Study on Parallel Algorithms for Power System Small Signal Stability Based on PC Clusters

With the interconnection of power systems, inter- area low frequency oscillation occurs frequently, insufficient damping of the weak-interconnected power grid has become a serious issue affecting power system stability. Considering the high rank of algebraic and differential equations and the geographical distributed data of interconnected power systems, it is necessary to analyze small signal stability problems in parallel mode with distributed computers. Based on the ´Multi-port Inverse Matrix parallel algorithm´ proposed by China EPRI, the parallel implementation methods of the simultaneous iteration algorithm and the Inverse Iteration/Rayleigh Quotient Iteration algorithm are described in detail in this paper. With the characteristics of limited communication time and good load balancing, the proposed parallel algorithms can run well on distributed PC clusters. Numerical simulation results on actual large-scale power systems show that the proposed algorithms are correct and efficient.