The ability of power systems to maintain functionality under earthquakes is critical to societal operation and post-disaster recovery. However, insufficient characterization of component failure correlations and simplified system operation modeling limit the accuracy of seismic performance assessment. This study proposes a seismic performance assessment framework for regional power systems that jointly considers spatially correlated ground motions and correlations among component seismic capacities. Component capacity correlations are constructed using correlation matrices and Cholesky decomposition, and are integrated with the spatial correlation of ground motions to generate post-earthquake component damage states reflecting these interdependencies. By integrating islanding identification with optimal power flow analysis, the post-earthquake operational performance of the system is quantitatively evaluated. Furthermore, a Gaussian process regression (GPR)–based surrogate model is introduced to quantify performance bias under different correlation structures. The results show that under low-intensity earthquakes, direct current optimal power flow (DC-OPF) and alternating current optimal power flow (AC-OPF) provide similar assessment results, whereas under high-intensity earthquakes, DC-OPF significantly overestimates the seismic performance of the power system. System residual power is strongly governed by component capacity correlations and seismic intensity. Assuming fully independent component capacities results in underestimation of both the mean and variability of system power output under strong earthquake conditions. The GPR model enables rapid prediction of the mean and standard deviation biases of system residual power under different correlation structures and earthquake magnitudes. These studies provide a feasible framework for assessing the seismic performance of complex power systems.
Qiao et al. (Mon,) studied this question.