A Review of the Application of Machine Learning in the Prediction of Electric Vehicle Charging Demand

Yixin Wang

doi:10.54691/n3njkv72

Authors

Yixin Wang

DOI:

https://doi.org/10.54691/n3njkv72

Keywords:

Electric Vehicles (EVs); Charging Demand Forecasting; Machine Learning; Spatiotemporal Modeling; Data-driven; Smart Grid.

Abstract

This paper systematically reviews the current application status, technical paths, and future trends of machine learning in electric vehicle (EV) charging demand forecasting. Firstly, starting from the background of global energy transformation and EV popularization, it points out the crucial role of accurate charging demand forecasting in grid stability and facility optimization. Secondly, it summarizes the research progress of machine learning technology in three major scenarios: intelligent management of charging piles, battery state evaluation, and charging load forecasting. It focuses on analyzing the advantages and limitations of models such as random forest, LSTM, GCN, and Transformer in capturing spatiotemporal features and improving prediction accuracy. Furthermore, it discusses core challenges such as data quality, model generalization, real-time performance, and multi-source data fusion, and proposes future research directions such as lightweight model design, transfer learning, federated learning, and V2G collaborative optimization. Finally, through multi-model comparison and technical route analysis, it provides theoretical support and practical reference for building an integrated "vehicle-pile-grid" smart energy system.

Downloads

Download data is not yet available.

References

[1] Statista. (2025). Electric vehicles worldwide - Statistics & facts. https://www. statista. com/ topics/ 1010/electric-mobility/#statisticChapter.

[2] National Development and Reform Commission. (2023). Building a high-quality charging infrastructure system to support the high-quality development of new energy vehicles - Expert interpretation of the Guiding Opinions on Further Building a High-Quality Charging Infrastructure System. http://www.ndrc.gov.cn/xxgk/jd/jd/202306/t20230629_1358020.html.

[3] Ministry of Public Security of the People's Republic of China. (2025). The number of new energy vehicles reaches 31.4 million. https://www.gov.cn/ zhengce/jiedu/ tujie/202501/ content_ 6999527. htm.

[4] Gan, Z. H. (2022). Research on SOC estimation method of pure electric vehicle based on machine learning [Master’s thesis, South China University of Technology].

[5] Li, Y. T. (2024). Research on electric vehicle charging load prediction based on big data. Automobile Forum, 155, 155–157.

[6] Qian, Q., Chen, H., Zheng, Y., et al. (2023). Foreign object detection method for wireless charging of electric vehicles based on machine learning. Modern Electronics Technique, 46(13), 43–48.

[7] Hou, L. N., Zhang, X., Luo, X. F., et al. (2024). A review of data-driven modeling and optimization for electric vehicle charging scheduling. Electric Engineering, 2024(21).

[8] Su, J. S., Li, Y. H., Qin, H., & Fan, J. K. (2020). Electric vehicle driving range prediction based on machine learning model. In *Proceedings of 2020 SAE - China Congress* (pp. 685–689).

[9] Gao, L. X., Gao, Y. P., Zhang, L., et al. (2024). Dynamic charging demand prediction of electric vehicles integrating spatio-temporal big data and user psychological analysis. In 2024 (2nd) Urban Power Grid Technology Innovation Conference.

[10] Geng, P., Yang, H. J., Shi, Z. X., et al. (2024). Electric vehicle charging demand prediction based on spatio-temporal graph convolutional network. Journal of Traffic Engineering, 24(11), 37–45.

[11] Zhao, W. B., Fan, X. D., Sun, S. B., et al. (2025). Scenario-based analysis and ultra-short-term prediction method for electric vehicle charging load. Electric Power Information and Communication Technology, 23(2), 28–37.

[12] Straka, M., Piatriková, L., van Bokhoven, P., et al. (2021). A matrix approach to detect temporal behavioral patterns at electric vehicle charging stations. Transportation Research Procedia, 55, 1353–1360.

[13] Wang, R. J., Gao, H. J., Luo, L. B., et al. (2025). A review of optimal operation research of new distribution systems based on deep reinforcement learning. Electric Power Automation Equipment. DOI:10.16081/j.epae.202505008

[14] Wei, S., & He, H. L. (2024). Research on fault diagnosis method of electric vehicle power battery under big data background. Automotive Safety Technology, (4), 74–76.

[15] Yang, L. (2023). Research on intelligent charging detection and automatic control system design for pure electric vehicles. Automobile Design, (11), 110–112.