The Application of Improved Independent Component Analysis in Identification of Harmonic Sources

Harmonic pollution has always been an important issue in power qualities. The key to harmonic administrator is to identify the harmonic sources in the power grid and find the locations as well. In this paper, the harmonic source identification problem is regarded as the blind source separation problem. The mathematical model and basic principle of independent component analysis (ICA) are discussed. The particle swarm optimization (PSO) algorithm based on gradient acceleration is introduced into independent component analysis. The principle and basic clue of improved independent component analysis are proposed. The improved independent component analysis method is applied here to identify the harmonic source. Finally, the simulation is carried out in the IEEE-14 node harmonic test system. The results show that the method is effective for harmonic sources identification when the network parameters and topologies are unknown.