The risk assessment and prediction of water inflow into tunnels of Guangzhou Metro Line 21

4th International Conference on Energy Equipment Science and Engineering

Feng, Yong^1 ; Chen, Lingqiang^2 ; Zhu, Liubing^2 ; Hu, Wenyi^2 ; Min, Youwei^2^3

China Railway 19 Bureau Group Co. Ltd., China^1

Guangzhou Metro Group Co. Ltd., China^2

China University of Geosciences, Hubei Province, Wuhan

430074, China^3

关键词: BP neural networks;    Evaluation indicators;    Geological factors;    Hydrologic conditions;    Integrated prediction models;    Practical projects;    Required precision;    Tunnel construction;   

Others  :  https://iopscience.iop.org/article/10.1088/1755-1315/242/6/062047/pdf
DOI  :  10.1088/1755-1315/242/6/062047

来源: IOP

【 摘 要 】

In order to decrease the negative influence of water inflow on tunnel construction, the risk of water inflow should be evaluated and predicted. Based on the method of experts' evaluation, combined with practical project, an assessment system of water inflow into mountain tunnel was set up, in which 9 evaluation indicators were proposed from the geological factors, hydrologic conditions and engineering three aspects, and every index weight was counted with the method of AHP. With BP neural network as an assessment tool, an integrated prediction model for water inflow into mountain tunnel was established. The assessment and prediction results for water inflow into tunnel of Guangzhou Metro line 21 were compared with results of expert classification. It shows that the error of BP neural network forecasting structure by means of AHP is not more than 3%, and it matches the required precision.