APPLICATION OF CLUSTERING AND FACTORIZATION TREE TECHNIQUES FOR PARALLEL SOLUTION OF SPARSE NETWORK EQUATIONS

Through the use of diakoptical techniques it is possible to solve the network equations for a power system using a parallel computer archite cture. Traditional parallel solution techniques give an unimpressive r eturn on investment as more processors are applied to the problem owin g to unacceptably large sequential sections of the algorithm and ill-b alanced computational loads. The paper analyses the diakoptical method using the factorisation tree as a tool. New algorithms are proposed f or network splitting, node ordering and load balancing between process ors. This, in connection with a novel recursively parallel solution al gorithm, gives improved speedup over the conventional parallel solutio n method. Simulation results are presented which confirm the usefulnes s of the method.