TBM construction risk analysis based on fuzzy Bayesian network

Fangdi Xie,Yong Zhou
DOI: https://doi.org/10.1080/13467581.2024.2329353
IF: 0.904
2024-03-24
Journal of Asian Architecture and Building Engineering
Abstract:Aiming at the characteristics of various TBM construction risk factors, fuzzy and complex systems, and frequent shutdown accidents, a TBM construction risk analysis methodology based on a fuzzy Bayesian network is proposed. This method mainly selects eighteen risk factors to construct an interpretive structural model. These factors affect the TBM construction from the four aspects of the geological risk, equipment risk, management risk, and technical risk encountered during the construction of a full-face mechanical excavation of the tunnel. A five-level hierarchical directed graph is extracted by matrix operation to determine the relationship between factors. In the case of incomplete data, the Bayesian network model can be established by employing the prior probability to describe the state of the root node through expert scoring, ideal for system predictions lacking data. Finally, the main risk factors affecting the TBM construction are introduced and discussed, and the TBM construction of the West Second Tunnel of the Beijing Water Supply Phase II Project is taken as an example to compare the evaluation results with those of the actual construction status. The obtained results are not much different from those of the actual situation, which proves that the proposed method is scientific and can provide a reliable basis for analogous projects.
construction & building technology
What problem does this paper attempt to address?
This paper mainly discusses the risk analysis method of tunnel boring machine (TBM) construction based on fuzzy Bayesian network. In view of the diversified risk factors, complex systems, and frequent shutdown accidents in TBM construction, the author proposes a method of selecting 18 risk factors to construct an interpretive structural model. These factors affect the construction of full-face mechanical excavation tunnel from four aspects: geological risk, equipment risk, management risk, and technical risk. The relationship between factors is determined by matrix operation to extract a five-level hierarchical directed graph. In the case of incomplete data, the Bayesian network model is established using expert scoring, which is suitable for system prediction with insufficient data. Finally, the evaluation results are compared with the actual construction conditions of the West Second Line Tunnel of the Beijing Water Supply Phase II project to demonstrate the scientificity and feasibility of the proposed method and provide reliable basis for similar projects. The study also compares various risk analysis methods and selects fuzzy mathematics comprehensive evaluation method and Bayesian network method, emphasizing the advantages of Bayesian network in dealing with complex systems and incomplete data.