Comparison of Neural Network and Fast Fourier Transform Based Selective Harmonic Extraction and Total Harmonic Reduction for Power Electronic Converters

A new strategy to estimate harmonic distortion from an AC line is presented for power electronic converters. An Adaptive linear neural network (ADALINE) is used to determine precisely the necessary currents in order to cancel harmonics. The proposed strategy is based on an original decomposition of the measured currents to specify the neural network inputs. This new decomposition is based on the Fourier series analysis of the current signals and Least Mean Square (LMS) training algorithm carries out the weights. This new estimation strategy appreciably improves the performances of traditional compensating methods and is valid both for single-phase and three-phase systems. The proposed strategy also allows extracting the harmonics individually. The method is based on the extraction of fundamental components of distorted line current using an ADALINE network. The output of the ADALINE is compared with distorted supply current to construct modulating signals and to generate PWM pulses for active line conditioner. Speed and accuracy of ADALINE results in improved performance of the active power line filter. In this paper converter is used as a non-linear load. The performance of ADALINE is verified with single phase fully controlled AC-DC converter simulated with neural network based active filter using MATLAB/SIMULINK. Also in this paper, harmonic components can be selectively extracted using ADALINE network and harmonic content analyzed by artificial neural network (ANN) is compared with the Fast Fourier Transform (FFT).