A hybrid particle swarm optimisation (psO) and genetic algorithm (GA) aided Kalman filter (psO-GA-KF) method for dynamic harmonic state estimation in the power system is investigated. The initial choice of states and ...
详细信息
A hybrid particle swarm optimisation (psO) and genetic algorithm (GA) aided Kalman filter (psO-GA-KF) method for dynamic harmonic state estimation in the power system is investigated. The initial choice of states and KF's parameters, such as process and measurement error covariance matrices significantly affects the estimation precision. Aiming at this problem, a combination of psO and GA (psO-GA) is proposed to optimise KF's parameters. In this hybrid algorithm, by replacing the gene mutationoperator of GA, the ps algorithm mutation operator is constructed, which can adjust the evolution direction and range based on historical records and swarm records, leading to avoiding the blindness of GA mutation and reducing the probability of devious mutation and enhancing the velocity of evolution. Thus, the global search ability of the hybrid psO-GA can ensure KF has more accurate estimation of harmonics. To test the effectiveness of the algorithm, several time-varying signals are simulated with harmonics and decaying DC components in the presence of amplitude drift and white noise. Simulation results show that the proposed hybrid algorithm has more accurate estimation, faster convergence and better robustness against noise in comparison with the conventional Kalman filter (KF), KF based on the maximum likelihood and psO aided KF (psO-KF).
暂无评论