Risk management for safety in mountain tunnel construction is of great significance. However, existing research lags behind engineering applications. In this paper, the risk of mountain tunnel collapse is used as an example to illustrate a new assessment method based on case-based reasoning, advanced geological prediction, and rough set theory. First, the risk surroundings and risk factors involved in tunnel collapse are integrated and summarized, and a risk assessment index system is established for tunnel collapse. At the same time, because the dynamic response parameters obtained by the advanced geological prediction usually indicate a typical geological structure, sensitive response parameters are introduced in the assessment index system. Advanced risk assessment can be performed for tunnel sections at a certain distance ahead of the tunnel face. Second, the major risk surroundings and the advanced geological prediction results are analyzed for the tunnel under assessment. Cases with similar attribute characteristics are selected via comparison with previous cases. Attribute reduction and calculation of weights are subsequently performed for the risk surroundings and risk factors of similar cases based on the attribute significance theory of rough sets. Finally, index screening and objective weights are applied in the fuzzy comprehensive assessment model. The results of this paper can be used to improve the theoretical level and reliability of risk assessment in tunnel safety and serve as a reference for tunnel construction.
Generative adversarial network for geological prediction based on TBM operational data
