For the collaborative demand of new energy power system and electric vehicle charging load
this paper proposes an attribute-augmented spatiotemporal-graph convolutional network- gated recurrent unit is proposed neural network for charging station forecasting (ASC-GCN-GRU) model to accurately predict charging station load in new energy scenarios. Firstly
in view of the limitation of the traditional method of pre-constructing a fixed adjacency matrix
the data adaptive graph generation (DAGG) module is introduced to dynamically mine the spatio-temporal correlation features between charging stations. node adaptive parameter learning (NAPL) module is used to adapt to the diversity of charging models in new energy scenarios. Secondly
considering the environmental factors closely related to new energy output
such as temperature and precipitation
as well as external factors such as date and point of interest (POI)
the node features with enhanced attribute characteristics are established by combining the temporal load of the charging station. Finally
recursive feature elimination with cross-validation (RFECV) was introduced to select the optimal collection with the best performance of ASC-GCN-GRU load forecasting model. The numerical results show that the proposed ASC-GCN-GRU model has better prediction performance than the GCN-GRU
gated recurrent unit neural network (GRU) and sparrow search algorithm-extreme learning machine (SSA-ELM).
QIN Y, WANG J, REN S, et al.Prediction of EV random charging load based on monte carlo simulation method[C]//2023 3rd International Conference on New Energy and Power Engineering, Hangzhou, China, 2023: 295-298.
LIU J, HUANG Y, YUAN H,et al.Residential electric vehicle charging load prediction method based on DBO-ELM[C]//2024 IEEE 4th New Energy and Energy Storage System Control Summit Forum, Hohhot, China, 2024:133-137.
BAI L, YAO L, LI C,et al.Adaptive graph convolutional recurrent network for traffic forecasting[J]. Advances in neural information processing systems, 2020, 33: 17804-17815.
ZHU J, YANG Z, MOURSHED M,et al.Electric vehicle charging load forecasting:a comparative study of deep learning approaches[J]. Energies, 2019, 12(14): 2692.