Risk assessment of tunnel segment uplift during construction based on variable weight-cloud model

This paper presents a novel risk assessment framework for uplift movement of segmental lining during shield tunnelling. This framework integrates multiple objective weighting methods into cloud model theory, which has methodological reliability and engineering safety. A comprehensive index system is first established which comprise five categories of risk factors: geological, operational, design, material, and management. Objective factor weights are determined using the entropy weight method, coefficient of variation (CoV) method, and criteria importance through intercriteria correlation (CRITIC) method. Then, factor weights are synthesized using cooperative game theory to derive robust constant weights to overcome the limitations of single-method biases. Furthermore, a real-time variable weight approach dynamically adjusts indicator importance and improves the reliability of risk predictions. Additionally, cloud model theory and the maximum membership principle convert monitoring data into probabilistic risk levels to quantify tunnelling uncertainty. The proposed framework is validated through its application to the Pazhou branch tunnel of the Sui-Guan-Shen Intercity Line, where the predicted segment uplift risk levels show strong agreement with the field observation. The results demonstrate the capability of the proposed framework for dynamic, data-driven risk assessment and its practical value in guiding safe shield tunnelling operations.