Risk uncertainty analysis in shield tunnel projects

Abstract

Subway shield tunnels usually pass through the complex geological environment of a city. There are many uncertain factors in shield tunnel construction risk, undoubtedly increasing the difficulty of construction safety control. Risk evaluation prior to project construction is necessary for the implementation of tunnel projects and reduces risk levels during construction. However, tunnel risks are uncertain and are characterized by random and fuzzy uncertainty. To account for these two uncertainties in the quantitative transformation of risk qualitative concepts and to improve the evaluation accuracy, in this study a new risk evaluation method based on a two-dimensional cloud model (TDCM) is proposed for shield tunnel projects. First, the risk evaluation criteria are quantified with the harmfulness and the probability of risk as the evaluation dimensions of the two-dimensional cloud, and the conversion relationship between the risk level and the quantitative domain is improved. A risk evaluation index system based on construction accidents and risk types for shield tunnel projects is established. Then, the Delphi method is used to evaluate the probability and harmfulness of the risk, the index weight and numerical characteristics of the two-dimensional cloud model are determined, and a shaped two-dimensional comprehensive cloud of thousands of cloud droplets is generated by a positive cloud generator. Finally, the domain is defined as the overlap between the comprehensive evaluation cloud and the standard cloud at different risk levels, and the contribution of comprehensive evaluation cloud drops to the comprehensive evaluation cloud in each domain is analysed. The risk level of the project is predicted based on the principle of maximum membership degree. The risk evaluation method proposed in this study is applied to two subway shield sections, and the risk level of these projects is accurately predicted. The TDCM evaluation method is compared with the one-dimensional cloud model (ODCM) evaluation method and the fuzzy comprehensive evaluation method (FCEM), and the advantages and applicability of the TDCM evaluation method are discussed.