Model Predictive Control Methods For Single-Phase Grid-Tied Inverter

The efficiency and quality of the generation system are much dependent on the topologies and control strategies of the grid-tied inverter,which is the interface between the generation system and the grid.The key of the grid-tied inverter control technique is the tracking of the sinusoidal current.It requires that the grid-tied inverter tracks the amplitude,phase and frequency of the current command based on the power demand by the grid and the power output by the renewable energy.Besides,low total harmonic distortion(THD)of the current should be achieved and the pollution caused by harmonics to the grid can be reduced accordingly.Aiming to solve the above problems,three strategies applied to the single-phase grid-tied inverter are studied based on the model predictive control(MPC).Firstly,the finite control set model predictive control(FCS-MPC)of the single-phase grid-tied inverter is researched.During every sampling period,the output currents of the next period corresponding to all of the switching states are evaluated by the cost function for an optimal switching state to the system.This method can obtain the simple,effective and good current tracking performance as well as the quick dynamic response.Secondly,the ripple feed-forward control based on FCS-MPC is developed.This thesis proposes a ripple feed-forward controller based on the characteristics of the output current ripple in a single-phase grid-tied inverter under the FCS-MPC.The amplitude and RMS values of the output current ripple can be effectively reduced without the change of the switching frequency.Finally,in order to solve the defects of the high output current ripple and the variable switching frequency caused by FCS-MPC,a fixed switching-frequency continuous control set model predictive control(CCS-MPC)is proposed.With the relatively low sampling frequency,CCS-MPC achieved almost equal switching frequency as the FCS-MPC.Moreover,the smooth output current with low THD can also be obtained.With the MATLB/Simulink simulation and experimental platform,the methods developed in this paper are implemented and the proposed performance are all demonstrated with simulation and experimental results.