Research On Forecast Of Grid-connected Photovoltaic Generation Based On D-S Evidence Theory

Photovoltaic power generation will gradually replace fossil fuel energy due to its own advantages.Affected by various factors such as the external climate environment and the equipment itself,the power generation has periodicity and uncertainty.The accurate prediction of power generated by photovoltaic grid-connected systems will help improve the stability and reliability of grid operation.Based on the historical data collected by the 5.6kWp photovoltaic grid-connected power generation system of Beijing University of Civil Engineering and Architecture,a D-S evidence theory photovoltaic grid-connected power generation prediction model based on solar radiation,ambient temperature,and temperature of photovoltaic modules was established.Using the decision tree's characteristic of large and simple data containment,the historical data is divided,and the data samples obtained from the decision of different attribute conditions are used as the evidence of the DS evidence theory to perform data fusion,and the prediction is obtained after multiple evidence fusions.power generation.Based on the analysis of the effects of various weather types in the four seasons on the prediction of photovoltaic power generation,the use of Python programming tools to achieve ultra-short-term prediction of photovoltaic power generation.Comparing the predicted and measured values,the forecast results show that the maximum root mean square error of all forecasted time periods is within 5%,less than the 15% error indicator specified in the technical requirements,and the qualified rate reaches 100%.The feasibility and accuracy of the D-S evidence theory as a predictive model are proved.