Scholarlink

<p>Recently, the rise of unconventional renewable energy sources has emphasized the requirement for effective energy storage solutions, with lithium-ion batteries which are playing a pivotal role in addressing this demand. To ensure the reliable and safe functioning of lithium-ion batteries, accurate state of health predictions is essential. This research focuses on proposing a model for predicting the State of health for lithium-ion batteries utilizing an explainable artificial intelligence-based transformer model adapted for tabular features. The proposed model performance is assessed through two training approaches which are Single lithium-ion battery and Multi lithium-ion battery. The single lithium-ion battery approach offers personalized analysis, while the Multi lithium-ion battery approach focuses on generalization. Evaluation metrics demonstrate the model’s effectiveness in capturing battery behavior, and the Shapley additive explanations model enhances interpretability by highlighting the importance of the capacity feature in state of health prediction. Results indicate battery B005 performs best at 70 epochs with Mean Square Error of 7.83E-06 in Single lithium-ion battery, 0.00082365 in Multi lithium-ion battery, battery B006 at 100 epochs with Mean Square Error of 1.39E-05 in Single lithium-ion battery, 0.00460376 in Multi lithium-ion battery, battery B007 consistently maintains low errors at 100 epochs with Mean Square Error of 2.71E-05 in Single lithium-ion battery, 0.00034584 in Multi lithium-ion battery, and battery B0018 presents a balanced trade-off at 70 epochs with Mean Square Error of 8.78E-06 in single lithium-ion battery, 0.0043917 in Multi lithium-ion battery. This research contributes valuable insights into lithium-ion battery health prediction that enhances reliability and efficiency in energy storage systems, contributing to the advancement of sustainable energy solutions.</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

Explainable AI-driven prognostics for battery health in sustainable energy systems

  • 期刊论文

  • 发表日期：

    2026年5月28日

查看全文
  • Heba Mamdouh Farghaly,
  • Asmaa S. Abdo,
  • Ashraf Darwish,
  • Aboul Ella Hassanein

摘要

Recently, the rise of unconventional renewable energy sources has emphasized the requirement for effective energy storage solutions, with lithium-ion batteries which are playing a pivotal role in addressing this demand. To ensure the reliable and safe functioning of lithium-ion batteries, accurate state of health predictions is essential. This research focuses on proposing a model for predicting the State of health for lithium-ion batteries utilizing an explainable artificial intelligence-based transformer model adapted for tabular features. The proposed model performance is assessed through two training approaches which are Single lithium-ion battery and Multi lithium-ion battery. The single lithium-ion battery approach offers personalized analysis, while the Multi lithium-ion battery approach focuses on generalization. Evaluation metrics demonstrate the model’s effectiveness in capturing battery behavior, and the Shapley additive explanations model enhances interpretability by highlighting the importance of the capacity feature in state of health prediction. Results indicate battery B005 performs best at 70 epochs with Mean Square Error of 7.83E-06 in Single lithium-ion battery, 0.00082365 in Multi lithium-ion battery, battery B006 at 100 epochs with Mean Square Error of 1.39E-05 in Single lithium-ion battery, 0.00460376 in Multi lithium-ion battery, battery B007 consistently maintains low errors at 100 epochs with Mean Square Error of 2.71E-05 in Single lithium-ion battery, 0.00034584 in Multi lithium-ion battery, and battery B0018 presents a balanced trade-off at 70 epochs with Mean Square Error of 8.78E-06 in single lithium-ion battery, 0.0043917 in Multi lithium-ion battery. This research contributes valuable insights into lithium-ion battery health prediction that enhances reliability and efficiency in energy storage systems, contributing to the advancement of sustainable energy solutions.