Power system state estimation solution using modified models of PSO algorithm: Comparative study

Abstract

Abstract The objective of all power system state estimation (PSSE) is to determine, by statistical projections, the best estimate of state variables represented by the voltage magnitudes and voltage angles of all the buses. Due to the complexity and non-linearity of the power system (PS), it is necessary to use more advanced methods for its analysis and control in real-time environment. This research discusses the application and the comparison of hybrid models of one of the algorithm using artificial intelligence (AI) technique (particle swarm optimisation ‘PSO’) in minimising the raw measurement errors in order to estimate the optimal point of the PS when certain sensitive data are incomplete. The effectiveness of the hybrid models are demonstrated and compared with the original PSO, artificial bee swarm optimisation (ABSO) algorithm and genetic algorithm (GA) using IEEE 14, 30, 118 and 300 bus test systems. Newton-Raphson load flow solution is taken as benchmark. Two different objective function formulations assessed by PSWV (Particle swarm without velocity equation), EPSOWP (Enhanced particle swarm optimiser incorporating a weighted particle), PSO-RF (PSO with repulsion factor) and CLPSO (Comprehensive learning PSO). The first formulation is the Weighted Least Square (WLS) and the second one is the Weighted Least Absolute Value (WLAV).