Harmonic Current Detection Based On BP ANN And Its Application In Active Power Filter Design

With the improvement of national economy in recent years, the electric power industry is developing with high-speed. But, the rapid development of power electronic technology makes a large number of power electronic devices applied to the society, and it makes harmonic pollution problems become more and more serious in power system. Harmonic current will produce serious harm to the power system and unnecessary losses. A variety of conventional compensation of harmonic suppression measures has been difficult to meet the needs of modern power system. Active power filter is proposed to adapt to the development of electric power system. As a new novel device compensating harmonics, APF is the focus in the field of power quality control.This thesis focuses on the three-phase shunt active power filter. The active power filter owes two parts. One is harmonic current detection, and the other is controlling the compensation current. Harmonic current detection is a key to improving the harmonic current compensation effects of Active Power Filter. Firstly, this thesis introduces the basic principle and topological structure of APF. Then, this thesis makes analysis among some conventional harmonic current detection methods, and introduces the harmonic current detection method based on the theory of instantaneous reactive power. In this thesis, harmonic current detection method based on ANN is presented. This method is brought for combining the noise canceling technology with artificial neural network. Not only its configuration is simple, but also its algorithm is easy. At last, the simulation results show the method is precise and fast.Compensating current is also an important key element for the APF. In this thesis, hysteresis voltage space vector technology is proposed. This method combines the hysteresis current control with voltage space vector control. It can overcome the problems that switching frequency based on hysteresis current control is not fixed and switching loss is large. It is also simpler than the traditional voltage space vector control.In the environment of MATLAB software, simulation models of three-phase shunt active power filter are built, and the simulation results show that harmonic current detection method based on BP neural network is effectively to compensate the harmonic current. It can effectively reduce total harmonic distortion.