On the modeling of ground-motion field for assessment of multiple-location hazard,damage, and loss: example of estimation of electric network performance during scenario earthquake

Consideration of within-earthquake ground-motion correlation is essential for the estimation of seismic hazards, damage, and loss for spatially distributed systems. In many seismically active regions, the strong motion data of real engineering significance are completely unavailable or very scarce. The absence of necessary data does not allow developing regional empirical correlation models, and the models obtained for other regions should be used with correspondent sensitivity analysis. The level of within-earthquake correlation may vary in broad range; therefore, development of correspondent criteria for selection from available models is important. In this paper, we analyzed the performance of a system of critical elements of electric power network (substations) depending on variations in within-earthquake correlation. The performance is described as probability of different levels of non-functionality, i.e., percentage of area suffering power outage, and probability of expected number of customers without power. We have shown that the proper choice of the within-earthquake correlation model is critical in comprehensive estimations of functionality of substations in electrical power system. Skipping the ground-motion variability and within-earthquake correlation may lead to unreliable results. Relevance of geology-based within-earthquake correlation models has been demonstrated, and a scheme, which allows reducing uncertainty in the choice of realistic correlation, has been proposed.