Parallel processing of sparse power system equations

Algorithms for parallel computation of the sparse matrix equations found in power system modelling are discussed and evaluated. A number of matrix factoring algorithms have been applied to a 60-busbar power system network with useful results. The greatest speed-up was achieved by a method that matches the number of parallel activities to the number of processors available. The PERT-based scheduling algorithm works well for small problems and should give useful results for larger networks