Machine Learning Based Prediction of Surrounding Rock Deformation and Collapse Risk Assessment in Tunnel Entrance Excavation
The geological and engineering environment faced by tunnel construction is extremely complex,covering various factors such as geological structures,lithology,groundwater level,and seismic activity.The evaluation and prediction of the stability of the surrounding rock during its construction process are particularly crucial.Especially during the excavation stage of tunnel openings,due to the impact of excavation operations on the surrounding geological environment,the risk of surrounding rock collapse significantly increases,which may lead to serious safety accidents and project delays.This article is based on machine learning methods,especially supervised learning and deep learning algorithms,for predicting the deformation of surrounding rock during tunnel excavation and evaluating the risk of collapse on this basis.The research results show that both the horizontal convergence and arch settlement predicted by ARMA algorithm have significant differences from the measured values,and the fluctuation amplitude is large.The maximum difference between the horizontal convergence predicted value and the measured value is 1.7 mm,and the maximum difference between the arch settlement predicted value and the measured value is 1.2 mm,indicating poor prediction performance.The difference between the predicted values of LSTM algorithm and RNN algorithm and the measured values is relatively small,and the prediction results of LSTM algorithm are significantly more stable compared to RNN algorithm.As time goes on,the risk of surrounding rock collapse increases.But the risk of collapse is less than level Ⅱ.
machine learningtunnel engineeringexcavation of tunnel entrancedeformation predictionrisk assessment
NETL
NSTL
万方数据
