Application Of Intelligent Optimization Algorithm In Photovoltaic Parameter Estimation

The energy crisis and environmental pollution have prompted us to seek new energy alternatives.Solar energy is abundant,widely distributed,clean and pollution-free,and is one of the most promising renewable energy sources.Photovoltaic power generation system is a new power generation system that converts solar energy into electrical energy.As an important part of photovoltaic power generation system,photovoltaic cells can accurately identify its parameters,provide guarantee for the technical design of photovoltaic cell fault diagnosis and maximum power point tracking.The evaluation and control of photovoltaic power generation systems are of great significance.Intelligent optimization algorithm is a new type of photovoltaic cell parameter identification method.Compared with traditional identification methods,it has the advantages of simple operation,less restrictive conditions,strong robustness and suitable for solving various complex environments.It has gradually been widely used in photovoltaic cell and photovoltaic module parameters estimation.However,traditional intelligent optimization algorithms also have certain shortcomings.The convergence speed and accuracy cannot meet the ideal requirements when identifying photovoltaic cell parameters.Thus,this research proposes various enhanced intelligent optimization algorithms and estimates parameter of photovoltaic cells and modules.The research content of the thesis mainly includes the following three parts:Aiming at the premature and slow convergence of the original atom search algorithm,a new arcsine-cosine mechanism is designed to improve the convergence speed and accuracy of the atom search algorithm.The proposed arcsine-cosine atom search algorithm is to identify parameters of two monocrystalline silicon and polycrystalline silicon photovoltaic cell models under fixed conditions,and compared with other methods.Results show the arcsine-cosine atom search algorithm can more accurately identify the parameters of single crystal and the polycrystalline silicon photovoltaic module under fixed environmental conditions.Aiming at the shortcomings of the gradient-based optimizer algorithm,such as being fall into the local optimum,a novel random learning mechanism is designed to boost the exploitation ability.The proposed random learning gradient optimizer algorithm is to identify unknown parameters of three monocrystalline silicon photovoltaic cell models and polycrystalline silicon photovoltaic modules under fixed environmental conditions,and compare experiments with other algorithms.The results show the random learning gradient optimizer can more accurately solve the parameter identification problem of different photovoltaic cells and modules under fixed environmental conditions.In view of the slow convergence speed and low accuracy of the backtracking search algorithm,a sine-cosine mechanism backtracking search algorithm is proposed to identify the parameters of two thin-film and single-crystal photovoltaic modules at different illumination and temperature conditions,and the experiments are compared with other algorithms.The results show that the sine and cosine backtracking search algorithm can obtain the best results more accurately and reliably in environments such as excessively high temperature and insufficient light.