Optimal selection of scalar and vector-valued intensity measures for improved fragility analysis in cross-fault hydraulic tunnels

Abstract

Seismic intensity measures (IMs) are crucial because they are significantly connected to earthquake hazards and structural responses. It means that selecting appropriate IMs is necessary for underground structures under unfavourable earthquake hazard conditions. The seismic design code and dynamic analysis have conventionally employed the peak ground acceleration (PGA) as an optimal IM, mainly referring to the seismic design code of aboveground structures. However, seismic investigation of underground structures using PGA is still controversial. Simultaneously, directly employing the PGA as an ideal IM may increase the uncertainty associated with ground motions, which in turn increases the uncertainty for probabilistic seismic demand analysis of hydraulic tunnels. As a result, it is important to understand the roles played by different IMs while analyzing the fragility of hydraulic tunnels. This work aims to conduct a comprehensive analysis to identify the optimal scalar- and vector-valued IMs for fragility investigation of cross-fault hydraulic tunnels. Therefore, the 1080 complete nonlinear dynamic time history analysis by surrounding rock-concrete lining-fluid interaction system is conducted to express the earthquake-induced drift ratio of a cross-fault hydraulic tunnel by incremental dynamic analysis. Subsequently, a refinement process is performed to determine the optimal scalar-valued IMs by comparing the correlation, efficiency, practicality, and proficiency of the 7 examined IMs. Meanwhile, the probabilistic seismic demand model of optimum vector-valued IMs is also developed and compared, followed by the three scalar IMs. Eventually, the difference between the fragility curves of the tunnel produced using the optimal scalar- and vector-valued IM is compared. The results demonstrate that compared with scalar two-dimensional fragility curves, the vector three-dimensional fragility surface may increase the reliable probability and minimize the uncertainty in analyzing the structural response hazard curve. The generated vector fragility surface can also estimate the seismic fragility of identical cross-fault hydraulic tunnels in an approximative manner.