RMZ MATERIALS AND GEOENVIRONMENT MATERIALI IN GEOOKOLJE Volume co .. ... _ . No. Letnik Ljubljana, December 2005 RMZ - Materials and Geoenvironment RMZ - Materiali in geookolje Old title: Rudarsko-metalurški zbornik (Mining and Metallurgy Quarterly), ISSN 0035-9645,1952-1997. Is issued quarterly by the Faculty of Natural Science and Engineering, Ljubljana, the Institute for Mining, Geotechnology and Environment Ljubljana and Premogovnik Velenje, Velenje. Izdaja Naravoslovnotehniška fakulteta Univerze v Ljubljani in Inštitut za rudarstvo, geotehnologijo in okolje Ljubljana, štirikrat letno. Financially supported also by Ministry of Education, Science and Sport of Republic of Slovenia. Pri financiranju revije sodeluje Ministrstvo za šolstvo, znanost in šport Republike Slovenije. Editor-in-Chief (Glavni urednik) Jože Pezdič Editorial Management Jakob Likar Advisory Board Uredniški odbor Evgen Dervarič, Premogovnik Velenje Tadej Dolenec, Univerza v Ljubljani Stevo Dozet, GeoZS, Ljubljana Jadran Faganeli, Univerza v Ljubljani Vasilij Gontarev, Univerza v Ljubljani Mariusz Orion Jedrysek, University of Wroclaw František Kavička, Technical University of Brno Klaus Koch, Technische Universität Clausthal Tomaž Kolenko, Univerza v Ljubljani Jakob Lamut, Univerza v Ljubljani Jakob Likar, Univerza v Ljubljani David John Lowe, British Geological Survey Jernej Pavšič, Univerza v Ljubljani Andrej Paulin, Univerza v Ljubljani Jože Pezdič, Univerza v Ljubljani Simon Pire, Univerza v Ljubljani Esad Prohić, Sveučilište, Zagreb Anton Smolej, Univerza v Ljubljani Janez Stražišar, Univerza v Ljubljani Andrej Šubelj, IRGO Ljubljana France Šušteršič, Univerza v Ljubljani Rado Turk, Univerza v Ljubljani Milivoj Vulić, Univerza v Ljubljani Editorial Office (Uredništvo): Barbara Bohar Bobnar Iztok Anželj Nives Vukič Digital Layout (Priprava za tisk): Tomaž Sterniša s.p., Ljubljana Print (Tisk): R-TISK d.o.o, Ljubljana RMZ - Materials and Geoenvironment Aškerčeva cesta 12, p.p. 312 1001 Ljubljana, R. Slovenija Telefon: +386 (0)1 470 45 00 Telefaks: +386 (0)1 470 45 60 Indexation bases of RMZ-M&G: CA SEARCH-Chemical Abstracts, METADEX, GeoRef, Energy Science and Technology, PASCAL. Main of 23 bases registrated Baze v katerih je RMZ-M&G indeksiran: CA SEARCH-Chemical Abstracts, METADEX, GeoRef, Energy Science and Technology, PASCAL. Najpomembnejši med 23 registriranimi The authors themselves are liable for the contents of the papers. Za mnenja in podatke v posameznih sestavkih so odgovorni avtorji. _Copyright © 2001 RMZ - MATERIALS AND GEOENVIRONMENT Naklada 700 izvodov. Printed in 700 copies. Letna naroč za posameznike: 4.000,00 SIT, za institucije: 8.000,00 SIT Yearly subscribtion 20 EUR, institutions 40 EUR Tekoči račun pri Novi Ljubljanski banki, d.d. Ljubljana: UJP 01100-6030708186 Davčna številka: 24405388 ISSN 1408-7073 RMZ - MATERIALS AND GEOENVIRONMENT PERIODICAL FOR MINING, METALLURGY AND GEOLOGY RMZ - MATERIALI IN GEOOKOLJE REVIJA ZA RUDARSTVO, METALURGIJO IN GEOLOGIJO RMZ-M&G, Vol. 52, No. 4 pp. 661-825 (2005) Ljubljana, December 2005 Historical Review More than 80 years have passed since in 1919 the University Ljubljana in Slovenia was founded. Technical fields were joint in the School of Engineering that included the Geologic and Mining Division while the Metallurgy Division was established in 1939 only. Today the Departments of Geology, Mining and Geotechnology, Materials and Metallurgy are part of the Faculty of Natural Sciences and Engineering, University of Ljubljana. Before War II the members of the Mining Section together with the Association of Yugoslav Mining and Metallurgy Engineers began to publish the summaries of their research and studies in their technical periodical Rudarski zbornik (Mining Proceedings). Three volumes of Rudarski zbornik (1937, 1938 and 1939) were published. The War interrupted the publication and not untill 1952 the first number of the new journal Rudarsko-metalurški zbornik - RMZ (Mining and Metallurgy Quarterly) has been published by the Division of Mining and Metallurgy, University of Ljubljana. Later the journal has been regularly published quarterly by the Departments of Geology, Mining and Geotechnology, Materials and Metallurgy, and the Institute for Mining, Geotechnology and Environment. On the meeting of the Advisory and the Editorial Board on May 22nd 1998 Rudarsko-metalurški zbornik has been renamed into "RMZ - Materials and Geoenvironment (RMZ -Materiali in Geookolje)" or shortly RMZ - M&G. RMZ - M&G is managed by an international advisory and editorial board and is exchanged with other world-known periodicals. All the papers are reviewed by the corresponding professionals and experts. RMZ - M&G is the only scientific and professional periodical in Slovenia, which is published in the same form nearly 50 years. It incorporates the scientific and professional topics in geology, mining, and geotechnology, in materials and in metallurgy. The wide range of topics inside the geosciences are wellcome to be published in the RMZ -Materials and Geoenvironment. Research results in geology, hydrogeology, mining, geotechnology, materials, metallurgy, natural and antropogenic pollution of environment, biogeochemistry are proposed fields of work which the journal will handle. RMZ - M&G is co-issued and co-financed by the Faculty of Natural Sciences and Engineering Ljubljana, and the Institute for Mining, Geotechnology and Environment Ljubljana. In addition it is financially suported also by the Ministry of Education Science and Sport of Slovenian Government. Editor in chief Table of Contents - Kazalo Simulation of microbiological pollution in the unsaturated zone of karstified limestone aquifers - tracing with bacteriophages Bricelj, M., Čenčur Curk, В........................................................................................................661 Comparative Analysis of Single Well Aquifer Test Methods on the Mill Tailing Site of Boršt Zirovski vrh, Slovenija Primerjalna analiza metod obdelave hidravličnih poizkusov v črpanem vodnjaku na odlagališču hidrometalurške jalovine Boršt, Zirovski vrh, Slovenija Ratej, J., Brenčič, M.....................................................................................................................669 Investigations of flow system and solute transport at an urban lysimeter at Union Brewery, Ljubljana, Slovenia Proučevanje tokovnega sistema in prenosa snovi v urbanem lizimetru Pivovarne Union, Ljubljana, Slovenija Trček, В........................................................................................................................................685 Vpliv mineralne sestave in strukture na obstojnost apnencev kot naravnega kamna The weathering durability of limestones as a function of their mineral composition and texture Jarc, S., Mirtič, В.........................................................................................................................697 Razlikovanje apnencev s pomočjo statističnih metod The distinction of limestone by statistical methods Jarc, S............................................................................................................................................711 Ustrezne analize črpalnih poizkusov v razpoklinskih vodonosnikih Appropriate analysis methods of pumping tests in fractured aquifers Verbovšek, T.................................................................................................................................723 Modeliranje napajanja vodonosnika v zaledju izvira Rižane z območja Brkinov Modelling the recharge of the aquifer in the Rižana catchment from Brkini area Janža, M........................................................................................................................................737 Toplo stiskanje jekla za poboljšanje CF53 Hot compression of CF53 tempering steel Terčelj, M., Perus, I., Kugler, G., Turk, R.................................................................................753 Feeding behaviour of graphite containing material Wojtas, H.J....................................................................................................................................765 Autor's Index, Vol. S2, No. 4..........................................................................................801 Autor's Index, Vol. S2.....................................................................................................802 Instructions to Authors...................................................................................................818 Template...........................................................................................................................820 Number of paper indexing in diferent bases................................................................825 Število indeksiranih člankov v posameznih bazah Simulation of microbiological pollution in the unsaturated zone of karstified limestone aquifers - tracing with bacteriophages Mihael Bricelj1, Barbara Čenčur Curk2 'NIB - National Institute of Biology, Slovenia; mihael.bricelj@nib.si 2IRGO - Institute for Mining, Geotechnology and Environment, Slovenia; E-mail: barbara.cencur@irgo.si Received: June 30, 2005 Accepted: November 24, 200S Abstract: The purpose of the research was to study the infiltration and migration of health-hazardous human viruses, such as enteroviruses, in the unsaturated zone of fractured and karstified rock, since these rocks present important aquifers in Slovenia. As a possible model for behavior of health-hazardous viruses, we used the salmonella bacteriophage P22H5. After injection, bacteriophages remain in the fractures (channels) and microfracture systems of the unsaturated zone and are rinsed by subsequent larger precipitation events even up to several months after the injection. The field experiments have shown different flow patterns depending on the fractured rock structure. In the research area some fast conduits (large fractures or faults) exist where water runs faster than in the total conductive part of the rock. On the other hand the tracer delay in microfracture system areas was observed. Key words: pollutant transport, fractured and karstified rocks, bacteriophage, experimental field site, Sinji Vrh (Slovenia) Introduction The purpose of the research was to study the infiltration and migration of health-hazardous human viruses, such as enteroviruses, in the unsaturated zone of fractured and karstified rock, since these rocks present important aquifers in Slovenia. As a possible model for the behavior of health-hazardous viruses, salmonella bacteriophage P22H5 was used. Phages have served as useful models for the behavior of human enteric viruses in water treatment processes, groundwater viral transport, inactivation and attachment studies on various subsurfaces, because of their similarity to enteric viruses in structure, size, and resistance to inacti-vation (Hedberg & Osterholm, 1993; Harvey & Ryan, 2004). Better knowledge of the pollutant transport and persistence of tracer in environment enables us to determine vulnerability and define protected areas for drinking water resources. The bacteriophage P22H5 is a virulent mutant that propagates in mouse typhoid fever bacteria Salmonella typhimurium and rarely occurs in waters (Seeley, 1982). From previous tracing experiments (Bricelj, 1986) it is well known that coliphages are a common constituent of faecally polluted waters and for this reason are not a suitable tracer, especially in the case of very high dilutions of the tracer, when a high or low background of coliphages may interfere with the tracer curve. The phage tracer P22H5 was injected at ten locations in 14 tracer experiments in running water and into the unsaturated zone in a karstic area where no background of phages for its host bacteria were present (Bricelj, 2003). The tracer experiment was carried out in the subsurface zone, since microbial activity is assumed to be most active in the upper parts of the unsaturated zone. Experimental field site Sinji Vrh A tracer experiment with bacteriophages was performed on the experimental field site Sinji Vrh (Čenčur Curk, 1997), Slovenia. It is located in the unsaturated zone of fractured and karstified Jurassic limestone at the edge of the Trnovski Gozd plateau (mean altitude 900 m a.s.l.), which is an overthrust of carbonate rock over Eocene flysch (Fig. 1). The subvertical Avče fault with a Dinaric direction NW-SE and several parallel faults cross this territory. These faults are interwoven with numerous connecting faults extending in the general direction N-S. Their intensity varies from open wide fractured zones to crushed and broken zones (Janež, 1997). The groundwater horizon lies extremely deep and appears on the surface at the lowest point of the impermeable flysh border (Fig. 1) at altitudes between 219 and 235 m as the karstic Hubelj spring. Its catchment area is estimated to about 50 km2 with an average annual precipitation of 2450 mm (Trček, 2003). The experimental field site Sinji Vrh presents a 340 m long artificial research tunnel, 5 to 25 m below the surface (Fig. 1). An agro-meteorological station has been installed on the surface near the tunnel entrance, where precipitation, evaporation, air temperature, air moisture, solar radiation, wind speed and direction (both at two levels) are continuously measured. A tracer experiment area (Fig. 1) was chosen close to the tunnel entrance on the north-western part. The main dip direction of fractures is NNE-SSW with subvertical dip because of the location of the area within a crushed zone of Avče fault. In the broken zone the tunnel is supported by concrete (Fig. 1). The Jurassic limestone of the tracer experiment area is composed of 99 % calcite and has a south-westerly dip direction and a gentle dip (of 5° to 30°). The unsaturated fractured and karstified limestone has a negligible matrix porosity and very high fracture density with some greater conduits (Veselič & Čenčur Curk, 2001). The injection hole was drilled through the soil cover in order to avoid tracer retardation because of sorption. A special construction for collecting water penetrating through the rock was developed. The water seeping from the ceiling of the research tunnel is gathered in 1.5 m long segments (MP1 - MP10; Fig. 1) with a gathering surface of 2.2 m2. Material and methods Bacteriophage P22H5 and salmonella mouse typhoid fever bacteria NIB22 (LT2 w.t. strain) were obtained from Dr. Miklavž Grabnar, Department of Molecular Biology, Biotechnical Faculty, University of Ljubljana. Figure I. Location of the experimental field site Sinji Vrh (EFS Sinji Vrh) with geological cross section of Trnovo plateau (Janež, 1997; Veselič et al., 2001). Below: Longitudinal section of the EFS Sinji Vrh with tracer tests area and tracer test sampling points in the research tunnel: MPI - MP10 (after Veselič and Cenčur Curk, 2001). The propagation of phages to obtain crude bacteriophage lysates was done by the method described in dissertation thesis of Bricelj @1994). The nutrient media Brain Heart Infusion Broth and Nutrient Agar were from Biolife, Milano. Water samples and phage suspensions were tested for viable phages (plaque forming units - pfu) according to the agar layer method of Adams (1966), using host bacteria as the indicator strains. On 29th September 2003 10.4 liters of tracing solution, composed of 11 tracers (salts, fluorescent dyes, deuterium, micro spheres and bacteriophage), was poured in 10 minutes into the new drilled borehole to the depth of 0.9 m. There were 1.2E+15 plaque forming units (pfu) of bacteriophage P22H5 as a part of tracer cocktail. It should be pointed out, that the phage concentration in injected tracing solution was calculated to predicted concentration at measuring points from the microfracture system, since there is slower flow with higher dispersion of the tracer. Because of that an overdose is reached in fast pathways such as MP4 and MP5 (see structure on Figure I below). Results The phage tracer appeared first at sampling point five (MP5) after 4.1 days with the peak value of 3.1 E+09 pfu/ml (Fig. 2, Table 1 and Table 2). One day later a positive result was obtained at MP4, the phage appeared with the peak value of 1.1 E+08 pfu/ml. The tracer appeared at all sampling points within 22 days. Peak values occurred at the time of appearance at MP2 and MP8. At MP3 the tracer appeared after 8 days, but the peak value was not reached until 50 days. At the furthest sampling points from the injection hole (MP1, MP2 and MP9, MP10) the peak values were within the lowest pfu values (Fig. 2). The first sampling campaign was completed in September 2004, after about one year (Fig. 2): at MP4 after 324 days and at MP5 after 347 days. At that time, there were still 4.2E+02 pfu/ml in MP4 and 9.8E+02 pfu/ml in MP5 in ml of water sample. Unfortunately the samples were not taken in October and November 2004, since at that time were some significant precipitation events. In winter there were no samples due to freezing of the ground and snow cover. The first water, seeping through the system, was obtained after snow melting at the end of March 2005. After 591 days (May 2005) of collecting the samples there Figure 2. The presence of bacteriophage P22H5 at sampling points MPI- MPIO. The concentrations in MPI - M3 and MP6 - MPIO are grouped below and marked alike. They are depicted only for comparison with MP4 and MPS. were positive results at MP4, MP5 and MP6 with the following concentrations: 9.5E+01, 6.2E+02 and 2.0E+00 pfU/ml, respectively (Fig. 2). The sampling of water is still going on, but with lower frequency. The recovery value was calculated only for MP4 and MP5 and was 0.95 % of the injected quantity at both sampling points (0.04 % for MP4 and 0.91 % for MP5). The results of appearance of the phage tracer, the appearance of peak values and time of the last sample containing phage tracer at different sampling points are summarized in Table I. Table 2 presents time sequence of the phage tracer appearance at sampling points. Discussion After the injection of bacteriophages, they remain in the fractures (channels) and microfracture systems of the unsaturated zone and are rinsed by subsequent larger precipitation events even up to year and a half after the injection. Some authors refer that the edges in subsurface structures could be one of the principal causes of charge heterogeneity. Such conditions could permit that negatively charged bacteriophages attach to otherwise repulsive surfaces, especially if the edges of crystals are oppositely charged (Bickmore et al., 2002; Flynn et al., 2004). The results from Sinji Vrh have shown that the unsaturated zone in the fractured and karstified rocks plays an important role in pollution retardation and storage. The rinsing of pollutants to deeper parts of the karst aquifer depends on the saturation rate of the soil and the unsaturated zone (precipitation events). The field experiments have shown different flow patterns depending on the fractured rock appearance peak appearance Last sample containing phages sampling point of tracer value of peak value (last positive result) days pfu/ml days days MPI 7 2.1E+03 11 63 MP 2 7 2.9E+02 7 63 MP 3 22 5.7E+02 50 550 MP4 5 1.1E+08 5 591 MP5 4 3.1E+09 4 591 MP6 8 4.3E+04 11 591 MP 7 13 1.4E+04 31 550 MP 8 7 6.5E+04 7 550 MP9 22 1.7E+03 25 214 MP10 8 4.9E+03 40 177 Table I. Appearance and presence of the phage tracer at MPI to MPIO. The time is in days and quantity of phages in plaque forming units (pfu) in ml of sample. Table 2. Time sequence of the phage tracer appearance at measuring points. days measuring point 4 MP 5 5 MP 4 7 MP 2 7 MP 8 11 MPI 11 MP 6 25 MP 9 31 MP 7 40 MPIO 50 MP3 structure. In the research area some fast conduits (large fractures or faults) exist where water runs faster than in the total conductive part of the rock, as in the case of MP4 and MP5. Tracer delay in microfracture system areas was also observed, especially at MP3, MP7, MP9 and MPIO (see Pig. I and 2). At these points the appearance of the peak value was delayed for 5O, 31, 25 and 4O days respectively. A very low recovery rate is due to the dispersion of the tracer in directions where it could not be sampled and the decay of tracer, dependent upon removal mechanisms such as filtration, sedimentation and irreversible adsorption (Slnton et al., 1997). Recent research with bacteriophages MS2 (Zhuang et al., 2OO3), PRDI (Blanford et al., 2OO5) and TU (plynn et al., POOR} is much more concerned with breakthrough percentage, recovery calculations of peak values of tracer curve, kinetics of virus surface inactivation and analytical models than with the longevity of phages in environment. Some notions on persistence of phages in the environment are referred by Deborde et al., 2OO3, for the phages MS2 and PRDI in floodplain aquifer. The breakthrough curves of phages contained The appearance of phage P22H5 12 & WRL - 14 11 ? WLL-9 10 3 MP10-177 9 I, 0 MP9-914 1 • 1 7 I D MP8 - 550 I II MP7 - 550 o> . =§ 6 I I _U МРЛ - sfts a. E ; I I -1 MH5 - 585 » . I 4 I [J MP4 - 585 I 3 I, J MP3 - 550 I 2 I, 1 MP"1 63 I 1 E- P MPI-63 100 200 300 400 500 days 600 Figure 3. The presence of bacteriophage P22HS at the sampling points MPI - MPIO in Sinji Vrh. The sampling points WRL and WLD, represent the percolation of phage tracer through right (WRL) and left (WLL) soil lysimeter at Wagna experimental field near Graz. long tails, so the slow releases of phages have been observed over a period of more than six months. There is need for further research of mechanisms of phage persistence in the environment and health-related significance of the mechanisms of such processes. The results were compared (Fig. 2) with the results from tracer tests in soil and gravel at References Adams, M.H. (19S9): Bacteriophages, Methods of Study of Bacterial Viruses., 443-522. Interscience Publishers, New York. Bickmore, B.R., Nagy, K.L., Sandlin, P.E., Carter T.S. (2002): Quantifying surface areas of clays by atomic force microscopy. Am. Mineral. Vol. 87, No. 5-6, pp. 780-783. Blanford, W.J., Brusseau, M.L., Yeh, T.C.J., Gerba, C.P., Harvey, R. (2005): Influence of water chemistry and travel distance on bacteriophage PRD-1 transport in a sandy aquifer. Water Research, Vol. 39, No. 11, pp. 2345-2357. Bricelj, M., Kosi, G., Vrhovšek, D. (1986): Tracing with Salmonella-phage P22H5. Steir. Beitr. Z. Hydrologie Vol. 37/38, pp. 296-271. Bricelj, M. (1994): Underground Water Tracing with the Phages of Salmonella typhimurium. Dissertation Thesis, University of Ljubljana, Biotechnical faculty, Department of Biology, 113 p. Bricelj, M. (2003): Microbial Tracers in Groundwater Research. RMZ - Materials and Geoenvironment Vol. 50, No. 1, pp. 67-70. Deborde, D.C., Woessner, W.W., Kiley, Q.T., Ball, P. (1999): Rapid transport of viruses in a flood-plain aquifer. Water research Vol. 33, No. 10, pp. 2229-2238. Cenčur Curk, B. (1997): Experimental field sites as a basis for the study of solute transport in the vadose zone of karstified rock. Acta hydrotechnica Vol. 15/20, 1-111. Bricelj. M, Cenčur Curk. B. (2005): Bacteriophage transport in the unsaturated zone of karstified limestone aquifers. Proceedings of the conference "Karst 2005", in press the Wagna lysimeter station (Austria). The phage tracer was very quickly eliminated from the water trickling through I m soil and 0.5 m gravel and positive results were concluded after nine or fourteen days respectively, in the left and right lysimeter (WLL and WRL). The results in the lysimeter are consistent with the findings of Van E lsas et al., 1991 and Powelson et al., 1991 and Van Cuyk et al., 2001. Flynn, R., Cornaton, F., Hunkeler, D., Rossi, P. (2004): Bacteriophage transport through a finning-upwards sedimentary sequence, laboratory experiments and stimulation. Journal of Contaminant Hydrology Vol. 74, pp. 231-252. Harvey, R.W., Ryan, J.N. (2004): Use of PRD1 bacte-riophage in groundwater viral transport, inacti-vation, and attachment studies. PEMS Microbiology Ecology Vol. 42, pp. 3-16. Hedberg, C.W., Osterholm. M.T. (1993): Outbreaks of Foodborne and Waterborne Viral Gastroenteritis. Clinical Microbiology Reviews Vol. 6, 199-210. Janež, J. (1997): Geological structure and hydrogeological position of the Hubelj spring. In: Karst hydrogeological investigations in south-western Slovenia. Acta carsologica , Vol. XXVI, No. 1, 82-86. Powelson, D.K., Simpson, J.R., Gerba, .CP. (1991): Effect of organic matter on virus transport in unsaturated flow. Applied and Environmental Microbiology. Vol. 57, No. 8, pp. 2192-2196. Seeley, N.D., Primrose, S.B. (1982): The Isolation of the Bacteriophage from the Environment. Journal of Applied Microbiology Vol. 53, pp. 1-17. Sinton, L.W., Finlay, R.K., Pang, L., Scott, D.M. (1997): Transport of bacteria and bacteriophages in irrigated effluent into and through and alluvial gravel aquifer. Water, Air and Soil Pollution Vol. 98, No. 1-2,.17-42. Trček, B. (2003): Epikarst zone and the karst aquifer behaviour: a case study of the Hubelj catchment, Slovenia. Ljubljana. Geološki zavod Slovenije, 1-100. Van Cuyk, S, Siegrist, Я., Ligan, A., Masson, S., Fischer, E., Figueroa, L. (2001): Hydraulic and purification behaviours and their interactions during wastewater treatment in soil infiltration system. Water Research Vol. 3S, No. 4, pp. 9S3-964. Van-Elsas, J.D., Trevors, J.T., Van-Overbeek, L.S. (1991): Influence of soil properties on the vertical movement of genetically-marked Pseudomonas fluorescens through large soil microcosms. Biology andPertility of Soils Vol. 10, No. 4, pp. 249-2SS. Veselič, M., Čenčur Curk, B. (2001): Test studies of flow and solute transport in the unsaturated fractured and karstified rock on the experimental field site Sinji Vrh, Slovenia. In: New approaches characterizing groundwater flow, Seiler & Wohnlich (eds), Balkema, Lisse, pp. 211-214. Veselič, M., Čenčur Curk, B., Trček, B. (2001): Experimental field site Sinji Vrh. In: Tracers in the unsaturated zone = Markierungsstoffe in der ungesättigten Zone, Berg et al. (eds), Beitraege zur Hydrogeologie Vol. S2, pp. 4S-60. Comparative Analysis of Single Well Aquifer Test Methods ______v on the Mill Tailing Site of Boršt Zirovski vrh, Slovenija Primerjalna analiza metod obdelave hidravličnih poizkusov v črpanem vodnjaku na odlagališču hidrometalurške v jalovine Boršt, Zirovski vrh, Slovenija Jože Ratej, Mihael Brenčič Geological Survey of Slovenia, Dimičeva 14, ÌOOO Ljubljana, Slovenia; E-mail: joze.ratej@geo-zs.si, mbrencic@geo-zs.si Received: June 02, 2005 Accepted: November 24, 200S Abstract: Several aquifer tests were performed during hydrogeological investigations on mill tailings Boršt of uranium mine Zirovski vrh and large data set was generated. These results were used for comparative study of several analytical models for hydraulic test evaluation and for comparison of the results provided by them. In the present paper, methods of Jacob (Cooper & Jacob, 1967), Papadopulos et al. (1967), Theis (193S), Hvorslev (19S1) and Cooper et al. (1967) are compared. In most cases the highest values are obtained by Papadopulos method. In some cases results of Hvorslev and Cooper method are up to two and a half decades lower than results of other three methods. This is mostly due to long duration of pumping. The critical values of pumping times, where results of different slug test and pumping test methods coincide, were also defined. Izvleček: Med hidrogeološkimi raziskavami na odlagališču hidrometalurške jalovine Boršt rudnika urana Zirovski vrh je bilo izvedenih več hidravličnih poizkusov, na podlagi katerih smo dobili velik nabor podatkov. Le-te smo uporabili za primerjalno analizo več analitičnih modelov za obdelavo hidravličnih testov. Med seboj smo primerjali metode Jacoba (Cooper & Jacob, 1967), Papadopulos-— et al. (1967), THEis-a (193S), Hvorslev-a (19S1) in CooPER-ja et al. (1967). V večini primerov smo najvišje rezultate dobili s Papadopulosovo metodo. V nekaterih primerih so bili rezultati Hvorsleva in Coopeija za dve in pol dekade nižji od rezultatov ostalih metod, kar je v glavnem posledica trajanja črpanja. Opredelili smo tudi t.i. "kritične čase", torej tiste čase trajanja črpanja, ob katerih se rezultati različnih metod še razmeroma skladajo. Keywords: comparative analysis, slug test, pumping test, single well test Ključne besede: primerjalna analiza, impulzni poizkus, črpalni poizkus, poizkus na črpanem vodnjaku Introduction During in-situ hydrogeological tests, a large number of factors affect the final outcome and only the obvious ones can be included in analytical models in order to ensure its relative simplicity. With comparative analysis one tries to assess analytical models with respect to their behavior in certain conditions, their resilience toward unexpected factor influences, and their ability to obtain representative results. Potential sources of error are usually related to local variation in permeability, leakage through pipe fittings and between pipe and adjacent soil, undetected impervious boundary close to the test, hydraulic fracturing by excessive differences in water heads, soil remolding or clogging or uprising during the test, and time lag in the piezometric responses (Chapuis, 1990). In addition, errors can arise from the selection of analytical model. During hydrogeological investigations on mill tailings Boršt of uranium mine Zirovski vrh (50 km W of Ljubljana) several pumping and slug tests were performed and large data set was generated. The characterization by hydraulic tests was made with the aim to define hydraulic permeability of the mill tailings and bedrock. These results were later used for hydrological balance calculations and groundwater numerical modeling of the mill tailings, which is positioned on large landslide. The large data set gives us opportunity to study several analytical tests for hydraulic test evaluation and to compare the results provided by them. In the present paper, methods of Jacob (Cooper & Jacob, 1967), Papadopulos et al. (1967), Theis (1935), Hvorslev (1951) and Cooper et al. (1967) are compared. Previous studies on the subject of comparative analysis of aquifer test methods have mostly discussed relations between different slug test methods, since they are the cheapest and therefore the most widely used methods for field evaluation of hydraulic conductivity. Herzog & Morse (1994) and Herzog (1994) compared methods of Hvorslev (1951), Cooper (1967) and Nguyen & Pinder (1984), and pointed out the importance of using several methods for calculating hydraulic conductivity, since no method can be applied at all times and employing of other methods is needed. Mace (1999) compared methods of Herbert & Kitching (1981), Barker & Herbert (1989), Hvorslev (1951), Bouwer & Rice (1976) and Cooper (1967) method for slug tests in large-diameter, hand-dug wells. He suggests that due to substantial well storage considerable pumping time may be required to lower the water level to the desired position. However, this shouldn't affect the calculation result provided that recovery time is considerably longer than pumping time. The relations between Hvorlev's (1951) and Cooper's (1967) model has been discussed by Chirlin (1989), who provides a rather assertive statement that for slug-tests the method of Cooper gives correct values of hydraulic conductivity, and Hvorslev's result deviate from the these real values due to neglecting aquifer storativity. In contrast, Chapuis et al. (1990), Chapuis (1998) and Chapuis & Chenaf (2002) provides several independent proofs (mathematical, physical, numerical and experimental), that the theory of Cooper et al. (1967) does not adequately represent slug-test conditions and thus can not give values of T and S. Karanth & P rakash (1988) studied the relations between slug tests and pump tests. They indicate that transmissivity values obtained by those two methods vary mostly within a factor of three, except for pump-test transmissivity values less than 1.16x 10"4 m2/s. Methods Construction of boreholes On the mill tailing site Boršt and its surroundings 21 new boreholes have been installed between May and October 2003. They were organized in two groups to determine the values of hydraulic conductivity of hydrometallurgical tailing as well as the bedrock that forms the base of tailings. The depths of piezometers ranged from 5 to 25 m and from 26 to 105 m, respectively. After the drilling, every well was rinsed with clean water and then activated with the airlift method. Activation took place until clean water flew out of the well, which was hard to achieve because in most cases dirtiness of water was a consequence of mud rinsing of hydrometallurgical tailings instead of drilling residue removal. No filter packs were installed in any of the piezometers. Shallow piezometers are equipped with cemented ф168 mm wide surface casing and have plastic inner casing PVC-U DN 100, slot 0.75 mm. Usually this casing goes up to 3 m in the base of the aquifer, where a plug and a sink are installed as well. Deep piezometers are installed in the same manner as the shallow ones with the exception of piezometers number 3, 4 and 5 (Table 1), which have one additional inner casing and piezometer number 12, which has two additional cemented inner casings due to its greater depth. Pumping well performance In all 20 piezometers, hydraulic tests were performed between 4th of September and 4th of November 2003 and were interrupted by heavy rain in the mean time. In all wells saturated thickness wasn't big enough to create a sufficient pressure head drop to develop test correctly. In addition, a large part of the material deposited in Boršt has relatively low permeability and consequently wells have very low yield (less than 0,1 l/s). Therefore, regular pumping tests weren't feasible in most of the wells, since the pump's lower limit of operation is approximately at discharges 1-2 dcl/s. Improvised slug tests were performed instead. Wells were pumped at a middle rate, which was defined on the basis of previous hydrogeological interpretation. So, basic condition for slug tests, water being removed from the well nearly instantaneously, was satisfied and drawdown response wasn't damaged. Such tests enabled the use of slug test methods as well as pumping test methods, since they satisfied all conditions for the two methods. Prior to testing pressure probes connected to automatic data loggers were inserted in the pumping well plus in other boreholes in its vicinity. Selection of probes with sensitivities that ranged from I to 3 bars depended on expected maximal drawdown (10-30 m). Time intervals between measurements were constant throughout the test but they varied between the tests from 1 to 24 seconds and were adjusted according to the duration of measurement and the expected rate of water level lowering. After a couple of tests it proved that there were no water level changes detected in the adjacent wells, therefore, we continued with testing with only one probe. Due to absence of water level change in the wells adjacent to the pumping well we had to limit our selection of analytical methods to those that describe single-well tests (i.e. tests that don't use any other piezometer for water level changes observation except the pumping well itself). Some of those methods were primarily developed for ordinary pumping tests with one or more piezometers (Jacob's (1967) and Theis's (1935) methods), but can also be used for single-well tests provided that certain additional assumptions and conditions are met. Analytical methods All methods discussed can not be applied to all performed tests in the same degree of reliance, since each method has different underlying assumption. Therefore, differences in basic conditions and assumptions that have to be satisfied for models to be successfully applied are presented in the following section as well as the governing equations. Models In this paper, methods of Jacob (Cooper & Jacob, 1967), Papadopulos (1967), Theis (1935), Hvorslev (1951) and Cooper (1967) are compared. These methods were selected while they are the most used in engineering practice. They assume boundary conditions that in most cases can reasonably be met in the field. Methods differ from one another in the given solution to the partial differential equation of groundwater flow as well as in their underlying assumptions. The later are incorporated, since field conditions in real world are way too complicated to be described with relatively simple analytical equations, therefore, certain assumptions and conditions have to be met to reduce the practical problem to mathematical constraints. Only some of them underlie all discussed methods, what leads to differences in results arising from the very fundaments of the application. Methods of Jacob, Papadopulos and Theis presume that (a) the well is pumped at a constant rate, whereas Hvorslev's and Cooper's methods suppose that (b) the water is added into or removed from the well instantaneously. In order to satisfy both conditions, improvised tests were performed where the wells were pumped at a constant rate (a) for a short time (b). If the pumping times are too long (longer than some critical time) empirical data doesn't coincide with the slug test model. According to Mace (1999), pumping times shorter than one day do not affect the performance of slug tests. However, in case of Boršt, these critical times were proved to be much shorter, about 30 -90 minutes. All five methods count for the storage in the well either by using only late time data in calculations or by incorporating this consideration directly in the type curves. On the other hand, Hvorslev method assumes incompressible water and soil, which in other words means that it does not count for storage in the aquifer as oppose to other methods discussed. As a consequence of this assumption flow toward the well is to be quasi-stationary by the Hvorslev theory, while in other methods flow is described as non-stationary. In addition, Hvorslev's model doesn't presume the penetration of entire aquifer and is in all less rigorous than other methods. Governing equations The general equation of groundwater flow d2s lds S ds —2+--=----(I) dr r 8r T dt was among the presented authors first solved by Theis (1935), who produced the following equation, derived from analogy between groundwater flow and conduction of heat: Q f e~ydy Q _ ч s = -— = ———W(u) (2) 471KOI у 4nKD (2) where и2 и3 u4 r2S W(u) = -0.5712-hiu + u--+---..., where u =- (3) 2.2! 3.3! 4.4! 4 KDt This equation was derived for fully penetrating pumping wells in homogeneous and isotropic non-leaky confined aquifers. An approximation u < 0.01 by Jacob is then applied to this equation, so that the terms beyond ln u in equation (3) become so small that they can be neglected. The equation for recovery data is then rewritten as: ^ 2,3 0ß К = ' ^ (4) and , 25 r2 t,t'>-^ (5) KD У where t is the time since pumping started and t' is the time since cessation of pumping, As' is change of head per one log cycle of time on semi-log plot s vs. t/t' (t/t' on the logarithmic scale). The most frequently used method for pumping tests, namely the Jacob (Cooper & Jacob, 1946) method is also based on the presented approximation of Theis's formula (2), except that it's applicable to drawdown data. The later can be simplified to give the following equation for calculating hydraulic conductivity from pumping test data: ^ 2,30 Q К = ' ^ (6) where As is change of head per one log cycle of time on semi-log plot s vs. t (t on the logarithmic scale). An additional assumption for single well test should be satisfied: 25 r2 t >-e- (7) KD v The most widely used method for slug tests has to be that of Hvorslev (1951), and is based on the following equation, derived for fully penetrating well in a confined aquifer, where a quasi-stationary flow and incompressible water and soil (i.e. zero aquifer storage) are assumed: v A ^ where A is the cross-sectional area of the well, l is the length of the tested portion, hj and hp are two values of water elevation in the well at the end and at the beginning of time interval At, and F is the shape factor that equals: F= Ш ln(-) (9) Hvorslev presented numerous such factors for different geometries. Among those, the shape factor for cased hole and uncased or perforated extension into aquifer of finite thickness was selected and used in our calculations. Papadopulos et al. (1967) suggested a solution developed directly for single well pumping tests. It is based on a large-diameter wells method, where well storage cannot be neglected. The governing equation for this method is: К = d0) where , fa? C(ß) Jn C(P) = 1-exp f v 4u [70(ßpM(ß)-r0(ßp)Ä(ß)] (12) Л(р) = р70(Р)-2аВД) (1Q) Д(Р) = Р/0(Р)-2о/1(Р) (1R) D(ß) = [_4(ß)]2 + [S(ß)]2 (1S) rlS riß , r К represents hydraulic conductivity, Q is pumping rate, D is aquifer thickness, sw is drawdown in the well, r is effective radius of screened part of the well, r is radius of the unscreened ' ew r 7 с part of the well, where the water is changing, t is time since pumping started, and F is well function, which is represented by the type curves and J0 (and Y0) and Yj are zero-order and first-order Bessel functions of the first and second kind, respectively. Cooper et al. (1967) presented also another overlapping graphs solution for slug tests. It is based on the following equation: f = (17) "o where riß _ KDt V a=-ssr P =-t- K=— (18) rc rc nrc V h0 is the initial water elevation in the well, ht is the water elevation at time t, b is the dimensionless time, V is the volume of water added to the well, and F(a,b) is the well function which is represented with a set of type curves (Figure 1, left) and is described by the equation. 8a г exp(-ßt/2 /a) F(a'P)=^| ц/е,а) du <19> where f(u,a) = [uJo(u) - 2aJj(u)]2 + [uYo(u) - 2aYj(u)]2 (20) and J0(u), Jj(u), Y0(u), Yj(u) - are zero-order and first-order Bessel functions of the first and second kind. Data processing In general, there are two types of numerical data processing, that both base on regression principle (methods will further be addressed only by the leading author). Curve fitting methods Papadopulos and Cooper try to find best possible fit between empirical data and type curves. They read four parameters (two in each graph) that enable the calculation of hydraulic conductivity of the aquifer (Figure I, left). Straight-line methods Jacob, Theis and Hvorslev derive the drawdown equation with certain operations and/or simplifications to the form that yields a straight-line graph (Figure I, right). The slope of the line and its intersection with the ordinate axis facilitate the computation of aquifer parameters. To define the subjective factor which is always present when applying different methods to aquifer test data, a short description of data segment selection criteria is given here. For drawdown data, the early time data doesn't coincide with the Jacob model, which is a consequence of the wellbore storage. Therefore, only late time data (later than approximately 2 - 5 minutes after the start of the pumping) was used for calculation. Similarly, first part of the recovery data was excluded from the calculation as well. Results obtained from this part of recovery data are erroneous because water rises faster than normally due to water coming back from the pipe to the well after the cessation of pumping. After this water returns, the recovery data should correspond to the Theis, Hvorslev and Cooper models. However, the subjective factor remains, since curve-matching as well as straight line matching processes were done solely by visually estimating the position of the best fit, and it was not quantified by the least-squares error function. Figure I. Data processing - left: Curve-fitting Cooper method; right: Straight-line Hvorslev method (Batu , 1998) Slika I. Obdelava podatkov - levo: Prilagajanje tipskim krivuljam pri Coopeijevi metodi; desno: Metoda Hvorsleva s premico (Batu ; 1998) Results and discussion Field results The calculation of hydraulic conductivity values from aquifer tests on Boršt site gave the following results as presented in Table I. As can bee seen from this table not all the methods could be applied to all tests. In cases number 15 and 17 to 20, slug tests with addition of water were performed, so the first three methods were inapplicable. Nevertheless, these cases were included in the analysis because of the study of the relationship between Hvorslev and Cooper method. In case number 14 the experimental data couldn't be fitted with sufficient certainty to type curves in Papadopulos method. Since no "real" values of hydraulic conductivity are available, we can only compare these results relatively towards one another and thus their rank numbers are given in the brackets for each test. Ranks of the hydraulic conductivity values for several cases show some typical arrangements (Figure 3). It is evident that in almost all cases the highest or at least second highest values are obtained by Papadopulos method. Moreover, cases number 1, 3, 4, 6, 8, 11 and 16 exhibit additional similarities. Results of Hvorslev and Cooper method are up to two and a half decades lower than results of other three methods. Among those two, Hvorslev's method results are usually somewhat higher than those of Cooper's Well No. Kjacob ^Papadopulos Kxheìs ^Hvorslev ^Cooper 1 6.38E-06 (1) 4.41E-06 (2) 1.84E-06 (3) 3.64E-07 (5) 4.30E-07 (4) 2 8.11E-07 (4) 1.75E-06 (2) 2.58E-06 (1) 7.06E-07 (5) 1.02E-06 (3) 3 6.46E-08 (4) 2.70E-07 (1) 2.15E-07 (2) 1.25E-07 (3) 2.58E-08 (5) 4 5.65E-07 (2) 8.11E-07 (1) 5.35E-07 (3) 1.40E-08 (4) 1.03E-09 (5) 5 6.62E-08 (5) 1.85E-07 (1) 8.28E-08 (3) 7.77E-08 (4) 1.63E-07 (2) 6 2.25E-06 (2) 3.66E-06 (1) 5.03E-07 (3) 6.51E-08 (5) 7.58E-08 (4) 7 4.82E-08 (3) 2.04E-08 (5) 4.42E-08 (4) 9.75E-08 (2) 1.42E-07 (1) 8 3.62E-07 (2) 3.72E-07 (1) 1.72E-07 (3) 1.31E-07 (4) 4.37E-08 (5) 9 1.43E-07 (5) 3.74E-07 (2) 1.76E-07 (4) 2.21E-07 (3) 5.09E-07 (1) 10 4.06E-07 (4) 5.55E-07 (2) 3.40E-07 (5) 4.12E-07 (3) 7.27E-07 (1) 11 7.63E-08 (2) 9.11E-08 (1) 2.79E-08 (3) 2.75E-09 (4) 1.23E-09 (5) 12 8.32E-07 (3) 2.90E-06 (1) 6.59E-07 (4) 3.69E-07 (5) 1.20E-06 (2) 13 8.42E-06 (4) 3.01E-06 (5) 5.46E-05 (2) 1.18E-04 (1) 5.10E-05 (3) 14 4.67E-05 (1) 1.30E-05 (4) 2.27E-05 (3) 4.50E-05 (2) 15 5.85E-07 (2) 7.41E-07 (1) 16 1.22E-05 (3) 5.14E-05 (1) 4.91E-05 (2) 4.02E-06 (4) 1.52E-06 (5) 17 4.43E-05 (2) 7.44E-05 (1) 18 6.72E-05 (2) 1.45E-04 (1) 19 1.49E-05 (2) 7.09E-05 (1) 20 1.23E-07 (2) 1.71E-07 (1) Table I. Hydraulic conductivity values and their ranks for 20 aquifer tests in Boršt, Žirovski vrh (I - 12: deep piezometers; 13 - 20: shallow piezometers) Tabela I. Vrednosti koeficientov prepustnosti in njihovi rangi za 20 hidravličnih testov na Borštu (1 - 12: globoki piezometri; 13 - 20: plitvi piezometri) method. Results of Theis and Jacob methods are in between Hvorslev and Cooper method results of one side and Papadopulos method values on the other. In addition, Jacob method gives a bit higher values than Theis model. Differences between tests can also be noticed on the scatter point graphs, where pumping test method is plotted against slug test method. They are most evident in the relationship between Jacob and Hvorslev method, which are believed to be the most resilient and straight forward solutions among pumping and slug test methods, respectively. Most of the cases coincide with the equal values line, except in some cases, which fall on a line parallel to the equal values line and where results of Jacob are greater than values of Hvorslev. Differences in results also occur between method that use drawdown data (Jacob, Papadopulos) and those that use recovery data (Theis, Hvorslev, Cooper). Nevertheless, they can only be observed where pumping times were short, otherwise they are blurred by the differences between the pumping and slug test. Overall, the recovery results show greater variability in values than those calculated from drawdown data. 5E-04 5E-05 5E-06 5F 1, 5E-07 * 5E-08 5E-09 5E-10 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 Piezometer Figure 2. Ranges of results obtained by methods of Jacob et al. (1946), Papadopulos et al. (1967), Theis (1935), Hvorslev (1951) and Cooper et al. (1967) Slika 2. Razponi dobljenih vrednosti po metodah Jacoba et al. (1946), Papadopulosa et al. (1967), Theisa (1935), Hvorsleva (1951) in Coopera et al. (1967) Ranges of hydraulic conductivity values Д « « А о О Д О о А Д u о 0 □ 1 □ я □ о ff □ ■fr 8 □ f А # 5 д о i б s д О о д □ д 8 8 0 A д * д 0 □ О О Kjacob □ К "^Papadopulos О KTheis д к ,xHvorslev * к Cooper 1 1 1 — « А « Figure 3. Relation between hydraulic conductivity values of Jacob and Hvorslev method Slika 3. Razmerje med koeficienti prepustnosti po Jacobovi metodi in metodi Hvorsleva In Table 2 correlations between results from different analitical models are presented. The most profound relation among the pairs of five discussed methods in case of Boršt is that of the two slug test methods - Hvorslev and Cooper. Somewhat lower are the correlations between the three pumping test methods. In general, pumping test methods do not correlate with slug test methods at such high degree. Discussion Results obtained by using different methods for calculation of hydraulic conductivity from aquifer tests showed a clear separation between tests with longer pumping times and those with shorter pumping timer. A term of "critical time" was introduced to distinguish between the two types of tests which describes the longest time of pumping at which slug tests methods can be employed and the results of these methods don't diverge significantly from those of pumping test methods. Critical time was only determined approximately. Mace (1999) showed that the instant addition/removal of water condition is quite flexible and can be extended to one whole day of pumping. However, this doesn't prove to be the case in aquifer tests on mill Table 2. Correlations between results from different analitical models Tabela 2. Korelacije med rezultati različnih metod R Kjacob Kpapadopulos K-Theis KjHvorslev K-Cooper Kjacob l Kpapadopulos 0.92 1 Kxheis 0.90 0.86 1 ^Hvorslev 0.66 0.57 0.84 l K<üooper 0.54 0.47* 0.68 0.93 l * statistically insignificant tailing of Boršt. Differences in results start to appear in tests with pumping times longer than half an hour, and the longer the pumping time, the greater the distinction between pumping and slug test method results. To be precise, in such cases slug test methods underestimate hydraulic conductivity values by as much as two orders of magnitude. Differences between "pumping" and "slug" tests can be seen in typical arrangements of ranks of the hydraulic conductivity values as described in previous section. Such typical arrangement is associated largely to those cases that resemble pumping tests more than slug tests, namely cases number I, 3, 4, 6, 8, 11 and 16. Moreover, cases of pumping times longer than "critical time" can be further distinguished from other cases as can be seen in Figure 5. "Pumping" test cases (number 1, 4, 6, 11) are the ones that diverge from the line of equal hydraulic conductivity towards the pumping test method axis, whereas "slug" test cases coincide well with this line. This shows that the results of these five methods can only be compared when a slug test was performed, and that the instant addition/removal of water condition is the most rigorous and thus the key condition to be satisfied in this matter. Furthermore, systematic differences seem to appear between Theis's recovery method for pumping tests and Hvorslev's method for slug-tests. Since both are straight-line methods a clear separation between straight and non-straight line portion of the recovery data can be defined (i.e. the point where experimental data start to concur with the model) in data processing procedure. It is apparent, that in cases when our improvised aquifer test resembles slug test in higher degree than pumping test, the experimental data coincides with Hvorslev's model earlier than with Theis's model. On the other hand, field data falls in Theis's model earlier, in cases with longer pumping times when performed test resembles pumping test. As a remark, Karanth & Prakash (1988) showed that with decreasing values of hydraulic conductivity the transmissivity values of slug tests exceed values of pumping test. Since similar trends, but less extent, have been noted in this study using different methods on the same set of aquifer test data, we can conclude that a part of the slug-pumping test relationship arises solely from model structure. Discrepancies between method that use drawdown data (Jacob, Papadopulos) and those that use recovery data (Theis, Hvorslev, Cooper) are a consequence of a couple technical facts. Firstly, not fully developed wells where activation of the well took place during the pumping resulted in lower values from drawdown data, and secondly, presence of well clogging with the fine-grained hydrometallurgical tailing which resulted in lower values from recovery data. Although these differences are small compared to the pumping/slug test differences, they add to the importance of model selection. The greater variability in results from recovery data also implies that well development is in fact an important factor. Somewhat higher results can also be obtained by Papadopulos method as can be seen in Table 1. This is due to many cases when experimental values coincide with the straight portion of the type curves, where ambiguous results are usually gathered. This portion of type curves represent the water being pumped from wellbore storage and is employed in cases of small well diameter and/or short pumping times as it was the case in Boršt. Nothing similar can be observed with the other curve-fitting method - Jacob method doesn't account for the wellbore storage and thus doesn't have such straight portions of type curves and the experimental and theoretic graphs can only be adjusted toward one another on x-axis. Since results of Hvorslev's method are as rule lower than those of any other method a reason for this was sought as well. Lower values could be a consequence of the difference between pumping and slug test conditions with respect to Jacob's, Papadopulos's and Theis's method on one hand, and the fact that it does not account for the aquifer storage with respect to Cooper's model on the other. Namely, after the depression due to pumping is formed, the Hvorslev model assumes that the water must fill the gaps between the soil particles in the drained portion of the aquifer, whereas storage accounting Cooper model requires additional water to compensate for the aquifer storage. Consequently, water in Cooper's model should flow to the well slower than in Hvorslev's model, resulting in higher values of hydraulic conductivity using the Cooper's model given that the actual flow to the well used for calculation is in fact unique and therefore the same for both cases. In addition, the difference between the Cooper and Hvorslev results in case of Boršt varies with the order of magnitude of the hydraulic conductivity. Since those methods could be applied to most performed tests and thus produced more results, we get a more thorough insight in their correlation. Figure 6 shows that in more permeable materials, results of Cooper exceed those of Hvorslev, 1.E-04 1.E-05 1.E-0S 1 E-07 1 E-06 1 E-03 1.E-09 j у = 6.475x11M I R2 = 0.922 -ye. /S » • k. У S /V 7 * 1 E-08 1 E-07 1 E-06 1 E-OS 1£- - ■ RIV-3 - - A- - - RV-3 - - - A- - - RVI-3 ——precipitation Figure 7. Time-trend plot of electroconductivity values for water sampled on the right side of the lysimeter beneath the industrial railway tracks Slika 7. Časovno nihanje električne prevodnosti v vodi, vzorčeni na desni strani lizimetra alluvial gravel aquifer. However, the role of vertical flow is quite the opposite, because it is the main factor controlling contaminant transport towards the aquifer saturated zone. Hence, investigation of the occurrence and frequency of rapid recharge events represents one of the main themes of the next research phase. With this regard, the monitoring of chlorides, of heavy metals and of herbicides has been established at the beginning of2005 and the first tracing test was undertaken at the end of March 2005. о-10 RIV-2 RIII-2 * precipitation i • RVI- ^ Д, Figure 8. Box plots of SI80 values in sampled water Slika 8. Skatlasti diagram SI80 vzorčene vode Jul-03 Aug-03 Sep-03 0ct-03 Nov-03 Dec-03 Jan-04 Feb-04 Mar-04 Apr-04 May-04 Jun-04 Jul-04 -precipitation —•—RI-2 —e—RII-2 ------- Ri-3 Figure 9. Time-trend plot of SI80 values in water sampled from the lysimeter upper levels Slika 9. Časovno nihanje SI80 v vodi, vzorčeni v zgornjih nivojih lizimetra -16 -1-1-1-1-1-1-1-1-1-1-1- Jul-03 Aug-03 Sep-03 0ct-03 Nov-03 Dec-03 Jan-04 Feb-04 Mar-04 Apr-04 May-04 Jun-04 Jul-04 precipitation — ■ ■ ■ — RIII-1 - ■ й ■ - RV-1 —■—RIII-2 —в—RIV-2 —A—RVI-2 --■■--■ RIII-3 - --и-- ■ RIV-3 - - ■ A- ■ ■ RV-3 - - A- ■ ■ RVI-3 —ж— LIII-6 Figure IO. Time-trend plot of 8I80 values in water sampled from the lysimeter lower levels Slika IO. Časovno nihanje SI80 v vodi, vzorčeni v spodnjih nivojih lizimetra Povzetek Proučevanje tokovnega sistema in prenosa snovi v urbanem lizimetru Pivovarne Union, Ljubljana, Slovenija V urbanem lizimetru Pivovarne Union (si. I in 2) potekajo raziskave toka in prenosa snovi v nezasićeni coni pleistocenskega prodnega vodonosnika (sl. 3), ki je vse pomembnejši vir pitne vode, ne le za Pivovarno Union, ampak tudi za mesto Ljubljana. Glavni cilj raziskav je študij vplivov industrije in prometa na omenjen vodonosnik, ki omogoča, da se prouči možnost onesnaženja virov podzemne vode na območju Pivovarne Union ter oceni vlogo nezasičene cone pri njihovi zaščiti. V prvi raziskovalni fazi so se izvajale tedenske meritve vodne bilance in osnovnih fizikalno-kemičnih parametrov vode, poleg tega pa je potekalo tudi mesečno vzorčenje vode za analizo izotopske sestave kisika (8180). V lizimetru se je vzorčila voda s pomočjo keramičnih svečk. 18 svečk je vgrajenih na koncu vrtin, na desni strani lizimetra (RI-1 do RI-6, RII-1 do RII-6 itd.), ki leži pod industrijskimi železniškimi tiri, 3 pa so vgrajene v vrtine na levi strani lizimetra (LI-4, LII-5 in LIII-6), ki leži pod asfaltnim območjem (sl. 2 in 3). Geološki prerez na sliki 3 kaže, da vrtine predirajo štiri različne plasti, vzorčna mesta pa so razporejena v 3 kolone in 6 nivojev na globinah 0,3-4 m. Za prvo raziskovalno leto je prikazana vodna bilanca vzorčnih mest lizimetra v tabeli 1 ter na slikah 4 in 5. Pretoki vzorčnih mest so močno odvisni od količine in intenzivnosti padavin (sl. 4 in 5), iz tabele pa je mogoče razbrati, da največja količina vode priteka v vzorčna mesta na nivoju III. Predvideva se, da je to posledica razvoja lateralne komponente toka v bližini kontakta med dvema plastema z različno strukturo ter, posledično, različno hidravlično prevodnostjo (sl. 3). Na sliki 5 je mogoče opaziti tudi pojavljanje vertikalnega toka iz nivoja III v nižja območja - povečana količina vode v vzorčnih mestih nižjih nivojev, zlasti na nivoju IV (oktober 2003, april in junij 2004). Lastnosti električne prevodnosti vzorčenih vod so prikazene na slikah 6 in 7. V lizimetru so najnižje vrednosti parametra vezane na nivoja I in II, medtem ko so najvišje vrednosti vezane na nivo III in ne na nižje nivoje, kar odseva pomembno vlogo lateralne komponente toka v bližini nivoja III. Po drugi strani pa slika 7, ki prikazuje časovno nihanje električne prevodnosti v vzorčenih vodah, kaže, kdaj in kje je bilo izrazito vertikalno napajanja spodnjih nivojih lizimetra (oktobra 2003 in aprila 2004). Lastnosti 5180 so ilustrirane na slikah 8, 9 in 10. Glede na padavine imajo vode zgornjih nivojev lizimetra (I, II in III) največje razpone vrednosti, kar odseva intenzivno dinamiko in s tem kratek zadrževalni čas. Nihanje parametra je veliko bolj dušeno v spodnjem delu lizimetra ( IV, V in VI), kar odseva manj intenzivno dinamiko in daljši zadrževalni čas. Na slikah 9 in 10 je treba pozornost nameniti odstopanjem od običajnih trendov. Le-ta opozarjajo na References Berg, W., Čenčur Сток, В., Frank, J., Feichtinger, F., Nutzmann, G., Papesch, W., Rajner, V., Rank, D., Schneider, S., Seiler, K.P., Steiner, КН., Stenitzer, E., Stichler, W., Trček, В., Vargay, Z., Veselič, M., Zojer, Н. (2001):Trac-ers in the unsaturated zone. Steir. Beitr. Hydrogeol.: 52, pp. 1-102. vertikalni tok in prenos snovi v vodonosniku med glavnimi hidrološki dogodki - oktobra 2003 in aprila 2004. Aprila 2004 so npr. padavine izpodrinile v spodnji del lizimetra vodo, ki je bila izotopsko osiromašena (sl. 10), kar je mogoče pripisati topljenju snega. Primerjava slik 9 in 10 kaže, da je vpliv topljenja snega opazen v zgornjem delu lizimetra že mesec prej. Rezultati prve faze raziskav v urbanem lizimetru Pivovarne Union so opisali osnovne lastnosti toka in prenosa snovi v opazovanem okolju. Opozorili so na hierarhijo toka v nezasičeni coni vodonosnika in odziv okolja nanjo. Identificirani sta bili dve pomembni vrsti tokov - lateralni in hitri vertikalni tok. Lateralni tok ima pomebno vlogo pri zaščiti podzemnih vodnih virov pleistocenskega prodnega vodonosnika. Vloga vertikalnega toka je povsem nasprotna, saj je le-ta glavni faktor za prenos in širjenje onesnaženja proti zasičeni coni vodonosnika. Glede na to, je glavna tema druge raziskovalne faze proučevanje hitrega vertikalnega napajanja. V ta namen se je vzpostavil na začetku leta 2005 monitoring kloridov, težkih kovin in herbicidov, konec marca 2005 pa se je izvedel prvi sledilni poskus. Burgman, J.O., Calles, В., Westman, F. (1987): Conclusions from a ten year study of oxygen-18 in precipitation and runoff in Sweden. In: Publication of IAEA Symposium 299 on Isotope techniques in Water Resources Development. IAEA, Vienna, pp. 579-590. Clark, I.D. and Fritz, P. (1997): Environmental Isotopes in Hydrogeology. Lewis Publishershers, New York, 311 p. Fank, J., Ramspacher P., Zojer, H. (2001): Art der Sickerwassergewinnung und Ergebnisinterpretation. In: Bericht der BAL über die Lysimetertagung, Gumpenstein, 16-17 April 1991. BAL, Gumpenstein, pp. SS-62. Helsel, D.R. & Hirsch, R.M. (1992): Statistical Methods in Water Resources, Elsevier, Amsterdam, S22 p. Juren, A., Pregl, M., Veselič, M. (2003): Project of an urban lysimeter at the Union brewery, Ljubljana, Slovenia. RMZ-Materials and Geoenvironment. S0/3, pp. 1S3-1S6. Kendall, С. & Mcdonell, J.J. (1998): Isotope tracers in catchment hydrology, Elsevier, Amsterdam, 722 p. Klotz, D., Seiler, K.P., Scheunert, I., Schroll, R. (1999): Die Lysimeteranlagen des GSF-Forschungscentrums fuer Umwelt und Gesundheit. In: Bericht uber die 8. Lysimetertagung. BAL, Gumpenstein, pp. 1S7-160. Nimmo, J.R. (2002): What Measurable Properties Can Predict Preferential Flow? In: Proc. of 17th World Congress of Soil Science, Bangkok, 14-21 Augut 2002. International Union of Soil Science, Bankog. Pezdič, J. (1999): Isotopes and geochemical processes, University book, Paculty of natural sciences, Department of geology, Ljubljana, , 269 p. Raimer, D., Berthier, E., Andmeu, H. (2003): Development and results of an urban lysimeter. Geophysical Research Abstracts. S, pp. 37-92. Seiler, K.P., Loewenstern, S., Schneider, S. (2000): The role of by-pass and matrix flow in the unsaturated zone for the groundwater protection. In: Sililo et al. (eds.), Groundwater: Past Achivements and Future Challeng. Balkema, Rotterdam, pp. 43-76. Trček, B. (2003): Epikarst zone and the karst aquifer behaviour, a case study of the Hubelj catchment, Slovenia, Geološki zavod Slovenije, Ljubljana, 100 p. Vpliv mineralne sestave in strukture na obstojnost apnencev kot naravnega kamna The weathering durability of limestones as a function of their mineral composition and texture Simona Jarc, Breda Mirtič Oddelek za geologijo, NTF, Univerza v Ljubljani, Aškerčeva 12, Ljubljana, Slovenija; E-mail: simona.jarc@ntfgeo.uni-lj.si Received: October 17, 2004 Accepted: November 24, 200S Izvleček: Določali smo vpliv mineralne sestave nekaterih vrst slovenskih apnencev na njihovo uporabno vrednost. Vzorci so bili izbrani iz nekaterih aktivnih in občasno delujočih kamnolomov apnenca: iz Hotavelj, Lesnega Brda, Drenovega Griča in Lipice. Odvzeti so bili vzorci sveže kamnine in že prepereli vzorci iz opuščenih delov kamnoloma, ki so bili okoli trideset let izpostavljeni vremenskim pogojem in delovanju različnih organizmov. Mineralna sestava je bila določena s pomočjo rentgenske difrakcije in vrstičnega elektronskega mikroskopa. S primerjavo površin svežih in že nekoliko preperelih vzorcev smo opazovali učinke preperevanja; predvsem gre za raztapljanje kamnine, vpliv insolacije in delovanja organizmov. Pogosto se različni načini preperevanja prepletajo in so medsebojno odvisni, tako da njihovih učinkov ne moremo razlikovati. Obstojnost kamnine je odvisna predvsem od strukture apnenca, ki jo pogojujeta način nastanka in vrsta diagenetskih procesov, in ne toliko od same mineralne sestave. Abstract: The influence of the mineral composition on weathering durability of some Slovenian building limestones has been investigated. Samples were taken from some active or occasionally active quarries: Hotavlje, Lesno Brdo, Drenov Grič and Lipica. The comparison between fresh limestone samples and about 30 years weathered samples was made. The mineral composition was determined by X-ray diffraction and scanning electron microscope. The combine effects of weathering, such as dissolution, insolation and biological activity, have been documented. The weathering durability is above all the result of limestone texture, therefore the origin and diagenetic processes play more important role than mineral composition itself. Ključne besede: apnenec, mineralna sestava, vrstični elektronski mikroskop (SEM), preperevanje Key words: limestone, mineral composition, scanning electron microscope (SEM), weathering Uvod Kalcitne in dolomitne kamnine predstavljajo dobro desetino vseh sedimentnh kamnin (Morse & Mackenzie, 1990), v Sloveniji pa zavzemajo karbonatne kamnine skoraj polovico ozemlja, samo apnenci okoli 35 % (Gams, 1974). Z apnenci se srečujemo praktično na vsakem koraku. Opazujemo jih v njihovem naravnem okolju ali pa kot gradbene in okrasne elemente, tu pa se nujno pojavi vprašanje njihove obstojnosti. Glavni dejavnik obstojnosti kamnine je njena sestava. Apnenci imajo enostavno kemično in mineralno sestavo, kljub temu pa obstajajo med njimi velike razlike v obstojnosti in uporabni vrednosti. Izbrani slovenski apnenci so razmeroma čisti; vsebujejo preko 91 mas.% CaC03. Z elektronskim mikroskopom opazujemo obliko mineralov, strukturo, vezivo v kamninah veliko bolj natančno kot z optičnim mikroskopom. Omenjene lastnosti močno vplivajo na obstojnost kamnine. Elektronski mikroskop se zelo veliko uporablja prav za opazovanje preperelih površin kamnin (Viles & Moses, 1998). Preiskovane vrste apnencev Apnenec iz Hotavelj je cordevolske starosti, izrazito pisan in mineralno nehomogen. Kamnino po Folku imenujemo intrabio-mikrosparit (Bilbija & Grimšičar, 1987, Ramovš, 1987). Obarvanost hotaveljskega apnenca je posledica določenih primesi, to pomeni, da se kemična in mineralna sestava nekoliko spreminjata, zato smo poskušali dobiti čimbolj enakomerno obarvane vzorce apnenca. Ločili smo tri barvne različke: sivega (v nadaljevanju ga označujemo kot HS, rožnatega (HR) in rdečega (HRd). Zaradi pogostih leč ali plasti nekarbonatnega materiala v hotaveljskem apnencu smo le-te poskušali analizirati kot samostojni vzorec (HV). Pisani apnenec z Lesnega Brda je po nastanku, starosti, sestavi in lastnostih skoraj enak hotaveljskemu (Mirtič et al, 1999). Tudi lesnobrdski apnenec je zelo pisan, od sive do rožnate barve, sivega različka je relativno manj, zato smo v preiskavah uporabili samo rožnati različek (LBRd). Črni karnijski apnenec z Drenovega Griča (vzorec DGC) je nekoliko mlajši od pisanega cordevolskega apnenca z Lesnega Brda. V Lipici ločimo glede na velikost skeletnega drobirja dve vrsti apnenca, enotni in rožasti. V obeh primerih gre za rekristalizirana biomikritna apnenca zgornjekredne starosti, ki se razlikujeta v teksturi. To razlikovanje, ki ga uporabljajo v kamnarski industriji, smo ohranili: vzorec enotnega apnenca označujemo kot LiU, vzorec rožnatega apnenca kot LiF (Jarc, 1996, Jarc, 2000). Metode dela Kemično analizo preiskovanih apnencev so opravili v laboratoriju ACME v Kanadi. Določili so količino Si02, Alp03, Fe203, MgO, CaO, Na20, Kp0, Ti02, P205, MnO, Crp03, Sr, žarilno izgubo (LoI - loss of ignition) z metodo induktivno vezane plazme emisijske spektrometrije (ICP-ES), celotni ogljik in žveplo pa z Leco (ACME, 1999). Za silikatno analizo so 0,2 g vzorca talili z 1,2 g LiB02 in raztopili v 100 ml 5 % HN03. Žarilno izgubo (LOI) so določili glede na spremembo mase vzorca po 1 uri žganja pri temperaturi 1000 °C. Rezultati kemične analize so v tabeli 1. Difraktogrami vseh vzorcev so bili posneti na rentgenskem difraktometru Philips, na Oddelku za geologijo. Pogoji snemanja so bili naslednji: sevanje CuKa, Ni filter, napetost 40 kV, tok 20 mA, hitrost snemanja 2,5°/minuto, območje snemanja 2q od 2° do 70°. Z metodo rentgenske difrakcije so bili v vzorcih določeni sledeči minerali: Tabela I. Kemična sestava vzorcev. Količine oksidov, žarilne izgube (LOI), ogljika in žvepla so v masnih odstotkih, stroncija v ppm. Table I. Chemical composition of investigated samples. All oxides, carbon, sulfur and loss of ignition content are in %, strontium in ppm. Si02 A1203 FeA MgO CaO Na20 КгО Ti02 P2Os MnO Cr203 Sr LOI C/TOT S/TOT HSI 0,21 0,17 0,05 0,91 54,91 0,03 <0,04 0,01 0,04 0,02 0,004 101 43,6 11,8 0,01 HSlp 0,38 0,15 0,05 0,94 54,56 <0,01 <0,04 <0,01 0,02 0,02 0,004 100 43,7 11,4 0,01 HS2 0,31 0,17 0,07 0,69 55,42 0,02 0,04 0,01 0,03 0,02 0,003 98 43,2 11,7 0,01 HRI 1,12 0,48 0,21 0,67 53,85 0,04 0,12 0,01 0,05 0,03 0,002 96 43,4 11,6 0,01 HR2 1,33 0,64 0,24 0,62 54,03 0,02 0,19 0,02 0,04 0,03 0,001 95 43 11,4 0,01 HRdI 0,58 0,31 0,59 0,52 54,24 <0,01 0,06 0,01 0,04 0,05 0,006 99 43,6 11,3 0,01 HRd2 0,84 0,42 0,75 0,45 54,08 0,04 0,12 0,01 0,03 0,05 0,006 94 43,1 11,5 0,01 HRd2p 0,95 0,42 0,72 0,46 54,16 <0,01 0,12 <0,01 0,02 0,05 0,003 95 43,1 11,5 0,01 HRd3 0,83 0,42 0,79 0,47 54,49 0,01 0,1 0,01 0,01 0,05 0,001 96 42,9 11,8 0,01 HVI 17,29 8,36 5,91 5,09 29,75 0,03 3,07 0,18 0,05 0,08 0,01 59 30,2 7,61 0,01 HV2 18,44 8,77 6,09 4,78 28,96 0,01 3,26 0,19 0,07 0,08 0,01 55 29,4 7,07 0,03 LBRdl 1,26 0,77 0,29 0,42 54,00 0,03 0,12 0,04 0,08 0,01 0,003 133 42,9 11,2 0,01 LBRd2 1,66 0,99 0,34 0,44 53,82 0,02 0,18 0,04 0,01 0,01 0,005 110 42,5 11,4 0,01 DGCl 1,47 0,77 0,51 0,83 51,75 0,03 0,12 0,03 0,02 0,01 0,004 972 44,4 13,1 0,23 DGC2 1,73 0,88 0,82 0,77 51,00 0,05 0,11 0,04 0,01 0,01 0,002 1055 44,5 14,2 0,37 LiUl <0,02 0,12 0,07 0,35 55,63 0,02 <0,04 0,01 0,01 <0,01 0,001 189 43,7 11,7 0,01 LiU2 0,02 0,1 0,1 0,3 55,96 0,01 0,04 0,02 0,03 <0,01 0,01 188 43,4 11,7 0,03 LiFl <0,02 0,09 0,07 0,31 55,76 0,03 <0,04 0,01 0,03 <0,01 0,004 261 43,6 11,6 0,01 LiF2 <0,02 0,08 0,09 0,28 56,03 0,02 <0,04 <0,01 0,02 <0,01 0,006 220 43,4 11,7 0,04 • Hotavlje sivi (HSI, HS2): kalcit, dolomit, vermikulit, muskovit. • Hotavlje rožnati (HRI, HR2): kalcit, muskovit. • Hotavlje rdeči (HRdI, HRd2, HRd3): kalcit, muskovit, pirit, vermikulit. • Hotavlje vložki (HVl, HV2): kalcit, klorit, hematit, dolomit, muskovit. • Lesno Brdo (LBRdl, LBRd2): kalcit, klorit, muskovit. • Drenov Grič (DGCl, DGC2): kalcit, klorit, kremen. • Lipica enotni (LiUl, LiU2): kalcit, klorit. • Lipica rožasti (LiFl, LiF2): kalcit, klorit. Pripravo vzorcev in samo mikroskopiranje na vrstičnem elektronskem mikroskopu (SEM) smo opravili na Inštitutu Jožef Stefan. Vzorce smo spolirali z diamantno pasto granulacije 0,25 mm in jih naparili z grafitom. Na posameznih mestih smo s pomočjo odbitih elektronov skušali ugotoviti elementno sestavo. Površine svežih in preperelih vzorcev smo opazovali s pomočjo sekundarnih elektronov. Prepereli vzorci so bili odvzeti na površini opuščenih delov kamnolomov, kjer so bili okoli 30 let izpostavljeni naravnim procesom preperevanja. Vzorcev nismo spolirali, predhodno smo jih samo ultrazvočno očistili, zvakumirali in naparili z grafitom in zlatom. Rezultati in razprava V apnencu iz Hotavelj prevladujejo euhedralna romboedrska zrna kalcita z gladkimi ploskvami, velikosti okoli 10 mm. Na posameznih mestih jih sekajo večje žile in leče drobnozrnatega kalcita, glinenih mineralov, glinencev (ortoklaz in albit) ter pirita. Te minerale smo potrdili tudi z elementno analizo. Sivi različek hotaveljskega apnenca ima največ MgO, ki pa kljub vsemu ne presega I mas.%, kar ustreza približno 0,6 mas.% magnezija (tabela I). Z rentgensko difrakcijo smo zasledili tudi dolomit. Velikost kalcitnih sparitnih kristalov je v povprečju okoli 20 mm. Gre za zelo čist CaCO3, ki vsebuje le okoli 0,36 mas.% magnezija. Količina magnezija je tako majhna, da ne moremo ugotoviti, ali je enakomerno razporejen v kalcitni rešetki ali je koncentriran na posamezna področja. Na redkih mestih opazimo več kot 100 mm velika sparitna zrna, ki niso homogena, vsebujejo nekoliko več magnezija (okoli 0,73 mas.%). Razporeditev z magnezijem bogatejših delov je popolnoma naključna čez celoten presek preiskovanega vzorca in jo lahko opazujemo le, če je sparitni kristal dovolj velik. Pri kristalih manjših dimenzij pa dobimo le povprečno količino magnezija (0,36 mas.%). Apnenec je malo porozen, gre za zaprto poroznost, pore pa dobimo na mejah posameznih sparitnih polj. Vzorec je zelo nehomogen. Takoj, ko je poroznost nekoliko večja, so v apnencu tudi glineni minerali. Z uporabljenimi metodami ne moremo ugotoviti načina nastanka glinenih mineralov: lahko predstavljajo netopne ostanke, nastale pri raztapljanju apnenca zaradi delovanja pornih in meteornih vod, ali so sekundarnega nastanka. Oblika por pa kaže, da litifikacija kamnine še ni popolna. Elementarna analiza glinenih mineralov pokaže natrij, magnezij, aluminij in silicij ter kisik (glinene minerale vermikulitove skupine je pokazala tudi rentgenska difrakcija); natančnejša kvalitativna analiza glinenega minerala je nemogoča, ker so kristali premajhni in okolica moti. Analiza površine rdečega hotaveljskega različka pokaže, da je količina magnezija v celotnem vzorcu približno enaka (približno 0,27 mas.% Mg). Apnenec je, kljub majhni količini železa (0,33 mas.%), intenzivno obarvan. Rdeči različek hotaveljskega apnenca ima med vsemi vzorci največ FepO3 (tabela 1). Z rentgensko analizo smo ugotovili prisotnost pirita, vendar ga pod elektronskim mikroskopom nismo zasledili. Prav tako nismo opazili glinenih mineralov. Poleg karbonatnih mineralov pa so v vzorcu tudi nepravilno oblikovana, približno 70 mm velika kremenova zrna. Glede na rezultate rentgenske difrakcije ima rožnati različek hotaveljskega apnenca najmanj pestro sestavo. Sparitni deli se menjavajo z bolj poroznimi in z magnezijem nekoliko bogatejšimi mikritnimi deli. Mikritni deli vzorca imajo tudi več drobnih glinenih mineralov (sestave magnezij, kalij, aluminij, silicij in kisik) in podolgovate kristale muskovita. Poleg omenjenih elementov je v vzorcu tudi železo, ki povzroča značilno obarvanost tega hotaveljskega različka. Sparitni deli pripadajo čistemu CaCO3, magnezija ne zaznamo. Poleg kalcita in muskovita pa smo pod elektronskim mikroskopom opazili majhna, nekaj 10 mm velika, nepravilna kremenova zrna in večja, do 150 mm velika zrna kalijevega glinenca. Opazovanje vzorca HV pod optičnim mikroskopom pokaže večja sparitna polja (800 mm) relativno čistega kalcita, ki so med seboj ločena z velikimi razpokami. Te so zapolnjene z glinenimi minerali, sekun- darnim karbonatom in železovimi minerali. V sparita dobimo nepravilno oblikovana zrna kremena (80 mm), podolgovata zrna muskovita in pirit. V vezivu prevladujejo glineni minerali, muskovit, veliko je tudi železa. Ce elementarno analizo primerjamo z rezultati rentgenske difrakcije, je železo vezano na hematit, ki to vezivo značilno obarva. Poleg kalcitnih kristalov so v vzorcu tudi lepi kristali 900 mm velikega conarnega dolomita. Verjetno gre za poznodiagenetski dolomit. Conarnost ni rezultat različne sestave; razmerje magnezija in kalcija je povsod približno enako, 1:3. Razlika je torej v optični orientaciji dolomita. Sklepamo, da gre za sintaksialni dolomitni cement, ki je precipitiral v okolni prazni prostor, ki je nastal s predhodnim raztapljanjem CaCOQ (Krinsley et al., 1998). Opazovanje obrusov pokaže, da barvnih različkov hotaveljskega apnenca med seboj ne moremo ločiti. Izstopa le različek HV, kar pa je razumljivo, saj je tudi kemično in mineraloško popolnoma drugačen. Površina preperelega apnenca ima popolnoma drugačen videz (slika 1). Že pri majhni povečavi lahko opazimo pore, ki jih pri svežem vzorcu apnenca še nismo opazili. Nekatere pore imajo obliko kalcitnih zrn. V takšnih primerih so nastale pore lahko rezultat fizikalnega preperevanja: temperaturne razlike in zmrzali so razrahljale vezi med kristali, zrno pa je izpadlo. Močno je tudi biogeno preperevanje. Izrazito okrogle pore so nastale z vrtanjem organizmov v substrat. Gre torej za kombinacijo kemičnega, fizikalnega in biogenega preperevanja (Summerfield, 1994). Naravne vode imajo navadno rahlo kisel do rahlo alkalen pH (Ford & Williams, 1994, Drever, 1997). V tem območju je raztapljanje omejeno predvsem s hitrostjo reakcij na površini trdne faze in ne toliko s hitrostjo izmenjave ionov med trdno fazo in Slika I. Preperela površina vzorca apnenca iz Hotavelj (HR). Izrazito okrogle pore so rezultat vrtanja organizmov. Figure I. Weathered surface of the Hotavlje limestone (HS). Rounded pores are the result of biogenic activity. raztopino (Morse, 1983). Kadar je raztapljanje odvisno predvsem od reakcij na površini, je rezultat povečan relief trdne faze. Raztapljanje v tem primeru poteka namreč samo na nekaterih mestih na površini. Na večjih kalcitnih kristalih lahko opazujemo primer takšnega raztapljanja vzdolž razkolnih ploskev (slika 2), ki predstavljajo mesta povečane reaktivnosti. Ce pa je hitrost raztapljanja omejena s hitrostjo transporta izmenjave ionov med trdno fazo in okolno raztopino, torej z difuzijo, je navadno raztapljanje na površini bistveno hitrejše in enakomernejše, posledica je obljenje površine (Berner, 1981, Morse, 1983). Primer selektivnega raztapljanja prikazuje slika 3. Površina večjega kalcitnega kristala ni ravna, temveč ima luskast videz. Tudi meje med posameznimi zrni so vse bolj izrazite, razpoke se širijo, obenem je močnejši vpliv fizikalnega preperevanja in končni učinek je izpad celega kalcitnega kristala, nastala pora ima obliko karbonatnega zrna. Elementna analiza je pokazala, da je v kalcitu tudi manjši delež magnezija. Kvantitativne analize sestave na nepoliranih vzorcih niso pravilne, saj so odboji elektronov tudi posledica reliefa in ne le elementne sestave. V obravnavanem primeru lahko govorimo le o kalcitu z manjšim deležem magnezija. Obenem z raztapljanjem prihaja tudi do obarjanja; na površini večjih kalcitnih zrn opazujemo drobne, pod 5 mm velike kristale kalcita. Se večji je vpliv preperevanja na bolj mikritnem delu hotaveljskega apnenca; zaobljenost kristalov je večja, topografija je še izrazitejša. Mikrit ima večjo specifično površino kot sparit in je zato tudi bolj reaktiven. Topnost minerala narašča eksponencialno z njegovo specifično površino (Morse & Mackenzie, 1990). Drobni delci se raztapljajo drugače kot veliki, manjša zrna se lahko adhezivno vežejo na večja in tako preprečujejo njihovo raztapljanje (Mackenzie et al., 1983). Zelo pomembna je tudi heterogenost velikosti U5B 20kU lBQwm х20в Slika 2. Raztapljanje vzdolž razkolnih ploskev (vzorec HV). Figure 2. Dissolution on cleavage planes (sample HV). Slika 3. Večje kalcitno zrno ima luskast videz, ki je posledica selektivnega raztapljanja. Mikritni deli so bolj prizadeti in gradijo nižje dele površine (vzorec HS). Figure 3. As a result of selective dissolution the surface of bigger calcitic grain is rough. Micrite is even more weathered and is found on concave parts of the sample (sample HS). mineralnih zrn. Čim večje so razlike velikosti zrn v strukturi, hitreje poteka raztapljanje. Tako je biomikrit bolj topen kot čisti mikrit. Najbolj topni so biomikritni apnenci, hitrost raztapljanja pa se bistveno zmanjša, če jih sestavlja več kot 40 do 50 % sparita (Ford & Williams, 1994). Obnašanje kamnine v vlažnem okolju je v splošnem določeno z mikrostrukturo, še posebej s strukturo por (Meng, 1992). Pri študiju preperevanja kamnin je pomembna t.i. efektivna velikost por (to so tiste pore, ki so pomembne za transport vode v kamnini). Voda povzroča tako kemijsko kot fizikalno razpadanje (Amoroso & Fassina, 1983). Kamnina, izpostavljena fizikalnemu razpadanju, je bolj podvržena tudi večjim kemičnim in biološkim spremembam, ker povečana površina in olajšan dostop zraku, vodi in drugim snovem, pomeni intenzivnejše kemične reakcije. Velika razpokanost kamnine torej močno pospeši preperevanje in tako zmanjšuje njeno obstojnost. V lesnobrdskem apnencu opazujemo polja sparita, ki jih sekajo številne žilice in žile, zapolnjene z glinenimi minerali (klorit) in muskovitom ter kremenom, elementarna analiza je pokazala, da je poleg kalija, magnezija, aluminja, silicija in kisika tudi železo, ki daje kamnini barvo, z elementno analizo pa smo ugotovili tudi malo fosforja. Lahko je detritičnega ali organskega izvora. Tudi kemična analiza vzorca LBRd pokaže nekoliko povečano količino P2O5 (0,08 mas.%) glede na ostale vzorce. Na posameznih mestih najdemo lepe kristale apatita, velikosti 10 mm, ki je najverjetnejši vzrok za povečano količino fosforja v vzorcu. Tudi v tem vzorcu je količina magnezija v sparitu premajhna, da bi lahko ločili mesta s povečano koncentracijo magnezija od mest čistega CaCO3. Površine subhedralnihkalcitnihzrn, velikosti 5 mm, so gladke, ravne, poroznost je zelo majhna. Na posameznih mestih v lesnobrdskem apnencu opazimo zelo velike (celo več kot 500 mm) kristale kalcita, ki so praktično nepoškodovani (slika 4), mikritni deli so bolj prizadeti in gradijo nižje dele površine. Najbolj intenzivno je raztapljanje na mejah zrn, velja: čim manjši so delci, večja je specifična površina, večje je raztapljanje. Raztapljanju so močno izpostavljene razkolne razpoke in razpoke nastale zaradi tektonskih procesov (slika 5). Okrogle pore (nimajo oblike kalcitnih kristalov) so rezultat vrtanja organizmov v substrat. Raztapljanje je bolj intenzivno vzdolž že obstoječih razpok kot vzdolž razkolnih ploskev. Relief apnenca je posledica različne velikosti kalcitnih zrn; dvignjeni deli so bolj debelozrnati in so manj prizadeti kot konkavni deli na površini vzorca. Le-ti so zgrajeni iz drobnih, navadno z nečistočami bolj bogatih kalcitnih kristalov. Nepravilnosti in primesi v kalcitni rešetki povzročajo napetosti v strukturi, kar povzroča večjo nestabilnost mineralne faze. Sistem teži k zmanjšanju te energije (Wenk et al., 1983), torej so ta mesta bolj dovzetna za raztapljanje. Procesi raztapljanja se izmenjujejo z vmesnimi fazami precipitacije mineralov. Novonastali minerali so navadno zelo drobnozrnati, njihovi kristali so slabo razviti, saj je rast hitra, substrat pa jim predstavljajo starejša mineralna zrna (sliki 4 in 5). Glede na velikost zrn in fosilnih ostankov ločimo dva tipa lipiškega apnenca: enotni in rožasti, ki pa se po kemijski in mineralni sestavi praktično ne razlikujeta med seboj. Apnenec iz Lipice je med vsemi preiskovanci najbolj čist. Vsebuje preko 99 mas.% CaCO3, Slika 4. Na površini kalcitnih romboedrov opazujemo selektivno raztapljenje in precipitacijo sekundarnega kalcita (Vzorec LBRd). Figure 4. The effect of selective dissolution on calcite rombohedra and the precipitation of secondary calcite at the same time. torej je delež primesnih mineralov zelo majhen. Kljub temu smo z rentgensko difrakcijo ugotovili prisotnost glinenih mineralov iz kloritove skupine. Količina MgO je, glede na kemično analizo, v obeh strukturnih različkih okoli 0,3 mas.%, sparitni deli skoraj nimajo magnezija (pod mejo detekcije), v mikritnih delih pa je magnezija nekoliko več; včasih tudi do 0,5 mas.%. Enotni apnenec je bolj enakomernozrnat. Poleg kalcita so posamezna zrna glinencev, gre za kalijeve in natrijeve glinence velikosti okoli 30 mm, ki se včasih tudi medsebojno preraščajo. Ugotovili smo tudi apatit, ki vsebuje malo žvepla. Redke žilice zapolnjujejo drobni kristalčki glinenih mineralov - klorita. Pore imajo obliko izpadlih kalcitnih kristalov; verjetno so nastale pri pripravi vzorca. Slika S. Močno preperela površina apnenca. Ločimo lahko tri vrste razpok: - najbolj intenzivno je raztapljanje v podolžni smeri, vzdolž že obstoječih razpok, nastalih zaradi tektonskih procesov; - razpoke vzdolž razkolnih ploskev; - okrogle pore so rezultat delovanja organizmov (vzorec LBRd). Figure S. Highly weathered limestone surface. Three types of fissures can be distinguished: - the dissolution along tectonic fissures is the most intensive; - fissures on the cleavage planes; - rounded fissures of biogenic origin (sample LBRd). Mineralna sestava rožastega različka je enaka kot pri enotnem. Bistvena razlika med apnencema je v zrnavosti. Rožasti apnenec je zelo nehomogen, v njem se menjavajo veliki sparitni kristali z mikritnimi območji (5 mm). Pore so bolj pogoste kot pri enotnem različku; gre za zaprto poroznost. Majhen del por je tudi tu nastal pri pripravi vzorca. Apnenec z Drenovega Griča je po sestavi zelo pester. S kemično analizo smo ugotovili zelo veliko primesi: baker, magnezij, aluminij, silicij, žveplo, železo... Reliefna preperela površina ni rezultat razlike v sestavi, pač pa je nastali relief posledica različne velikosti delcev. Mehansko in kemično preperevanje pospešujejo še organizmi. Pri večji povečavi opazimo, da so kristali kalcita nepravilnih oblik; gre za anhedralna zrna, velika okoli 5 mm. Površine kristalov niso ravne, nekateri robovi zrn so deloma že zaobljeni. Površine kristalov na konveksnem delu površine so bolj reliefne, njihovi robovi pa so ravni in do zaobljevanja še ni prišlo (tu je izrazitejše selektivno raztapljanje). Medtem ko so kalcitni kristali na konkavnih mestih na površini vzorca bolj zaobljeni, njihove površine pa so manj poškodovane. Primer močno prizadetih kristalov kalcita je na sliki 6. Po sestavi gre za kalcit z malo magnezija. Na površini so že nastale manjše razpoke, ob katerih je raztapljanje intenzivnejše. Stiki med zrni so lepo vidni, zrna so tudi že dobro zaobljena. Kristalni defekti predstavljajo točke začetka raztapljanja. Kjer so kristalne ploskve pravilne, so njihove površine gladke, robovi so ostri, v nasprotju s tistimi mesti, kjer so kristalne ploskve slabše razvite ali so bile nepravilnosti v kristalni strukturi in sestavi. Kemična analiza je pokazala, da je v kristalih kalcita na takšnih mestih še magnezij, kalij, železo, titan, aluminij in silicij. Izvora teh elementov z omenjenimi metodami nismo mogli določiti, verjetno pa je bil vsaj del teh elementov vnešen v kamnino kasneje. Možno pa je, da je del kristala, ki je bolj poškodovan, imel povišano količino magnezija in železa, ki sta pospešila raztapljanje. Idealni, stehiometrijski kristal brez defektov je najbolj stabilna oblika minerala. Zaradi substitucije kationov in anionov ter mrežnih defektov se obstojnost mineralov zmanjšuje (Blatt & Tracy, 1995). Slika б.Мобпо preperei kalcit s povečano vsebnostjo magnezija (Vzorec DGC). Figure 6. Weathered calcite with increased magnesium content (Sample DGC). Sklepi Obstojnost in s tem uporabna vrednost apnencev, ki se uporabljajo kot naravni kamen, je odvisna od kemične in mineralne sestave ter njihovih strukturnih značilnosti. Preiskovani slovenski apnenci, z izjemo vzorca glinenih vložkov v hotaveljskem apnencu, so kemično relativno čisti CaCO3. Z izjemo apnenca z Drenovega Griča (91 mas.% CaCO3) imajo preko 95 mas.% CaCO3. Po sestavi so si zelo podobni, delež primesnih prvin je relativno majhen. Z elektronskim mikroskopom lahko zelo učinkovito ločimo nekarbonatne minerale v karbonatnih kamninah. Nekarbonatni minerali nastopajo kot detritična zrna (kremen, glinenci, sljude, apatite) ali kot kasneje nastali, diagenetski minerali (kremen, pirit, hematit, apatit, barit, anhidrit). Količina magnezija v kalcitu je premajhna, da bi lahko s pomočjo elektronskega mikroskopa točno določili lego primesi v kalcitni rešetki. Na splošno večji sparitni kristali vsebujejo manj magnezija kot mikrosparitni ali mikritni deli in tudi poroznost je v sparitnem delu bistveno manjša kot v mikritnem delu. Predvsem velikost in habitus kristala določata njegovo odpornost. Konveksne dele preperele površine gradijo v glavnem sparitni kristali z lepo razvitimi ploskvami, kjer je navadno tudi manj magnezija v strukturi kot v mikritu, ki je raztapljanju in fizikalnim učinkom površinskega spreminjanja bolj podvržen. To ugotovitev smo lahko potrdili pri vseh preiskovanih vrstah apnenca. Iz oblikovanosti por lahko sklepamo na njihov nastanek. Okrogle pore so biogenega nastanka, pore, nastale zaradi izpadanja kristalov, imajo obliko le-teh, pore z nepravilnimi zaobljenimi robovi pa so nastale zaradi raztapljanja mineralnih zrn. Na površini kalcitnih kristalov pa lahko ločimo tri vrste sekundarnih razpok: razpoke, ki so posledica naključnega raztapljanja, razpoke vzdolž razkolnih ploskev (z značilno usmerjenostjo) in razpoke, ki nastanejo zaradi lomljenja in drobljenja mineralnih zrn. Le-to povzroča raztapljanje in fizikalno preperevanje. Tudi rezultati rentgenske difrakcije ne pokažejo večjih razlik med apnenci. Mineralna sestava apnencev ne more povzročiti tako velikih razlik v reaktivnosti oziroma obstojnosti kamnine. Obstojnost kamnine zaradi procesov preperevanja je odvisna predvsem od strukture apnenca, ki jo pogojuje način nastanka in vrsta diagenetskih procesov. Summary The weathering durability of limestones as a function of their mineral composition and texture The effects of atmospheric weathering on some limestones from Slovenia were investigated by electron microscope. Samples were chosen from several active or temporary active limestone quarries: Hotavlje, Lesno Brdo, Drenov Grič and Lipica. Samples of fresh rock and weathered rock from abandoned parts of the quarries were taken as well. The weathered rock samples have been exposed to atmospheric conditions and effects of different organisms for about thirty years. Investigated limestone samples are very homogenous; they are chemically relatively pure CaC03. The samples HV represent the clay layer within the Hotavlje quarry. Limestones consist of more than 95 % of CaC03. The only exception is the limestone from Drenov Grič with the 91 % CaC03. Being so alike, they are very hard to be distinguished. Some rare noncarbonate minerals represent the detritic grains (quartz, feldspars, mica, apatite...) and/or younger diagenetic minerals (quartz, pyrite, hematite, apatite, mica, anhydrite). Magnesium content in calcitic grains is also detectable. The bigger the sparite crystal is the smaller is the magnesium content. Therefore, the micritic parts have more variable chemical composition and the porosity is higher, too. The weathered surfaces display the influence of crystal size and habit on the limestone durability. The convex parts of the weathered sample consist of well shaped bigger sparitic crystals with fine developed crystal planes, Reference ACME (1999): ACME analytical laboratories brochure. Acme Analytical Laboratories, Ltd., Vancouver; 18 p. Amoroso, G.G. & Fassina, V. (1983): Stone decay and conservation, atmospheric pollution, cleaning, consolidation and protection. Elsevier Science, Amsterdam; 4S3 p. Berner, R.A. (1981): Kinetics of weathering and diagenesis. Reviews in mineralogy, 8, 111-134, Washington. Bilbija, N. & Grimšičar, A. (1987): Obstojnost arhitektonskega naravnega kamna iz Slovenije. Geološki zbornik 8, 1S1-160, Ljubljana. Blatt H. & Tracy, R.J. (199S): Petrology: Igneous, sedimentary and metamorphic. W.H. Preeman and Company, New York; S29 p. Drever, J.I. (1997): The geochemistry of natural waters. Surface and groundwater environments. Prentice Hall, New Jersey; 436 p. whereas the concave parts are more micritic, with higher content of magnesium and other impurities. The pores in limestone are of different origin: rounded pores are the result of biological activity, some pores have the shape of the calcitic grains and some of them have irregular shapes and are the result of selective dissolution of carbonate because of the higher impurity content in calcite lattice. Also, the results of X-ray diffraction do not show some major differences between investigated limestones. Therefore, the high differences in weathering durability of the limestones are not caused by the chemical and mineralogical composition, rather are the result of limestone texture. The size of the calcite crystal and its habit are the most important features in durability of stone. The content of impurities in calcite lattice is also important, too. Ford, D.C. & Williams, P.W. (1994): Karst geomor-phology and hydrology. Chapman & Hall, London; 601 p. Gams, I. (1974): Kras. Slovenska matica, Ljubljana; 3S9 p. Jarc, S. (1996): Toplotne lastnosti naravnega kamna v Sloveniji. Diplomsko delo. Ljubljana: Univerza v Ljubljani; 90 p. Jarc, S. (2000): Vrednotenje kemične in mineralne sestave apnencev kot naravnega kamna. Magistrsko delo. Ljubljana: Univerza v Ljubljani; 88 p. Krinsley, D.H., Pye, K., Boggs, S.Jr., Tovey, N.K. (1998): Backscattered scanning electron microscopy and image analysis of sediments and sedimentary rocks. Cambridge University Press, Cambridge; 193 p. Mackenzie, F.T., Bischoff, W.D., Bishop, F.C., Loijens, M., Schoonmaker, J., Wollast, R. (1983): Mag-nesian calcites: low-temperature occurence, solubility and solid-solution behavior. Reviews in Mineralogy, 11, 97-144, Washington. Meng, B. (1992): Moisture - transport - relevant characterization of pore structure. Proceedings of the 7'h International congress on deterioration and conservation of stone, 387-397, Lisboa. Mirtič, B., Mladenovič, A., Ramovš, A., Senegačnik, A., Vesel, J., Vižintin, N. (1999): Slovenski naravni kamen. GeoZS, ZAG, Oddelek za Geologijo NTP, Ljubljana; 131 p. Morse, J.W. (1983): The kinetics of calcium carbonate dissolution and precipitation. Reviews in mineralogy, 11, 227-264, Washington. Morse, J.W. & Mackenzie, F.T. (1990): Geochemistry of sedimentary carbonate. Developments in sedimetology 48, Elsevier, Amsterdam; 177 p. Ramovš, A. (1987): Triasne gradbene in okrasne kamnine v severni Sloveniji. Geološki zbornik 8, 2S-3S, Ljubljana. Summerfield, M.A. (1994): Global geomorphology, an introduction to the study of landforms. John Wiley 8 Sons, New York; S37 p. Viles, H.A. & Moses, C.A. (1998): Experimental production of weathering nanomorphologies on carbonate stone. Quarterly Journal of Engineering Geology, 31, 347-3S7, Oxford. Wenk, H.R., Barber, D.J., Reeder, R.J. (1983): Mi-crostructures in carbonates. Reviews in Mineralogy, 11, 301-367, Washington. Razlikovanje apnencev s pomočjo statističnih metod The distinction of limestone by statistical methods Simona Jarc Oddelek za geologijo, NTF, Univerza v Ljubljani, Aškerčeva 12, Ljubljana, Slovenija; E-mail: simona.jarc@ntfgeo.uni-lj.si Received: October 20, 200S Accepted: November 24, 200S Izvleček: Preiskani slovenski apnenci imajo enostavno kemično in mineralno sestavo, so razmeroma čisti; vsebujejo preko 91 mas.% CaCOQ. Kljub majhnim razlikam v kemični sestavi in majhni vsebnosti magnezija v kalcitni rešetki (manj kot 0,6 mas.%), pa lahko apnence z natančno kemično analizo med seboj razlikujemo. Glede na rezultate t-testa in clusterske analize se preiskovani apnenci najbolj razlikujejo glede vsebnosti SiO2, AlpOQ, MgO in Sr. Na osnovi detajlne kemične analize in uporabe ustreznih statističnih metod bi tako apnencem lahko celo določili njihov izvor. Abstract: The analysed Slovenian limestones have relatively simple chemical and mineral composition. They contain above 91 wt.% CaCOQ, respectively. However, these small differences in composition and the very small amount of magnesium in calcite lattice (less than 0'6 wt.%) may aid in differentiating the limestone. The results of t-test and cluster analysis are in good agreement. The content of SiOP, AlpOQ, MgO and Sr have the largest influence in distinction of investigated limestones. Therefore, the use of statistical methods is very helpful in their identification. Key words: limestone, chemical composition, t-test, cluster analysis Ključne besede: apnenec, kemična sestava, t-test, clusterska analiza Uvod Statistične metode nam lahko pomagajo pri identifikaciji kamnin in določanju njihovega izvora. Provenienco apnenca sem skušala določiti na osnovi njegove detajlne kemične analize in uporabo t-testa enakosti populacijskih povprečij ter clusterske analize. Vzorci so bili izbrani iz nekaterih aktivnih in občasno delujočih kamnolomov apnenca: iz Lipice (LiUl, LiU2, LiFI in LiF2), Hotavelj (HI do H7), Lesnega Brda (LBl in LB2) in Drenovega Griča (DGl in DG2). Mineralna sestava preiskovanih apnencev je opisana v Jarc & Mirtič (v tisku). Za primerjavo sem analizirala še dva vzorca glinenih primesi v hotaveljskem apnencu (HVl in HV2). Analitika Kemična analiza je bila opravljena v laboratoriju ACME v Kanadi. Vsebnosti oksidov in prvin ter žarilne izgube (LOI -loss of ignition) so bile določene z metodo induktivno vezane plazme emisijske spektro- Tabela I. Kemična sestava vzorcev. Količine oksidov, žarilne izgube (LOI), ogljika in žvepla so v masnih odstotkih, ostalih prvin v ppm. Table I. Chemical composition of investigated samples. All oxides, carbon, sulfur and loss of ignition content are in weight %, other elements are in ppm. Si02 A1203 Fe203 MgO CaO na^o k20 Ti02 p2o5 MnO Cr203 LiUl <0,02 0,12 0,07 0,35 55,63 0,02 <0,04 0,01 0,01 <0,01 0,001 LiU2 0,02 0,1 0,1 0,3 55,96 0,01 0,04 0,02 0,03 <0,01 0,01 LiFl <0,02 0,09 0,07 0,31 55,76 0,03 <0,04 0,01 0,03 <0,01 0,004 LiF2 <0,02 0,08 0,09 0,28 56,03 0,02 <0,04 <0,01 0,02 <0,01 0,006 LB1 1,26 0,77 0,29 0,42 54,00 0,03 0,12 0,04 0,08 0,01 0,003 LB2 1,66 0,99 0,34 0,44 53,82 0,02 0,18 0,04 0,01 0,01 0,005 HI 0,21 0,17 0,05 0,91 54,91 0,03 <0,04 0,01 0,04 0,02 0,004 Hip 0,38 0,15 0,05 0,94 54,56 <0,01 <0,04 <0,01 0,02 0,02 0,004 H2 0,31 0,17 0,07 0,69 55,42 0,02 0,04 0,01 0,03 0,02 0,003 H3 1,12 0,48 0,21 0,67 53,85 0,04 0,12 0,01 0,05 0,03 0,002 H4 1,33 0,64 0,24 0,62 54,03 0,02 0,19 0,02 0,04 0,03 0,001 H5 0,58 0,31 0,59 0,52 54,24 <0,01 0,06 0,01 0,04 0,05 0,006 H6 0,84 0,42 0,75 0,45 54,08 0,04 0,12 0,01 0,03 0,05 0,006 H6p 0,95 0,42 0,72 0,46 54,16 <0,01 0,12 <0,01 0,02 0,05 0,003 H7 0,83 0,42 0,79 0,47 54,49 0,01 0,1 0,01 0,01 0,05 0,001 HV1 17,29 8,36 5,91 5,09 29,75 0,03 3,07 0,18 0,05 0,08 0,01 HV2 18,44 8,77 6,09 4,78 28,96 0,01 3,26 0,19 0,07 0,08 0,01 DG1 1,47 0,77 0,51 0,83 51,75 0,03 0,12 0,03 0,02 0,01 0,004 DG2 1,73 0,88 0,82 0,77 51,00 0,05 0,11 0,04 0,01 0,01 0,002 Ba Ni Sr Zr Y Nb Sc LOI С/ТОТ БЯОТ LiUl <5 <20 189 <10 <10 <10 <10 43,7 11,7 0,01 LiU2 <5 <20 188 100 <10 <10 <10 43,4 11,7 0,03 LiFl <5 <20 261 <10 <10 <10 <10 43,6 11,6 0,01 LiF2 <5 <20 220 <10 <10 <10 <10 43,4 11,7 0,04 LB1 5 <20 133 139 <10 <10 <10 42,9 11,2 0,01 LB2 9 <20 110 <10 <10 <10 <10 42,5 11,4 0,01 HI 5 <20 101 <10 <10 <10 <10 43,6 11,8 0,01 Hip <5 <20 100 13 <10 <10 <10 43,7 11,4 0,01 H2 <5 <20 98 <10 <10 <10 <10 43,2 11,7 0,01 H3 <5 <20 96 125 <10 <10 <10 43,4 11,6 0,01 H4 5 <20 95 <10 <10 <10 <10 43 11,4 0,01 H5 8 <20 99 <10 <10 <10 <10 43,6 11,3 0,01 H6 5 <20 94 10 <10 <10 <10 43,1 11,5 0,01 H6p 7 <20 95 <10 <10 <10 <10 43,1 11,5 0,01 H7 7 <20 96 <10 <10 <10 <10 42,9 11,8 0,01 HV1 69 25 59 77 <10 <10 <10 30,2 7,61 0,01 HV2 71 24 55 95 10 <10 <10 29,4 7,07 0,03 DG1 13 <20 972 12 <10 <10 <10 44,4 13,1 0,23 DG2 13 <20 1055 29 <10 <10 <10 44,5 14,2 0,37 metrije (ICP-ES), celotni ogljik in žveplo pa z Leco (ACME, 1999). Za silikatno analizo so 0,2 g vzorca talili z 1,2 g LiB02 in raztopili v 100 ml 5 % HN03. Žarilno izgubo so določili glede na spremembo mase vzorca po 1 uri žganja pri temperaturi 1000 °C. Rezultati kemične analize so v tabeli 1. Rezultati kemične analize kažejo, da so količine niklja, cirkonija, itrija, niobija in skan-dija v vseh vzorcih pod mejo detekcije, zato ti rezultati v nadaljevanju niso upoštevani. Pravilnost analitike je bila ugotovljena s pomočjo standardov S0-15 in CSA, natančnost ali ponovljivost izbrane metode pa z večkratnimi analizami (vzorci označeni s p) istega vzorca (Swan & Sandilands, 1995, Zupančič, 1994). Pravilnost metode določanja kemične sestave je zelo dobra, saj odstopanja analitskih in priporočenih vrednosti niso večje od 6 %, v splošnem pa so analitske nekoliko nižje od priporočenih vrednosti, prav tako je natančnost metode zadovoljiva (Jarc, 2000). Slabšo natančnost zasledimo le pri tistih oksidih in prvinah, katerih količine so zelo majhne, še posebej, če je količina na meji detekcije. Pri višjih količinah (npr. CaO) je natančnost dane metode določanja kemične sestave dobra (Jarc, 2000). Osnovne statistike, normalnost PORAZDELITVE IN KORELACIJE Osnovne statistike količin oksidov in prvin 19 vzorcev (N) so podane v tabeli 2. Srednje vrednosti so podane z aritmetično sredino (x), mediano (Me), geometrično sredino (xG), razpršenost podatkov pa z razponom (razlika med najvišjo (max) in najnižjo (min) vrednostjo) ter z aritmetičnim (s) in geometrijskim (sG) standardnim odklonom ter z aritmetično (s2) in geometrično (sG2) X *G Me min-max Q25-Q75 s s5 sa sa2 ?b, b2 ?b1L Ьгь Si02 2,550 0,465 0,840 0,01-18,44 0,21-1,47 5,431 29,49 9,399 88,340 2,746 6,353 0,299 1,376 ai2o3 1,269 0,428 0,420 0,08-8,77 0,15-0,77 2,588 6,70 3,788 14,347 2,739 6,309 10,469 10,222 FeA 0,935 0,304 0,290 0,05-6,09 0,07-0,75 1,807 3,26 4,265 18,187 2,679 6,075 4,560 1,132 MgO 1,016 0,655 0,520 0,28-5,09 0,42-0,83 1,397 1,95 2,228 4,964 2,697 6,172 56,535 952,899 CaO 51,705 50,911 54,160 28,96-56,03 53,82-55,42 7,980 63,69 1,216 1,479 -2,669 6,030 0,002 1766453 Na20 0,022 0,018 0,020 0,005-0,05 0,01-0,03 0,013 0,00 2,079 4,321 0,358 0,538 0,208 4,571 K20 0,408 0,094 0,110 0,02-3,26 0,02-0,12 0,974 0,95 4,408 19,430 2,788 6,499 19,854 67,504 Ti02 0,035 0,017 0,010 0,005-0,19 0,01-0,04 0,054 0,00 2,923 8,543 2,561 5,624 12,139 4,807 p2o5 0,032 0,027 0,030 0,01-0,08 0,02-0,04 0,020 0,00 1,920 3,685 1,002 0,689 0,641 6,499 MnO 0,028 0,012 0,020 0,0005-0,08 0,01-0,05 0,026 0,00 6,050 36,601 0,861 0,221 0,095 2,083 cr2oj 0,005 0,004 0,004 0,001-0,01 0,002-0,006 0,003 0,00 2,090 4,367 0,865 0,065 0,374 3,189 Ba 12,526 6,624 5,000 3,0-71,0 3,0-9,0 20,500 420,26 2,633 6,933 2,680 6,066 39,271 148,143 Sr 216,632 141,791 100,000 55,0-1055,0 95,0-189,0 286,255 81942 2,223 4,943 2,634 5,939 44,475 294,868 LOI 41,979 41,723 43,400 29,4-44,5 42,9-43,6 4,321 18,67 1,127 1,269 -2,736 6,317 0,002 2570946 с 11,331 11,207 11,600 7,07-14,2 11,4-11,7 1,574 2,48 1,173 1,377 -1,580 4,004 0,008 79428,6 s 0,044 0,017 0,010 0,01-0,37 0,01-0,03 0,094 0,01 2,990 8,938 3,068 9,080 133,374 5490,109 Tabela 2. Osnovne statistike: aritmetična (x) in geometrična (xG) sredina, mediana (Me), razpon vrednosti (min-max), kvartilni razpon (Q25-Q75), aritmetični (s) in geometrični (sG) standardni odklon, aritmetična (s2) in geometrična (sG2) varianca, asimetričnost (Vbj) in sploščenost (b2) naravnih vrednosti ter asimetričnost (VbIL) in sploščenost (b2L) logaritmiranih vrednosti. Število vzorcev (N) je 19. Table 2. Basic statistics: arithmetic (x) and geometric (xG) mean, median (Me), range (min-max), quartiles (Q2S-Q ), arithmetic (s) and geometric (sG) standard deviation, arithmetic (s2) in geometric (sG2) variance, skew-ness (VbI), kurtosis (b2) and skewness (VbIL) and kurtosis (b2L) of logaritmic values. The number of samples (N) is 19. varianco. Zaradi t.i. "outlierjev" (nenormalno visoke ali nizke vrednosti) je podan še kvartilni razpon (Q25-Q75), kjer zanemarimo 25 % najvišjih in 25 % najnižjih vrednosti (Swan & Sandilands, 1995), ki daje realnejšo sliko razpršenosti podatkov. Metode parametrične statistike lahko uporabimo samo, če so podatki porazdeljeni normalno. Ce porazdelitev podatkov ne ustreza normalni, jih moramo z ustrezno transformacijo prilagoditi temu pogoju (Koch & L ink, 1970). Normalnost porazdelitve naravnih in logaritmiranih vrednosti sem ugotavljala s testi sploščenosti (b2, b2L) in asimetričnosti (Vbp Vb1L). Za normalno porazdeljene podatke se vrednost asimetričnosti približije 3, vrednost sploščenosti pa 0. Iz rezultatov je razvidno, da normalni porazdelitvi bolj ustrezajo naravne vrednosti. Vsi parametri so izračunani s pomočjo računalniškega programa CSS Statistica. Zaradi primerljivosti rezultatov (podatki so v različnih enotah) sem pri izračunu korelacijskih koeficientov uporabila logaritmirane vrednosti. S tem sem tudi zmanjša napaka zaprtega niza podatkov (vrednosti so podane v odstotkih), kjer so vrednosti že same po sebi medsebojno odvisne (Swan & S andilands, 1995). Izračunane koeficiente korelacije sem primerjala s kritično (tabelirano) vrednostjo koeficienta na 95 % ravni zaupanja, r005 17knt a 0,456 (Petz, 1985). Korelacijska matrika (tabela 3) kaže, da je CaO z večino oksidov drugih elementov negativno koreliran. CaO je vezan predvsem na mineral kalcit, ki je kemično relativno čist. Ostale prvine so torej vezane na druge minerale, predvsem glinene minerale. To potrjuje tudi dobra korelacija SiO2, MgO, Al2O3 in Fe2O3 ter K2O, ki so vezani na minerale smektitove in kloritove skupine ter na muskovit (Jarc & Mirtič, v tisku). V nekaterih Tabela 3. Matrika korelacijskih koeficientov logaritmiranih vrednosti (N=19). * - prvini sta značilno medsebojno odvisni na 9S % ravni zaupanja. Table 3. Matrice of correlation coefficients of logaritmic values (N=19). * - correlations are significant on 9S % confidence interval. Si02 Si02 1 ai2o3 ai2o3 0,91* 1 fea fea 0,80* 0,90* 1 MgO MgO 0,75* 0,81* 0,66* 1 CaO CaO -0,64 -0,85* -0,78* -0,91* 1 Na20 najo 0,03 0,13 -0,01 0,02 -0,01 1 K20 к2о 0,85* 0,97* 0,92* 0,78* -0,87* 0,07 1 Ti02 тю2 0,66* 0,87* 0,71* 0,72* -0,82* 0,28 0,84* 1 P205 p2o5 0,31 0,32 0,20 0,41 -0,41 0,07 0,38 0,33 1 MnO MnO 0,88* 0,66* 0,63* 0,61* -0,43 -0,16 0,62* 0,28 0,31 1 Cr203 Cr203 0,14 0,24 0,28 0,36 -0,47* -0,13 0,30 0,36 0,41 0,02 1 Ba Ba 0,77* 0,92* 0,90* 0,84* -0,90* 0,01 0,90* 0,84* 0,22 0,54* 0,40 1 Sr Sr -0,29 -0,24 -0,19 -0,32 0,34 0,39 -0,34 -0,05 -0,46* -0,49* -0,17 -0,14 1 LOI LOI -0,58* -0,80* -0,73* -0,87* 0,98* 0,05 -0,85* -0,77* -0,45 -0,39 -0,49* -0,84* 0,47* 1 с с -0,49* -0,68* -0,61* -0,77* 0,90* 0,18 -0,76* -0,64* -0,55* -0,37 -0,50* -0,70* 0,66* 0,96* 1 S 0,08 0,15 0,20 0,09 -0,08 0,28 0,07 0,28 -0,28 -0,19 0,09 0,27 0,84* 0,06 0,30 primerih (n.pr. korelacija koeficienta med CaO ali MgO in Ba) je korelacijski koeficient relativno visok in je posledica posameznih nenormalno visokih ali nizkih vrednosti. Test enakosti populacijskih povprečij S t-testom enakosti populacijskih povprečij sem poskušala ugotoviti, ali lahko preiskovane apnence med seboj statistično razlikujemo. Tabela 4. t-test enakosti populacijskih povprečij. F - rezultat testa populacijskih varianc posameznih skupin, t - rezultat testa enakosti populacijskih povprečij posameznih skupin, * - skupini se na 9S % ravni zaupanja statistično razlikujeta. Table 4. t test of equality of population means. Fx y - population variance result of individual groups, tx y - t-test of equality of population means result of individual groups, * - the groups are statistically different on 9S % confidence interval. f,.2 tl-2 f1.3 tl-3 fm tl-4 f1.5 ti-s f2-3 t2-3 Si02 3200,0 -11,813* 5871,78 -3,643* 26450,0 -50,698* 1352,0 -19,919* 1,835 2,509* ai2o3 82,971 -11,412* 95,143 -2,991* 288,170 -67,102* 20,743 -20,189* 1,147 4,073* Fe203 5,556 -12,238* 459,346 -1,839 72,000 -105,200* 213,556 -6,094* 82,682 -0,298 MgO 4,333 -5,237* 40,154 -3,401* 55,440 -47,459* 2,077 -17,060* 174,000 -1,503 CaO 2,076 13,059* 7,273 5,500* 9,280 95,199* 8,362 16,699* 15,101 -1,381 N^O 1,333 -0,730 3,135 0,071 3,000 0 3,000 -2,309 4,181 0,514 к2о 18,000 -6,299* 32,444 -2,138 180,500 -53,532* 2,000 -11,110* 1,803 1,433 Ti02 0,000 -6,093* 2,111 0,421 1,260 -30,889* 1,263 -4,222* 0,000 9,400* P2O5 26,727 -0,995 1,758 -1,202 2,180 -3,974* 1,833 0,961 15,207 0,872 MnO - - 0,000 -4,804* - - - - 0,000 -2,435* Cr203 7,125 0,432 4,071 1,258 - -1,678 7,125 0,777 1,750 0,467 Ba 0,000 -3,266* 0,000 -2,168 - -109,411* - - 2,215 1,193 Sr 4,468 3,480* 193,394 10,807* 147,710 6,102* 2,915 -22,071* 43,282 5,287* LOI 3,556 4,961* 3,827 1,498 14,220 50,919* 4,500 -7,934* 1,076 -2,577* С 8,000 5,222* 13,111 1,269 58,320 25,569* 242,000 -5,828* 1,639 -1,846 S 0,000 1,111 0,000 2,655* 1,130 0,195 43,556 -6,262* - - FM tM F2-5 t2-5 FM t3.4 F3-5 t3-5 F4-5 t4-s Si02 8,266 -26,947* 2,367 -0,587 4,505 -48,541* 4,343 3,045* 19,564 -27,591* ai2o3 3,473 -33,033* 4,000 0,447 3,029 -56,963* 4,587 3,790* 13,893 -36,467* Fe203 12,960 -60,862* 38,440 -2,229 6,380 -23,467* 2,151 1,147 2,966 -29,766* MgO 240,250 -29,004* 9,000 -11,700* 1,381 -28,870* 19,333 1,184 26,694 -26,191* CaO 19,262 60,611* 17,361 6,573* 1,276 63,845* 1,150 -7,799* 1,110 40,430* NajO 4,000 0,447 4,000 -1,342 1,045 -0,049 1,045 1,823 1,000 1,414 к2о 10,028 -30,263* 36,000 1,151 5,563 -56,295* 64,889 0,648 361,000 -32,061* Ti02 0,000 -29,000* 0,000 1,000 2,667 -47,488* 2,667 6,784* 1,000 -21,213* p2o5 12,250 -0,412 49,000 0,849 1,241 -2,873* 3,222 -1,690 4,000 -4,025 MnO - - - - 0,000 -4,235* 0,000 -2,435* - - Cr203 0,000 -6,000* 1,000 0,707 0,000 -4,835* 1,750 -0,233 0,000 -7,000* Ba 4,000 -28,175* 0,000 -3,000 1,806 -44,805* 0,000 5,632* 0,000 -57,000* Sr 33,063 5,526* 13,023 -20,713* 1,309 20,409* 563,646 59,500* 430,563 23,022* loi 4,000 28,845* 16,000 -8,489* 3,716 51,537* 17,222 5,349* 64,000 36,342* С 7,290 13,754* 30,250 -4,204 4,448 25,326* 18,458 8,631* 4,150 10,299* S 0,000 -1,000 0,000 -4,143 0,000 -2,714* 0,000 11,242* 49,000 3,960 Si02 0,801 0,301 - AI203 : :: ; :: t=o=r::: 0,071 ;; 0,021 и Lipica Lesno Brdo Hotavlje Hot-pnmesi Drenov Grič Fe203 1,005 ; 0,505 ••• Lipica Lesno Brdo Hotavlje Hot.-primesi Drenov Grlo CaO Lipica Lesno Brdo Hotavlje Hot-primeel DrenovGrlc K20 Ш Ш ' Lipica Lesno Brdo Hotavlje HoL-pri mesi Drenov Gric Lipica Lesno Brdo Hotavlje MgO Hot.-primesi Drenov Gric :: r-p-i ; L_u_J S Lipica Lesno Brdo Hotavlje HoL-primesi Drenov Gric Na20 T Upica Lesno Brdo Hotavlje Hot.-primeel Drenov Gric P205 ....... ~r .....1 ° l 0 .. □ i :: Lipica Lesno Brdo Hotavlje HoL-primesi Drenov Gric Sr 1 5 t П ...........ш.......... ........._ .. ............ ..........1 И 1.......... Lipica Lesno Brdo Hotavlje HoL-primesi Drenov Gric Lipica Lesno Brdo Hotavlje HoL-primesi Drenov Gric Vzorce sem razporedila v pet skupin glede na posamezna nahajališča. V prvi skupini so vzorci iz Lipice (LiUl, LiU2, LiFI in LiF2), v drugi skupini je lesnobrdski apnenec (LBl in LB2), sledi skupina hotaveljskega apnenca (Hl - H7). Vzorec HV predstavlja svojo (četrto) skupino, ker se že po sestavi močno razlikuje od vseh ostalih; gre za glinene primesi v hotaveljskem apnencu. V peto skupino sem uvrstila črni apnenec z Drenovega Griča (DGl in DG2). Dvostranski t-test sem izvajala z računalniškim programom CSS z naravnimi vrednostmi. Vrednosti t so izračunane na podlagi izidov testa F, ki ugotavlja podobnost populacijskih varianc. Rezultati (tabela 4) so pokazali, da se skupine med seboj statistično razlikujejo po količini večine oksidov in prvin na ravni zaupanja 95 %. Torej bi lahko že z geokemično analizo apnenca določili njegov izvor oz. nahajališče. Seveda bi bilo potrebno narediti bistveno večje število kemičnih analiz vseh apnencev iz različnih nahajališč, da bi lahko postavili statistične meje količin posameznih oskidov in prvin, na osnovi katerih bi ugotavljali izvor apnenca. Nekatere vrednosti F so zelo visoke in nam pri tako majhnem številu vzorcev kažejo na "nestabilnost" podatkov; varianca znotraj posamezne skupine je ob tako majhnem vzorcu prevelika, zato so vrednosti nezanesljive; primer apnenca iz Hotavelj, ki ima pri večini analiziranih prvin zelo širok vrednosti (slika l). V našem primeru je bilo vzorcev apnencev vsekakor premalo, da bi lahko točno definirali kemično sestavo apnencev iz posameznih nahajališč. Kljub temu pa je test enakosti populacijskih povprečij pokazal, da se vzorec HV, ki predstavlja glinene primesi v hotaveljskem apnencu, popolnoma razlikuje od vseh ostalih (vrednosti t so zelo visoke). Glede na izračunane vrednosti t za vse prvine v dani skupini se statistično najmanj razlikujeta apnenca z Lesnega Brda in iz Hotavelj (vrednosti t so malo višje od kritične vrednosti t na ravni zaupanja 95 %), kar je tudi razumljivo, saj sta genetsko podobna (Jarc, 2000). Apnenec z Drenovega Griča se od ostalih apnencev razlikuje predvsem po povečani količini organske snovi. Posamezna nahajališča se razlikujejo predvsem po vsebnosti Si02, Al203, MgO in Sr, medtem ko se vrednosti Fe203 (kljub visokim vrednostim t), CaO, K20, P20S, Nap0 lahko prekrivajo in jih ne moremo uporabili za identifikacijo apnencev (slika l). Za ločevanje različnih tipov apnencev so vsekakor zanesljivejši tisti oksidi in prvine, ki nastopajo v večjih količinah, medtem ko oksidi in prvine v manjših količinah pokažej o tudi manjše statistične razlike. Kot primer lahko navedemo Na20, katerega vsebnost je v vseh vzorcih na meji detekcije, t-test pa ne da nobenih statistično značilnih razlik med izbranimi skupinami glede na njegovo količino. B Slika I. Prikaz srednjih vrednosti, standardnih odklonov ter maksimalnih in minimalnih vrednosti nekaterih oksidov oz. prvin v posameznih nahajališčih. Skala je logaritemska. Figure I. Box-Whiskers diagrams of means, standard deviations and ranges (max-min) of some analysed oxides and elements in investigated limestone quarries. Logarithmic scale. Clusterska analiza Cilj clusterske analize je združevanje posameznih objektov v skupine glede na podobnost med njimi. Uporabila sem clustersko analizo tipa k-mean, ki se od običajne loči po tem, da vnaprej določimo število skupin (k). Osnova izračuna te metode je "obrnjena" analiza variance. Računalniški program CSS razdeli vzorce v vnaprej določeno število skupin (k) tako, da je varianca znotraj skupine čim manjša, varianca med skupinami pa čim večja. Statistično značilnost rezultatov določi s testom F. Višja kot je vrednost F, bolj učinkovito spremenljivka vpliva na ločevanje vzorcev v skupine. Uporabljeni podatki pa morajo biti standardizirani tako, da je povprečna vrednost posamezne spremenljivke 0, standardni odklon pa I (Zupančič, 1994). Ker sem poskušala ugotoviti, ali se apnenci iz posameznih nahajališč statistično razlikujejo, sem tudi clustersko analizo tipa k-mean izvedla s petimi skupinami. Rezultati so podani v tabeli 5. Cim večja je vrednost F, večji je vpliv spremenljivke na uvrstitev vzorca v določeno skupino. Na uvrstitev v posamezne skupine najbolj vplivajo spremenljivke CaO, K2O, Sr ter SiO2 in Al2O3, kar je pričakovano, saj so to oksidi in prvine, ki so vezani na karbonate in glinene minerale. Medtem ko količine NapO, Cr2O3 in P2O5 na uvrstitev v skupine ne vplivajo. Podobne rezultate dobimo s t-testom enakosti populacijskih povprečij, kjer so vrednosti SiO2, Al2O3, MgO in Sr tiste, ki so najbolj značilne za posamezne apnence. Rezultati obeh statističnih metod tako potrjujejo, da na obilnost prvine vpliva predvsem prisotnost minerala v kamnini oz. vsebnost "nečistoč" v kalcitni rešetki. Tabela S. Rezultati analize variance clusterske analize tipa k-mean za S skupin. Table S. k-mean cluster analysis results for S groups. varianca med skupinami znotraj skupine F Si02 528,0 2,7 697,2 A1203 120,0 0,7 618,9 Fe203 58,0 0,9 226,6 MgO 35,0 0,4 332,6 CaO 1143,0 3,0 1347,2 Na20 0,0 0,0 1,2 к2о 17,0 0,1 1144,7 Ti02 0,0 0,0 154,7 P205 0,0 0,0 7,7 MnO 0,0 0,0 15,3 Cr203 0,0 0,0 3,7 Ba 7520,0 44,5 591,5 Sr 1467763,0 7195,9 713,9 LOI 334,0 1,6 713,1 C 44,0 1,0 146,2 S 0,0 0,0 48,1 V prvo skupino so uvrščeni vsi vzorci iz Lipice (LiUl, LiU2, LiFI in LiF2), v drugo oba vzorca z Drenovega Griča (DG1 in DG2), sledi samostojen vzorec iz Lesnega Brda (LB1), četrta skupina je najbolj številčna in združuje en vzorec iz Lesnega Brda in vse vzorce apnenca iz Hotavelj (LB2, HI, Hlp, H2, H3, H4, H5, H6, H6p, H7), svojo skupino pa predstavljata tudi vzorca glinenih vložkov iz Hotavelj (HV1 in HV2). Rezultati clusterske analize se ujemajo z rezulati t-testa enakosti populacijskih povprečij, kjer sem skupine definirala sama. Izjema je le vzorec lesnobrdskega apnenca (LB2), ki ga je clusterska analiza uvrstila v skupino apnencev iz Hotavelj. Oba tipa apnenca (iz Hotavelj in z Lesnega Brda) sta genetsko podobna (Jarc, 2000) in torej glede na kemično sestavo podobna. Tudi rezultati t-testa enakosti populacijskih povprečij pokažejo najmanjšo statistično razliko med tema razredoma apnencev. Rezultati in razprava Število vzorcev je za statistično obdelavo majhno, kljub temu pa sem poskušala dobiti ocene količin oksidov in prvin, ki bi mi pomagale pri identifikaciji apnencev. Pri določanju provenience apnenca sem uporabila dve neodvisni statistični metodi: clustersko analizo in t-test enakosti populacijskih povprečij. Preiskovani slovenski apnenci, z izjemo vzorca glinenih primesi v hotaveljskem apnencu, so kemično relativno čisti CaC03. Z izjemo apnenca z Drenovega Griča (91 mas.% CaC03) vsebujejo preko 95 mas.% CaC03. Po sestavi so si zelo podobni, delež primesnih prvin je relativno majhen. Kljub majhnim razlikam v sestavi, pa apnence z natančno kemično analizo in statistično obdelavo podatkov med seboj lahko razlikujemo, predvsem po vsebnosti Si02, Al203, MgO in Sr. Korelacija med analiziranimi oksidi in prvinami pokaže, da je Ca0 vezan predvsem na mineral kalcit, ki je kemično relativno čist (negativna korelacija z ostalimi oksidi in prvinami). 0stale prvine so vezane predvsem na druge minerale, gre za minerale smektitove in kloritove skupine ter na muskovit (Jarc & Mirtič, v tisku}. S pomočjo t-testa enakosti populacijskih povprečij sem preverila uvrščanje vzorcev v "geokemične" skupine, glede na posamezna nahajališča: Lipica, Lesno Brdo, Hotavlje -apnenci, Hotavlje - glinene primesi, Drenov Grič. Rezultati t-testa so pokazali, da se skupine med seboj statistično razlikujejo glede na količine večine analiziranih oksidov in prvin na 95 % ravni zaupanja. Glede na izračunane vrednosti t za vse prvine v dani skupini se statistično najmanj razlikujeta apnenca z Lesnega Brda in iz Hotavelj, ki sta tudi genetsko podobna. Vzorec glinenih vključkov iz Hotavelj pa se, po pričakovanjih, popolnoma razlikuje od vseh ostalih. Apnenec z Drenovega Griča se od ostalih apnencev razlikuje predvsem po povečani količini organske snovi; ta mu daje tudi značilno, črno barvo. Tudi clustersko analizo sem izvedla s petimi skupinami. V tem primeru se na podlagi uporabljenega računalniškega programa vzorci razporedijo v skupine na osnovi podobnosti med njimi. Rezultati clusterske analize se dobro ujemajo z rezultati t-testa enakosti populacijskih povprečij. Apnenci se, kljub majhnim razlikam v kemični sestavi, statistično razlikujejo. Nahajališče apnenca bi torej lahko določali z detajlno kemično analizo. Sklepi Preiskovani apnenci so razmeroma čisti, vsebujejo preko 91 mas.% CaC03, izjema je le vzorec glinenih primesi iz Hotavelj, kjer vsebnost karbonata znatno nižja (približno 52 mas.% CaC03). Ca0 je z večino oksidov in prvin negativno koreliran, kar kaže na relativno čist kalcit. Dobra korelacija med Si02, Al203, Mg0, Fe203 in K20 torej kaže na prisotnost drugih mineralov, predvsem gre za glinene minerale. Rezultati t-testa enakosti populacijskih povprečij apnencev iz 4 nahajališč so pokazali, da se apnenci med seboj statistično razlikujejo na 95 % ravni zaupanja. Največje razlike so v vsebnosti Si02, Al203, Mg0 in Sr, torej tistih prvin, ki so vezane na prisotnost glinenih mineralov, sljud in čistost kalcita. Statistično se najmanj razlikujeta apnenca iz Hotavelj in Lesnega Brda, ki sta tudi genetsko podobna. Vzorec glinenih primesi iz Hotavelj pa se, po pričakovanjih, popolnoma razlikuje od vseh ostalih. Clusterska analiza je potrdila rezultate t-testa. Največja je podobnost med apnencema iz Hotavelj in Lesnega Brda. Kljub temu da so preiskovani apnenci zelo čisti (preko 95 mas.% CaC03, apnenec z Drenovega Griča pa vsebuje le 91 mas.% CaC03) in tudi po sestavi zelo podobni, bi jih z natančno kemično analizo lahko ločevali. V tem primeru gre le za pilotske vzorce, s pomočjo katerih sem poskušala samo pokazati, da se apnenci, kljub majhnim razlikam v kemični sestavi, statistično razlikujejo. Seveda bi bilo potrebno narediti bistveno večje število kemičnih analiz vseh apnencev iz različnih nahajališč, da bi lahko postavili statistične meje količin posameznih oksidov in prvin, na osnovi katerih bi ugotavljali provenienco apnenca. Summary The distinction of limestone by statistical methods The identification of limestone by two independent statistical analysis, t-test and cluster analysis was tested on some Slovenian samples from Lipica, Hotavlje, Lesno Brdo and Drenov Grič. The results of detailed chemical analysis are in Table 1. The accuracy and precision of the analytical technique are satisfactory (Jarc, 2000). The mineral composition of investigated samples is described in Jarc & Mirtič (in press). The composition of the investigated limestones is very similar. The limestone from Drenov Grič contains the lowest amount of CaC03 - above 91 wt.%, whereas the others contain above 95 wt.% of CaC03. The analysis of clay layer in Hotavlje limestone (sample HV) was made for the comparison. Mineral calcite is relatively pure, as seen in the negative correlations between Ca0 and other oxides (Table 3). Therefore, the other elements are due to minerals like smectite or chlorite and muscovite (Jarc & Mirtič, in press). The t-test is used in testing the difference between two population means based on differences found between sample means and considering variance and number of observations. The tested groups were determined accordingly to the limestone quarry: in first group are samples from Lipica quarry (LiUl, LiU2, LiFl and LiF2), in second group are samples from Lesno Brdo (LB1 and LB2), in third group is limestone from Hotavlje (Hl - H7), the claylayer from Hotavlje quarry (HV1 and HV2) represents the fourth group and in fifth group are samples from Drenov Grič (DGl and DG2). Two-way t test was calculated by programme CSS Statistica. The results are in Table 4. The defined groups are statistically different as to almost all analysed oxides and elements on 95 % confidence interval, but the greatest differences are in content of Si02, Al203, Mg0 and Sr. Therefore, t-test is very helpful in the identification of the limestone regardless of their very similar chemical compositions. Cluster analysis is a useful technique for grouping objects into unknown groups. The k-mean cluster analysis was used, where the characteristics of the groups are to be derived from the data on the basis of the smallest variance in the same group and the largest variance between the groups. For the cooperation with t-test the cluster analysis with five groups were performed. The largest influence on the classification of investigated samples (Table 5) have Si02, Al203, CaO, K20 and Sr, whereas the content of Na20, Cr203 and P20S have no influence on grouping. First group represent the samples from Lipica (LiUl, LiU2, LiFl and LiF2), in second group are samples from Drenov Grič (DGl and DG2), in third group is the sample LBl, fourth group represent samples from Hotavlje (Hl - H2) and sample LB2 Viri ACME (1999): ACME analytical laboratories brochure. Acme Analytical Laboratories, Ltd., Vancouver; 18 p. Jarc, S. (2000): Vrednotenje kemične in mineralne sestave apnencev kot naravnega kamna. Magistrsko delo. Ljubljana: Univerza v Ljubljani; 88 p. Jarc, S. & Mirtič, B. (v tisku): Vpliv mineralne sestave na obstojnost apnencev kot naravnega kamna. RMZ- Materials and Geoenvironment, Ljubljana. Koch, G.S. & Link, RE (1970): Statistical analysis of geological data. Willey & Sons, New York; 37S p. and in fifth group are both samples of clayish layer from Hotavlje (HVl and HV2). The results of cluster analysis are in good agreement with t-test analysis. The only exception is sample LB2, which was classified into Hotavlje limestone group. Nevertheless, the Hotavlje and Lesno Brdo limestones are of the same origin and very similar chemical composition, which explain the result of cluster analysis and the smallest difference in t-test. The origin of limestone can be determined by thorough chemical analysis and statistical analysis of the data. Limestones differ statistically significantly despite of their similar chemical composition, so the statistical methods could be useful in their identification. Petz, B. (198S): Osnovne statističke metode za nematematičare. SNL, Zagreb; 409 p. Swan , A.R.H. & Sandilands, M. (199S): Introduction to geological data analysis. Blackwell Science, Oxford; 446 p. Računalniški program CSS Statistica. Zupančič, N. (1994): Petrološke in geokemične značilnosti pohorskih magmatskih kamnin. Doktorska disertacija. Ljubljana: Univerza v Ljubljani; 178 p. Ustrezne analize črpalnih poizkusov v razpoklinskih vodonosnikih Appropriate analysis methods of pumping tests in fractured aquifers Timotej Verbovšek Univerza v Ljubljani, NTF, Oddelek za geologijo, Aškerčeva 12, 1000 Ljubljana, Slovenija; E-mail: timotejverbovsek@ntfgeo.uni-lj.si University of Ljubljana, Faculty of Natural Sciences and Engineering, Aškerčeva 12, Ljubljana, Slovenia; E-mail: timotejverbovsek@ntfgeo.uni-lj.si Received: June 8, 2005 Accepted: November 24, 200S Izvleček: Kljub dejstvu, da so nekatere metode črpalnih poizkusov v razpoklinskih vodonosnikih že uveljavljene, se v Sloveniji za analizo razpoklinskih kamnin še vedno uporabljajo neustrezne metode, razvite za medzrnske vodonosnike. Namen prispevka je opozoriti na uporabo ustreznih postopkov, podati kratek pregled uveljavljenih metod ter na nekaj primerih analizirati rezultate črpalnih poizkusov v dolomitih predvsem z modeli dvojne poroznosti. Abstract: Despite the fact that methods for analyzing pumping test data in fractured aquifers are at the present time developed to the practical level of use, in Slovenia they are still not used correctly. Even for the fractured aquifers the preferable analysis methods are Theis and Cooper-Jacob equations, which are applicable only to single-porosity homogenous aquifers. The purpose of this paper is to warn against erroneous applications of inappropriate methods and to give a brief summary of suitable methods. Finally few examples of pumping tests in dolomites are analyzed and interpreted with methods based on double porosity model. Ključne besede: črpalni poizkusi, vodonosnik, razpoklinska poroznost, model dvojne poroznosti, dolomit Key words: pumping tests, aquifer, fracture porosity, double porosity model, dolomite. Uvod Karbonatni vodonosniki pripadajo razpoklinskemu ali kraško-razpoklinskemu tipu. Tako pri nas kot drugod so ti vodonosniki čedalje bolj zanimivi zaradi izkoriščanja pitne vode, v tujini pa so pomembni tudi kot naftni rezervoarji. Trenutni svetovni trendi kažejo, da se raziskave razpoklinskih vodonosnikov šele začenjajo (Neuman, 2005). Najbolj zanesljive podatke o vodonosnikih, t.j. hidravlične parametre pridobimo s pomočjo črpalnih poizkusov. Tudi za kamnine z razpoklinsko poroznostjo se uporabljajo metode črpalnih poizkusov, razvite za medzrnske vodonosnike; predvsem Theisova in Cooper-Jacobova. Те metode za razpoklinske kamnine večinoma niso primerne, zato so z njimi določeni rezultati pogosto napačni in tudi nelogični. Namen prispevka je opozoriti na ustrezno uporabo metod v razpoklinskih vodonosnikih ter podati njihov pregled, prav tako pa na izbranih primerih podati obdelave in komentarje nekaterih metod za razpoklinske vodonosnike. Metode V strokovni literaturi je v primerjavi z metodami črpalnih poizkusov, razvitimi za kamnine z medzrnsko poroznostjo, metod za razpoklinske kamnine bistveno manj oz. jih nekateri celo v celoti izpuščajo (Batu , 1998). Pri prvih je namreč mogoče upoštevati popolnost vodnjakov, stacionarnost toka, anizotropnost, različno debelino vodonosnika, večplastni sistem, odprtost vodonosnikov itd. Za razpoklinske kamnine pa se uporablja precej manj metod, saj so te mlajše in še ne dovolj uveljavljene, predvsem zaradi: • težko določljivih lastnosti razpok (terensko kartiranje, geofizikalne raziskave), • zapletenih izračunov, ki zahtevajo precej parametrov, • uporabe posebne tehnologije pri črpalnih poizkusih (npr. tesnil oz. 'packerjev'), • dejstva, da so modeli dvojne poroznosti računalniško podprti šele malo časa. Številni računalniški programi za obdelavo črpalnih poizkusov podpirajo predvsem najenostavnejše metode za medzrnske vodonosnike, le redki pa tudi za razpok-linske, npr. Aquifer Test (Waterloo Hydrogeologic Inc., 2001), AQTESOLV (HydroSOLVE, Inc., 2003) in AquiferWin32 (Rumbaugh & Rumbaugh, 2003). V prispevku je poudarek na metodah v razpoklinskih vodonosnikih in ne tudi na kraško-razpoklinskih, ker ti zahtevajo poseben pristop. Kraški vodonosniki so izredno heterogeni, saj je bilo ugotovljeno, da se lahko transmisivnost blokov matriksa kamnine in razpok razlikuje tudi za faktor 100.000 (Krivic, 1983). Modeli razpoklinskih vodonosnikov Stanje v naravi opišemo z modelom, ki predstavlja poenostavitev realnega sistema. Za razpoklinske kamnine je razvitih precej modelov, v glavnem pa jih lahko ločimo na diskretne modele, na modele kontinuuma in multikontinuuma ter na kombinirane hibridne modele (Čenčur Curk, 2002). Pri diskretnih modelih moramo poznati geometrijo ter različne lastnosti razpok (položaj v prostoru, gostoto, povezanost, odprtost, hrapavost ipd.), razpoke pa nato obravnavamo deterministično, stohastično ali s fraktalnimi metodami. Pri diskretnih modelih je geometrijo razpok ponavadi zelo težko opisati, zato uporabljamo modele kontinuuma. Pri teh obravnavamo prostor kot ekvivalenten izotropen homogen prostor na makroskopskem nivoju tako, da upoštevamo povprečne vrednosti merjenih parametrov. Podoben pristop velja za modele multikontinuuma, kjer je prostor razdeljen na dva ali več prekrivajočih se homogenih podsistemov kontinuuma. Najbolj znan model iz te skupine je model dvojne poroznosti, in je zaradi široke uporabe natančneje opisan v nadaljevanju. Hibridni modeli združujejo modele kontinuuma ter diskretne modele. Model dvojne poroznosti Pri tem modelu upoštevamo koncept dvojne poroznosti (B Arenblatt et al., I960). Zanj je značilno, da prostor ločimo na dva prekrivajoča se dela, na razpoke in na vmesne bloke matriksa, tako da tok v njih obravnavamo ločeno. Geometrija razpok je odvisna od modela, v vsakem primeru pa je zelo poenostavljena. Največkrat se uporabljata tridimenzionalni model ortogonalnih razpok, katere razdelijo prostor v kocke enakih dimenzij (Warren & Root, 1963) ali pa model med seboj vzporednih horizontalnih razpok, ki razdelijo prostor v ploščate bloke (slab-shapedblocks). Redkeje se uporabljajo tudi modeli dveh sistemov razpok. Ti ločijo prostor na bloke v obliki navpičnih stolpcev (npr. pri modeliranju razpok v bazaltih; Aguilera, 1987). Za bloke matriksa je značilno, da imajo primarno poroznost ter veliko sposobnost vskladiščenja, tok v njih pa je majhen. Za razpoke velja nasprotno. Imajo namreč majhno sposobnost vskladiščenja, tok skozi vodonosnik pa teče večinoma po njih. Dodatni predpostavki sta, da je tok vode iz blokov matriksa možen le v razpoke, ter da je tok v vodnjak mogoč le iz razpok in ne iz matriksa zaradi bistveno večje prepustnosti razpok. Ločimo lahko dva režima toka. Pri psevdo-stacionarnem toku je prisoten tok iz matriksa v razpoke in se nivo gladine v matriksu ne spreminja. Pri nestacionarnem toku gladina v blokih matriksa ni stalna in se spreminja s časom. Kvantitativno opišemo tok v obeh sistemih z modificirano difuzijsko Slika I. Razpoklinska poroznost v realni kamnini (A) ter poenostavitev z modelom dvojne poroznosti s sistemom treh pravokotnih sistemov (B) in vzporednim sistemom razpok (C) Kruseman & deRidder, 1991) Figure I. Fractured rock formations. (A) A naturally fractured rock formation, (B) Warren-Root's idealized three-dimensional, orthogonal fracture system, (C) Idealized horizontal fracture system (Kruseman & deRidder, 1991) enačbo (MoENCH, 1984; Domenico & Schwartz, 1998): Tok v razpokah: KV2h = Ss^ + q (1) ot dh' Tok v matriksu: K*V2h = Ss,— -q (2) ot q predstavlja tok iz blokov matriksa v razpoke, ki ga za psevdostacionarne razmere opišemo kot: q = -oK\h'-h) (3) h'= prostorska povprečna vrednost hidravličnega nivoja v blokih matriksa a = geometrijski faktor [dolžina"2] K predstavlja koeficient prepustnosti sistema razpok (K = Kf • Vf), K' pa analogno koeficient prepustnosti sistema blokov matriksa (K' = Km • V J. (4) K = koeficient prepustnosti razpok V = razmerje med prostornino razpok ter prostornino celotnega volumna K = koeficient prepustnosti matriksa V = razmerje med prostornino matriksa ter prostornino celotnega volumna Podobno sta definirana koeficient specifičnega elastičnega vskladiščenja za sistem razpok Ss = S sf • Vf ter za sistem blokov matriksa S ' = S • V . (5) s sm m v ' Metode črpalnih poizkusov Pri analizi rezultatov črpalnih poizkusov v razpoklinskih kamninah moramo izbrati enega od naštetih modelov razpoklinskih vodonosnikov, nato pa upoštevati še lastnosti črpalnih in opazovalnih vodnjakov ter toka v vodonosniku (popolnost vodnjaka, število opazovalnih vodnjakov, režim toka, odprtost vodonosnika itd.). Metode glede na izbrane lastnosti ločimo v več skupin: 1. Modeli z eno razpoko (horizontalna ali vertikalna razpoka). Ti primeri so v naravi precej redki, izjeme so le umetno povzročene razpoke, ki nastanejo zaradi namernega povečanja izdatnosti vodnjaka s hidravličnim razpokanjem (hydraulic fracturing). Tok je sprva pravokoten na ploskev razpoke, pozneje pa se spremeni v psevdoradialnega. Med metodami obdelav črpalnih poizkusov so najbolj znane (Kruseman & de Ridder, 1991): • metoda Gringarten-Witherspoon (1972) za opazovalne vodnjake v homogenih izotropnih zaprtih vodonosnikih, • metoda Gringarten & Ramey (1974) za črpalni vodnjak v homogenih izotropnih zaprtih vodonosnikih, • metoda Ramey-Gringarten (1976) za črpalni vodnjak v homogenih izotropnih zaprtih vodonosnikih z upoštevanjem neplanarne razpoke s sposobnostjo uskladiščenja. 2. Modeli dvojne poroznosti. Med vsemi modeli za razpoklinske vodonosnike so te metode najbolj razširjene. Pri večini metod predpostavimo, da je vodonosnik izotropen, homogen, zaprt, enake debeline ter se razširja neskončno daleč. Bistvenega pomena je tudi, da je črpani pretok ves črpalni poizkus konstanten. Najbolj uveljavljene metode slonijo na naštetih predpostavkah, razlikujejo pa se predvsem v režimu toka in v številu opazovalnih vodnjakov: • Warren-Root (1963): psevdostacio-naren tok (nivo gladine v matriksu se ne spreminja), velja za črpalni vodnjak, • B ourdet & G ringarten (1980): psevdostacionaren tok iz matriksa v razpoke, velja za opazovalne vodnjake, • Kazemi et al. (1969): nestacionaren tok (nivo gladine se v matriksu spreminja), velja za opazovalne vodnjake, • Boulton & Streltsova (1977): črpalni vodnjak, vodonosnik razdeljen na porozne horizontalne bloke, ločene z razpokami, • Moench (1984): psevdostacionaren in nestacionaren tok, uvedba koncepta tanke mineralne plasti. Prve tri metode so si med seboj dokaj podobne, saj za vse veljajo naslednje matematične predpostavke. Znižanje v vodnjakih opišemo z enačbo, analogno Theisovi (Kruseman & de Ridder, 1991): s — Q 4nT F(u*,X,(o) / (6) Theisovo fUnkcijo vodnjaka torej nadomešča fUnkcija F(u*, X, ю), ki je odvisna od treh parametrov. Prvi, u* predstavlja modificirano vrednost Theisovega parametra u in je podan kot ter vskladiščenjem v celotnem sistemu (oba parametra sta brez dimenzije): 2 K. Ke со = sf + psm (8) (9) a = geometrijski faktor, odvisen od odnosa med razpokami in matriksom [površina"1] Tipičen primer odziva znižanja v vodnjaku oz. v piezometrih v odvisnosti od logaritma časa je prikazan na sliki 2. Krivuljo lahko v idealnem primeru ločimo na tri dele: • Pri zgodnjih časih črpanja (začetni del krivulje) je odvisnost s - log t linearna, saj je ß=0 in takrat se enačba poenostavi v Theisovo enačbo, za katero lahko uporabimo Cooper-Jacobovo poenostavitev. Tok v vodnjak prihaja tedaj le iz razpok: 2,3Q, 2,257}* s = ——log-— 4nT, f Sfr2 (10) #* -. Tft 0sv + ßs>2 (7) T2 predstavlja transmisivnost razpok, Sf in Sm koeficienta elastičnega vskladiščenja v razpokah ter v blokih matriksa in ß koeficient, ki je pri zgodnjih časih črpanja enak 0, pri poznih pa enak 1/3 (ortogonalni bloki) ali 1 (plasti). Dodatna parametra, s katerima opišemo tok v vodonosniku z dvojno poroznostjo, sta koeficient interporoznega toka X, odvisen od oblike, velikosti in prepustnosti blokov matriksa, ter ю, razmerje med vskladiščenjem v razpokah Po določenem času črpanja (srednji del krivulje) se ustvari prehodno obdobje, ko začne zaradi ustvarjenega gradienta voda teči iz blokov matriksa v razpoke. Znižanje poteka tedaj počasneje, kar se odraža v manjšem naklonu krivulje oz. tudi horizontalnem poteku pri večjih prepustnostih matriksa. V zadnjem delu krivulje pri poznih časih črpanja priteka voda tako iz matriksa kot tudi iz razpok, krivulja se podobno poenostavi v Theisovo, le da je faktor ß=1/3 oz. 1 in znižanje enako 2,3Q . 2,25Tft - (11) s = ■ log 4nTf "(Sf + W5my Moenchova metoda (1984) predstavlja nadgradnjo metod Warren-Root (1963) za psevdostacionaren tok in Kazemi et al. (1969) za nestacionaren tok. Z uvedbo koncepta tanke plasti na površini razpok (angl. fracture skin) je Moench razložil, zakaj prihaja tako do psevdostacionarnega kot tudi nestacionarnega toka. Ta plast predstavlja delno prepusten material, ki zavira tok iz matriksa v razpoke. Ce je plast zelo slabo prepustna, je največji gradient hidravličnega nivoja med matriksom in razpokami prisoten na površini plasti in nestacionarni tok se poenostavi v psevdostacionarnega. Vsi računalniški programi, ki podpirajo obdelavo črpalnih poizkusov v vodonosnikih z dvojno poroznostjo, podpirajo le Moenchovo metodo, saj ta združuje koncepte vseh prej omenjenih metod. Za računalniško reševanje metode je potrebno imeti podatke o opazovalnih vodnjakih. 3. Diskretni oz. stohasticni modeli. Za analizo črpalnih poizkusov se precej manj kot zgoraj opisani modeli uporabljajo diskretni modeli, ki temeljijo na modeliranju realne geometrije razpok z vsemi izmerjenimi razpokami (naklon, vpad, odprtost, raztezanje itd.). Hidravlične lastnosti razpoklinskega sistema se nato izračunajo s statističnimi oz. stohastičnimi metodami (National Research Council, 1996). Metode temeljijo na povezavi posameznih razpok v modele mreže razpok (DFN, Discrete Fracture Network) in prav tako kot fTaktalne še niso širše uveljavljene. 4. Fraktalni modeli. Predvsem v zadnjih letih so za analizo črpalnih poizkusov v razpoklinskih kamninah pričeli razvijati fraktalne metode (Acuna & Yortsos, 1995; Hamm & Bidaux, 1996; Leveinen, 2000). Metodam je skupna predpostavka, da se nek pojav pojavlja v enaki obliki v različnih merilih. Tako lahko npr. na bloku kamnin določen vzorec razpok pričakujemo v regionalnem merilu. Večina teh metod je razvitih za nestacionaren tok, rešitve pa podobno kot pri Theisovi in ostalih metodah omogoči prilagajanje podatkov tipskim krivuljam. Pri fTaktalnih metodah je lahko dimenzija toka tudi realno število n, ki ima vrednosti med 1 (enodimenzijski tok), 2 (cilindrični dvodimentionalni tok) ali 3 (radialni sferni tok v prostoru). Splošno difuzijsko enačbo, ki opisuje nestacionaren tok za te primere, je določil Barker (1988) v svojem Generalized Radial Flow (GRF) modelu: Slika 2. Odvisnost s - log t pri modelu dvojne poroznosti Kruseman & deRidder, 1991) Figure 2. Semi-log time-drawdown plot in a fractured rock formation of the double porosity type (Kruseman & deRidder, 1991) кf д n-\ dr n-idh dr c dh (12) Podatki iz literature kažejo, da taki primeri v naravi obstajajo, saj so rezultate črpalnih poizkusov v različnih kamninah prilagodili krivuljam z različno dimenzijo n, npr. n=l,8 v granitih, l,7 v gabrih (Lods & Gouze, 2004). Za n=2 se Barkerjeve enačbe poenostavijo v znani formuli Theisa in Cooper-Jacoba za nestacionaren tok ter v Thiemovo formulo za stacionaren tok. Rezultati in razprava V nadaljevanju je podanih nekaj značilnih primerov rezultatov in analiz črpalnih poizkusov v dolomitih. Podobne grafe bi lahko analizirali tudi v nekaterih drugih kamninah, za katere veljajo predpostavke modela dvojne poroznosti (npr. razpokani peščenjaki, razpokane magmatske kamnine itd.). Za dolomite je značilno, da spadajo med zelo razpokane kamnine, saj so razpoke v njih precej bolj pogoste kot v peščenjakih in apnencih (Aguilera, 1980). To potrjujejo tudi opazovanja dolomitnih plasti v Sloveniji. Poroznost matriksa lahko v dolomitih niha v precej širokem razponu, od skoraj nič do nekaj odstotkov in pripada večinoma medkristalni poroznosti, nastali pri dolomitizaciji, redkeje pa poroznosti, nastali z raztapljanjem (vugs), ter fenestralni poroznosti, značilni za medplimske dolomite (Moore, 2001). Na sliki 3 je prikazan graf znižanja v odvisnosti od časa pri črpalnem poizkusu v dolomitu cordevolske starosti pri konstantnem pretoku Q = 15 l/s (Verbovšek, 2003). Na njem je prikazana tridelna krivulja, značilna za model dvojne poroznosti. V začetnem delu krivulje (prvih nekaj minut) je opaznih nekaj odstopanj od premice, nato sledi položni del in kasneje (od 1000 minut dalje) spet počasno naraščanje znižanja. Podatki veljajo za črpalni vodnjak, zato je bila za analizo primerna obdelava po metodi Warren-Roota (1963), ki je dala naslednje rezultate: Tf = 4,43-10"3 m2/s, Sf = 1,85-10"2 t (min) Slika 3. Diagram znižanja s v odvisnosti od časa (črpalni poizkus v dolomitu cordevolske starosti, Q=1S l/s) (Verbovšek, 2003) Figure 3. Semi-log plot for pumping test data in dolomite (Cordevolian age, discharge Q = IS l/s) (Verbovšek, 2003). S = 0,34 in koeficienta X = 1,35-10"6 ter m 7 7 ю = 0,05. Primerljive vrednosti smo dobili tudi pri analizi Moenchovih podatkov za črpalni vodnjak po metodi Warren-Roota (Kruseman & de Ridder, 1991). Trans-misivnosti blokov matriksa po tej metodi ni mogoče določiti. Obnašanje po modelu dvojne poroznosti je posledica dejstva, da ima dolomit cordevolske starosti veliko medkristalno poroznost, saj je nastal s pozno diagenezo. Na drugem diagramu (sl. 4) je vidno znižanje v dolomitu zgornjetriasne starosti pri črpanju Q = 0,1 l/s (Verbovšek, 2003). Črpanje je trajalo bistveno manj časa kot pri prejšnjem primeru, zato iz oblike krivulje ni jasno, kakšen bi bil nadaljni potek podatkov na grafu. Mogoče je namreč, da so meritve zajele le začetni del krivulje, ko je prisoten tok le iz razpok, saj leži večina podatkov na linearnem delu. To potrjuje nekaj zadnjih meritev, kjer se znižanje počasi izravnava v bolj položno krivuljo, ki je tipična za model dvojne poroznosti. Druga možnost pa je, da ima dolomit zelo majhno poroznost matriksa in sta tako tok kot vskladiščenje prisotna le v razpokah in ne tudi v matriksu. V tem primeru velja izračunani koeficient elastičnega vskladiščenja S = 0,31 tako za bloke matriksa kot tudi za razpoke (Sf + ßSm), zadnjih nekaj meritev pa lahko predstavlja vpliv slabše prepustne bariere ali pa manjše merske napake. Za analizo linearnega dela je lahko tedaj primerna Cooper-Jacobova metoda, kateri izračunani podatek T = 6,54-10"6 m P/s predstavlja transmisivnost razpok Tf in ne transmisivnosti kamnine z medzrnsko poroznostjo Tm. Na sliki 5 (Verbovšek, 2003) je prikazan primer črpalnega poizkusa, kjer modela dvojne poroznosti žal zaradi prevelikih nihanj pretoka ne moremo direktno uporabiti na s-t diagramu. Zato moramo v ta namen uporabiti za pretok normirani s/Q-t diagram, na kar se često pozablja. Takih situacij je precej, saj se pogosto uporabljajo t.i. step testi, s katerimi t (min) Slika 4. Diagram znižanja s v odvisnosti od časa (črpalni poizkus v dolomitu zgornjetriasne starosti, Q=0,1 l/s) Figure 4. Semi-log plot for pumping test data in dolomite (Norian-Rhaetian age, discharge Q = 0.1 l/s) (Verbovšek, 200з) Slika S. Diagram znižanja s v odvisnosti od časa (črpalni poizkus v "glavnem dolomitu" norijsko-retijske starosti), variabilen pretok (Verbovšek, 2003) Figure S. Semi-log plot for pumping test data in dolomite (upper Triassic age), variable discharge (Verbovšek, 2003) določujejo učinkovitosti delovanja vodnjakov. Kadar sta v takem obravnavanem sistemu prisotni obe vrsti poroznosti, so efekti dvojne poroznosti zaradi nihanja pretokov zabrisani. Zaključki Temeljna razlika med analizami črpalnih poizkusov v medzrnskih in v razpoklinskih vodonosnikih je v izračunu hidravličnih parametrov, saj dobimo v prvem primeru podatka o transmisivnosti (T) in specifičnem elastičnem vskladiščenju vodonosnika (Ss), v drugem pa ločene rezultate transmisivnosti (Tf) in specifičnega elastičnega vskladiščenja razpok (Ssf) ter transmisivnosti (Tm) in specifičnega elastičnega vskladiščenja blokov matriksa (S ). Razlika je opazna tudi na grafu odvisnosti znižanja od časa (s - log t), kjer je za medzrnske kamnine odvisnost linearna, za razpoklinske pa se kaže v obliki tridelne krivulje v obliki črke S (sl. 2 in sl. 3). Za prvi del krivulje je značilen tok le iz razpok (Tf >> Tm), nato pride do upočasnitve naraščanja znižanja zaradi toka iz matriksa v razpoke, na koncu pa se blažilni učinek dotoka vode iz matriksa izgubi. V začetnem delu so zaradi kratkih merskih časovnih intervalov na krivulji mogoče manjše napake pri meritvah, prav tako pa na ta del krivulje vplivajo efekti vskladiščenja vode v vodnjaku ter "kožni" oz. skin efekti, zato premici (sl. 2) skoraj nikoli nista vzporedni. Tridelna krivulja je lahko zabrisana tudi, kadar je parameter X izredno majhen, saj je tedaj transmisivnost blokov matriksa zanemarljiva in tok je v celotnem črpalnem poizkusu prisoten le v razpokah. Kamnina se tedaj obnaša podobno kot homogen vodonosnik z medzrnsko poroznostjo. Analiza izbranih črpalnih poizkusov v dolomitnih vodonosnikih kaže, da nekatere krivulje sledijo modelu dvojne poroznosti in jih je zato potrebno tako tudi razlagati. Rezultati samodejnega računalniškega prilagajanja rezultatov krivulj za medzrnsko poroznost skozi vse točke na sliki 3 so nelogični, dobljene vrednosti pa so napačne, saj naj bi parametre v razpokah in v matriksu obravnavali ločeno (Rumbaugh and Rumbaugh, 2003). Nekateri črpalni poizkusi v dolomitih na grafu ä - log t vseeno ne prikazujejo omenjene krivulje v treh delih, temveč le v linearni odvisnosti, podobno kot pri medzrnskih kamninah (sl. 4). Vzrokov je več. Če je črpanje prekratko, pridobimo podatke le za prvi del krivulje, saj tedaj še ne prihaja do izcejanja vode iz matriksa v razpoke. Črpalni poizkusi bi v takih primerih torej morali biti daljši (primer na sliki 2). Če so razpoke drobne in je kamnina z njimi gosto prežeta ter če zanjo veljajo pogoji laminarnega toka oz. Darcyjev zakon, potem lahko za razpoklinske vodonosnike uporabimo tudi metode, razvite za medzrnske vodonosnike (Vasvari & Kriegl, 2003). Model dvojne poroznosti tedaj namreč izgubi pomen in se poenostavi v model z enojno poroznostjo, za katerega veljajo metode Theisa in Cooper-Jacoba. Seveda moramo upoštevati, da predstavlja izračunani podatek T transmisivnost razpok T у hkrati pa podatek o elastičnem koeficientu vskladiščenja S velja za razpoke (S). Tretja razlaga za linearen potek krivulje je mogoča pri analizi podatkov iz opazovalnih vodnjakov. Tridelna krivulja je namreč značilna le za bližnjo okolico črpalnega vodnjaka, saj je na določeni oddaljenosti od vodnjaka prisoten le združen tok iz matriksa in iz razpok (Kazemi et al., 1969). Če je iz podatkov črpalnega poizkusa razvidno, da ima vodonosnik dvojno poroznost, lahko zaradi metode samodejnega računalniškega prilagajanja krivulj za medzrnsko poroznost nastanejo temeljne napake. Namesto realnih vrednosti hidravličnih parametrov dobimo prevelike ali premajhne vrednosti, kar je odvisno od načina samodejnega prilagajanja. Tako lahko npr. izračunamo večjo transmisivnost ali specifično elastično vskladiščenje vodo-nosnika, kot je dejanska, s tem pa precenimo sposobnost izkoriščanja vode iz vodo-nosnika. Napačni so lahko tudi izračuni hitrosti pretakanja vode v vodonosnikih, dejansko širjenje polutantov je lahko npr. mnogo hitrejše, kot smo izračunali, ker se polutanti v razpokah in v matriksu gibljejo s precej različno hitrostjo. Čeprav so obdelave črpalnih poizkusov v razpoklinskih vodonosnikih bolj zapletene in manj razširjene kot metode za medzrnske kamnine, jih je vsekakor potrebno uporabljati, ko analiziramo črpalne poizkuse v razpoklinskih kamninah. Ker moramo včasih potrebne parametre določiti že med črpanjem, je za pravilne rezultate ključnega pomena predhodno poznavanje ustreznih metod. Zahvale Hvala izr. prof. dr. M. Veseliču za kritične pripombe pri izdelavi prispevka. Summary Appropriate analysis methods of pumping tests in fractured aquifers Carbonate aquifers belong to fractured or karstic-fractured type. Although the appropriate methods for analyzing pumping test data are those developed for fractured aquifers, unsuitable methods like Theis or Cooper-Jacob are still currently used for analysis in Slovenia. Use of these methods leads to erroneous calculations of hydraulic parameters. However, even in some recent aquifer test data books the methods for fractured reservoirs are completely omitted (Batu , 1998). The reason for their absence is probably the fact that they are still "being developed" and are not widely in use, mostly because of complex mathematical theory used for analysis, hard-to-obtain fracture properties and the need for special drilling equipment. Modeling of fractured aquifers can be divided in four categories: dicrete models, continuum, multicontinuum and hybrid models. Most aquifer test analysis methods are based on the double porosity model (Barenblatt et al., I960), which belongs to the multicontinuum type. Characteristic for this model is that the heterogeonous space is divided into two overlapping media: the fractures and the matrix blocks (Fig. 1), both of them having their own characteristics. The fractures have high permeability and low storage capacity and the matrix blocks the opposite, low permeability and high storage capacity. There are several aquifer test data analysis methods, which fall into different categories: • Single vertical or horizontal fractures: methods of Gringarten-Witherspoon (1972), Gringarten & Ramey (1974) and Ramey-Gringarten (1976). • Double porosity models: Warren-Root (1963), Bourdet & Gringarten (1980), Kazemi et al. (1969), Boulton & Streltsova (1977) and Moench (1984). • Discrete and stochastic models: different methods, based on Discrete Fracture Network (DFN) models (National Research Council, 1996). • Fractal analysis: most methods are based on Barker's (1988) Generalized Radial Flow (GRF) model, which allows non-integer values for flow dimension. Most methods, used for analysis of dolomite and other fractured-type aquifers, belong to double porosity models. Drawdown in these systems can be divided in three time periods (Fig. 2): early pumping time, when all the flow comes from the storage in the fractures, medium pumping time, which represents transition period during which the matrix blocks feed the water to the fractures and late pumping times, when the water comes both from the storage in fractures and matrix blocks (Kruseman & de Ridder, 1991). On Figure 3 is shown the semi-log plot for aquifer test data of dolomite of cordevolian age (Verbovšek, 2003). The curve follows the ideal double porosity type curve of early, middle and late pumping times. The parallel lines however do not occur, because the early time linear relationship is hidden by well-storage and skin effects. Data have been analyzed by Warren-Root (1963) method, which supports single pumping well conditions with no observation wells and it showed the following results: T = 4.43-10"3 m2/s, = 1.85-10"2, 5 = 0.34, f ' f ' rn ' X = 1.35-10"6 and Ю = 0.05, which are comparable to those of Moench (Kruseman & De Ridder, 1991). The double porosity behaviour can be explained by large intercrystalline porosity of dolomite, which was developed by burial diagenesis. On the next figure (Fig. 4; Verbovšek, 2003) is shown the plot which does not show the typical double porosity curve. In comparison with previous data the pumping times are much shorter, which can result in "missing" transition and late pumping times parts of the curve. Pumping test should be longer in this case. Another explanation is that dolomite has very low matrix porosity, which simplifies the double porosity model into single-porosity one, having only fracture porosity. In this case the usual Theis or Cooper-Jacob methods can be used. The last plot (Fig. 5; Verbovšek, 2003) shows the data obtained by using variable discharge. Drawdown significantly deviates from the ideal three-part curve and the double porosity methods can generally not be used due to the too short later steps. Viri Acuna, J. A. & Yortsos, Y. C. (1995): Application of fractal geometry to the study of networks of fractures and their pressure transient.- Water Resources Research Vol. 31, No. 3,: pp. 527-540. Aguilera, R. (1980): Naturally Fractured Reservoirs.-Pennwell Books, 703 p. Aguilera, R. (1987): Well Test Analysis of Naturally Fractured Reservoirs.- SPE Pormation Evaluation, Society of Petroleum Engineers. Barenblatt, G.I., Zheltov, IU.P., Kochina, I.N. (I960): Basic concepts in the theory of seepage of homogenous liquids in fissured rocks. - Journal of Applied Mathematics and Mechanics 24 (5): pp. 1286-1303. Barker, J. A. (1988): A Generalized Radial Flow Model for Hydraulic Tests in Fractured Rock.-Water Resources Research.: 24, No. 10, pp. 1796-1804. Batu, V. (1998): Aquifer Hydraulics. A comprehensive guide to hydrogeologic data analysis.- John Wiley 8 Sons Inc., 727 p. The main difference between the single porosity and double porosity models is that the last ones give the results of transmissivities and storativities of both fractures and matrix. We can also detect the double porosity behaviour on the semi-log plot of drawdown versus time, which shows two lines (usually not paralel due to well-storage and skin effects) with transition time in between (Fig. 2). Use of automatic computer-fitting linear curve to all the data in Figure 2 or Figure 3 would lead to erroneous and illogical results, as there are actually two transmissivities and storativities of the system (fracture and matrix). One can therefore calculate incorrect transmisivities and storativities, and this could lead to misinterpreted flow and pollutant velocities. The proper use of fractured aquifer methods is therefore of utmost importance. Boulton, N.S. & Strelstova, T.D. (1977): Unsteady flow to a pumped well in a fissured water-bearing formation.- J. Hydrol.: 30, pp. 29-46. Bourdet, D. & Gringarten, A. C. (1980): Determination of fissure volume and block size reservoirs by type-curve analysis.- Paper SPE 9293, Presented at 1980 SPE Annual Pall Techn. Conf andExhib., Dallas Cenčur Curk, B. (2002): Tok in prenos snovi v kamnini s kraško in razpoklinsko poroznostjo: Doktorska disertacija. Univerza v Ljubljani, NTF, 253 p. Domenico, P. A. and Schwartz, F. W. (1998): Physical and Chemical Hydrogeology, 2nd ed.- Wiley, 528 p. Gringarten, A. C. & Witherspoon, P. A. (1972): A method of analyzing pump test data from fractured aquifers.- Int. Soc. Rock Mechanics and Int. Ass. Eng. Geol., Proc. Symp. Rock Mechanics, Stuttgart, Vol. 3-B, pp.1-9. Hamm, S-Y. & Bidaux, P. (1996): Dual-porosity fractal models for transient flow analysis in fissured rocks.- Water Resources Research. 32, No. 9, pp. 2733-2745. National Research Council (1996): Rock fractures and fluid flow. Contemporary Understanding and Applications.- National Academy Press, SSI p. Neuman, S.P. (2005): Trends, prospects and challenges in quantifying flow and transport through fractured rocks.- Hydrogeology Journal. 13, 124-147, DOI: 10.1007/s10040-004-0397-2 Rumbaugh, D. & Rumbaugh, J. (2003): Program AquiferWin32, Version 3.00.- Environmental Simulations, Inc. Vasväm, V. & Kmegl, C. (2003): Determination of hydraulic properties in fractured aquifers in Austria.- V: Krasny, J., Zbynek, H., Bruthans, J.: International Conference on Groundwater in Practured Rocks, pp. 109-110, Prague. Verbovšek, T. (2003): Izdatnost vodnjakov in vrtin v Sloveniji - skupina dolomitnih vodonosnikov.-Diplomsko delo, Univerza v Ljubljani, NTF, 206 p. Warren, J.E. & Root, P.J. (1963): The Behavior of Naturally Fractured Reservoirs.- Soc. of Petrol. Engrs. J. 3, pp. 24S-2SS. Waterloo hydrogeologic inc. (2001): WHI AquiferTest, Version 3.01, Waterloo Hydrogeologic Inc., Canada (Interna pomoč programa), http://www.waterloohydrogeologic.com. Amsterdam. HYDROSOLVE INC. (2005): AQTESOLV for Windows. Version 3.5. http://www.aqtesolv.com. Kazemi, H., Seth, M.S., Thomas, G.W. (1969): The Interpretation of Interference Tests in Naturally Fractured Reservoirs with Uniform Fracture Distribution.- Soc. of Petrol. Engrs. J., pp. 463-472. Krivic, P. (1983): Studija hidrodinamike kraškega vodonosnika (Slovenski povzetek).- Geologija.: 26, pp. 149-186. Kruseman, G. P. & De Ridder, N. A. (1991): Analysis and Evaluation of Pumping Test Data, 2nd ed. -International Institute for Land Reclamation and Improvement, Wageningen, (ILRI Publication 47), 377 p. Leveinen, J. (2000): Composite model with fractional flow dimensions for well test analysis in fractured rocks.- Journal of Hydrology. 234, pp. 116-141. Lods, G. & Gouze, P. (2004): WTFM, software for well test analysis in fractured media combining fractional flow with double porosity and leakance approaches. Computers & Geo-sciences: 30, pp. 937-947. Moench, A. F. (1984): Double porosity Models for a Fissured Groundwater Reservoir With Fracture Skin.- Water Resources Research. 20, No. 7, pp. 831-846. Moore, C. H. (2001): Carbonate Reservoirs, Porosity Evolution and Diagenesis in a Sequence Strati-graphic Framework, Elsevier Science B. V., Modeliranje napajanja vodonosnika v zaledju izvira Rižane z območja Brkinov Modelling the recharge of the aquifer in the Rižana catchment from Brkini area. Mitja Janža Geološki zavod Slovenije, Dimičeva 14, 1000 Ljubljana; Slovenija, E-mail: mitja.janza@geo-zs.si Received: June 22, 200S Accepted: November 24, 200S Izvleček: V članku je opisan dinamičen hidrološki model z distribuiranimi parametri (MIKE SHE-MIKE II), izdelan na območju Brkinov - delu napajalnega območja vodonosnika v zaledju izvira Rižane. Rezultat modela je prostorsko in časovno porazdeljeni simulirani odtok z območja Brkinov, ki posredno napaja vodonosnik. Značilnosti odtoka, določene na podlagi simuliranih dnevnih odtokov za enajstletno obdobje (od 1.1.1983 do 31.12. 1993), so opisane s statistikami: povprečni odtok 0,970 mQ/s, standardni odklon odtokov 2,327 m3/s, minimalni odtok 0,012 mQ/s, maksimalni odtok 30,S34 m3/s, mediana 0,288 mQ/s, petindvajseti percentil 0,129 m3/s, petinsedemdesetini percentil 0,S31 mQ/s, cenilka asimetričnosti S,230 in cenilka sploščenosti 36,943. Abstract: In this paper a dynamic distributed hydrological model (MIKE SHE-MIKE 11) that was developed in the area of Brkini (part of the recharge area of the aquifer of Rižana spring) is described. The result of the model is spatially distributed and temporally variable simulated outflow from the Brkini area that indirectly recharges the aquifer. Characteristics of the outflow, defined on the daily simulated outflows for eleven years period (from 1. 1. 1983 to 31. 12. 1993), are described by statistics: average outflow 0.970 m3/s, standard deviation 2.327 mQ/s, minimum outflow 0.012 m3/s, maximum outflow 30.S34 mQ/s, median 0.288 m3/s, 2S percentile 0.129 mQ/s, 7S percentile 0.S31 m3/s, skewness S.230 and kurtosis 36.943. Ključne besede: napajanje vodonosnika, hidrološki model, MIKE SHE, Brkini, izvir Rižane, Slovenija. Key words: aquifer recharge, hydrological model, MIKE SHE, Brkini, Rižana spring, Slovenia. Uvod Izvir reke Rižane je najpomembnejši vir pitne vode na območju slovenske Obale. Pojavlja se na kontaktu med vodonosnimi paleogenskimi apnenci in slabo prepustnimi eocenskimi flišnimi plastmi. Po svojih značilnostih je tipični kraški izvir. Zaledje izvira je del obsežnega kraškega sistema, ki se začenja na zahodu s Krasom in na vzhodu končuje v Kvarnerskem zalivu. Sistem izvira Rižane tvori osrednji del, od koder odtekajo podzemne vode proti Tržaškemu zalivu, Kvarnerju, proti jugu pa izvirajo kot Rižana ali pa obnavljajo vodonosne plasti apnencev pod flišnim pokrovom proti Dragonji (Prestor, 1992). Vodonosnik v zaledju izvira Rižane neposredno napajajo padavine, posredno pa voda potokov, ki pritečejo z območja Brkinov in poniknejo na stiku fliša s karbonatnimi kamninami. Dotok ponikajočih voda potokov v izvir Rižane dokazujejo sledilni poskusi (Krivic et al., 1987; Krivic et al., 1989; Novak, 1963). Zaradi hitrega dospetja sledila iz območja ponikanja do izvira (štiri do šest dni) je to območje izjemnega pomena za varovanje vodnega vira in je uvrščeno v ožje vodovarstveno območje. od 1. 11. 2001 do 31. 3. 2003. Prednost uporabe modela je možnost simulacije prostorske in časovne spremenljivosti napajanja. S pomočjo modela pridobljene nove informacije so bile kasneje vključene kot robni pogoj v model vodonosnika v zaledju izvira Rižane (Janža, 2003), s katerim so bili simulirani pretoki izvira v obdobju od 1. 1. 1983 do 31. 12. 1993. Obravnavano območje V članku je opisana ocena količine posrednega napajanja vodonosnika v zaledju izvira Rižane z območja Brkinov. Ocena temelji na hidrološkem modelu (MIKE SHE-MIKE 11), ki je bil izdelan na osnovi novo pridobljenih podatkov o pretokih v obdobju Obravnavano območje Brkinov obsega severovzhodni del z varstvenimi pasovi varovanega zaledja izvira Rižane (sl. 1). Sestavljeno je iz osmih povodij s skupno površino 44,39 km2 (sl. 2). Gradijo ga flišne kamnine in se po hidrogeoloških lastnostih Slika I. Obravnavano območje. Figure I. Study area. Tabela I. Osnovne reliefne značilnosti povodij. Table I. Basic relief characteristics of the catchments. Oznaka povodja P1 P2 P3 P4 P5 P6 P7 P8 Površina [km2] 6,52 5,90 2,28 1,46 3,33 6,84 7,88 10,18 Naklon Povprečni [°] 13,3 14,5 10,9 11,1 11,4 12,1 12,3 10,5 Standardni odklon [°] 6,5 6,6 4,6 4,8 5,2 5,4 5,5 5,0 Nadmorska višina Povprečna [m] 632 632 632 618 604 604 612 582 Minimalna [m] 497 475 547 543 524 501 499 492 Maksimalna [m] 811 808 745 741 747 760 763 764 Standardni odklon [m] 78 81 49 52 53 58 63 57 Slika 2. Digitalni model višin z območji povodij in mestom meritve M2. Figure 2. Digital elevation model with catchments and measurement location M2. bistveno razlikuje od pretežno karbonatnega dela zaledja. Večina vode odteče z območja Brkinov površinsko in podpovršinsko, z značilnim hitrim povečanjem odtokov po deževju. Skupna značilnost potokov je, da pritekajo iz slabo prepustnih flišnih kamnin in ponikajo na delu povodja, kjer je matična podlaga apnenec. Dno dolin na tem delu je pokrito z aluvialnimi sedimenti. Potoki se večinoma končajo s ponori. Voda potokov jih doseže le ob visokem vodostaju, drugače ponikne že v strugi. Osnovne reliefne značilnosti povodij, ki so bile izdelane z analizo digitalnega modela višin - DMV (ZRC SAZU & Mobitel, d. d., 2000), so prikazane v tabeli I. Zaradi različnih poimenovanj potokov na obravnavanem območju so povodja označena z oznakami od PI do P8 (od severozahoda do jugovzhoda). Natančneje je te slepe doline morfološko analiziral Mihevc (1991). Hidrološki model obravnavanega območja Uporabljeni podatki Izdelava modelaje zahtevala številne vhodne podatke. V nadaljevanju so na kratko opisani najpomembnejši med njimi. Natančneje je uporabljene podatke opisal Janža (2003). Meritve pretokov Za namene modeliranja so bile opravljene meritve pretokov na izbranih mestih potokov. Meritve so bile občasne, razen na merskem mestu M2, kjer je bil nameščen sistem za zvezno meritev pretoka (sl. 2). Za zvezno meritev pretoka je bilo izbrano mesto, ki po svojih značilnostih odražala lastnosti vseh povodij. Meritev je bila opravljena v umetnem kanalu, ki služi kot prepust potoka pod gozdno cesto. S tem se je izognilo vplivu spremembe merskega preseka, kar je pogosto težava pri meritvah pretoka v naravnih koritih. S konstrukcijo ob straneh kanala je bil zmanjšan njegov presek na iztoku in preoblikovan v trapezoidno obliko, ki omogoča natančnejšo meritev nizkih pretokov. Pretok je bil ocenjen posredno preko nivoja vode v kanalu, ki je bil merjen v petnajstminutnih intervalih s tlačno sondo z natančnostjo 6,2 mm. Sonda je bila postavljena na dno kanala. Privzeto je bilo, da pretoka ni, ko se gladina vode v kanalu zniža do nivoja sonde. Za določitev odnosa med pretokom in višino vode so bile uporabljene meritve pretoka s kemijsko integracijsko metodo in takrat izmerjeni nivoji. Na podlagi teh podatkov je Slika 3. Pretočna krivulja. Figure 3. Rating curve. bila določena pretočna krivulja (sl. 3), ki ima obliko fUnkcije: q = gh", kjer so: (I) q pretok; h višina gladine vode v kanalu; g, u umeritvena koeficienta. Natančnost meritev, uporabljenih za določitev pretočne krivulje je prikazana na slikah 4 in 5, ki prikazujeta absolutno in relativno razliko med posameznimi meritvami in njihovim povprečjem. Na sliki 6 so prikazane razlike med modelom pretočne krivulje in povprečji merjenih pretokov. Ustreznost uporabljenega modela pretočne krivulje je bila preverjena z analizo variance, ki temelji na razmerju med eksperimentalno napako posameznih meritev in odstopanjem modela pretočne krivulje od povprečnih meritev v posameznih točkah. V obravnavanem primeru znaša povprečni kvadrat čiste napake MSpE = 3707 l2/s2, povprečni kvadrat napake prilagajanja pa MSL0F = 5740 l2/s2. Razmerje F™ = MSL0F/MSpE znaša 1,55, kar je manj kot tabelarična kritična vrednost porazdelitve F(ak k) = 2,0 (k = 13 - število točk na katere je bil model prilagojen, Slika 4. Razlika med posameznimi meritvami pretokov in njihovim povprečjem. Figure 4. Difference between measured discharges and their averages. Slika S. Relativna razlika med posameznimi meritvami pretokov in njihovim povprečjem. Figure2S. Relative difference between measured discharges and their averages. -100 -I -80 -60 -40 - -20 n JC 0 Й 20 ae 40 60 80 100 ^ XIJL- 0.5 1.5 2 Višina vodne gladine (dm) 2.5 3.5 Slika 6. Razlika med modelom pretočne krivulje in povprečji merjenih pretokov. Figure 6. Difference between the rating curve model and average discharges. p = I - število parametrov modela, N = 51 - število vseh meritev) za mejo zanesljivosti a = 0,05. Z zadostitvijo pogoja: 1 < Fizr< F(0 05 1P je bila potrjena ustreznost modela pretočne krivulje. Z opisanim modelom pretočne krivulje izračunani povprečni dnevni pretoki na merskem mestu M2 so prikazani na sliki 8. Zaradi tehničnih razlogov in občasno zamrznjene struge potoka meritve niso zvezne za celotno obdobje od 1. 11. 2001 do 31. 3. 2003. Okrog 20 % obravnavanega obdobja je brez podatkov. Za obdobja razpoložljivih podatkov znaša ocenjena mediana dnevnih pretokov 25 l/s. Meritve naravnih vodotokov so podvržene številnim napakam. V opisanem primeru jih je težko ovrednotiti, so pa predvsem posledica poenostavljene izvedbe merskega mesta in omejenega števila neposrednih meritev. Zaradi manjšega števila meritev za stanja visokih vod in manjše natančnosti teh meritev (sl. 4) je zanesljivost izračuna visokih pretokov s pomočjo izdelane pretočne krivulje manjša. Meteorološki podatki Modeliranje je bilo izvedeno za dve obdobji. V vsakem obdobju so bili uporablj eni podatki takrat delujočih meteoroloških postaj, ki so bile najbližje obravnavanemu območju. Za prvo obdobje modeliranja (od 1. 1. 1983 do 31. 12. 1993) so bile uporabljene dnevne višine padavin iz padavinskih postaj Matavun, Kozina in Podgrad ter višine potencialne evapotranspiracije iz meteorološke postaje Ilirska Bistrica. V drugem obdobju modeliranja (od 1. 11. 2001 do 31. 3. 2003) so bile uporabljene višine padavin iz padavinskih postaj Kozina in Podgrad ter višine potencialne evapotranspiracije iz meteorološke postaje Godnje. Za prostorsko porazdelitev padavin je bila uporabljena korigirana Thiessenova metoda (Janža, 2003). Vrednosti višin potencialne evapotranspiracije so bile uporabljene brez porazdelitve - enotne vrednosti na celotnem obravnavanem območju. Digitalni model višin Za izdelavo modela površja obravnavanega območja je bil uporabljen digitalni model višin - DMV (ZRC SAZU & Mobitel, d. d., 2000). Velikost celic DMV je 25 m, povprečna višinska natančnost okoli 2 m za ravninska območja, za zmerno razgiban relief okoli 5 m in za hribovit relief okoli 10 m (Oštir, 2000). Za potrebe modeliranja je bila spremenjena velikost celic DMV, tako da ustrezajo ostalim vhodnim prostorskim podatkom. V ta namen je bila uporabljena bilinearna interpolacija. Pedološki podatki Podatki za modeliranje nezasičene cone temeljijo na pedološki karti v merilu 1 : 25000 in izbranih pedoloških profilov (CPVO, 2001). Pedološka karta je poligonski informacijski sloj, sestavljen iz pedokarto-grafskih enot (PKE), ki so osnovne kartografske enote. Posamezna PKE je sestavljena iz ene ali več pedosistemskih enot, ki v naravi značilno nastopajo skupaj in jih zaradi merila karte ni mogoče prikazati ločeno. Poligoni PKE se med seboj razlikujejo po zastopanih pedosistemskih enotah (tipih tal) in njihovem medsebojnem razmerju (Vrščaj & Tič, 1998). Na obravnavanem območju so bile, glede na vrsto in zastopanost pedosistemskih enot v posamezni PKE, le-te združene v sedem pedoloških enot, ki so uporabljene v modelu. Hidravlične lastnosti teh enot so bile opredeljene na podlagi tipičnih pedoloških profilov (za posamezno enoto), izbranih v širši okolici obravnavanega območja. Posamezni pedološki profil je sestavljen iz različnih horizontov, ki imajo teksturne podatke (deleže peska, melja in gline). Z uporabo pedotransfer funkcij so bile na podlagi teh podatkov ocenjene hidravlične lastnosti pedoloških enot, ki so uporabljene v modelu. Vegetacijski podatki Porazdelitev vegetacijskih razredov oziroma raba tal v modelu je bila določena s klasifikacijo satelitske podobe LANDSAT-5 TM (Janža, 2005). Lastnosti posameznih vegetacijskih razredov so bile opredeljene s predhodno določenimi (modeliranimi) vrednostmi vegetacijskih parametrov (Kristensen et al., 2000). Hidrološki model MIKE SHE MIKE SHE je programski paket za modeliranje celotnega hidrološkega kroga (Abbot et al., 1986; Refsgaard & Storm, 1995). Je integriran sistem komponent ali modulov, ki omogoča modeliranje posameznih procesov hidrološkega kroga. Kompleksnost naravnega sistema oziroma njegovo konceptualno razumevanje in zahtevana zanesljivost modela pogojujeta uporabo (vključitev) različnih komponent. Hidrološki procesi so opisani z diferencialnimi enačbami, ki jih program rešuje numerično z uporabo metode končnih razlik. Ena od komponent modela MIKE SHE je MIKE 11, ki omogoča modeliranje hidrodinamičnih procesov površinskih voda. Uporablja se lahko kot samostojni model (reke, potoka, jezera) ali združen z modelom MIKE SHE, kar omogoča modeliranje celotnega hidrološkega kroga na obravnavanem območju. Teoretične osnove modeliranja hidroloških procesov so natančneje opisane v DHI (2000a; 200b) in Janža (2003).V nadaljevanju so podane enačbe modela za opis dinamike podzemne vode, ki je najpomembnejši del modela na obravnavanem območju. Parametri modela zasičene cone so najobčutljivejši v modelu, zato je bila kalibracija omejena na te parametre. V modelu opisuje trodimenzionalni tok podzemne vode v zasičeni coni enačba: IHHhiHNb-. I kjer so: (2) K . K . K koeficienti prepustnosti vzdolž xx уу zz г г koordinatnih osi [m/s]; h piezometrični nivo [m]; Qe volumski pretok na enotski volumen (dotok/iztok) [s"1]; Ss specifični koeficient elastičnega uskladiščenja [m-1]. Model omogoča tudi simulacijo drenažnega toka v zasičeni coni. Ta se pojavi, ko je nivo podzemne vode nad nivojem drenaže (sl. 7). Odtok je odvisen od razlike med nivojema (Л) in časovne konstante (cdr ), ki določa A \/\/ \/ drenaže ООО površina V- h 4~ zdr Slika 7. Shematski prikaz koncepta drenaž v modelu (po DHI, 2000a). Figure 7. Schematic presentation of drains in the model (after DHI, 2000a). gostoto drenaž. Drenažni odtok je modeliran kot linearen rezervoar z izrazom: q = (hn-zdrn)cdrn (3) kjer so: q drenažni odtok [m3] hn nivo podzemne vode (v n celici) [m] ; zdrn nivo drenaž [m]; cdr drenažna časovna konstanta rs_1l. n Zasnova in parametrizacija modela na obravnavanem območju Model na območju Brkinov je zasnovan na povodju P2A - delu povodja P2, ki leži vzvodno od merske točke M2, kjer so bile opravljene zvezne meritve pretoka (sl. 2). Enak pristop je bil uporabljen za kalibra-cijske parametre. Njihove vrednosti so bile določene v fazi kalibracije na povodju P2A in nato uporabljene na celotnem območju modela. Kalibracija modela je bila opravljena na podlagi vizualne primerjave merjenega (s pretočno krivuljo izračunanega) in modeliranih hidrografov. Najpomembnejša razloga za uporabo enotnega koncept na vseh povodjih sta: • povodja imajo podobne (hidrogeološke, topografske, vegetacijske) značilnosti; • pomanjkanje ustreznih merskih podatkov na ostalih povodjih, ki bi omogočili kalibracijo vsakega povodja posamezno. Modeliranje je bilo izvedeno za dve ločeni obdobji: • prvo obdobje modeliranja, od I. I. 1983 do 31. 12. 1993; • drugo obdobje modeliranja, od 1. 11. 2001 do 31. 3. 2003. Postopek modeliranjaje bil razdeljen na več faz: 1. Modeliranje pretoka na merskem mestu M2 - odtoka iz povodja P2A. Ta model je bil izdelan za obdobje izvedenih meritev pretokov (od I. II. 2001 do 31. 3. 2003). 2. Vrednosti parametrov, ki so bile določene v procesu kalibracije tega modela, so bile nato uporabljene pri izdelavi modela za vsa povodja potokov na celotnem obravnavanem območju. 3. Ta skupni model povodij je bil nato uporabljen za modeliranje količine ponikajočih vod potokov v obdobju med 1983 in 1993, ki ustreza obdobju, uporabljenem v modelu vodonosnika v zaledju izvira Rižane (JanŽA, 2003). Za potoke na območju Brkinov je značilno hitro povečanje pretokov po deževju. Večina vode odteče površinsko in podpovršinsko. Ta tokova sta v modelu simuliran z uporabo fUnkcije drenaž. Nivo drenaž je bil postavljen 0,5 m pod površino. Predpostavljeno je bilo, da je to globina manjših kanalov, ki delujejo kot drenažni sistem in niso ustrezno opisani z uporabljenim DMV (zaradi njegove premajhne natančnosti). Model deluje tako, da v primeru, ko podzemna voda v računski celici naraste nad nivo drenaž, odvede presežek vode do sosednje celice z nižjim nivojem. Postopek se nadaljuje, dokler ni dosežen vodotok. V manjšem delu povodij (predvsem njihovem nižjem delu), kjer se pojavljajo lokalne depresije, je bila uporabljena funkcija drenaže, ki ne upošteva naklona drenažnega nivoja in odvaja drenirano vodo neposredno do najbližjega vodotoka. Dinamika odvajanja vode je definirana s časovno konstanto, ki je bila določena v fazi kalibracije in ima v modelu vrednost 6xl0"6 s_1. Tako odvedena (drenirana) voda tvori skupaj z osnovnim (baznim) tokom, ki je modeliran kot medzrnski tok v zasičenem delu računske (geološke) plasti, vhodne podatke za model odtoka po strugi potokov. Območje Brkinov je sestavljeno iz slabo prepustnih flišnih plasti, zato so hidrološki procesi, ki vplivajo na odtok z območja omejeni na zgornji del teh plasti. V modelu so njihove lastnosti opredeljene z eno računsko plastjo, ki ustreza geološki plasti debeline 5 m. S kalibracijo določena vrednost koeficienta prepustnosti plasti v horizontalni smeri je 10"7 m/s, v vertikalni pa 10~8 m/s. Spodnji meji plasti je pripisana zelo nizka vertikalna prepustnost (10~9 m/s) in deluje praktično kot neprepustna plast. Horizontalna velikost računske celice modela je 90 x 90 m. Odtok po strugi potokov je modeliran s programskim orodjem MIKE 11, ki je integrirano z modelom MIKE SHE. Položaj strug potokov v modelu (sl. 2) je določen z digitalizacijo rečne mreže topografske osnove 1 : 5000 (Geodetska uprava RS). Korita potokov so definirana poenostavljeno s trikotnimi preseki. Na začetku potokov (v najvišjem delu) je nastavljen robni pogoj ni dotokov. Za modeliranje toka v koritih je uporabljena metoda enostavnega hidravličnega izračuna (ang. kinematic routing), kjer temelji postopek izračuna hidrografa v določeni točki na dotoku in hidrografih vzvodno ležečih pritokov (DHI, 2000b). Prednost metode je stabilnost in nezahtevnost glede vhodnih parametrov, ki v obravnavanem primeru niso bili na razpolago. Rezultati in razprava Povprečni simulirani dnevni pretok za celotno drugo obdobje modeliranja (I. II. 2001 do 31. 3. 2003) na merskem mestu M2 znaša 71 l/s (P50 = 29 l/s). Grafična primerjava modeliranih povprečnih dnevnih pretokov na merilnem mestu M2 (sl. 9) z neposrednimi in z modelom pretočne krivulje izračunanimi zveznimi meritvami pretoka 5 2 0.4 Q. 0.2 - izračunani pretok C neposredne meritve i 1.01 31.12.01 2.3.02 2.5.02 1.7.02 31.8.02 31.10.02 30.12.02 1.3.03 Slika 8. Neposredne meritve pretoka in izračunani povprečni dnevni pretoki (s pretočno krivuljo) na merskem mestu M2 (od I. II. 2001 do 31. 3. 2003). Figure 8. Direct discharge measurements and calculated average daily discharges (with rating curve) on measurement location M2 (from 1. 11. 2001 to 31. 3. 2003). Slika 9. Simulirani povprečni dnevni pretoki na merskem mestu M2 (od 1. 11. 2001 do 31. 3. 2003). Figure 9. Simulated average daily discharges on measurement location M2 (from 1. 11. 2001 to 31. 3. 2003). Qopaz - Opazovani pretoki (m3/s) Slika IO. Korelacija med opazovanimi in simulirani povprečni dnevni pretoki na merskem mestu M2 (od I. II. 2001 do 31. 3. 2003). Figure IO. Correlation between observed and simulated average daily discharges on measurement location M2 (from 1. 11. 2001 to 31. 3. 2003). (sl. 8) kaže zmožnost modeliranja nizkih pretokov, kakor tudi dinamiko povečanja in upadanja pretoka. Vendar pa se določeni pretoki (vrhovi) hidrografov ne ujemajo. To odstopanje bi se lahko pripisalo predvsem vhodnim padavinskim podatkom, ki so najpomembnejša vhodna spremenljivka v modelu. Relativno majhno območje povodja P2A (3,6 km2) je podvrženo lokalnim vremenskim razmeram, ki se težko opišejo z oddaljenimi padavinskimi postajami. Korelacija med simuliranimi in opazovanimi (merjenimi) pretoki na merskem mestu M2 je prikazana na sliki IO. Smerni koeficient za linearno odvisnost pri kateri nastopa simulirani pretok kot odvisna, opazovani pretok pa kot neodvisna spremenljivka znaša 1,152, koeficient korelacije med primerjanima pretokoma pa O,75. Glavne bilančne komponente modela vseh povodij za enoletno obdobje (od 3O. 3. 2OO2 do 3O. 3. 2OO3) so prikazane na sliki 11. V modelu je polovica padavinske vode prenesena nazaj v ozračje kot posledica evapotranspiracije. Preostala voda iz povodij odteče do potokov predvsem v obliki površinskega in podpovršinskega hitrega odtoka, ki je v modelu simuliran s funkcijo drenaž. Približno trikrat manjši je počasen osnovni - bazni odtok. Hidrograf simuliranih skupnih odtokov s celotnega območja modela za prvo obdobje modeliranja in statistike odtokov po povodjih in skupno so prikazane na sliki 12 ter v tabelah 2 in 3. Krivulja trajanja simuliranih povprečnih dnevnih skupnih odtokov iz vseh povodij v prvem obdobju modeliranja je prikazana na sliki 13. Celoten razpon pretokov je bil razdeljen na trideset enakih intervalov. Iz grafa je razvidno, da močno prevladujejo nizki odtoki. Le okrog 15 % simuliranih Slika II. Glavne bilančne komponente modela vseh povodij (skupno) za obdobje od 30. 3. 2002 do 30. 3. 2003 (lmm ustreza 1,4 l/s). Figure II. Main water balance components of the model of all catchments for the period from 30. 3. 2002 to 30. 3. 2003 (lmm corresponds to 1.4 l/s). 32 28 24 ■ Я 20 "e Š 1 Slika 7. Prostorska predstavitev krivulj tečenja za jeklo CFS3 - eksperimentalne in napovedane vrednosti pri uporabi nekonstatnega parametra gladkosti (we = 0,03, wr = 0,10, wt w^^r0»03) ■ ■(e-0.02) 0,01, Diskusija Iz maksimalnih napetosti tečenja za različne temperature in hitrosti deformacije smo izračunali konstante hiperbolične sinusne enačbe 7 [26]. Z = e exp(ß/ RT) = ^(sinh acj )" (7) To enačbo najprej logaritmiramo in takole preuredimo ln(sinh(aa)) = iln(8) + -^---ln(,4), (8) n RnT n nato pa definiramo funkcijo c2, ki minimizira razliko med izračunanimi in izmerjenimi vrednostmi napetosti tečenja [25] „ 2 _ Y1 (z,- ~ a\xi ~ Utfi ~ fl3 )2 ь Li 2 d (W) i=l kjer je N število meritev, z,. = ln(sinhaa), xt =lne,. in y. =104r_1- Ostale oznake so at =n~\ a2 =10^Qn~lR~x in аъ =n~l\a.A. Pri napaki upoštevamo samo napako napetosti z. , ki jo lahko izrazimo kot ei =a e" cothaa,, kjer so napake izmerjenih napetosti. Podrobnosti postopka minimizacije zgornje enačbe 9 najdemo v [23]. Funkcija c2 ima minimum pri Q = 316,86 kJ mol1, a = 0,00945 MPa-1, n = 5,3 in A = 1,88T012 s-1. Ta vrednost je primerljiva z območjem vrednosti (Q = 280 - 330 kJ mol-1) sorodnih jekel za poboljšanje, ki so bile dobljene po tangentni metodi na osnovi rezultatov, dobljenih iz torzijskih preizkusov [16-17]. Primerjavo odvisnosti maksimalne napetosti tečenja od temperature, med izračunanimi in izmerjenimi vrednostmi za tri različne hitrosti deformacije, prikazuje slika 8a. Na sliki 8b pa je podana primerjava med izračunanimi in izmerjenimi maksimalnimi napetostmi tečenja. Iz obeh slik je razvidno, da za obravnavano jeklo izbrana empirična enačba 8 odlično opiše zvezo med hitrostjo deformacije, temperaturo in maksimalno napetostjo tečenja. Kemična sestava jekel od šarže do šarže, čeprav v dovoljenih mejah, stalno niha. Vpliv kemične sestave (posebno karbidotvornih elementov) na napetost tečenja avtorji pri njihovem napovedovanju rešujejo z vpeljavo ogljikovega ekvivalenta, saj pri uporabi BP nevronskih mrež veliko število vhodnih vplivnih parametrov vpliva na natančnost napovedovanja. V primeru uporabe CAE nevronske mreže pa teh težav nimamo, saj lahko uporabimo poljubno število vhodnih parametrov, torej lahko upoštevamo tudi vsak legirni element posebej [24-25]. Pri uporabi običajnih BP nevronskih mrež moramo najprej določiti optimalno arhitekturo nevronske mreže, t.j. določiti število plasti in število nevronov v teh plasteh, saj doslej še ni jasnih navodil za izbiro teh parametrov. Pri uporabi CAE nevronske mreže imamo fiksno število skritih plasti, število nevronov v plasti pa je odvisno od števila modelnih vektorjev. Inženir se lahko posveti modeliranju pojava in ne izgublja časa z določanjem abstraktnih parametrov nevronske mreže. Bistvena prednost CAE metode v prikazanem primeru je enostavnost in relativno dobra natančnost predlaganih modelov. Pomembno dejstvo, ki dodatno opravičuje uporabo CAE nevronskih T [°C] "p [MPa] (izmerjeno) Slika 8. Primerjava med izmerjeno in izračunano odvisnostjo maksimalne napetosti tečenja od temperature za tri različne hitrosti 0,1 s"1, I s"1, 8 s"1 (a.) in primerjava med izmerjenimi in izračunanimi maksimalnimi napetostmi (b.). mrež je, da se parametri med preizkusom spreminjajo; CAE nevronske mreže so namreč na osnovi tako dobljene baze podatkov sposobne določiti realne krivulje tečenja pri konstantnih pogojih. V obravnavanem primeru so bile pri meritvah upoštevane samo tri različne hitrosti deformacij. Znano je, da je mogoče skozi tri točke v najboljšem primeru napeljati kvadratno krivuljo oz. polinom drugega reda, ki opisuje pojav. Brez vnaprej znane zakonitosti, ki bi jo upoštevali v CAE modelu, je opisovanje pojava v smeri hitrosti deformacij relativno slabo. Predvidevamo, da bomo v prihodnjih raziskavah CAE metodo lahko dopolnili tako, da bo v primeru premajhnega števila podatkov za katerikoli vhodni parameter pojava (v opisanem primeru hitrosti deformacije) mogoče upoštevati vnaprej predpostavljeno oz. poznano zakonitost. To dejstvo se namreč s pridom izkorišča pri optimizaciji experimentalnega dela, težje pa ga je upoštevati v matematičnih modelih krivulj tečenja brez a-priori predpostavk. 5. Zaključki Za potrebe optimiranja tehnologije toplega preoblikovanja jekla za poboljšanje CF53, namenjenega predvsem za strojne dele, ki so lahko izpostavljeni visokim mehanskim obremenitvam, smo na Gleeble 1500 izvedli tople stiskalne preizkuse. Temperaturno, deformacijsko in hitrostno deformacijsko območje ustreza območju toplega valjanja. Preizkusili smo metodo napovedovanja krivulj tečenja s pomočjo umetne inteligence (CAE NN). Oblike krivulj tečenja kažejo na procese dinamične rekristalizacije med toplo deformacijo. Študija je potrdila odlično napovedno sposobnost CAE nevronske mreže za napovedovanje krivulj tečenja Pri tem smo uporabili dva pristopa in sicer metodo konstantnega parametra gladkosti ter metodo nekonstantnega parametra gladkosti. Slednji daje boljše rezultate predvsem zaradi boljše sposobnosti modeliranja fizikalnih zakonitosti v področjih velikih gradientov. Dosežene natančnosti so praktično znotraj območja 5 %, v povprečju znaša napaka 3 %. Pri uporabi CAE nevronske mreže ne potrebujemo določevanja optimalne arhitekture plasti mreže, s »poskus-napaka« postopkom pa enostavno določamo optimalne vrednosti parametra gladkosti pri razpoložljivi bazi podatkov. Postopek napovedovanja z CAE NN je enostavnejši v primerjavi BP NN, natančnost napovedovanja pa je na isti ravni. Izračunali smo tudi aktivacijsko energijo za CF53, ki znaša Q = 316,86 kJ mol1. Zahvala Avtorji se zahvaljujejo Unior-u Zreče za izdelavo vzorcev. Literatura [1] Koop, R., Luce, R., Leisten, В., Wolske, M., Tschirnich, M., Rehrmann, Т., Volles, R., (2001): Flow stress measuring by use of cylindrical compression test and special application to meta forming processes, Steel Research, 72, 394-401. [2] Schoten, K., Bleck, W., Dahl, W. (1998): Modelling of flow curves for hot deformation, Steel Research, 69, 193-197. [3] Kugler, G., Turk, R. (2004): Modeling the dynamic recrystalization under multi-stage hot deformation, Acta Materialia, 52/16, 46594668. [4] Liu, J., Chang, H., Hsu, T.Y., Ruan, X. (2000): Prediction of the flow stress of high speed steel during hot deformation using a BP artificial neural network, Journal of Materials Processing Technology, 103, 200-205. [5] Hata, N., Kakado, J.I., Kikuchi, S., Takuda, H. (1985): Modelling on flow stress of plain carbon steel at elevated temperatures, Steel research, 56/11, 575-582. [6] Rao, K.P., Hawbolt, Е.В. (1992): Development of constitutive relationships using compression testing of medium carbon steel, Trans. ASME 114, 116-123. [7] Kliber, J., Schindler, I. (1997): Matematical description of stress-strain curve in metal forming processes, Metalurgija, 36/1, 9-13. [8] Davenport, S.B., Silk, N.J., Sparks, C.N., Sellars, C.M., (2000): Development of constitutive equations for modelling of hot rolling, Materials Science and Technology, 16, 539-546. [9] Liu, J., Chang, H., Wu, R., Hsu, T.Y., Ruan, X. (2000): Investigation on hot deformation behaviour of AISI Т1 high-speed steel, Materials characterization, 45, 175-186. [10] McQueen, H.J., Ryan, N.D. (2002): Constitutive analysis in hot working, Materials Science and Engineering A, 322, 43-63. [11] Phaniraj, M.P., Lahiri, A.K. (2003): The applicability of neural network to predict flow stress for carbon steels, Journal of2Materials Processing Technology, 141, 219-227. [12] Rao, K.P., Prasad, Y.K.D.V. (1995): Neural network approach to flow stress evaluation in hot deformation, Journal of Materials Processing Technology, 53, 552-566. [13] Hodgson, P.D., Kong, L.X., Davies, C.H.J. (1999): The prediction of the hot strength of steels with an integrated phenomenological and artificial neural network model, Journal of Materials Processing Technology, 87, 131-138. [14] Kong, L.X., Hodgson, P.D. (1999): The application of constituitive and artificial neural network model to predict the hot strength of steels, ISIJInt. 39/10, 991-998. [15] Cheng, Y.W., Tobler, R.L., Filla, B.J., Coakley, K.J. (1999: Constitutive behaviour modeling of steels under hot rolling conditions, NIST Technical Note 1500-6, Materials Reliability Series, 39-49. [16] McQueen, H.J., Ryan , N.D. (2002): Constitutive analysis in hot working, Materials Science and Engineering A, 322, 42-63. [17] McQueen, H.J., Yue, S., Ryan, N.D., Fry, Е. (1995): Hot working characteristics of steels in austenitic state, Journal of Materials Processing Technology, 53, 293-310. [18] Grabec, I., Sachse, W. (1997): Synergetics of Measurement, Prediction and Control. ISBN 3540-57048-9, Springer-Verlag. [19] Grabec, I. (1990): Self-Organization ofNeurons Described by the Maximum-Entropy Principle. Biol.Cybern, 63, 403. [20] Perus, I., Fajpar, P., Grabec, I. (1994): Prediction of the seismic capacity of RC structural walls by non-parametric multidimensional regression, Earthquake Eng. Struct. Dyn. 23, 1139-1155. [21] Fajjfar, P., Perus, I. (1997): A non-parametric approach to attenuation relations. Journal of Earthquake Engineering, 1(2), 319. [22] Perus, I., Fajpar, P. (1997): A non-parametric aproach for empirical modelling of engineering problems, Engineering Modelling, 10(1-4), 7. [23] Kugler, G., Knap, M., Palkovski, H., Turk, R. (2004): Estimation of activation energy for calculating the hot workability properties of metals, Metalurgia, 43/4, 267-272. [24] Terčelj, M., Perus, I., Turk, R. (2003): Suitability of CAE neural network and FEM for predicting of wear on die radii in hot forging, Tri-bology International, 36, 573-583. [25] Turk, R., Perus, I., Terčelj, M. (2004): New starting points for prediction of tool wear in hot forging, Int. J. of Machine Tools & Manufacture, 44/12-13, 1319-1331. [26] Sellars, C.M., Mcg. Tegart, W.J. (1972): Int. Metall. Rev., 17, 1-24. Feeding behaviour of graphite containing material Heinz-Josef Wojtas Fakultät fur Ingenieurwissenschaften, Institut fur Angewandte Materialtechnik, Universität Duisburg-Essen, Lotharstr. 65, D-47057 Duisburg; E-mail: hk225wo@uni-duisburg.de Received: June 30, 200S Accepted: November 24, 200S Abstract: On casting and cooling of metals in a mould energy in form of heat will be transferred from the liquid metal to the moulding material. The mechanism of energy transfer and the corresponding capture capacity of the moulding material define the energy transfer in form of heat by cooling of the metal at contemporaneous heating of the moulding material. Only the exact knowledge of the temporary and quantitative process of heat flow will lead a solidification simulation to correct results. According to second law of thermodynamics energy in the form of heat may only be transferred from a colder to a warmer material if this is forced by mechanical work. Self-transfer of heat is only possible from an area of higher temperature in direction to areas with lower temperature. Only on the premises of a temperature drop a self-transfer of heat is possible. The energy transfer between materials of different temperatures is finished when an energetic balance is reached, i.e. when after the heat transfer there is the same temperature at all materials. Energy in the form of heat is transferred by means of heat conductivity, heat transfer (heat convection) or heat radiation. Key words: metal/mould material, graphite, heat transfer, conductivity, energy balance, feeding Part I: Definition of various influencing variables Physical values defining the solidification process Density (p) The density of a substance is the relation of mass to volume, some times called "specific mass". Although most of the solid bodies and liquids expand on heating, these volume changes are relatively small and nearly independent of temperature and pressure. Thermal conductivity (A,) Thermal conductivity means the energy transport inside a substance by interaction of atoms and molecules, which are not transported themselves (for example within the single silica grain, the solidified crystal or the convection-free melt). The energy transport is done in the way, that quicker and larger swinging molecules of a more heated substance area continue to transfer energy to adjoining and less heated substance regions, until - after energy transfer is finished - the same average swinging condition and the same temperature is set. Thermal conductivity of various materials is different. It is expressed by thermal conductivity X. Thermal conductivity of different materials is experimentally investigated and depends on temperature. Heat transfer (sometimes called "heat convection") Under the term "heat transfer" is to be understood an energy transport between several materials (for example between two touching silicate grains or the solidified casting surface and the mould material) with different temperature and without material transport. Energy transfer is done in the touching area of both substances and implies a temperature drop. After heat transfer is finished both substances have the same temperature T within the touching area. Real heat convection is combined with a material transfer. This materials transfer is called forth within liquid and gaseous materials by a thermal change of volumes and density. Heat convection within liquid metals is detectable only at an early stage of mould filling. Heat convection of air has not been detected within a mould. The pressure build-up within pore volume of the mould by means of evaporation or combustion processes does not belong to the physical values, but to the variable moulding material values defining the solidification process. Heat2transition If liquids or gases of different temperatures are separated by a solid wall, an energy transfer will take place, consisting of heat conduction and heat transfer. This combined form of heat transfer is called heat transition. Heat2radiation Between bodies of different temperatures heat is not only transferred by means of thermal conductivity or heat convection but also at same time by means of heat radiation (for example across the gap between the silicate grains). Everywhere at heat transfer processes the kinetic energy of the molecules is partially transferred into radiation energy and radiated. The percentage of radiation energy at complete energy transfer is small at low temperatures. Heat radiation belongs to electro-magnetic waves and they are within visible frequency range only at higher temperatures. Frequency range, diffusion, reflection and refraction follow the valid regularities of luminous radiation. The radiation energy impacting a radiated body may be absorbed reflected or transmitted. The absorbed part of the radiation energy will be transferred again into kinetic energy of the molecules and will heat the radiated body which therefore becomes source of own radiation. A body absorbing all impacting radiation energy is called an absolute black body. Absorption and emission are on highest level. The heat Q radiated from an absolute black body with a surface A within a time t depends on the body temperature T and results from the law of Stefan Boltzmann. Amount of heat The amount of heat Q is the quantity of energy in the form of heat, which can be feed to a material or taken away. On thermo-technical calculation there is often a reference to material quantity of I kg. The amount of heat referring to I kg is called specific heat q. Those materials that extend noticeable during heating is to differ between heat capacities at constant pressure and constant volume respectively. This is irrelevant for liquids and solid bodies due to the unimportant volume modification. Although it is normally calculated with heat capacity at constant pressure. Specific heat capacity (cp) The specific heat capacity (cp) characterizes the various heating-up of materials. It indicates the amount of energy in the form of heat, which is necessary to heat a material quantity with a mass of I kg for I К (or 1°C) by keeping the respective aggregate state. If a material quantity with the mass mj and the temperature T shall be heated to a temperature Tp by supply of heat this adding heat results of: Q = mcp (Tp - Tj) (1) Variable moulding material VALUES defining THE SOLIDIFICATION process Density, bulk density, packing density (p) At production of moulds a loose, granular material conglomerate is compacted on machines. The voids fraction between the grains decreases. Generally you have to assume considerable density fluctuation for all moulds. The density fluctuation varies with type and sort of sand, binder, moulding box and method of compacting. By this the voids fraction fluctuates, too, and - as a result Figure I. View of a mould wall - the density of the material conglomerate. This "packing density" significantly influences the energy transport through the mould (Fig. I). Therefore the average density of the mould should be considered at least. The movement of the mould wall is mainly caused by the thermal expansion of the mould sand. But this expansion is only approx. I % within relevant temperature range, so that the amendment of packing density caused by this may be neglected on further calculation of heat flow. Heat capacity "apparent heat capacity" (c) The heat capacity, that enters into the equations 2 in this case, is the heat capacity of a granular material conglomerate of mould basis material and binder system (Fig. I). It is an "apparent heat capacity" applying at first only to the volume element with a certain voids fraction and a certain packing density, which underlies the measurement. Я T pc(f) — = div [Х(Г)gradT] + W(T,r, t) (2) The influence of the inorganic binder to the heat capacity of this volume element may be neglected, as their portion is small and as there is not a big difference within heat capacity between mould basis material and binder. The portion of organic binders within mould material systems is even smaller and their contribution to the heat capacity of the system, in which an oxygen deficiency inside of the pore volume avoids the complete combustion of the binders, results from required energy for disintegration and the energy set-free from combustion. This energy contribution in the form of a heat source within the volume element might be negligible. The enthalpy for the a-ß allotrope change of the quartz is 10.5 kJ/kg and - compared with the heat capacity of the mould - is consequently negligible small too. Inside wet sand the vaporization, the condensation and the migration of the condensation zone goes through the volume element at low temperatures at the beginning of the solidification. Therefore, at first a heat source (releasing condensation heat) goes through the volume element followed by a heat lowering (depriving heat of vaporization). These effects enter into the apparent heat capacity and there is no need to consider them specially. This apparent specific heat or heat capacity respectively would be independent from T but would depend on the temperature at transition point metal / mould material and considers the heat absorbed from the mould completely. Thermal conductivity "apparent thermal conductivity" X In the same meaning as at heat capacity there results an "apparent thermal conductivity". These "apparent thermal conductivities" would be different at the same sand and different casting alloys, such as aluminium, cast iron and steel. It results in a dependence of the apparent thermal conductivity from packing density and from the melting temperature of the alloy. The apparent thermal conductivity decreases on increasing voids fracture. Within a lot of papers the specific thermal conductivity is defined as apparent specific thermal conductivity between room temperature and transition temperature at transition point metal / mould. But here the above-mentioned definition is used, as it considers the heat absorbed from the mould. Variable material values, which DEFINE the solidification PROCESS Average carbon activity The average carbon activity of cast iron is mainly defined by the content of carbon graphite or carbide stabilizers as well as of the actual temperature. Within a certain range the degree of saturation and the carbon equivalent are rules for it. Local carbon activity This means the carbon activity directly at the place of solidification. It is defined by the local composition which is defined by separation processes, liquation and undercooling. Diffusion coefficient The diffusion coefficient of a cast iron is defined by the content of carbon, graphite and carbide stabilizers as well as by the actual temperature. There is no simple term like the degree of saturation to describe this value, but it defines the separation process of the two phases austenite and graphite during the eutectic solidification. Diffusion path length The diffusion path length during the separation of the eutectic melt to austenite and graphite is arranged by the number of nuclei forming and able to growth. The diffusion path length only depends on the number of nuclei and not on their formation (theory of nuclei formation). Growth of austenite and graphite The growth is defined by the thermodynamic values volume energy and interfacial tension. Influence of material values Material values and their definable structure If - in a first step - the shrinkage volume resulting from a mould wall movement are regarded as constant and not influenceable by the material, the total shrinkage volume results as additive sum of contraction volume of the hypoeutectic austenite the eutectic austenite and the volume increase of graphite at its precipitation. The size of the hypoeutectic solidification contraction is mainly defined by the chemical composition, especially by the carbon and silicon content and values like the degree of saturation or the carbon equivalent respectively, in which further elements are considered. They are a degree how much the actual composition differs from the eutectic one (Fig. 2). Therefore the quantity of hypoeutectic solidification may be fixed by the thermal analysis by means of the temperatures, which define the beginning and the end of the hypoeutectic solidification. A contraction volume can be derived from the weight or volume of the predictable hypoeutectic solidification. This contraction volume can be equalized either by feeding or can be minimized by changing the chemical composition. The minimization of the contraction volume may be done by ЛТ- Unterkühlung der Schmelze Te: Temperatur eter stabilen elektischer: Erstarrung Fo: Oberfläche des Keims ìGq: Freie Enthalpie des Keims Nr Zahl der kugelförmigen Cluster MEG: Menge des eutektisefren Graphits Figure 2. Influence of the solidification manipulation for controlling the solidification contraction. addition of carbon or silicon or by an amended inoculation modus. The hypoeutectic solidification portion, defined by this way, fixes the quantity of the eutectic solidification as a complementary sum to the total volume or to the weight considered. The eutectic solidification consists of the eutectic austenite with a volume contraction and the graphite with a volume dilatation. Eutectic austenite and graphite are fixed in their mutual relationship according to the multiple component system. By this the contraction sum of the eutectic austenite becomes a calculable value (Fig. 2a). The only value, which is now decisive for the decrease of the eutectic shrinkage volume, is the quantity of the precipitated graphite with the corresponding volume increase. At first you have to avoid the possibility of the meta-stable solidification of eutectic that would be combined with a non-compensation of the eutectic austenite contraction. As a rule this is done by adjustment of the carbon and silicon contents to the wall thickness and by adding inoculants that initiate the precipitation of graphite. The quantity of eutectic graphite and consequently the size of volume increase is a function of carbon and silicon contents. The number of spheroids is a question of the number of nuclei, as the same graphite volume may divide to one great spheroid or to any number of small ones. The efficient number of nuclei results of the content of inoculants and the combination of efficient components within inoculants (Fig. 2a). Without entering nucleation theories it is possible - by means of known data from literature of surface tension and enthalpy of fusion - to make calculations, which may lead by support of the thermal analyse to statements about number of nuclei and the approximate graphite volume (Fig. 2b). Influence of moulding material values The real feeding demand of a casting consists of: • volume deficit of the hypoeutectic contraction, • volume deficit of the eutectic contraction, • volume increase of the eutectic graphite precipitation and • "apparent shrinkage". The "apparent shrinkage" is a value that is only found within a few reflections for feeding calculations so far. To carry out this paper it was consequently necessary to define the possible influence factors of the "apparent shrinkage" exactly. The "apparent shrinkage" is a magnification or enlargement of the mould cavity due to the casting process, the solidification behaviour of the cast iron and the attributes of the mould material (mould wall movement). These influences may increase the feeding demand of the casting. The "apparent shrinkage" consists of two main influence factors: • the load defining the pressure onto the mould wall produced by the casting process and the solidification behaviour of the cast iron (graphite precipitation), • the enlargement of the mould cavity or the stability of the mould that withstands the pressure (mould wall movement). Figure 3. Interaction between solidification and shrinkage in a mould: a) the apparent shrinkage does not appear, b) the apparent shrinkage appears. Pigure 3 shows shrinkage cavities appears when the feeder does not have enough liquid iron for compensation of apparent shrinkage, a) no mould wall movement, shrink-free casting, exactly according to pattern, sufficient feeding volume b) great mould wall movement, casting larger than pattern, shrinkage within casting, insufficient feeding volume. The graphite precipitation or the solidification behaviour of cast iron with spherical graphite is not the only parameter that defines the pressure onto the mould wall. Other loads developing on casting and solidification process are big figures too. These loads result of: • the ferrostatic pressure, • the casting heat, • the composition of alloy. The second main influence factor defining the apparent shrinkage is the mould movement or the mould attributes respectively. The dependences of the mould wall movement result of: • the type of mould material, • the composition of mould material, • the compaction, • the geometry of the mould, • the attributes of the mould box. Part 2: Solidification structure of cast iron AND efficiency OF FEEDERS Thermal Analysis From the Ist derivative of the cooling curve results the gradient of the cooling rate over the complete solidification process. The point where the cooling rate (the first derivate) shows the first maximum and starts to decrease again corresponds to the liquidus temperature TL. The following first minimum is the temperature for the beginning eutectic solidification, the eutectic start temperature TES. The TES following point of intersection of the Ist derivative and the zero line marks TElow. At this point the dissipated heat and the released latent heat are in balance. From this point the released latent heat of graphite precipitation prevail and this leads to temperature increase. After a now following maximum of the first derivative the cooling rate decreases again and shows a zero value corresponding to TE After this the ongoing eutectic solidi- 0 TL 1 liquid temperature, it is the temperature where the first solid particles appear, normally austenite as a consequence of cooling in the cast iron 2 TES 3 Start temperature of beginning eutectic solidification 4 TE,™ 5 The minimum eutectic temperature after that temperature increases as a result of released latent heat 6 ТЕНЮН 7 the maximum eutectic temperature as a result of the released latent heat 8 TS 9 solidus temperature, the temperature when the solidification process is finished 10 Tmetast 11 Temperature of metastable eutectic solidfication 12 R 13 recalescence - difference of temperatures telow and tehks„ Figure 4. Significant points within cooling curve of cast iron fication ends at point TS, indicated by a minimum at the first derivative. The first derivative is also in a position to define the periods of solidification. The period of hypoeutectic austenite solidification results up to the first minimum, which is marked by the length growth of the dendrites. The first minimum represents the moment when the eutectic solidification begins and between minimum and crossing zero-line the austenite dendrites are growing thicker instead of growing in length. This process is coupled with a eutectic solidification at the same time. This pure eutectic solidification is defined by this zero crossing and the zero crossing of the second derivative at the end of solidification. The possibility of metastable eutectic solidification is marked by the metastable eutectic temperature, which can be calculated with different factors. If TE, falls below low this specific temperature, which is strongly influenced by Si and elements stabilizing carbide, this may lead to a metastable solidification with a corresponding higher volume deficit. Due to carbide decomposition during further solidification process this chill must not be positively recognizable at room temperature. Therefore T should r metast be lower than TE . low Likewise there can arise a shift to metastable solidification at the end of the solidification process by an enrichment of alloying elements within remaining melt. The final temperature of the eutectic solidification should be above those of the metastable eutectic temperature. The height of (TL) liquidus temperature as well as the period of time until temperature TElow will be reached are a degree for the quantity of hypoeutectic austenite solidification and - proportional to this - for a feeding demand. This feeding demand can only be equalized by a feeder. This volume Abkühlkurven Quick-Cup U L» ^ 1200.0 100,0 150,0 Zeh [e] -ungeimpft Ferr. -ungeimpft Ferr. Tellur -GSI Perl. GSI Peri. Tellur -Spherix Ferr. -Spherix Ferr. Tellur -Spherix Peri. -Spherix Perl. Tellur Figure S. Some real cooling curves. Spherix Ferr. Tell. ung. Tellur GSI Peri. Spherix Peri. deficit of solidification cannot be reduced by the following graphite precipitation. Within this phase - after the austenite dendrites have built a spongy network structure and thereby have defined the future surface resp. wall structure - there is at first the diameter growth of the dendrites and the beginning precipitation of graphite between the dendrites with a corresponding volume increase. In this connection exists the possibility resp. the danger of an enlargement (growth) of casting areas with an increasing inner volume deficit. Therefore TL should be as low as possible and the period between TL and TElow should be as short as possible. The recalescence R is the difference between TE and TE . It represents the released low nigh r latent heat of the eutectic solidification. The recalescence should be as low as possible, and TE as well as TE should be clearly low high above Tmetast. Steady and good graphite precipitation can be expected then that would correspond to the conditions for self-feeding or riserless pouring technique. Real cooling curves show characteristic differences with regard to their nucleation condition and the cooling conditions (Fig. 5). From reflections on solidification and solidification morphology the different significant temperatures and time periods of the cooling curves have to point out, which structure as well as which volume deficit have to be expected (Fig. 6). With today's computer facilities it is possible to define the single periods of time and temperature also by a Ist and 2nd derivative of curve already during recording the solidification curve and to Impfmittel A 0,4 % Impfmittel A 0,2 % Impfmittel В 0,4 % Impfmittel В 0,2 % Anzahl der Sphärolithen 310 Anzahl der Sphärolithen 274 Anzahl der Sphärolithen 196 Anzahl der Sphärolithen 175 Bereich 1 Beieich 2 Bereichs Die Gußstücke sind dicht und Die Gußstücke sind dicht aber nicht mehr so konturenscharf. Im konturenscharf, Speiser gut und Anschnittbereich zeigen sich erste Fehler. Die Speiser sind tief aber glattwandig eingefallen. ungleichmäßig eingefallen Die Gußstücke zeigen Innenlunker, Ein merkliches Einfallen im Oberkasten wird erkennbar. Die Speiser fallen Figure 6. Real cooling curves in comparison to casting results. Figure 7. Scheme of points and areas of evaluation in comparison with structure evaluation transfer them to a analysis of variance, from which a correction measure has to be derived. According to Figure 7 evaluation programmes have been made for the melts in question, and it was tried to state dependences. Time frame of feeder efficiency Feeding is a transport phenomenon Feeding is founded on transport processes of flowable material from the feeder into the contraction areas of the casting to compensate there the arising volume deficit. This means that in the casting, too, corresponding transport processes have to take place. These necessary transport processes are restricted by the solidification morphology of the alloys. At a cast iron with spheroidal graphite several crystallization processes with different morphologies take place one after the other. The primary solidification of cast iron is always exogenous spongy up to mushy. Already at an early stage a dendrite network passes through the casting wall and builds a spongy plaiting, in which the remaining eutectic melt solidifies. Spheroidal eutectic solidify endogenous mushy. The graphite spheroids will be enclosed by growing austenite and will be isolated from remaining melt. By diffusing carbon through the austenite to the graphite spheroid this spheroids can grow and can affect a strong pressure to their surroundings. This effect will be stronger and for a longer period of time the longer the solidification time is or the longer the wall / casting remains in liquid condition. Feeding types differ according to: • mass feeding, • interdendritic feeding, • feeding by sucking solidified shells inwards (surface sinks), • self-feeding - compensation of the metallic solidification contraction by the volume increase of precipitating graphite. Mass feeding is the movement of a mixture of crystals and melt from the feeder into the casting due to gravity. This process come to an end as soon as the crystals hinder each other in their movement, hook or jam, this is the so-called "pour point". There are a direct coherence between feeding capacity and solidification morphology. The mass feeding is mainly responsible for the formation of macro-shrinkage within feeder. Interdendritic feeding is the movement of melt through a crystal network. With their quantity portion at feeding it is defined by the ramification degree of the crystals and the resulting flow channels. Within interdendritic area conditions are to assume like within a filter, at which the size of flow channels at a point of time X is defined by volume relation between crystal network and melt. Portion of interdendritic feeding at total volume of a macro-shrinkage has to be estimated as small. Feeding by sucking solidified shells inwards (surface sinks), as movement of two opposite surfaces to each other, can only avoid the formation of an inner deficit and for the present it has no perceptible influence to the formation of macro-shrinkage within feeder. Surface sinks often are not macroscopic perceptible offhand. Self-feeding as a compensation of metallic volume contraction by volume increase of self-precipitating graphite within a melt has to be regarded in a direct coherence between the quantity of self-precipitating graphite and the quantity of hypoeutectic and eutectic austenite, so that the contraction of austenite can partially or completely be compensated by the volume increase of graphite, or may be even exceeded. The latter is known as a swelling or growing of castings combined with a volume increase of the castings. Eingusskern Überlaufkern Oberkasten Brechkern Unterkasten Ohne exotherme Hülse Mit exothermer Hülse Figure 8. The assembly - pour point investigation. Eingusskern Überlaufkern Exotherme Hülse Oberkasten Brechkern Unterkasten ^ш Figure 9. Several Points of time during solidification. Pour point Figure 8 schematically shows the experimental set-up. This set-up was oriented on an experimental arrangement, which already has been used in a similar way. The test-piece, in principle representing the feeder, was arranged within a cope box, a cam with separating core were arranged within a drag box. Sprue-core and flow-off core are set on the cope box. The sprue-core, flow-off-core and separating-core were made of C02 sand or furan resin sand. There were used common production resins without additives. The sprue-core had the function to enable a uniform filling of the test-piece and to position the flow-off-core above the test-piece. The flow-off-core took care for a constant filling level of the test-piece. Surplus material from the sprue-core flew-off over the rear side of cope box after effected filling of the test-piece. This led to uniform test conditions (no feeding effects). The cam and the test-piece heat the separating core (principle of a breaker core below a feeder) and thereby decrease the cooling speed at the bottom side of the test-piece. In this area the test-piece solidifies more slowly. Now it is guaranteed that material which is still being moveable inside the feeder will flow-out by taking off the feeder. Figure 9 shows several points of time during solidification. Conclusions of the results of the pour point test (Fig. IO) Under same conditions pearlitic and ferritic iron become immovable approx. at the same time. For both cast iron qualities the possibility of a mass feeding is finished already before the beginning of eutectic solidification. The volume deficit occurring during further solidification of metallic Ferrltlsche Gusselsen, Impfmittel GSI 0,3% keine exotherme Hülse V31 1,07 4,36 1375'C Perlltlsches Gusselsen, Impfmittel GSI 0,3% keine exotherme Hülse V33 1,07 4,301419°C I-Ohne Hülsa Mitte -»-OH Festanteil | I-Ohne Hülse Mitte -»-OH Festante» | Ferritisches Gusseisen, Impfmittel Spherix 0,3% keine exotherme Hülse V34 1,10 4,48 1470'C Perlltlsches Gusselsen, Impfmittel Spherix 0,3% keine exotherme Hülse V29 0,97 4,01 137ГС 1340,00 1320,00 1300,00 1280,00 0 1260,00 - 1240,00 1 1220,00 £ 1200,00 1 1180.00 I-Ohne Hülse Mitte -»-OH Festantall | -Ohne Hülse Mitte OH a) without exothermic sleeve Ferritisches Gusselsen, Impfmittel GSI 0,3% exotherme Hülse V 32 1,11 4,53 1440*C Perlltlsches Gusselsen, Impfmittel GSI 0,3% exotherme Hülse V35 1,02 4,25 1451 "C EE Exotherme Hülse Mitte -»-MH Festanteil EE Exotherme Hülse Mitte - Ferritisches Gusselsen, Impfmittel Spherix 0,3% exotherme Hülse V36 1,05 4,29 1393°C exotherme Hülse V30 1,04 4,26 1440*C "§ 1220,00 EE Exotherme Hülse Mitte - EE Exotherme Hülse Mitte -*- MH Festentell b) with exothermic sleeve Figure IO. Quantity of the solidified portion a) without and b) with using an exothermic sleeve, registered in the solidification curve. matrix may then only be compensated by • interdendritic feeding, • feeding by surface sinks, • self-feeding, i.e. by volume increase of the eutectic self-precipitating graphite. This should have the main influence with regard to the density of a casting, because the feeding possibility within interdendritic area for longer distances is not possible without corresponding pressure differences (analogy to filter), and a sinking of the walls should become impossible, too, with a dendrite self-supporting network existing all over the complete cross-section of the wall. So a feeder at cast iron with spheroidal graphite can only compensate the liquid contraction from casting temperature up to the beginning of solidification as well as a part of the volume contraction of the hypoeutectic austenite solidification. After starting eutectic solidification there is no moveable metal in the feeder. The main rest of volume contraction of the solidifying metallic matrix has to be compensated by the volume increase of the self-precipitating graphite. This quantity may - among others - be controlled by composition, graphiti-zation potential and cooling speed. With regard to the solidification simulation it results that the early immovability of the mass within feeder has to be considered, as it is only possible to give an accurate prediction when it is known how long a feeder can supply movable material at different iron qualities. Resulting change of shifting thermal centres due to gravity ends at this time. The principle of directional solidification and it's corresponding simulative translation cannot be maintained for cast iron after this point of time, this means approx. 20 - 30 % of the total solidification time at both ferritic and pearlitic cast iron. Thermal centre still existing or still arising at this point of time have to be criticized with the consideration of volume increase of graphite at a controllable eutectic graphitization potential. This must be different to the eutectoid graphitization potential with the consequence of ferritic and pearlitic cast iron. Part 3: Influence of cooling conditions as well as of inoculation to the solidification structure Casting method The hot-wire method as a measuring method to analyse thermo-physical data of refractory material is only conditionally suitable to analyse thermo-physical data of the moulding material, as the values analysed by this method do not reflect the conditions within the mould during casting. The long heating times, necessary for this method due to the necessity of a temperature balance, lead to modifications within mould material. At each measurement at more than 100 °C or even at nearly 100 °C the moulding material will be completely dried. At higher temperatures the lustrous carbon former pyrolyse, and the bentonite, too, changes due to loss of chemically combined water. These processes, which may have effects to the heat transfer within the mould, do not take place within the mould in this way, at least not at a greater distance to the casting or within the time between pouring and formation of a rigid outer shell. Furthermore, it is hardly possible to effect measurements under operation conditions according to the hot-wire method. Due to the fact that the thermo-physical data also depend on the packing density of the mould material and therefore depend on its compaction, the mould - as a consequence -has to be produced by a moulding machine under operation conditions. Afterwards, the hot wire would have to be built into the mould, and the complete mould box would have to be put into the furnace for analyse of the temperature dependence. An alternative to analyse thermo-physical data - especially under operation conditions - is the casting method. This method is useful for analysis of thermal diffusivity. This value describes the speed, by which a temperature modification spreads within mould material. The analysis of the thermal diffusivity is made by the casting method by measuring the spreading of a temperature front within mould material. Although in general an analytic solution of the heat conductivity equation is not possible a solution can be found for simple special cases, especially for one- or two-dimensional problems. These cases can be approached to a test by using a geometrical simple body - for example a cube, a plate or a cylinder. Such a mould is poured and during solidification and cooling process the temperature is measured at different distances to the casting wall within mould material. Design and size of the test body are chosen in such a way to achieve a temperature distribution within mould material, which will be easily to describe. This will only take place if the distance between measuring position and casting wall will be small compared with the expansion of the casting wall releasing heat. If a one-dimensional temperature distribution is achieved (i.e. the temperature only depends on the distance from measuring position to the easting wall) it is possible - by laying down the heat balanee - to ealeulate thermal diffUsivity aeeording to (Tlb-T2ay \2 s 2-At f T, - T T -T С \ r \ In h. ln - V U J U J (4) diameter within mould material. Due to the small mass of these thermo eouples they ean quiekly follow up modifieations of mould material temperature. The thermo eouples are set into blind holes, whieh are eut into the mould material from the mould joint by using a striekle. As the thermo eouples are led from the measuring point (tip of the thermo element) up to the mould joint at a parallel way to the easting At whieh rp r2, rQ are radii of the three measuring positions, At is the relevant period of time, Tj, Tp, TQ are the average temperatures of the three measuring points within this seetion, and T and Tu are the ' 2a 2b temperatures of the middle measuring point at the beginning and at the end of this period of time. There are possible measuring points within mould material in front of the middle of a plate or a eube surfaee; easier measuring eonditions are given at a eylindrieal test body. By using a rotation symmetry several measuring points may be arranged around the test body (Fig. II), and by taking the average of measured eurves it is possible to reduee the effeets of inseeurities of measuring positions and other faults of measuring. The figured test arrangement represents the optimum. Due to mould teehnieal reasons the test body eannot be a perfeet eylinder, as a eertain inelination of mould is neeessary to remove the pattern from the produeed mould. For measuring the temperature distribution within mould material there are used thermo eouples without proteetion eovering, whieh are set into boreholes of less than 2 mm Figure II. Arrangement of measuring positions at a rotationally symmetrical experimental set-up. wall and by that alongside the isothermal line, a falsification of the measuring values by the heat conduction of the thermoelectric wire can be avoided. The lead-in wire of the thermo couples are led to the outside alongside the mould joint and are connected with a multichannel data logger (Fig. 12). Due to the grain structure of the mould material as well as due to its rather low strength the positioning of the thermo couples cannot be done as accurate as desired. This is the reason for using 6 trial equipments (each 4 thermo couples) equally placed around the test pattern (it was measured at a distance of 5, 10, 15 and 20 mm from mould wall). The average of these 6 temperature curves achieved by this method was taken, whereby the highest as well as the lowest measuring value have been ignored. Within mould material result the following temperature curves (Fig. 13). Figure 12. Design of experiments. 1000 0 -.-.-.-.-.-.-.-.-.- 0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000 Figure 13. Average temperature curves within mould material. Figure 14. Calculated temperature diffusivity during test phase. From this temperature curves the thermal diffusivity during test phase can be continuously calculated. Only the area of cooling phase has been evaluated, because after pouring the liquid metal into the mould there are further processes, which influence the measurement and the evaluation. So an arrest point of 100 °C can clearly be recognized at the beginning of all temperature curves, after that the temperature rises again and shows a course, which corresponds to a heat transfer by means of thermal conductivity. The arrest point is caused by the displacement of water, which was used for bonding purposes. This is the reason for not using this area within the calculation of thermal diffusivity. A second area at which the measurement cannot be evaluated is within figure 14 approx. 2000 up to 2400 seconds. During this time period a latent heat that compensates the cooling for a certain period of time is set free within the casting. During the further course of the measurement temperature differences become smaller and smaller, so that the relative measuring error continuously rises, until at least, at low temperatures, no more evaluation will be possible. 200 2SO 300 400 450 500 550 Temperatur [°C] 650 700 Figure 15. Calculated temperature dependence of temperature diffusivity Gießversuch Lohmann, Rotationssymmetrischer Aufbau, Form 2 IBìiffiSliijSi I 1|ю°< 'M бооо|| ^ j 7c|» 411 i|n[l Z.IIM -T1a-Т2я ТЗа-a | Figure 16. Temperature curves of steel If the calculated temperature diffusivity is figured above the average temperature taken within mould material (in the area of measurement points) it results to Figure 14. The gap, within calculated temperature diffusivity curve, corresponds to the above explained arrest point due to set-free latent heat. But it is easily possible to define a compensation curve that bridges this area (Fig. 15). By using a single test it is possible to define the thermal diffusivity of mould material for a large temperature range (here: 200 up to 700 °C). Nevertheless the evaluation of the used mould material - consisting of silica sand with 8 % bentonite and adjusted to a compactability of 45 % - shows a rather small temperature dependence of thermal diffusivity. Similar results have been achieved on further tests, among other things with production mould materials of participating foundries. As expected, the steel tests do not show the described arrest point of temperature curves. The reason is the different solidification behaviour of steel. At each participating foundry two measurements had been done (Fig. 16). It is conspicuous that at each second measurement were always analysed higher thermal diffusivity in the three foundries for all test runs. This may eventually refer to the fact, that the moulds were produced directly one after the other, i.e. with an interval of several minutes, according to machine cycle time. But the positioning of the thermo couples and the connection to the data logging device took approx. 30 minutes per mould, so that - although the second mould had been immediately covered - a drying resp. modification of mould material by souring effect could not be excluded in the course of time. Influence of different cooling conditions to solidification The different thermal diffusivities of the mould material, i.e. its different cooling effect and by that a different cooling rate within casted metal - even at same conditions - may cause different structures and by that different shrinkage sensitivities. The effects, which may have even small differences within cooling rate, are shown in Figure 17. In this diagram the relationship is shown between the volume contraction and the degree of saturation. It should be noted that a volume expansion can only take place in a small range of the degree of saturation, defined by Si-contents, and this only for the stable eutectic solidification, i.e. for low cooling rate. At higher cooling rates the curves rise up, so that no more volume expansion can take place within short time, even not at an optimum degree of saturation of 1. Therefore a limitation of cooling rate is necessary to secure a self-feeding. Figure 17. Theoretic volume modification of cast iron It is apparent that the concept of the degree of saturation is not only sufficient for description of volume modification and by that for the formation of shrinkage, i.e. the figure of carbon with regard to the graphite precipitation cannot be replaced for example by silicon, as it has been done at corresponding calculations of CE and Sc values. But it also shows that at same chemical composition in a first step there cannot be differences between a pearlitic and a ferritic cast iron with regard to structure, volume modification and thereby shrinkage behaviour. Ferritic and pearlitic amount will not be fixed before eutectoid transformation in solid condition. Between the group of curves of stable and metastable solidification there is the area in which the structure will be defined by the cooling rates thereby the shrinkage behaviour, and that will be the result of thermal diffusivity of mould material and the set-free heat quantity of the test piece. But this process is - at same carbon activity within the melts - highly defined by diffusion, diffusion length and the growth rate, because in opposite to the ledeburite formation the graphite precipitation is a kinetic problem. The lower group of curves shows the optimum of an inoculated stable solidification with the effect of self-feeding. A continuous transition is possible between both groups of curves, which may be expressed at the castings in form of hard spots or carbides within structure. Each part of carbide or ledeburite within the structure increases the feeding demand in opposite to the optimum. Due to the following transformation and precipitation processes in solid it might be these processes during cooling will not be recognized within metallographic evaluations at room temperature. The adjustment of different cooling rates by variation of thermal diffusivity of mould material - for example by different compatibilities or mould material compositions - is only hardly to realize in practice. A pattern plate with test castings had been made to control the dependences. By one pouring there could be casted plates with a thickness of IO, 20, 30 and 40 mm, a cube with module 1.6 as well as a transverse link for cars (Fig. 18). Figure 18. Cluster with test castings ■ Figure 19. Stepped wedge RMZ-M&G 2005, S2 The different plate thicknesses led to different moduli and by that to different reproducible cooling rates. The transverse links for cars were sawed, and the shrinkage was compared with known fault protocols of the foundry, but the results will not be shown at this place. Parallel to this stepped wedges were casted and compared (Fig. 19). During the tests there were casted: • ferritic and pearlitic adjusted cast iron, • with each 2 different inoculants. By these it was varied: • carbon activity, • diffusion coefficient, • diffusion length, • austenite resp. graphite growth. Whereby further parameter were hold on a constant level for example degree of saturation, mould material composition etc.. Three test pieces were taken from each cube or plate, each one of the top (near mould joint), from the bottom and from the middle. These test pieces were polished and they were evaluated by using a image analysis program with regard to the parameter: • number of spheroids (recognized spheroids per mm2) and • graphite quantity (percentage at surface). As an example, Figures 20 and 21 show the influence of different wall thicknesses (and by that different cooling rates) to the formation of spheroids. There are remarkable differences within the single structures. By increasing wall thickness and by that by decreasing cooling speed the number of spheroids decreases, but their diameter increases. Especially inoculant В seems to have its optimum effect at a wall thickness of 30 mm. Casting in plate design At left side the Figure 22 shows pearlitic cast iron and in comparison ferritic cast iron at right side. The upper diagram shows the Ferritisches Gusseisen, 0,2% Impfmittel A ? 'l ' . * • ' ' • - ' - * ч » , „ « ■ • ■ » ■- •'.•i.,: 'r j ti ':■ ,•'■' i "■■. i'1'. ». * P ». • •* <; • - > . ■ - *' ■■ ■ . r .■ ' ■■'■ . ■ /v-^ri'. ' ; )),{ V.; ; - ■*■ -, * Л * -Ч ' ; Я ' :>•*•' V ■ ■ ■ ; • . , • * ■ % WOrtel Unten tOmm Wand Unten 20 mm Wand Unten 30 mm Wand Unten 40 mm Wand Unten Ferritisches Gusseisen, 0,3% Impfmittel A è, TI ■ • » « < * , i ». j , л / . ( ' 1 » » • # * ц / 1 1 -l * ' " ^ 1 * * - . . "X • > ♦ ^ v . .. ^ ^ ■ ». « > т - ' ■ 4- . - Iii . ir zи • 1 M * " '. ■■ * V v•• * ' ■ ^ ^ г. • v - - Wurtel Unten lOrrun wand unlsn 20 rnm Wand Unten 30 mm Wartd Untóri 40 mm wand unten Ferritisches Gusseisen, 0,4% Impfmittel A ■ V é • » ' • ч » „ • ' - ; + * -i ' * * * .» v,- • . ^ ; - «ti. " . * - i k , _ - * t -. * ■■ ■"■ . ■ .. ". ■. " ' -■- ; -v t ■ ...* ■.. * j* _ r * .'■ V-, ■ ' - t-. : У;'.' v -'v:.' »i • ' л • ** : л . t ••• • . v ». » • ' • T» + * wortei unten lOrniTì wand Unten 20 mm wend Unten 30 mm wand unten 40 mm Wand Unten Ferritisches Gusselsen, 0,2% Impftnittel Б . " ' 4 * i - V , i' r Würfel Unten 10mm Wand Unten 20 mm Wand Unten lo mm Ward Unten 40 mm Wand Unten Ferritisches Gusselsen, 0,3% Impfmittel В V ж * f rr, > * 't r . л, 4 •, , * * > ,i V» t " ' f • i .i,, V J ' ', '' . -. V;, VL IV:- ■ X. :■. % • - '.f ''/i-.1- t ■ "i • . ^ , ' » i; .i v;: ** ■ ::»» v/1 -H'. 1 -• • ■ ■*. t v . " Wurf el Unter 10tnm Ward Unten 20 mm wand unten 30 mm ward unter 40 mm Wand Unten Ferritisches Gusselsen, 0.4% Impfmittel В »K . vv * * . - - - ^X*: -; : :"'*- ■ ji 1 4 . .. 4 . ' 4s . * , ' '.* 1 ч • " . . . -V *. * г. ,.*"•' V. ;*> ••Л■•.■■".■ . . **'.'1 'V ) "* t —• *. 1 Würfel unten 10mm Ward unter го mm wand unten 30 mm ward urter 40 mm wand Unten Figure 20. Femtic ductile, inoculated with inoculante A and В number of spheroids, the lower diagram shows the surface component of spheroids in percentage. Within each diagram inocu-lants A and В are shown in comparison. At pearlitic cast iron it is clearly to see, that at same type and concentration of inoculant a higher cooling speed (at lower wall thickness) leads to an increase of the number of spheroids. The evaluation of surface component does not show a uniform tendency. Perl iti s с h< ìs Gusseisen, 0,2 % Im pfmittel А • '* -V • •* . * . ;V! . i -i*.- v ■ .:■ i;;'';.;;.-;.'./..-.: t. "f' ;■ ■ •'■ ' , ...... • - • - ■--..*■ о ; *■' >v'. *j. . ' % .*'" " f.'3.i \ '■ . . ' ! " • ' ■.. .■ ■•*■; . -, • ... . , v. : ■■ Л . " * .* - - - ■-.' шт • '' e: : Л ■ -* « - Wortsl unter 10m m wand Linton 20 mm ward unten 30 mm wand unten 4Q mm wand Unten Perlltlsches Gusseiser», 0,3% Impfmittel A v. ' ' *' i>. •>'; , :' " * -, - : * v , - ■ *.. * Л ■ - - ■ » ■■ 1 *."*•' V* -, . * I.,- . Ч>■ 'v'v'V1*>*-• -, /v • Г • . ' " . ■ ' «*. -,..•„"; • .*. :*:"'.-. Л'.ч ■vv'"-;-;.'.7':.^:' •'.. ' ... j . i : ■;■ •• V . ^^ i",. : ■ . *■ *■■. ' ' i* . . . - У - . VViafe) Linieri tQm m Wand Unten 20 mm Wand Unten 30 min Wand Uri tei 11) mm Wand Unten Perlltisches Gusselsen, 0,4% Impfmittel A ' Л ; ' ■ к . v v * , « ■ ■ • ■ • ». * » - * » 1 - 4 • ' .•* -" 5 • . . • , ■ "i* • ' 1 t* . ■ * Л . - ' ■ - i г"-.'f.. .*-..;. . ■»■/*.'■*■■•'■-.i'.i ш-тщ ■ • . -.и Wijrfe) Unter 10mm Wand Unten 20 mm Ward Unten 30 mm Wand Unten mm Wand Unten Perlltisches G u s s eisen, 0,2% Impfmittel В ; -*■ ■ • * * ■• • » *' * . ». * .V i»,.' • • . • ' •• •• - ■ - - ■ • t i. ' л i.*'""* V ,.■ 1 ■ * * * " .'.■>• 'Vi ■ » « •V- V,- • ;. Würfel Unlen 10mm Wand Untori 2C1 mm Wand Untori $0 mm Wand Urten 40 mm Wand Urlar Perlitisches C u ss eis en, 0,3% Impfmittel Б t -*.*■■' . >.' ■ Iii l'I . '+'■ * '. * \ - " ; ' ■.-:,** ■ ' • • « . ' V 11 j, ' .. . •• v;,i i , ■ • 1-* V . V ■ ■ • 3 : —; : * ■ ',* " t * , . s ■ - - » L * » * . ■ * • .- V - л"» • ' • *. .. ■ ■■ . t£ - Wurfe! Unten 10mm Wand Unten ÌD mm Wand Unten 3D mm Wand Unten 40 mm wand Unten Perlitisches Gusseisen, 0,4% Impfmittel В -t i-i -1 .r - . J- * « . » » .. ♦ * - , ' t : ■. *. , . . * ' • - : " ' . . " "т » • , f t ' . 1 ' m , S • ... m' 9 » »* ' t v . ; * -» Würfel unten 10mm Wand Urten 20 mm Wand Unten 30 mm Wand Urtar 40 mm Wand Unten Figure 21. Pearlitic ductile iron, inoculated with inoculante A and В RMZ-M&G 2005, S2 At ferritic cast iron the dependence of the number of spheroids from cooling rate is remarkably lower. Here, too, the surface component does not show a uniform relationship between wall thickness or concentration of inoculant. At both types of iron an inoculant concentration of 0.3 % results to an optimum number of spheroids. At higher inoculant concentration the number of spheroids decreases again. A and B. Inoculant В tends a little to a lower number of spheroids. On increasing addition of inoculant the number of spheroids decreases at pearlitic cast iron. At ferritic cast iron inoculant A shows the same tendencies like pearlitic cast iron, but at a remarkable lower number of spheroids. At ferritic cast iron and at different inoculant additions inoculant B shows contrary behaviour to the use with pearlitic cast iron. The number of spheroids increases with increasing inoculant quantity. Stepped wedge Giving same inoculant at same concentrations the quicker type of cooling shows a remarkable higher number of spheroids (Fig. 23). Beginning at a wall thickness of 30 mm this tendencies efface a little and are not longer clearly. No basic difference is to recognize between inoculant On increasing inoculant addition the surface component of spheroids decreases at both inoculants. For inoculant A there is no difference to recognize for all wall thicknesses with an inoculant addition of 0.4 %, and for inoculant В the surface component decreases even more. For both inoculant quantities the influence of wall thickness is shown. The lowest wall thickness has the highest surface component, Pertitisches Gusseisen, Sphgrolithenzahl [I/mm! Ferritisches Gusseisen, Sphirelithenzahl [l/mm1] 10 mm 20 mm 30 mm 40 mm Würfel 10 mm 20 mm 30 mm 40 mm Würfel - 0,2% 0,3% -*— 0,4% 10 mm 20 mm 30 mm 40 mm Würfel 10 mm 20 mm 30 mm 40 mm Würfel Wandstärke -0,2% 0,3% 0,4% Perlltlsches Gusselsen, Flächenantell Sphärollthsn [%] Impfmittel A_Impfmittel В 10 mm 20 mm 30 mm 40 mm Würfel 10 mm 20 mm 30 mm 40 mm Würfel - 0,2% -■- 0,3% 0,4% Ferrltlsches Gusselsen, Flächenantell Sphärollthsn [%] 10 mm 20 mm 30 mm 40 mm Würfel 10 mm 20 mm 30 mm 40 mm Würfel Wandstärke - 0.2% -Ш- 0,3% -*- 0.4% Figure 22. Graphite formation at test castings in plate design. which equally lowers down above 30 mm up to 50 mm. On ferritic cast iron the relation within wall thicknesses are not as unequivocally as at pearlitic cast iron, and the wall thickness of 30 mm seems to have a special position. The effect of inoculant A and В clearly differs on pearlitic and ferritic cast iron. It has been shown that the thermal diffusivity as a value for cooling of a casting is very different within the different foundries, and by that it has to be clearly defined at each foundry as a decisive value for the solidification simulation. Variations of more than 30 % between the values of the different foundries are the rule. At the tests for definition of the influence of different variable material values to the graphite precipitation in form of: • number of spheroids, • graphite quantity. It has been shown that these values cannot alone give sufficient information about the quantity of eutectic graphite and by that about volume deficits or a shrinkage volume to be expected at solidification. To be able to predict shrinkage volume by means of: • physical values which define the solidification process, • variable mould material values which define the solidification process, • variable material values which define the solidification process. It is necessary to get - on the side of material - one or more new quantitative parameter, which have to result from thermodynamic criterions of solidification. Ferritisches Gussalsen, Sphflrollthendchte PerlKlsches Gusselsen, Sphgrollthendlchte A 0,3% A 0,4% B0,3% BO.4% ■ 10mm e 20mm ^^30mm " 40mm 50mm Femtischee Gusseieen, Sphärolithen Flächenanteil ■ 10mm ^^20mm ^^ 30mm ^^ 40mm ^^ 50mm Peilltlsches Gussdsen, Sphärolithen Flächenanteil I * 10mm Ш 20mm 30mm ^^ 40mm ^^ 50mm] Figure 23. Graphite formation on casting of stepped wedges. RMZ-M&G 2005, S2 Ъ) Filled in with resin bonded mould material Figure 24. Construction of pattern plate c) With mould box and charged Figure 2S. Simulation of mould box attributes Part 4: Influence of mould and mould material Besides the metal wall movement (Part 2) the mould wall movement (the attributes of the mould) is the second main influence factor that has an effect on "apparent shrinkage". The dependence of the mould wall movement results of: • type of mould material, • composition of mould material, • compaction, • mould box attributes. Influence of mould box attributes Four test cubes were positioned on a square pattern plate, each two of them with a modulus of 1.7 and 1.6 (Fig. 24). Each cube was equipped with the corresponding exothermic feeder. a) Without mould box By using this pattern equipment there were formed and cast three moulds, under simulation of different mould box attributes. All the moulds were made of a clay bonded mould material (sand F32, 8 % bentonite A and ~ 2 % water) and compacted with a pneumatic hand rammer. There was used spheroidal graphite cast iron SR = I as casting material: • The first mould was poured without mould box (Fig. 25a). There was used a conical wooden frame during ramming. The frame was removed after compaction of the mould. • The second mould was compacted by means of wooden frame, too. After removing this wooden frame there was set a metal frame around the mould. To increase the stability of the mould resin bonded mould material was filled in between mould and metal frame (Fig. 25b). • To achieve an even higher mould stability it was chosen a rigid mould box for the third mould and the mould was charged (Fig. 25c). There were taken photos from the castings with sprue, runner and feeder as well as from the actual cubes. The mass of feeders and castings (cubes) were graphically evaluated, just as the surface of the cubes, sawed through the midst. The evaluation of the surfaces should document the size of the enlargement of the three different moulds. For the purpose of comparison the contours of the sawed cubes were copied on a sheet of paper, scanned into the computer and calculated. The surfaces shown above were defined by means of the average values of the two indentic cubes of a mould. Figure 26 shows the changes of the sawed cube surfaces for the three cast moulds in dependence of the "mould box stability". The surfaces decrease on increasing stability of the mould, and there is to remark no difference within their behaviour between both moduli. Fläche der Würfel M1,7 11000 10900 10800 г4 I 10700 ~ 10600 10500 10400 — 10771 10939 ohne Formkasten mit mit Formkasten hangebundenem Formstoff Surface of cubes Ml,7 Fläche der Würfel M1,6 9250 9225 9212 9050 8991 8900 ohne Formkasten mit harzgebundenem Formstoff mit Formkasten Surface of cubes Ml,6 Figure 26. Surface of the cubes (Ml,6 and Ml,7) Masse der Würfel M1,7 In [g] 7830 7780 7730 [g]7680 7630 7580 7530 7788 7720 I 7565 I ohne Formkasten mit Formkasten harzgebundenem Formstoff Mass of cubes Ml,7 (g) Masse der Würfel M1,6 in [g] 6040 6020 [g] 6000 5980 5960 5940 ohne Formkasten mit heizgebundenem Formstoff mit Formkasten Mass of cubes Ml,6 (g) Figure 27. Mass of the cubes (Ml,6 and Ml,7) By evaluating the mass of cubes and feeders there should also be defined the size of the enlargement of the three moulds. For the purpose of comparison the feeders were sawed off, and feeders and cubes were separately weighed. Figure 27 shows the mass of the cubes (M1.6 and 1.7) and Figure 28 shows the total masses (M1.6 and 1.7) of the three cast moulds. Just like the evaluation of the surfaces the masses shown in the figures were defined by means of the average values of the two identic cubes of a mould. On the X-axle there are shown the three moulds with the different box attributes, on the Y-axle there are shown the masses (g). The mass of the cubes decreases on increasing stability of the mould, and there is to remark no difference within their behaviour between both moduli. The evaluation of the mass of cubes clearly shows, that the mass of the cube or the enlargement of mould is the smallest one by using a rigid mould box. The difference of enlargement is much less between the moulds without any mould box and the one with resin bonded mould material filled in between mould and metal frame. Gesamtmasse M 1,6 Figure 28. Mass of the cubes and feeders First of all, you would assume that the cast quantity of iron or the total mass is identical for all the three moulds. But the representation of total mass (Fig. 28) shows that the mould enlarges already while pouring the cast iron into the mould. In this connection it becomes clear, that the enlargement is the smallest one by using a rigid mould box. This test makes clear that the mould box attributes have influence on the mould wall movement. By comparing the differences between the masses extending of mould wall movement it is obviously mould wall movement can unequivocally define the feeding behaviour or the quality of a casting. The mould box attributes tested in this experiment are relative easy to stabilize in practice or should always be constant on a plant. The influence of type of bentonite / type of mould material The pattern equipment or the arrangement of the cubes and the pouring system of the preceding tests have been taken over, and the cubes and the pouring system have been fixed Figure 29. Pattern equipment on a pattern plate, which could be installed in a moulding machine. As the distance of the cubes to the box border is important for the valuation of the influence of the mould wall movement, this distance has been the same at each cube (75 mm) (Fig. 29). Three different types of bentonite were used. The relevant mould material attributes were measured and stated for each mixture, but they are not shown in this short version. There were produced four moulds with different contents of bentonite (5, 7, 9, 11 %) from each type of bentonite. There was added coal dust to each mould material Kantenvolumina M1,7 — M 1088000 £ 1004000 " 1082000 1080000 1078000 1078000 1074000 1072000 p — 1083050 1084114 — — - 1077787 - _П— u □ + + Bentonit [%] J? □ BentonltA □ Bentonit В у Bentonit С Figure 30. Edge volumes mixture, the quantity was half of bentonites's content. Just like the preceding tests silica sand F32 was prepared with the different bentonite types in a turbo-mixer. The cast iron (SR = I) was melted in a medium frequency coreless crucible induction furnace. In opposite to the preceding tests the cubes were not sawed and their surfaces were not evaluated, but the cube-edges were measured and the volume was calculated. The evaluation of the edge-volume (Fig. 30) shows that for the three types of bentonite the average values of volume of the cast cubes was bigger than those of the original pattern volume, whereby there was no large difference between the three average values of edge-volume. It can be assumed that the enlargement of a mould is not only defined by bulge of surfaces, but also by the enlargement of the "edge-frame" of the mould. Influence of the type of bentonite The influence of the type of bentonite to the enlargement of the mould cavity is clearly to recognize in the evaluation of the mass of cubes (Fig. 31). The enlargement of mould cavity is the biggest one at type A and the smallest one at type C, whereby the difference of enlargemet of mould cavity is not as large between bentonite type В and C as to bentonite type A. The unequal behaviour of the three types of bentonite (Fig. 31) is hardly to interpret by means of the methylene blue value, as a comparison of these values is only possible for clay-types of the same origin. To give reasons for this behaviour an exact analysis of the attributes of the three bentonite-types would be necessary. Influence2of2bentonite2content The influence of the bentonite content to the enlargement of mould cavity is not easy to interpret for the three tests (Fig. 31-33). Masse der Würfel/ M1,7 Bwitonit [%] □ BentonltA □ Bentonite в Bentonite Figure 31. Masses of cubes with M 1.7 Masse der Speiser/ M1,7 » 411 410 410 414 - Ш 273 ■ 278 293 -H—- 5% 7% 9* 11% 5% 7% 9% 11% 5% 7% 9% 11% Bentonit [%] □ Baritoni!A □ Bentonite □ Bentonite Figure 32. Masses of feders An important influence of the bentonite content to the enlargement of the mould cavity is not recognizable at the three cast tests. But it has to be considered that the mechanical properties increase, too, with increasing bentonite content. But on all cast tests this overservation could not be connected with the enlargement of mould cavity. To achieve the respective mould condition of each mixture at different bentonite contents each of the mould material mixtures need a different quantity of water. This observation is easy to recognize, too, within the mould material analysis, and it directs to the fact that - besides the influence of the type of bentonite - the compactability (water content) seems to be the determining parameter for the size of the mould wall movement. "Apparent shrinkage" The graphic evaluation of the total masses (Fig. 33) and the mass of the feeders (Fig. 32) shows that already while pouring, the moulds enlarge in different ways due to the ferrostatic pressure and the casting heat. Dependent bentonite-type the mass of the mould of bentonite-type С is a little smaller than those of bentonite-types A and B. The enlargement of mould cavity and - as a consequence - an increase of cube mass is mainly marked at bentonite-type A. Here it flows the most material from feeder for compensation of mould space enlargement, whereby at type of bentonite B at approx. same total mass (Fig. 33) the mass of the cubes are clearly less, but therefore the mass Gesamtmasse/ M1,7 7250 L-l—UJ—UJ——— , I , I , I I , I::; I , 1**1 , I , I : I , I I 5% 7% 9% 11% 5% 7% 9% 11% 5% 7% 9% 11% Bentonit [%] □ Würfel Bentonit A □ Speiser Bentonit A □ Würfel Bentonit В □ Speiser Bentonit В □ Würfel Bentonit С в Speiser Bentonit С Figure 33. Total mass and partition to feeder and cube. of the feeders are clearly larger than it is at bentonite type A. The mould space enlargement was smaller at this type of bentonite В with the consequence that less material from feeder was necessary for compensation of mould space enlargement. At same type of feeder bentonite В has a greater security against "apparent shrinkage". The enlargement of mould cavity at each type of bentonite does not show any tendency on change of bentonite contents. As the influence of bentonite contents to the mould wall movement is not so important, it was taken an average value of eight cubes (M1.7) of each type of bentonites, which were produced with different contents of bentonite. At same time it was taken an average value of the corresponding eight compactabilities measured of the mould sand mixtures. Figure 34 shows the average value of the mass of the cubes (M1.7) for each type of bentonite above the average value of the compactabilities. The influence of the type of bentonite to the enlargement of mould cavity is clearly to see by means on Figure 34. By comparing the mass of cubes of the bentonite-types В and С the difference of masses is more than 60 g. Between type A and С there is even a difference of nearly 200 g. This is approx. 3 % of the cube-volume, this value could absolutely define the quality of a casting. As a consequence of this test it becomes clear, that the type of bentonite and the compactability (water content) are two important parameters that define the enlargement of mould cavity and thereby define the "apparent shrinkage". Masse der Würfel (Mittelwert)/ M1,7 7550 7500 7450 7400 7350 7300 7542 7411 7348 39 □ Benton it А 45 Verdichtbarkeit [%] □ Bentonit В 49 □ Bentonit С Figure 34. Average values of mass of cubes as function of average values of compactabilities of moulds Summary At first several physical values defining the solidification process are collected. These physical values are characterised as variable moulding material values after that. Material values defining the solidification process are afterwards described and than their influence on feeding behaviour by chemical composition explained. The real feeding demand of a casting is described by influence of the mould and than the idea of the "apparent shrinkage" is explained. With thermal analysis results an assessment tool for quality and solidification structure of cast iron. Feeding behaviour of a melt can be specified by this method. Feeding is a transport phenomenon and there are several different feeding mechanisms during solidification time. It is shown that feeders are only active till 20 - 30 % of the solidification time. This report will be continued. References Koppe, W.; Engler, S. (1962): Giesserei 49, Nr. 10/11, S. pp. 26S-306. Engler, S.; Dette, M. (1974): Giesserei 61, Nr. 26, S. pp. 769-774. Engler, S.; Wojtas, H. J.(1979): Giessereiforschung 31, Nr. 1, S. pp. 37-44. Feichtinger, H. K.(1980): VSE-Tagung 22.01.1980 „Methoden der Qualitätsprüfung von GraugussSchmelzen" (Institut für Metallurgie der ETH Zürich). Autor's Index, Vol. 52, No. 4 Brenčič Mihael mbrencic@geo-zs.si 669 Bricelj Mihael mihael.bricelj@nib.si 661 Cenčur Curk Barbara barbara.cencur@irgo.si 661 Janža Mitja mitja.janza@geo-zs.si 737 Jarc Simona simona.jarc@ntfgeo.uni-lj.si 697, 711 Kugler Goran goran.kugler@ ntf.uni-lj.si 753 Mirtič Breda breda.mirtic@guest.arnes.si 697 Perus Iztok iperus@siol.net 753 Ratej Jože joze.ratej@geo-zs.si 669 Terčelj Milan milan.trcelj@ ntf.uni-lj.si 753 Trček Branka branka.trcek@geo-zs.si 685 Turk Rado rado.turk@ntf.uni-lj.si 753 Verbovšek Timotej timotejverbovsek@ntfgeo.uni-lj.si 723 Wojtas Heinz-Josef hk225wo@uni-duisburg.de 765 Autor's Index, Vol. 52 Abu Zeid Mahmoud 328 Abu-Zreig Majed majed_abuzreig@yahoo.ca, majed@just.edu.jo 172 Achiari Hendra d02sc191@ynu.ac.jp 173 Adeleye Mutiu A. 123,127 Aharonov E. 251 AhmadianR. 281,354 Akbulut Aydogan 289 Al Bakri Dhia 313 Al Qubaisi Nawal 328 Alajbeg Anđa andja.alajbeg@ina.hr 115 Alary Claire 371 Albayrak Ismail ismail.albayrak@epfl.ch 9 Albertazzi Sonia 284 Aleksandryan Anahit analeks@freenet.am 174, 175 Al-Enezi Eqbal 209 Al-Ghadban Abdul Nabi 209 Al-Kaabi Namaa 327 Aljinović Dunja daljin@rgn.hr 581 Allah2Abd samyabdallah@hotmail.com 171 Allott Tim 320 Almeida Paola p.almeida@ismar.cnr.it 176 Alsharhan2Abdul2Rahman 327 Andersen Frede 0 foa@biology.sdu.dk 177, 179, 247 Andersen H. E. 178, 265 Andersen Troels tandersen@biology.sdu.dk 179, 264 Ansaloni Ivano 309 Antonelli2 Christelle 25 Antonić Jan 235 Arheimer B. 264 Armeanu A. 203 Asplund Lillemor 294 Auvray2Franck 63 Auvray Isabelle Isabelle.auvray@limos.uhp-nancy.fr 182,213,286 Ayyoubzadeh Seyed Ali ayyoub@modares.ac.ir 370 Bacani Andrea abacani@rgn.hr 115 Badr2Nadia2B. 277 Badura Hannes h.badura@tugraz.at 183, 257, 331 Baharuddin2 Norshidah Bajt Oliver 81 Ballantine Deborah J. 185 Balogun Saka A. 123 Banasiak Robert Robert.Banasiak@ugent.be 186 Baneschi Ilaria i.baneschi@igg.cnr.it 187 Barcellos Roberto L. rlb@usp.br 188 Barešić Jadranka 241 Bargant Stanislav 115 Barnsley Michael J. 362 Barthe Jean-Francois 356 Bartholini G. gabriella.bartholini@fg.ismar.cnr.it 342, 343 Bass Jonathan A. B. jabb@ceh.ac.uk 189, 200, 201, 367 Baudu Michel 63 Behrendt H. 264 Beldowski Jacek hyron@iopan.gda.pl 5 Bellucci Luca Giorgio luca.bellucci@bo.ismar.cnr.it 218, 309, 322 Belzunce Maria Jesüs jbelzunce@pas.azti.es 190 Bendixen Tine 247 Berthelin Jacques 182, 286 Berto D. d.berto@icram.org 191 Bertola Paolo 229 Biberhofer Johann Hans.Biberhofer@ec.gc.ca 192 Bidin K. 193 Bierl Reinhard bierl@uni-trier.de 222 Bignert Anders 294 Billon Gabriel 219, 220, 283, 296, 356 Blake William H. william.blake@plymouth.ac.uk 212, 297, 330, 362 Blažič Andrej Andrej.Blazic@rlv.si 495 Boers Paul P.Boers@riza.rws.minvenw.nl 237, 264 Boldrin Alfredo alfredo.boldrin@ismar.cnr.it 284 Bole B. 1 Bonardi Maurizio m.bonardi@ismar.cnr.it 176 Bonté Philippe 371 Borja Angel 190 Boukhary Mohamed 327 Bozau Elke 259 Brancelj Anton anton.brancelj@uni-lj.si 287 Brenčič Mihael mbrencic@geo-zs.si 549, 669 Briansó José Luis 241 Bricelj Mihael mihael.bricelj@nib.si 661 Brizzotti Marizilda M. mariz oceano@yahoo.com.br 188 Bura-Nakić Elvira 196 Butler Patricia J. Callegari Giovanni Calligaris Chiara Capellacci Samuela Casper Peter Castelli Alberto Castro Raùl Catarino Joana B. Catona Francesco Chan Wai-ying Charlton Murray Chazal Philippe M. Choy Satish C. Church Tony Ciglenečki Irena Cioni Roberto Clarisse O. Clarke Stewart J. Cokgor Sevket Collins Adrian L. Cooper Michelle Cotton Jacqueline A. Covelli Stefano Cutter S. L. Czerniak Katarzyna Carman Magda Cenčur Curk, Barbara Cermelj Branko Cosović Božena da Silva José Figueiredo Daneu Nina Das D. K. Dautović Jelena David E. Dawson Alistair de Camargo Plinio B. De Vittor Cinzia Dedkov A. P. Deeks Lynda K. Defloor Griet Deloffre J. Deluchat Véronique 340, 341 308 calligar@univ.trieste.it 176 291 364 castelli@discat.unipi.it 309 190 111 308 194 278 63 254 313 irena@irb.hr 195, 196, 283, 373 187 296 stewart.clarke@english-nature.org.uk 197 cokgor@itu.edu.tr 9, 13 185, 248 michelle.cooper@jcu.edu.au. 198 jacot@qmul.ac.uk 189, 197, 200, 201 covelli@univ.trieste.it 17, 55, 176 238 key-la@o2.pl 202 magda.carman@geo-zs.si 607 barbara.cencur@irgo.si 661 cermelj@mbss.org 1, 91 195 jfs@ua.pt 111 nina.daneu@ijs.si 429 dkdasl231@rediffmail.com 312 283 david@icsi.ro 203 310 pcamargo@cena.usp.br 188 devittor@univ.trieste.it 17 dedkov@mail.ru 204 301 186 296 63 Denis Lionel Lionel.denis@univ-lillel.fr 205, 315 Dervarič Evgen evgen.dervaric@rlv.si 485 Desroy Nicolas 205 Dewhurst Rachel E. 344 Didier Bourlcs bourles@cerege.fr 319 Dinelli Enrico dinelli@ambra.unibo.it 206, 343 Dobnikar Meta meta.dobnikar@ntfgeo.uni-lj.si 397, 429 Doerr Stefan H. 212 Dolenec Matej matej.dolenec@ s5.net 397, 429, 437 Dolenec Tadej tadej.dolenec@ntfgeo.uni-lj.si 115, 397, 429, 523, 523 Dolinar Bojana bojana.dolinar@uni-mb.si 419 Dong Jingmei 365 Donini Filippo 206 Droppo Ian G. Ian.Droppo@ec.gc.ca 263 Duck Robert W. r.w.duck@dundee.ac.uk 111 Duffa Céline 25 Dyrynda Peter 362 Eckelhart Alexandra office@w-summer.org 207 Egorov I. E. 21 Ekholm Petri petri.ekholm@ymparisto.fi 208, 270 El-Sammak Amr asammak@kisr.edu.kw 209, 327, 328 Ericsson Ulla 294 Evans Martin 320 Eyrolle Frédérique 25, 319 Fabbri Elena 206 Facchinelli Aurelio aurelio.facchinelli@unito.it 31,302 Faganeli Jadran jadran.faganeli@uni-lj.si 1, 17, 35, 55, 81, 103 Faghihirad S. 210 Faithful John W. 198 Fajon Vesna 75 Fang2Don 361 Farguell Joaquim 211 Farwing Victoria J. 257715@swan.ac.uk 212 Faure Pierre Pierre.Faure@g2r-uhp.nancy.fr 182, 213, 286 Fiesoletti2F. 342,2343 Fischer J. C. 296 Fischer2Jean-Claude 356 Ford Phillip 313 Förstner Ulrich u.foerstner@tu-harburg.de 215 Foster2I.2D.2L. 306 Foster Ian Ian.Foster@coventry.ac.uk 214, 310, 372 Franco Javier Fraternali Michaela Friberg Nikolai Frignani Mauro Frömmichen René Furtado Valdenir V. Gabelle Cedric Gachanja A. N. Galeano Ester Gallé Tom Galović Ines Gangol Pradeep Garbett-Davies Hannah R. Geertrui Uyttendaele Geller Walter Geraldene Wharton Giani M. Gibičar Darija Gierlowski-Kordesh Elizabeth Giordano P. Globevnik Lidija Gogus Mustafa Goharzadeh Afshin Golobočanin Dušan Golterman Han L. Gonsiorczyk Thomas Gosar Mateja Grace Michael R. Grace Mike Granger Steven J. Grisenti Paolo Gruca - Rokosz Renata Grujičić D. Gu Ji-Dong Guidi Massimo Guo Shenglian Gusarov A.V. Haag Ingo Halas Stanislaw Hanafi Sulfikar Hart Barry T. mauro. frignani@bo. ismar. cnr. it vfurtado@usp.br cedric.gabelle@ed.univ-lillel.fr agachanjah@yahoo.com gess@enet.com.np g.wharton@qmul.ac.uk patrizia. giordano@bo. ismar. cnr. it lidija.globevnik@guest.arnes.si golterman@wanadoo.fr gosar@geo-zs.si paolo.grisenti@ing.unitn.it jdgu@hkucc.hku.hk slguo@whu.edu.cn avgusarov@mail.ru ingo.haag@ludwig-wawi.de halas@tytan.umcs.lublin.pl sulfikar@sci.monash.edu.au 190 206 268 218, 284, 305, 309, 322 259 188 91, 219, 220 221 218 222 115 223, 224 362 41 259 367 191 71 273 191, 342 45 253 225 226 227 364 571 233 323 340 229 357 469 266, 230, 272 187 368 204,231 51 245 233 233 Hassell Kathryn 254 Haygarth Phillip M. 340, 341 He Mengchang 234 Heath Ester ester.heath@ijs.si 55, 235 Hebel Bernd bernd.hebel@env.ethz.ch 236, 252, 366 Heigerth Günther 183, 331 Hejzlar J. 264 Heppell Catherine M. c.m.heppell@qmul.ac.uk 200, 201, 238, 240, 367, 318, 329 Herzsprung Peter 259 Hines Mark E. mark_hines@uml.edu 239 Hintze Thomas 261 Hlkanson Lars 232 Hoare Daniel J. d.hoare@ucl.ac.uk 240 Hoffmann C. C. 265 Hollenkamp Carol 273 Horvatinčić Nada nada.horvatincic@irb.h 241 Humphreys2 Geoff 212 Hupfer2Michael hupfer@igb-berlin.de 242,2267 IanA. Sanders 367 Ibrahimpašić2 Haris 115 Illarionov2A.2G. 21 Imamaura Masahiro mima@criepi.denken.or.jp 243 Imberger Jörg imberger@cwr.uwa.edu.au 275 Inzelt György 196 Iroume2Andres 41 Isazadeh2S. Siavash_i@mehr.sharif.edu 244 Ishii2Takashi 243 Ivey2Greg 316 Jaćimović Radojko 71 Jacobs2Patric 360 Jacqueline2A.2Cotton 367 Jacques Berthelin 213 James J. Gareth 362 Jana2K. 312 Janson Anne-Laure 205 Janža Mitja mitja.janza@geo-zs.si 737 Jarc Simona simona.jarc@ntfgeo.uni-lj.si 697, 711 Jarvie Helen, P. 341 Jean-Luc Potdevin 220 Jeanneau Laurent 182, 213 J^drysek Mariusz-Orion morin@ing.uni.wroc.pl 245, 246 Jensen Henning S. hsj@biology.sdu.dk 177, 247, 274 Jensen J. P. 264 Jones Paula A. Paula.A.Jones@exeter.ac.uk 248 Jonsson Per per.jonsson@itm.su.se 294 Josef Hejzlar hejzlar@hbu.cas.cz 237 Joynes Adrian J. 341 Jrrgensen Michael 177 Jung Goo B. 269 Jurkovšek Bogdan bogdan.jurkovsek@geo-zs.si 581 Kaasalainen H. first.lastname@fimr.fi 249,271 Kadir Selahattin 289 Kaligarič Mitja mitja.kaligaric@uni-mb.si 45 Kamau J. N. 221 Kanduč Tjasa tjasa.kanduc@ijs.si 67, 363 Karhu J. juha.karhu@helsinki.fi 271 Karlsson 0. Magnus magnus.o.karlsson@af.se 232 Kastelic Vanja vanja.kastelic@ntfgeo.uni-lj.si 447 Katolikov Victor shi@sp.ru 349 Katsman R. Regina.Katsman@weizmann.ac.il 251 Katterfeld Christian c.katterfeld@unibas.ch 236, 252, 332, 366 Kayaturk Yurdagül 253 Kazungu M. 221 Kefford Ben J. 254 Khachatryan Artak 174, 175 Khalili Arzhang 225 Kim Cha-kyum kick@namhae.ac.kr 255 Kim Jin H. 269 Kim Won I. 269 Kitanidis Peter K. 325 Kleeberg Andreas kleeberg@igb-berlin.de 256 Kniewald Goran 283 Helmut Knoblauch helmut.knoblauch@tugraz.at 183,257,331 Kobayashi Hiroshi 260 Koch Georg 115 Kocman David 71 Koff Tiiu koff@eco.edu.ee 258 Kokpinar M. Ali 253 Kolahdoozan M. 210 Kolar-Jurkovšek Tea tea.kolar@geo-zs.si 581 Kominkova Dana kominkova@lermo.cz 290 Kompare Boris boris.kompare@fgg.uni-lj.si 235 Kononets Mikhail 299 Autor's Index 809 Koponen Jorma koponen@eia.fi 87 Koschel Rainer 364 Koschorreck Matthias 259 Koshimizu Satoshi koshi@yies.pref.yamanashi.jp 260 Kosjek Tina 235 Kotilainen A. aarno.kotilainen@gsf.fi, 271 Kotnik Jože joze.kotnik@ijs.si 75 Koutsoyiannis Demetris 157 Kovač Nives kovac@mbss.org 81 Kozerski Hans-Peter kozerski@igb-berlin.de 261 Krajcar Bronić Ines 241 Kramar Sabina bintza@email.si 447 Krein Andreas krein@uni-trier.de 262 Krishnappan Bommanna G. Bommanna.Krishnappan@ec.gc.ca 263 Križanovski Andrej 75 Kronvang Brian BKR@DMU.DK 178, 237, 264, 265, 268 Kucukali Serhat 13 Kugler Goran goran.kugler@ ntf.uni-lj.si 475, 753 Kumar S. Sathish mcs@nitw.ernet.in 333 Kummu Matti matti.kummu@iki.fi 87 Kurtenbach Andreas 222 Lafitte R. 296 Lai Ho Yan 230 Lai Jessie 338 Lai M.Y. 266 Langone Leonardo 218, 284, 285 Lartiges Bruno 182, 213, 286 Laskov Christine laskov@igb-berlin.de 267 Lauridsen Rasmus B. r.lauridsen@ucc.ie 268 Lee Jong S. jongslee@rda.go.kr 269 Leeks Graham J. L. 185 Leermakers Martine 219 Lehtoranta Jouni 270 Leivuori Mirja 249, 270, 271, 274 Lenzi Mario 41 Leonidakis J. 1 Lespes Gaetane 352 Lewandowski Jörg 242 Li Jiaxi 272 Li Xingru 234 Lojen Sonja sonja.lojen@ijs.si 91 Lopez Dina L. 273 RMZ-M&G 2005, S2 Luke L. Warren 367 Lukkari Kaarina Kaarina. Lukkari@fimr.fi 249, 270, 274 Magda Cotman magda.cotman@ki.si 199 Magnoni Mauro m.magnoni@arpa.piemonte.it 31 Maksić Aleksandar 226 Malenković Vladimir vladimir.malenkovic@rlv.si 485 Mansuy Laurence 182, 213, 286 Marcie Christophe 352 Marini M. mauro .marini@an. ismar. cnr. it 342 Marion Andrea marion@idra.unipd.it 298 Mark Trimmer 367 Markič Miloš milos.markic@geo-zs.si 67, 276 Markus Meili Markus.Meili@itm.su.se 280 Marti Clelia cmarti@fich.unl.edu.ar 275 Massoud A.H. Saad 277 Mayer Janez Janez.Mayer@rlv.si 495 Mayer Tatiana tanya.mayer@ec.gc.ca 278 McDowell Richard W. richard.mcdowell@agresearch.co.nz 279 Mcintosh Jennifer 67 McKelvie ian D. 323 Medved Jožef jozef.medved@ntf.uni-lj.si 629 Menhaj M. В. 281, 354 Merten G. H. 307 Mesić Saša 95 Meyers Philip A. pameyers@umich.edu 282 Meys Joris F. A. 360 Mighall Tim 214, 306, 310 Mikac Nevenka 283 Miko Slobodan smiko@igi.hr 95, 115 Milačič Radmila radmila.milacic@ijs.si 75, 288, 352, 353 Milena Horvat milena.horvat@ijs.si 59, 71, 75, 119, 165 Milivojevič Nemanič Tadeja 352 Miljević Nada 226 Mirtič Breda breda.mirtic@guest.arnes.si 697 Miserocchi Stefano stefano .miserocchi@bo. ismar. cnr. it 284, 185 Mišič Miha miha.misic@geo-zs.si 1, 419 Mohamed Samy samyabdallah@hotmail.com 171 Montar Ges-Pelletier Emmanuelle Emmanuelle.montarges@ensg.inpl-nancy.fr 182, 213, 286 Morche David david.morche@geo.uni-halle.de 99 Mori Nataša natasa.mori@nib.si 287 Mostaert Frank 360 Autor's Index 811 Motelica-Heino M. m.motelica@brgm.fr 249,271 Mozzherin V. I. 204 Mrvar Primož primoz.mrvar@ntf.uni-lj.si 629 Münster Uwe 208 Muri Gregor Gregor.Muri@nib.si 141 Murko Simona 288 Mutlu Halim hmutlu@ogu.edu.tr 289 Muxika Inigo 190 Nabelkova Jana nabelkova@lermo.cz 290 Naden Pamela S. 340, 341 Nadezdić Milica goldus@vin.bg.ac.yu 226 Nair K. Shadananan nair59@yahoo.com 335 Nastro Gian Marco 291 Nazari A. 244 Neal Colin 341 Ngila, C. J. 221 Ni Jinren nijinren@iee.pku.edu.cn 292, 365, 369 Niculescu V. 203 Nielsen Kai Kai.Nielsen@geo.ntnu.nu 615 Nikolaevich Nicholas 293, 350 Nikolai Friberg NFR@DMU.DK 217 Nugegoda Dayanthi 254 Nützmann Gunnar 349 Nylund Kerstin 294 Obelić Bogomil 241 Ogorelec B. 1 Ogrinc Nives nives.ogrinc@ijs.si 17, 35, 55, 75, 103, 363 Oldham Carolyn c.oldham@cwr.uwa.edu.au 275, 316 Olsson Mats 294 Onodera Shin-ichi 127 Orel Boris boris.orel@ki.si 81 Osterc Andrej 295 Ouddane Baghdad 219, 296 Ovesen N. B. 265 Owens Philip N. Philip.owens@bbsrc.ac.uk 297, 301, 303 Packman Aaron I. a-packman@northwestern.edu 298 Paez-Osuna Federico paezos@ola.icmyl.unam.mx 305, 322 Parilkova Jana 137 Pavlovec Rajko rajko.pavlovec@ntf.uni-lj.si 597 Pedersen Morten Lauge MLP@DMU.DK 217, 265, 300 Pedersen Ole 179 Pekkarinen Jouko 208 Pempkowiak Janusz 5 Penna Nunzio cebiam@uniurb.it 291 Penna2Paolo 344 Pérez Victor 190 Perin Guido guiper@unive.it 176 Perrone Ursula ursula.perrone@unito.it 31, 302 Perus Iztok iperus@siol.net 753 Petticrew Ellen L. ellen@unbc.ca 297 ellen.petticrew@plymouth.ac.uk 303 Pezdic Jože joze.pezdic@ntfgeo.uni-lj.si 67, 276 Phillips M. R. m.phillips@sihe.ac.uk 304 Piani Raffaela piani@univ.trieste.it 17 Piazza Rossano piazza@unive.it 305 Pichler Srđan 195 Pitkänen Heikki 270 Pittam N. J. apy097@coventry.ac.uk 306 Plesnicar Ales ales_plesnicar@yahoo.com 403 Poleto C. cristiano_poleto@hotmail.com 307 Porto Paolo P.Porto@exeter.ac.uk 308 Pramanik B. Ray 312 Prevedelli Daniela prevedelli.daniela@unimore.it 309 Price N. Brian fsl@glg.ed.ac.uk 218 Proffitt Helen helen@nel25.free-online.co.uk 214, 310 Prohić Esad 95, 195 Punning Jaan-Mati mati@eco.edu.ee 311 Puskas Zsofia 196 Puste A. M. ampuste_bckv@yahoo.co.in 312 Queralt Ignasi 211 Raco Brunella 187 Rahman AKM 313 Rajchel Jacek 107 Rajchel Lucyna rajchel@geolog.geol.agh.edu.pl 107 Rajic Milica milica @polj.ns.ac.yu 314, 346 Rajic Milorad mrajic@ifvcns.ac.yu 314, 346 Ramelli Marion 182, 213 Rampazzo F. 191 Ranzinger Marko 67 Ratajczak Tadeusz 107 Ratej Jože joze.ratej@geo-zs.si 669 Rauch Mathieu mathieu.rauch@ed.univ-lille1.fr 205, 315 Ravaioli Mariangela 218, 309 Rawlins Barry G. 301 Read Deborah read@cwr.uwa.edu.au 316 Rečnik Aleksander aleksander.recnik@ijs.si 429 Reddy K. Surender 334 Reid Miriam K. m.warner@qmul.ac.uk 317 Rhodes Ed 212 Ricci Fabio 291 Ridgway John 362 Righetti Maurizio 229 Ritesh Arya 180 Roberts S.E.t ts.e.roberts@qmul.ac.uk 318 Roger S. Wotton 367 Roje Vibor 283 Rolland Benoìt benoit.rolland@irsn.fr 25, 319 Romano Stefania 218 Rosa Fernando 278 Ross Jan-Henning 332 Rothwell James j.j.rothwell@student.manchester.ac.uk 320 Rouillier Marie-Claude 286 Rozanov Alexander G. rozanov@sio.rssi.ru 299, 321 Rubinic Josip jrubinic@gradri.hr 161 Rudnicka Anna 293 Ruiz-Fernandez Ana Carolina caro@ola.icmyl.unam.mx 305, 322 Rumhayati Barlah barlah.rumhayati@sci.monash.edu.au 323 Rysin I. I. rysin@uni.udm.ru 21 Rzepa Grzegorz 107 Sacchi Elisa elisa.sacchi@manhattan.unipv.it 31 Sachsenhofer Reinhard F. sachsenh@unileoben.ac.at 276 Sadeghi Seyed Hamidreza sadeghi@modares.ac.ir 324 Saenger Nicole nsaenger@stanford.edu 325 Safina G. R. ggf@mail.ru 204, 326 Saim Nor'ashikin 184 Sakhkalyan Edward 174, 175 Sala Maria 211 Sanders Ian A. i.a.sanders@qmul.ac.uk 329 Sandru C. 203 Sangiorgi Francesca 206 Sarkkula Juha juha.sarkkula@environment.fi 87 Sasaki Jun 173 Sayer Aimee M. ggsayer@swan.ac.uk 193, 330 Sayer C.D. 240 Schindl Georg 207 Schmid Gerhard 51 Schneider Josef schneider@tugraz.at 183, 257, 331 Schneider Philipp philipp.schneider@unibas.ch 252, 332, 366 Schutter Jan de 360 Schwartz Renc schwartz@tu-harburg.de 261 Sekhar. M. Chandra mcs@nitw.ernet.in 333 mcs@recw.ernet.in 334 Serafimovski Todor seraft@rgf.ukim.edu.mk 397, 523, 523 Shakesby Rick A. 212 Sharma S. K. SKS105@rediffmail.com 336 Sharma U. C. 337 Sharma Vikas 337 Shen Pingping 230, 338 Shields Graham 198 Shin Joung D. 269 Silveira A. L. L. 307 Sion E. Roberts 201 Skarp Jonny 280 Skei Jens jens.skei@niva.no 339 Skrzypek Grzegorz 246 Smith Barnaby P. G. bpgs@ceh.ac.uk 340, 341 Smolej Anton smolej@ntf.uni-lj.si 643 Solaun Oihana 190 Soon Chen Sau 184 Spagnoli F. 342, 343 Specchiulli A. 342, 343 Spencer Kate L. k.spencer@qmul.ac.uk 238, 317, 344, 351 Sperle Marcelo sperle@uerj.br 176 Srivastava Ajai ajaibhu@yahoo.com 345 Stewart Ian 279 Stibilj Vekoslava 295 Stojiljković Dragica dragica@polj.ns.ac.yu 314, 346, 347, 348 Stojiljković Milica 347 Street Robert L. 325 Studnicka Markus 207 Stuparević Leonida 463 Sturman V. I. 21 Sukhodolov Alexander alex@igb-berlin.de 293, 349, 350 Sukhodolova Tatiana 350 Summer Wolfgang office@w-summer.org 207 Sun Liying 292 Sun Weiling sunweiling@iee.pku.edu.cn 365 Sun Yan 234 Suzuki2Keiko k.suzuki@qmul.ac.uk 351 Svendsen L. M. 178 Svetlana Pakhomova s-pakhomova@yandex.ru 299 Szramek Kathryn bkszramek@umich.edu 363 Sajn Robert robert.sajn@geo-zs.si 561, 571 Sčančar Janez janez.scancar@ijs.si 75, 288, 352, 353 Sekularac Gordana gordasek@tfc.kg.ac.yu 347, 348 Solar Slavko slavko.solar@geo-zs.si 615 Sömen Joksič Agnes 119 Sparica Marko mksparica@ igi.hr 115 Sparica-Miko Martina 95, 115 Strbac Nada 469 Surca Vuk Angela 81 Tabatabai M. R. M. mrmtabatabai@pwit.ac.ir, mrmtabatabai@yahoo.com 210 Taher-shamsi A. 281, 354 Tajrishi M. 244 Tallberg Petra petra.tallberg@ymparisto.fi 355 Tasev Goran tasevg@rgf.ukim.edu.mk 397, 523, 523 Tateda Yutaka 243 Tea Zuliani 353 Terasmaa Jaanus 311 Terčelj Milan milan.trcelj@ ntf.uni-lj.si 643, 753 Tesi Tommaso 285 Thoumelin Guy 356 Tiess Günther guenther.tiess@notes.uniloeben.ac.at 615 Tijani Moshood N. tmoshood@yahoo.com 123, 127 Toman Mihael Jožef 165 Tomaszek Janusz A. 357 Tomura Kenji 260 Trček Branka branka.trcek@geo-zs.si 685 Trimmer Mark 329 Trung Nguyen Nhu nguyen_nhutrung@hotmail.com 359 Turchetto Margherita 284 Turk Rado rado.turk@ntf.uni-lj.si 475, 643, 753 Ulaga Florjana Florjana.ulaga@gov.si 131 Valencia Victor 190 van der Perk Marcel m.vanderperk@geog.uu.nl 301 van Hullebusch Eric D. e_vanhullebusch@yahoo.fr 63 Vanlierde Elin 360 Verbovšek Timotej timotejverbovsek@ntfgeo.uni-lj.si 723 Verhoeven Ronny 186 Vershinin2Andrei 299 Vesely Jaroslav vesely.j@fce.vutbr.cz 137 Vetter2Thomas thomas.vetter@geo.uni-halle.de 358 Vilicić Damir dvilici@biol.pmf.hr 115 Vončina Maja maja.voncina@ntf.uni-lj.si 629 Vrabec Marko 67 Vreča2Polona polona.vreca@ijs.si 141,549 Wada2 Eitaro 245 Wagner2Erich erich. wagner@verbund. at 257 Wagner2Horst horst.wagner@notes.uniloeben.ac.at 615 Walling2Desmond2E. d.e.walling@exeter.ac.uk 185, 214, 248, 308, 318, 361 Walsh2Rory2P.2D. r.p.d.walsh@swansea.ac.uk 193,330, 362 Walter2Lynn2M. almwater@umich.edu 35, 76, 363 Wang Yingying 272 Wartel Michel 91 Wauer Gerlinde gerlinde@igb-berlin.de 364 Vdović Neda 220 Weibel Matthias 236 Weisshaidingera Rainer rainer.weisshaidinger@unibas.ch 236, 252, 366 Wendt-Potthoff Katrin 259 Westrich Bernhard bernhard.westrich@iws.uni-stuttgart.de 51 Wharton Geraldene g.wharton@qmul.ac.uk 189,197, 200 Will H. Blake william.blake@plymouth.ac.uk 193 Williams Erika L. cerikalw@umich.edu 363 Wojtas Heinz-Josef hk225wo@uni-duisburg.de 765 Woodward J. Jamie.c.Woodward@man.ac.uk 372 Xu Gaohong xugh@cjh.com.cn 368 Xu Nan 369 Yan Yan Yang Zhifeng 234 Yilmaz Levent lyilmaz@itu.edu.tr 145,153 Yun Sun G. 269 Zachoval Zbynek 137 Zagorc-Končan Jana 199 Zakiaghl Hajar zakiaghl@modares.ac.ir 370 Zangrando Roberta rozangra@ve.idpa.cnr. it 305 Zarris Demetris zarris@itia.ntua.gr 157 Zavšek Simon simon.zavsek@rlv.si 67 Zebracki Mathilde 371 Zhai, M. 221 Zhao Zhenye 230 Zhou Hong Zorana Zrnif zrzoka@yahoo.com 348 Zorzou Maroulia Zorzou@jbaconsulting.co.uk 372 Zuliani Tea 352 Zupančič-Kralj Lucija lucija.zupancic@fkkt.uni-lj.si 235 Zupančič Nina nina.zupancic@ntfgeo.uni-lj.si 403 Zwicker Gordana zsc.gordana@np-plitvicka-jezera.hr 161 Žibret Gorazd gorazd.zibret@geo-zs.si 561 Žic Vesna vzic@irb.hr 373 Zivkovič Dragana dzivkovic@tf.bor.ac.yu 463, 469 Živković Živan jmm@eunet.yu 469 Žižek Suzana 165 Žumer Jože 35 (NEW) INSTRUCTIONS TO AUTHORS (from Sep. 2003) RMZ-MATERIALS & GEOENVIRONMENT (RMZ- Materiali in geookolje) is a periodical publication with four issues per year (established 1952 and renamed to RMZ-M&G in 1998). The main topics of contents are Mining and Geotechnology, Metallurgy and Materials, Geology and Geoenvironment. RMZ-M&G publishes original Scientific articles, Review papers, Technical and Expert contributions (also as short papers or letters) in English. In addition, evaluations of other publications (books, monographs, ...), short letters and comments are welcome. A short summary of the contents in Slovene will be included at the end of each paper. It can be included by the author(s) or will be provided by the referee or the Editorial Office. * Additional information and remarks for Slovenian authors: English version with extended "Povzetek", and additional roles (in Template for Slovenian authors) can be written. Only exceptionally the articles in the Slovenian language with summary in English will be published. The contributions in English will be considered with priority over those in the Slovenian language in the review process. Authorship and originality of the contributions. Authors are responsible for originality of presented data, ideas and conclusions as well as for correct citation of data adopted from other sources. The publication in RMZ-M&G obligate authors that the article will not be published anywhere else in the same form. Specification of Contributions Optimal number of pages of full papers is 7 to 15, longer articles should be discussed with Editor, but 20 pages is limit. Scientific papers represent unpublished results of original research. Review papers summarize previously published scientific, research and/or expertise articles on the new scientific level and can contain also other cited sources, which are not mainly result of author(s). Technical and Expert papers are the result of technological research achievements, application research results and information about achievements in practice and industry. Short papers (Letters) are the contributions that contain mostly very new short reports of advanced investigation. They should be approximately 2 pages long but should not exceed 4 pages. Evaluations or critics contain author's opinion on new published books, monographs, textbooks, exhibitions .(up to 2 pages, figure of cover page is expected). In memoriam (up to 2 pages, a photo is expected). Professional remarks (Comments) cannot exceed 1 page, and only professional disagreements can be discussed. Normally the source author(s) reply the remarks in the same issue. Supervision and review of manuscripts. All manuscripts will be supervised. The referees evaluate manuscripts and can ask authors to change particular segments, and propose to the Editor the acceptability of submitted articles. Authors can suggest the referee but Editor has a right to choose another. The name of the referee remains anonymous. The technical corrections will be done too and authors can be asked to correct missing items. The final decision whether the manuscript will be published is made by the Editor in Chief. The Form of the Manuscript The manuscript should be submitted as a complete hard copy including figures and tables. The figures should also be enclosed separately, both charts and photos in the original version. In addition, all material should also be provided in electronic form on a diskette or a CD. The necessary information can conveniently also be delivered by E-mail. Composition of manuscript is defined in the attached Template The original file of Template is temporarily available on E-mail addresses: joze.pezdic@ntfgeo.uni-lj.si, joze.pezdic@guest.arnes.si barbara.bohar@ntfgeo.uni-lj.si References - can be arranged in two ways: - first possibility: alphabetic arrangement of first authors - in text: (Borgne, 1955), or - second possibility: [11 numerated in the same order as cited in the text: example'11 Format of papers in journals: Le Borgne, E. (1955): Susceptibilite magnetic anomale du sol superficiel. Annales de Geophysique, 11, pp. 399-419. Format of books: Roberts, J. L. (1989): Geological structures, MacMillan, London, 250 p. Text on the hard print copy can be prepared with any text-processor. The electronic version on the diskette, CD or E-mail transfer should be in MS Word or ASCII format. Captions of figures and tables should be enclosed separately. Figures (graphs and photos) and tables should be original and sent separately in addition to text. They can be prepared on paper or computer designed (MSExcel, Corel, Acad) Format. Electronic figures are recommended to be in CDR, AI, EPS, TIF or JPG formats. Resolution of bitmap graphics (TIF, JPG) should be at least 300 dpi. Text in vector graphics (CDR, AI, EPS) must be in MSWord Times typography or converted in curves. Color prints. Authors will be charged for color prints of figures and photos. Labeling of the additionally provided material for the manuscript should be very clear and must contain at least the lead author's name, address, the beginning of the title and the date of delivery of the manuscript. In case of an E-mail transfer the exact message with above asked data must accompany the attachment with the file containing the manuscript. Information about RMZ-M&G: Editor in Chief prof. dr. Jože Pezdič (tel. ++386 1 4704-633) or Secretary Barbara Bohar Bobnar, un. dipl. ing. geol. (++386 1 4704-630), Aškerčeva 12, Ljubljana, Slovenia or at E-mail addresses: joze.pezdic@ntfgeo.uni-lj.si, joze.pezdic@guest.arnes.si barbara.bohar@ntfgeo.uni-lj.si Sending of manuscripts. Manuscripts can be sent by mail to the Editorial Office address: • RMZ-Materials & Geoenvironment Aškerčeva 12, 1001 Ljubljana, Slovenia or delivered to: • Reception of the Faculty of Natural Sciences and Engineering (for RMZ-M&G) Aškerčeva 12, Ljubljana • E-mail - addresses of Editor and Secretary • You can also contact them on their phone numbers. These instructions are valid from September 2003 TEMPLATE The title of the manuscript should be written in bold letters (Times New Roman, 14, Center Name Surname1, .... , & Name Surnamex (Times New Roman, 12, Center) xFaculty of ... , University of ... , Address..., Country, e-mail: ... (Times New Roman, 12, Center) THE LENGTH OF FULL PAPER SHOULD NOT EXCEED TWENTY (20, INCLUDING FIGURES AND TABLES) PAGES (OPTIMAL 7 TO 15), SHORT PAPER FOUR (4) AND OTHER TWO (2) WITHOUT TEXT FLOWING BY GRAPHICS AND TABLES. Abstract(Times New Roman, Bold/Normal, II): The text of the abstract is placed here. The abstract should be concise and should present the aim of the work, essential results and conclusion. It should be typed in font size 11, single-spaced. Except for the first line, the text should be indented from the left margin by 10 mm. The length should not exceed fifteen (IS) lines (10 are recommended). Key words: a list of up to S key words (3 to S) that will be useful for indexing or searching. Use the same styling as for abstract. Introduction (Times New Roman, Bold, 12) Two lines below the keywords begin the introduction. Use Times New Roman, font size 12, Justify alignment. There are two (2) admissible methods of citing references: 1. by stating the first author and the year of publication of the reference in the parenthesis at the appropriate place in the text and arranging the reference list in the alphabetic order of first authors; e.g.: "Detailed information about geohistorical development of this zone can be found in: Antonijević (1957), Grubić (1962), ..." "... the method was described previously (Hoefs, 1996)" 2. by consecutive Arabic numerals in square brackets, superscripted at the appropriate place in the text and arranging the reference list at the end of the text in the like manner; e.g.: "... while the portal was made in Zope[3] environment." Results and discussion (Times New Roman, Bold, 12) Tables, figures, pictures, and schemes should be incorporated (inserted, not pasted) in the text at the appropriate place and should fit on one page. Break larger schemes and tables into smaller parts to prevent extending over more than one page. Conclusions (Times New Roman, Bold, 12) This paragraph summarizes the results and draws conclusions. Acknowledgements (Times New Roman, Bold, 12, Center - optional) This work was supported by the ****. References (Times New Roman, Bold, 12) Regardless of the method used, in the reference list, the styling, punctuation and capitalization should conform to the following: FIRST OPTION - in alphabetical order Casati, P., Jadoul, F., Nicora, A., Marinelli, M., Fantini-Sestini, N. & Fois, E. (1981): Geologia della Valle del'Anisici e dei gruppi M. Popera - Tre Cime di Lavaredo (Dolomiti Orientali). Riv. Ital. Paleont., Vol. 87, No. 3, pp. 391-400, Milano. Folk, R. L. (19S9): Practical petrographic classification of limestones. Amer. Ass. Petrol. Geol. Bull.; Vol. 43, No. 1, pp. 1-38, Tulsa. SECOND OPTION - in numerical order [1] Trček, B. (2001): Solute transport monitoring in the unsaturated zone of the karst aquifer by natural tracers. Ph.D. Thesis. Ljubljana: University of Ljubljana 2001; 12S p. [2] Higashitani, K., Iseri, H., Okuhara, K., Hatade, S. (199S): Magnetic Effects on Zeta Potential and Diffusivity of Nonmagnetic Particles. Journal of Colloid and Interface Science 172, pp. 383-388. Citing the Internet site: CASREACT-Chemical reactions database [online]. Chemical Abstracts Service, 2000, updated 2.2.2000 [cited 3.2.2000]. Accessible on Internet:. Povzetek (Times New Roman, 12) A short summary of the contents in Slovene (up to 400 characters) can be written by the author(s) or will be provided by the referee or by the Editorial Board. TEMPLATE for Slovenian Authors The title of the manuscript should be written in bold letters (Times New Roman, 14, Center) Naslov članka (Times New Roman, 14, Center) Name Surname1, .... , & Name Surnamex (Times New Roman, 12, Center) Ime Priimek1, ..., Ime Priimekx (Times New Roman, 12, Center) xFaculty of ... , University of ... , Address..., Country; e-mail: ... (Times New Roman, 12, Center) xFakulteta..., Univerza., Naslov., Država; e-mail: ... (Times New Roman, 12, Center) THE LENGTH OF ORIGINAL SCIENTIFIC PAPER SHOULD NOT EXCEED TWENTY (20, INCLUDING FIGURES AND TABLES) PAGES (OPTIMAL 7 TO 15), SHORT PAPER FOUR (4) AND OTHER TWO (2) WITHOUT TEXT FLOWING BY GRAPHICS AND TABLES. DOLŽINA IZVIRNEGA ZNANSTVENEGA ČLANKA NE SME PRESEGATI DVAJSET (20, VKLJUČNO S SLIKAMI IN TABELAMI), STROKOVNEGA ČLANKA ŠTIRI (4) IN OSTALIH PRISPEVKOV DVE (2) STRANI. Abstract(Times New Roman, Bold/Normal, II): The text of the abstract is placed here. The abstract should be concise and should present the aim of the work, essential results and conclusion. It should be typed in font size 11, single-spaced. Except for the first line, the text should be indented from the left margin by 10 mm. The length should not exceed fifteen (IS) lines (10 are recommended). Izvleček(TNR, B/N, II): Kratek izvleček namena članka ter ključnih rezultatov in ugotovitev. Razen prve vrstice naj bo tekst zamaknjen z levega roba za 10 mm. Dolžina naj ne presega petnajst (1S) vrstic (10 je priporočeno). Key words: a list of up to S key words (3 to S) that will be useful for indexing or searching. Use the same styling as for abstract. Ključne besede: seznam največ S ključnih besed (3-S) za pomoč pri indeksiranju ali iskanju. Uporabite enako obliko kot za izvleček. Introduction - Uvod (Times New Roman, Bold, 12) Two lines below the keywords begin the introduction. Use Times New Roman, font size 12, Justify alignment. All captions of text and tables as well as the text in graphics must be prepared in English and Slovenian language. Dve vrstici pod ključnimi besedami se začne Uvod. Uporabite pisavo Times New Roman, velikost črk 12, z obojestransko poravnavo. Naslovi slik in tabel (vključno z besedilom v slikah) morajo biti pripravljeni v slovenskem in angleškem jeziku. There are two (2) admissible methods of citing references - obstajata dve sprejemljivi metodi navajanja referenc: I. by stating the first author and the year of publication of the reference in the parenthesis at the appropriate place in the text and arranging the reference list in the alphabetic order of first authors; e.g.: 1. z navedbo prvega avtorja in letnice objave reference v oklepaju na ustreznem mestu v tekstu in z ureditvijo seznama referenc po abecednem zaporedju prvih avtorjev; npr.: "Detailed information about geohistorical development of this zone can be found in: Antonijević (1957), Grubić (1962), ..." "... the method was described previously (Hoefs, 1996)" 2. by consecutive Arabic numerals in square brackets, superscripted at the appropriate place in the text and arranging the reference list at the end of the text in the like manner; e.g.: 2. z zaporednimi arabskimi številkami v oglatih oklepajih na ustreznem mestu v tekstu in z ureditvijo seznama referenc v številčnem zaporedju navajanja; npr.; "... while the portal was made in Zope[3] environment." Results and discussion - Rezultati in razprava (Times New Roman, Bold, 12) Tables, figures, pictures, and schemes should be incorporated (inserted, not pasted) in the text at the appropriate place and should fit on one page. Break larger schemes and tables into smaller parts to prevent extending over more than one page. Tabele, sheme in slike je potrebno vnesti (z ukazom Insert, ne Paste) v tekst na ustreznem mestu. Večje sheme in tabele je potrebno ločiti na manjše dele, da ne presegajo ene strani. Conclusions - Sklepi (Times New Roman, Bold, 12) This paragraph summarizes the results and draws conclusions. Povzetek rezultatov in zaključki. Acknowledgements - Zahvale (Times New Roman, Bold, 12, Center - optional) This work was supported by the ****. References - Viri (Times New Roman, Bold, 12) Regardless of the method used, in the reference list, the styling, punctuation and capitalization should conform to the following: Ne glede na uporabljeno metodo pri seznamu citiranih referenc upoštevajte naslednjo obliko: FIRST OPTION - in alphabetical order (v abecednem zaporedju) Casati, P., Jadoul, F., Nicora, A., Marinelli, M., Fantini-Sestini, N. & Fois, E. (1981): Geologia della Valle del'Anisici e dei gruppi M. Popera - Tre Cime di Lavaredo (Dolomiti Orientali). Riv. Ital. Paleont. ; Vol. 87, No. 3, pp. 391-400, Milano. Folk, R. L. (19S9): Practical petrographic classification of limestones. Amer. Ass. Petrol. Geol. Bull.; Vol. 43, No. 1, pp. 1-38, Tulsa. SECOND OPTION - in numerical order (v numeričnem zaporedju) [1] Trcek, B. (2001): Solute transport monitoring in the unsaturated zone of the karst aquifer by natural tracers. Ph.D. Thesis. Ljubljana: University of Ljubljana 2001; 12S p. [2] Higashitani, K., Iseri, H., Okuhara, K., Hatade, S. (199S): Magnetic Effects on Zeta Potential and Diffusivity of Nonmagnetic Particles. Journal of Colloid and Interface Science 172, pp. 383-388. Citing the Internet site: CASREACT-Chemical reactions database [online]. Chemical Abstracts Service, 2000, updated 2.2.2000 [cited 3.2.2000]. Accessible on Internet:. Citiranje internetne strani: CASREACT-Chemical reactions database [online]. Chemical Abstracts Service, 2000, obnovljeno 2.2.2000 [citirano 3.2.2000]. Dostopno na svetovnem spletu: . Povzetek - Summary (Times New Roman, 12) An extended summary of the contents in Slovene (from one page to approximately 1/3 of the original article length). Razširjeni povzetek vsebine prispevka v Angleščini (od ene strani do približno 1/3 dolžine izvirnega članka). Number of SCI search (število SCI citatov).....................................................................185 Number of paper indexing in different bases (Število indeksiranih člankov v posameznih bazah) CA SEARCH - Chemical Abstracts (1967 - present).....................................................375 METADEX: Metal Science..............................................................................................135 GeoRef .............................................................................................................................125 Inside Conferences..............................................................................................................76 PASCAL..............................................................................................................................30 Energy Science and Technology.........................................................................................27 Aluminium Industry Abstracts............................................................................................18 Ei Compendex.....................................................................................................................13 EngineeredMaterials Abstracts.............................................................................................3 Analytical Abstracts..............................................................................................................1 FLUIDEX (Fluid Engineering Abstracts)............................................................................1 TULSA™ (Petroleum Abstracts).........................................................................................1 Indexing also in (number of indexing not yet available) (Indeksiran tudi v (števila vpisov še nimamo) Alloys Index, Bibliography and Index of Geology, Chemical Titles, IMM Abstracts and Index (Institution of Minimg and Metallurgy), INIS Atomindex, Metals abstracts Index, Nonferrous Metals Alert, Polimers, Ceramics, Composites Alert, Steel Alerty