Abstract：Effective prediction of shield boring parameters under complex geological conditions can be used to guide the shield tunneling pertinently. The study is based on the Φ7m shield boring parameters of Chegongmiao station~Hongshuwan station and Nanshan station~Qianhaiwan station in Shenzhen metro line 11, and the parameters are obtained by field monitoring. Firstly, by using BP artificial neural network method, the prediction model of shield boring parameters in a mixed ground is established. Then, the formation parameters are used as input group and shield boring parameters as output group, after training the samples, the output values of the boring parameters are basically consistent with the original values, which shows that the model has good nonlinear mapping ability. Finally, based on the formation parameters of typical section in shield tunneling, the model is used to predict the shield boring parameters in a mixed ground. The predictive value fits the actual value, and the average error is less than 15%. The BP neural network model established in this paper can be used to predict similar types of shield boring parameters in mixed ground.
李超 李涛 李正 詹金武. 基于BP神经网络的复合地层盾构掘进参数预测与分析[J]. 土木工程学报, 2017, 50(S1): 145-150.
Li Chao Li Tao Li Zheng Zhan Jinwu. Prediction and analysis of shield boring parameters in a mixed ground based on BP neural network. 土木工程学报, 2017, 50(S1): 145-150.
