Prediction equations for the standardized version of cumulative absolute velocity as adapted for use in the shutdown of U.S. nuclear power plants

The U.S. Nuclear Regulatory Commission (USNRC) uses the standardized version of the cumulative absolute velocity (CAV) together with the response spectra of the recorded ground motion at a site to determine whether a nuclear power plant must be shut down for inspection after an earthquake. In order to better understand the impact of these criteria on potential plant shutdowns, we used several subsets of the PEER-NGA strong motion database to develop empirical prediction equations between a CAV intensity measure that incorporates these criteria and the geometric mean horizontal component of CAV. This particular approach was used because, after applying the USNRC shutdown criteria, we found that there were an insufficient number of records remaining to reliably develop a ground motion prediction equation (GMPE) for this new CAV intensity measure directly from the physical parameters of an earthquake. We use these prediction equations to demonstrate how a lower-bound value of this intensity measure can be used to prevent non-damaging ground motions from both nearby small-magnitude earthquakes and distant large-magnitude earthquakes from contributing to the ground motion hazard computed from a probabilistic seismic hazard analysis (PSHA). We suggest that the use of a CAV criterion, in addition to a criterion based on either peak ground motion parameters or response spectral ordinates, could also be used to prevent the unnecessary shutdown of nuclear power plants outside of the U.S. Furthermore, with some adjustment, a similar approach could be used to rapidly assess the potential damage to conventional structures after an earthquake to aid in emergency response and loss assessment.

► A duration-based ground motion damage potential parameter is defined (CAVDP).
► CAVDP incorporates shutdown criteria used by U.S. Nuclear Regulatory Commission.
► Prediction equation for CAVDP is developed using PEER NGA strong motion database.
► CAVDP is shown to be a more reliable parameter than standardized version of CAV.
► Using a minimum CAVDP removes non-damaging events from contributing to a PSHA.