PROCEDURE OF SLOPE FAILURE PREDICTION DURING RAINFALL BASED ON THE BACK ANALYSIS OF ACTUAL CASE RECORDS

ABSTRACT The method of slope stability analysis during rainfall employed in the present study is based on the combination of simplified unsaturated seepage analysis and usual circular arc stability analysis, and the method can evaluate not only both the reduction of soil strength along slip surface and the increase of self weight of slope due to seepage but also some other factors such as rainfall intensity, antecedent precipitation, vegetation, slope protection, weathering and spring water. The effects of these factors on the stability of the slope are evaluated in terms of the reduction of soil strength through introducing new parameter of strength reduction. The linear regression analysis will be made for evaluating the new strength reduction parameter using field observations and the data of the past case records of slope failures due to rainfall. By applying the Bayesian method to the linear regression analysis the probabilistic slope failure prediction is also shown to be available. The numerical simulations for the developed procedure has made the each effect of those factors on the stability of slopes clear and distinct. The developed procedure is applied for a case record of slope failure using actual rainfall observations. The risk of slope failure calculated by the developed method increases rapidly just before the observation of actual slope failure. The effect of the slope protection by the use of concrete spraying on the stabilization of slope is also shown to be well evaluated by the developed procedure.