The Research Of Active Power Filter Based On Neural Network Harmonic Detection

With the rapid development of the industrialization and informatization process and the wide application of power electronic technology, harmonic pollution is more and more serious for power grid. Its impact mainly includes the following two aspects:first of all, the quality of the power grid would been reduced, which will cause some threats to the property and safety of the user; second, in the power grid harmonic magnetic field might produce an inestimable impact for the electrical equipment in the around. So it is necessary to take some measures to suppress harmonics. In the mass harmonic suppression methods, in the power grid with harmonic adding active power filter is relatively simple and compensation effect is also very good.In this paper, innovation research the harmonic detection method, make harmonic detection and compensation effect be further improved.This paper firstly introduces the working principle of active power filter and topological structure, further analyses some traditional harmonic detection methods, Summarizes their advantages and disadvantages. On this basis this paper sets up the active power filter system whose harmonic detection is neural network detection method. In other words, take the load current and its corresponding phase in each cycle as the input of BP neural network, and the fundamental current as its output, then make the fundamental current and load current subtract and get harmonic current; The compensation current part adopts hysteresis control method which is simple and good real-time. Finally, in Matlab/Simulink environment simulation, the simulation results show that this method has a good compensation performance, Effect is better than the traditional harmonic detection method based on instantaneous reactive power theory, so that the performance of the active power filter is improved.