Education, Science, Technology, Innovation and Life
Open Access
Sign In

New Energy Electric Vehicle Charging Load Forecasting Based on the SSA-CNN-LSTM Model

Download as PDF

DOI: 10.23977/jeeem.2025.080113 | Downloads: 1 | Views: 270

Author(s)

Wenting Ning 1, Guangyun Li 2

Affiliation(s)

1 School of Automation and Information Engineering, Xi'an University of Technology, Xi'an, 710048, China
2 Institute of Electrical Engineering, Chinese Academy of Sciences, Beijing, 100190, China

Corresponding Author

Wenting Ning

ABSTRACT

To reduce power consumption and optimize the charging-discharging compatibility between electric vehicle (EV) charging stations and EVs, this study addresses the challenge of insufficient load forecasting accuracy caused by the stochastic nature of EV charging behavior. A short-term EV charging load forecasting model is proposed based on the Sparrow Search Algorithm-Convolutional Neural Network-Long Short-Term Memory (SSA-CNN-LSTM) hybrid architecture. The model constructs input features incorporating charging time and historical load characteristics. CNN is employed to extract spatial-temporal features from the input data, while the LSTM network enhances temporal prediction accuracy. By establishing a single-input single-output framework, the SSA optimizes critical hyperparameters of the hybrid CNN-LSTM model. Comparative experiments with benchmark models, including Multi-Layer Perceptron (MLP), standalone LSTM, and CNN-LSTM, demonstrate that the optimized SSA-CNN-LSTM model achieves superior short-term forecasting precision. Results indicate significant improvements in prediction accuracy, validating the effectiveness of the proposed method in addressing the uncertainty of EV charging loads and enhancing grid operational efficiency. Innovations: First integration of SSA with CNN-LSTM for EV charging load prediction. Adaptive hyperparameter optimization replacing manual tuning. Practical feasibility: The model is deployable in smart grid management systems to reduce peak-load risks and enhance renewable energy integration, with potential applications in vehicle-to-grid (V2G) interaction scenarios.

KEYWORDS

Deep learning, long short-term memory, Convolutional Neural Networks, Sparrow Search Algorithm, Prediction model

CITE THIS PAPER

Wenting Ning, Guangyun Li, New Energy Electric Vehicle Charging Load Forecasting Based on the SSA-CNN-LSTM Model. Journal of Electrotechnology, Electrical Engineering and Management (2025) Vol. 8: 103-110. DOI: http://dx.doi.org/10.23977/jeeem.2025.080113.

REFERENCES

[1] Cabessa J, Strozecki Y. Refined Kolmogorov complexity of analog, evolving and stochastic recurrent neural networks[J]. Information Sciences,2025,118-122.
[2] Yang M, Li M, Li G. On memory-augmented gated recurrent unit network[J]. International Journal of Forecasting, 2025,41(2):844-858.
[3] Wei J, Tang W, Ditthakit P, et al. Enhancing Drought Forecast Accuracy Through Informer Model Optimization[J]. Land,2025,14(1):126-129.
[4] Sukprasert A, Dananjoyo R, Supachaiwat J, et al. Developing a Model to Predict the Value of Thailand's Gem and Jewelry Exports Using Multiple Linear Regression Analysis[J]. Mathematical Modelling of Engineering Problems, 2024,11(12).
[5] Smyl S, Bergmeir C, Dokumentov A, et al. Local and global trend Bayesian exponential smoothing models[J]. International Journal of Forecasting,2025,41(1):111-127.
[6] Carrillo A J, Hoffmann F, Stuart M A, et al. The Mean-Field Ensemble Kalman Filter: Near-Gaussian Setting[J]. SIAM Journal on Numerical Analysis,2024,62(6):2549-2587.
[7] Wang R, Zhu J, Wang S, et al. Multi-modal emotion recognition using tensor decomposition fusion and self-supervised multi-tasking[J]. International Journal of Multimedia Information Retrieval,2024,13(4):39-44.
[8] Lee J H, Park J D. Collision evasive action timing for MASS using CNN–LSTM-based ship trajectory prediction in restricted area[J]. Ocean Engineering,2024,294-301.
[9] Lu T, Hou S, Xu Y. Ultra-Short-Term Load Forecasting for Customer-Level Integrated Energy Systems Based on Composite VTDS Models[J]. Processes,2023,11(8).
[10] Suqi Z, Ningjing Z, Ziqi Z, et al. Electric Power Load Forecasting Method Based on a Support Vector Machine Optimized by the Improved Seagull Optimization Algorithm[J]. Energies,2022,15(23):9197-9206.

Downloads: 5126
Visits: 242453

Sponsors, Associates, and Links


All published work is licensed under a Creative Commons Attribution 4.0 International License.

Copyright © 2016 - 2031 Clausius Scientific Press Inc. All Rights Reserved.