Analysis on the active power filter of PI iterative control strategy

During recent decades, heavy industry (mining, petrochemical, aluminium, pulp and paper ,etc) has installed a large number of static converters for the control of large loads. Nonlinear loads such as traditional diode/thyristor rectifiers with capacitive and inductive loads generate harmonic and reactive current, which leads to poor power factor, low energy efficiency, and harmful disturbance to other appliances. As a result, harmonics is the most noticeable power quality problem, and many other power quality issues are associated with the variation of reactive power consumption. Various types of compensators or conditioners have been proposed for this purpose. Passive filters were used to suppress harmonic and improve power factor in industry system conventionally, but it has some serious problems .Series and parallel resonance may occur between the system impedance and passive filter, which will amplify the harmonic current and voltage and sometimes result in damaging passive filters as well as make high voltage breaker trip off occasionally. Active power filter (APF) is one of them and has been put into many field installations especially in Japan because of its almost perfect performance. Active power filter offers a kind of important way for harmonic suppression, which can compensate harmonic current to meet impedance and frequency change automatically. To suppress effectively harmonic, a kind of control strategy of APF is important. In this paper, the iterative learning control strategy based on PI learning law in APF is proposed. In this new type, the system robust is strengthened by using forgetting factor learning law, at the same time the D-type learning law forward-feed loop of fuzzy current reference errors under optimizing aim is introduced, both of which improve the tracking precision of the system. Finally the results of simulation and industrial application are given to verify that the proposed control strategy has good characteristic of reactive pow- - er compensation and harmonic filter.