Abstract:

Foundation monitoring was necessary methods to ensure the implementation of foundation engineering safety in mines.Because the different models for the settlement of foundation monitoring exists certain difference，therefore，how to select an effective portfolio model that can predict the settlement of foundation pit accurately at a certain time in the future was the main problems.In this research，the time series prediction model and grey model（gray sequence combination forecast model） were employed to a deep foundation pit（5.7~13.7 m deep foundation pit） subsidence monitoring data analysis，and the predicted results were accurate and reliable. Meanwhile，compared with the predicted results of single model（such as ARIMA and GM （1，1）），prediction accuracy of the gray timing sequence model was much higher.The predicted results that we obtained were the closest to the measured values，which presented it was a extremely effective prediction method for foundation pit.

Key words: deformation monitoring, subsidence data, time series model, grey model, deep foundation pit, mine