In improve the recognition ability of equipment power quality complex disturbances

a new approach is proposed to deal with identification and classification of power quality complex disturbances based on Particle Swarm Optimization（PSO）and Extreme Learning Machine（ELM）with combined feature extraction.The combination of S-transform and Empirical Mode Decomposition（EMD）is employed to construct a set of composite feature vectors in order to represent the complex disturbances comprehensively.ELM training error is used as the fitness function of PSO to optimize the hidden layer neuron number.The initial parameters of PSO are set up to complete the classifier design

and the higher classification accuracy is maintained on the basis of improving the classification speed.Test results show that the new approach can accurately recognize seven types of power quality complex disturbances of the equipment power system

and has strong noise immunity.Compared with the non-optimized ELM

the new approach reduces the training and classification time

and improves the recognition accuracy.