A Computational Framework for Predictive Risk Assessment of Shield Tunnel Construction

Shield tunnelling is a complex and challenging process that is prone to risks. This study proposes a novel risk assessment framework on the basis of geological conditions and shield parameters to enhance the safety and efficiency of tunnelling. The proposed framework integrates the multi-grained cascade forest (gcForest) and fuzzy set pair analysis (FSPA). The former is utilized for the prediction of shield operational parameters based on the geological conditions, while the latter is employed to assess the risk level based on the data characteristics of predicted operational parameters. The objective weights used in the FSPA are obtained from the intercriteria correlation. A case study of a field tunnel project in Guangzhou is conducted to validate the assessment framework. The results indicate that the predicted risk levels agree well with the real engineering observations. The proposed framework is a practical tool for assessing risk during tunnelling, contributing to improved project management and risk mitigation.