Scholarlink

<p>In recent years, deep learning techniques have become essential for predicting the Remaining Useful Life (RUL) of Lithium-ion batteries. This study introduces a novel approach to RUL prediction, establishing a new benchmark against existing methodologies. Our main contribution is the development of an advanced method that exceeds previous works in RUL prediction accuracy. Notably, our approach shows good performance compared to hybrid models using the same NASA and CALCE datasets. We created a hybrid deep learning model that achieves high accuracy by integrating key Neural Network architectures: Convolutional Neural Networks (CNN), Deep Neural Networks (DNN), and Gated Recurrent Units (GRU). This innovative model, termed "improved-CGD," was rigorously assessed using metrics such as Mean Absolute Error (MAE) and Root Mean Square Error (RMSE). Additionally, we evaluated battery reliability through Mean Time To Failure (MTTF) based on NASA cycle life data. Our results demonstrate that the improved-CGD model sets a new standard for RUL prediction accuracy in Lithium-ion batteries.</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

RUL prediction for lithium-ion batteries using improved-CGD hybrid model

  • 期刊论文

  • 发表日期：

    2025年6月4日

查看全文
  • Brahim Zraibi,
  • Mohamed Mansouri,
  • Salah Eddine Loukili

摘要

In recent years, deep learning techniques have become essential for predicting the Remaining Useful Life (RUL) of Lithium-ion batteries. This study introduces a novel approach to RUL prediction, establishing a new benchmark against existing methodologies. Our main contribution is the development of an advanced method that exceeds previous works in RUL prediction accuracy. Notably, our approach shows good performance compared to hybrid models using the same NASA and CALCE datasets. We created a hybrid deep learning model that achieves high accuracy by integrating key Neural Network architectures: Convolutional Neural Networks (CNN), Deep Neural Networks (DNN), and Gated Recurrent Units (GRU). This innovative model, termed "improved-CGD," was rigorously assessed using metrics such as Mean Absolute Error (MAE) and Root Mean Square Error (RMSE). Additionally, we evaluated battery reliability through Mean Time To Failure (MTTF) based on NASA cycle life data. Our results demonstrate that the improved-CGD model sets a new standard for RUL prediction accuracy in Lithium-ion batteries.