Evaluation of Water Inrush Hazard in Karst Tunnel Based on Improved Non-Linear Attribute Variable Weight Recognition Model

Water inrush in karst tunnels will cause casualties and economic losses. Thus, it is significant to objectively assess the water inrush risk level and adopt valid preventive measures to reduce losses from this disaster. The relationship between the factors affecting water inrush in the dynamic coupling system is strong nonlinear, so the attribute recognition model, which lessens the mutation points and error and causes the evaluation results to be more reasonable and accurate, is improved nonlinearly in this paper. Firstly, the assessment system was established by selecting seven factors: formation lithology, unfavorable geological conditions, attitude of rock formation, landform and physiognomy, contact zones of dissolvable and insoluble rock, layer and interlayer fissures, and groundwater level. Secondly, the multi-factor interaction matrix, C-OWA operator, and variable weight theory are used to calculate the constant weight and variable weight of each evaluation index. In addition, the linear attribute measurement function of the attribute identification model is optimized by using the non-linear trigonometric function to distinguish the risk level of the water inrush. Finally, the proposed model was successfully used in Qiyueshan Tunnel. The evaluation results of the risk level are more accurate than other methods, and they are in agreement with the excavation results. The proposed model provides a valuable reference for the risk assessment and project management of tunnel construction.