Abstract: To reduce the computational complexity and long training time in relevance vector machine(RVM),this paper proposed an optimized algorithm based on fast relevance vector machine(FRVM),which not only greatly reduces the training time of relevance vector machine,but also improves its classification accuracy.This method is applied to the classification of power quality disturbances.Firstly,the wavelet transform is applied to analysis the time-frequency features of the power quality disturbances,and the difference of the energy of the wavelet transform signal in each layer and the standard signal energy is used as feature vector.Secondly,FRVM is used to classify the feature vector to realize power quality disturbances classification based on wavelet transform and FRVM.The simulation verifies that this method can classify all kinds of power quality disturbances,and has higher classification efficiency and accuracy than the classical RVM.

Key words: Power quality,FRVM,Disturbance classification