Solução de problemas de fluxo de potencia mal condicionados através do método GMRES incluindo controladores FACTS

Abstract

In this work is investigated the performance of the GMRES Method in studies of power flow extremely loaded, ill-conditioned and including FACTS Controllers, focusing on Jacobian matrix with high numbers of conditioning and real eigenvalue close to zero, in order to consolidate the virtues of GMRES in ill-conditioned problems. From this information you can set operating scenarios that may indicate points of robustness and GMRES efficiency. Consolidating test systems (hypothetical and actual) of difficult convergence associated with extremely high loaded, is also part of research to create s cenarios close to reality, or that may occur in practice. The primary purpose of the inclusion of FACTS is to assess the impact that these devices can cause the Jacobian matrix. These characteristics are informed through eigenvalues and conditioning number. In addition, we propose an iterative solver developed in MATLAB platform for power flow studies and a detailed investigation of the influence of the reordering and the e limination of non-null elements in the quality of the preconditioner of incomplete factors. The proposed methodology was applied to hypothetical and real systems with FACTS controllers, and the results of the experiments showed the greater efficiency of the iterative solver GMRES in most simulations. It was found that the larger the size of the electrical system and the worse their numeric characteristics, the better its performance.