Research On The Control Strategy Of Three-phase Four-leg Active Power Filter

As a large number of non-linear loads are applied to the low-voltage three-phase four-wire power system, the power quality problems about current harmonic and three-phase unbalanced have become increasingly prominent. Active power filter(APF) receives much concern for having good dynamic property, compensation flexible and compensation feature which is not affected by power grid impedance. This thesis focuses on the harmonic current detection method, current tracking control strategy and the DC-link voltage balancing control strategy of three-phase four-leg shunt APF.Firstly, an improved dq harmonic current detective method based on virtual flux and integral filter is given. On one hand, the virtual flux is used in phase locking method, in which the positive sequence voltage extractor is set in front of flux observer and the cross compensation is set in low pass filter’s input terminal, in order to improve the algorithm’s adaptability in complex power grid conditions. On the other hand, this method has been applicable to three-phase four-wire system by replacing LPF function with integration filter and modifying the integral cycle, it also can reduce the detection delay and improve the detection accuracy. The simulation results confirm that this improved method is effective.Secondly, two kinds of current control methods are compared. Then, the vector control method based on quasi-PR+3D-SVPWM is researched, and its control flow is analyzed. The design approach of MPCC controller is given. The method of two step predictions is adopted to compensate control delay. Aiming at the problem that model mismatch will influence compensation effect, so a simple inductance parameter identification method is applied to modify parameter. The simulation results are analyzed under three conditions. They are normal operation, model mismatch operation and the operation after introducing inductance observation, respectively. The simulation results show that compared with vector control strategy, the four-leg APF based on MPCC strategy has more prominent advantages of dynamic response and control process simplification.Thirdly, referencing on the thought of 3D-SVPWM modulation based on αβγ coordinate, a improved model predictive voltage control(MPVC) based on partition judgment is given. The given voltage vector position is determined by making two partition judgment processes, and the optimal subset is selected from 24 groups of alternative switch state subsets. In this way, the optimizing efficiency is improved and the time-consuming of calculation is reduced. The influences of delay coefficient and inductance mismatch coefficient on MPVC control system stability are discussed. The evaluation function which contains constraint switch frequency is established, and multi-objective optimization control method is given. The simulation results confirm that the improved MPVC strategy is effective, and multi-objective optimization control method can effectively reduce the switching frequency.Finally, the DC-link voltage balancing control strategy considering grid voltage fluctuation is given. Combined with the advantage of fuzzy control which can realize the online parameters adjustment, a kind of fuzzy-PI controller is designed. Compared with PI control, the fuzzy-PI control method has the advantages of fast dynamic response and strong robustness. In view of the problem that the grid voltage’ fluctuation will influence compensation effect, a kind of droop fuzzy-PI control strategy is proposed. According to the fluctuation of the grid voltage, the fuzzy-PI controller is responsible for the tracking control of the given DC-link voltage and the droop regulator optimizes this voltage. Simulation results show that the proposed method can improve the adaptability and the dynamic performance of voltage loop, at the same time, can also reduce the overshoot of the DC-link voltage effectively.