Scholarlink

<p>Reliable battery health prediction is essential for sustainable Electric Vehicle (EV) operation. Inaccurate estimation of battery condition can lead to premature battery replacement or inefficient utilization, thereby impacting resource efficiency and overall sustainability. This study presents a data-driven Quantum Machine Learning (QML) framework for battery health prediction that captures complex nonlinear degradation behavior under dynamic operating conditions. The proposed approach is based on a Hybrid Quantum Long Short-Term Memory (Hybrid-QLSTM) model that integrates variational quantum circuits with classical LSTM networks to effectively model both nonlinear feature interactions and temporal dependencies in battery data. The framework is evaluated on multiple benchmark lithium-ion battery datasets using a battery-wise data splitting strategy to ensure realistic and unbiased validation. The results demonstrate that the proposed Hybrid-QLSTM model consistently outperforms classical models across all datasets. Specifically, the model achieves Root Mean Square Error (RMSE) values of 0.0158, 0.0165, and 0.0151 for the NASA, CALCE, and MIT datasets, respectively, while also demonstrating strong performance in terms of Mean Absolute Error (MAE), Coefficient of Determination (<InlineEquation ID="IEq1"> <EquationSource Format="TEX">\(R^2\)</EquationSource> <EquationSource Format="MATHML"><math> <msup> <mi>R</mi> <mn>2</mn> </msup> </math></EquationSource> </InlineEquation>), and Average Absolute Shift (AAS). The analysis further indicates that the integration of quantum feature representation with temporal learning enhances the modeling of nonlinear interactions between battery parameters such as voltage and current, which are inherently coupled through electrochemical processes. Overall, the current study demonstrates the potential of QML for reliable real-world battery health prediction, enabling its application in sustainable EV battery management systems.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Multi-dataset battery health prediction model for sustainable electric vehicle systems using quantum machine learning

  • 期刊论文

  • 发表日期：

    2026年5月29日

查看全文
  • Monika,
  • Sandeep kumar Sood

摘要

Reliable battery health prediction is essential for sustainable Electric Vehicle (EV) operation. Inaccurate estimation of battery condition can lead to premature battery replacement or inefficient utilization, thereby impacting resource efficiency and overall sustainability. This study presents a data-driven Quantum Machine Learning (QML) framework for battery health prediction that captures complex nonlinear degradation behavior under dynamic operating conditions. The proposed approach is based on a Hybrid Quantum Long Short-Term Memory (Hybrid-QLSTM) model that integrates variational quantum circuits with classical LSTM networks to effectively model both nonlinear feature interactions and temporal dependencies in battery data. The framework is evaluated on multiple benchmark lithium-ion battery datasets using a battery-wise data splitting strategy to ensure realistic and unbiased validation. The results demonstrate that the proposed Hybrid-QLSTM model consistently outperforms classical models across all datasets. Specifically, the model achieves Root Mean Square Error (RMSE) values of 0.0158, 0.0165, and 0.0151 for the NASA, CALCE, and MIT datasets, respectively, while also demonstrating strong performance in terms of Mean Absolute Error (MAE), Coefficient of Determination ( \(R^2\) R 2 ), and Average Absolute Shift (AAS). The analysis further indicates that the integration of quantum feature representation with temporal learning enhances the modeling of nonlinear interactions between battery parameters such as voltage and current, which are inherently coupled through electrochemical processes. Overall, the current study demonstrates the potential of QML for reliable real-world battery health prediction, enabling its application in sustainable EV battery management systems.