Large-scale Wind and Solar Grid-connected Power Prediction Based on Causal Delay Spectrum and LSTM

Development of renewable energy puts forward higher requirements for effective power prediction, which is critical to economic dispatch and operation of power system. In order to improve the prediction precision, considering characteristics of large-scale hybrid wind and solar grid-connected power, a method (CDS-LSTM) based on causal delay spectrum and long short term memory is proposed. Firstly, build the multilayer perceptron based on CDS. The number of weather factor and maximum delay are determined. Then, build the LSTM model of renewable energy, and the delay order of power output factors is determined by autocorrelation function. Finally, get the prediction of renewable power by merging the output of CDS model and LSTM model. Experimental results based on the data from an actual grid in one city verifies effectiveness of the method.