Parallel implementation of predictor-corrector methods for power system transient characteristic studies

For providing detailed power system transient characteristic analysis, more rigorous system representations are generally required. Consequently, the computational burden will become heavier. To speedup the solution time while maintaining the required accuracy, an attractive way is to implement the parallel computation techniques on multiprocessor architecture. This paper investigates the applications of two parallel predictor-corrector algorithms, namely the multistep predictor-corrector algorithm (MPG) and the block predictor-corrector algorithm (BPC), on a generalized power system transient simulation program running on both the 2-CPU Alliant FX/80 and the 4-CPU Convex C3840 systems. The IEEE Second Benchmark Model System 2 is selected for detailed numerical verifications, and the cross comparisons and discussions are also provided. From these illustrations, valuable guidance can be obtained for the associated multiprocessor parallel processing schemes applied to power system transient characteristic analyses.