Harmonics Current Detection in Three-phase Circuit using Neural Network

Abstract

Abstract Harmonics seriously affect the safety and economy of power system. In order to control harmonics effectively, harmonic current and harmonic power must be detected quickly. Based on the analysis of the components of three-phase nonlinear current, a neural network detection method of fundamental and harmonic current in three-phase circuit is proposed. The method uses a neural network instead of the low-pass filter in the instantaneous reactive power theory. The example results show that the proposed method has the advantages of less computation, better real-time performance and high detection accuracy. The method can be used to detect the positive-sequence (negative-sequence) fundamental active current, positive-sequence (negative-sequence) fundamental reactive current, harmonic, distortion current and harmonic power.