Adaptive Predictive Current Control Of Active Power Filter Based On Parameter Identification

In terms of the issues of the increasing harmonic pollution,the active power filter(APF)has higher compensation precision and faster response capability in real time compared with the generalized passive power filters.The APF,which consists of three-level H-bridge topology can significantly improve the system compensation capacity and effect so that it is widely applied in the high-voltage applications.Hence,a diode-clamped three-level APF with related control technologies is preferred and studied in this dissertation.First,based on the basic working principle of APF,the working modes and the mathematical model of APF are analyzed and established respectively.Subsequently,the control system transfer function of the predictive current control method is calculated,from which the instability of system is demonstrated as the inductance value of the inductor changes to degrade its harmonic compensation performance.The simulation results confirm the aforementioned issue.In addition,based on the aforementioned issue,a least squares method-based online parameter identification scheme is studied in this dissertation.With respect to the condition of changed inductance of the inductor,the recursive least squares method with forgetting factor is employed to establish the APF model of identifying ac-side inductance value.Furthermore,a parameter identification-based adaptive predictive current control strategy is proposed.The proposed strategy can maintain the awesome harmonic compensation performance of the APF even when the inductance value of the inductor changed.Its effectiveness is verified by simulation results.Per analyzing the calculation structure of the parameter identification module,the proposed method is totally coded using Verilog-HDL including the IP core of the matrix multiplication module and the parameter identification module.Prior to the overall control program design,the correctness of the designed modules is demonstrated by simulation results.Finally,a small-scaled prototype is setup to further validate the proposed method.Based on the experimental results,the parameter identification module can accurately identify the inductance value at ac side and the adaptive predictive current control strategy can maintain great harmonic compensation performance when the inductance value of the inductor changes.Simultaneously,the controller system has the capability to compensate unbalanced loads,which confirm its great dynamic performance as well.