Deprem etkisindeki şevlerde stabilitenin incelenmesi

Abstract

Bu çalışmada doğal veya insan eliyle oluşturulmuş şevlerde stabilite incelenmektedir, özellikle deprem riski taşıyan bölgelerde oluşturulacak şevlerde olası deprem büyüklüklerine göre stabilite analizleri yapmak, doğal şevlerde ise gene bu büyüklüklere göre şevin stabilitesi hakkında karar vermek çoğu zaman gerekli olmaktadır. Depremin zeminin yapısında meydana getirdiği kayma mukavemetindeki azalmalar ve çözülmeler stabilite analizlerinde göz önüne alınmamaktadır. Bu çalışmada, depremin neden olduğu tekrarlı kayma gerilmeleri etkisi altında zemin şevlerinin stabilite analizi için bir yöntem önerilmektedir. Olası deprem büyüklükleri önce deprem ivmesi şeklinde daha sonra deprem manyitüdü ve faya uzaklıklar şeklinde ele alınmıştır. Önerilen bu şev stabilite analiz yöntemi ile daha önce bir deprem sırasında heyelan görülen bölgeler için şev kaymalarına göre mikrobölgeleme yapılmış, gerçekte de bu yöntemle şev kaymasına göre riskli gruba giren bölgeler heyelanların olduğu bölgeler olmuştur. Stabilite yönteminin bir mikrobölgeleme yöntemi olarak da uygulanabileceği görülmüştür. Depremler sırasında etkiyen tekrarlı kayma gerilmelerinin zeminin yapısında meydana getirdiği kayma mukavemeti azalmalarının mekanizması incelenmiş, çeşitli azalma nedenleri de araştırılarak, bilgisayar programları ile bu azalmalar deprem büyüklüklerine, zemin ve şev özelliklerine göre modellenmiştir. Modellenen bu azalmalar önerilen şev stabilite analizinde hesaba katılmıştır. Bu azalmaların güvenlik hesaplarını da nasıl etkilediği araştırılmış ve bununla ilgili, deprem manyitüdüne, faya uzaklıklara, zeminin mukavemetine bağlı stabilite grafikleri hazırlanmıştır. Şev stabilite analizlerinde çok kullanılan, dilim yöntemlerinden Bishop ve Janbu yöntemlerine depremin neden olduğu mukavemet azalmaları da ilave edilmiştir. Çalışmada önerilen yöntem ile Bishop ve Janbu yöntemlerinin belli şev kesitleri için karşılaştırmaları yapılmış ve yöntemin diğer yöntemlere yakın sonuçlar verdiği gözlenmiştir.

Earthquakes may trigger slides and may cause severe damages in cut slopes, earth dams, embankments and natural slopes. Large number of landslides that have taken place during recent earthquakes have demonstrated that instabilities of natural and man-made slopes is one of the major causes of damage. In order to estimate susceptible zones for slope instabilities during earthquakes, various zonation methods were developed in the literature. These methods may be considered in three levels depending on the comprehensiveness of the approach. The first level methods defined as Grade-1 Methods, were developed based on number of landslides observed in earthquakes with respect to earthquake magnitude and epicenter or fault distance. Due to approximate nature of these methods, geotechnical and topographical conditions were not taken into account. The methods proposed by Tamura (1978), Yasuda and Sugitani (1988), Keefer et al (1978), and Ishihara and Nakamura (1987) are some of the procedures in this category. The second level method or Grade-2 Methods, are more reliable since geological characteristics as well as the topographical aspects of the region were taken into consideration in addition to earthquake magnitude and epicentral distance. Methods adopted by Kanagawa Prefectural Government (1986) and proposed by Mora and Vahrson (1992) were the two alternatives in this category. The third level Grade-3 Methods, are the most comprehensive microzonation procedures to identify areas with different degrees of susceptibility of slope instabilities. In this category geotechnical properties of soil and rock layers as well as the slope geometry were used along with the peak ground acceleration. The approach suggested by Koppula (1984) was modified to be used for zonation. The suggested approach is a pseudo-static evaluation of slope stability based on seismic coefficient to account for earthquake induced horizontal forces. The potential failure surface is assumed as a circular arc and the geometry of the slope and configuration of the failure surface are considered in the formulation of stability number N., as Nj = 3 (a +cot5- a cotacotS) / DEN ( 1 ) DEN = sin2a sin28( Dl + D2 ) (2) Dj = 1-2 cot2P - 3 cota cotp + 3 cotp cot8 + 3 cot5 cota - - 6n cot{3- 6n2 - 6n cota + 6ncot 8 (3) D, = A(cotP+cot38+3 cota cot28-3 cota cotp cot8-6n cota cot8) (4) where a is the central angle, 5 is inclination angle of the secant, and n is the distance from the bottom of the slope to the toe of the failure arc, and p is the slope angle. Assuming linear variation of shear strength with depth the factor of safety is obtained as ; Fs = tan <() * Nj. (5) where <|> is the average angle of shear strength for the soil and rock layers in the region. Thus safety factor depends on the angle of shear strength and stability number, N^ representing the configuration of the slope and failure surface. The minimum values of the stability number were determined by carrying out a parametric study in terms of a, 8, and n to find the most critical failure surface and the variation of minimum N, can be expressed as a function of p (slope angle) and A(earthquake acceleration). On July 22, 1967 an earthquake of magnitude M=7.1 took place along the North Anatolian Fault Zone in Adapazari-Mudurnu region causing 80 km of fresh faulting. Maximum relative displacements of 190 cm lateral and 120 cm vertical were measured. A significant number of slope failures occurred during the earthquake. North Anatolian Fault zone is a belt of few kilometers. The Mudurnu epicentral area is located on the block of Pontides mostly composed of crystalline metamorphics. The crystalline series form the basement of the region. Eocene flish is found mostly on the west part of the fault zone lying over Upper Cretaceous floes. Pliocene rocks consisting of marls, weakly cemented sandstones and hard clays are located at the southern boundary. Slope derbis, derived mostly from Lower Cretaceous limestone covers the central part of the zone (Yılmaz et al., 1981). Based on the available information three areas where major slope failures have taken place (Ambraseys, et al 1967) were selected to carry out microzonation. In the first stage based on the geologic map of the region XI with a scale of 1:500 000, Grade-1 Methods suggested by Tamura (1978), Yasuda and Sugitani (1988), Keefers et al. (1978), Ishihara and Nakamura (1987) were applied and zones of different degrees of susceptibilities were determined as a function of distance from the faults and epicenters for earthquake magnitudes corresponding to return periods of 200 and 500 years. Among these methods the procedure proposed by Ishihara and Nakamura (1987) have yielded realistic results. Most of the observed landslides during 1967 Mudurnu Earthquake were in the highest risk zone. Similar results were also obtained by the procedure suggested by Tamura (1978). In the second stage more detailed Grade-2 methods adopted by Kanagawa Prefectural Government (1986) and proposed by Mora and Vahrson (1991) were used for zonation. The region with two major slope failures was selected for zonation and based on the topographical map of 1:25 000, slope failures susceptibilities were determined using meshes of 500x500m. The zonation map obtained by Mora and Vahrson method for an earthquake magnitude of 6.9 correponding to 200 year return period, the two major landslides that have taken place during Mudurnu 1967 earthquake were in moderate risk zones. The procedure developed based on the suggestions of Koppula (1984) is utilized to calculate the factor of safety for slope stability. A zonation in terms of factor of safety is carried out on a map with scale 1:10 000 in areas where major slope failures were observed during 1967 earthquake. The area were divided into 500x500 m. meshes, the distribution of shear strength angle, <\> is estimated based on the geological and geotechnical investigations conducted in the region with a scale of 1:5,000 or 1:10,000 (Yılmaz, et al, 1981). The slope angle, J3 for each mesh is considered separately and the steepest slope, determined from the topographical map, is selected as the slope angle for that mesh. The peak acceleration at the site is estimated based on seismic risk studies and using a suitable attenuation relationship. This procedure is applied for microzonation of the area around Akyokuş Village in Adapazarı region. In this example based on the seismic risk studies for the region, an earthquake magnitude of M=6.9 corresponding to 200 year return period is selected for estimating the peak ground acceleration. Since the North Anatolian Fault is within 20 km distance, the peak ground acceleration is taken as 0.30g based on an attenuation XII relationship proposed for Turkey. The calculated factors of safety are considered in three groups with respect to risk levels as: a. High for Fs.<1 b. Moderate for 1 > Fs >1.5 c. LowforFs> 1.5. The region around Akyokuş village is calculated to have mostly, moderate risk levels for slope instability. During 1967 Adapazari-Mudumu earthquake of magnitude M= 7.1, slope failures were observed around Akyokuş village located approximately 28 km from the epicenter and about 3 km from the surface ruptures which substantiates the applicability of the suggested approach. Cyclic stresses induced by earthquakes may lead to increase in pore pressures and reductions in shear strength. However, additional reduction in shear strength is also expected due to degradation of soil stiffness with respect to amplitude of cyclic strains generated by the earthquake excitations. These factors have to be taken into consideration and must be analytically evaluated in determining the factor of safety for slope stability. There are two possible failure mechanism that may lead to slope instabilities. One of them is the reduction in the quasi-static shear strength which most of the time would trigger slope failures following an earthquake. This case can be evaluated by introducing parameters to account for pore pressure increase and shear strength reduction into the pseudo-static slope stability analysis. An important issue encountered in the field is the reduction of the static shear strength due to cyclic loading. There are differences in the results reported in the literature. The results obtained from series of tests conducted on undisturbed soil samples under undrained conditions have indicated that the strength reduction following cyclic loading without allowing the dissipation of excess pore pressures, was more dependent on number of cycles than cyclic stress ratio. And it appears possible to consider a critical number of cycles below which strength reduction would be limited and can be considered as independent of cyclic stress ratio." For larger number of cycles, it was possible to observe a more distinct pattern for reduction in the shear strength post cyclic loading as a function of cyclic stress ratio. The factors such as soil stratification, properties of soil layers, distance to epicenter and the magnitude of the earthquake is taken into consideration X1U and semi-emprical relationships were developed based on laboratory cyclic simple shear tests and field observations after earthquakes to estimate the magnitude of pore pressure accumulation and shear strength reduction. The second possible failure mechanism may be due to the magnitude of cyclic stresses generated during an earthquake. In this case an approach has to be adopted to assess the cyclic yield stress that would produce excessive deformations in the slope which may initiate slope failures. A procedure is developed to estimate cyclic yield stress based on cyclic laboratory tests and to determine the dynamic safety factor to evaluate the susceptibility of slope instability. The response patterns observed from cyclic tests indicate that it appears possible to consider a critical shear stress ratio which can be defined as "the critical level of repeated stress". If the soil samples are subjected to cyclic shear stresses with stress ratios larger than the critical level, pore pressure will accumulate continuously and the sample will undergo large cyclic shear deformations. However, if the applied stress ratio is smaller than the critical stress ratio, the accumulated pore pressure will be limited and the sample will experience relatively small shear deformations. The effect of pore pressure increase on shear strength with respect to earthquake magnitude and epicentral distance was determined. The corresponding decrease in the shear strength was expressed as a function of shear strength angle.