A Study On The Methodology Of Generation Forecasting For Photovoltaic Power Generation

With the increasing maturity of photovoltaic technology and the great support of the national policy, the cost of photovoltaic power generation is gradually lowered, the application of solar photovoltaic power generation has become a universal concern. The output power of the photovoltaic system is intermittent and high volatility, which makes grid photovoltaic power plant access to the entire security of the power grid, power supply and demand balance, and power quality can not be ignored.Accurate prediction of the output power of photovoltaic system, there are important significance for optimizing the scheduling and the safe and stable operation of power system.Select the method of prediction of photovoltaic power generation as research content, with a high practical significance and academic value.The amount of output power of the photovoltaic system has high degree of randomness and fluctuation amplitude due to the influence of various factors, such as weather conditions and geographical environment, which increases the difficulty of forecasting power of its output.In the current photovoltaic power prediction research work in the field of photovoltaic power, the existing prediction models using different data information from different angle to make a forecast for photovoltaic power, the prediction accuracy is to be improved.Characteristics of combination forecast methods of each single model can, effectively using various information, provides the possibility to improve the prediction accuracy of the existing model.Firstly, influence on the PV output power factors are discussed, and analyze the correlation between the output power of the PV system and its impact factors by path analysis method. On the basis of correlation analysis, comprehensive utilization factor of historical data, output power of the photovoltaic system of meteorological factors as the input data samples. Then research the single forecasting model and the combined forecasting method such as chaotic neural network, support vector machine, multiple linear regression model and the grey model etc. Based on the single model, the combination forecasting model was established using the natural selection particle swarm optimization algorithm to determine the weight coefficients, and predict the instance of a PV power plant using the model above. What’s more, the validity and the accuracy of the model were verified.