Solving the nonlinear power flow equations with an inexact Newtonmethod using GMRES

摘要

This paper presents a detailed investigation into theneffectiveness of iterative methods in solving the linear systemnsubproblem of a Newton power flow solution process. An exact Newtonnmethod employing an LU factorization has been one of the most widelynused power flow solution algorithms, due to the efficient minimum degreenordering techniques that attempt to minimize fill-in. However, the LUnfactorization remains a computationally expensive task that can benavoided by the use of an iterative method in solving the linearnsubproblem. An inexact Newton method with a preconditioned GeneralizednMinimal Residual (GMRES) linear solver is presented as a promisingnalternative for solving the power flow equations. When combined with angood quality preconditioner, the Newton-GMRES method achieves a betternthan 50% reduction in computation, compared to Newton-LU, for twonlarge-scale power systems: one with 3493 buses and 6689 branches,nanother with 8027 buses and 13765 branches