Power Forecasting Model For Photovoltaic Plants Generation Based On The Output Of WRF Mode

Photovoltaic power generation,as an important form of renewable energy development,has the advantages of being clean,pollution-free,and sustainable,and has a broad prospect for development.Due to the influence of meteorological factors and the system level,photovoltaic power is characterized by fluctuation,intermittence and nonlinearity.Large-scale photovoltaic grid-connected can easily impact the safe operation of power grid.The existing Photovoltaic power prediction research mainly relies on the historical information,which is not conducive to the prediction of photovoltaic power generation in complicated weather.In view of it,based on the short-term forecast output from WRF,this paper proposes a photovoltaic power generation forecasting method.The main work is as follows:Firstly,the mathematical model of solar cells was established according to the basic principles of photovoltaic power generation.Using Simulink simulation to establish a solar energy project model considering the dust deposition of the panel,the output characteristics of solar cells were obtained and discussed,and the influence of meteorological factors,such as actual irradiance considering dust deposition,temperature,humidity,and weather type on photovoltaic output was analyzed.Secondly,we use the WRF model to predict the meteorological factors of photovoltaic power generation,and analysis the influence of the setting parameters on the WRF prediction results,and carries out some sensitive experiments of forecasting irradiance according to different model resolution,microphysical processes and cumulus parameterization schemes.It is found that increasing the mode resolution is beneficial to reduce the forecast error.The choice of physical process will also affect the forecast results.Based on the optimized WRF scheme,we forecast the weather from June to July in Zhongwei City of Ningxia.The forecast results show that the WRF model has different forecasting ability for different type's weathers,and has poor prediction ability on rainy and cloudy weather.Considering these conditions,the MOS correction equations for the solar radiance were established according to the type of weather,which strongly reduced the prediction error.Finally,based on WRF output results,a new prediction model for the photovoltaic power generation is founded,which combined the WRF prediction results and the GA-BP neural network.The BP neural network algorithm were introduced and analyzed in detail,whose weights and thresholds were optimized by genetic algorithms(GA),and the BP neural network modified by genetic algorithm was used to predict the photovoltaic power generation in different weather conditions.After compared with the actual results,this method is proved to be reliable.