Journal of Shanghai Jiaotong University ›› 2020, Vol. 54 ›› Issue (8): 820-830.doi: 10.16183/j.cnki.jsjtu.2020.99.011

Previous Articles     Next Articles

Evaluation of Tunnel Collapse Susceptibility Based on T-S Fuzzy Fault Tree and Bayesian Network

CHEN Wu1,2, WANG Hao1(), ZHANG Guohua1, WANG Chengtang1,2, ZHONG Guoqiang3   

  1. 1. State Key Laboratory of Geomechanics and Geotechnical Engineering, Institute of Rock and Soil Mechanics, Chinese Academy of Sciences, Wuhan 430071, China
    2. University of Chinese Academy of Sciences, Beijing 100049, China
    3. Shandong Provincial Communication Planning and Design Institute, Jinan 250031, China
  • Received:2019-01-14 Online:2020-08-28 Published:2020-08-18
  • Contact: WANG Hao


Collapse is one of the common disasters in the process of drilling and blasting tunnel construction. In order to effectively prevent the collapse accident and provide decision-making basis for safety risk analysis and management of tunnel construction, the T-S fuzzy fault tree and Bayesian network were combined, and an evaluation method based on both of them was proposed to calculate tunnel collapse possibility. According to the transformation of T-S fuzzy fault tree to Bayesian network, the Bayesian network model and conditional probability table were determined. In addition, fuzzy number and fuzzy subset were used to describe the fault state and probability of nodes respectively, and the bidirectional reasoning algorithm of Bayesian network was used to calculate. This method can use two different forward inferences to calculate the possibility of tunnel collapse, including the fuzzy subset of root node fault probability and the actual fault state in construction. At the same time, it can check the fault according to the result of importance analysis of root nodes, and can calculate the posterior probability of fault diagnosis of root nodes by back inference. Finally, the application of two engineering examples shows that the method can scientifically evaluate the possibility of tunnel collapse and determine key risk factors. This method can be used as a decision-making tool for tunnel construction safety assurance and management.

Key words: construction tunnel by drilling and blasting, collapse possibility, T-S fuzzy fault tree, Bayesian network, fuzzy number, fuzzy subset, importance

CLC Number: