Research On Parameter Identification Method Applied To Photovoltaic System

With the widespread deployment and use of photovoltaic power in China’s power supply,the scale of China’s photovoltaic industry is increasing year by year,and the larger and larger power generation scale has gradually strict requirements on the performance indicators of photovoltaic system.Accurately establish the photovoltaic power generation system model,obtain the parameters under the model to achieve the fitting of photovoltaic data,in order to achieve the photovoltaic power station performance monitoring,power calculation,maximum power point tracking and power output and other functions.However,the photovoltaic system has many monitoring points,large and complex data,and the existing parameter identification methods are unable to meet the needs,so the parameter identification field has become the focus of research by scholars at home and abroad.At present,the research direction of scientific research institutions and universities is focused on the method of parameter identification of photovoltaic model.This dissertation is devoted to optimizing the parameter identification optimization algorithm and exploring the method to improve the accuracy of the parameter identification model.The main research contents are as follows:(1)The FOA algorithm applied light v component parameter identification,and put forward tabu search algorithm and fruit flies(TS-FOA)parameter identification method to optimize the ability of FOA algorithm,the algorithm is based on the complex relationship between fruit flies foraging and derived a iteration algorithm of TS-FOA under the global scope,the proposed algorithm not only ensure the previous global search ability,At the same time,TS concept is introduced to optimize the traditional FOA algorithm to further reduce the later iteration time,which can avoid the problem of falling into the local optimal solution in the later iteration and improve the optimization efficiency.The results show that the predicted value curve of TS-FOA algorithm is more consistent with the actual power curve,and the prediction accuracy is higher than the traditional algorithm.(2)In order to further solve the problem of parameter identification of large-scale data sets under the double-diode model,a fractional-order Darwinian particle optimization algorithm(RFOD-PSO)based on the reward and punishment mechanism is proposed.Upgrade and adjustment,an improved method using fractional calculus is designed to optimize the PSO algorithm,which avoids the problem that the PSO algorithm falls into local optimum earlier when the amount of data is large or the number of iterations is large.The reward and punishment mechanism(RP)further improves the search process of the particle swarm.Finally,the empirical data is used to test,and the results show that the RFOD-PSO algorithm has better parameter identification results under the conditions of multiple iterations and double-diode models.(3)According to the requirements of the demonstration photovoltaic power station system,based on the technical framework of Idea + Tomcat + Springboot + Vue,a demonstration power station monitoring and management system is designed,and modules such as login,monitoring,retrieval,error warning,and parameter identification of the photovoltaic power station system are developed,and the realization of A photovoltaic power station system that can easily realize parameter identification information analysis,and improves the management ability of power information.