chemicals (wall cake cracking material). Oil-based mud is the traditional spotting Àuid (Krol, 1981; Ayers et al., 1989). Because of concern about mud disposal, spotting Àuids used offshore are either synthetic-based Àuids or benign water-based Àuids (such as drill-in Àuids and salt solutions) (Kercheville et al.,1986). Drill-in Àuids are low and ultra-low solids Àuids; the sealing mechanism is generated inside the rock, so they are leaving just a thin ¿lm on the borehole walls. Salt solutions with a low activity coef¿cient combined with environmentally-safe lubricants (two-phase spot) produce low torque levels. 4 LABORATORY RESEARCH Laboratory tests were carried out to determine the effectiveness of mud system additives in minimizing differential pressure sticking of drill pipe. Differential sticking tendency of the tested lignosulfonate mud was evaluated using differential sticking tester marketed by OFI Testing Equipment International (Figure 5). The test device consists of ¿ltration cell capable of holding 200 mL of Àuid, perforated bottom capable of holding ¿lter paper and screen, plate (on a plunger) and torque wrench. Torque necessary to break the plate free is measured. Tests were carried out at room temperature and pressure of: 3 292.2 kPa (477.5 PSI). Differential sticking tester measures the stuck pipe tendency coef¿cient of drilling Àuids, and also determines how effective lubricants or treatments might be with any given drilling Àuid.

190

Annual 2010/2011 of the Croatian Academy of Engineering

Figure 5. OFITE Differential sticking tester

This coef¿cient takes into account both the friction, or “stickiness”, of the ¿lter cake, as well as the amount of cake building that must occur in order to freeze or stick the pipe in the hole. By measuring the area of cake building during a test, the bulk sticking coef¿cient is obtained and read directly at the end of the test. How likely a given Àuid will be to produce a stuck pipe situation and how effective a given treatment may be, can be immediately determined on-site. The bulk sticking coef¿cient (Ksc) is calculated by dividing the sliding force (Fs) by the normal force (Fn): Ksc = Fs / Fn

(Eq. 3)

For radius of plate r = 25.4 mm (1”): The sliding force (Fs) of the plate is a function of the measured torque (Tu). Fs = 1.5 × Tu

(Eq. 4)

The normal force (Fn) on the plate is determined by multiplying the area by the differential pressure (This assumes that a pressure of 477.5 PSI was used during the test) (Eq. 5) Fn = 1 500 × r2

(Eq. 5)

Finally, the bulk sticking coef¿cient (Ksc) is equal to: Ksc = 0.001 × Tn

(Eq. 6)

Annual 2010/2011 of the Croatian Academy of Engineering

191

Where Tu is torque reading (lbf -inch or 0.1129 N·m). The stuck tendency coef¿cient (Kst) is equal to the bulk sticking coef¿cient (Ksc) multiplied by the variable stuck area (Eq. 7): Kst = Ksc × (Variable Stuck Area) (Eq. 7) Laboratory tests were run to evaluate the effectiveness of mud system additives: Carboxymethylcellulose - CMC (¿ltration control additive) and lubricant on differential sticking tendency of the tested Àuids (Gaurina-Meÿimurec et al., 2010.). Formulations of tested lignosulphonate mud is shown in Table 1. To determine inÀuence of CMC and lubricant on mud properties, especially on differential sticking tendency, amount of CMC and lubricant (Lube 167) in base lignosulphonate mud was varied. The numbers included in the mud sign implies CMC concentration in grams per 1 L of the mud (C0, C3, C5) and lubricant concentration in ml per 1 L of the mud (L0, L20, L40). Mud properties such as API Àuid loss, cake thickness, rheological properties and pH value were determined according to API RP13B (API Recommended Practice Standard Procedure for Testing Drilling Fluids). Spotting Àuid was composed of 620 ml Diesel oil, 80 ml Pipe Lax W, 280 ml water and 73 g barite. Spotting time was 16 hours. Composition Water Bentonite

Units

C0L0 C3L0 C5L0 C0L20 C3L20 C5L20 C0L40

C3L40 C5L40

ml

1000

1000

1000

1000

1000

1000

1000

1000

1000

g L -1

80

80

80

80

80

80

80

80

80

-1

FCL

gL

20

20

20

20

20

20

20

20

20

NaOH

g L -1

3

3

3

3

3

3

3

3

3

CMC

gL

-1

0

3

5

0

3

5

0

3

5

Viscosifer

g L -1

5

5

5

5

5

5

5

5

5

0,6

0,6

0,6

0,6

0,6

0,6

0,6

0,6

0,6

0

0

0

20

20

20

40

40

40

-1

Biocide

gL

Lubricant

ml L-1

Defoamer

ml L

-1

Barite

g L -1

1

1

1

1

1

1

1

1

1

500

500

500

500

500

500

500

500

500

Table 1. Formulation of the tested lignosulphonate mud

Laboratory results are shown in ¿gures from 6 to 15. Torque increases with time regardless of concentration of CMC and lubricant, but for speci¿ed time decreases with increasing concentration of CMC and lubricant. In addition, torque decreases after placement of spotting Àuid for16-hours (Fig. 6).

192

Annual 2010/2011 of the Croatian Academy of Engineering

Figure 6. Effect of time and additives on torque

Bulk sticking coef¿cient after 300 min sticking time is decreased with increasing concentration of CMC and lubricant. For instance, the value of bulk sticking coef¿cient of C5L40 mud is 35,3 % lesser than of C0L0 mud, and for C5L40 bulk sticking coef¿cient is 2,3 times less after 16 hours spotting time (Fig. 7).

Figure 7. Effect of CMC and lubricant on bulk sticking coef¿cient

Bulk sticking coef¿cient is decreased with increasing concentration of CMC but increased with time regardless of concentration of CMC (mud without lubricant) (Fig 8).

Annual 2010/2011 of the Croatian Academy of Engineering

193

Figure 8. Effect of time and concentration of CMC on bulk sticking coef¿cient

Bulk sticking coef¿cient of mud without lubricant and with 5 g CMC after 300 min is 9,2 % less than without CMC but 2,4 times higher than after 60 min (Fig. 9). The bulk sticking coef¿cient of mud with 5 g CMC and with 4% lubricant after 300 min test is 29 % less than without lubricant.

Figure 9. Effect of lubricant on bulk sticking coef¿cient

The corelation between torque after 60 min and mud properties (Àuid loss, cake thicknes, plastic viscosity, gel strengts and pH value) is observed. The lower Àuid loss value the lower value of torque (Fig. 10). In addition, API Fluid loss decreases with addition of CMC

194

Annual 2010/2011 of the Croatian Academy of Engineering

and lubricant.

Figure 10. The corelation between torque and Àuid loss of tested muds

Cake thickness and torque decreases with increasing concentration of CMC and lubricant. The thiner cake the lower torque (Fig. 11).

Figure 11. The corelation between torque and cake thickness of tested muds

Plastic viscosity and yield point increase with increasing concentration of CMC and

Annual 2010/2011 of the Croatian Academy of Engineering

195

lubricant (Fig. 12 and 13).

Figure 12. The corelation between torque and plastic viscosity of tested muds

Figure 13. The corelation between torque yield point of tested muds

Gel strenght increases and pH value decreases with increasing concentration of CMC and lubricant (Fig. 14 and 15).

196

Annual 2010/2011 of the Croatian Academy of Engineering

Figure 14. The corelation between torque and gel strenght of tested muds

Figure 15. The corelation between torque and pH value of tested muds

5 CONCLUSIONS The sticking tendency of mud can be estimated using differential sticking tester. Use of the test at the well-site allows the mud engineer to identify the problem before the sticking

Annual 2010/2011 of the Croatian Academy of Engineering

197

occurs, and to select and implement the most effective treatment option. According to performed tests torque and bulk sticking coe¿cient constantly increase with sticking time regardless of concentration of CMC and lubricant in tested lignosulphonate mud, but for any speci¿ed sticking time (60 min, 120 min, 180 min, 240 min or 300 min) their values decrease with increasing concentration of CMC and lubricant. After placement of spotting Àuid for16-hours torque signi¿cantly decreases. In addition, the corelation between mud composition, torque and mud properties is observed. API Àuid loss, ¿lter cake thickness and pH value decrease but plastic viscosity, yield point and gel strenght increase with increasing concentration of CMC and lubricant. Fluid loss and ¿lter cake thickness have a direct inÀuence on values of the torque and sticking coef¿cient (reducing sticking area). The lower Àuid loss and the thiner cake the lower values of the torque and the bulk sticking coef¿cient. Therefore using the lignosulphonate mud signed C5L40 with 5 gL-1 CMC and 40 mlL-1 lubricant sticking pipe can be avoided or if it still occurs the force required to free differentially stuck pipe will be less. 6 REFERENCES 1.

Ayers, R.C., O´Reilly, J.E., Henry, L.R. (1989.) Offshore Operators Committee Gulf of Mexico Spotting Fluid Survey, paper SPE/IADC 18683, presented at the 1989 SPE/ IADC Drilling Conference, New Orleans, Louisiana, February 28-March 3, 517-519. 2. Bushnell-Watson, Y.M., Panesar, S.S. (1991) Differential Sticking Laboratory Tasts Can Improve Mud Design, paper SPE 22549 presented at the 66th Annual Technical Conference and Exhibition of the Society of Petroleum Engineers, Dallas, TX, October 6-9.147-156. 3. Dupriest, F.E., Elks Jr., W.C., Ottesen, S. (2010.) Design Methodology and Operational Practices Eliminate Differential Sticking, paper IADC/SPE 128129, presented at the 2010 IADC/SPE Drilling Conference and Exhibition, New Orleans, Louisiana, February 2-4. 5-13. 4. Gaurina-Meÿimurec, N., Pašiü, B., Matanoviü, D. (2010.) Spotting Fluids for Freeing Differentially Stuck Pipe, The 1st Annual Congress of Oil Field Chemicals, Conference Proceedings, Beijing, China, November 15-17, 57. 5. Isambourg, P., Ottesen, S., Benaissa, S., Marti, J. (1999.) Down-Hole Simulation Cell for Measurement of Lubricity and Differential Pressure Sticking, paper SPE/IADC 52816, presented at the 1999 SPE/IADC Drilling Conference, Amsterdam, Holland, March 9-11, 1-12. 6. Kercheville, J.D., Hinds, A.A., Clements, W.R. (1986.) Comparison of Environmentally Acceptable Materials with Diesel Oil for Drilling Mud Lubricity and Spotting Fluid Formulations, paper IADC/SPE 14797, presented at the 1986 IADC/SPE Drilling Conference, Dallas, Texas, February 10-12, 611-615. 7. Krol, A.D. (1981.) Laboratory Evaluation of Stuck Pipe Spotting Fluid Effectiveness, paper SPE 10096, presented at the 56th Annual Fall Technical Conference and Exhibition of Society of Petroleum Engineers of AIME, San Antonio, Texas, October 5-7.1-13. 8. Montgomery, J.K., Keller, S.R., Krahel, N., Smith, M.V. (2007) Improved Method for Use of Chelation to Free Stuck Pipe and Enhance Treatment of Lost Returns, paper SPE/IADC 105567 presented at the 2007 SPE/IADC Drilling Conference,Amsterdam, The Netherlands, 20-22 February, pp 1-7. 9. Outmans, H.D. (1974.) Spot Fluid Quickly to Free Differentially Stuck Pipe, The Oil and Gas Journal, July 15, 65-68. 10. Pal, K., Mangla, V.K., Joshi, N.P., Tewari, P.P., Bose, U.N. (2000.) A New Approach

198

Annual 2010/2011 of the Croatian Academy of Engineering

Solves Differential Sticking Problems in Geleki and Ahmedabad Fields, paper SPE 63055, presented at the 2000 SPE Annual Technical Conference and Exhibition, Dallas, Texas, October 1-4.1-8. 11. Reid, P.I., Meeten, G.H., Way, P.W., Clark, P., Chambers, B.D., Gilmour, A., Sanders, M.W. (2000.) Differential-Sticking Mechanisms and a Simple Wellsite Test for Monitoring and Optimizing Drilling Mud Properties, SPE Drilling & Completion, Vol. 15, No. 2, June, 97-104. 12. Simon, K., Gaurina-Meÿimurec, N.; Pašiü, B. (2005) Drilling Fluids Differential Sticking Tendency Determination, Rudarsko-geološko-naftni zbornik, Vol.17, Faculty of Mining, Geology and Petroleum Engineering, Zagreb, 31-35. 13. ***, (1998) FLUID Facts – Engineering Handbook, Baker Hughes Inteq, Houston, TX, March 1998.

Annual 2010/2011 of the Croatian Academy of Engineering

199

Use of Plasma Technology for Modification of Textiles Prof. dr. Ružica Čunko Faculty of Textile Technology, University of Zagreb Prilaz baruna Filipovića 28a, Zagreb [email protected] tel. 01/3712-523

Dr. Sanja Ercegović Ražić Faculty of Textile Technology, Univesity of Zagreb Prilaz baruna Filipovića 28a, Zagreb

Abstract The use of non-thermal, oxygen and argon low-presser plasmas for modi¿cation of cellulose textiles is discussed. The emphasis is given to the characterization of the complex micromorphology and physical-chemical changes of the textile surface after plasma treatments, as well as on de¿ning the impacts of these changes on the obtained ¿nal properties of textile materials treated in this way. SEM and AFM results con¿rmed the thesis that using oxygen and argon plasmas, two, essentially different, processes of textile surface ablation occurred; the ¿rst is chemical etching and the second physical sputtering. Using the AFM, the ¿bre surface morphology was estimated at nanoscale. The measurements of the vertical rise of water in fabric sample and the results of the wetting time (vertical test and drop test) indicate the improvement of hydrophilic properties of all tested samples after lowpressure oxygen plasma treatment. The presented results con¿rm that the low-pressure plasma treatment is an acceptable and appropriate method of textile surface modi¿cation and an eco-friendly method at the same time. Keywords: low-pressure plasma, surface functionalization, cellulose fabric modi¿cation. 1 Introduction The textile and clothing industries in Europe, USA, Japan and some other developed countries are facing some big challenges today, largely because of the globalization process. Therefore, the shift to high- and multifunctional, added value, luxury and technical textiles is deemed to be essential for their sustainable growth. The growing environmental and energy-saving conditions will also lead to the gradual replacement of many traditional wet chemistry-based textile processing, using large amounts of water, energy and efÀuents, by various forms of low-liquor and dry-¿nishing processes. Plasma technology, when developed at a commercially viable level, has strong potential to offer in an attractive way achievement of new functionalities in textiles. The signi¿cant research work on introduction of plasma technology in the textile ¿eld has been going since the early 1980s and recently the application of plasma treatments for textile modi¿cation has become more and more signi¿cant. In many laboratories across the world the researches are dealing with plasma treatments of variety of ¿brous materials in order to improve their functional properties. A variety of low-pressure and atmospheric plasma machines, mostly in prototype form, have been offered for industrial processing of

200

Annual 2010/2011 of the Croatian Academy of Engineering

textiles. In recent years, considerable efforts have been made by many plasma technology suppliers, as well as researches, to impart broad range of new functionalities and textile properties by plasma treatment [2, 7, 10, 23, 24, 27, 28]. In the recent point of view, a desired ¿ber surface characteristics and textiles modi¿cation by plasma seem very promising. Plasma treatments can be used both in substitutions of conventional processes and for the production of innovative textile materials with properties that cannot be achieved via wet processing. They are applicable to different textile substrates, even to those that cannot be modi¿ed by conventional methods. It is important to note, that plasma agency is limited to the top surface layer and no signi¿cant alteration of bulk properties of the ¿ber is produced. Plasma treatments are fast and extremely gentle, as well as environmentally friendly, being dry processes characterized by low consumption of chemicals and energy [3, 15]. If plasmas are used as pre-treatment, they can reduce the amount of chemicals required by the conventional process and the concentration of pollutants in the efÀuents. The ¿rst investigation of possibilities and phenomenology of textiles modi¿cation by plasma treatment in Croatia was carried out within the scope of the research project Multifunctional textile materials for personal protection (Faculty of Textile Technology, Department of materials, ¿bres and textile testing, 1997-2011). The results of a part of this research are presented in this paper. 2 Possibilities and phenomenology of cellulose textiles modification by plasma treatment 2.1 What is plasma?

Plasma was ¿rst identi¿ed (as “radiant matter”) by Sir W. Crookes in 1879; Sir J. J. Thomson identi¿ed the nature of the matter in 1897, and in 1928. I. Langmuir has introduced the term “plasma”. In physics plasma is de¿ned as distinct phase of matter, separate from the traditional solids, liquids, and gases. It comprises a dynamic mix of ions, electrons, neutrons, photons, free radicals, meta-stable excited species and molecular fragments. Since the particles in plasma are electrically charged (generally by being stripped of electrons), it is frequently described as an “ionized gas.” Energy is needed to strip electrons from atoms to make plasma. The energy can be of various origins: thermal, electrical, or light (ultraviolet light or intense visible light from a laser). With insuf¿cient sustaining power, plasmas recombine into neutral gas. Plasma is electrically conductive and can be manipulated by magnetic ¿elds. It is odd to consider that plasma is actually the most common phase of matter, especially since it was the last one discovered. Flame, lightning, interstellar nebulae, stars, and even the empty vastness of space are all examples of the plasma state of matter.

Annual 2010/2011 of the Croatian Academy of Engineering

201

Figure 1 Plasma lamp [13] The colors are a result of relaxation of electrons in excited states to lower energy states after they have recombined with ions. These processes emit light in a spectrum characteristic of the gas being excited.

2.2 Principles of plasma processes and application of plasma in the field of textiles Plasmas are generally classi¿ed as thermal and non-thermal. Thermal (“hot”) plasmas are characterized by a condition of thermal equilibrium between different species contained in the gas and their temperature raise of several thousand degrees. It is clear that these plasmas due to their destructive nature, are not suitable for textile treatment. Contrary to thermal plasmas, non-thermal plasmas are “cold” plasmas and are produced at room temperature or little above room temperature. In this case, electrons have higher energies than ions and molecules (their energies ranging from 0.1 to some electron volts), and, due to the low density of the gas, collisions with the other species are relatively rare and thermal equilibrium is not reached. The bulk temperature of the gas is comparable to room temperature. Electron collisions with neutral species produce additional electrons and ions. Thanks to the low operating temperatures, cold plasmas are suitable for textile treatment. The most important for this purpose are low-pressure plasma and atmospheric pressure plasma [11, 17, 20, 22, 23]. Using of high vacuum pumps, the low-pressure plasma reactors are working in the pressure range of 10-2 to 10-3 mbar (vacuum vessel). The gas which is then introduced in the vessel is ionized with the help of high frequency generator. For radiofrequency range (typically 40 KHz or 13, 56 MHz), normally the working gas pressure is kept in the lower 0.1 mbar range, whereas for micro-wave sources (915 MHz or 2.45 GHz), a working pressure between 0.5 and 1 mbar is often used. The interaction between the very active chemical species and photons present in the plasma gas and a textile substrate is the basis of dissimilar industrial applications. As a consequence of the very complex and non-equilibrium nature of cold plasmas, a multiplicity of very different phenomena can occur, depending on the art of the gas and the operating conditions [11, 12, 17, 18, 28]. The most important plasma/textile interaction processes are: cleaning or etching, surface activation, grafting, polymerization and deposition (Fig. 2). All these phenomena occur at the ¿bre surface and are limited to the most external layer of the substrate. Normally, the effects do not involve layers deeper than 10 -100 nm.

202

Annual 2010/2011 of the Croatian Academy of Engineering

Figure 2 Cold plasma processes [19] 2.3 Mechanisms of interactions between plasmas and cellulose fibres

Cellulose is the most widely known ¿brous polysaccharides and the basic building polymer of agricultural ¿bres (cotton, Àax, hemp) as well as of various man-made ¿bres (viscose, modal, lyocell, bamboo). The basic structural repeating unit for cellulose is the cellobiose unit, composed of two anhydroglucose rings e.g. anhydroglucose units. Each unit contains three alcohol hydroxyls groups (-OH). These hydroxyls form hydrogen bonds inside the macromolecule itself (intra-molecular) and between other cellulose macromolecules (intermolecular). Cellulose in ¿bres is highly crystalline polymer with tight packing of ordered chains wherein the hydrogen bonding is very strong. Cellulose ¿bre also contains amorphous regions in which there can be a variation from complete disorder of the cellulose chains to some order. Owing to the hydroxyl groups, cellulose can be traditionally modi¿ed through reactions of esteri¿cation and etheri¿cation. It has been determined that the accessibility of the -OH groups is controlled by the relative ratios of crystalline and amorphous regions of the cellulose ¿bre superstructure. Amorphous cellulose can be functionalized much more readily than their crystalline counterparts. Cellulose modi¿cation using cold plasma differs signi¿cantly from the traditional esteri¿cation and etheri¿cation reaction. Reactive plasma gas components (particles) can react directly with the anhydroglucose units by abstraction of hydrogen atoms from either carbon or oxygen atoms in the plasma reactor. Thus, cellulose can be oxidized, reduced and/or substituted in new ways. XPS data collected from plasma treated cellulose samples indicate that new functionalities are formed compared with unmodi¿ed cellulose. In addition to C-OH and C-O-C, the existence of O=C=O, O-CO-O, O=C-O, C=O and O-CO-O functionalities are detected [6, 27]. The interaction of the active species of plasma with cellulose in ¿bre surface, involves several electron-mediated processes, as well as positive ion-induced reactions. Some of these reactions can promote hemolytic cleavages leading to the formation of free-radical

Annual 2010/2011 of the Croatian Academy of Engineering

203

sites. These reactive centers can lead to a variety of functionalization mechanisms. Cellulose can be grafted by monomer introduced in plasma gas and the grafting process is restricted to the surface layer [2, 4, 12, 21, 26]. By using speci¿c molecules, a process known as plasma-enhanced chemical vapor deposition (PE-CVD) may occur, as well as metal nanoparticles can be deposed on ¿bre surface using plasma gas. All these processes are by-product free. The reactions are not accompanied by the deposition of undesired macromolecular layers on the reactor walls. Plasma treatments usually do not create signi¿cant environmental problems. 3 Experimental, materials and methods The emphasis in this investigation is given to the characterization of the complex micromorphology and physical-chemical changes of the textile surface after plasma treatments, as well as on de¿ning the impacts of these changes on the obtained ¿nal properties of textile materials treated in this way. 3.1 Materials

The textile substrate used was the cellulose fabrics woven in plain wave, using threads spun from lyocell and modal man-made cellulose ¿bres and from cotton. The structural characteristics of sample fabrics are presented in Table 1. Before plasma treatment, the raw cotton fabric was undergoes to desizing and scouring treatments according to the industrial process conditions (sample CO sc). Table 1: Basic characteristic of tested fabric samples woven in plain weave

Samples

Cotton (CO sc) Lyocell (CLY I) Modal (CMD I)

Yarn ¿neness (warp/ weft) / tex 29.4/25.0 31.3/25.0 31.3/25.0

Numb. of threads (warp/weft) / cm

25/20 21/19 23/19

Volume mass / gcm-3

Mass per unit area / gm-2

UT

O2-p

Ar-p

UT

O2-p

Ar-p

0.369

0.367

0.362

130.6

128.1

128.3

0.409

0.431

0.444

120.1

126.8

129.8

0.420

0.429

0.444

121.8

123.9

126.0

UT = untreated sample, O2-p = oxygen plasma treated, Ar-p = argon plasma treated sample 3.2 Plasma treatment of fabric samples

Plasma system Low-pressure (LP) - plasma NANO LF laboratory system (Diener Electronic GmbH, Germany, Fig 3) of 40 kHz low-frequency generator and maximum power of 300 W was used for fabric treatments. One electrode and the four trays were placed inside the cylindrical vacuum chamber of 24 L volume. A standard part of the plasma system was a conventional rotary vane pump, type D8B, suction power of the approx. 8 m3/hour, with an electromagnetic valve, which prevented the return of the oil vapor into the vacuum chamber. Brass needle valves (max. Àow rate of 400 sccm), with pressure reducers speci¿ed for each type of gases applied, provided constant gas Àow rate during plasma processes performed. The plasma treatments were controlled fully automatically.

204

Annual 2010/2011 of the Croatian Academy of Engineering

Figure 3 Plasma system LP-Nano LF-40kHz, Diener Electronic GmbH

Surface activation with oxygen and argon plasma Before plasma treatment, the fabric samples (dimensions 340x200 mm) were dried to remove the moisture and afterwards treated (separately with oxygen and argon plasma) under the optimum process conditions, determined by preliminary work [6]: the gas Àow rate of 40 sccm, working pressure of 0.34 - 0.4 mbar, operating power of 300 W and treatment time of 5 min. 3.3 Method of analysis

Surface Analysis The ¿bre surface micro-morphology characteristics of untreated and plasma treated fabrics were investigated using JEOL 6060LV Scanning Electron Microscopy (SEM) technique. Before analysis all the samples were coated with carbon and a 90% Au/10%Pd alloy layer. Magni¿cation of 1000x and 7000x was used for analysis of surface changes depending on the tested ¿bre type and applied plasma gas. The topography of ¿bre surface roughness caused by oxygen and argon plasma treatment was analyzed using the most recent Atomic force microscopy (AFM). The concept on which AFM is based is the generation of the images of the ¿bre surface by measuring the physical interactions (forces) between a sharp tip of AFM cantilever and the sample (Fig 4). When the tip is brought into proximity of a sample surface, forces between the tip and the sample lead to a deÀection of the cantilever according to Hooke´s law. Depending on the situation and surface topography, forces that are measured in AFM include mechanical contact force, Van der Waals forces, capillary forces, chemical bonding, electrostatic forces, magnetic forces, etc. Registered values of cantilever deÀection are electronically converted into pseudo three-dimensional (3D) image of a sample.

Annual 2010/2011 of the Croatian Academy of Engineering

205

Figure 4 Basic schematic principle of the AFM in contact mode [14]

An important feature of the AFM technique is based on the fact that, apart from the qualitative characterization, it also allows the quantitative assessment of ¿bre surface topography at the nano-scale, where the changes caused by plasma action are expected. AFM imaging was performed in contact mode, using a Multimode AFM with Nanoscope IIIa controller (Veeco Instruments, Santa Barbara, CA) with a vertical engagement (JV) 125 ȝm scanner. Contact mode imaging was performed using silicon-nitride tips (NP20, Veeco, nom. freq. 56 KHz, nom. spring constant of 0.32 N/m); with the highest scan resolution of 512 samples per line. Processing and analysis of images were carried out using the NanoScope™ software (Digital Instruments, version V531r1). Wet ability properties Wet ability properties were analyzed using vertical test and drop test. The Vertical test [16] was used to measure the rate of vertical capillary rise of distilled water in a woven specimen strip suspended in the distilled water (at 15r2 mm below the water surface). The vertical rise of distilled water in fabric sample was measured in weftand warp direction. In order to achieve the better visibility of the water front rise, the methylene-blue as a water colorant was used. All test specimens were conditioned at the standard atmosphere, according to HRN ISO 139. Test was performed in accordance with EN ISO 9073-6:2000. In order to improve reproducibility and reliability of measurement results, all the measurements were recorded by digital video camera recorder SONY DCRDVD304E. The Drop test was carried out as a measure of absorbency of fabric samples. The absorbency is de¿ned as the propensity of a material to take in and retain a liquid, usually water, in the pores and interstices of the material. Test was performed in accordance with AATCC e.g. TEGEWA prescription [1, 25]. A drop of water of a de¿ned volume is allowed to fall from a ¿xed height onto the taut surface of a test specimen. The time required for full absorption of water drop (e. g. time to disappear of drop) is measured and recorded as wetting time (seconds, stopwatch). The shorter the time, the more absorbent is the textile material. Five seconds or less is generally considered to represent adequate absorbency.

206

Annual 2010/2011 of the Croatian Academy of Engineering

4 Results and discussion Surface analysis by Scanning Electron Microscopy SEM analysis of the lyocell and modal cellulose ¿bres in the fabric sample surface were carried out with magni¿cation of 1000x and 7000x. A part of obtained results [9] are presented in Fig. 5 and 6.

Untreated lyocell ¿bres

Oxygen-plasma treated lyocell ¿bres

Argon-plasma treated lyocell ¿bres

Figure 5 SEM photographs of untreated, oxygen and argon plasma treated lyocell ¿bres

Untreated modal ¿bres

Oxygen treated modal ¿bres

Argon treated modal ¿bres

Figure 6 SEM photographs of untreated, oxygen and argon plasma treated modal ¿bres

The obtained results with SEM technique indicate that the oxygen plasma treatment certainly clean the ¿bres surface, removing the organic contaminants from surface. The treatment with argon plasma results in sputtering of the ¿bre surface layer increasing the roughness and ¿brillations-effect (especially by lyocell) of the ¿bres. Thereby the speci¿c area of ¿bres increases and the size of ¿bre area accessible to reactions increases to. Surface topography analysis by atomic force microscopy (AFM) The examples of ¿bre surface topography images of untreated, as well as oxygen- and argon-plasma treated lyocell ¿bres, observed by AFM [6, 8], are shown in Fig. 7 - 9. The results are presented in a standard form, e.g. as 3D plot of ¿bre surface (a), the height data image (b), and section analysis of ¿bre along marked lines (c). The vertical distance of scanned topographic elements, are given as a measure of roughness of imaged surface area of ¿bre.

Annual 2010/2011 of the Croatian Academy of Engineering

a

207

b

Figure 7 The AFM of untreated lyocell ¿bres: a - Surface plot - 3D topographic image; b - Section analysis along the marked line vertical to the axis of the tested ¿bre (imaged area is 25 ȝm × 25 ȝm with vertical scale of 3000 nm). Vertical distance marked on the c: approx. 122 nm

a

b

Figure 8 The AFM of oxygen plasma treated lyocell ¿bres: a - Surface plot - 3D topographic image; b - Section analysis along the marked line vertical to the axis of tested ¿bre (imaged area is 15 ȝm x 15 ȝm with vertical scale of 1500 nm) Vertical distance marked on the c: approx. 24 nm

a

b

Figure 9 The AFM of argon plasma treated lyocell ¿bres: a - Surface plot - 3D topographic image; b - Section analysis along the marked line vertical to the axis of tested ¿bre (imaged area is 10 ȝm × 10 ȝm with vertical scale of 2000 nm) Vertical distance marked on the c: size range from approx. 23 nm to 59 nm

208

Annual 2010/2011 of the Croatian Academy of Engineering

The results of surface roughness noticeable on the surface plots (Fig. 7 - 9) and the results obtaining by section analysis point to a different effect of oxygen and argon plasma on the tested ¿bres, as compared to the untreated ones. Surface roughness of the oxygen plasma treated ¿bres was considerably lower (approx. 24 nm) compared to the untreated samples, where ¿bre surface roughness was approximately 122 nm. These results additionally prove that oxygen, as a reactive gas, changes the outermost ¿bre surface layers, resulting in the removal of surface impurities and leading to smoother surface layer, as can be seen in Fig. 8. As opposed to oxygen, argon, as an inert gas, caused increased ¿bre surface roughness and ruptures along the longitudinal axis of the tested ¿bres, as can be seen in Fig. 9. The surface of the argon plasma treated sample was altered and ruptures appeared, visible on the surface and oriented in the direction of ¿bre axis. Their size was in the range from 23 to 59 nm at the imaged area as well. AFM results con¿rmed the thesis that using oxygen and argon plasmas, two essentially different processes of textile surface ablation occurred; the ¿rst is chemical etching and the second physical sputtering. Capillary rise of water by vertical test The capillarity method measures the rate of vertical capillary rise in a specimen strip suspended in the distilled water. The part of obtained results [5] is presented on the graphs, Fig. 10 – 12. 150 140

Height of the water front, h [mm]

130 120 110 100 90 80 70 60 50 40 30 20

CLY I ut (warp) CLY I O2 plasma, 5 min (weft) CLY I Ar plasma, 5 min (weft)

CLY I O2 plasma, 5 min (warp) CLY I Ar plasma, 5 min (warp) CLY I ut (weft)

10 0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 Time, t [min]

Figure 10 Results of the capillary rise of water measured on untreated and plasma treated lyocell fabrics in warp and weft direction

Annual 2010/2011 of the Croatian Academy of Engineering

209

150 140 130

Height of the water front, h [mm]

120 110 100 90 80 70 60 50 40 30 20

CMD I ut (weft) CMD I Ar plasma, 5 min (warp) CMD I O2 plasma, 5 min (weft)

CMD I O2 plasma, 5 min (warp) CMD I Ar plasma, 5 min (weft) CMD I ut (warp)

10 0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 Time, t [min]

Figure 11 Results of the capillary rise of water measured on untreated and plasma treated modal fabrics in warp and weft direction 150 140

Height of the water front, h [mm]

130 120 110 100 90 80 70 60 50 40

CO I O2 plasma, 5min (warp)

CO I Ar plasma, 5min (warp)

30

CO I ut (warp)

CO I ut (weft)

20

CO I O2, 5min (weft)

CO I Ar plasma, 5min (weft)

10 0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 Time, t [min]

Figure 12 Results of the capillary rise of water measured on untreated and plasma treated cotton fabrics in warp and weft direction

It is important to emphasize that several of aspects affect the wettability of treated samples, such as: ¿bre superstructure, ¿bre diameter, ¿bre surface irregularity and surface chemical composition. Micro- and macrostructure of fabric is very important to. In this respect, the increase of fabric structure density during plasma treatment cud be reason of the capillary force raising and better vertically rise of water through fabric.

210

Annual 2010/2011 of the Croatian Academy of Engineering

Absorption time of water drop by dropt test Results of the water drop absorption time of tested fabrics are presented in Tab. 2. Table 2: Results of the water drops absorption time [s] of tested fabrics Sample Treatment

Cotton (Co raw)

Cotton (Co sc)

Lyocell (CLY I)

Modal (CMD I)

Untreated

> 3600

5.1

2

1.2

O2 plasma, 5 min

10.0

0.8

1

0.8

Ar plasma, 5 min

> 3600

37.8

0.8

0.8

Obtained results presented in Table 2 indicate the improvement of wetting and wicking properties of all tested samples by low-pressure oxygen plasma treatment. The oxygen plasma treatment is especially effective for raw cotton fabric. In this case the oxygen plasma treatment could be recommended as eco-friendly substitution for the wet scouring process, in order to achieve good soak properties (wettability). Plasma treatments with argon as an inert gas are not enough effective to improve wettability of cotton fabrics; the drop absorption time (s) determined for raw cotton fabric is extremely high (characteristic for hydrophobic surface). Hydrophilic property of scoured cotton fabric after argon plasma treatment still decreases (drop absorption time was increased from 5.1 to 37.8 s). Could be supposed that super¿cial primary wall of cotton ¿bres is thick enough to protect ¿bers of plasma ionized particles attacks. Contrary, the low-pressure argon plasma treatments of lyocell and modal fabrics additionally improve their original high hydrophilic properties. 5 Conclusions Presented results are a part of investigation of the nature of low-pressure plasma inÀuence on the ¿bre surface characteristics and their effects on the properties modi¿cation of cellulose based textile materials. • The results obtained using the SEM technique indicate that the oxygen plasma treatment leads to the cleaning and etching of the ¿bre surface, in contrary to the argon plasma treatment which resulting in sputtering of the cellulose ¿bre surface layer. This results in increasing the roughness of the ¿bres and contributes to the enhancement of ¿bre speci¿c area (ratio: surface area/volume). • The AFM results con¿rmed the thesis that using oxygen and argon plasmas, two essentially different processes of textile surface ablation occurred; the ¿rst is chemical etching and the second physical sputtering. The estimated vertical altitude of surface topographic elements of the oxygen plasma treated ¿bres was considerably lower compared to the untreated samples (i.e. 24 nm in relation to 122 nm for untreated sample). The surface of the argon plasma treated sample was altered and the visible ruptures appeared on the surface. The cracks at the imaged area are oriented in the direction of ¿bre axis and their vertical altitude range of 22 to 59 nm as well. • The results of the vertical test and the drop test con¿rm the improvement of hydrophilic properties of all tested samples after low-pressure oxygen plasma treatment. Plasma treatment with inert gas argon, at the given process parameters, was not enough effective to improve wettability of cotton fabrics. Contrary, the argon plasma treatments of lyocell and modal fabrics additionally improve their original high hydrophilic surface

Annual 2010/2011 of the Croatian Academy of Engineering

•

211

properties. It could be concluded that presented investigation results con¿rm that the plasma treatment is an acceptable and appropriate method of textile surface modi¿cation, avoiding ¿bre damage, and an eco-friendly method at the same time.

References 1. 2.

3. 4. 5.

6. 7. 8. 9.

10. 11. 12.

13. 14. 15. 16.

AATCC Test Method 79:2000 - Drop test Chen P, Wang J, Wang B, Li W, Zhang C, Li H, Sun B (2009) Improvement of interfacial adhesion for plasma-treated aramid ¿ber-reinforced poly(phthalazinone ether sulfone ketone) composite and ¿ber surface aging effects. Surface and Interface Analysis 41 (1) pp 38-43 Demir A, Özdoßgan E, Özdil N, Gürel A (2011) Ecological materials and methods in the textile industry: Atmospheric-plasma treatments of naturally colored cotton. Journal of Applied Polymer Science 119 (3) pp 1410-1416 Denes F, Young RA (1998) Surface modi¿cation of polysaccharides under coldplasma conditions. In: Dumitriu S (Ed.) Structural Diversity and Functional Versatility of Polysaccharides, New York, Basel, Hong Kong, Marcel Dekker Inc Ercegoviü Ražiü S, ýunko R, Bautista L, Mota J, Crespo L (2010) Ageing Effect on Wettability Properties of Low-Pressure Plasma-Treated Cellulosic Fabrics. In: Dragþeviü Z (Ed) Proceedings of the 5th International Textile, Clothing & Design Conference - Magic World of Textiles, Faculty of Textile Technology, University of Zagreb, Zagreb, pp 570-575 Ercegoviü Ražiü S (2010) Ciljana modi¿kacija svojstava tekstilija primjenom plazme i metalnih spojeva. PhD Thesis, University of Zagreb, Faculty of Textile Technology, Zagreb, Croatia Ercegoviü Ražiü S, ýunko R (2009) Modi¿kacija svojstava tekstilija primjenom plazme. Tekstil 58 (3) pp 55-74 Ercegoviü Ražiü S, ýunko R, Svetliþiü V, Šegota S (2011) Application of AFM Microscopy for Identi¿cation of Fibres Surface Changes after Plasma Treatments. Materials Technology (accepted for publishing) Ercegoviü Ražiü S, ýunko R, Bukošek V, Rolich T (210) Hydrophilicity Improvement of Cellulose Based Materials by Plasma. In: Proceedindgs of the 41th International Symposium on Novelties in Textiles, Faculty of Natural Sciences and Engineering, University of Ljubljana, Ljubljana, pp 344-350 Guimond S, Hanselmann B, Amberg M, Hegemann D (2010) Plasma processing of textiles: Perspectives. Melliand International 16 (4) pp 182-183 Hauser P J, El-Shafei A (2011) Atmospheric pressure plasma treatments for repellent textiles. AATCC Review 11 (1) pp 70-74 Horrocks AR, Nazaré S, Masood R, Kandola B, Price D (2011) Surface modi¿cation of fabrics for improved Àash-¿re resistance using atmospheric pressure plasma in the presence of a functionalized clay and polysiloxane. Polymers for Advanced Technologies 22 (1) pp 22-29 http://en.wikipedia.org/wiki/File:Plasma-lamp_2.jpg (visited 14.01.2011.) http://www.scientec.fr/include/surface/agilent/options/EN-mode_contact_friction. html (visited 14.01.2011.) Ibrahim NA, Hashem MM, Eid MA, Refai R, El-Hossamy M, Eid BM (2010) Ecofriendly plasma treatment of linen-containing fabrics. Journal of the textile Institute101 (12) pp 1035-1049 ISO 9073-6:200 Textiles – Test methods for nonwovens – Part 6: Absorption (Liquid

212

Annual 2010/2011 of the Croatian Academy of Engineering

wicking rate) 17. Wei QF (2005) Functinalization of textile ¿bres using plasma-based technology. Technical Textiles int. 2005 International Newsletter pp 27-29 18. Jimenez M, Bellayer S, Duquesne S, Bourbigot S (2010) Improvement of heat resistance of high performance ¿bers using a cold plasma polymerization process. Surface and Coatings Technology 205 (3) pp 745-758 19. Marchandalli B, Riccardi C (2007) Plasma treatments of ¿bres and textiles. In: Shishoo R (Ed) Plasma technologies for textiles Woodhead Publishing Ltd, Cambidge, pp 282298 20. Mehta R (2010) Plasma Treatment in the textile industry. Colourage 57 (7) pp 45-48 21. Morent R, De Geyter N, Vershuren J, De Clerck K, Kickenes P, Leys C (2008) Nonthermal plasma treatment of textiles. Surface &Coatings Technology 202 pp 3427 -3449 22. Nath K (2010) Atmospheric pressure plasma systems in textile application. International Dyer 195 (9) pp 25-31 23. Severich B (2008) Atmospheric pressure plasma - A new technology for modifying textile fabrics. Melliand International 14 (2), pp 120-121 24. Shahidi S, Rashidi A, Ghoranneviss M, Anvari A, Wiener J (2010) Plasma effects on anti-felting properties of wool fabrics. Surface and Coatings Technology 205 (SUPPL. 1) pp S349-S354 25. TEGEWA Bericht (1987) Tropftest - eine Methode zur schnellen Bestimung der Saugfähigkeit an textile Flächengebilden Melliand Textilberichte 68 pp 581-583 26. Tseng HJ, Hsu SH, Wu MW, Hsueh TH, Tu PC (2009) Nylon textiles grafted with chitosan by open air plasma and their antimicrobial effect. Fibers and Polymers 10 (1) pp 53-59 27. Vesel A (2008) XPS study of surface modi¿cation of different polymer materials by oxygen plasma treatment. Informacije MIDEM 38 (4) pp 257-265 28. Vesel A, Mozetic M, Strnad S, Peršin Z, Kleinschek S, Hauptman N (2009) Plasma modi¿cation of viscose textile. Vacuum 84 (1) pp 79-82

Annual 2010/2011 of the Croatian Academy of Engineering

213

Importance of salt content reduction in bakery products Žaneta Ugarčić-Hardi Josip Juraj Strossmayer University of Osijek, Faculty of Food Technology F. Kuhača 20, Osijek [email protected]

Abstract Studies have shown that the excessive intake of sodium/salt has a negative effect on health, primarily on the cardiovascular system.The average daily salt intakes of the European population are high (10-20 g salt/day) and surpass quantity of physiological requirements (5-6 g salt/day). Therefore, many international and national bodies set an aim to reduce the salt intake in diet. Since the average daily salt intake of Croatian population is 1316g/day, the Croatian Academy of Medical Science launched national programme for reducing salt intake. The ¿rst task of Croatian programme is to obtain the exact data on salt consumption. It is estimated that 70-75% of total salt comes from manufactured foods. Since the salt intake from bread is signi¿cant (~30%), the bakery industry is joining the national initiative to salt reduction. The aim of this paper was to estimate the salt content in 23 samples of commercial bread and 20 samples of different rolls taken from the market in Osijek-Baranja County (East Croatia). The obtained results showed that the salt content in bread varied in the range from 0.96 to 2.05%. This share is even higher in products strewed with salt (2.04 to extremely 4.76). Key words: salt reduction, bakery product, salt content in bakery product 1 Introduction Cardiovascular diseases including heart disease and stroke cause more than 50% of all deaths in the world. The major risk factors include high blood pressure, high cholesterol level obesity and smoking (Skupnjak, 2007, Reiner, 2008, Drenjanþeviü-Periü, 2010, EFSA 2005). High salt consumption can raise blood pressure and lead to cardiovascular diseases. Too much salt in food carriees also the risk of kidney disorders, osteoporoses or stomach cancer. Studies indicate that it is possible to lower blood pressure by cutting sodium/salt intake in diet. Salt and sodium are currently one of the top issues for healthy reformulation efforts by food and beverage manufactures (Heidolph et al, 2011). On a global basis, the population has a dietary intake of sodium that exceeds the required level. The current daily dietary guideline for sodium is 2,300 mg/day (5-6g/day salt) (IOM, 2004). This amount of salt intake should be much lower for children depending on age. In most countries the average daily sodium consumption has increased to 10-20 g salt (MacGregor, 2008). Males in general, consume more sodium than females (Dietary Guidelines Advisory Committee, 2010). In Croatian population the average daily salt intake accounts 13-16 g/day. Average daily intakes of salt are 13.3 ± 4.3 g for men and 10.2 ± 4.2 g for women (Premužiü et al, 2010, Miškulin et al, 2010).

214

Annual 2010/2011 of the Croatian Academy of Engineering

1.1 Salt sources

10% from total daily consumption of sodium is coming from natural food that contains salt in their origin (meat, ¿sh, eggs, vegetables). 15-20% salt in our daily diet is originates from cooked food or meal consumption (salt shaker). 70-75% of the sodium in the diet in Europe, 77% in the U.S. and 88% in Australia (Heidolph et al 2011), comes from processed foods („Hidden salt“), for which the consumers have little control. Salt intake from bread and bakery products, according to some investigations accounts for 30-35%. Table1. Share of salt intake from food Products

%

Bread and bakery products

34

Meat and meat products

28

Cheese, cream, eggs

10

Fish and ¿sh products

7

Milk and milk products

5

Fruits and products

5

Fat, confectionary, drinks

11

Source: Bundesforschungsanstalt für Getreide–und Kartoffelverarbeitung (1989)

As the cost of health care rises, with cardiovascular disease, health experts, health agencies, and medical societies are looking for ways of improving health of their nations. It is believed that reduction in sodium may help in diminishing health care costs by reducing the risks of cardiovascular-related diseases and other sodium-impacted illnesses, such as kidney damage, stomach cancer, and osteoporosis. Although the direct cause and effect with regard to high blood pressure and sodium is not clear, there are many international and national bodies that set an aim to reduce the salt intake in diet. Among them, Consensus Action on Salt and Health – CASH, provides recommendations, follows developments, looking for consensus among medical professionals and the public. The sodium-reduction strategies have been adopted in Finland, the United Kingdom, Ireland, France, the European Union, Switzerland, Canada and USA and they emerge globally (IOM 2004, EFSA 2005). Finland is the only country that has implemented a sodium-reduction policy that is considered to be “mandatory“. In 1993, Finland adopted a more aggressive sodiumreduction strategy which established mandatory labeling requirements for food products. Currently, the most comprehensive such “voluntary” program is being implemented in the United Kingdom, and employs a strategy consisting of: a) b) c)

programs designed to encourage the food industry to reformulate products to reduce sodium levels on a voluntary basis, a public relations campaign aimed at increasing public awareness concerning the relationship between sodium intake and HTN risk and ways to reduce sodium in take, and “voluntary” front-of-package “traf¿c light” labeling, which employs a colorcoded signaling system to inform consumers whether the sodium levels in food are deemed to be “high” (red), “medium” (amber), or “low” (green) by public health of¿cials (Heidolph et al, 2011)

Annual 2010/2011 of the Croatian Academy of Engineering

215

In the UK salt intake has already fallen from 9.5 to 8.6 g/day (i.e. 10% reduction) from when salt reduction ¿rst started in 2004 to the end of 2006 (MacGregor ,2008). In 2006 Croatian Action on Salt and Health (CRASH) was established by Croatian Academy of medical Sciences and an alliance of societies including Croatian Hypertension Society, Croatian Atherosclerosis Society and Croatian Cardiac Society, launching an initiative to reduce salt intake in Croatian population and developing a National program for salt intake reduction ( Reiner, 2008, Jelakoviü, 2008). Croatian Ministry of Health and Social Care supports these activities, and CRASH is included in the action of the World Health Organization on mapping sodium intake in European countries. Since the salt intake from bread is signi¿cant (~30%), the bakery industry is joining the national initiative for salt reduction. The bakery manufacturers have been the ¿rst that joined this action gaining information on the “Flour-Bread ‘07” Congress in October 2007 (Ugarþiü-Hardi et al, 2007). In order to encourage and educate the manufacturers to produce food and meals with low or reduced salt content, Croatian Food Agency, Faculty of Food Technology in Osijek, Medical Faculty and Croatian National Institute of public Health in Osijek have organized seminars for bakers and public. The levels of salt used in baking are variable and it is very dif¿cult to obtain absolute levels of salt in bread. The ¿rst task of the Croatian programme is to obtain the exact data on salt consumption in different parts of Croatia (Reiner, 2008). Therefore, the aim of this paper was to estimate the salt content in bakery products in Osijek-Baranja County in Croatia. 2 Why is salt added to bread dough? Salt content in bakery products varies from country to country from 1 – 2%. In Croatia salt addition amounts from 2 to 2.5%, (a20g/kg Àour = 13g/kg bread; 5g Na/kg bread). The primary reason to salt addition is to improve the taste. Unsalted bread is tasteless. Salt has also a signi¿cant technical effect on the properties of wheat gluten by strengthening the gluten network in dough. It inÀuences gluten development, the rheology of dough and the speed of fermentation. (Slumier, 2005). Salt increases the resistance, extensibility and elasticity of the wheat gluten (Hlinka, 1962, He et al, 1992, Fisher et al, 1994.). Salt decreases the amount of water required to produce dough of ¿xed consistency (Cauvain and Young, 1998). The investigation conducted by the German Research Center for Food Chemistry and the Technical University of Munich shows that salt appears to increase the osmotic pressure leading to a reduction in yeast activity and a decrease in the crumb pore size. Salt also increased the mixing time and dough resistance. A maximum loaf volume was achieved at 0.5% salt per 100 g of Àour. Sensory considerations show that salt affects the overall Àavor pro¿le of wheat bread. Sodium recognition strongly depends on salt content (Heidolph et al, 2011). Although the salt addition has an impact on sensory and technological properties, investigations showed that salt addition in bakery products can be reduced up to 25% without negative effect on quality (Forschung hilft dem Backgewerbe, 1989). Currently the investigation shows that bread and bread products can be manufactured satisfactorily at levels of 1% salt on a Àour basis. (Linko et al, 1984, EFSA, 2005). 2.1 Sodium replacement

Replacement of sodium chloride is the next best method to reduce salt in a product. Cations similar to sodium such as potassium work best. However, a bitter or metallic off-Àavor is often sensed when potassium chloride is used. Other mineral salts can partially replace

216

Annual 2010/2011 of the Croatian Academy of Engineering

salt in a cost-effective manner as well, but, they often deliver off-Àavors which can limit their use. Replacement with NH3Cl tastes of plant licorice, and replacement with CaCl has a negative effect on dough rheology. Although arti¿cial salt does not exist, the most common strategy is to replace sodium salts by other inorganic salts such as potassium chloride. It has been demonstrated that replacing 20% of sodium by potassium results in bread with an accept-able taste, whereas 40% replacement results in unacceptable off-taste (Salovaara, 1982). As a consequence, sodium replacement is limited. With the growing market of salt reduced food products, there is also an increasing demand for salt substitutes or salt taste enhancing agents. In the recent years a lot of commercial mixtures of NaCl and KCl can be found on the market. Such a commercial products are “Pansalt”, “Sub4salt”and “Sanusol”. They should give 25-50% reduction in sodium. 3 Material and methods The salt content was estimated in different bakery products taken from the market in “Osijek-Baranja” County, (East Croatia): 23 samples of commercial bread and 20 samples of different rolls strewed with salt. Chloride ion was determined by Volhard’s titration. This method determines the salinity of foods based on the concentration of the chloride ion titrated with silver nitrate solution. The analyses were provided in the Croatian National Institute of Public Health in Osijek and the Faculty of Food Technology in Osijek. 4 Results and discussion Table 2. Salt content in different bread types Bread types Wheat French bread (White baguette) Wheat Easter braided bread “Sovital” bread Wheat water bread “Krunovit” bread “Drava vital” bread Family bread “Monastery sun” bread I “Monastery sun” bread II “Grandmother mix” bread Bread with onion Corn baguette White bread with inulin Extra white bread White ciabatta White bread I Roll bread White bread II White bread III White bread IV Brown bread Baguette Water bread Average

Salt content (%) 1.75 1.41 1.79 1.49 1.57 1.31 1.30 1.90 2.01 1.01 0.96 1.76 1.29 1.52 1.73 1.58 2.02 1.58 1.03 2.05 1.61 1.84 1.29 1.56

Annual 2010/2011 of the Croatian Academy of Engineering

217

The obtained results showed that the salt content in bread in the Osijek-Baranja County varied in the range from 0.96 to 2.05%, (Table 2). The average is 15.6 g /kg bread, which is higher than the average value for salt content in bread in EU. It means that eating only 2 slices of bread (a100 g), salt intake amounts to 1.56 g. According to some investigations the most frequently consumed type of bread is the white wheat one (>4 slices), (PucarinCvetkoviü, et al, 2008). This share of salt is even higher in products strewed with salt (from 2.04 to extremely 4.76%, average value 2.84), (Table 3). Consumption of 1 roll bread (a70 g) gives the salt intake amounts of 2.09 g, which is 1/3 of the total recommended daily intake (5-6 g). Most consumers of those products (breakfast rolls) are younger, who get a habit of eating salty food from childhood The investigation of the diet habits of school children in Osijek has shown that 32.0% of school children population has a habit to eat daily some of the bread products, 28.3% (154/545) girls and 35.9% (191/532) boys. Sample was strati¿ed by settlements population, size 1077 pupils, 545 girls and 532 boys (Miškulin et al, 2009). The high salt intake in childhood leads to higher blood pressure and an increased risk of stroke and heart disease in later life. Table 3. Salt content in rolls strewed with salt Type of bakery products strewed with salt Finger stick (Baguette ¿ngers) Pretzel Salt stick Finger stick (Baguette ¿ngers) Pretzel Salt stick Roll (Baguette roll) Finger stick (Baguette ¿ngers) Pretzel Salt stick Salt stick Pretzel Finger stick (Baguette ¿ngers) Roll (Baguette roll) Pretzel Finger stick (Baguette ¿ngers) Salt stick Roll (Baguette roll) Pretzel Roll (Baguette roll) Average

Salt content (%) 4.76 3.04 5.98 2.20 2.41 2.60 2.04 2.08 2.14 2.51 2.28 3.44 2.45 2.40 4.57 2.17 2.53 2.34 2.26 2.71 2.84

Very high salt content was determined in snack products, up to 5%. Average salt content of these products is 2.81%. Eating 100 g (1 package) of such products, 1/2 of total salt quantity of physiological requirements will be taken, without being conscious of that fact (Ugarþiü-Hardi et al, 2010). Therefore, it is important to raise consumer awareness abut the association between salt consumption and health in order for people to assume more responsibility for themselves. One of the preconditions for this is the improved labelling of foods with nutrition and health information, including details of salt content. Salt labelling

218

Annual 2010/2011 of the Croatian Academy of Engineering

should detail the salt content per serving with the recommended intake per day. Some products in Croatia are labelling on this way. Fig.1.

Fig.1. Recommended daily intake of nutrition’s and percentage of each nutrition in product

The bakery and other food industries are invited to support the national action on salt reducing. The bakery industry is recommended to provide that gradually in two steps. The ¿rst step should be 10% salt reduction, which means addition of 1.8 g salt/100 g Àour. In the second step the reduction should be up to 25% (1.5 g salt/100 g Àour). 25% salt reduction would not affect the taste and technical properties. This target could be achieved across the next three years. 5 Conclusion The results on salt content in different bakery products show that reduction on salt amounts is necessary. The salt content in bread in the Osijek-Baranja County varied in the range from 0.96 to 2.05%, and in special products from 2.04 to extremely 4.76%. Although the salt addition has an impact on sensory and technological properties, investigation showed that salt addition in bakery products can be reduced up to 25% without negative effect on quality. It may be possible to set an average target of salt reduction in bakery products gradually from 10 to 25% in the next three years, with the aim of getting used the food with less salt, which should contribute to long-term prevention of cardiovascular diseases. Discussions with the bakery products industry and the government would be needed to de¿ne a way to achieve the reduction in salt level consistent with technological and sensory needs. 6 References 1. 2. 3. 4. 5. 6.

Cauvain SP, Young LS (1998): Technology of Breadmaking. Blackie Academic and Profesionals, London Dietary Guidelines Advisory Committee Draft Report (2010) Drenjanþeviü-Periü I (2010) Prekomjeran unos soli kao þimbenik rizika za razvoj kardiovaskularnih bolesti, Znanstveno - struþni skup „Sol i zdravlje“, Faculty of Food Technology in Osijek, Osijek, 28. 1. 2010. EFSA- European Public Health Alliance (2005) EFSA highlights the dangers of eating too much salt, Published online at www.epha.org/a/1912./ Accessed 3 March 2011 Fisher MH, Aitken TR, Anderson JA (1994) Effects of mixing, salt and consistency on extensograms. Cereal Chemistry 26, 81–97. Forschung hilft dem Backgewerbe (1989): Nutzen und Möglichkeiten einer natriumreduzierte Ernährung, Bundesforschungsanstalt für Getreide–und Kartoffelverarbeitung,, Detmold, pp 10 – 23.

Annual 2010/2011 of the Croatian Academy of Engineering

7. 8. 9. 10. 11. 12. 13. 14. 15. 16.

17. 18. 19. 20. 21. 22. 23.

24.

219

FSAI (2010) Salt and health: Review of the scienti¿c evidence and recommendations for public policy in Irland He H, Roach RR, Hoseney RC (1992) Effect of nonchaotropic salts on Àour breadmaking properties. Cereal Chemistry 69 (4), 366–371 Heidolph B B, Ray DK, Koehler P, Weiber, Slocum S, Noort MW (2011): Looking for my lost shaker of salt…replacer: Àavor, funktion, future, Cereal, Food World, Vol. 56, No.1, 5-19. Hlynka I (1962) InÀuence of temperature, speed of mixing, and salt on some rheological properties of dough in the farinograph, Cereal Chemistry 39, 286–303. IOM-Institute of Medicine (2010) Strategies to reduce sodium intake in the United States, Appendix C: International efforts to reduce sodium consumption IOM- Institute of Medicine (2004) Dietary reference intakes for water, potassium, sodium, chloride and sulfate, United States Institute of Medicine. Jelakoviü B (2008) Salt and hypertension, Znanstveni skup, Kardiovaskularno zdravlje, Prehrana i sol, Akademija medicinskih znanosti, 31, Zagreb, 21.11.2008, Abstract, 31 Linko P, Harkonen H, Linko YY (1984) Effects of sodium chloride in the processing of bread baked from wheat, rye and barley Àours. Journal of Cereal Science 2 (1), 53–62 MacGregor GA (2008) Salt-from evidence to implementation, Znanstveni skup, Kardiovaskularno zdravlje, Prehrana i sol, Akademija medicinskih znanosti, 3, Zagreb, 21.11.2008. Abstract, 3 Miškulin M, Periš D, Ugarþiü-Hardi Ž, Dumanþiü G (2010) Skrivena sol u prehrani djece školske dobi - ima li mjesta zabrinutost?, 34. struþni skup s meÿunarodnim sudjelovanjem, “Zdravstvena ekologija u praksi”, Zagreb, 21.-23. 4. 2010 Proceedings, pp 103-109 Premužiü V, Erceg I, Jovanoviü A, Reiner Ž, Jelakoviü B (2010): Unos soli u odrasloj populaciji, Hrvatski þasopis za javno zdravstvo HCJZ, On.line 6, (21), Pucarin-Cvetkoviü J, Polašek O, Kern J, Vuletiü S, (2008) Regional features of Croatian nutrition, Znanstveni skup, Kardiovaskularno zdravlje,Prehrana i sol, Akademija medicinskih znanosti, Zagreb, 21.11.2008. Abstract, 8 Reiner Ž (2008) National programme for the reduction of salt intake in Croatia,Znanstveni skup, Kardiovaskularno zdravlje, Prehrana i sol, Akademija medicinskih znanosti, Zagreb, 21.11.2008. Abstract, 5. Salovaara H (1982) Sensory limitations to replacement of sodium with potassium and magnesium in bread. Cereal Chemistry 59(5):427 Skupnjak B(2007): Inicijativa za nacionalnu kampanju smanjenja utroška soli, Hrvatski þasopis za javno zdravstvo HCJZ, Online, Vol 3, 10. Slumier P (2005) Principles of breadmaking, Amercan Association of Cereal Chemists, Inc. St. Paul, Minnesota, USA, pp46-47. Ugarþiü-Hardi Ž (2007) Salt reduction in Croatian baked products: Yes or no? 4th International Congrss „Flour-Bread Ғ07“ and 6th Croatian Congress of Cereal Technologists, Faculty of Food Technology in Osijek, Josip Juraj Strossmayer University of Osijek, Opatija, October 24-27, 2007, Abstract, 31 Ugarþiü-Hardi Ž, Dumanþiü G, Pitlik N, Koceva Komleniü D, Jukiü M, Kuleš A, Sabo M (2010) The salt content in bakery products in Osjecko-baranjska County, Proceedings of 5th International Congress „Flour-Bread Ғ09“ and 7th Croatian Congress of Cereal Technologists, Ugarþiü-Hardi Ž., Faculty of Food Technology Osijek, Josip Juraj Strossmayer ty of Osijek, ty of Osijek, 551-556.

Podupirajuüi þlanovi: Arhitektonski fakultet, Zagreb, www.arhitekt.hr Centar za vozila Hrvatske d. d., www.cvh.hr Elektrotehniþki fakultet, Osijek, www.etfos.hr Energetski institut Hrvoje Požar, Zagreb, www.eihp.hr Fakultet elektrotehnike i raþunarstva, Zagreb, www.fer.unizg.hr Fakultet elektrotehnike, strojarstva i brodogradnje, Split, www.fesb.hr Fakultet kemijskog inženjerstva i tehnologije, Zagreb, www.fkit.hr Fakultet prometnih znanosti, Zagreb, www.fpz.hr Fakultet strojarstva i brodogradnje, Zagreb, www.fsb.hr Geodetski fakultet, Zagreb, www.geof.hr Graÿevinski fakultet, Osijek, www.gfos.hr Graÿevinski fakultet, Rijeka, www.gradri.hr Fakultet graÿevinarstva, arhitekture i geodezije, Split, www.gradst.hr Graÿevinski fakultet, Zagreb, www.grad.hr Gra¿þki fakultet, Zagreb, www.grf.hr Hrvatska elektroprivreda d. d., Zagreb, www.hep.hr Hrvatska zajednica tehniþke kulture, Zagreb , www.hztk.hr Kemijsko-tehnološki fakultet, Split , www.ktf-split.hr Konþar - Institut za elektrotehniku d. d., Zagreb , www.koncar-institut.hr Odašiljaþi i veze d.o.o., Zagreb, www.oiv.hr Pliva Hrvatska d.d., Zagreb, www.pliva.hr Pomorski fakultet, Rijeka, www.pfri.hr Prehrambeno-biotehnološki fakultet, Zagreb, www.pbf.hr Prehrambeno-tehnološki fakultet, Osijek, www.ptfos.hr Rudarsko-geološko-naftni fakultet, Zagreb, www.rgn.hr Strojarski fakultet, Slavonski Brod, www.sfsb.hr Supra Net, d.o.o., www.supranet.hr Sveuþilište u Dubrovniku, www.unidu.hr Sveuþilište J. J. Strossmayera u Osijeku Tehniþki fakultet, Rijeka, www.riteh.hr Tehniþko veleuþilište u Zagrebu, www.tvz.hr Tekstilno-tehnološki fakultet, Zagreb, www.ttf.hr Uljanik brodogradilište d. d., Pula, www.uljanik.hr