annual 2010/2011 of the croatian academy of engineering - HATZ

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 Kai St., 10000 Zagreb, Croatia 

Editor-in-Chief 

Prof. Vilko Žiljak, Ph.D., 

Vice president of the Croatian Academy of Engineering 

Editorial Board 

Prof. Stanko Tonkovi, 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. Pribievi, D. Medak, A. apo 

GEODETIC CONTRIBUTION TO THE GEODYNAMIC RESEARCH 

OF THE AREA OF THE CITY OF ZAGREB .................................................... 11 

Drago Katovi, Andrea Katovi, Marija Krnevi 

SPANISH BROOM (SPARTIUM JUNCEUM L.) .............................................. 23 

Matanovi Davorin, Moslavac Bojan, Nediljka Gaurina-Meimurec 

THE WAY TO REDUCE PIPE WEARING WHILE DRILLING ...................... 38 

Dinko Mikuli, Vladimir Koroman 

FUNDAMENTALS OF MILITARY PRODUCTION DEVELOPMENT IN 

THE CROATIAN WAR FOR INDEPENDENCE FROM 1991 TO 1993 .......... 46 

Darko Stipaniev 

INTELLIGENT FOREST FIRE MONITORING SYSTEM – FROM 

IDEA TO REALIZATION ................................................................................... 58 

Igor Petrovi, Davorin Kovai 

LABORATORY TESTING AND NUMERICAL MODELLING OF MBT 

WASTE DEFORMABILITY ............................................................................... 74 

Darko Dujmovi, Boris Androi, Ivan Lukaevi 

BEAM-TO-COLUMN JOINT MODELLING TO EC3 ...................................... 89 

Darko Dujmovi, Boris Androi, Josip Piskovic 

MODELLING OF JOINT BEHAVIOUR IN STEEL FRAMES ...................... 109 

Juraj Božievi, 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 Crnkovi, Darko Stipaniev, Maja Štula, Juraj Feljan, Luka 

Lednicki, Josip Maras, Ana Petrii 

DICES: DISTRIBUTED COMPONENT-BASED EMBEDDED 

SOFTWARE SYSTEMS ................................................................................... 154 

Martin Žagar, Hrvoje Mlinari, Josip Knezovi 

FRAMEWORK FOR 3D MOTION FIELD ESTIMATION AND 

RECONSTRUCTION........................................................................................ 168 

Gaurina-Meimurec Nediljka, Pašic Borivoje, Matanovi Davorin 

LABORATORY EVALUATION OF MUD DIFFERENTIAL STICKING 

TENDENCY AND SPOTTING FLUID EFFECTIVENESS ............................ 184 

Ružica unko, Sanja Ercegovi Raži 

USE OF PLASMA TECHNOLOGY FOR MODIFICATION 

OF TEXTILES ................................................................................................... 199 

Žaneta Ugari-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 scientic 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 Josipovi, 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 

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 Trafc 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 Ruer Boškovi 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 Kai St., Zagreb. 

Bulletin of the Academy in Croatian and English "Tehnike znanosti / Engineering 

Power vol. 10 (1) 2011 was published. 

2012 

Bulletin of the Academy in Croatian and English "Tehnike znanosti / Engineering 

Power vol. 10 (special issue) 2011 entirely devoted to Ruer Boškovi and 300th 

anniversary of his birth was published. 

In March 2012 Croatian Academy of Engineering signed Agreement on Scientic 

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 Scientic Committee for Agriculture and Forestry of the Croatian 

Academy of Sciences and Arts and Biotechnical Center of the Croatian Academy 

of Engineering, a scientic conference "Food as the Foundation of Health and 

Longevity" was held on June 18th 2012.


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. Peri, 

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 


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 reect upon important 

development issues. 

Information about the events associated with the Academy could be found on our 

web site: www.hatz.hr


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 densied 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 scientic 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 coefcient 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 scientic insight about geodynamics of the City of Zagreb area. The 

results of this scientic 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, 

(Prelogovi, Cvijanovi 1981) (Kuk et al. 2000), (Kuk et al. 2000a), (Tomljenovi, 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” (oli et al. 1999) (Medak, 

Pribievi 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 specic 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, Pribievi 2001), (Medak, Pribievi 2002), (Medak,

12 


Pribievi 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 (Pribievi 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 scientic software GAMIT/GLOBK show signicant movement on GPS points 

as a result of geodynamic activity in the research area (Medak, Pribievi 2004), (Medak, 

Pribievi 2006), (Pribievi et al. 2007), (apo 2005), (apo 2009). From the analysis results 

the geodetic model of tectonic movements has been created and scientic comparison 

with geologic model created on the basis of age-long research. The correlation coefcient 

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 (apo et al. 2009), (apo 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 signicant 

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.


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 oor 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. (apo 2005) 

Figure 3. Example of cracks due to movements in the zone of Zagreb fault - Vidovec – cracks 4 

cm wide (apo 2009)

14 


Figure 4. a) Donja Planina, Planinarska Street 21. Whole house has been cut with the crack of 

NNW-SSE direction. On the facade a signicant 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 (apo 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 (apo 2009). 

Figure 6. a) In Gornji Mikulii 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 (apo 2009).


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 (apo 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 km 2 . 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 (apo 2005)

16 


Thanks to the signicant 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 (Solari, 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 (oli et al 1999), (Medak, 

Pribievi 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, Pribievi 2002) 

(Pribievi 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 densied 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, 

Mikulii, Vugrovec and Kašina. In year 2005, ve new geodynamic points were established 

in that part of Zagreb (Medak et al. 2007a).


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) 

(apo 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 densication points 

2005, 2007 (11 and 21 points) (Pribievi et al. 2007) (apo 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 


Figure 11: Velocity rates and their error ellipses for period 2006-2007 (apo 2009), (Wessel 2004) 

Table 2: Statistical representation of absolute values of velocity rates for 2006–2007 in mm/yr. 

v 

v 

v hz 

mm/yr 

v H 

mm/yr 

min. 0,2 0,57 1,43 0,60 

max. 10,70 19,36 19,37 50,27 

avg. 3,14 4,86 6,43 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. 

v v 

v hz 

v H 

mm/yr mm/yr 

min. 0,03 0,04 0,12 0,02 

max. 3,93 3,42 4,34 17,48 

avg. 0,83 1,03 1,45 1,97


Figure 13: Geodetic velocity model created using IDW interpolation 

with respect to faults model 

(1) 

Table 4: Correlation coefcient by Spearman formula 

Calculating of correlation coefcient by above given formula (1) for Spearman rank correlation 

we get rs=0.94 with 1% signicance level (or 99% in Gaussian model). 

The correlation coefcient between geodetic and geologic model shows high degree of correlation 

thus giving credibility to both methods of research (apo 2009).

20 

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 scientic GAMIT/GLOBK software, designed at MIT 

specically 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 signicant 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 coefcient has been calculated 

for the geodetic and geological model of the geodynamic movements, by using 

Spearman correlation coefcient formula. The resulting value of 0.94 with signicance 

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 dene 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 dene the zones of landslides 

caused by tectonic movements. 

REFERENCES 

1. Altiner, Y. (1999). Analytical Surface Deformation Theory for Detection of the Earth’s 

Crust Movements. Springer Verlag. 

2. Altiner, Y., Marjanovi-Kavanagh, R., Medak, D., Medi, Z., Prelogovi, E., Pribievi, 

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.


Reports on Geodesy (p. 225-229). Nice, France: Warsaw University of Technology. 

3. 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. 

4. 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 

5. oli, K., Prelogovi, E., Pribievi, B., Švehla, D. (1999). Hrvatski geodinamiki 

projekt CRODYN i GPS mreža Grada Zagreba. In A. Baji (Ed.), Znanstveni skup 

Andrija Mohorovii - 140. obljetnica roenja: zbornik radova (p. 141-152). Zagreb: 

Državni hidrometeorološki zavod. 

6. Dong, D., Herring, T., & King, R. (1998). Estimating regional deformation from a 

combination of space and terrestrial geodetic data. Journal of Geodesy, 72 (4), 200- 

214. 

7. apo, A. (2005). Obrada i interpretacija geodetskih mjerenja na geodinamikoj mreži 

Grada Zagreba. magistarski rad. Zagreb: Geodetski fakultet, 12.05. 2005., 174 str. 

Voditelj: Medak, Damir. 

8. apo, A (2009): Korelacija geodetskog i geološkog modela tektonskih pomaka na 

primjeru šireg podruja Grada Zagreba. doktorska disertacija. Zagreb: Geodetski 

fakultet, 08.05. 2009, 199 str. Voditelj: Pribievi, Boško. 

9. apo, Almin; Pribievi, Boško; Medak, Damir; Prelogovi, 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. 

10. 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 . 

11. Gerasimenko, M. D., Shestakov, N. V., Teruyuki, K. (2000). On optimal geodetic network 

design for fault-mechanics studies (Vol. 52). Earth Planets Space. 

12. Herring, T., King, R., McClusky, S. (2006a). Documentation for the MIT Global Kalman 

lter VLBI and GPS analysis program: GLOBK 10.3. Cambridge. 

13. Herring, T., King, R., McClusky, S. (2006b). Documentation for the MIT GPS analysis 

software: GAMIT 10.3. Cambridge. 

14. Kuk, V., Prelogovi, E., Dragievi, I. (2000). Seismotectonically Active Zones in the 

Dinarides. Geol. Croatica, 53 (2), 295-303. 

15. Kuk, V., Prelogovi, E., Sovi, L., Kuk, K., Mari, K. (2000a). Seizmološke i seizmotektonske 

znaajke šireg zagrebakog podruja. Graevinar, 52 (11), 647-653. 

16. Medak, D., Pribievi, 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. 

17. Medak, D., Pribievi, B. (2002). Geodinamika mreža Grada Zagreba. In T. Baši 

(Ed.), Zbornik Geodetskog fakulteta Sveuilišta u Zagrebu povodom 40. obljetnice 

samostalnog djelovanja 1962-2002. pp. 145-156. Zagreb. 

18. Medak, D., Pribievi, B., Prelogovi, 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 podroja geodezije in geozike. p. 57-64. Ljubljana: 

Slovensko združenje za geodezijo in geoziko. 

19. Medak, D., Pribievi, B. (2003). Processing of Geodynamic GPSNetworks with 

GAMIT Software. Reports on Geodesy, Warsaw University of Technology, 64 (1), 

75-84.

22 


20. Medak, D., Pribievi, 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. 247- 

256). Veszprem, Hungary: Springer. 

21. Medak, D., Pribievi, 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., Pribievi, B., apo, A. (2007a). Progušenje toaka Geodinamike mreže 

Grada Zagreba u podsljemenskoj zoni. Geodetski list , 61(84)(4), 247-258. 

23. Medak, D., Pribievi, B., Prelogovi, E., apo, A. (2007b). Primjene geodetskogeodinamikih 

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. Prelogovi, E., Cvijanovi, D. (1981). Potres u Medvednici 1880. godine. Geol. vjesnik 

(34), 137-146. 

26. Pribievi, B., Medak, D., Prelogovi, E. (2004). Geodinamika prostora Grada Zagreba. 

Geodetski list , 58(81)(1), 51-65. 

27. Pribievi, B., Medak, D., Prelogovi, E., apo, A. (2007). Geodinamika prostora 

Grada Zagreba. Zagreb: Geodetski fakultet Sveuilište u Zagrebu. Znanstvena monograja. 

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. Solari, M. (1999). Suradnja srednje europskih zemalja u geodeziji i geodinamici. 

In A. Baji (Ed.), Znanstveni skup Andrija Mohorovii - 140. obljetnica roenja: 

zbornik radova (p. 165-177). Zagreb: Državni hidrometeorološki zavod. 

31. Šoši, I. (2004). Primijenjena statistika. Zagreb: Školska knjiga, Udžbenici Sveuilišta 

u Zagrebu. 

32. Tomljenovi, B. (2002). Strukturne znaajke 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.).


Drago Katović, Ph.D. 

Spanish Broom (Spartium Junceum L.) 

University of Zagreb 

Faculty of Textile Technology 

Zagreb, Prilaz baruna Filipovića 28A 

e-mail: dkatovic@ttf.hr 

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: katovic@unical.it) 

Marija Krnčević B.Sc. of Ethnology and Archeology, 

Museum of Šibenik, Gradska vrata 3 

marija.krncevic@muzej-sibenik.hr) 

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 benets 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 ax and hemp have mostly been used as textile raw material of cellulosic origin

24 


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


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 


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);


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 ower 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,… ores apibus gratissimi (Spanish Broom is used for tying and making crumples,… 

and bees often land on its owers). 

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-sufciency, 

reached 700,000 tones. It was planted on 300,000 hectares and there were 61 processing 

factories [2].

28 


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 

signicatu. 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 ax 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 ax 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 ax and Spanish Broom. Provisions 

for the protection of useful plants were prescribed quite often. One of the examples 

is the Statute of Bra 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 rened 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 Rajkovi 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, benet 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


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. Ani 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 Premuži 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 Rajkovi, 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 Zakuac 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 

oating 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 unprotability. 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 


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 Petriani 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


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 


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 .


Figure 14 - 20 SEM images of the fuse (variations in zoom)

34 


The images show bers covered with ashes, but not completely burned (grey part of the 

sample). Otherwise, they would not be identiable 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 3 rd or 4 th century AD in our parts: 

cotton, ax, 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 insufcient for a denite 

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 19 th 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 difcult 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 scientic studies


relating to the production of Spanish Broom as raw material for composite materials [6- 

8]. The advantage of Spanish Broom over ax 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 ax and hemp demand high quality soil each 

year. Its cultivation on rocky areas would have several benets. 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 signicantly 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 signicant 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. 

Cotton Flax Hemp Spanish broom 

Tensile strength (MPa) 300-700 500-900 310-750 750 

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 


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, 

ax 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 ax 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 CH 2 

and CH groups of 

long alkyl chains of waxes, are present in the spectra of ax 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.


Figure 25 FTIR spectrum of bers obtained from Spanish Broom applying four different procedures 

compared to the cotton ber. 

REFERENCES 

1. Kovaevi Z., Krnevi M., Katovi A. Katovi 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. Stojanovi A.: Brnestra (Žuka- Spartium Junceum) (1962) Etnološki Zavod Filozofskog 

fakelteta Sveuilišta u Zagrebu, 4, 3-45 

4. Katovi D., Katovi A., Krnevi 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 efcient 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 978-953- 

307-194-7

38 


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: dmatan@rgn.hr 

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 coefcient is essential in dening the project technical 

feasibility. As the part of drilling system, drilling uid can impact on the reduction of friction 

coefcient by implementation of lubricants. Selection of lubricants, their compatibility 

with drilling uid 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 difculties at logging jobs. From that 

reason, the principle factors that require extra consideration for choosing a uid for drilling 

a horizontal well are: formation damages, hole cleaning, hole stability and lubricity. Good 

lubricating properties of the drilling uid can improve greatly drilling operation efciency. 

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 difcult to detect what is causing increased down 

hole friction. 

Regardless of the drilling uid or lubricant used, down hole friction can be reduced by


conditioning the uid 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 uid 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 efciency. The method must be systematic, logical and must have quantiable 

approach. The applicable method is one that denes and quanties 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 inuence 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 inuence in drilling process. Bit design and used down hole motor combination 

efciency according to observed formation characteristics are essential. To obtain optimal 

performance of bit-motor combination optimal ow rate must be obtained and proper uid 

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 


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:


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: (F b,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 

uid, (r) is radial clearance between well bore and tubular, () is coefcient of friction and 

(H) is vertical well bore depth. 

The compressive axial load at the end of the build section (EOC) is: 

(2) 

Where: (F eoc 

) is axial load at the end of build section, (F kf) 

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: (L h 

) is horizontal section length. 

It is obvious that coefcient 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 benets. 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 coefcient of friction and normal force: 

Where: (F t 

) is axial tension acting at lower end of element, ( ) is average inclination 

angle of element, (F n 

) is normal force acting on element, (M) is increase in torsion over 

length of element, and (r p 

) 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. 

(4) 

(5)

42 

LUBRICANTS 


Friction reduction can be achieved by using a drilling uid with good lubricating qualities 

such as oil-base uids, polymer uids, and glycol-derivative systems [4]. Oil-based uids 

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 uid 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 emulsiers, antifoaming 

agents and for warehouse purposes pour point depressants. 

Lubricants are designed to reduce torque to increase horsepower at the bit by reducing the 

coefcient of friction. In water based drilling uids certain oils, graphite powder, glycols 

and soaps are used for this purpose. To be economic, a drilling uid 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 uid, lubricants: 

• can coat metal surfaces, reducing the adhesion of steel to the lter cake; 

• can be incorporated into the lter cake and provide better uid-loss control (resulting in 

thinner cakes); and 

• can be incorporated into the lter cake to reduce the yield stress of the cake. 

In many uid 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 uids were carried out. They are water-dispersible lubricants designed to 

decrease the coefcient 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]. 

Inuence of addition different concentrations (0%, 2%, 3%) of three lubricants on rheological 

and ltration properties in different drill-in uids were carried out according API 

standard using Fann viscometer and API lter press. The three unweighted, water based


drill-in uid systems were tested. Fluids are designated as WBM 1, WBM 2 and WBM 3 

(Table 1). 

Lubricating properties were studied using the Lubricity Tester. Briey 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 uid 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 coefcient 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 inuence on plastic 

viscosity and yield point in comparison to uid without lubricants. For all three lubricants 

there is an signicant reduction of coefcient of friction (COF). Fig. 3 is an illustration of 

the inuence of lubricant concentration on COF in selected water based muds.

44 


Fig. 3. Efciency of selected lubricants in mud WBM 1 

The comparison of different lubricants inuence 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. 

Conclusions 

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

All used lubricants have certain but not important impact on rheological and ltration properties 

of tested uids, 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


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 

F b,max 

- maximal axial compressive load at the bottom of vertical section, N 

F eoc 

-axial load at the end of build section, N 

F n 

- axial load acting on element, N 

F kf 

- axial load at the kickoff point, N 

F t 

- axial tension acting at lower end of element, N 

H - vertical well bore depth, m 

I - moment of inertia of tubulars, m4 

L h 

- horizontal section length, m 

M - increase in torsion over length of element, N•m 

r - radial clearance between well bore and tubulars, m 

r p 

- characteristic radius of drill string element, m 

R - radius of well bore curvature, m 

w e 

- tubular weight in uid, N/m 

- friction factor 

- Ludolf’s number, 3,14 

References 

1. 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 

2. 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 

3. Johancsik CA, Friesen DB, Dawson R. (1984) Torque and Drag in Directional Wells- 

Prediction and Measurement, Journal of Petroleum Technology, June, p. 987-992 

4. 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 22- 

24, p. 401-410 

5. Bol GM. (1986) Effect of Mud Composition on Wear and Friction on Casing and Tool 

Joints, SPE Drilling Engineering, October, p. 369-376 

6. Evans N, Langlois B, Audibert-Hayet A, Dalmazzone C, Deballe E. High Performance 

Emulsiers for Synthetic Based Muds, SPE 63101, SPE Annual Technical Conference 

and Exhibition, Dallas, TX, October 1-4, p. 1-11 

7. Gaurina-Meimurec N. (1998) Horizontal Well Drill-In Fluids, The Mining, Geological, 

Petroleum Engineering Bulletin, vol. 10, Zagreb, Croatia, p. 73-76

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: dinko.mikulic@vvg.hr 

Tel. 01/6230-761, Mob. 099 6222-503 

Vladimir Koroman 

Brodarski institut d.o.o., 

Av. V. Holjevca 20, 10020 Zagreb, Croatia 

e-mail: vladimir.koroman@hrbi.hr 

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 Scientic 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 fullled 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.


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 ofces 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 rnomerec 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 Scientic Research. President of the Republic of Croatia Franjo 

Tuman, PhD, issued the Decision on the Appointment of Gojko Šušak as the Defence 

Minister on 16 September 1991, while Croatian Prime Minister Franjo Greguri, PhD, 

passed the Decision on the Appointment of Vladimir Volarevi, 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 

Scientic 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 rnomerec in Zagreb and experts from the Army 

Depot Mainentance Bregana, who as volunteers enlisted for Croatian military service, and 

who were competent for specic eld of military equipment and armaments development. 

These were the following Croatian army ofcers: 

Dinko Mikuli, MSc, motor vehicles; Zdenko Matijaši, MSc, motor vehicles; Vjekoslav 

Stojkovi, PhD, motor vehicles; Mijo Vrhovski, PhD, motor vehicles; Simeon Kovaev, 

PhD, motor vehicles; Nikola Gambiroža, PhD, explosives; Mirjana Gambiroža Juki, 

PhD, explosives, Milan Prpi, MSc, explosives, Mladen Pleše, PhD, explosives; Mladen 

Barkovi, PhD, electronic devices; Artur Gorianin, MSc, electronic devices; Aljoša 

Božikovi, MSc, radars; Mirko Jukl, MSc, radars; Mihalj Strmeki, MSc, air defence; 

Vladimir Lebinac, MSc, electronic devices, Goran Prokopec, MSc, telecommunication; 

Ivan Pokaz, MSc; Mirko Kukolj, MSc, armaments; Milan Ivanuševi, MSc, NBC equipment; 

Josip Petrovi, electronic devices; Josip Vinter, overhaul; Vinko Turk, artillery 

weapons; Ivan Peica, air defence, Vitoš Bebi, engineering; Zdravko Paveli, armaments; 

Petar Radonji, navy. 

The cooperation was established between The Sector for Development, Production and 

Scientic 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 


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, .akovi Specijalna vozila, Sl.Brod, elik/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 Scientic 

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 . akovi, 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-


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, Otoac, 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. Volarevi 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 Volarevi, PhD

50 

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 Mikuli 

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 sufcient 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.


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 battleeld 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 Scientic Research intensies 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 sufcient 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 scientic-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 scientic-research institution. The complete 

preparation of the project was led by professional adviser in The Sector for Development, 

Production and Scientic Research, Dinko Mikuli, 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 


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 Konar – 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 ofcers 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 ermak 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 Scientic 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 Tuman and Minister of Defence Gojko Šušak came to visit 

Brodarski Institute on 31 March 1993.


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 Vukovi, MSc and Zdenko Novak, MSc, reconstructed 

the cabin and cargo crate, which was completely produced in Progres rm-Jastrebarsko. 

Nedjeljko Ški, MSc, proposed the tting of 130 HP. domestic diesel-engine Torpedo, 

which would be technically and logistically better than 110 HP engine. Feasibility study 

justied the truck domestic production. A new manufacturing department for vehicle production 

was prepared within the factory. In the production system Director Neven Dedi 

and Head of Procurement Department Marelo Jurman worked out a plan for mounting

54 


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 Šljunara 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 protability 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 unication 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 

intensied 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 Šljunara 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 dened 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


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 satised 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, Zeevo, Šibenik, 1993.

56 


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 

sufcient 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 Scientic 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 

Matijaši, Col. Vjekoslav Stojkovi, Col. Josip Petrovi and other military ofcers 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 signicantly 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.


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 justied their purpose. Regarding the direct commanders' demand and soldiers' information 

from the eld, the Croatian Army was satised. 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 Scientic 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 fullled the military missions and reconciled the 

requests, thus serving the purpose of the country's liberation. 

4 References 

1. Ivkovi Z (2001) Hrvatski oklopnjaci 1991.-1993.: Ministarstvo obrane Republike 

Hrvatske, Zagreb, 2001. 

2. Poster (1996) Runo izraena oklopna vozila Hrvatske vojske 1991-1992: Galerija 

Zvonimir, Zagreb, 1996. 

3. Gambiroža N (1992) Proizvedeno u Hrvatskoj: Hrvatski vojnik, br. 17/92, Zagreb,1992. 

4. Bandula D (1997) Hrvatska vojna industrija: Velebit, 31.12.1997, Zagreb. 

5. Mallet J (1998) Croatian tanks and rocket launchers hits of Eurosatory 1998: DAILY, 

Janes Defence Weekly, Le Bourget, Pariz, 1998. 

6. Mikuli D (1991/1992) Kolegij blok: vlastite bilješke, Zagreb, 1992. 

7. Mikuli D (1991/1992) Oklopljena vozila: slikovni album, Zagreb, 1992. 

8. Mikuli D (1992/1993) Oklopna borbena vozila: slikovni album, Zagreb, 1993.

58 


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 

darko.stipnaicev@fesb.hr 

Abstract 

Wildres, including forest res, are natural phenomena that cause signicant 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 wildre damage minimization. The re season 2003 was quite severe one, 

particularly in Split and Dalmatia County. Provoked by great damages caused by 2003 

wildres 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 wildre 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 wildre 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: wildre, wildre detection, smoke detection, intelligent system 

1 Introduction 

Wildre 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 specic names could be used like forest re, grass re or bush- 

re. Wildres represent a constant threat to ecological systems, infrastructure and human 

lives. According to prognoses, wildres, 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 wildre 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


1992 to 2007 there were 4.851 wildres in Croatia and the burning area was 251.910 ha. 

Fire seasons 2000 and 2003 were particularly severe with 706 and 532 wildres and the 

total damage caused by wildres was huge. For example in 2003 only in Split and Dalmatia 

County there were 130 wildres, the total burned area was 9.700 ha, the direct damage 

caused by wildres (reghters 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.€ (Stipaniev et al., 2004, Stipaniev et al. 2007). Wildres in 

2003 were particularly catastrophic on Split and Dalmatia County islands. For example, 

on islands Hvar and Bra 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 – Wildres on Dalmatian islands in 2003 

Provoked by great damages caused by 2003 wildres 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 wildre prevention and protection. After detailed survey of research 

topics and implementation of ICT systems in wildre prevention task in other countries 

subjected to wildres, we have focused ourselves primarily on these three topics: 

- Automatic early wildre detection, 

- Calculation of micro location wildre risk index, and 

- Wildre behavior and propagation simulation. 

The rst topic is the most important one, because the only effective way to minimize damage 

caused by wildres is wildre early detection and fast and appropriate reaction, apart 

from preventive measures. Great efforts are therefore made to achieve early wildre detection, 

which is traditionally based on human wildre surveillance, realized by 24 hours observation 

by human observers located on selected monitoring spots. In Croatia the human 

wildres 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 wildre surveillance is video 

cameras based human wildre 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 

distant video presence on re remote location. 


Figure 2 – The difference between human wildre surveillance (left) and video camera based 

human wildre surveillance (right) 

The video cameras based human wildres 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 magnication) 

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 wildre 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) (iForest- 

Fire, 2011), belongs to the last generation of advanced automatic wildre 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 signicant export opportunities. 

In the rest of this paper, a short introduction to automatic wildre 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 wildre surveillance system 

was extended in the last couple of years. There are two main types of automatic wildre


surveillance systems: satellite systems based on satellite wildre monitoring and terrestrial 

systems based on wildre 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 wildre surveillance and monitoring system has to fulll. Beside 

the automatic detection of wildres, the distance video presence on re remote location is 

almost of equal importance and today’s satellite systems could not fulll that task. Satellite 

systems could be used only for wildre detection with limited spatial and time resolution. 

Therefore the automatic wildre surveillance and monitoring ground-based or terrestrial 

system used as an enhancement of human based wildre 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 wildre detection is based on smoke 

recognition during the day and re ames recognition during the night. 

- Infrared (IR) thermal imaging cameras where wildre detection is based on detection of 

heat ux from the re (Arrue et al, 2000). 

- Optical spectrometry that identies 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 efcient system for 

wildre 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 efcient 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 ames. Because of 

that in commercial optical spectrometry based systems, the video cameras in visible spectra 

are also included. Infrared systems are good choice for wildre 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 


contemporary systems for wildre 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 wildre 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 wildres is the most close to 

human based wildre detection. Human wildre observers primarily use his (or her) vision 

sensor (eyes) to detect wildre. Sometimes humans use additional visual enhancement 

devices like binocular to check suspicious areas, but humans never use other sensors for 

early wildre detection. During our research and development phase we have spoken with 

a lot various human wildre observers and no one of them said that he (or she) is using 

for example hearing (ears) to detect wildres. That was the reason why our efforts from 

the beginning were to develop ground-based wildre 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 wildres, various terrestrial systems based on cameras sensitive 

in visible spectra were developed and proposed. In all of them automatic wildre detection 

was based on smoke recognition during the day and ame 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 reections and human activities. Therefore, 

systems are usually designed as semi-automatic systems, which means that a human 

operator supervises the automatic wildre 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 conrms 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 wildre surveillance 

mentioned in previous section, but only to conrm 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 efciency 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 wildre 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 wildre surveillance and monitoring 

system. Wildres 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 wildres, particularly wildre smoke during the day and 

wildre ames during the night. If something suspicious is found, pre-alarm is generated


and appropriate image parts are visibly marked. The operator inspects suspicious image 

parts and decides is it really the wildre 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 wildre observer 

network theory based on three-layer sensor network architecture, formal theory of 

perception and notation of observer (Stipaniev et al, 2007a; Šeri et al, 2009). Wildre 

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 – Wildre 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 wildre 

protection. For such environment behavior the term environmental intelligence (EI) was 

introduced (Stipaniev et al, 2007a; Šeri et al, 2009).

64 


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 wildre 

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 wildre risk calculation during the 

monitoring phase and wildre 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-


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 reghting management 

activities. GIS data is essential for Simulation Mode and Fire Risk Calculation Mode. 

iForestFire is intelligent because it is based on articial 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) (Šeri 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 wildre 

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 (Krstini et al,2009; Krstini 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 (Jakovevi et al, 2009), 

rule-based expert system, data fusion algorithms (Stipaniev et al, 2007b) and integration 

of re risk index calculation with automatic adjustment of detection sensitivity (Bugari 

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 wildre observer detection quality evaluation 

(Jakovevi 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 reghters 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 


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 reghting intervention, both the early re detection and appropriate 

judgment about the potential re danger are important. That is the reason why from the 

reghters’ point of view, both automatic detection of wildres 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 (Stipaniev 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-dened 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-


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 wildre risk calculation and simulation of wildre propagation. 

Figure 6 – Experimental system for micro-location wildre risk calculation (left) and simulation 

of wildre 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 Krstini (who was also the voluntary reghter at island Hvar) enrolled 

the course Digital Image Processing and Analyses. The results of his seminar entitled 

“Wildre 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 Stipaniev as a team 

leader, Maja Štula as a designer of overall Web based information system, Damir Krstini

68 


as a designer of wildre detection algorithms and Ljiljana Šeri as a designer of wildre 

observer agent architecture. From the beginning the main reghters adviser was Tomislav 

Vuko, the vice commander of Croatian reghters for Adriatic Coast and Islands. 

Research started in 2003 with wildres video materials collecting necessary for detection 

algorithm development. Controlled res were burned at island Hvar in cooperation with 

voluntary reghters. 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 wildre 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 wildre alarm reduction has been applied so it was important to have 

manually segmented, not only wildre 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


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 Bra 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 signicant contributions were given also by 

Toni Jakovevi (advanced IP based meteorological station, detection algorithm), Marin 

Bugari (Web GIS based system features), Josip Maras (component based system architecture), 

Petar Jeri (software developement) and Kaja Radi (electronic components design 

and realization).

70 


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 (Stipaniev 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 scientic 

research, too. As a result of research connected with system development three PhD theses 

were written and defended (D.Krstini, Lj.Šeri and T.Jakovcevic), two PhD theses are in 

preparation phase and lot of scientic papers were published in journals, books and conference 

proceedings. A specialized Web portal dedicated to wildre observers and smoke 

recognition has been created and maintained (http://wildre.fesb.hr), but also our interest 

in more general scientic wildre research has resulted in establishment of Center for 

Wildre Research (http://cipop.fesb.hr).


Figure 11 – Istria iForestFire Net – advanced wildre monitoring network in Istria County 

5 Conclusion 

The only effective way to minimize damages caused by wildres is wildre 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 

wildre detection and monitoring. It describes the advanced wildre 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 dened 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 scientic research project, but today it is 

commercial system widely used for advanced wildre 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 scientic research activities resulting in a number of PhD thesis and published papers 

in journals, books and conference proceedings.

72 

References 


1. 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. 

2. Bugari M, Jakovevi T, Stipaniev D (2009) Automatic adjustement of detection 

parameters in forest re video monitoring system, MIPRO, May 2009, Opatija, pp. 

270-275 

3. 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 

4. Forest Fire Finder NGNS IS, http://www.ngns-is.com/html/orestas/fff_intro_eng. 

html (Accessed March, 1 2011) 

5. iForestFire (2011) Itelligent Forest Fire Monitorig System, http://iforestre.fesb.hr 

(Accessed March, 1 2011) 

6. iMeteo (2011) Advance Ultrasound IP based Mini Meteorological Station, http://imeteo.fesb.hr 

(Accessed March, 1 2011) 

7. Jakovevi T, Stipaniev D, Krstini D, (2009) False alarm reduction in forest re 

video monitoring system, MIPRO 2009, May 25-29, Opatija, 264-269 

8. Jakovevi T, Šeri Lj, Stipaniev D, Krstini D, (2010) Wildre smoke detection algorithms 

evaluation, VI Int. Conf. on Forest Fire Research, Coimbra, Nov 15-18 2010. 

9. Krstini D, Jakovevi T, Stipaniev D, (2009) Histogram-Based Smoke Segmentation 

in Forest Fire Detection System, Information Technology and Control, Vol.38, 

No.3, 2009. 237-244 

10. Krstini D, Kuzmani Skelin A, Slapniar I (2011) Fast Two-Step Histogram-Based 

Image Segmentation, IET image processing. 5 (2011), 1; 63-72 

11. 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 

12. Sahin Y, Animals as Mobile Biological Sensors for Forest Fire Detection, Sensors 

2007, 7, 3084-3099 

13. Sahin Y G, Turker Ince S, Early Forest Fire Detection Using Radio-Acoustic Sounding 

System, Sensors 2009, 9, 1485-1498 

14. Stipaniev D, Hrasnik B (2004) Integral, holistic model of wildre prevention in Split 

and Dalmatia County, Expert Study for Split and Dalmatia County, FESB Split, 2004. 

(231 pages) (in Croatian) 

15. Stipaniev D, Hrastnik B, Vuji R (2007) Holistic Approach to Forest Fire Protection 

in Split and Dalmatia County of Croatia, Wildre 2007 Int.Conference, Sevilla, Spain, 

May 2007. 

16. Stipaniev D, Bodroži 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 

17. Stipaniev D, Bodroži 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 

18. Stipaniev D, Bugari M, Bodroži 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 

19. Stipaniev D, Štula M, Krstini D, Šeri Lj, Jakovevi T, Bugari M (2010) Advanced 

automatic wildre surveillance and monitoring network, VI International Conference 

on Forest Fire Research, Coimbra, Porugal, Nov. 15 – 18, 2010, 053, 15 pages


20. Šeri Lj, Stipaniev 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. 708904.1- 

708904.12 

23. Žaja D (2008) The role of Croatian Forests d.o.o. in ghtings again wildres, International 

Workshop – new Methodes and Approaches to Wildres Prevention and Protection, 

Dec, 8-11 2008., Makarska (in Croatian)

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: igor.petrovic@gfv.hr 

Prof. Davorin Kovačić, Ph.D. 

University of Zagreb, Faculty of Geotechnical Engineering 

Hallerova aleja 7, Varaždin, Croatia 

e-mail: davorin.kovacic@gfv.hr 

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 simplied 

hypoplastic model was used. For the calibration of this model it was sufcient 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 


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 landlling 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.


An ideal project of waste management in the Republic of Croatia corresponds to the socalled 

non-landll 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 landlls which will remain in operation, 

- construction of new waste management centres in which the inevitable waste should be 

landlled. 

Table 1. Quantitative targets for landlls 

Targets 

Year 

2005. 2010. 2015. 2020. 2025. 

Regional centres for waste 

management 

0 1-2 2-3 3 4 

County centres for waste management 

0 3-7 7-10 10-14 14-21 

“Ofcial landlls” * 187 100 50 30 14-21 

Percentage of remedied landlls 

(% from the number determined 

in 2000). 

5 65 75 85 100 

*Group “Ofcial landlls“ consists of the following categories dened above: legal land- 

lls, landlls in the process of legalization, ofcial landlls, and landlls by agreement 

A waste landll 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 


Figure 1. Stability and deformation of a landll 

The study of the mechanical behaviour of municipal solid waste by monitoring the existing 

landlls 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 Landlls (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 landll of waste categories and operational requirements for waste 

landlls (2007), which is based on Council Directive 1999/31/EC of 26 April 1999 on the 

landll waste, only pre-treated waste is allowed to be dumped in landlls. 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 landll 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 landlled, 

compared with the wastes which have been landlled before. It can be concluded that the 

knowledge of mechanical parameters of MBT waste for the design of landlls, which will 

be built within waste management centres, is necessary. The extended operation period of 

a landll due to the reduction of landll 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 conned 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


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 scientic project “Characterization of municipal solid waste”. The basic outline of the 

project was given by Petrovi (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 landll 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 fulls 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 Unied Soil Classication 

System (USCS), the MBT waste can be classied as a coarse grained material, as 

the particle sizes lie in the range between 0.05 and 30 mm. The uniformity coefcient and 

coefcient 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 


The average density of the solid particles, s, 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 

s = 2147 [kg/m 3 ] was determined. 

Detailed description of the laboratory testing of basic MBT waste parameters is given by 

Petrovi 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 dening the technical specications 

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 

specications of new oedometer were dened 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 satised. Maximum vertical pressure on sample is 

2000 kN/m 2 . 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


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 


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 dened initial void ratio. 

The basic geotechnical parameters for both samples at the beginning of the test are presented 

in Table 2 respectively. Herein denotes the density, w the moisture content, d the 

dry density, s the average density of solid particles, e the void ratio and S the degree of 

saturation. 

Table 2 Basic geotechnical parameters of tested samples 

w d s e0 S 

[kg/m 3 ] [%] [kg/m 3 ] [kg/m 3 ] / [%] 

Sample A 1366 65 828 2147 1,59 88 

Sample B 1093 31 834 2147 1,57 42 

4.2 Consolidation procedure 

The initial pressure obtained from the pressure plate was 4,7 kN/m 2 . 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/m 2 , followed 

by unloading to 36 kN/m 2 , then reloading to 365 kN/m 2 and nally unloading to 4,7 kN/m 2 .


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 4,7-50 50-100 100-200 200-360 

Increment [kN/m 2 ] [kN/m 2 ] [kN/m 2 ] [kN/m 2 ] 

Sample A 566 1190 1515 3666 

Sample B 657 1316 1639 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/m 2 ] [kN/m 2 ] [kN/m 2 ] [kN/m 2 ] [kN/m 2 ] [kN/m 2 ] [mm] 

Duelmann 

(2002) 

730 1480 2460 4920 / / 0-30 

840 1990 1870 3290 / / 0-20 

800 1000 1800 / / / 0-20 

Kuehle- 1070 1590 1680 2880 / / 0-40 

Weidemeier 

(2003) 

500 1100 1600 2800 / / 0-40 

940 1490 2440 3030 / / 0-60 

600 1300 2000 2800 / / 0-60

82 

Bidlingmaier 

et al. (1999) 

Ziehmann 

(1999) 


355 1670 1947 2904 4514 / < 60 

239 905 1570 2973 4977 / < 100 

/ / / / / 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/m 2 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 dened by Duelmann (2002) has been also conrmed. 

Detailed description of the oedometric compression tests of MBT waste samples is given 

by Petrovi et al. (2010) and Petrovi et al. (2011c). 

4.4 Friction angle 

In order to estimate the critical friction angle c, 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.


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 dened 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 = - (11+ 22+ 33)/3 = -(1+2K0) 11/3, where 11 is the vertical 

stress, 22 = 33 the lateral stress under zero lateral strain and K0 denotes the pressure 

coefcient at rest, i.e. K0= 22 / 11= 33 / 11. 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 c according to: K0 

= 1 - sin c. 

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 simplied version of the hypoplastic model by Bauer (1996) and 

Gudehus (1996). 

(3) 

(4)

84 


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) 

Herein, is the time derivative of the Cauchy stress are normalized 

quantities, is the strain rate, and is the rate of void ratio. Factor is related to the critical 

friction angle c 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) 

K 0 

in Eq. (8) is also called the pressure coefcient at rest. It depends on the critical friction 

angle c and the density factor . From Figure 10 it can be seen that Eq. (8) allows a ner 

adaptation of K 0 

. In particular higher friction angles c and lower values of the density 

factor give signicantly lower values for K 0 

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 K 0 

, a critical friction angle of c = 39º and =1.1, the values e o 

, 

h oe 

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.


Figure 10. Comparison of K 0 

values obtained with the hypoplastic constitutive model (HM) for 

different density factor and friction angles c with the function by Jaky (1944) 

Table 5. Constants used for numerical simulations 

hs 

[kN/m 2 ] 

n e0 fd 

[0] 

Sample A 3519 0.389 1.75 1.1 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 Petrovi (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 


Figure 12. Oedometric compression and extension of Sample A 

Detailed description of the numerical modelling of MBT waste deformability is given by 

Petrovi 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 scientic 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 conrmed that the 

MBT waste can be classied 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 veried 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 conrmed. Simplied 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 Petrovi 

(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.


References 

1. Bauer E (1996) Calibration of a comprehensive hypoplastic model for granular materials. 

Soils and Foundations., Vol. 36(1), pp 13-26 

2. 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. 

3. Bauer E (2000) Conditions for embedding Casagrande’s critical states into hypoplasticity. 

Mechanics of Cohesive-Frictional Materials, 5. pp 125-148 

4. 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 

5. 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 

6. Casagrande A (1936) The Determination of Preconsolidation Load and its Practical 

Signicance. Proceedings, 1st International Conference on Soil Mechanics Foundation 

Engineering, Cambridge, Vol. 3, pp. 60–64. 

7. Council Directive 1999/31/EC of 26 April 1999 on the landll waste, Ofcial Journal 

of the European Communities, L 182, 16.7.1999. 

8. Domanovac T, Orašanin R (2002) Municipal solid waste composition for continental 

and coastal part of Croatia. In: Milanovi Z. (ed.) Proceedings of VIIth International 

Symposium Waste Management Zagreb 2002, mtg-topgraf, Zagreb, pp 61-68 (in Croatian) 

9. 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 

10. Geotechnics of Landll Design and Remedial Works Technical Recommendations – 

GLR, Second Edition, Ernst & Sohn, Berlin, 1993. 

11. Gudehus G (1996) A comprehensive constitutive equation for granular materials. Soils 

and Foundations, Vol 36(1), pp 1-12 

12. Jaky J (1944) The Coefcient of Earth Pressure at Rest. Journal for Society of Hungarian 

Architects and Engineers, pp 355-358 

13. Jessberger HL, Kockel R (1995) Determination and assessment of the mechanical properties 

of waste materials. In: Sarsby (ed.) Waste Disposal by Landlls – GREEN'93, 

Balkema, Bolton, pp 313-322 

14. Kuehle-Weidemeier M. (2003) Landlling and properties of MBP waste, Proceedings 

Sardinia 2003, Ninth International Waste Management and Landll Symposium, S. 

Margherita di Pula, Cagliari, Italy 

15. Niemunis A, Herle I (1997) Hypoplastic model for cohesionless soils with elastic 

16. strain range. Mechanics of Cohesive-Frictional Materials, Vol. 2(4), pp 279-299 

17. Ordinance on the methods and conditions for the landll of waste, categories and 

operational requirements for waste landlls (2007) Ofcial Gazette of the Republic of 

Croatia, No. 117/07 (in Croatian) 

18. Petrovi I (2008) Deformability of mechanically and biologically treated municipal 

solid waste, In: Milanovi Z. (ed.) Proceedings of Xth International Symposium Waste 

Management Zagreb 2008, Gospodarstvo i okoliš, Zagreb, pp 329-334 (in Croatian)

88 


19. Petrovi 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. Petrovi I, Kovai D (2010) Measurement of stiffness modulus of mechanically and 

biologically treated waste, In: Milanovi Z. (ed.) Proceedings of XIth International 

Symposium Waste Management Zagreb 2010, Udruga za gospodarenje otpadom, 

Zagreb,(in Croatian) 

21. Petrovi 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. Petrovi I, Szavits-Nossan V, Štuhec D (2011b) Laboratory testing of waste after biomechanical 

treatment. Graevinar 63(1), pp 43-53 (in Croatian) 

23. Petrovi I, Szavits-Nossan V, Kovai D (2011c) Deformability of mechanically and 

biologically treated municipal solid waste, Graevinar 63(3), pp 255-264 (in Croatian) 

24. Waste Management Strategy of the Republic of Croatia (2005) Ofcial 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) 

Ofcial 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


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- 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 identied 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 efcient 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 dene 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 congurations. 

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 simplied mechanical spring and rigid links model very well. 

The joint is represented by adequate rigid and exible components. They represent individual 

joint parts, Fig. 1.

90 


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

Typical bolted joint components are: 

• column web panel in shear, 

• end-plate in bending, 

• beam ange in bending, 

• beam web in tension, 

• beam ange 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-) curve can be used to show individual joint behaviour, 

Fig. 2. 

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

approximation 

Simpler behaviour idealisations are possible. The simplied 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.


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: 

• Identication of the active components from a global list of components (20 different 

components currently codied in EN 1993-1-8). 

• Evaluation of the force deformation response (F-) 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- curve). 

These steps are illustrated in Fig. 4. in simple case of a beam-to-column welded joint.

92 


Understanding the mechanical behaviour of different joint components allows the analysis 

of a large number of different joint congurations. The purpose of the component method 

is to determine the F- 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- 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 classication: 

• 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- 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, V wp,Rd 

, for 

single-sided framed joints: 

(1) 

where: 

f y,wc 

A vc 

M0 

the yield strength of the column web, 

the shear area of the column, 

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: 

A c 

b c 

t fc 

t wc 

r c 

the total area of the column, 

the ange width, 

the ange thickness, 

the web thickness, 

the root-radius of the web-ange 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 coefcient partially solves this problem, which has also


been accepted into EN 1993-1-8. The resistance of the unstiffened column web in shear for 

single-sided joints, V wp,Rd 

, is: 

(3) 

The shear deformation of the column web, s 

, in the overall initial rotation of joint is given 

by: 

(4) 

where: 

V 

G 

the shear force on the column web taken as 2F i 

(F i 

denoting the force in each 

bolt row and i the bolt row), 

the shear modulus. 

Thus the corresponding axial stiffness, K wp 

, is: 

(5) 

where: 

z 

E 

the lever arm between the compressive and the tensile areas, 

the modulus of elasticity. 

Dividing the axial stiffness, K wp 

, with the modulus of elasticity E the Eq. (5) is: 

(5a) 

where: 

k wp 

the stiffness coefcient 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 modied by a transformation 

parameter , EN 1993-1-8, to deal with this effect. For different joint congurations and 

unbalanced bending moments, the stiffness coefcient, k 

WP 

, Eq. (5a), must be modied by 

the transformation parameter :

94 


(6) 

For stiffened column web shear deformation can be disregarded (k wp 

= ). 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 ange are assumed, as shown in Fig. 

6. 

The following failure modes are possible: 

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

Mode 1: End-plate yielding without bolt failure


(7a) 

Mode 2: Simultaneous yielding of end-plate with bolt failure 

(7b) 

Mode 3: Bolt failure without end-plate yielding 

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, 

F t,Rd 

the design tension resistance of a bolt, 

the total value of F t,Rd 

for all bolts in the T-stub. 

F t,Rd 

Design bending resistance of an end-plate depending on failure mode is: 

(7c) 

(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 ange 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 coefcient analytical equations can be derived from classical beam 

theory, when the effective width is correctly estimated, so that: 

(9)

96 


where: 

k b,p 

the stiffness coefcient of the end-plate in bending, the designation in EN 1993- 

1-8 is k5, 

t p 

is the thickness of a end-plate. 

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

• Column ange 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 ange is stiffened. The 

unstiffened ange 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 ange yielding without bolt failure 

(10a) 

Mode 2: Simultaneous yielding of column ange with bolt failure 

(10b) 

Mode 3: Bolt failure without column ange yielding 

(10c) 

For column ange bending resistance the value which is used is the minimum out of the 

three failure modes.


The component stiffness coefcient 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 ange characteristics are used. The designation of this 

stiffness coefcient 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: 

b eff,t,wb 

t wb 

f y,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, 

the thickness of beam web, 

the yield strength of the beam web. 

The initial stiffness for this component can be taken to be innite. 

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 anges. 

Using the von Mises yield criterion, the design crushing resistance, Fc,wc,Rd, is: 

(13) 

The effective width of the column web in compression, b eff,c,wc 

, for bolted end-plate joints 

is:

98 


(14) 

where: 

a p 

the effective thickness of the weld, Fig. 8, 

s = r for rolled column sections, 

the length obtained by dispersion at 45º through the end-plate. 

s p 

Fig. 8. Transverse compression on an unstiffened column 

Reduction factor, k wc 

, accounts for the inuence of vertical normal stress due to axial force 

and the bending moment, com,Ed 

, and is given in the following: 

(15) 

Reduction factor accounts for the shear interaction, and equals:


(16) 

with 

(17a) 

(17b) 

where: 

is the transformation parameter from EN 1993-1-8. 

The buckling resistance, F c,wc,Rd 

, is taken approximately using the Winter equation: 

(18) 

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

or 

(19a) 

(19b) 

The normalised plate slenderness (column web panel), 

, is: 

(20) 

The initial deformation of the c component is:

100 


(21) 

where: 

N 

A c 

d c 

h c 

the resultant compressive force, taken as 2Fi (Fi denote the force in each bolt 

row and i the bolt row), 

the effective web area in compression zone, Ac = twc•beff,c,wc, 

the depth between column llets, 

the beam depth minus beam ange thickness. 

So that the initial (axial) stiffness becomes: 

(22) 

Dividing the axial stiffness, K c,wc 


(22a) 

It must be noted that for the calculation of the stiffness coefcient, kc,wc, a reduction of the 

effective width used for the resistance calculation is adopted (0,7×b eff,c,wc 

). The designation 

of this stiffness coefcient in EN 1993-1-8 is k 2 

. 

• 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, F t,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 inuence of vertical stresses arising from 

the column. 

Analogous to the preceding case, the initial deformation of this component w 

is: 

(24) 

where: 

T the resultant tensile force, taken as 2Fi (Fi denoting the force in each bolt row 

and i the bolt row), 

A t 

the effective web area in the tensile zone, A t 

= t wc 

×b eff,t,wc 

,


d c 

h t 

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, K t,wc 


(25a) 

This means that Eq. (25a) provides the stiffness coefcient, k t,wc 

, of the column web in tension. 

The designation of this stiffness coefcient in EN 1993-1-8 is k3. 

• Beam ange and beam web in compression 

This component, the beam ange 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 innity. 

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: 

A s 

f ub 

L b 

the tensile area of the bolt, 

the ultimate tensile strength of bolts, 

the sum of the thickness of the connected plates (the column ange and the endplate), 

the thickness of the washers and the half thickness of the nut and the bolt 

head.

102 


• Welds 

Welds are virtually undeformable (Kw = ). A rigid-plastic model is adopted as adequate, 

and the design resistance is: 

(29) 

where: 

a 

f u 

w 

the effective thickness of the weld, 

the ultimate tensile strength of the weld, 

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 coefcient, k i 

. The springs are connected serially and/or parallely, 

and the initial rotational stiffness, S j,ini 

, is: 

(30) 

The bending resistance of the joint, M j,Rd 

, equals: 

(31) 

where: 

F tr,Rd 

the effective design tension resistance of bolt-row r, 

h r 

the distance from bolt-row r to the centre of compression, 

r the bolt-row number.


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 ( = 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 S j,ini = 260 863 kNm/rad, 

• Bending resistance M j,Rd = 554 kNm. 

In order to agree with the designer’s expectations, every value has to be higher than: 

• the inexibility limit, based on the rotational stiffness 129 769 kNm/rad, 

• the maximum bending moment, transferred to the joint in global frames analysis, equals 

M j,Ed 

= 416 kNm. 

Since both conditions were fullled, the joint design can be considered satisfactory.

104 


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 inexible 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 

dened, 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 

signicant 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 congurations. They are under consideration 

with the goal of simplifying joint geometry, taking into account the frame analysis 

requirements.


• 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. 1, which the designer has suggested, fulls 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 signicantly 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 


Table 1. Joint congurations with corresponding resistance and stiffness values, as 

well as frame analysis requirement fullment 

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 fullled. 

This analysis shows very clearly that even for rigid joints, the new approach to joint design 

based on EC 3 can have serious benets. This is the result of the clearer denition of the 

inexibility 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


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 

simplied 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. Zoetemeijer, P. (1974) A design method for the tension side of statically-loaded bolted 

beam-to-column joints, Heron, pp 1-59 

2. Yee, Y.L.; Melchers, R.E. (1986) Moment-rotation curves for bolted connections, 

Journal of Structural Engineering 112 3, pp 615-635 

3. 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 

4. 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 

5. European Convention for Constructional Steelwork (1992) Analysis Design of Steel 

Frames with Semi-rigid Joints, Publication 67, Brussels: ECCS 

6. Weinand, K. (1992) SERICON – Databank on Joints in Building Frames, Proceedings 

of the 1st COST C1 Workshop, Strasbourg, pp 28-30 

7. 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 

8. 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 17- 

42 

9. 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 

10. Huber, G.; Tschemmernegg, F. (1998) Modelling of steel connections, Journal of Constructional 

Steel Research 45 2, pp 199-216 

11. 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 

12. 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 

13. Kuhlmann U. (1999) Inuence of axial forces on the component: web under compression, 

Proceeding of COST-C1 Working Group Meeting, C1/WG2/99-01, Thessaloniki 

14. Faella, C.; Piluso, V.; Rizzano, G. (2000) Structural Steel Semirigid Connections, The-

108 


ory, Design and Software, CRC Press LLC, 2000 N.W. Corporate Blvd., Boca Raton, 

Florida 33431 

15. CoP Software, (2002) Version 2002R03, RWTH Aachen, MSM Liège, ICCS Hoofddorp 

16. 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 

17. European Committee for Standardization (CEN) (2002) “EN 1990: 2002, Eurocode 0: 

Basis of structural design“, Final draft, Brussels 

18. Dujmovi, D.; Androi, B.; Skeji, D. (2003) Modeliranje prikljuaka elinih okvirnih 

konstrukcija, Graevinar 55 6, pp 339-348 

19. Dujmovi, D.; Skeji, D.; Androi, B. (2003) Modeliranje prikljuka nosa-stup prema 

Eurokodu 3, Graevinar 55 7, pp 397-405 

20. Girão Coelho, A.M. (2004) Characterization of the ductility of bolted end plate beamto-column 

steel connections, PhD Thesis. University of Coimbra, Coimbra, Portugal 

21. 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 

22. 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 

23. Skeji, D.; Dujmovi, D.; Androi, 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


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, 


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, classication, 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 


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 classied 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 . 

Because of insufcient 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 insufciency was a particular problem in 

terms of global structural analysis. By introducing a new engineering discipline called 

structural modelling, the insufciencies of the traditional approach had began to be consid-

110 


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 classies joints 

regarding stiffness (nominally pinned, semi-rigid and rigid), and resistance (nominally 

pinned, partial-strength and full-strength). Also, joints can be classied regarding ductility, 

although the criteria for this classication have not clearly been prescribed within Eurocode 

3 yet. 

2 STRUCTURAL BEHAVIOUR OF JOINTS 

A connection is dened 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 congurations, Figure 

1.b), are dened. 

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. Classication 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


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 innite. The moment that the joint transfers will have as a consequence 

the difference of 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 M j 

with the appropriate rotation , which in turn is the difference between absolute rotations 

of two connected structural elements. Rotational stiffness is dened as the slope of curve 

M j 

- 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 innite, 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 M j 

as well as the appropriate 

rotation 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 simplication, 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 modication 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 


• 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 - curve that describes its behaviour, 

• classication of joints: provides boundary conditions for the application of traditional 

joint modelling types, i.e. pinned or rigid, 

• joint idealisation: derivation of a simplied Mj - curve so that it is adjusted to the specic 

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, 

• simplied 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 identication 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 congurations. 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 simplied mechanical 

model consisting of springs and rigid links. In this the joint is simulated by means of the appropriate 

selection of rigid and exible components. These components represent specic 

parts of the joint that, depending on the type of loading, make an identied 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 ange in bending 

5. End-plate in bending 

6. Flange cleat in bending 

7. Beam or column ange and web in compression


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 ange, column ange, 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 identied. In 

general, each of these components is characterized by a non-linear relationship between 

the force and deformation (F-), 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 dened 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 -) 

joint response, by means of a representative mechanical model. 

Its application can suit various levels of element “processing”, e.g. a simplied 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 congurations with a relatively low number of 

components. The key element in the component method therefore relates to the characterisation 

of the F- 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- relationship 

of each component. Classication of joints 

3.2.1 Classification according to stiffness 

Classication 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 simplication, 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 classication according to stiffness are as follows:

114 


• rigid joint 

(unbraced frames), 

(braced frames), 

• semi-rigid joint 

(unbraced frames), 

(braced frames), 

• pinned joint 

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

Fig. 4. Boundaries of classication according to joint stiffness 

3.2.2 Classification according to resistance 

Classication according to resistance consists of comparing the design joint resistance moment, 

M j,Rd 

, with the boundaries of “full-strength” and “pinned”, Figure 5. 

Boundaries of classication according to resistance are as follows: 

• full strength joint 

• partial-strength joint 

• pinned joint 

M full strength 

denotes the design resistance of the weaker structural element in the connection. 

It shall be noted that the classication based on experimental Mj- joint characteristics was 

not taken into consideration because only the design properties were considered.


3.2.3 Classification according to ductility 

Fig. 5. Boundaries of classication of joint stiffness 

Adequate experience and appropriate detailing result in the so-called pinned joints that 

exhibit a sufcient 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 classication of cross-sections 

according to rotation capacity. Thus, joints can be classied into three classes concerning 

ductility, see Figure 6. 

Fig. 6. Ductility classes of joints – classication according to rotation capacity

116 


Class 1 - Able to reach MRd and with a sufciently 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 verication 

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 

Resistance 

Full-strength Partial-strength Pinned 

Rigid Continuous Semi-continuous - 

Semi-rigid 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


Table 2: Joint modelling and frame analysis 

Type of frame analysis 

Joint modelling 

Elastic-perfectly 

Elastic 

Rigid plastic plastic and elastoplastic 

Continuous Rigid Full-strength Rigid/full-strength 

Semi-continuous Semi-rigid Partial-strength 

Rigid/partialstrength 

Semi-rigid/fullstrength 

Semi-rigid/partial 

strength 

Simple Pinned Pinned Pinned 

Fig. 7. Simplied joint modelling for frame analysis 

3.4 Joint idealisation 

The Mj- curve that describes non-linear behaviour of joints is not convenient for everyday 

practice. However, it can be idealised without a more signicant 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 M j,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- curve and the yield plateau in 

the idealisation. 

The value of constant stiffness of joint S j.ini/ 

has been explained in a number of papers, 

and practical values are provided in EC 3, Part 1.8. The coefcient results from a highly

118 


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

elements. 

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

Actually there are different possibilities of Mj- 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 classication 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- 

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 inuences 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- 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 justication of a structure in which the joints have 

been considered as semi-rigid. To obtain the minimum price of a steel structure two design


strategies are possible regarding the execution of the joint: 

• Simplication of joint details which results in reducing the costs of workshop execution. 

It is important for unbraced frames where joints transfer signicant moments and those are 

rigid joints. 

• By reducing cross-sections of elements which results in decreasing the expenses for material. 

It is signicant 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 classication 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 


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 specic 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. Eurocode 3, Design of Steel Structures, Part 1.1: General Rules and Rules for Buildings, 

European Prestandard ENV 1993-1-1, Brussels: CEN, 1992. 

2. Revised Annex J of Eurocode 3. Joints in Building Frames, European Prestandard 

ENV 1993-1-1: 1992/A2, Brussels: CEN, 1998. 

3. Eurocode 3, Design of Steel Structures, Part 1.8: Design of Joints, European Prestandard 

prEN 1993-1-8, Final Draft, Brussels: CEN, 2001. 

4. Maquoi, R.; Chabrolin, B.: Frame Design Including Joint Behaviour, ECSC Report 

18563, Luxembourg, Ofce for Ofcial Publications of the European Communities, 

1998. 

5. 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 

6. European Convention for Constructional Steelwork, Analysis Design of Steel Frames 

with Semi-rigid Joints, Publication 67, Brussels: ECCS, 1992. 

7. Zoetemeijer, P.: A design method for the tension side of statically-loaded bolted beamto-column 

joints, Heron 1974; 1-59 

8. 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 

9. Huber, G.; Tschemmernegg, F.: Modelling of steel connections, Journal of Constructional 

Steel Research 45 (1998) 2, 199-216 

10. 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) 

11. Giro Coelho, A. M.; Bijlaard, F. S. K.; da Silva, L. S.:Experimental assessment of the 

ductility of extended end plateconnections. Engineering Structures 26 (2004), 1185- 

1206. 

12. 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. 

13. 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. 

14. Cruz, P. J.S; da Silva, L. S.; Rodrigues, D. S; Simes, 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 

15. Robot Millennium Program - Version 17.5: User's Manual, RoboBAT, June 2004.


Development strategy of 

ECOINDUSTRAL PARK RAŠA – EIPR 

Prof.dr.sc. Juraj Božičević 

Elitech Ltd. 

Trg kralja Tomislava 18 Zagreb 10000, Croatia 

juraj.bozicevic1@zg.t-com.hr, Tel. 00385 1 4922264 

Dr.sc. Marijan Andrašec 

Ecooleum Ltd. 

Trg kralja Tomislava 18, Zagreb 10000, Croatia 

marijan.andrasec@zg.t-com.hr, 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 Introduction 

2.0 The Community of Raša 

2.1 Geography 

2.2 Present situation 

3.0 Ecoindustrial Park Raša – EIPR - conceptualization 

3.1 What is an Ecoindustrial Park – EIP? 

3.2 Basics for conceptualization of EIPR 

3.3 Projects in EIPR proposed for realization 

3.4 Ecoindustrial Park and the Community of Raša 

3.5 Stakeholders 

3.6 Knowledge and experience on which venture relies 

4.0 Ecoindustrial Park Raša – EIPR - realization 

4.1 The way of achieving realization of EIPR 

4.2 Economy of the project EIPR 

4.3 Risks 

4.4 Development effects achieved with EIPR 

4.5 Priority tasks

122 

5.0 Conclusion 

6.0 References 



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 middle- 

European 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 Maari, 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


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 fulll 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 trafc 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 trafc 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 benecial 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 

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 signicant 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 hidromorc (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-Boljunica“. 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 

purication facility. 

Across the Community of Raša there is state road D21 Pula-Labin, number of regional and 

local roads. Numbers of unclassied roads are under community’s jurisdiction. 

Railroad Lupoglav-Štalije is used exclusively for transport; it’s not electried, it’s in group 

of mountain railways and is qualied 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 classied for international sea shipping (state-level port). The port of Koromano 

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


the past it was mostly used to support local industry, and therefore never had signicant 

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 trafc, 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 overow 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 ows. Acting together, the members of business units take care of the 

common prot, which is bigger than amount of single prot that the single company would 

achieve optimizing proper activity independently. In that way economic success of the 

companies increases and minimizes its inuence on environment. That is a new concept 

of sustainable development of the local community which protects the economy of local 

society from undesired inuences of globalization. 

EIP has rich and different possibilities of development including diversity of environment 

cultivation, infrastructure, single outt 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 (brownelds development). 

New industrial site development.

126 


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 difcult 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 

renery or other types of operation with large-scale material ows, 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 “by products” 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 efciency 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 specic 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 ow 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-specic 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 benet 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


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 difculties 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 trafc 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 Uka, 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-prot 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 benets 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 dened 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 dene 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 denition 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 prot 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 difculties: 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 


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 fulll 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 purication 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 protable) phase of technological development in that eld.


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 exibility: 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 puried water for agricultural irrigation, 

we would create a system that should economically and ecologically friendly bring 

several benets: purication of waste water, produced energy could be used for pumping 

water for irrigation in agriculture, puried 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 ower 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 protable. 

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 


are included from the beginning through the public discussions in the project and its inuence 

on living space, and number of benets: 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 diversication 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 efcient and economically sustainable and efcient ecoindustrial 

park of manufacturing and service industries, which would be based on principles of industrial 

symbiosis and realizing benets 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 efcient 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.


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 Uka-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, dene 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 


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 inuence 

Investors are loath to the risks, so it is important to decrease their exposure to the risks 

through wisely dened conditions of investments. Let’s emphasize that, for example, public 

private partnership cannot signicantly 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 exible 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,


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 dened, 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 


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 purication 

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 trafc infrastructure 

Renewal and furnishing of the port-area Bršica-Štalije 

Management of storm water and inundation protection 

Ecologically efcient 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. Andrašec M. (2009), Tehnologijske platforme, hrvatski razvojni projekti u EU 

okruženju. In: Božievi J. (eds) Inovacijska kultura i tehnologijski razvoj, CROSS, 

Zagreb, pp 109-114 

2. Baas L. (2008), Industrial Symbiosis in the Rotterdam Harbour and Industrial Complex, 

Bus. Strat. Env 17, pp 330-340 

3. Božicevic J. (1993), Private communication, Zagreb 

4. Božievi J. (ed) (1999), Luka kao složen sustav, Rijeka – glavna hrvatska luka, 

HATZ-CROSS, Zagreb


5. Božievi J. (ed) (1999), Povezani Rijeka i Zagreb hrvatsko i europsko gospodarsko 

žarište, HATZ, Zagreb 

6. Božievi J., Andrašec M. (2009), Opina Raša – Mudra zajednica, Opina Raša, 

Raša 

7. Božievi J., Andrašec M. (2009), Napredak prema održivom razvoju, Opina 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 Ofcials 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 


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. 


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


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 1- 

10, 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 specically 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: 

dF ve 

/dt =-k 1 

F ve 

/1/

138 


dF it 

/dt =k 1 

F ve 

–k 2 

F it 

–k 3 

F it 

/2/ 

if neglecting chemotherapeutic drug resorption rate value in ventricular part (Such 

oversimplication 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 

dF fc 

/dt =k 2 

F it 

/3/ 

If, however, a chemotherapeutic agent resorption rate is enough signicant already in 

ventricular part, then instead of equation /1/ the following equation should be applied: 

dF ve 

/dt =-k 1 

F ve 

–aF ve 

/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 

dF sng 

/dt =k 3 

F it 

–(k 4 

+k 6 

+k 7 

+k 9 

+k k 

) F sng 

+k 5 

F cr 

+k 8 

F hp 

+k a 

F t 

+ 

+k i 

F men 

/4/ 

or by equation 

dF sng 

/dt =a F ve 

+k 3 

F it 

–(k 4 

+k 6 

+k 7 

+k 9 

+k k 

) F sng 

+k 5 

F cr 

+k 8 

F hp 

+k a 

F t 

+ 

+k i 

F men 

/4a/ 

Communication with nephritic-urinary system is mainly irreversible and therefore drug 

transfer rate towards this system can be presented by applying the equation 

dF nu 

/dt =k 6 

F sng 

/5/ 

The liver has important role and communicates reversibly with blood. The equation 

dF hp 

/dt =k 7 

F sng 

–k 8 

F hp 

/6/ 

is considered to be mathematically adequate. 

For the main part of tissue one can apply the equations: 

dF t 

/dt =k 9 

F sng 

–(k a 

+k d 

) F t 

+k e 

F tr 

/7/ 

dF tr 

/dt =k d 

F t 

–k e 

F tr 

/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


respiratory system is necessary and is a prerequisite of body survival. The equations 

and 

dF cr 

/dt =k 4 

F sng 

–(k 5 

+k b 

) F cr 

+k c 

F rsp 

/9/ 

dF rsp 

/dt =k b 

F cr 

–k c 

F rsp 

/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 

dF men 

/dt =k k 

F sng 

–k i 

F men 

/11/ 

Presence of chemotherapeutic substance in the cerebral-neural system can induce positive 

and/or negative effects inuencing their intensities. One of possibilities to express these 

consequences could be by applying the equations 

dMenp/dt =k u 

F men 

(1-Menp/Mpm) /12/ 

dMenn/dt =k v 

F men 

(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 F men 

quantities. Since 

it is much more probable that the inhibitory effects would reect rather on positive than 

negative consequences, it seems to be sufcient to modify the equation /12/ with taking 

into account inhibitory effects of increasing F men 

quantities. The equation, 

dMenp/dt =k u 

F men 

(1-fF men 

-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 

efcacious 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 dened body viability, Bv, and body non-viability, Bnv. In 

present approach it is supposed that differential equations 

dB v 

/dt =-k s 

.B v 

F sng 

+k r 

S rv 

B nv 

/14/ 

dB nv 

/dt =k s 

B v 

F sng 

–k r 

S rv 

B nv 

/15/ 

dS rv 

/dt =-k r 

S rv 

B nv 

+k rv 

(1-S rv 

) /16/

140 


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”, S rv 

, 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 dened 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 inuence. No doubt, both the efciency 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 =wPs+xMenp-yMenn /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 efciently 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 inuence cannot be recommended for detailed 

consideration in this work, because of actual insufcient knowledge to dene and estimate 

exactly the effects of their inuence. 

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 

sufcient 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:


Dp + Fmen Fmni 

/Sh-1/ 

If the expressed reduction rate of positive effects becomes proportional to Dp amount, then 

the applicability of following equations can be expected: 

dD p 

/dt =b dMenn/dt –c D p 

F men 

/19/ 

dF mni 

/dt =c D p 

F men 

/20/ 

dF men 

/dt =k k 

F sng 

–k i 

F men 

- c D p 

F men 

/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 =k u 

F men 

(1-fF men 

-Menp/M pm 

) –d D p 

/22/ 

Such a consideration does not ask detailed discussion referring to the kinetics of inactivated 

drug, F mni 

. If both positive and negative effects will spontaneously reduce in proportion to 

their specic rates, then instead of equation /22/ and equation /13/ one can propose the 

equations: 

dMenp/dt =k u 

F men 

(1-fF men 

-Menp/M pm 

) –d D p 

–g Menp /23/ 

dMenn/dt =k v 

F men 

(1-Menn/M nm 

) –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 

simplications can be introduced in order to have more convenient nal model. It seems 

that lumping of the terms F ve 

and F it 

into common F it 

, as well as of F hp 

, F tr 

and F t 

into 

common term F t 

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: 

dF it 

/dt =-k 1 

.F it 

–k 2 

.F it 

/25/ 

dF sng 

/dt =k 1 

F it 

–(k 4 

+k 6 

+k 9 

+k k 

) F sng 

+k 5 

F cr 

+k a 

F t 

+k i 

F men 

/26/ 

dF t 

/dt =k 9 

F sng 

–k a 

F t 

/27/ 

2.4 Modelling of the psychotherapeutic substance transformation 

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

142 


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, signicant 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 efciency which could be expressed 

by formed metabolites. In this work, pharmacokinetics of formed therapeutically active 

product is considered to demonstrate how some specic 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 specic 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 

reected to the state in cardiovascular system. Therefore, in the extended mathematical 

model, instead of equation /26/ the equations 

dF sng 

/dt =k 1 

Fit-(k 4 

+k 6 

+k 9 

+k k 

+j)F sng 

+k 5 

F cr 

+k a 

F t 

+k i 

F men 

/26a/ 

dF tsng 

/dt =jF sng 

-(k 6 

+k k 

)F tsng 

+k a 

F tt 

+k i 

F mt 

/28/ 

should be applied, if neglecting the formed product afnity to the hearth and if supposing 

its transfers concerning other body compartments happen with same specic rates as 

those referring to transfers of the original substance. In addition, the corresponding new 

equations should be also included: 

dF nut 

/dt =k 6 

F tsng 

/29/ 

dF tt 

/dt =k 9 

F tsng 

-k a 

F tt 

/30/ 

dF mt 

/dt =k k 

F tsng 

-k i 

F mt 

/31/ 

It is supposed that formed product inuences 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 =k u 

(F men 

+F mt 

)(1-fF men 

-Menp/M pm 

) /32/ 

dMenn/dt =k v 

(F men 

+F mt 

)(1-Menn/M nm 

) /33/ 

dB v 

/dt =-k s 

B v 

(F sng 

+F tsng 

)+k r 

S rv 

B nv 

/34/ 

dB nv 

/dt =k s 

B v 

(F sng 

+F tsng 

)-k r 

S rv 

B nv 

/35/ 

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

them now.


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 efciency. 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 scientic 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 (k u 

, k v 

, w, x, y, z) on the psychotherapy 

efciency (Ut). Ut values refer to the 50 th hour of the computer simulation process. 

Parameter 

name 

Values 

ku kv w x y z Ut 

0.02 0.015 0.001 0.001 0.001 0.001 0.00492 

0.022 0.025 0.001 0.001 0.001 0.001 -0.000525 

0.02 0.02 0.001 0.001 0.001 0.001 0.00123 

0.02 0.02 0.001 0.002 0.001 0.001 0.0218 

0.02 0.02 0.001 0.001 0.002 0.001 -0.01934 

0.02 0.02 0.002 0.001 0.002 0.001 -0.01811 

0.02 0.02 0.002 0.001 0.002 0.002 -0.01573 

0.02 0.01 0.002 0.001 0.002 0.002 0.000668 

0.02 0.01 0.002 0.002 0.001 0.002 0.033608 

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

144 


way of its application and its applied amount, afnity 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 reect 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. 

Y (mg) 

Psychotherapy pharmacokinetics 

10.0 

1.00 0.200 

9.5 

9.0 

8.5 

8.0 

7.5 

0.75 0.149 

7.0 

6.5 

6.0 

5.5 

5.0 

0.50 0.098 

4.5 

4.0 

3.5 

3.0 

2.5 

0.25 0.046 

2.0 

1.5 

1.0 

0.5 

0.0 

0.00 -0.005 

0 10 20 30 40 50 

Time (h) 

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 

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

Ut,Ps (units) 

Y (mg) 

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 

Psychotherapy pharmacokinetics 

0 10 20 30 40 50 

Time (h) 

1.00 

0.65 

0.30 

-0.05 

-0.40 

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

0.200 

0.138 

0.075 

0.012 

-0.050 

Ut,Ps (units) 

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) b) 

Fig. 1a,b 

Simulated psychotherapy course for applied same computer simulation parameters during 

two consecutive cycles: k 1 

=0.3; k 2 

=0.01; k 3 

=0.8; k 4 

=0.1; k 5 

=0.5; k 6 

=0.02; k 7 

=0.05; k 8 

=0.08; 

k 9 

=0.1; k a 

=0.1; k b 

=0.05; k c 

=0.03; k d 

=0.1; k e 

=0.05; K k 

=0.1; k i 

=0.15; k u 

=0.02; k v 

=0.01; 

M pm 

=1.0; M nm 

=1.0; k s 

=0.005; k r 

=0.1; k rv 

=0.5; w=0.002; x=0.002; y=0.001; z=0.002; f =0.65. 

Initial value for F ve 

was the same for both a) and b) cycles; other initial independent variable 

quantities of the 2 nd cycle (b) were those at the 50 th 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 inuence 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


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. 

Y (mg) 

Psycho- therapy pharmacokinetics 

6.0 

1.00 0.050 

5.7 

5.4 

5.1 

4.8 

4.5 

0.75 0.038 

4.2 

3.9 

3.6 

3.3 

3.0 

0.50 0.025 

2.7 

2.4 

2.1 

1.8 

1.5 

0.25 0.013 

1.2 

0.9 

0.6 

0.3 

0.0 

0.00 0.000 

0.0 4.8 9.6 14.4 19.2 24.0 

FIT_CALC vs T 

Time (h) 

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 

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

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

Y (mg) 

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 


0.0 4.8 9.6 14.4 19.2 24.0 

Time (h) 

1.00 

0.75 

0.50 

0.25 

0.00 


0.100 

0.075 

0.050 

0.025 

0.000 


FIT_CALC vs T 

FF_CALC vs T 

FSG_CALC vs T 

FC_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) 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: k 1 

=0.8; k 2 

=0.01; k 4 

=0.1; k 5 

=0.5; k 6 

=0.02; k 9 

=0.1; 

K k 

=0.1; k a 

=0.1; b=0.01; c=0.03; d=0.0001; k i 

=0.15; k u 

=0.02; f=0.5; k v 

=0.01; g=0.0001; 

h=0.1; M pm 

=1.0; M nm 

=1.0; k s 

=0.005; k r 

=0.1; k rv 

=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 classied 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 classication 

criterion seems to be by considering every patient as individual asking specic 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 


As demonstrated, slowed resorption leads to benecial 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 k 1 

) on values 

of pharmacokinetics parameters at 10th hour after drug application. 

Parameter name Initial Value at 10th hour for given k 1 

value 

value 

k 1 

=0.8/h k 1 

=0.08/h 

FIT 100.0 0.030354 40.657 

FF 0.0 1.2342 6.5937 

FSG 0.0 16.699* 15.413** 

FC 0.0 3.8810 2.8076 

FTSG 0.0 17.915 8.5609 

FNU 0.0 6.2270 2.5503 

FT 0.0 18.144 8.6117 

FM 0.0 14.235 7.2429 

FR 0.0 2.1823 0.82074 

MEP 0.0 -5.8849 -1.1067 

MEN 0.0 0.59911 0.28760 

BV 0.99 0.33840 0.57298 

BNV 0.01 0.66160 0.42702 

SRV 1.0 0.88626 0.93331 

U 0.0 -0.058650 -0.005600 

PS 0.0 -0.022766 0.014274 

DP 0.0 0.0011259 0.0016591 

FMI 0.0 0.0048652 0.0012169 

FTT 0.0 9.0919 3.1410 

FMT 0.0 7.7490 2.75743 

FUT 0.0 2.6062 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 dened with k 1 

= 

0.08/h.


Quantity 

name 

Initial 

value 

Simulated value 

after 8 th hour 

Simulated value 

after 25 mg active 

substance addition 

at 8 th hour and after 

16 hours 

Simulated values after 

50 mg active substance 

additions at 8 th and at 16 th 

hour 

Value after 

16 th hour 

Value after 

24 th hour 

FIT 100.0 48.675 35.861 48.030 47.716 

FF 0.0 5.7028 9.9043 11.330 16.920 

FSG 0.0 15.600 17.548 21.448 25.419 

FC 0.0 2.6998 3.3293 4.0042 4.8585 

FTSG 0.0 7.0583 13.705 15.469 24.329 

FNU 0.0 1.9287 4.7860 5.2682 9.3740 

FT 0.0 7.0789 13.049 14.819 20.868 

FM 0.0 6.1546 10.210 11.749 15.583 

FR 0.0 0.58643 1.5909 1.7375 3.0097 

MEP 0.0 -0.54687 -3.2256 -3.7939 -6.4175 

MEN 0.0 0.19988 0.51099 0.53363 0.69866 

BV 0.99 0.63995 0.42220 0.38628 0.27897 

BNV 0.01 0.36005 0.57780 0.61372 0.72103 

SRV 1.0 0.94798 0.90219 0.89739 0.87648 

U 0.0 -0.0018992 -0.034447 -0.037620 -0.12630 

PS 0.0 0.013684 0.0042820 0.0022673 -0.039698 

DP 0.0 0.0014112 0.0013068 0.0012736 0.00039878 

FMI 0.0 0.00058767 0.0038032 0.0040628 0.0065879 

FTT 0.0 2.0975 7.0013 7.5256 14.736 

FMT 0.0 1.8886 5.7933 6.2655 11.627 

FUT 0.0 0.52811 2.2175 2.3495 5.5524

148 


Y (mg) 

100 

95 

90 

85 

80 

75 

70 

65 

60 

55 

50 

45 

40 

35 

30 

25 

20 

15 

10 

5 

0 


0 2 4 6 8 10 

Time (h) 

1.0 

0.8 

0.5 

0.3 

0.0 


0.100 

0.075 

0.050 

0.025 

0.000 


FIT_CALC vs T 

FF_CALC vs T 

FSG_CALC vs T 

FC_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 

Y (mg) 

100 

95 

90 

85 

80 

75 

70 

65 

60 

55 

50 

45 

40 

35 

30 

25 

20 

15 

10 

5 

0 


0 2 4 6 8 10 

Time (h) 

1.00 

0.75 

0.50 

0.25 

0.00 


0.100 

0.075 

0.050 

0.025 

0.000 


FIT_CALC vs T 

FF_CALC vs T 

FSG_CALC vs T 

FC_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) b) 

Fig. 3a,b. 

Effect of active substance resorption rate on pharmacokinetics quantities: a) k 1 

=0.8; b) 

k 1 

=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


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 dene 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 B v 

, B nv 

, S rv 

, Menp, Menn, etc. No doubt, a reliability

150 


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 dene, at least approximately, the maximal specic 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 efciency of 

psychotherapy treatments. Of course, evaluations should be performed in the right time. 

One cannot recommend e.g. a postponing of B nv 

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 trafc 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 ) 

B v 

body vitality, dimensionless 

B nv 

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 ) 

F cr 

pharmacokinetic substance mass in hearth, mg (M) 

F fc 

pharmacokinetic substance mass in faeces, mg (M) 

F hp 

pharmacokinetic substance mass in liver, mg (M) 

F men 

pharmacokinetic substance mass in neural-brain system, mg (M) 

F mni 

inhibited F men 

, mg (M) 

F mt 

transformed pharmacokinetic substance mass in neural-brain system, mg (M) 

F nu 

pharmacokinetic substance mass in nephritic–urinary system, mg (M) 

F nut 

transformed pharmacokinetic substance mass in nephritic–urinary system, mg (M) 

F rsp 

, F plm 

pharmacokinetic substance mass in respiratory system, mg 

(M) 

F sng 

pharmacokinetic substance mass in blood, mg (M) 

F tsng 

transformed pharmacokinetic substance mass in blood, mg (M) 

F t 

pharmacokinetic substance mass in tissue, mg (M) 

F tt 

transformed pharmacokinetic substance mass in tissue, mg (M) 

F tr 

pharmacokinetic substance mass in tissue receptor, mg (M) 

F ve, 

F it 

pharmacokinetic substance mass in ventricular (F ve 

)-intestinal (F it 

) compartment, 

mg 

(M) 

g kinetic constant, h -1 (T -1 ) 

h kinetic constant, h -1 (T -1 ) 

j kinetic constant, h -1 (T -1 ) 

k 1 

kinetic constant with reference to F ve 

, h -1 (T -1 ) 

k 2 

kinetic constant with reference to F it 

, h -1 (T -1 ) 

k 3 

kinetic constant with reference to Fit, h -1 (T -1 ) 

k 4 

kinetic constant with reference to F sng 

, h -1 (T -1 ) 

k 5 

kinetic constant with reference to F cr 

, h -1 (T -1 )


k 6 


, h -1 (T -1 ) 

k 7 


, h -1 (T -1 ) 

k 8 

kinetic constant with reference to F hp 

, h -1 (T -1 ) 

k 9 


, , h -1 (T -1 ) 

k a 

kinetic constant with reference to Ft, h -1 (T -1 ) 

k b 

kinetic constant with reference to F cr 

, h -1 (T -1 ) 

k c 

kinetic constant with reference to F rsp 

, h -1 (T -1 ) 

k d 

kinetic constant with reference to Ft, h -1 (T -1 ) 

ke kinetic constant with reference to F tr, 

h -1 (T -1 ) 

ki kinetic constant with reference to F men 

,h -1 (T -1 ) 

k k 

kinetic constant with reference to F sng, 

,h -1 (T -1 ) 

k r 

kinetic constant with reference to B nv 

and Srv, h -1 (T -1 ) 

k rv 

kinetic constant with reference to Srv, h -1 (T -1 ) 

k s 

kinetic constant with reference to Bv and F sng 

, mg -1 h -1 (M -1 T -1 ) 

k u 

kinetic constant with reference to F men 

and Menp, mg -1 h -1 (M -1 T -1 ) 

k v 

kinetic constant with reference to F men 

and Menn, mg -1 h -1 (M -1 T -1 ) 

Menn normalised quantity of negative effects caused by F men 

, dimensionless 

Menp normalised quantity of positive effects caused by F men 

, dimensionless 

Mnm maximal value of Menn, dimensionless 

Mpm maximal value of Menp, dimensionless 

Ps normalised effect of psychiatrist and personnel, dimensionless 

S rv 

normalised quantity of revitalisation substrate, dimensionless 

Ut normalised psychotherapy efciency, dimensionless 

w kinetic constant with reference to Ps, h -1 (T -1 ) 

x kinetic constant with reference to Menp, h -1 (T -1 ) 

y kinetic constant with reference to Menn, h -1 (T -1 ) 

z kinetic constant with reference to Ps, h -1 (T -1 ) 

References 

1. Bošnjak M. (1982) Mathematical modelling of biochemical reaction systems (in 

Croatian) Kem. Ind. 31 545 -559 

2. Bošnjak M. (2009) Introduction to Kinetics of Microbial Processes ( in Croatian: 

Uvod u kinetiku mikrobnih procesa), Graphis, Zagreb, (2A= pp. 275-306, Appendix) 

3. Bošnjak M., Topolovec V., Vrana M. (1978), Growth kinetics of Streptomyces 

erythreus during erythromycin biosynthesis, J. Appl. Chem. Biotechnol. 28 791- 798 

4. 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 

5. Bošnjak M., Bago Joksovi A., Pigac J., Bošnjak Cihlar Ž., Hranueli D., (2006) 

Applicability of mathematical models in dening the behaviour kinetics distinction 

among microbial strains, Chem. Biochem. Eng. Q. 20 375 -388 

6. Bošnjak M., Udikovi Koli N., Petri I., Cihlar D., Hršak D., (2010) Integrated 

approach to mathematical modelling of atrazine degradation in different reaction 

systems, Food Technol. Biotechnol.48 392-403 

7. 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 

8. Karba R., Mrhar A., Bremšak F., Kozjak F., (1977) Estimation of pharmacokinetics

152 


parameters of amoxicillin and ampicilline by identication (in Slovenian: Doloevanje 

farmakokinetinih parametrov amoksicilina in ampicilina s pomojo identikacije), 

In: Informatica 77, Proc. Yug. Intern. Symp. 6 218, Slov. Soc. Informatika, Ljubljana 

9. 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 

10. 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, 29 th Symp. Soc. Gen. Microb., Cambridge, University 

Press, Cambridge, , pp. 327 -35 

11. LILLY USA, LLC Internet sites. 

12. Li R. C., Nix D. E., Schentag J. J., (2006) Pharmacodynamic modelling of bacterial 

kinetics: -Lactam antibiotics against Escherichia coli, J. Pharm. Sciences 83 970 

-975 

13. 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 

14. 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 

15. 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 specic 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


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: marijan.bosnjak@hatz.hr 

Home address: 

HR-10000 Zagreb, Slovenska 19, Croatia 

Tel. (385)-1-3702507

154 


DICES: Distributed Component-based Embedded 

Software Systems 

Mario Žagar 

Faculty of Electrical Engineering and Computing 

University of Zagreb, Zagreb, Croatia 

mario.zagar@fer.hr 

Ivica Crnković 

Mälardalen Real-Time Research Centre 

Mälerdalen University, Västerås, Sweden 

ivica.crnkovic@mdh.se 

Darko Stipaničev 

Faculty of Electrical Engineering, Mechanical Engineering and Naval Architecture 

University of Split, Split, Croatia 

darko.stipanicev@fesb.hr 

Maja Štula 



maja.stula@fesb.hr 

Juraj Feljan 

Mälardalen Real-Time Research Centre 

Mälerdalen University, Västerås, Sweden 

juraj.feljan@mdh.se 

Luka Lednicki 



luka.lednicki@fer.hr 

Josip Maras 



josip.maras@fesb.hr 

Ana Petričić 



ana.petricic@fer.hr 

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,


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 


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 efciency 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 signicant importance in Croatia, since many leading companies 

in Croatia either produce such systems (e.g. Konar, Ericsson Nikola Tesla) or use such 

systems (e.g. Pliva, or many small companies). 

DICES is addressing efcient 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 


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 efciency, 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.


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 modied 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 difcult to establish the minimum amount of code responsible

158 


for implementing the desired behavior [8]. 

In the web application domain, reuse is especially difcult: 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 dened 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 dened by a 

combination of HTML, CSS, JavaScript and resources (images, videos, fonts, etc.) that are 

intermixed with code and resources dening 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 denes 

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.


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 exible because any modications 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 


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 congured 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 conguration 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 dened 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


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 notications 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 

exible and eliminating the need to change the code of devices to adapt them to the needs 

of the developed system. Also, run-time modication of systems is much easier. System’s 

behaviour can be modied 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 


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. 

5 An overview of the PRIDE tool 

Figure 5: Screenshot of UComp Developer 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, verication 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.


In the following list we give some overall reections 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 signicantly 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 specic 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-specic 

technologies that provide support not only for functional development but also for analysis 

and verication 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 specic tool support that enables: 

Efcient system design by using existing components, 

Seamless integration of different tools to provide the analysis and verication 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 verication techniques. In difference 

to similar approaches [1, 2, 17], PRIDE puts emphasis on EFPs during the entire lifecycle 

- from early specication to deployment and synthesis. 

PRIDE has been designed to support four design strategies that are especially important to 

consider for having an efcient 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 specication, 

to very “concrete” with a xed specication, 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 efcient 

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 


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 efcient 

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 identied 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 verication 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 dened. 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


manually modied 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 signicantly 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 predened 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 specication 

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 


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 modications 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 specic port, 

service or the component as a whole. Attributes are dened by attribute types, and include 

attribute values with metadata and the specication 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 dening of early estimates for EFPs even before the component has been 

implemented and use it for analysis in early stages of system development. New, userdened 

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 efcient 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.


References 

1. 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 

2. Articus Systems, Rubus Software Components, http://www.arcticus-systems.com , 

Accessed 7 March 2011 

3. Bellettini C., Marchetto A., Trentini A. (2004) WebUml: reverse engineering of web 

applications, In: Symposium on Applied computing, ACM New York, NY, USA, 1662- 

1669 

4. 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 

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 

6. 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 

7. Feljan J., Lednicki L., Maras J., Petricic A., Crnkovi I. (2009) Classication and survey of 

component models, MRTC report ISSN 1404-3041 ISRN MDH-MRTC-242/2009-1-SE, 

Mälardalen Real-Time Research Centre, Mälardalen University 

8. Holmes R.(2008) Pragmatic Software Reuse, Ph.D. Dissertation, University of Calgary 

9. 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 

10. Ricca F., Tonella P. (2001) Understanding and Restructuring Web Sites with ReWeb, IEEE 

Multimedia, Vol 8, Issue 2, pp. 40-51 

11. Satoshi Konno, CyberLink for Java, http://www.cybergarage.org/twiki/bin/view/Main/ 

CyberLinkForJava , Accessed 7 March 2011 

12. Sentilles S., Štpán P., Carlson J., Crnkovi 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 

13. Sentilles S., Vulgarakis A., Bureš T., Carlson J., Crnkovi 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 

14. The Eclipse Foundation, Eclipse Modeling Framework Project, http://www.eclipse.org/ 

emf/ , Accessed 7 March 2011 

15. The Eclipse Fundation, Graphical Modeling Framework, http://http://www.eclipse.org/ 

gmf , Accessed 7 March 2011 

16. UPnP forum, Universal Plug and Play, http://www.upnp.org , Accessed 7 March 2011 

17. 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 

18. Maras J., Štula M., Carlson J, Crnkovi I. (2011). Firecrow web page, http://www.fesb. 

hr/~jomaras/?id=Firecrow, Accessed 7 March 2011

168 


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. 


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 ow. 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


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 ow 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 ow. 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 ow. The basic assumption is that the motion eld is well approximated by a 

constant vector eld velocity v r 

within any small region of the volume plane. The optical 

ow is the approximation of the motion eld which can be computed from time-varying 

volume sequences. For each point p i 

within a small patch Q of size N x N x N, it can be 

written 

∂E 

( ∇E) T r 

v + = 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 


p , p , p ∈ 

1 2 

K , 3 Q . Therefore, the optical ow can be estimated within Q as 

N 

the constant vector, v r that minimizes the functional 

2 

T E 

 

v E 

v . (2) 

 

p Q 

t 

 

i 

The solution of the least squares problem can be found by solving the linear system 

T 

T 

A Av = A b [3]. The i-th row of the N 3 x 3 matrix A is the spatial gradient evaluated 

at point p i 

: 

⎡∇E 

⎢ 

⎢ 

∇E 

A = 

⎢M 

⎢ 

⎣∇E 

( p1 

) 

( p ) 

2 

⎤ 

( p ) ⎥⎥⎥⎥ NxNxN ⎦ 

and b is the N 3 -dimensional vector of partial temporal derivatives of the volumetric 

brightness, evaluated at p , p2 , K , p 3 after a sign change: 

1 N 

∂E 

⎤ 

( ) ( ) T 

p , 

⎡∂E 

b = − 

⎢ 1 

K , p NxNxN 

⎣ ∂t 

∂t 

⎥ . (4) 

⎦ 

The least squares solution of the constrained system can be obtained as 

r 

v = 

T −1 

T 

( A A) A b 

(3) 

. (5) 

ψ [] v r is the optical ow (the estimate of the motion eld) at the centre of patch Q. By 

repeating this procedure for all cubic patches, a dense optical ow is obtained. This 

algorithm can be summarized as follows. The input is a time-varying sequence of n volumes 

E 1 

, E 2 

, …, E n 

. Let Q be a sub-volume region of N x N x N voxels. 

1. Filter each image of the sequence along each spatial dimension with a Gaussian lter 

of standard deviation equal to σ s . 

2. Filter each image of the sequence along the temporal dimension with a Gaussian 

lter of standard deviation σ t . 

3. 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 ow 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 ow 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


sequence. We will now describe and adapt two different feature-based matching techniques 

from 2D to 3D: the constant ow algorithm and the feature tracking tracking. 

2.2.1 Constant Flow Algorithm 

Features can be dened as centres of those regions for which the smallest eigenvalue of A T A 

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 ow algorithm [4]. 

The input is formed by two frames of a volumetric sequence, I 1 

and I 2 

and a set of 

corresponding feature points in the two frames. Let Q 1 

, Q 2 

, and Q’ be three N x N x N cubic 

regions. Let d be the unknown displacement between I 1 

and I 2 

of a feature point p on which 

Q 1 

is centred. Feature point p is centre point of the shape segmented and dened 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 Q 2 

is estimated through constant ow, 

and added to the current displacement estimate (initially set to 0). Second, the patch Q 2 

is 

wrapped according to the estimated ow. This means that Q 2 

is displaced according to the 

estimated ow, and the resulting patch, Q’, is restored in the voxel grid of frame I 2 

. If the 

estimated ow normalized in time equals (v x 

, v y 

, v z 

), the grey value at voxel (i, j, k) of Q’ 

can be obtained from the values of the voxels of Q 2 

close to (i – v x 

, j – v y 

, k – v z 

). 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 Q 1 

on p. 

2. Estimate the displacement d 0 

of p, centre of Q 1 

, and let d = d + d 0 

. 

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 I i-1 

and I i-2 

, I i-2 

and I i-3 

and so on, can be used to make predictions on the disparities 

between I i-1 

and I i 

, before observing frame I i 

. 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 ti 

= ti 

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, [ ] T 

p = x , y , z is considered, in the frame acquired at instant t i 

, moving 

i 

i 

i 

i

172 


with vector eld velocity v [ ] T 

i 

= vx, i 

, v 

y, 

i 

, v , the motion on the cubic plane is described 

z, 

i 

with the state vector s [ ] T 

i 

= xi 

, yi 

, zi 

, vx, i 

, vy, 

i 

, v . 

z, 

i 

i 

A physical system is modelled by a state vector s , often called simply the state, and a set 

s 

of equations, called the system model. The state is a time-dependent vector, i 

is stated 

as s ( t i 

) , the components of which are system variables, in a sufcient number to capture 

the dynamic properties of the system. The system model is a vector equation describing 

the evolution of the state in time. Assuming a sufciently small sampling interval (and 

therefore constant feature velocity between the frames), the system model can be written as 

p ξ (6) 

i 

= pi− 1 

+ vi− 

1 

+ 

i−1 

v η (7) 

= + 

i 

v i − 1 i−1 

where ξi− 

1 and η 

i−1 

are zero-mean, random, vector-modelling additive system noise. In 

terms of the state vector s 

i rewrites si 

= φ 

i− 1si−1 

+ wi− 

1 with 

⎡1 

0 0 1 0 0⎤ 

⎢ 

⎥ 

⎢ 

0 1 0 0 1 0 

⎥ 

⎢0 

0 1 0 0 1⎥ 

φ 

i−1 

= ⎢ 

⎥ 

(8) 

⎢0 

0 0 1 0 0⎥ 

⎢0 

0 0 0 1 0⎥ 

⎢ 

⎥ 

⎣0 

0 0 0 0 1 

⎦ 

and 

w i −1 

⎡ξ 

i−1 

⎤ 

= ⎢ ⎥ 

⎣η 

i−1 

⎦ 

. (9) 

The difference between the predicted state and the real state is measured with the state 

correction via the lter gain factor K t 

, 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 dened 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


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 classication 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 proles. 

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 

determined by computing the magnitude gradient components G x = 

∂G 

∂x 

, G y = 

∂y 

∂G 

and G z = 

∂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 

2 2 2 

d ( x, 

y, 

z) 

Gx 

+ Gy 

Gz 

e = + 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 specied by angles and ϕ

174 


that can be computed by equations 

G y 

tan θ = 

(10) 

Gx 

and 

Gz 

tanϕ = 

. (11) 

( G ) 2 + ( G ) 2 

x 

y 

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, e 1 

, e 2 

and e 3 

, by angles , and 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, R j 

, rotating vectors around e j 

, 

that is, 

⎡1 

0 0 ⎤ 

R ( ⎢ 

⎥ 

1 

α) 

= 0 cosα 

− sinα 

⎢ 

⎥ 

(12) 

⎢⎣ 

0 sinα 

cosα 

⎥⎦ 

⎡ cos β 0 sin β ⎤ 

R 

⎢ 

⎥ 

2( β ) = 0 1 0 

(13) 

⎢ 

⎥ 

⎢⎣ 

− sin β 0 cos β ⎥⎦ 

⎡cosγ 

− sin γ 0⎤ 

R = 

⎢ 

⎥ 

3( 

γ ) sin γ cosγ 

0 . (14) 

⎢ 

⎥ 

⎢⎣ 

0 0 1⎥⎦


The matrix R describing the overall rotation is the product of the R j 

(the order of overall 

multiplications matters; different sequences give different results with the same triplet of 

angles): 

R = R R 

1 2R3 

= 

⎡ cos β cosγ 

= 

⎢ 

sinα 

sin β cosγ 

+ cosα 

sinγ 

⎢ 

⎢⎣ 

− cosα 

sin β cosγ 

+ sinα 

sinγ 

− cos β sinγ 

− sinα 

sin β sinγ 

+ cosα 

cosγ 

cosα 

sin β sinγ 

+ sinα 

cosγ 

sin β ⎤ 

− sinα 

cos β 

⎥ 

⎥ 

cosα 

cosγ 

⎥⎦ 

(15) 

Second parameterization is as follows. According to Euler’s theorem, any 3D rotation 

can be described as a rotation by an angle, , around an axis identied by a unit vector 

r 

n = [ n n n ] T . The corresponding rotation matrix R can then be obtained in terms 

1 2 3 

of 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 

2 2 2 

the constraint that n has a unit norm, that is by dividing each n i 

by n 

1 

+ n2 

+ n3 

[5]. 

The matrix R in terms of and n is given by 

2 

⎡ n1 

n1n2 

n1n3 

⎤ ⎡ 0 − n3 

n2 

⎤ 

⎢ 

2 ⎥ 

R = I cosθ + ( 1 − cosθ 

) 

+ 

⎢ 

− 

⎥ 

⎢ 

n2n1 

n2 

n2n3⎥ 

sinθ 

n 0 

⎢ 

3 

n1⎥ 

(16) 

2 

⎢ 

⎥ 

⎣n3n1 

n3n2 

n3 

⎦ ⎢⎣ 

− n2 

n1 

0 ⎥⎦ 

Conversely, both and n can be obtained from the eigenvalues and the eigenvectors of R. 

The three eigenvalues of R are 1, cosθ 

+ i sin θ and cosθ 

− 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 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 p 1 

, p 2 

, ..., p n 

, 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 p [ ] T 

ij 

= xij 

, yij 

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

ij 

and think of the x ij 

, y ij 

and z ij 

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 

⎥ 

⎢ ⎥ . (17) 

⎢⎣ 

Z ⎥⎦

176 


For use in the rank theorem, the mean of the entries on the same row is subtracted from 

each x ij 

, y ij 

and z ij 

: 

~ xij 

= xij 

− x 

(18) 

~ i 

yij 

= yij 

− y 

(19) 

~ i 

(20) 

z 

ij 

= z 

ij 

− z 

i 

where 

1 

x = 

n 

∑ 

i 

x ij 

n j= 

1 

1 

y = 

n 

∑ 

i 

y ij 

n j= 

1 

(21) 

(22) 

1 

z = 

n 

∑ 

i 

z ij 

n j= 

1 

are the coordinates of p 

i , the centroid of the voxels in the i-th frame. Furthermore, x~ , 

ij 

y~ 

ij 

and z~ can be denoted as entries of three N x n matrices, X ~ , Y ~ and Z ~ , and form the 3N x 

ij 

n matrix W, called the registered measurement matrix: 

(23) 

~ 

⎡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 p i 

. 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


the centroid of p 1 

, ..., p n 

(Figure 2), and let i i 

, j i 

and k i 

denote the unit vectors of the i-th 

frame (at time instant i). It can be seen from Figure 2 that 

x 

y 

z 

ij 

ij 

ij 

T 

i 

( p − d ) 

= i 

(25) 

T 

i 

j 

( p − d ) 

j 

i 

= j 

(26) 

T 

i 

( p − d ) 

j 

i 

i 

= k 

(27) 

where d i 

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, 

n 

1 

∑ p j 

= 0 . (28) 

n 1 

j= 

Now, it can be rewritten 

n 

1 T 

T 

( p − d ) − i ( p − d ) ii 

p 

j 

~ x = 

T 

ij 

= ii 

j i ∑ i 

n m=1 

n 

1 T 

T 

( p 

j 

− di 

) − ji 

( pm 

− di 

) ji 

p 

j 

~ y = 

T 

ij 

= ji 

∑ 

n m=1 

n 

1 T 

T 

( p 

j 

− di 

) − ki 

( pm 

− di 

) ki 

p 

j 

~ z = 

T 

ij 

= ki 

∑ 

n m=1 

Therefore, the rotation matrix R size of 3N x 3 can be dened as 

m 

i 

(29) 

(30) 

. (31) 

T 

⎡i 

⎤ 

1 

⎢ ⎥ 

⎢ M ⎥ 

⎢ 

T 

i ⎥ 

n 

⎢ T ⎥ 

⎢ 

j1 

⎥ 

R = ⎢ M ⎥ 

⎢ T ⎥ 

⎢ jn 

⎥ 

T 

⎢k 

⎥ 

1 

⎢ ⎥ 

⎢ M ⎥ 

⎢ 

T 

⎥ 

⎣kn 

⎦ 

(32) 

and a 3 x n shape matrix C as 

[ p K ] 

C = (33) 

so it can be written that 

1 

p n 

~ 

W = 

RC .

178 

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 ow algorithm. 

Rotations in space in time are approximated with the 3D rotation methods. The optical 

ow 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 

ow algorithm (Figure 4). The resulting displacement between the rst and the 

second frame is computed as d = [ ( i − vx 

), ( j − vy 

)( , k − vz 

) ] = [ 153,2,0] 

2 

2 

2 

d ( i − v ) + ( j − v ) + ( k − v ) = 153, 013 

= 

x 

y 

z 

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

and 

. The resulting displacement d is 

Figure 4 Example of motion between three consecutive frames. Motion between the second and the 

third frame is determined by the constant ow algorithm.


Figure 5 Example of motion between three consecutive frames. Motion between the second and the 

third frame is determined by the optical ow. 

In Figure 5 the same procedure of motion estimation between the second and the third frame 

is repeated by using of the optical ow algorithm. There is a signicant visual difference 

between the estimated and real position of 3D object. Also, in case of the constant ow 

algorithm, we have to code only the displacement d which is signicantly less (for three 

consecutive frames 151 kB) than coding the whole optical ow 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 p = [ 517 ,38, 92] T 

and [ ] T 

i 

v = − 5,0, 1 , the motion on the 

volume plane is described with the state vector [ ] T 

i 

i 

s = 517 ,38,92, −5,0, 

1 . Estimated

180 


state vector s 

i+ 1 , with assumed = 1 

si 

1 

= 512 ,38,94, −52,6, 

−14 

Estimated state vectors for consecutive 8 frames are stated below: 

ΔT , gives values [ ] T 

[ 512 ,38,94, 52,6, ] T 

+ 1 

= − s = [ 460,44,108, −37,2, 

1] T 

, 

s 14 

i 

i+ 

2 

− 

[ ] T 

, s 13 

i+ 3 

= 423,46,107, −11, 

−12, 

s [ ] T 

, 

i+4 = 412 ,34,120,1,0, 2 

[ ] T 

, s 1 

i+ 5 

= 413 ,34,122,9,6, − s [ ] T 

, 

i 6 

= 422,40,121,17 ,8, 5 

s [ ] T 

i+ 7 

= 439,48,126,6,24, −2 

s [ ] T 

, 

i 8 

= 445,72,124,0,0, 0 

+ 

, 

+ 

. 

+ 

. 

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 specic 

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


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 = R R 

1 2R3 

= 

cos( −13° 

) cos( 29° 

) − cos( −13° 

) sin( 29° 

) sin( −13° 

) 

( 8° 

) sin( −13° 

) cos( 29° 

) + cos( 8° 

) sin( 29° 

) − sin( 8° 

) sin( −13° 

) sin( 29° 

) + cos( 8° 

) cos( 29° 

) − sin( 8° 

) cos( −13° 

) 

( 8° 

) sin( −13° 

) cos( 29° 

) + sin( 8° 

) sin( 29° 

) cos( 8° 

) sin( −13° 

) sin( 29° 

) + sin( 8° 

) cos( 29° 

) cos( 8° 

) cos( 29° 

) 

⎡ 

⎢ 

⎢ 

sin 

⎢⎣ 

− cos 

⎡0,1478 

− 0,1478 

= 

⎢ 

⎢ 

0,8509 0,8813 

⎢⎣ 

0,2623 − 0,0015 

− 0,225 ⎤ 

− 0,1356 

⎥ 

⎥ 

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 

T 

⎡− 

2,84 ⎤ ⎡i 

⎤ 

1 ⎡− 

0,27 − 0,76 0,9 ⎤ 

⎢ ⎥ ⎢ T ⎥ ⎢ 

⎥ 

⎢ 

−1,13 

⎥ ⎢ j1 

⎥ ⎢ 

− 0,73 0,58 − 0,24 

⎥ 

⎢− 

2,13 ⎥ ⎢ 

T 

k ⎥ 

1 ⎢− 

0,82 0,24 0,31 ⎥ 

⎢ ⎥ ⎢ ⎥ 

T ⎢ 

⎥ 

⎢− 

3,065⎥ 

⎢i2 

⎥ ⎢− 

0,35 − 0,69 0,75 ⎥ 

⎡ 4 ⎤ 

~ 

W = ⎢ − 0,55 ⎥ , ⎢ T 

R = j 

⎥ 

= ⎢− 

0,69 0,72 − 0,31 ⎥ and C = 

⎢ ⎥ 

⎢ ⎥ ⎢ 

2 

⎥ ⎢ 

⎥ 

⎢ 

3,5 

⎥. 

T 

⎢− 

2,875 ⎥ ⎢k 

⎥ ⎢− 

0,91 0,15 0,24 

⎢ 

2 

⎥ 

⎣ 1 ⎥⎦ 

⎢− 

3,515 ⎥ ⎢ 

T 

⎥ ⎢− 

− 

⎥ 

⎢ ⎥ ⎢i 

⎥ 0,49 0,61 0,58 

3 

⎢ 

⎥ 

⎢ 0,015 ⎥ ⎢ T 

j ⎥ ⎢− 

0,59 0,79 − 0,39 

3 

⎥ 

⎢ ⎥ ⎢ ⎥ 

T 

⎣− 

3,515 

⎢ 

⎥ 

⎦ ⎢ ⎥ ⎣− 

0,97 0,07 0,12 

⎣k3 

⎦ 

⎦

182 


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 ow 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.


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 ow 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 


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 

nediljka.gaurina-medjimurec@rgn.hr, Tel. +385 1 5535 825 

Pašic Borivoje 



borivoje.pasic@rgn.hr, Tel. +385 1 5535 840 

Matanović Davorin 



davorin.matanovic@rgn.hr, 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 uid in a wellbore 

annulus to free differentially stuck pipe. Spotting uid 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 uids has been determined in laboratory 

using sticking tester as well as inuence of spotting uids on freeing differentially stuck 

pipe. Results of the testing are presented in the paper. 

Key words: differential sticking, stuck pipe, spotting uid, lter cake, drilling 


Pipe sticking is, for most drilling organizations, the greatest drilling problem worldwide. 

It results in a signicant 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


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 (Bushnell- 

Watson 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 

DOWN AND 

ROTATE 

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 

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 denition 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 signicant 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 


The pipe is held in the cake by a difference in pressures (P) between the hydrostatic 

pressure of the mud (P m 

) and the formation (pore) pressure (P f 

) in the permeable zone. 

This pressure (P) 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 signicant 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 signicant 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 = P·A· (Eq. 1) 

where 

F - pullout force, N 

P - differential pressure, Pa 

A - contacta area, m 2 

- coefcient 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 coefcient of friction 

() increases with time. 

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


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 P 

(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 inuenced 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), 

specic mud composition, and uid 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 uid properties in attempts to minimize the risk of sticking, and 

• Select a drilling uid that will yield smooth lter cake with low coefcient 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 coefcient of friction of drilling 

uid 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 


the adhesion of steel to the mud cake, (b) be incorporated into the lter cake and provide 

better uid-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 uid 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 uid in a wellbore annulus. Before 

placement of spotting uid, 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 oor. Knowing the stretch L and the forces applied F 1 

and F 2 

, Hooke’s 

law, the length of the drill string from the surface to the free point (L f 

) 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 (m 2 ), 

• L 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 uid (spot) is a small volume or pill of any substance, oil or water base, that is 

positioned in the wellbore to achieve a specic purpose. Most service companies provide 

multiple spotting uid options. The purpose of the spotting uid is to dissolve or break 

down the lter cake so the pipe can be freed. Possible mechanisms of spotting uid 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 uids 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 uid. When pipe sticking


occurs, pupming of spotting uid 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 uid (Fig. 4d) and drill collar is freed (Fig. 4e). 

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

Spotting uids could be devided into four main categories: (a) water-based spotting uids, 

(b) diesel-based spotting uids, (c) synthetic-based spotting uids, and (d) acid based 

spotting uid. They could be unweighted or weighted (spotting uid, viscosier and 

weighting material) and consist of detergents, soaps, oils, surfactants and other chemicals 

(wall cake cracking material). Oil-based mud is the traditional spotting uid (Krol, 1981; 

Ayers et al., 1989). Because of concern about mud disposal, spotting uids used offshore 

are either synthetic-based uids or benign water-based uids (such as drill-in uids 

and salt solutions) (Kercheville et al.,1986). Drill-in uids are low and ultra-low solids 

uids; 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 coefcient 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 uid, 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 coefcient of drilling 

uids, and also determines how effective lubricants or treatments might be with any given 

drilling uid.

190 


Figure 5. OFITE Differential sticking tester 

This coefcient 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 coefcient is 

obtained and read directly at the end of the test. How likely a given uid 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 coefcient (K sc 

) is calculated by dividing the sliding 

force (F s 

) by the normal force (F n 

): 

K sc 

= F s 

/ F n 

(Eq. 3) 

For radius of plate r = 25.4 mm (1”): 

The sliding force (F s 

) of the plate is a function of the measured torque (T u 

). 

F s 

= 1.5 × T u 

(Eq. 4) 

The normal force (F n 

) 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 × r 2 (Eq. 5) 

Finally, the bulk sticking coefcient (Ksc) is equal to: 

K sc 

= 0.001 × T n 

(Eq. 6)


Where T u 

is torque reading (lb f 

-inch or 0.1129 N·m). 

The stuck tendency coefcient (K st 

) is equal to the bulk sticking coefcient (K sc 

) multiplied 

by the variable stuck area (Eq. 7): 

K st 

= K sc 

× (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 uids (Gaurina-Meimurec et al., 2010.). Formulations 

of tested lignosulphonate mud is shown in Table 1. To determine inuence 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 uid loss, cake thickness, rheological properties and pH value were determined 

according to API RP13B (API Recommended Practice Standard Procedure for Testing 

Drilling Fluids). Spotting uid 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 Units C0L0 C3L0 C5L0 C0L20 C3L20 C5L20 C0L40 C3L40 C5L40 

Water ml 1000 1000 1000 1000 1000 1000 1000 1000 1000 

Bentonite g L -1 80 80 80 80 80 80 80 80 80 

FCL g L -1 20 20 20 20 20 20 20 20 20 

NaOH g L -1 3 3 3 3 3 3 3 3 3 

CMC g L -1 0 3 5 0 3 5 0 3 5 

Viscosifer g L -1 5 5 5 5 5 5 5 5 5 

Biocide g L -1 0,6 0,6 0,6 0,6 0,6 0,6 0,6 0,6 0,6 

Lubricant ml L -1 0 0 0 20 20 20 40 40 40 

Defoamer ml L -1 1 1 1 1 1 1 1 1 1 

Barite g L -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 specied time decreases with increasing 

concentration of CMC and lubricant. In addition, torque decreases after placement of 

spotting uid for16-hours (Fig. 6).

192 


Figure 6. Effect of time and additives on torque 

Bulk sticking coefcient after 300 min sticking time is decreased with increasing 

concentration of CMC and lubricant. For instance, the value of bulk sticking coefcient of 

C5L40 mud is 35,3 % lesser than of C0L0 mud, and for C5L40 bulk sticking coefcient is 

2,3 times less after 16 hours spotting time (Fig. 7). 

Figure 7. Effect of CMC and lubricant on bulk sticking coefcient 

Bulk sticking coefcient is decreased with increasing concentration of CMC but increased 

with time regardless of concentration of CMC (mud without lubricant) (Fig 8).


Figure 8. Effect of time and concentration of CMC on bulk sticking coefcient 

Bulk sticking coefcient 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 

coefcient 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 coefcient 

The corelation between torque after 60 min and mud properties (uid loss, cake thicknes, 

plastic viscosity, gel strengts and pH value) is observed. The lower uid loss value the 

lower value of torque (Fig. 10). In addition, API Fluid loss decreases with addition of CMC

194 

and lubricant. 


Figure 10. The corelation between torque and uid 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


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 


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


occurs, and to select and implement the most effective treatment option. 

According to performed tests torque and bulk sticking coecient constantly increase with 

sticking time regardless of concentration of CMC and lubricant in tested lignosulphonate 

mud, but for any specied 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 uid for16-hours torque signicantly decreases. In addition, the corelation between 

mud composition, torque and mud properties is observed. API uid 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 inuence on values of the torque and sticking coefcient (reducing sticking 

area). The lower uid loss and the thiner cake the lower values of the torque and the bulk 

sticking coefcient. 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-Meimurec, N., Paši, B., Matanovi, 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 


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-Meimurec, N.; Paši, 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.


Use of Plasma Technology for Modification of Textiles 

Prof. dr. Ružica Čunko 

Faculty of Textile Technology, University of Zagreb 

Prilaz baruna Filipovića 28a, Zagreb 

rcunko@ttf.hr 

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 modication 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 dening the impacts of these changes on the obtained nal properties of 

textile materials treated in this way. SEM and AFM results conrmed 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 conrm that the low-pressure 

plasma treatment is an acceptable and appropriate method of textile surface modication 

and an eco-friendly method at the same time. 

Keywords: low-pressure plasma, surface functionalization, cellulose fabric modication. 


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 efuents, 

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 signicant 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 modication has become more and more signicant. 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 


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 modication 

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 modied by conventional methods. It is 

important to note, that plasma agency is limited to the top surface layer and no signicant 

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 efuents. 

The rst investigation of possibilities and phenomenology of textiles modication 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 identied (as “radiant matter”) by Sir W. Crookes in 1879; Sir J. J. Thomson 

identied the nature of the matter in 1897, and in 1928. I. Langmuir has introduced the 

term “plasma”. In physics plasma is dened 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 insufcient 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.


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 classied 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 


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, ax, 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 

modied through reactions of esterication and etherication. 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 modication using cold plasma differs signicantly from the traditional 

esterication and etherication 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 unmodied 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


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 specic 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 

signicant 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 dening 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 

Yarn 

neness 

(warp/ 

weft) / tex 

Numb. of threads 

(warp/weft) / cm 

Volume mass / gcm -3 

Mass per unit area / 

gm -2 

UT O 2 

-p Ar-p UT O 2 

-p Ar-p 

Cotton 

(CO sc) 

Lyocell 

(CLY I) 

Modal 

(CMD I) 

29.4/25.0 25/20 0.369 0.367 0.362 130.6 128.1 128.3 

31.3/25.0 

21/19 

0.409 0.431 0.444 120.1 126.8 129.8 

31.3/25.0 23/19 0.420 0.429 0.444 121.8 123.9 126.0 

UT = untreated sample, O 2 

-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 m 3 /hour, with an 

electromagnetic valve, which prevented the return of the oil vapor into the vacuum chamber. 

Brass needle valves (max. ow rate of 400 sccm), with pressure reducers specied for each 

type of gases applied, provided constant gas ow rate during plasma processes performed. 

The plasma treatments were controlled fully automatically.

204 


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 ow 

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. 

Magnication 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 deection 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 deection are electronically converted 

into pseudo three-dimensional (3D) image of a sample.


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 m scanner. Contact mode imaging was performed using silicon-nitride tips (NP- 

20, 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 152 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 DCR- 

DVD304E. 

The Drop test was carried out as a measure of absorbency of fabric samples. The absorbency 

is dened 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 dened 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 



Surface analysis by Scanning Electron Microscopy 

SEM analysis of the lyocell and modal cellulose bres in the fabric sample surface were 

carried out with magnication 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 specic 

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.


a 

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 m × 25 m with vertical scale of 3000 nm). 

Vertical distance marked on the c: approx. 122 nm 

b 

a 

Figure 8 The AFM of oxygen plasma treated lyocell bres: 


axis of tested bre (imaged area is 15 m x 15 m with vertical scale of 1500 nm) 

Vertical distance marked on the c: approx. 24 nm 

b 

a 

b 

Figure 9 The AFM of argon plasma treated lyocell bres: 


axis of tested bre (imaged area is 10 m × 10 m with vertical scale of 2000 nm) 

Vertical distance marked on the c: size range from approx. 23 nm to 59 nm

208 


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 conrmed 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. 

Height of the water front, h [mm] 

150 

140 

130 

120 

110 

100 

90 

80 

70 

60 

50 

40 

CLY I ut (warp) 

CLY I O2 plasma, 5 min (warp) 

30 

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

20 

CLY I Ar plasma, 5 min (weft) 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



150 

140 

130 

120 

110 

100 

90 

80 

70 

60 

50 

40 

30 

20 

10 

CMD I ut (weft) 

CMD I Ar plasma, 5 min (warp) 

CMD I O2 plasma, 5 min (weft) 

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] 

CMD I O2 plasma, 5 min (warp) 

CMD I Ar plasma, 5 min (weft) 

CMD I ut (warp) 

Figure 11 Results of the capillary rise of water measured on untreated and plasma treated modal 

fabrics in warp and weft direction 


150 

140 

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 


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 

O 2 

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 supercial 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 inuence 

on the bre surface characteristics and their effects on the properties modication 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 specic area (ratio: surface area/volume). 

• The AFM results conrmed 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 conrm 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


properties. 

• It could be concluded that presented investigation results conrm that the plasma 

treatment is an acceptable and appropriate method of textile surface modication, 

avoiding bre damage, and an eco-friendly method at the same time. 

References 

1. AATCC Test Method 79:2000 - Drop test 

2. 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 

3. Demir A, Özdogan 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 

4. Denes F, Young RA (1998) Surface modication of polysaccharides under coldplasma 

conditions. In: Dumitriu S (Ed.) Structural Diversity and Functional Versatility of 

Polysaccharides, New York, Basel, Hong Kong, Marcel Dekker Inc 

5. Ercegovi Raži S, unko R, Bautista L, Mota J, Crespo L (2010) Ageing Effect 

on Wettability Properties of Low-Pressure Plasma-Treated Cellulosic Fabrics. In: 

Dragevi 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 

6. Ercegovi Raži S (2010) Ciljana modikacija svojstava tekstilija primjenom plazme 

i metalnih spojeva. PhD Thesis, University of Zagreb, Faculty of Textile Technology, 


7. Ercegovi Raži S, unko R (2009) Modikacija svojstava tekstilija primjenom 

plazme. Tekstil 58 (3) pp 55-74 

8. Ercegovi Raži S, unko R, Svetlii V, Šegota S (2011) Application of AFM 

Microscopy for Identication of Fibres Surface Changes after Plasma Treatments. 

Materials Technology (accepted for publishing) 

9. Ercegovi Raži S, unko 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 

10. Guimond S, Hanselmann B, Amberg M, Hegemann D (2010) Plasma processing of 

textiles: Perspectives. Melliand International 16 (4) pp 182-183 

11. Hauser P J, El-Shafei A (2011) Atmospheric pressure plasma treatments for repellent 

textiles. AATCC Review 11 (1) pp 70-74 

12. Horrocks AR, Nazaré S, Masood R, Kandola B, Price D (2011) Surface modication 

of fabrics for improved ash-re resistance using atmospheric pressure plasma 

in the presence of a functionalized clay and polysiloxane. Polymers for Advanced 

Technologies 22 (1) pp 22-29 

13. http://en.wikipedia.org/wiki/File:Plasma-lamp_2.jpg (visited 14.01.2011.) 

14. http://www.scientec.fr/include/surface/agilent/options/EN-mode_contact_friction. 

html (visited 14.01.2011.) 

15. 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 

16. ISO 9073-6:200 Textiles – Test methods for nonwovens – Part 6: Absorption (Liquid

212 


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 282- 

298 

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 modication 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 

modication of viscose textile. Vacuum 84 (1) pp 79-82


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 

zaneta.ugarcic-hardi@ptfos.hr 

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 13- 

16g/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 signicant (~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 


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, Drenjanevi-Peri, 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 (Premuži et al, 

2010, Miškulin et al, 2010).

214 


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 im plemented 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) programs designed to encourage the food industry to reformulate products to 

reduce sodium levels on a voluntary basis, 

b) a public relations campaign aimed at in creasing public awareness concerning the 

relationship between sodium intake and HTN risk and ways to reduce sodium in 

take, and 

c) “voluntary” front-of-package “trafc 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 ofcials (Heidolph et al, 2011)


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, Jelakovi, 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 signicant (~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 (Ugari-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 difcult 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%, (20g/kg our = 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 signicant technical effect on the properties of wheat gluten by strengthening 

the gluten network in dough. It inuences 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 Ger man 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 our. Sensory considerations show that salt 

affects the overall avor 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 our 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 prod uct. Cations 

similar to sodium such as potassium work best. However, a bitter or metallic off-avor 

is often sensed when potassium chloride is used. Other mineral salts can partially replace

216 


salt in a cost-effective manner as well, but, they often deliver off-avors which can limit 

their use. Replacement with NH 3 

Cl tastes of plant licorice, and replacement with CaCl 

has a negative effect on dough rheology. Although articial 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 re sults 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. 


Table 2. Salt content in different bread types 

Bread types Salt content (%) 

Wheat French bread (White baguette) 1.75 

Wheat Easter braided bread 1.41 

“Sovital” bread 1.79 

Wheat water bread 1.49 

“Krunovit” bread 1.57 

“Drava vital” bread 1.31 

Family bread 1.30 

“Monastery sun” bread I 1.90 

“Monastery sun” bread II 2.01 

“Grandmother mix” bread 1.01 

Bread with onion 0.96 

Corn baguette 1.76 

White bread with inulin 1.29 

Extra white bread 1.52 

White ciabatta 1.73 

White bread I 1.58 

Roll bread 2.02 

White bread II 1.58 

White bread III 1.03 

White bread IV 2.05 

Brown bread 1.61 

Baguette 1.84 

Water bread 1.29 

Average 1.56


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 (100 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), (Pucarin- 

Cvetkovi, 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 (70 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 stratied 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 Salt content (%) 

Finger stick (Baguette ngers) 4.76 

Pretzel 3.04 

Salt stick 5.98 


Pretzel 2.41 


Roll (Baguette roll) 2.04 


Pretzel 2.14 



Pretzel 3.44 



Pretzel 4.57 




Pretzel 2.26 


Average 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 

(Ugari-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 


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 our. 

In the second step the reduction should be up to 25% (1.5 g salt/100 g our). 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 

dene a way to achieve the reduction in salt level consistent with technological and sensory 

needs. 

6 References 

1. Cauvain SP, Young LS (1998): Technology of Breadmaking. Blackie Academic and 

Profesionals, London 

2. Dietary Guidelines Advisory Committee Draft Report (2010) 

3. Drenjanevi-Peri I (2010) Prekomjeran unos soli kao imbenik rizika za razvoj 

kardiovaskularnih bolesti, Znanstveno - struni skup „Sol i zdravlje“, Faculty of Food 

Technology in Osijek, Osijek, 28. 1. 2010. 

4. 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 

5. Fisher MH, Aitken TR, Anderson JA (1994) Effects of mixing, salt and consistency on 

extensograms. Cereal Chemistry 26, 81–97. 

6. Forschung hilft dem Backgewerbe (1989): Nutzen und Möglichkeiten einer 

natriumreduzierte Ernährung, Bundesforschungsanstalt für Getreide–und 

Kartoffelverarbeitung,, Detmold, pp 10 – 23.


7. FSAI (2010) Salt and health: Review of the scientic evidence and recommendations 

for public policy in Irland 

8. He H, Roach RR, Hoseney RC (1992) Effect of nonchaotropic salts on our breadmaking 

properties. Cereal Chemistry 69 (4), 366–371 

9. Heidolph B B, Ray DK, Koehler P, Weiber, Slocum S, Noort MW (2011): Looking 

for my lost shaker of salt…replacer: avor, funktion, future, Cereal, Food World, Vol. 

56, No.1, 5-19. 

10. Hlynka I (1962) Inuence of temperature, speed of mixing, and salt on some rheological 

properties of dough in the farinograph, Cereal Chemistry 39, 286–303. 

11. IOM-Institute of Medicine (2010) Strategies to reduce sodium intake in the United 

States, Appendix C: International efforts to reduce sodium consumption 

12. IOM- Institute of Medicine (2004) Dietary reference intakes for water, potassium, 

sodium, chloride and sulfate, United States Institute of Medicine. 

13. Jelakovi B (2008) Salt and hypertension, Znanstveni skup, Kardiovaskularno zdravlje, 

Prehrana i sol, Akademija medicinskih znanosti, 31, Zagreb, 21.11.2008, Abstract, 31 

14. Linko P, Harkonen H, Linko YY (1984) Effects of sodium chloride in the processing of 

bread baked from wheat, rye and barley ours. Journal of Cereal Science 2 (1), 53–62 

15. 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 

16. Miškulin M, Periš D, Ugari-Hardi Ž, Dumani G (2010) Skrivena sol u prehrani 

djece školske dobi - ima li mjesta zabrinutost?, 34. struni skup s meunarodnim 

sudjelovanjem, “Zdravstvena ekologija u praksi”, Zagreb, 21.-23. 4. 2010 Proceedings, 

pp 103-109 

17. Premuži V, Erceg I, Jovanovi A, Reiner Ž, Jelakovi B (2010): Unos soli u odrasloj 

populaciji, Hrvatski asopis za javno zdravstvo HCJZ, On.line 6, (21), 

18. Pucarin-Cvetkovi J, Polašek O, Kern J, Vuleti S, (2008) Regional features of 

Croatian nutrition, Znanstveni skup, Kardiovaskularno zdravlje,Prehrana i sol, 

Akademija medicinskih znanosti, Zagreb, 21.11.2008. Abstract, 8 

19. 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. 

20. Salovaara H (1982) Sensory limitations to replacement of sodium with potassium and 

magnesium in bread. Cereal Chemistry 59(5):427 

21. Skupnjak B(2007): Inicijativa za nacionalnu kampanju smanjenja utroška soli, 

Hrvatski asopis za javno zdravstvo HCJZ, Online, Vol 3, 10. 

22. Slumier P (2005) Principles of breadmaking, Amercan Association of Cereal Chemists, 

Inc. St. Paul, Minnesota, USA, pp46-47. 

23. Ugari-Hardi Ž (2007) Salt reduction in Croatian baked products: Yes or no? 

4th International Congrss „Flour-Bread 07“ 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 

24. Ugari-Hardi Ž, Dumani G, Pitlik N, Koceva Komleni D, Juki M, Kuleš A, 

Sabo M (2010) The salt content in bakery products in Osjecko-baranjska County, 

Proceedings of 5th International Congress „Flour-Bread 09“ and 7th Croatian Congress 

of Cereal Technologists, Ugari-Hardi Ž., Faculty of Food Technology Osijek, Josip 

Juraj Strossmayer ty of Osijek, ty of Osijek, 551-556.

Podupirajui lanovi: 

Arhitektonski fakultet, Zagreb, www.arhitekt.hr 

Centar za vozila Hrvatske d. d., www.cvh.hr 

Elektrotehniki fakultet, Osijek, www.etfos.hr 

Energetski institut Hrvoje Požar, Zagreb, www.eihp.hr 

Fakultet elektrotehnike i raunarstva, 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 

Graevinski fakultet, Osijek, www.gfos.hr 

Graevinski fakultet, Rijeka, www.gradri.hr 

Fakultet graevinarstva, arhitekture i geodezije, Split, www.gradst.hr 

Graevinski fakultet, Zagreb, www.grad.hr 

Graki fakultet, Zagreb, www.grf.hr 

Hrvatska elektroprivreda d. d., Zagreb, www.hep.hr 

Hrvatska zajednica tehnike kulture, Zagreb , www.hztk.hr 

Kemijsko-tehnološki fakultet, Split , www.ktf-split.hr 

Konar - Institut za elektrotehniku d. d., Zagreb , www.koncar-institut.hr 

Odašiljai 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 

Sveuilište u Dubrovniku, www.unidu.hr 

Sveuilište J. J. Strossmayera u Osijeku 

Tehniki fakultet, Rijeka, www.riteh.hr 

Tehniko veleuilište u Zagrebu, www.tvz.hr 

Tekstilno-tehnološki fakultet, Zagreb, www.ttf.hr 

Uljanik brodogradilište d. d., Pula, www.uljanik.hr

annual 2010/2011 of the croatian academy of engineering - HATZ

Create successful ePaper yourself

Delete template?

Save as template?