Developing an evaluation model based on unascertained measurement for evaluation of tunnel squeezing

Tunnel squeezing brought great difficulties to the construction and severely threatened the safety of on-site operators. The researches regarding large deformation evaluation have been widely developed, but actual conditions of tunnels are considerably complex, producing a large variety of uncertainty information existing in the evaluation process. Therefore, we constructed an unascertained measurement model incorporating four membership functions for evaluation of tunnel squeezing based on the collected datasets. Simultaneously, information entropy was introduced to objectively calculate the index importance for each index. For the first group data (GPI), the accuracy associated with four membership functions are 100%, 83.33%, 50%, and 83.33%, respectively, while the accuracy of GPII are 70%?77.5%?67.5%, and 70%, respectively. Linear function and parabolic function show better performance on uncertainty information interpretation according to the classification results. The results revealed that the uncertainty model constructed in this study can enrich the available uncertainty evaluation system.