Predicting slope collapse using Slope Stability Radar deformation data

We selected six open-pit wall collapses, representing a variety of failure sizes, failure types and rock mass characteristics, in order to assess the accelerating creep theory by Voight and Fukuzono as an early forecasting tool. Time-displacement monitoring data were collected shortly before ultimate failure using the GroundProbe Slope Stability Radar (SSR), a ground based interferometric radar able to detect surface displacements with a sub-millimetre accuracy. Non-linear estimation techniques were employed to relate slope displacement behaviours to Voight's empirical relationship in order to investigate the trend of its controlling parameters and possibly identify a time window where forecasting output were found to be stable. Analyses were performed taking into account different amounts of survey points involved in the failure mechanism and various values of the velocity at time of failure parameter. Finally we evaluated Fukuzono's inverse velocity approach paying particular attention to update linear fits whenever a trend change was identified. © 2010 Taylor & Francis Group, London