Improved unascertained measure model for risk evaluation of collapse in highway tunnels

Tunnel collapse is a serious hazard that poses a substantial threat to worker safety and construction expenses. The results of risk evaluations can provide valuable information for collapse prevention and risk management. For this purpose, an improved unascertained measure function for collapse risk evaluation was proposed. Firstly, a risk evaluation index system was established using major indexes, including geological conditions, design and construction factors. Secondly, a combination method of the analytic hierarchy process (AHP) and the anti-entropy weight (AEW) based on the maximum deviation (MD) concept was proposed. The subjective and objective weights were linearly optimized to assign the index weights. Finally, the conventional linear unascertained measure function was improved by normalizing. By combining the weights, the values of the evaluation indexes were regarded as intervals for calculating the measure values. Confidence recognition analysis was used to determine the collapse risk level. Jinzhupa Tunnel was evaluated to have a high collapse risk and experienced a small-scale collapse during excavation, verifying the feasibility of the improved unascertained measure model. This method considers the uncertainty of the evaluation index and can effectively predict the risk level of tunnel collapse.