To improve the accuracy of photovoltaic power generation forecasting,a photovoltaic power prediction model(VMD-CIWOA-BPNN)is proposed based on Variational Mode Decomposition(VMD)and an improved Whale Optimization Algorithm(CIWOA)coupled with a Back Propagation Neural Network(BPNN).First,data preprocessing and correlation analysis are performed,and the mutual information method is used to select environmental factors highly correlated with photovoltaic output power as the input variables of the model.VMD can be used to decompose complex power data into multiple intrinsic mode functions,reduce the impact of noise and make the prediction model more robust and accurate.Second,CIWOA initializes the whale positions using a Cubic map chaotic mapping.The whale positions are updated by an adaptive weight strategy to optimize the initial weights and thresholds of BPNN,and enhance the convergence speed and accuracy of BPNN.The prediction effect of the model is evaluated through the result evaluation indicators.The experimental results show that the VMD-CIWOA-BPNN model is superior to the BPNN model and the CIWOA-BPNN model in terms of the three evaluation indicators of MAPE,MAE and RMSE.In particular,the decline is more obvious in rainy and cloudy weather conditions,and it can predict the photovoltaic power generation more accurately.
维普
万方数据