Advance Rate Predictions of Tunnel Boring Machines Using Bayesian-Optimized CNN-LSTM
Xiaoxiong Men, Yuanfei Li, Baicun Guo, Lai Wang, Xinyu Ye, Qiujing Pan
Abstract
During the tunnelling process of a tunnel boring machine (TBM), accurately predicting the advance rate (AR) is highly desirable for enhancing construction efficiency and safety. Inaccurate AR estimates may lead to extended construction periods and, thus, increased project costs. This study introduces a hybrid deep learning method that combines the convolutional neural network (CNN) and the long short-term memory network (LSTM), optimized using Bayesian optimization, to predict the AR of a TBM. The proposed method includes feature selection, model establishment, and hyperparameter optimization. Data from two tunnel projects are used to validate the effectiveness of the proposed Bayesian-optimized CNN-LSTM model. The results show that the proposed model achieves higher accuracy in predicting AR, outperforming the artificial neural network (ANN), random forest (RF), and LSTM models.