Classification of single and combined power quality disturbances

In this paper, a new classification method that is based on discrete wavelet transform (DWT) and adaptive neuro-fuzzy inference system (ANFIS) is proposed to classify power quality disturbances. First, the multiresolution signal analysis technique of DWT and Parseval´s theorem are employed to extract discriminating features of the disturbance signal. Then, the proposed classifier system can identify the type of problem. Several simultaneous and combined disturbances can be successfully recognized using the proposed approach. The system design outlines are addressed and the diagnosis algorithm is described. The adopted feature vector consists only of four elements which much reduces the computational burden and speeds-up the system response. Four types of combined disturbances and eight types of single disturbances are efficiently diagnosed on testing the system with a large number of power quality events.