The Analysis Of Measurement For Power System Harmonics And Inter-Harmonics Based On Neural Networks

Along with the rapid development of the modern industry and the information technology, electric power's users more and more pay attention to the requirements of the quality of electric power day by day. But nowadays with extensively using of nonlinear electric equipment, the problems of electric power system harmonic and inter-harmonic pollution have become more and more serious. Because in the power system, nonlinear load produced not only the fundamental frequency's integral multiple harmonic beside, but also may produce the fundamental frequency's un-integral multiple, meaning inter-harmonic. Harmonic and inter-harmonic detection and analysis precondition of the harmonic treatment, precise harmonic and inter-harmonic detection will provide good basis for harmonic treatment.In recent years, the algorithm of neural network can be widely used in the harmonic detection, through multiple training the way renewal power vector repeatedly, and improves effectively the signal estimate of the power system.This paper describes the causes of harmonic and harmonic harm, and the current management method of harmonic treatment. Also introduces the development history of artificial neural network theory and its current application status.In order to solve the current problems existed in kinds of harmonic analysis method, a neural network approach (NNA) was proposed for estimating accurately harmonics and inter-harmonics parameters from the signals in power systems. It is aimed at the system in which the sampling frequency cannot be locked on the actual fundamental frequency. By training the frequencies, magnitudes and phases of the fundamental wave, the harmonics and the inter-harmonics based on the NNA, an accurate harmonics and inter-harmonics measurement result can be obtained as the algorithm converge. The simulating result shows that the estimated harmonics and inter-harmonics parameters can be measured at accuracy of 99.995% under actual fundamental frequency varying from 49.5 to 50.5 Hz.Meanwhile, in the article proposes a kind of method of combining FFT and neural network, meaning correct the results of harmonic analysis through FFT algorithm, so get the power system harmonic frequency, and by using the neural network to calculate interpreting value of harmonic wave amplitude and phase. Using this method we could measure accurately between the amplitude and phase of kinds of harmonics, and through the simulation example analysis proves that this method is efficient.