Study On Photovoltaic Array MPPT Control Based On BP Neural Network

At present, due to the rapid development of social economy, the energy shortage and environmental pollution problems have become more and more serious. Solar energy as a renewable clean energy has become the most development potential of new energy, and photovoltaic power generation technology as one of the most effective way to use the solar energy has been widely concerned by the community. Because of the shielding in photovoltaic array surrounding public facilities, buildings and mobile cloud, light intensity and ambient temperature change rapidly, the output characteristic curve of photovoltaic cells will appear the characteristic of multi peaks, it leads to the method of traditional maximum power tracking can not accurately track the control for photovoltaic cells and greatly reducing the photoelectric conversion rate, resulting in a waste of energy and the life of photovoltaic cells have been affected, meanwhile, it results in a waste of energy and the life of photovoltaic cells have been affected. Therefore, the maximum power tracking control of photovoltaic cells is very important to improve the overall efficiency of Photovoltaic Power Generation(PPG) under complex illumination conditions.In this paper, through the analysis of the working principle of photovoltaic cells, we built model of photovoltaic cells in MATLAB, the output characteristic curve of PV array are studied in different obscuration. On the principle basis of unipolar PPG system, the PV model of unipolar photovoltaic grid connected PPG is built.The traditional methods of maximum power tracking are studied and simulated, and the traditional method is easy to fall into the local maximum power point and the direction of the maximum power point. The new intelligent algorithms are introduced. The basic BP neural network algorithm can be able to accurately amplify the maximum power tracking control when fixed obstacles appear in the photovoltaic array. On the basis of this, an improved BP neural network algorithm is proposed. A neural network model is built on the MATLAB, which is applied in the single stage grid connected PPG system. This paper simulated when the mobile cloud through the photovoltaic array and tested the improved BP neural network algorithm for photovoltaic array for maximum power tracking control. In order to develop further verify the effectiveness of the improved BP neural network method, the hardware of semi-physical in the loop simulation is carried out by RT_LAB. In RT_LAB to build a unipolar photovoltaic model, the improved BP neural network algorithm and the production of DSP wave PWM program are completed. The improved BP neural network algorithm can complete accurately maximum power tracking control in the photovoltaic array appear mobile or fixed obstructions of photovoltaic array. Comparative analysis result shown that,the dynamic response time of proposed MPPT method is shorter than the basic neural network, and the tracking accuracy is improved effectively.