In order to evaluate the safety of the shallow buried tunnel cavern section after collapse treatment,based on the moni-toring data of the deformation and force of the surrounding rock after treatment,the Hampel filter was used to process the data anoma-lies,and combined with the time series prediction DeepAR model,the Hample-DeepAR data prediction model was established and applied to the analysis of the stability of the surrounding rock after collapse treatment to verify the effectiveness of the treatment meas-ures.The results show that the MAE,MSE,RMSE,and R2 values of the Hampel-DeepAR model for vault settlement prediction are 0.018 5,0.000 5,0.021 4,and 0.564 8,respectively,which have higher prediction accuracy and better prediction effect compared with DeepAR,ARIMA,SVM,LSTM,and BP models,indicating that the use of Hample.The values of surrounding rock displace-ment and force prediction show that the arch top settlement value is 6.58mm,the surrounding convergence value is 5.89 mm,the arch top surrounding rock pressure value is 0.186MPa,the maximum stress value of steel reinforcement is 152.47 MPa,and the max-imum stress value of second liner concrete is 1.73 MPa,which shows the necessity of using Hample filter to deal with pre-processed abnormal values,and verifies the applicability of Hampel-DeepAR model.The stability of the surrounding rock has been greatly im-proved after treatment,which can provide a theoretical basis for the evaluation of the effect of tunnel collapse treatment.
关键词
隧道塌方/变形预测/Hampel滤波/Hampel-Deep/AR模型/处治效果
Key words
tunnel collapse/prediction of deformation/hampel filtering/hampel-DeepAR model/treatment effect
维普
万方数据