Open AccessA Novel A-CNN Method for TBM Utilization Factor Estimation
Author(s) -
Honggan Yu,
Jianfeng Tao,
Chengjin Qin,
Hao Sun,
Chengliang Liu
Publication year - 2021
Publication title -
journal of physics. conference series
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.21
H-Index - 85
eISSN - 1742-6596
pISSN - 1742-6588
DOI - 10.1088/1742-6596/2002/1/012049
Subject(s) - computer science , convolutional neural network , factor (programming language) , dimension (graph theory) , artificial neural network , set (abstract data type) , artificial intelligence , data mining , pattern recognition (psychology) , machine learning , mathematics , programming language , pure mathematics
Utilization factor is one of the most important performance indicators of TBM, which affects the construction period and cost of the tunnel. However, there are few models to evaluate the utilization factor based on geological conditions and operational parameters. In this paper, a novel A-CNN neural network architecture for TBM utilization factor estimation is proposed. Firstly, the input dimension is expanded by a full-connected neural network. Secondly, a convolutional neural network is designed and added behind the expanded input to extract relevant features. Finally, a regressor is designed to build the mapping relationship between the extracted features and the utilization factor. The data collected from a Singapore tunnel project was utilized to verify the proposed method. The results show that the R 2 of the proposed method on the test set is 0.521, which is 52.33%, 15.60%, 34.10%, and 9.40% higher than the KNN-based, SVR-based, RF-based, and DNN-based methods, respectively. Therefore, the proposed method can estimate the TBM utilization factor more accurately.