Abstract: Subsea tunnels that traverse fault zones face significant structural safety risks due to the high variability of surrounding rock and the influence of overlying seawater. To address this challenge, the authors propose a comprehensive analytical framework for assessing structural safety risk. A case study is conducted on the Qingdao Jiaozhou Bay second subsea tunnel, in which six primary parameters from ten fault zones along the marine section are statistically analyzed, establishing a fault-zone database with 1 970 entries. Based on this database, the load and softness indices are introduced to quantitatively assess the magnitude of overburden loading and the deformation resistance of the surrounding rock, respectively. A validated three-dimensional numerical model is then employed to train a Kriging-based response surface, using the load and softness indices as inputs and the maximum crown settlement as the output, followed by a corresponding reliability analysis. By integrating the response surface with the reliability results, five structural safety risk grades are defined, and the relationship between the indices and the risk grades is established. Finally, the indices are applied to analyze the risk associated with the f12 fault zone crossed by the Jiaozhou Bay second subsea tunnel, resulting in the crown settlement and corresponding risk grades for the monitoring sections along the three tunnel lines. The main findings are as follows: (1) The load and softness indices approximately follow a normal and lognormal distribution, respectively, with a correlation coefficient of －0.483. (2) The exceedance probability of a 20 mm crown settlement when crossing a fault zone in the marine section is 5.28%. (3) The proposed rapid assessment method for structural safety risk grades demonstrates an accuracy of 90.47% in engineering applications.

Key words: subsea tunnel, fault zone, risk analysis, load index, softness index, monitoring data, Kriging-based response surface