Electric vehicle charging load prediction method based on multi-objective modal decomposition and NAHL neural network

To improve the accuracy of electric vehicle charging load prediction, a prediction method based on multi-objective variational mode decomposition (VMD) and automatic artificial neural network with an augmented hidden layer (NAHL) is proposed. The non-dominated sorting genetic algorithm II (NSGAII) is improved by using the simulated binary crossover (SBX) and linear decreasing mutation (LDM), known as the NSGAII-LDSBX algorithm. The improved NSGAII-LDSBX algorithm is used to optimize the parameters of VMD, decompose the signal into several subsequences, and reconstruct the subsequences through fuzzy entropy (FE). Furthermore, the NSGAII-LDSBX is used to optimize the NAHL model and predict each component. An experiment is conducted using the load of the electric vehicle charging station in Jiading District, Shanghai as an example. Analysis shows that compared with other models, the proposed model has better prediction accuracy and can effectively predict the charging load of electric vehicles.