High-precision state-of-charge (SOC) prediction is critical for electric vehicle (EV) safety and performance. To address the high computational complexity of existing data-driven methods, which rely on long historical sequences, this paper proposes a purely data-driven end-to-end SOC prediction framework based on sliding-window technology. The framework adopts an SCSSA-optimized convolutional neural networks (CNN)-long short-term memory (LSTM)-attention hybrid model that integrates a CNN for local feature extraction, an LSTM for modeling temporal dependencies, and an attention mechanism for adaptive feature weighting, with an improved sparrow search algorithm for global hyper-parameter optimization. Experiments are conducted using 29 months of operational data from 20 EVs. Results show that the proposed method achieves 14.8%, 8.9%, and 17.6% improvements in mean absolute error, root mean square error, and mean absolute percentage error, respectively, compared with the best benchmark model, with R2 consistently above 0.96. The method demonstrates excellent robustness across seasonal variations and diverse charging patterns, laying a solid technical foundation for SOC prediction in battery management systems.
Xing Zhang, Ju-Qiang Feng, Shunli Wang et al.· Engineering Research Express· 0 citations
Lithium-ion batteries have become the core energy carrier of electric vehicles and energy storage systems due to their high energy density and long cycle life. Their state of charge (SOC) is an important parameter in battery management systems, playing a key role in energy management, safety protection, and life prediction. However, the SOC cannot be measured directly, and it is difficult for traditional estimation algorithms to balance accuracy and real-time under nonlinear, non-Gaussian noise, multiworking conditions, and parameter time-varying conditions. This paper reviews the research progress of SOC estimation based on an improved particle filter (PF); systematically analyzes its comparison with direct measurement, data-driven, physical model, and mixed methods; and focuses on the fusion path of improved PF and the equivalent circuit model, online parameter identification, intelligent optimization algorithm, and deep learning. The results show that the improved PF algorithm can effectively alleviate the problem of particle degradation and significantly improve the estimation accuracy and robustness under the whole life cycle and complex working conditions. Among them, the nonlinear autoregressive neural network combined with particle filtering method has the best performance, achieving a root mean square error of about 0.02% and a maximum error of less than 0.07% under dynamic working conditions, which is significantly better than other methods. The results show that improving PF can not only break through the bottleneck of traditional methods but also provide an important direction for future high-precision SOC estimation. This paper suggests that subsequent research should further focus on intelligent optimization, multisource fusion, and embedded lightweight implementation to promote the large-scale engineering application of improved PF methods in electric vehicles and energy storage systems.
Shunli Wang, Liya Zhang, Mamadou Fall et al.· Journal of Energy Engineerin...· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
