Research On The Control Strategy Of Shunt Active Filters Based On Finite States Model Predictive Control

Currently,with the widespread using of non-linear loads on the grid supply,voltage and current are polluted.Power quality has become a serious problem,affecting industrial production and people’s lives.Active power filter(APF)is considered an ideal equipment to control power quality.While active power filter has been applied in the real world,but the treatment effect is not ideal,the root cause is time delay generated by digital controller reduces the accuracy of the system of compensation.Based on this,this thesis studies the time delay caused by digital control,adopting three-phase three-wire system parallel topology structure.This thesis firstly describes the topological structure of the active power filter,analyzes its working principle and from a mathematical point establishes its mathematical model.According to time delay problem,the thesis analyses the reasons and expounds its effect for the performance of the system of compensation.Model predictive control is considered to be the best means to solve the time delay,because the plant can be controlled in advance,but the computation of traditional model predictive is very large,and the active power filter is a quick real-time systems,sampling period is very short,so traditional model predictive must be improved before application.After a detailed analysis of the mechanism model predictive control,to solve above problems,by the inherent characteristics of converter(limited number of switch statuses),a control algorithm based on finite state model predictive control(FS-MPC)is adopted,significantly reducing the amount of calculation of traditional model predictive control,eliminating the delay of governance effect of impact by digital control.Traditional control of APF is based on the current control strategy;the whole system adopts double loop controllers,which respectively are the DC bus voltage control loop and harmonious reactive current tracking control loop.After the analysis of dispensability of double-loop control,this thesis designes a control algorithm based on direct power control(DPC)of the single-loop control strategy,combining the finite state model predictive control and direct power control.The power variables of the system are controlled variables,without the harmonic and reactive current detection process,reducing the complexity of the system.To validate the designed algorithm,in the environment of Matlab/Simulink a simulation is carried out.The simulation results show that after compensating the measured grid current distortion rate of 2.18%,proving the designed algorithm can effectively suppress harmonics and compensate reactive power,with a good dynamic characteristic.Finally,this thesis has set up a platform APF experiment in the laboratory environment,and uses C language programming.The experimental results show that the design of the FS-MPC-based algorithm can effectively eliminate APF control delay resulting by digital control.The proposed approach provides a straightforward and effective single loop solution,with excellent dynamic performance,not requiring grid synchronization or PWM schemes.