Experimental Study On The Power Prediction Of Distributed Photovoltaic Power Generation

The quantitative relationship between photovoltaic(PV)power output and meteorological factors is obtained by analyzing the meteorological information and PV power output data collected from a set of amorphous silicon photovoltaic power generation platform and meteorological information monitoring station.On the basis of this,a method for predicting meteorological information by time series method is proposed,and PV output power is predicted by the support vector machine method.Using MATLAB GUI,a set of PV output power prediction software is implemented.The main tasks of this paper are as follows:(1)A set of amorphous silicon PV power generation system and supporting meteorological information monitoring stations are designed.The correlation between data is obtained by analyzing the data of meteorological and PV power generation.(2)A PV power prediction method based on time series and support vector machine is proposed,and the prediction process is optimized by a similar daily method.The method is tested by typical day data and the prediction error is analyzed.(3)Design and implement a set of PV power prediction software based on MATLAB GUI,and it is used in the actual work of PV power prediction.