Driven by the dual carbon goals, the widespread adoption of electric vehicles (EVs) has been accelerated by the demand for low-carbon and green transportation. To enhance the accuracy of electric vehicle charging load forecasting, supporting stable grid operation and optimized scheduling of charging infrastructure, a decomposition ensemble learning prediction (DEP) framework based on two-layer optimization and the unscented Kalman filter (UKF) was proposed for EV charging load forecasting. Firstly, a BKA-VMD decomposition model was constructed. The black-winged kite algorithm (BKA) was employed to adaptively optimize the parameters of variational mode decomposition (VMD) to enhance decomposition stability and modal feature extraction capability. Secondly, the improved polar lights optimization (IPLO) algorithm was applied to optimize the hyperparameters of the forecasting models. The UKF-Transformer-LSTM, UKF-Transformer, and LSTM models were used to predict for high frequency, medium frequency and low-frequency components to achieve a better balance between global exploration and local exploitation and improve the convergence and robustness of the prediction process. Finally, the data from electric vehicle charging stations in Region H were employed for empirical analysis. The results show that the proposed model outperforms the benchmark models across multiple evaluation metrics, whose R2 can reach 0.994 1. The proposed model is suitable for ultra-short-term EV charging load forecasting and can provide effective support for real-time EV load dispatching and minute-level response applications in charging stations.
XU Xiao-min, YANG Meng-qi, LI Xiang-ying, XIAO Liang, GUAN Xiao, YANG Yi.
Electric Vehicle Load Prediction Based on Bilayer Optimization and Unscented Kalman Filter Crossover Framework[J].
Science Technology and Engineering, 2026, 26(9): 3817-3829 DOI:10.12404/j.issn.1671-1815.2508456
$\mu {{\overset{\scriptscriptstyle\frown}{x}}_{k}}=\overset{2n}{\mathop{\underset{i=0}{\mathop{\sum }}\,}}\,W_{m}^{\left( i \right)}x_{k}^{-\left( i \right)}$
$\sum{{{{\overset{\scriptscriptstyle\frown}{x}}}_{k}}}=\overset{2n}{\mathop{\underset{i=0}{\mathop{\sum }}\,}}\,W_{\text{c}}^{\left( i \right)}\left[ x_{k}^{-\left( i \right)}-\mu {{{\overset{\scriptscriptstyle\frown}{x}}}_{k}} \right]{{\left[ x_{k}^{-\left( i \right)}-\mu {{{\overset{\scriptscriptstyle\frown}{x}}}_{k}} \right]}^{T}}+\sum{Q}$
$\Sigma {{x}_{k}}{{Z}_{k}}=\overset{2n}{\mathop{\underset{i=0}{\mathop{\sum }}\,}}\,W_{\text{c}}^{\left( i \right)}[x_{k}^{-\left( i \right)}--\mu {{\overset{\scriptscriptstyle\frown}{x}}_{k}}]{{[Z_{k}^{\left( i \right)}-\mu {{Z}_{k}}]}^{\text{T}}}$
Shi RuifengChangtai, HeZhe, et al. Research on the planning method of integrated light-storage-charging electric vehicle charging station[J]. Journal of Solar Energy, 2025, 46(5): 109-118.
ShuZhengyu, LiuWencan, LiHuangqiang, et al. An orderly charging and discharging strategy for electric vehicles based on cooperative game and dynamic time-sharing tariff[J]. Electric Power Engineering Technology, 2025, 44(3): 179-187.
[5]
HuangY, WuS, WangZ, et al. Causality-aware multi-graph convolutional networks with critical node dynamics for electric vehicle charging station load forecasting[J]. IEEE Transactions on Smart Grid, 2025, 16(4): 3210-3225.
[6]
DanishS M, HameedA, RanjhaA, et al. Block-FeDL: electric vehicle charging load forecasting using federated learning and blockchain[J]. IEEE Transactions on Vehicular Technology, 2025, 74(2): 2048-2056.
LuXin, MuYunfei, HaoLu, et al. A coordinated optimisation planning method for multi-type charging facilities considering user charging behaviour decisions[J]. Transactions of the Chinese Society for Power Systems, 2025, 37(12): 1-14.
[9]
ShenX, ShenJ, WangY, et al. A deep reinforcement learning-based reactive power optimization method for distribution networks considering electric vehicle charging load prediction[J]. Sustainable Energy, Grids and Networks, 2025, 43: 101777.
LiHengjie, ZhuJianghao, FuXiaofei, et al. Ultra-short-term load forecasting for electric vehicle charging stations based on ensemble learning[J]. Journal of Shanghai Jiao Tong University, 2022, 56(8): 1004-1013.
[12]
XiaoG, TongH, ShuY, et al. Spatial-temporal load prediction of electric bus charging station based on S2TAT[J]. International Journal of Electrical Power & Energy Systems, 2025, 164: 110446.
LiuYizhou, LiZhenhua, WeiWei, et al. Short-term charging load forecasting for electric vehicles based on a combined LSTM-GNN model[J]. Journal of Power Sources, 2025, 23(8): 145-156.
[15]
BianR, WangL, LiuY, et al. Electric vehicle load forecasting based on convolutional networks with attention mechanism and federated learning method[J]. IET Generation, Transmission & Distribution, 2024, 18(13): 2313-2324.
[16]
ShangY, LiD, LiY, et al. Explainable spatiotemporal multi-task learning for electric vehicle charging demand prediction[J]. Applied Energy, 2025, 384: 125460.
ZhangYanyu; ZhangZhiming; LiuChunyang, et al. Electric vehicle charging load forecasting based on dynamic adaptive graph neural networks[J]. Automation of Electric Power Systems, 2024, 48(7): 86-93.
YuLei, TanZhukui, WangYang, et al. An optimisation method for ordered charging of electric vehicles based on deep reinforcement learning[J]. Southern Power Grid Technology, 2024, 18(12): 148-155.
[23]
PeiZ, ZhangZ, ChenJ, et al. KAN-CNN: a novel framework for electric vehicle load forecasting with enhanced engineering applicability and simplified neural network tuning[J]. Electronics, 2025, 14(3): 414.
[24]
XieT, ZhangY, ZhangG, et al. Research on electric vehicle load forecasting considering regional special event characteristics[J]. Frontiers in Energy Research, 2024, 12: 1341246.
[25]
TianJ, LüY, ZhaoQ, et al. Electric vehicle charging load prediction considering the orderly charging[J]. Energy Reports, 2022, 8: 124-134.
[26]
QuH, KuangH, WangQ, et al. A physics-informed and attention-based graph learning approach for regional electric vehicle charging demand prediction[J]. IEEE Transactions on Intelligent Transportation Systems, 2024, 25(10): 14284-14297.
[27]
ChenF, WangY, DengH, et al. Probabilistic prediction of electric vehicle load in service area incorporating highway spatiotemporal traffic information[J]. Energy, 2025, 329: 136683.
ZhangLin, GaoShengqiang, SongYu, et al. Short-term load forecasting model for power system based on ALIF-VMD quadratic decomposition and NGO-CNN-LSTM[J]. Science Technology and Engineering, 2025, 25(11): 4583-4597.
[32]
XieT, YunX, ZhangG, et al. Charging station cluster load prediction: spatiotemporal multi-graph fusion technology[J]. Renewable and Sustainable Energy Reviews, 2024, 206: 114855.
ZhangYusheng, CaoMin, LeiYu, et al. Short-term load forecasting model for electric vehicles based on SSA-BiGRU-CNN neural network and waveform data correction[J]. Power Grids and Clean Energy, 2025, 41(2): 67-74.
[35]
WangJ, WangW chuan, HuX xue, et al. Black-winged kite algorithm: a nature-inspired meta-heuristic for solving benchmark functions and engineering problems[J]. Artificial Intelligence Review, 2024, 57(4): 98.
XiaMeng, YuHuijun. Short-term power load forecasting based on VMD-IOOA-Bi GRU model and error compensation[J]. Science Technology and Engineering, 2025, 25(27): 11642-11650.
[38]
ZhouM, ShiC, HuF, et al. RC parameter identification and load aggregation analysis of air-conditioning systems: a multi-strategy improved black-winged kite algorithm[J]. Energy and Buildings, 2025, 337: 115641.
[39]
MohapatraS, KaliyaperumalD, PorselviT, et al. Optimal sizing of hybrid renewable energy systems relying on the black winged kite algorithm for performance evaluation[J]. Scientific Reports, 2025, 15(1): 20568.
[40]
GongR, JiangA, HuH, et al. Short-term multi-featured power load forecasting model based on GBKA-VMD-Mambaformer with noise separation error correction under scarce load data[J]. Electric Power Systems Research, 2025, 247: 111846.
[41]
DingC, YuD, LiX, et al. Transformer dissolved gas concentration prediction based on quadratic decomposition reconstruction and BKA-BiLSTM[J]. IEEE Transactions on Dielectrics and Electrical Insulation, 2025, 32(4): 2433-2442.
[42]
FuJ, SongZ, MengJ, et al. Prediction of lithium-ion battery state of health using a deep hybrid kernel extreme learning machine optimized by the improved black-winged kite algorithm[J]. Batteries, 2024, 10(11): 398.
[43]
YuanC, ZhaoD, HeidariA A, et al. Polar lights optimizer: algorithm and applications in image segmentation and feature selection[J]. Neurocomputing, 2024, 607: 128427.
[44]
XuZ, WangZ, LiuR, et al. Efficient multi-UAV path planning in dynamic and complex environments using hybrid polar lights optimization[J]. Journal of King Saud University Computer and Information Sciences, 2025, 37(6): 125.
[45]
HuangT, HuangF, QinZ, et al. An improved polar lights optimization algorithm for global optimization and engineering applications[J]. Scientific Reports, 2025, 15(1): 11581.
LuJinling, HuXinghua, ZhangXuezhe, et al. Dynamic state estimation of distribution networks based on load forecasting and trackless particle filtering[J]. Journal of Electric Power Systems and Automation, 2024, 36(4): 133-140, 158.
FuYueyue, ZhangZheng, YuanKai, et al. A hybrid prediction method of day-ahead electricity price based on Kalman filter[J]. Electrical Automation, 2022, 44(4): 57-60.
FengJiawei, WangHaixin, YangZihao, et al. Multivariate load short-term forecasting based on digital twin and multi-model fusion[J]. Journal of Solar Energy, 2024, 45(10): 97-106.
LiuYisong, KangYingwei. Heavy-duty gas turbine load prediction control based on LSTM-XGBoost integrated model[J]. Thermal Power Engineering, 2025, 40(4): 207-214.
YuDuo, CaoYing, WangHairong, et al. Short-term load forecasting based on ICEEMDAN-PE and IDBO-Informer combined model[J]. China Electric Power, 2025, 58(6): 19-32.
HanLuyang, ZhangWenhui, SongZize, et al. Charging station load prediction under fast and slow charging modes[J]. Journal of Taiyuan University of Technology, 2025, 56(3): 427-435.