Research On Power Optimization And Forecast Methods Of PV Arrays In Distributed Photovoltaic System

As the world energy exhausting and environmental damage gradually serious, green and renewable energy were paid more and more attention. Making full use of the energy to relieve the crisis and environmental destruction play an important role.Photovoltaic energy power is a kind of clean, green, renewable energy which is recognized as the the ideal energy to realize sustainable development. Distributed photovoltaic power system as a supplement architecture of centralized power supply, has many advantages so that has been widely applied in practice. However, due to the objective cause, distributed photovoltaic power system is not effeicient which causes the output power is usually lower than the actual maximum. Therefore, optimizing the distributed photovoltaic system power output has important significance to improve system efficiency and accelerate the development of photovoltaic industry.Distributed photovoltaic system usually was blocked and partial shaded which leads to the power-voltage characteristic has multiple output-peaks,so that conventional maximum power tracking (MPPT) algorithm was failure, a kind of global search ability algorithm was needed. In addition, when photovoltaic system working at morning, dusk, rainy and cloudy weather such as low light intensity conditions,the arrays output power is low and not matched nominal power of inverter which causes lower efficiency and output poor quality power that would influents power grid and load. In view of the two problems mentioned, a simulated annealing algorithm which has global convergence ability was presented in partial shaded photovoltaic system MPPT firstly, through simulation and experimental verification, the results shows that this algorithm can realize multi-peaks curve of MPPT and improve the output power influence of partial shaded photovoltaic arrays; Then,a kind of reconstructable switch topology used for photovoltaic array and inverter connecting was designed from the vision of distributed photovoltaic system topology structure, hierarchical clustering theory was used to control the reconstructable topology, joining with BP-ANN and ARMA algorithm of photovoltaic system short-term forecasting means as reconstruct operation condition. Distributed photovoltaic system platform and reconstructing topology were established in Matlab, topology reconstruction control strategy was programmed,and the photovoltaic system short-term prediction experiment platform was developed and simulated,the experimental results show that the reconstructable topology can improve the connection between the inverter and array, rises up distributed photovoltaic system output power on partial shaded and low light conditions, and the short-term power prediction algorithm has higher accuracy.The simulated annealing MPPT algorithm can solve the photovoltaic system partial shaded output problem, the reconstructable topology promoted distributed photovoltaic system output power through solving power matching problem of array and inverter on partial shadow and low light conditions. Of course, there are many aspects to optimise photovoltaic power system output power, such as Condensing and Sun Tracking system, photovoltaic battery and inverter, the equipment technology improvement also plays an important role to output optimization.