Research On Predictive Control Of Single Phase LCL-type Grid-tied Inverter In Weak Grid

With the rapidly increased new energy connected to the grid,LCL-type grid-tied inverters are more and more applied as the interface of new energy and power grid.For the grid impedance can not be ignored,the robustness of the current control of LCL-type grid-tied inverters in weak grid has received attention.Because of its strong robustness and fast dynamic response,predictive control has been applied in power electronics technology with the development of digital control technology.In this paper,predictive control method is used to solve the problem of a single phase LCL-type grid-tied inverter,in order to achieve a simple and feasible method,taking account of robustness and dynamic respons.As a typical kind of predictive control method,deadbeat control is widely used in the control of grid connected inverter.Based on deadbeat control theory,this paper proposes a simple and improved deadbeat control method,and digital control delay is considered.Through a simple difference equation,the control value of next time can be predicted to compensate for the delay.In this paper,a detailed theoretical analysis of the proposed method is presented to explain the characteristics of adapting weak grid.And the tracking ability of the reference and the ability to restrain the grid disturbance are analyzed in the frequency domain.Finite state control set(FCS)control method has the advantage of simple thought and good robustness,but the switching frequency is not fixed;Continuous control set(CCS)control method has the advantage of fixed switching frequency,but the robustness is relatively poor.A predictive method based on the modulation wave optimization is proposed in view of the advantages and disadvantages of FCS method and CCS method.This method has the characteristics of strong robustness and fixed switching frequency.Through the analysis the modulation optimization method,the selection of predictive model,the selection of cost function,the selection of search range and step length,time delay compensation are analyzed in detail.Finally,the robustness of the modulation optimization method is analyzed,and it is pointed out that the adaptability of the weak grid and the good robustness to the LCL parameter.The State-Average-Model is used as the predictive control model,so it is difficult to predict the disturbance of the nonlinear disturbances,such as deadtime,which will have a negative impact on the actual output waveform quality.Against the shortcomings of the control method,including the influence of dead time,the steady state error and DC voltage ripple effects respectively,a detailed analysis of the factors affecting output performance are analysed and their solutions are given.Add dead time compensation in modulation wave to eliminate the effect of dead time.Use quasi resonant regulator to eliminate the steady state error.Adjust modulation wave according to the real-time sampling value of DC voltage to eliminate the influence of DC ripple on steady state.Finally,the purpose of improving the quality of output waveform for predictive control is achieved.Based on the Matlab/Simulink simulation and experimental platform,the methods proposed in this paper including improved deadbeat method and modulation optimization method are validated,simulation and experimental results show that the proposed methods are reasonable and effective.