Study On LCL Active Power Filter Based On MPC Compound Control

With the rapid development of the power electronics industry,the harmonic pollution of the power grid is becoming gradulally serious.As a kind of power electronic device,active power filter can dynamically compensate the harmonics and reactive power of the power grid,and has received widespread attention.This thesis takes the LCL type APF as the research object and analyzes the stability of LCL filter and the range of active damping parameters.In order to improve the compensation performance of APF,on the basis of traditional model predictive control,the improved model predictive control method and the MPC compound control method embedded with repetitive controllers are studied,and theoretical analysis and simulation experiment verification are carried out.The main work can be summarized in the following four aspects.(1)Analysis of mathematical model,current feedback strategy and resonance suppression method of LCL type APF.According to the topological structure of the LCL active power filter,the mathematical model of the system in the two-phase static coordinate system and the two-phase rotating coordinate system is deduced,the parameter range of the LCL filter is calculated.Comparing the stability range of the system under the current feedback control strategy of the inverter side with the current feedback control strategy of the grid side.It is concluded that when there is a delay in the system,the grid-side current feedback control strategy may be stable without adding any active damping.And designed an active damping scheme based on the capacitive current feedback to suppress the resonance peak of the LCL filter.(2)Multi-vector improved MPC strategy and vector synthesis improved MPC strategy research.According to the current control algorithm,model predictive control is a control method which is easy to apply to nonlinear systems and its dynamic response speed is fast.However,the steady-state accuracy will be limited by control frequency.In order to improve the steady-state compensation accuracy of the active power filter,a multi-vector improved model prediction method and a vector synthesis improved model prediction method are proposed based on the traditional model predictive control.The improved multi-vector MPC method reduces the current tracking error by adding two sets of switching voltage vectors.The improved MPC method of vector synthesis optimizes the cost function of the improved multi-vector MPC method,eliminating the calculation process of cyclic optimization and further improving the current control accuracy.(3)Study on MPC compound control strategy based on repetitive control.The improved MPC methods of multi-vector and vector synthesis effectively enhance the current control accuracy.However,due to the inherent delay of the system,the steady-state accuracy is still not ideal.Therefore,the method of embedding repeated control links in the improved model predictive control of vector synthesis is proposed to improve the accuracy of steady-state compensation.At the same time,it ensuring the response speed of the system as much as possible.It also improves the accuracy of steady-state compensation and optimizes the performance of the active power filter.The design of the voltage outer loop adopts a typical type II system correction method and introduces a multi-rate link to reduce the utilization of digital system resources.(4)Building a simulation model and experiment platform of LCL type APF for simulation analysis and experimental verification.Building a simulation model of LCL type APF in Matlab/Simulink and combining the theory to simulate,analyze the system under different algorithms and parameters are the basis of an experimental prototype with TMS320F28335 as the control core was built for physical experiments.In the experiment,the method of model based design is used to generate the top-level software code.The experimental results verified the feasibility and effectiveness of the designed scheme.