Abstract:Considering the characteristics of multi-index, fuzziness, and uncertainty of collapse in mountain tunnels, the set pair analysis (SPA) method and information entropy theory were combined and applied for risk assessment of collapse. Firstly, the risk environment and risk factors of tunnel collapse were considered, and the main influencing factors, such as surrounding rock grade, groundwater, unfavorable geology, initial support stiffness, excavation method, construction technology level, and construction management level, were selected as evaluation indices. Second, the SPA method was introduced into the tunnel collapse risk assessment, and the certainty-uncertainty problem of tunnel stability-instability was analyzed from the perspectives of identity, discrepancy, and contrary, and a set pair system of evaluation indexes and classification standards was constructed. Then, the evaluation indices of collapse were divided into two main types: benefit type and cost type indices. Based on SPA, the single index connection degrees of evaluation indices were calculated. At the same time, the information entropy theory was introduced to determine the objective weight of the evaluation index, and it was coupled to the single index connection degree of the evaluation index, and the entropy weight-set pair comprehensive connection degree evaluation model of tunnel collapse risk assessment was constructed. Finally, the credible degree recognition criterion was used to determine the risk level collapse. Taking Binfang 2^# tunnel as an example, the shallow-buried section of the right tunnel K31+500～K31+660 was emphatically analyzed. The results showed that the determination uncertainty of tunnel stability instability based on the set pair analysis method could better solve the multi-index and uncertainty in the risk assessment of mountain tunnel collapse. The modified entropy weight method was used to determine the index weight, which reduced the influence of human factors; The confidence recognition criterion was introduced to identify the risk of tunnel collapse, which made up for the distortion of evaluation results caused by the maximum membership principle. The evaluation process of this method is reasonable, and the results are reliable. It is an effective collapse risk evaluation method and provides a new research idea for the collapse risk evaluation of mountain tunnel.