The Application Of The MPPT By SVR Model Based On The Improved PSO Algorithm

In today’s rapid development of society, the human demand for energy is more and more big, the solar energy as the most ideal alternative energy in new energy has been widely recognized, photovoltaic power generation technology to the attention of more and more people. Photovoltaic maximum power point tracking technology is a key method to improve the efficiency of PV system. How to effectively control the working mechanism of maximum power point of photovoltaic cells, and improve the efficiency of photovoltaic array component of photoelectric conversion, has become a research hotspot in the field of solar energy research. In this paper, after studying the photovoltaic cells on the basis of the principle of maximum power point tracking technology, combined with SVM and PSO algorithm, based on improved PSO algorithm is mainly studied the SVM regression model, and applies to the MPPT prediction problem, make the photovoltaic modules work in maximum power point, to achieve maximum output power.First, the paper describes in detail the working principle and mathematical model of solar photovoltaic cells, this paper discusses the principle and the existing photovoltaic maximum power point tracking control method. And the existing research results are analyzed and evaluated, points out the problems existing in the theoretical research and practical application.Secondly, this paper introduces the SVM regression algorithm and PSO algorithm, particle swarm optimization algorithm is mainly studied the problem of parameter Settings of the inertia weight of PSO algorithm and learning factor are studied respectively, puts forward the new method of particle swarm optimization algorithm parameter Settings.Then, combined with SVM regression algorithm, improved particle swarm optimization algorithm was applied to parameter optimization of support vector machine, according to the sample data to find a better punishment parameters C and kernel function g, in order to improve the generalization performance of SVM regression model. Therefore put forward the SVR model based on improved PSO, and applies the model of maximum power point of light volt battery track prediction, and compared with the SVR model based on basic PSO algorithm are compared, and the experimental results show that the model can accurately track the battery maximum power point, good prediction effect.Finally, this algorithm proposed in this paper puts forward the problems to be solved are many, in order to better implementation in the subsequent research photovoltaic maximum power point tracking control, and to expand the train of thought for the further study of the SVM method.