A new prediction model for rock slope failure time: a case study in West Open-Pit mine, Fushun, China

Establishing an accurate method for predicting the failure times of rock slopes subject to creep deformation is challenging, but at the same time crucial for preventing damage to properties and loss of life. In this paper, the Medium-short Term Prediction of Landslide by Polynomial (MsTPLP) model is proposed based on the Levenberg-Marquardt (LM) algorithm. The West Open-Pit mine in Fushun, NE China is currently the largest open-pit coal mine in Asia. The landslide on the southern slope of the West Open-Pit mine was selected as the study case. Global Positioning System (GPS) monitoring is employed in landslide displacement monitoring. Based on the analysis process of the MsTPLP model, the displacement time series derived from GPS monitoring points is selected as the input. The model parameters of the MsTPLP model are obtained using the Levenberg-Marquardt (LM) algorithm. The predicted failure time of a landslide, which is the output, can be determined according to the prediction criteria of the model. The prediction results show that the MsTPLP model can provide accurate landslide displacement predictions (correlation coefficient R (2) > 0.98 and average relative error ARE < 17 %). The forecasting results of the landslide show that the estimated failure time is Mar 5, 2014. Based on field investigation and displacement analysis, the landslide on the southern slope of the West Open-Pit mine occurred on Mar 9, 2014. The predicted and actual failure times are significantly close, demonstrating the potential of the new method in landslide prediction.