Research On Model Order Reduction Of Photovoltaic Grid-connected System

A large number of new energy sources are connected to the grid with power electronic interfaces,which makes the power system show a high power electronic development trend,which leads to a significant increase in the order of the non-linear high-order model used to describe the dynamic behavior of the power system,and an increase in the time scale span,and cause a series of problems such as difficulty in solving.At the same time,the simulation efficiency of large-scale new energy power generation systems is also low,and model order reduction is one of the effective methods to solve the above problems.Therefore,this paper studies the model order reduction and simplification of photovoltaic grid-connected systems in detail.The main work is as follows:First of all,the basic model of photovoltaic power generation system is briefly introduced.It is mainly analyzed from two aspects of physical components and control system.Take PV grid-connected system using PQ control strategy as an example.Based on the Matlab/Simulink simulation platform to build a model,the dynamic characteristics of the grid-connected system under different fault conditions are briefly analyzed.Secondly,the model reduction of photovoltaic grid-connected systems is performed.Based on the singular perturbation theory,appropriate perturbation parameters are selected for each state variable of the system,and the singular perturbation model is established to perform order reduction ignoring the fast dynamic form.Based on the eigenvalue sensitivity theory,state variables with strong correlation with each eigenvalue are calculated.Different dominant eigenvalues are selected to obtain different dominant state variables.Ignore the remaining state variables and obtain different reduced-order models.Thirdly,the time-domain simulation method is used by Matlab to verify and analyze each reduced-order model under the conditions of voltage disturbance,controller command value change and system frequency fluctuation.The dynamic characteristics of the reduced-order models are analyzed in detail,the differences and physical meaning between the reduced-order models and the detailed electromagnetic transient models are compared.Based on the simulation results,the model order reduction error and simulation efficiency between each order reduction model and the original model are qualitatively analyzed,which is in line with the actual model order reduction effect.Finally,the simulation efficiency of the reduced-order models is analyzed in detail.By setting different simulation steps,the reduced-order efficiency of each reduced-order model is calculated.The two-norm method and the Pearson correlation coefficient method are used to calculate and analyze the reduced order errors between reduced-order models and detailed electromagnetic transient models.At the same time,the influencing factors of the reduction error were studied.By changing the simulation step size and the controller parameters,the Pearson correlation coefficient method was used to calculate the reduction error.The correlation conclusion between order reduction error and the simulation step size and the controller parameters was obtained.The two reduced-order models of photovoltaic grid-connected systems proposed in this paper meet the balance of simulation accuracy and simulation efficiency to a certain extent.They can be applied to different applications,of which the ninth-order model is suitable for small-scale grid-connected systems with priority on accuracy,and the fifth-order model can be applied to large-scale grid-connected systems with high simulation efficiency requirements.