ࡱ; HI  !"#$%&'()*+,-./0123456789:;<=>?@ABCDEFGJKLMNOPQRSTUVWXYZ[\]Root Entry FCompObjbWordDocumentObjectPool  FMicrosoft Word 6.0 Document MSWordDocWord.Document.6;  Oh+'0 D h   @d C:\WINWORD\TEMPLATE\NORMAL.DOT"THE KOSTANJEK LANDSLIDE IN ZAGREBBoboBoboܥe- ec:.a(((PPPJLLLjTi=DM PJP(PPJv:JTHE KOSTANJEK LANDSLIDE IN ZAGREB Bogdan Stani} and Ervin Nonveiller Geotechnical Department, Civil Engineering Institute of Croatia, Janka Raku{e Road, Zagreb, HR SYNOPSIS The large landslide Kostanjek on the southern slopes of the mountain Zagreba~ka Gora, in the western suburb of Zagreb was activated in 1963 after some 2.1*106 m3 of marl for the cement factory Croatia was excavated at the foot of the slope. The geologic features (sinclinal structure, faults, hydrogeologic conditions, extension of the "Tripoli" marl strata) which had an important role in the formation of the slide, as well as the way of marl exploitation by means of mass blasting, are presented in the paper. The Kostanjek slide involves an urbanized area of some 100 ha. It is estimated that a sliding mass of some 32*106 m3 is involved, with a maximum depth of 90 m. Sliding occurs in three levels. The displacements of the surface are 3 to 6 m. The excavation of marl was stopped in 1988, when a total of 5,3*106 m3 of material was excavated, and the magnitude of surface displacement per year has decreased. The Kostanjek slide is still active. It is estimated that a natural stabilization of displacements would take a long time, some decades. Possible measures for the stabilization of movements are analyzed, and preventive drainage with continuous observations of the slide is proposed. Introduction On level ground at the foot of the slope of the Kostanjek landslide, a cement factory was constructed in 1907 (Fig.1). It utilized marl excavated manually from the nearby slope, and limestone from higher horizons of the mountain Zagreba~ka Gora, as raw materials, which were initially transported to the factory by cableway, and later, partly underground, by a conveyor belt. Fig. 1 After World War II the consumption of cement increased rapidly and in 1954 the excavation of marl started at the toe of the slope about 200 m away of the factory buildings. Due to the increased cement production the excavation technology was changed to mass blasting, and the retarded millisecond firing was introduced in May 1963. In 1963 some damages were observed on the factory buildings. The end pillar of the cableway started inclining toward the building, and some damages were observed on other buildings at the foot of the slope (Fig.2). Fig. 2 Exploration boring was undertaken and core samples were tested in the laboratory. It was determined that the marl contained some montmorillonite and that samples tested in oedemeters were swelling. It was predicted that the stress relief from the excavated marl would cause raising of the surrouding soil surface by 15 cm. It was observed that the bench mark (1) on the workshop building (Fig.3) raised 4 cm from 15.10.1963 to 13.2 1968 (Pehnec, 1967). Fig. 3 Following this finding some measures were undertaken in order to prevent the consequences of further swelling. The cable railway was replaced by a conveyor belt, which partly went underground (Fig. 2). Shortly afterwards deformations occurred in the tunnel and the conveyor belt was damaged, which again was assumed to be caused by the swelling of the underlying marl. Troughout this period no consideration was given to horizontal displacements which also occurred. In order to increase the cement production it was decided at the beginning of the 70-es to build a new modern cement factory east of the existing one. Formal soil explorations were undertaken, but the experience with the swelling of the soil were not considered. The facilities of the new factory were designed and their construction was completed by the end of 1976. At that time substantial horizontal displacements were already observed at the foot of the slope where marl was excavated. The new power house, the fuel reservoirs and the conveyor belts were affected, whereas the old workshop building was destroyed by horizontal and vertical displacements of the ground. Fig.4 shows the displacement of bench mark (1) measured in mid 1975. It is evident that the tendency of raising was still linear with time and it amounted to about four times the computed value of swelling due to the excavation of marl near the factory buildings between 1954 and 1960. Evidently the swelling of marl was not the cause of the damages of the new factory facilities. Fig. 4. Further investigations were entrusted to the authors and a geologist, M. Ortolan. It was immediately obvious that the slope of the Kostanjek region was affected by a large slide the contour of which was clearly visible from its foot at the east, along the slope, to its foot at the west, as shown in Fig 5. Fig. 5. Since the excavation of marl for the new factory was intensified, the deformations of the slope increased progressively, as well as the damages to public and private buildings in this very interesting urbanized region of Zagreb. It was thus decided by the end of 1988 to stop the production of the cement factory after 5.3*106 m3 , i.e. 13 millions of tons of marl were excavated. Geologic and tectonic conditions Lithostratigraphic sequence The Kostanjek landslide developed on the NW slopes of the mountain Zagreba~ka gora. At the base of the slide region there are thick Mesozoic layers of the Upper Trias T3, consisting dominantly of dolomite and some limestone. Above them are Cenozoic layers belonging to Neogene (periods Miocene - Torton, Sarmat and Panon) (Ortolan, 1990.) The layers of the Upper Torton M22, more than 50 m thick, consist of dolomitic breccias and conglomerates, Litotamnic limestones, and limy sandstone. Above the Upper Torton there are layers of Lower Sarmat M31, about 120 m thick. Their lower 75 m consist of silty limy marl, whereas limestone and sandstone beds a more frequent with depth. The upper 12 m consist of thinly laminated silty shists, Tripoli layers (intechange of grey and white thin laminae of the same mineralogical composition) and thinly laminated silty marls, which continuously surface on a large area along the perimeter of the slide along its northern and western rim. In the laboratory it was determined that the residual shear strength of the Tripoli beds is very low. The deep slide surface follows these beds on the whole area. Above these layers there are layers of the Lower and Upper Panon 1M23, 2M23. The Lower Panon, about 5 m thick, consists of layered hard marl and limy marl with thin intercalations of sandstone. The Upper Panon, about 50 m thick in the central part of the sliding mass, consists of light grey marl with intercalations of sand, sandstone and clay. The shallower slide planes follow clayey beds within the Upper Panon. Structural and tectonic relations The region of the Kostanjek landslide is a syncline dipping toward east, and its axis is in the direction NNE - SWS (Fig 5). Three deep faults are significant as shown in Fig. 5. One fault is in the direction N-S on the eastern border, the second is in the direction NE - SW on the NW border, and the third is regional in the direction ENE - WSW along which the base beds sink deep in the valley of the Sava river. These faults delimit the area of the slide which did not completely develop in the Sarmat beds. It was observed on the boundaries of the marl excavation and on samples from borings that in the discontinuities of the beds there are highly preconsolidated clay beds with striated shiny planes, an evidence of shearing deformations (shallower sliding). Hydrogeologic relations The investigation has shown that the deep thick beds of the Upper Trias T3, consisting of dolomite and limestone, are permeable. From the exploration of some termal bore holes the permeability was estimated between 10-2 and 10-3 cm/s. These beds reach the surface about 400 m north of the upper border of the slide, and on the slopes of Zagreba~ka Gora, where plenty of rain and melting snow penetrate the ground. Permeability tests were carried out in exploration boreholes on the slide (KS-2, 2', 3, 4, 5, 6, 7 & 8, shown in Fig 8). The boreholes were equipped with piezometers. The results of measurement piezometric level are shown in cross sections, Fig. 14(b). In some piezometers the ground water level is above the ground surface, and in all of them it is close to the ground surface. The yearly precipitation in the region averages 1.000 mm, partly in the form of snow, which causes the raising of the ground water level in spring, and the soil becomes saturated to the surface. Geotechnical characteristics of the materials in the slide The undisturbed cores from the bore holes in the Lower and Upper Panon layers, especially from the Tripoli bed in which the deepest sliding plane developed, were tested in the laboratory. The grain size distribution and classification data were determined (Fig. 6), as well as the unconfined strength, the shear strength in drained slow tests in the triaxial tests and in direct shear tests, and the residual shear strength in the rotary apparatus. In addition, the mineralogical composition of the samples was determined. The results of these tests are shown in Fig 6 and in Tables 1 and 2. Fig. 6. Tables 1. Tables 2. The results of shear strength tests, which were verified in the stability analyses shown in the following section show that under the given geometric setting and hydrogeologic relations, the residual shear strength defined by the parameters cr = 0, f = 9o, brings the slide to a stable condition with FS=1.0. Interpretation of slide movement and description of slide Since it was evident that the dominant slide movements were on the deepest sliding plane, all the results of the existing surveying of bench mark displacements were reinterpreted. The results of aerophotogrammetric surveys taken over several years, show the total displacement of the surface of the slide. The measurements of benchmark displacements on some factory buildings for some periods, as well as on the benchmark network of the region within the slide area in the periods 78/79, 79/88 and 88/94 were included. The results shown in Fig. 7 have a mean error of mx=0.20-0.33 m, my=0.14-0.27 m in the horizontal plane, which is satisfactory, while the error of mz=0.28-0.39 in the vertical plane is not satisfactory. Fig. 7 The results of the aerophotogrammetric survey show that contours of the displacements can be constructed on the slide area, and they fairly well match the boundary of the slide area established in the field. The displacements of the ground surface from 1963 to 1985 are in the range of 3 - 5 m, which indicates that sliding occurs on several sliding planes. In 1988 several benchmarks were installed on the surface of the slide for geodetic observations of the movement of the slide, and it was possible to establish a link with previous measurements on bench marks 6, 7 and 12 (Fig.8). Fig. 8 The results of these geodetic measurements of the slide displacements shows that the displacements can be dividet into three groups according to their magnitude (Fig. 9), which confirms the geologic indication of the existence of three sliding planes: the deepest sliding plane is shown in Fig. 15, the middle plane is 27 m above, and the shallowest about 40 m above the deepest plane. The analysis of the displacement vectors in Fig. 9 allows the following conclusions: -the sliding on the deepest and middle planes generally occurs soutwards, with a slight deviation toward the East by 5-10o; -the sliding on the shallowest sliding plane is 25o West of South; similar conclusions follow from the geodetic measurements in the period 1979/88 (Fig.10); -considering all the measurement results it can be concluded that the displacement vectors consist of about 50% movement on the deepest sliding plain and the rest occurs on the two shallower planes. The analysis of the geodetic measurements on the old and new benchmark networks leads to the following conclusions: -from the period 73/76 to 88/94 the magnitude of displacement per year is decreasing (Fig. 11); -the interpretation of the aerophotogrammetriccally measured displacements from 1963 to 1988, and the geodetic measurements from 1988 to 1994 give a total movement on the deepest plane of 3.4 m (or 0.11 m/year) and 6,5 m on all planes (or 0.21 m/year); -the displacements on the deepest plane computed from the diagram in Figure 11 for the period from 1975 to 1994, amount to 2 m (or 0.105 m/year); this indicates that out of the total movement of 3.4 m from 1963 to 1994, 1,4 m was developed from 1963 to 1975 (or 0.115 m/year); since the movement computed from curve 4 in Fig. 11 would result in 1.7 m, the actual movement must be below curve 4, which means that the slide was not initiated by a progressive failure but by some other external cause (probably by the mass blasting in the quarry); further movements were caused by marl excavation, probably of varying intensity, and by unfavorable hydrologic conditions; -the extrapolation of the curve of the magnitude of displacement per year shows that it will take many decades before the movements stop. Fig. 9 Fig. 10 Fig. 11 Stability analysis Three-dimensional analysis procedure Stability computations of slope failures are generally carried out with solutions for two-dimensional problems although slope failures are very often three-dimensional. At present there is no reliable generally accepted method for three-dimensional slope stability computation. Therefore the method of computation adopted has to be carefully chosen (Stani} et al. 1991). Two dimensional solutions result in the lowest safety factors and they are commonly used for the stability computation of failures on very wide slopes. Whereas on localized failures (localized loading or unloading, systems of discontinuities etc.) they lead to conservative solutions. Under such conditions the critical three-dimensional shape of the slide body can be chosen by one of the solutions by Azouz et al., (1978, 1981), Baligh et al., (1975), Gens et al., (1982) for a circular failure plane, or by Hungr (1987), Hovland (1977) for failure planes of general shape, which is cousidered by the authors as conservative. The analysis of real slope failures depends on the stiffness of the soil. Undrained failures occur mainly in soft cohesive soils which exibit large deformations, and the failure surfaces are of a regular shape, so that the above mentioned solutions by Azouz, Baligh or Gens are acceptable. For drained failures the shape of the failure surface depends mainly on the soil structure (stratification, fissures etc.). Under such conditions the failure surface does not develop in some regular shape, but it follows the soil structure and discontinuities. In the case of the slip surface of a regular shape the 3D contribution is minor (Skempton, 1985). Uncomplete 2D methods of slices can not be adapted to 3D solutions for irregularly shaped slip surfaces, because their solutions are not even realistic for 2D cases. Complete 2D solutions (Chen et al., 1982a and 1982b, Xing, 1988) introduce limitations regarding the geometry (symmetric ellipsoidal shapes) which makes it difficult to shape the model. Stani} et al., (1991) have analyzed the existing 3D methods of analysis and concluded that for undrained slope failures it is more adequate to use complete 2D solutions and take into account the contribution of the end effects rather than completing the uncomplete solutions. The slope stability analyses which use the methods of limiting equilibrium, give the minimum factor of safety for rigid sliding bodies. Bishop (1955) has defined the safety factor as the ratio of the available resisting shear force on the slip surface and the mobilized force needed to obtain equilibrium. According to this definition the safety factor in 2D analyses is given by Stfi i FS=------- (1) Stf i Summing the resisting and active forces on the slices along the slip surface of the cross section, the safety factor results in S [(si - ui) tanfi + ci] li i FS=---------------------------- (2) Sti li i Equation (2) is basic for the computation of the 3D safety factor for the defined sliding body. Mihalinec et al.,(1991) have presented a 3D solution for the analysis of slope slides of a given shape. The sliding body is divided into elements, each of which is defined by a representative cross section. The resisting and active forces, and the 2D safety factors are computed for all cross sections, and they form, along with the width of the elements, the necessary data for the computation of the 3D safety factor. Generally each element will have a different safety factor, so that it is possible for them to have mutually different displacements, which is not allowed in the whole sliding body. Such mutual movements are prevented by activating shear forces on the contacts between the elements. These forces act in the direction opposite to the displacements which are potentially initiated by different safety factors. It is assumed that the mobilized shear forces between the adjacent elements act together with the resisting forces on the slip surfaces, and that the equilibrium is established on the whole slidig body, which is defined by the 3D safety factor FS3. The elements with a greater 2D safety factor transmit a part of their mobilized resisting force from their base, to the elements with a lower safety factor through the shear forces activated on their contacts, so that the integrity of the sliding body is maintained. Equation (2) for an element changes for the 3D solution to [S [(si - ui) tanfi + ci] li ] Lm + T(e)m i FS3=-------------------------------------------- (3) Sti li Lm i The shear force on the cross section of element m needed to establish a unique safety factor is computed from Eq. (3) as T(e)m=[Sti li Lm] FS3 -[S [(si - ui) tanfi + ci] li] Lm (4) i i T(e)m>0 for elements the safety factors of which are less than the 3D one, and T(e)m<0 for elements with a safety factor which is greater than the 3D one. Fig. 12 The shear forces on all elements, needed to establish a unique 3D safety factor, are computed from Eq.(4). The shear forces between the elements are cumulatively transferred among the elements (Fig. 12) and the mobilized force on the m-th contact is given by m T(k)m= ST(e)i - T(k)0 (5). i=1 The safety factor F(k)v on each contact between the elements must be greater than FS3. If this condition is not satisfied, the physical compatibility of the whole sliding body is not satisfied either. A safety factor F(k)r, which must be greater than one, is defined as F(k)r = F(k)v / FS3 (6) According to Eq.(5), the following relationship must be satisfied for the first m successive elements ST(e)i=T(k)o+T(k)m (7). Considering the separation of m successive elements, using Eq. 7, and after carrying out the summation over elements 1 to m, Eq.3 changes to m  S[S [(si - ui) tanfi + ci] li] Lj + T(k)0+ T(k)m j=1 i FS3m=------------------------------------------------ (8) m S[Sti li] Lj j=1 i which is the cumulative safety factor for the first m successive elements, which all must be greater than the 3D safety factor for the whole sliding body. When any FS3mnHnInJnynznnnnnnnn{{FNcdjkmnoqrxy{|}P uDP]^cU]c]ch]c ]ch]cV]c]c]chPu"E_n   ]2W 5 !$Q$&&&&'!F!F!F!F!!!!! !!!!! !!!!!!!!!! !!! !!! !! !!!!!dddddddddd%'/(7*+ +q,W-^-6//Q01112566'6/60616F6m6V:y;<V?@@@@@@AAAAAAAOB!!!!!!!!!!!!!! !!!!!!!!!!!! !!!F!J!B!F!J!B!!s!B!F!s!B!ddddd)OBCEFGGGH:HLH[H\HHHI'I(IIIJJJJKL-LLL>MCM|MMMMMMMP3Q>QyQ!!!!!!s!B!{!s!B!!!{!s!B!!!!!F!J!F!!!!!!!F!s!B!{!F!s!B!F !!s!B!{ddxddd(yQQQQ RRSSSSSTTUtXX[\'\I\]m``abcccKeegghhhhh"ibi!s!B!F!!!!!!!!!! !! !!!!! !!!!!!!!!!!!!!!!!!dd dddddd&biiij?j|jjjjjklkkklClgllll m6mmmnHnnnnnbo ppTqqorrstt2uu!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!Vdddd)u,vvHwwoxxyz=zz{{{|H|?}~~?B"6vDNbcopq}~!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!@ &@%ddddV&!ddK@Normala "A@"Default Paragraph Font @ Header !)@ Page Number!                        !  _ "&+m/13y8=AE-IXMPVZ_(d|g6jm2r=wx?|v X#G   >fdkm[+@GIOMQTlDEFGHIJKL'OByQbiuMNOPQRS !!CJDJEJMMMh!QIvan%C:\PROJEKTI\PODSUSED\CLANAK\KOST2.DOCBobo%C:\PROJEKTI\PODSUSED\CLANAK\KOST3.DOCBobo%C:\PROJEKTI\PODSUSED\CLANAK\KOST3.DOCBobo%C:\PROJEKTI\PODSUSED\CLANAK\KOST3.DOCBobo%C:\PROJEKTI\PODSUSED\CLANAK\KOST3.DOCBobo%C:\PROJEKTI\PODSUSED\CLANAK\KOST3.DOCBobo%C:\PROJEKTI\PODSUSED\CLANAK\KOST3.DOCBobo%C:\PROJEKTI\PODSUSED\CLANAK\KOST3.DOCBobo%C:\PROJEKTI\PODSUSED\CLANAK\KOST3.DOCBobo%C:\PROJEKTI\PODSUSED\CLANAK\KOST3.DOC@HP LaserJet IIILPT2:hppcl5aHP LaserJet III 8DR HP LaserJet III 8DR 33 33aTimes New Roman Symbol &ArialCRO_Bookman-NormalCRO_Swiss-Normal!h2%f:E!=!THE KOSTANJEK LANDSLIDE IN ZAGREBBoboBoboࡱ;