Research Of MPPT Using Artificial Neural Network And Support Vector Machine For PV Module

As the result of lacking traditional resource (such as coal, oil, natural gas, etc.), peoplebegin to pay their attention to develop renewable energy sources. Solar energy, as a new typeof green renewable energy sources, compared with other new type of energy sources, it is veryhuge and widely distributed, is an ideal renewable energy sources. Especially, in recentdecades, along with the development of the science and technology, PV(photovoltaic) systemand related industries has become one of the fastest-growing industries. Therefore theresearches of PV system become more and more important. At present, because of the lowconversion efficiency, high price of PV cells and long cycle cost recovery, which is seriouslyhampering its promotion and application. There are two methods to maximize using the powergenerated by PV cells: reducing the circuit loss and making PV cells output maximum power.The method of maximum power point tracking makes the PV cells working at the maximumpower point to obtain the maximum output power, which will be research in emphasis in thisarticle.1) The operation principle and process of PV cell are introduced and analyzed in details.The structures and principle of maximum power point tracking (MPPT) aresystematically discussed. Some difficult problems in the applications and theories ofMPPT are pointed out.2) Based on the theory of data collecting system, data collecting in irradiance andtemperature are designed and realized. After analyzing factor affecting the change ofthe position of MPP, the theory and structure of Artificial Neural Network(ANN) arepresented. A new method is introduced which uses BP network to design this network,whose training algorithm is improved to get faster convergence and more accuratedirection, and to predict MPP. Compared with the method of Perturb and observeused in MPPT, the result shows better precision and efficiency.3) As a learning method based on SLT, Support Vector Machine(SVM) has theadvantages of global solutions, good adaptability and high generalization ability in theory. In this paper, SVM is used here in approaching mathematical model of PVmodule. The result of simulation indicate good performance. SVM is also used in theresearch of PV MPPT and the result is satisfying.