Currently, gas disasters in tunnels occur frequently and cause severe casualties and property losses. However, there are few studies on the risk assessment of gas tunnels, especially comprehensive analysis of the safety risk of gas tunnels considering multiple factors. This paper presents a comprehensive assessment method for safety risk of gas tunnel construction based on fuzzy Bayesian network. The results of safety risk assessment are determined by two aspects: risk probability and risk hazard. In the process of assessment, 12 risk indicators were extracted from 32 important factors by the principal component analysis method, and a fuzzy interpretative structural modeling was established and converted into a fuzzy Bayesian network to evaluate the risk probability. The risk index weight was obtained through principal component analysis, and the risk hazard level was calculated by using the fuzzy comprehensive assessment method in combination with the hazard degree of the risk index. After a comprehensive assessment of gas concentration and other field conditions, the relevant parameters such as the fuzzy interpretative structural modeling and fuzzy Bayesian network are determined. This paper selects two construction segments of the Like tunnel as case studies to verify the feasibility and application prospects of this method. The method establishes a safety risk assessment model for gas tunnels, which can allow management personnel to fully evaluate gas tunnel risk and provides a safe construction environment.

Blocking seepage channels in underground engineering construction can easily cause changes in groundwater level and lead to a series of problems of deterioration of the geological environment. This paper uses the physical model test system and the preparation method of the corresponding stratum similar materials to carry out the physical model test of the application of the water fluctuation and diversion measures represented by the subway station. In this way, the excavation process of the subway structure is restored and the influencing factors of the water level fluctuation are studied. Firstly, according to the similarity theory, clay-similar, sand-similar and rock-similar materials were prepared using sand, clay, blanc fixe, iron powder and other raw materials. The simulation of six test conditions under four hydraulic gradient conditions was carried out, and then the solution of water-level fluctuation was proposed and tested. Finally, the influencing factors of water-level fluctuation are analyzed. The results of the study show that the largest water-level fluctuation occurs at the monitoring points on the upstream surface. The diversion measures adopted in this paper have obvious effect on the relief of water level yong height. The measured variation of water level in this experimental environment is 0.54m.