<<
[an error occurred while processing this directive] | [an error occurred while processing this directive] >>
Research on Prediction of TBM Driving Parameters Based on Deep Learning
(1 School of Civil Engineering，Tongji University，Shanghai 200092；2 School of Electronics and Information Engineering，Tongji University，Shanghai 200092)
Abstract With continuous application of the full face tunnel boring machine (TBM) in tunnelling, the adaptive ad?
justment of TBM driving parameters becomes more and more important. Because it is difficult to accurately predict the geological conditions before excavation, it is important to predict some important driving parameters such as cutter head torque and total thrust based on the field monitoring data for the adaptive adjustment of TBM driving parameters. Taking the TBM3 lot of Jilin Water Supply Project as the research object, it introduces the long short-term memory network (LSTM) which can consider the data time correlation, and puts forward a model which can accurately predict the total thrust and cutter head torque of the stable section according to the driving parameters of first 30 s of the rising section of TBM driving cycle. The prediction results show that the prediction accuracy of the total thrust and the cutter head torque in the stable section is high, up to 91% and 89%, respectively. Compared with the ran?
dom forest, LSTM can make full use of a large number of TBM driving parameters, reflecting a better prediction ability. The research in this paper has a certain reference value for the real-time adjustment of TBM driving parameters.
Abstract：
With continuous application of the full face tunnel boring machine (TBM) in tunnelling, the adaptive ad?
justment of TBM driving parameters becomes more and more important. Because it is difficult to accurately predict the geological conditions before excavation, it is important to predict some important driving parameters such as cutter head torque and total thrust based on the field monitoring data for the adaptive adjustment of TBM driving parameters. Taking the TBM3 lot of Jilin Water Supply Project as the research object, it introduces the long short-term memory network (LSTM) which can consider the data time correlation, and puts forward a model which can accurately predict the total thrust and cutter head torque of the stable section according to the driving parameters of first 30 s of the rising section of TBM driving cycle. The prediction results show that the prediction accuracy of the total thrust and the cutter head torque in the stable section is high, up to 91% and 89%, respectively. Compared with the ran?
dom forest, LSTM can make full use of a large number of TBM driving parameters, reflecting a better prediction ability. The research in this paper has a certain reference value for the real-time adjustment of TBM driving parameters.
LI Gang1 LI Xiaojun1 YANG Wenxiang2 HAN Dong1
.Research on Prediction of TBM Driving Parameters Based on Deep Learning[J] MODERN TUNNELLING TECHNOLOGY, 2020,V57(5): 154-159
2020, Vol. 57 