A multi-indicator evaluation method of shale gas tunnel hazard based on attribute recognition model and its application

• A multi-indicator evaluation method based on attribute recognition and combined weighting is proposed for shale gas tunnel hazard assessment. • Six key parameters related to shale gas enrichment and genesis were selected, with data obtained from laboratory and field tests in northwestern Hunan, China. • Combined weights were calculated using AHP, CRITIC, and a genetic algorithm, revealing organic matter abundance as the most critical factor. • The method showed strong agreement with actual incidents in the Leijiashan Tunnel and identified 8 high-risk tunnels out of 12, outperforming traditional evaluation approaches. The combustion and explosion of shale gas pose serious safety hazards to tunnel construction. Due to its self-generation and self-storage characteristics, existing evaluation methods are poorly applicable. To address this issue, this paper proposes a multi-indicator evaluation method for shale gas tunnels based on the attribute recognition model and a combined weighting method. Based on the enrichment patterns and genesis mechanisms of shale gas, six key parameters were selected as critical indicators for hazard evaluation. These parameter data were obtained by conducting laboratory experiments and on-site tests on tunnel samples collected from existing projects in northwestern Hunan, China. The analytic hierarchy process and the criteria importance though intercriteria correlation were used to determine the subjective and objective weights, respectively. The genetic algorithm was employed to calculate the combined weights. The results indicate that organic matter abundance has the highest combined weight (0.2215), making it the most important indicator. By analyzing the impact of each indicator on shale gas risk, a hazard classification standard and corresponding attribute measurement function were developed. Combined with the weight evaluation results, a shale gas tunnel hazard evaluation method was constructed. A case study on the Leijiashan Tunnel was conducted to validate the model, demonstrating strong consistency between the evaluation results and actual incidents. The method was further applied to 11 additional tunnels, identifying 8 of them as relatively high-risk. Compared with existing evaluation approaches, the proposed method exhibited superior applicability and accuracy, highlighting its strong potential for broader implementation in shale gas tunnel risk management.