Distributed Photovoltaic System Short-term Forecasting Based On Optimized Neutral Network

Nowadays, haze pollution becomes more and more serious in many regions in China. And environmental issues are paid more and more attention. Our country is still taking coal as the primary energy but now is vigorously developing new energy technologies and applications. Distributed photovoltaic power generation uses solar energy. It becomes one of the most extensive development potential of new energy in twenty-first-century as it is clean and no pollution. In this context, National Grid of Liaoning Province is energetically developing solar power generation including distributed photovoltaic systems. In "13th Five-Year" power planning, it announced that province’s power installed capacity reached 43 million 710 thousand kilowatts by the end of 2015. The solar installed capacity of 320 thousand kilowatts. And to the end of the "13th Five-Year", Installed capacity comes to 56 million 800 thousand kilowatts, with 1 million 820 thousand kilowatts of Solar installed capacity.The power output of distributed photovoltaic system is affected by the environmental factors. And it can only work during the day under light conditions. So it has the disadvantages of random, intermittent and nonlinear. When access to the power grid it will bring greater difficulties to the power grid dispatching and safe operation. Mastering the fluctuation characteristics of the output of distributed PV system and realizing system output power prediction will be very important for the application and development of distributed photovoltaic system.In this paper, as there is always lack of research on the input information in most of the research on the prediction of the output power of photovoltaic power generation. This study aimed at the distributed photovoltaic system in Liaoning area and taking "Shenyang Agricultural University photovoltaic demonstration power station" as a case. In view of the characteristics of the distributed PV system in Liaoning area, the fixed PV modules are adopted. Firstly, based on the model of solar irradiance, the radiation illumination on the plane is converted to the radiation illumination which is accepted by the component plane with a certain angle and direction and a more accurate model of distributed PV system is established. Then the influence of different weather types, temperature, wind speed and other environmental factors on the PV output was quantified by the introduction of the PV output shielding factor ηt.The wave characteristics of distributed PV system under different environmental factors are obtained. And the input information of the forecast model are determined. In the end, a distributed photovoltaic system based on BP neural network is established. Research realized to predict the output power of distributed PV system more accurately under different weather types. And the accuracy of the forecast of the output power of the photovoltaic system is greatly improved. The design of the study can be used as a predictive model to effectively predict the power output of the distributed PV system model in Liaoning. And provide support for the electric power management department to make a more reasonable scheduling scheme for the grid connected PV system.