Research On Modeling And Parameter Identification For Grid-connected PV Generation System

As one of the important parts of the renewable energy source(RES),photovoltaic(PV)generation system has been receiving continuous attention in recent years and will continue its trend of rapid growth.Due to the characteristics of intermittent,random,fluctuating and control complexity for PV generation,large-scale PV systems access will bring about a series of adverse effects on the stability,reliability and economy for the power system.Modeling and parameter identification technology are the key technologies to obtain the operating characteristics of PV generation system and analyze the impact of large-scale PV system access on the power system.This article focuses on the modeling and parameter identification of PV generation system,and the main research content and innovation lies in:1.Taking a typical two-stage PV generation unit as the research object,a modular model of the PV generation system is established from complex to simple.And the adaptability of different models for inverter was analyzed,laying the foundation for model parameter design,parameter identification and model verification in later chapters.2.Aiming at the design problem of PV inverter control parameter,this paper presents a method of PI parameter design for universal controller based on analytical computation.First of all,the analytical expressions of PI parameters for the inner and outer loop of the controller are derived according to the transfer function.The conventional control PI parameters are determined according to the open-loop cut-off frequency and the phase margin.Secondly,according to per unitary mathematical model,the PI parameters of the controller after per-unitization are deduced by comparing the structure of the controller before and after per-unitization.Finally,based on the similar characteristics of the PI parameter expressions of controllers in different systems,the PI parameters of the unified controller are determined.The simulation results show that the PI parameters designed based on proposed method guarantee the stability of the DC bus voltage,and the THD of the rated grid-connected current is about 3%,which meets the system design requirements.3.In order to identify the controller PI parameters of PV inverters,the method of frequency domain identification based on Levenberg-Marquardt(LM)is adopted in this paper.Firstly,the identifiability and sensitivity of the controller parameters are analyzed based on PV inverter decoupling identification model.Secondly,the frequency response of the system is obtained by applying sinusoidal disturbance signals of different frequencies to the DC voltage reference value and LM is used for parameter fitting.Finally,the implementation of proposed method is elaborated in the MATLAB/Simulink.The simulation results show that there is a high degree of parameter fitting in outer loop due to its high sensitivity.Similarly,there is a certain degree of error for inner loop parameter identification due to its lower sensitivity,but the dynamic external characteristics of the identification model and the simulation model are basically the same.4.According to the detailed model and identified parameters of PV generation system,the simulations are performed with two scenarios of insolation change and voltage drop to observe the dynamic grid-connection characteristics of PV generation system.Secondly,due to the time-consuming problem in the long-term operation of the system,a super real-time simulation(SRTS)method for the power electronic system(PES)is proposed.Taking the phasor model of PV generation system as an example,the effectiveness of super real-time operating characteristics was verified through simulations at different time scales.