Comparative study of polar and Cartesian co-ordinate algorithms for power-system state-estimation problems

The paper presents a detailed study of the convergence characteristics and an assessment of the performance of three algorithms used to solve the power-system state estimation problem. These algorithms are the polar co-ordinate formulations of the full and fast-decoupled weighted least squares (WLS) and the Cartesian co-ordinate WLS. These algorithms are studied and compared to determine their convergence patterns on several measurement sets consisting of line flows, busbar injections, voltage magnitudes and the network ill-conditioning. The algorithms were tested on different IEEE test systems. Results are presented to illustrate the convergence performance.