基于ARIMA-LSTM的矿山排土场边坡沉降预测模型研究

In order to ensure the safety and stability of the slopes of the mine discharge site, firstly, the slope settlement of Huaxin Cement mine discharge site is monitored by using GNSS technology; for the historical data collected hourly between 0:00 on January 1, 2022 and 23:00 on April 31, 2022, the data are analyzed by box-and-line diagrams and graphs of the relationship between the cumulative settlement and the time series; then, taking into consideration of the shortcomings of the single model, the ARIMA-LSTM combined prediction model was established, and absolute relative error and root mean square error were introduced as evaluation indexes; finally, the prediction results of the combined model were compared with the actual values and the single model. The results show that according to the historical monitoring data, slope 3 is more stable compared with slope 1 and slope 2. The maximum absolute relative errors and average absolute relative errors of the combined ARIMA-LSTM predictions for the three slopes are 6.67%, 9%, 6.09% and 4.07%, 4.45%, 3.37%, respectively, which improves the prediction accuracy and meets the needs of engineers and technicians for mine engineering prediction.