Research On Dynamic Modeling Of Photovoltaic Power System

Achieving a larger share of renewable energy in the consumption of the power industry is the focus of China's energy transformation.With the rapid development of photovoltaic power generation technology,the total installed capacity of photovoltaic power generation is increasing in the power grid.Since the photovoltaic system needs to be connected to the power grid through power electronic equipment such as inverters,the characteristics of voltage fluctuation and flicker are both volatility and intermittent,which determines that the difference between the principle and the conventional generator set is very large.In order to solve this problem,it is very important to establish a simulation model that is simple,fast and accurate in response to the dynamic characteristics of the photovoltaic power generation system.In this paper,the mainstream two-stage three-phase grid-connected photovoltaic system is the research object,with the establishment of dynamic phasor model,and improved MPPT control strategy.At the same time,the neural network is used to predict the photovoltaic power output.Firstly,this paper introduces the definition of dynamic phasor method and the computational characteristics of dynamic phasors,and adopts dynamic phasor method to mathematically derive the photovoltaic power generation system including the control system,to establish the dynamic phasor model of each link of photovoltaic system.By comparing the results of the dynamic phasor model,this paper verifies the detailed time domain model,the traditional phasor model,and the correctness of the dynamic phasor model,and demonstrates the superiority of the dynamic phasor model.Secondly,this paper expounds the misjudgment and oscillation problems of the perturbation observation method in MPPT.The perturbation observation method based on power prediction and the perturbation observation method based on dichotomy are proposed to improve the misjudgment and oscillation problems respectively.This paper establishes the simulation models of two disturbance observation methods by using MATLAB to verify the effectiveness of the control strategy.Finally,combining the two methods,this paper proposes a variable step size perturbation observation method based on power prediction.Finally,this paper establishes the dynamic phasor model of the photovoltaic system that includes the controller,and verifies the correctness of the dynamic phasor model.At the same time,an LSTM memory network suitable for predicting solar radiation is established to predict the power of the photovoltaic system.Combined with the dynamic phasor model,the simulation of photovoltaic models in long time zones becomes a reality.