Study On Short Term Forecasting Method Of Photovoltaic Power Generation Based On Combination Model

Because of the fluctuation, intermittence and instability of the output of photovoltaic power station, it is bound to have an adverse impact on the safety and stability of the power grid. Therefore, research on photovoltaic short-term power forecast technology is particularly important, short-term power prediction accuracy can not only effectively reduce the harm to the power grid, but also can help carry out real-time scheduling and guide conventional energy generation planning to mitigate the adverse effects of photovoltaic power generation system, to ensure the power grid safe and stable operation. Hence, it is of great significance to select the short-term power forecasting method of photovoltaic power generation as the research content.This paper is to study the short-term photovoltaic power forecasting technoloogy. The paper briefly introduces the background of photovoltaic power generation and the domestic and foreign research status of the photovoltaic power forecasting technology. Because of the existing domestic photovoltaic power prediction model are the basic principle from different perspectives, different extraction information of sample data using the model characteristics,thus to make predictions of photovoltaic power station power, but the single prediction model of photovoltaic power under the influence of many factors is almost impossible to obtain satisfactory accuracy. Based on the example of a photovoltaic system, a detailed analysis of the relationship between the photovoltaic power and weather type, radiation, temperature and humidity etc. is presented. Through the correlation analysis and comparison, the paper puts forward the classification of weather types. The photovoltaic generation is random and fluctuated due to the influence of various factors, such as weather conditions and geographical environment, which increases the difficulty of forecasting its power. The existing prediction models for power prediction generally take single or few factors, so its accuracy is to be improved. Characteristics of combination forecast methods of each single model can,effectively using various information, provide the possibility to improve the prediction accuracy of the existing models. The paper details the grey GM forecasting method and BP artificial neural network forecasting method, multiple linear regression model and combination forecasting method based on using the entropy method. Combined with the actual photovoltaic power generation system, 3 typical weather types are put forward for example analysis. By comparing the forecasting results and evaluated by MAPE, the paper proves that the combination model is practical and effective in the short-term photovoltaic power prediction.