An optimization method based on LM-GA for parameter identification of photovoltaic cell

Abstract

Smart sensors are the core of condition monitoring of power equipment. However, energy supply for sensors in complex electric power field is difficult and in the spotlight. Solar energy as an easily available energy is a good solution to this problem. Accurate identification of photovoltaic cell model parameters can ensure the subsequent stable energy supply. The purpose of this paper is to realize the accurate identification of photovoltaic cell model parameters, so as to serve the energy supply of monitor devices. Firstly, the basic circuit of photovoltaic panel energy supply is built. Secondly, the U-I characteristic curve of photovoltaic cell under different loads is measured. Thirdly, the equivalent parameter model of photovoltaic cell is constructed. Finally, the model parameters are accurately identified based on Levenberg Marquarelt (LM)-Genetic Algorithm (GA). LM algorithm is used for fast global calculation to determine the approximate range of global optimal solution, and then GA is used to further iterate within this range to obtain high-precision local extremum. The proposed method achieved a better fitting effect and results has higher precision in parameter identification of photovoltaic cell. Moreover, under the rated working condition, the errors between the identified parameters and the numerical values provided by the manufacturer are within ±5%. This paper presents an optimization algorithm combining LM and GA, which can accurately identify the parameters of photovoltaic model under different solar radiation levels, and provide strong technical support for the energy supply of sensors and other monitoring equipment.