Recursive cooperative swarm optimization for harmonic analysis of time-varying power signal

Harmonics make a great impact on the stable and safe operation of power system in modern society. Empirical mode decomposition enhanced with masking signals could properly extract harmonic components from time-varying power signals, while the parameters for constructing masking signals display a strong influence on the decomposition accuracy. In this paper, recursive cooperative swarm optimization is proposed to search for the proper set of parameters. This approach could locate the region where the best solution exists discretely and recursively, and the best approximate set of values in this region could be achieved continuously. These procedures, which narrow down the optimizing space gradually, implement alternately according to the fitness changes to approach the optimal results. Meanwhile, the Chaos theory and the self-adaptive weighting are introduced to enhance the optimization accuracy and convergence rate. The experiments demonstrate that the introduced approach performs better in the accuracy and convergence rate in optimizing multidimensional parameters, and it effectively improves the accuracy of harmonic analysis.