Research On Short-Term Power Forecasting Method Of Wind Power Generation Based On Dynamic Adaptive Technology

Wind power prediction can reduce the adverse effects of wind power instability on the grid,and is help for the system dispatcher to adjust the schedule,to improve the power grid peaking capacity.Simultaneously,it can also enhance the competitiveness of wind power in the electricity market.The dynamic adaptive technology of wind power generation is proposed in this paper,and studied the following aspects:First,the technical classification of wind power was introduced at this stage,the data sources,information flow and forecast process of power forecast for wind power generation were analyzed.On this basis,the overall framework of wind power forecasting was discussed in this paper,six physical hierarchies of wind power clustering power with general significance were proposed,and the method of improving the short-term power prediction accuracy of the cluster was discussed from the various physical levels,which are data layer,mapping layer,feature layer,model layer,feedback layer and output layer.Second,A short-term power forecasting method for wind power cluster based on dynamic adaptive technology and deep learning were studied in the paper.DBN network structure was construct and short-term power forecasting model for wind power cluster was established based on the deep learning theory.Wind power data is pre-processed,The influencing factors of the deep learning prediction model based on adaptive technology are studied by adjusting the depth of the network structure of the network level,number of nodes per layer,sample set and test set number.Finally,Short-term power prediction model of wind power cluster based on dynamic adaptive technology and statistical scale-up is established in the paper.The dynamic division of wind power cluster,the dynamic selection of reference wind farm,the adaptive selection of samples,the adaptive selection of prediction model were studied.The dynamic clustering of the cluster by means of weather classification is beneficial to the sub-cluster wind farm to select the appropriate forecasting method according to the similar weather type in the paper,The prediction accuracy of the regional short-term power is improved,The dynamic selection of the benchmark wind farm can improve the situation which it is difficult to find a representative wind farm in the cluster.the reference wind farm is selected based on the correlation between the output of a single wind farm and the output of the cluster.The adaptive model of the baseline wind farm power prediction was built based on the weather classification method,which improved the prediction accuracy of the reference wind farm.The clustering output predicted by the dynamic selection model of the statistic scale method can predict the short-term trend of wind power cluster.