An overall optimization strategy for novel hybrid parallel active power filters based on genetic algorithm

Power quality issues such as reactive power control and harmonics mitigation have been gaining wide attentions these years. This paper introduces a novel hybrid parallel active power filter (APF), in which the passive power filter (PPF) and the transformer-coupled APF connected in series. In the proposed hybrid APF, there is a trade-off between compensation performance and system investment. Traditionally, the parameters of the LC filter and APF are decided in an approximate and experience-dependent way. In this paper, an overall optimization design strategy is proposed aiming to reduce the system cost with the same power factor and THD requirements of the supply. Firstly, the models of PPFs and the APF are established respectively for rating evaluation. Secondly, the objective function of the whole investment in the parallel system is obtained based on assumed unit price of the resistors, reactors, capacitors and APF. These coefficients can be modified easily according to different prices for different components. Finally, the minimum value of the objective function is found out through a genetic algorithm (GA). With the parameters based on GA overall optimization strategy, the whole cost of the system decreased about 2.26 percent with less computation and better compensation behavior comparing with traditional design method. The novel hybrid APF is applied in an arc furnace.