|Implementation of Temporal Parallelization for Rapid Quasi-Static Time-Series (QSTS) Simulations|
|Joseph A. Azzolini1, Matthew J. Reno1, Davis Montenegro2
1Sandia National Laboratories, Albuquerque, NM, United States
/2Electric Power Research Institute, Knoxville, TN, United States
Quasi-static time-series (QSTS) analysis of distribution systems can provide critical information about the potential impacts of high penetrations of distributed and renewable resources, like solar photovoltaic systems. However, running high-resolution yearlong QSTS simulations of large distribution feeders can be prohibitively burdensome due to long computation times. Temporal parallelization of QSTS simulations is one possible solution to overcome this obstacle. QSTS simulations can be divided into multiple sections, e.g. into four equal parts of the year, and solved simultaneously with parallel computing. The challenge is that each time the simulation is divided, error is introduced. This paper presents various initialization methods for reducing the error associated with temporal parallelization of QSTS simulations.