Study On Single Phase Active Power Filter Based On Current Prediction

Electric energy has become the most important energy in modern society. In the information age, single-phase electric equipment in all electrical equipment accounted for proportion is growing, especially in the civil power, nonlinear single-phase equipment accounted for a large part, causing great harmonic pollution to the single-phase grid. Single-phase active power filter (APF) is an effective solution to improve power quality. It can not only suppress the harmonic pollution effectively of all kinds of single-phase nonlinear devices, but also can be used in power quality control of a variety of three-phase unbalanced harmonic sources by combination. In this paper, the key technologies of the single-phase active power filter, harmonic detection and current tracking control algorithm are studied.The harmonic detection algorithm based on instantaneous reactive power theory has been widely used in three-phase APF. But this method cannot directly be used in single-phase circuit. In this paper, we first summarize some common harmonic detection algorithms used in single-phase circuits. Then three typical harmonic detection methods are studied, and compare their similarities and differences, and the performance of their low pass filter. Finally, the sliding window iterative DFT algorithm, which has high detection accuracy and low latency, is chosen as the harmonic detection algorithm of the single phase circuit.In this paper, deadbeat control is used as the current tracking control method. Firstly, the mathematical model of single-phase active power filter is established, and based on this model, the principle of deadbeat control is analyzed. It is concluded that the harmonic current prediction is the key of the control method.Based on repetitive prediction, the lateral prediction algorithm is proposed. According to the sampling value at the same time of the different power frequency cycle, the algorithm can predict the harmonic current value at the same time of next power frequency cycle. The lateral prediction algorithm has a higher prediction accuracy than repetitive prediction algorithm, especially for subharmonic current. The correctness and effectiveness of the algorithm are verified by the simulation in MATLAB.