Seismic Responses Of Rock Slopes In A Shaking Table Test

An earthquake always triggers in mountainous areas large amounts of secondarygeological disasters, such like landslides, debris flows, collapses, rock falls and etc.,which cause deaths and economic loss even greater than those directly caused byground motion. Field survey and seismic moniroing results from historical earthquakesshowed that, most of slope failures occurred in positions with abnormal topography, likein prominent and single ridges, where the acceleration, velocity and displacementresponses also reached very high values. The phenomina are called topograhicamplification effect. In China, the Wenchuan earthquake of2008showed us withunprecedented data that, the distribution of slope failures is much correlated totopography, geology and source parameters at the same time. Investigators have reacheda qualitative aggreement on the above phenomena. In order to interpret thosephenomena, they strive to uncover the response laws of slope acceleration, velocity,displacement and etc. before failures, herein namely seismic slope responses. However,due to the lack of field monitoring on slopes and the inconsistency between monitoringand numerical results, no quantitative aggrement has been reached for seismi sloperesponses. In the evaluation work of seismic slope stability, pseudo-static method isalways adopted which impractically simplifys the earthquake force and ignores itsamplification caused by irregular topography. In this context, just after the Wenchuanearthquake, when someone else devote themselves to the field monitoring work, theauthor also focuses on the research of seismic slope responses through physical modeltests. The present paper is expected to summarize the test results and extract somequantitative knowledge of seismic slope responses.With the help of large-scale shaking table tests, simulating the typical lithology androck structure developped in Wenchuan earthquake area, ten model slopes with differentfeatures were designed to explore the seismic responses of rock slopes. In terms oflithology, these models simulated soft rock slopes, hard rock slopes and rock slopeswith complex lithology. In terms of rock strtucture, these models simulatedhomogeneous slopes, horizontally layered slopes and slopes with one weak interlayer.Through a secquence of preparation work on determing similitude relations, sensorarrangement, input motions and a loading scheme, a set of design plan has beenestablished applicable to the rock slopes.The present paper focuses on analyzing the effects of topography, geology andsource parameters on model slope acceleration and displacement responses, based onsensor measurements before model slope failures. Among three kinds of effects, thetopographic effect will be emphasised. Finally, the basic laws about seismic sloperesponses can be formed. The main contents and results of the paper are as follows: (1) Topographic effect analysis. In this section, both horizontal and verticalcomponent acceleration and displacement responses were analyzed at differentelevations of each model slope. For each component, responses on slope surface andinside slope were compared. Results show that, the acceleration responses for bothcomponents demonstrate nonlinear amplification mode with increasing elevations,herein namely topographic amplification. The amplification facor relative to shakingtable response was calculated for peak ground acceleration (PGA) and thecorresponding Arias intensity at each elevation. Based on statistical results ofacceleration amplification factors of ten models under all test conditions, most factorsfor PGA have values lower than3.5, while most factors for Arias intensity have valueslower than10.0. The obvious difference between two component responses is that,topographic amplification is weak for horizontal component in the lower slope withrelative elevation h/H≤0.5, all facors keep within2.0for peak horizontal acceleration(PHA) and within5.0for horiozntal Arias intensity. Another statistic analysis wereperformed in ratios calculated between responses on slope surface and inside slope. Forhorizontal component responses, both PHA and Arias intensity, ratios concerntrate in arange of0.8~1.2, while0.5~2.0for vertical responses. Finally, acceleration analysis infrequency domain indicates that, the acceleration amplification in higher elevation ofmodel slope is related to the high-frequency motions in this positon.(2) Geological effect analysis. Based on analysis (1), in this section, effects oflithology and rock structure on both horizontal and vertical component responses arefurther disscussed. Results show that, in the lower slope part with h/H<0.5, the horizontallayer structures play the control role for both component responses. The difference in slope structurecauses a maxmum PHA ratio of1.2between homegeous model and horizontally layered model, aswell as a maximum PVA ratio of2.2. In the upper slope part, the lithology plays the control rolefor both component responses. The difference in lithology causes a maxmum PHA ratio of2.5between model simulating soft rock slope and model simulating hard rock slope, as well as amaximum PVA ratio of2.1.When a weak interlayer was considered in a model slope, in the upper slope partwith0.75≤h/H≤1.0and at the slope bottom, thickness of interlayer plays the control rolefor both component responses. The difference in thickness (3cm and15cm) causes amaxmum PHA ratio of1.7and2.6respectively in those two positions, as well as amaximum PVA ratio of1.3and3.1. In the middle and lower slope part with0.25≤h/H≤0.5, the dip angle of interlayer plays the control role for both componentresponses. The difference in dip angle (0oand20o) causes a maxmum PHA ratio of1.8aswell as a maximum PVA ratio of1.5.(3) Source effect analysis. In this section, emphasis was put on effect of excitationintensity on horizontal and vertical component acceleration responses. As excitationintensity increases, the inner structure of model slopes occurs deterioration. Accordingto the decrease degree of resonant frequecy and the time when model began to deformand finally failed, the slope structure deterioration can be graded to three levels: low, middle and high. For three levels of deterioration, excitation intensity demonstratesdifferent effects on acceleration responses. Nonetheless, based on the change tendencyof horizontal component responses with increasing excitation intensity, three stages canbe used to describe the response process of model slope: preparation stage, the upperlimit of excitation intensity is0.3g for PHA and0.15m/s for Arias intensity; triggeringstage, the range of excitation intensity is (0.3g，0.6g) for PHA and（0.15m/s，1.5m/s）for Arias intensity; after that, slope response enters into another stage with a suddenchange or with a final stability.It should be noted that, the above results are applicable to rock slopes, andapplication in soil slopes needs further validation. Limited by test conditions, modelslopes only simulate prototypes with small geometric size, that is to say, they behave aslow slope responses（H<0.2λ）[159]. Even so, the results are expected to provide areference to research in seismic problems of rock slopes and its engineering application.