Model Predictive Control Of T-type Inverters

The grid-connected inverter,as an extremely important part of new energy power generation,has a great influence on the performance of the energy generation system.This paper takes T-type grid-connected inverter as the research object,and mainly studies the application of model predictive control(MPC)in three-phase T-type grid-connected inverter and single-phase T-type grid-connected inverter.First,this paper analyzes the topology and working principle of the T-type three-level bridge arm,and on this basis,the mathematical models of the three-phase T-type grid-connected inverter and single-phase T-type grid-connected inverter are derived A three-level SVPWM algorithm is proposed,and a three-level SVPWM over-modulation processing strategy is proposed.Then this paper derives the current prediction equation and power prediction equation of the three-phase T-type inverter.According to these equations,finite control set model predictive current control,finite control set model predictive power control,model predictive current control based on three-level SVPWM modulation and model predictive power control based on three-level SVPWM modulation are discussed in detail.For finite control set model predictive control,this paper proposes a method of filtering out redundant vectors in advance,which further reduces the computational complexity of the finite control set model predictive control.For model predictive control based on three-level SVPWM modulation,this paper proposes a DC capacitor voltage difference feedback control method to balance the neutral point voltage,which can dynamically adjust the short vector effective time in the SVPWM algorithm according to the imbalance degree.For single-phase T-type grid-connected inverter,this paper derives the current prediction equation and power prediction equation.This paper proposes two vector synthesis model predictive power control and three vector synthesis model predictive power control.Different from the traditional limited control set model predictive power control,these two control methods select multiple switch states in a sampling period to synthesize the optimal control vector.Compared with the traditional finite control set model predictive control,the proposed control methods have more concentrated switching frequency and better suppression of the fluctuation of the center point.In addition,this paper proposes a model parameter adjustment method,which effectively reduces the impact of model parameter changes on model predictive control.Then in SIMULINK,the control method proposed in this paper is simulated.Finally,a T-type grid-connected inverter experiment platform was built.The hardware design and software design of the experimental platform are introduced,and the effectiveness of the analysis and simulation in this paper is verified by experiments.