Scholarlink

<p>Lithium-ion batteries play a crucial role in our daily lives. Accurately predicting the State of Health (SOH) and Remaining Useful Life of lithium-ion batteries is essential. To solve the limitations in prediction accuracy of traditional methods, a new approach is proposed that uses an adaptive convergence factor improved Gold Rush Optimizer (GRO) to optimize a Bi-directional Long Short-Term Memory (BiLSTM) neural network for predicting lithium-ion battery capacity. Since the capacity degradation of lithium-ion batteries follows time series patterns, BiLSTM has been chosen to solve this issue effectively. At the same time, to enhance the learning capability of BiLSTM, GRO is used to optimize the network's weights and bias parameters, resulting in higher prediction accuracy. In addition, an adaptive convergence factor is proposed to balance GRO's local and global search capabilities. Compared to other neural networks such as Feed-forward Neural Network (FNN), Long Short-Term Memory (LSTM) neural network, and BiLSTM, as well as different meta-heuristic optimization algorithms to optimize BiLSTM like Differential Evolution (DE), Harris Hawk Optimization (HHO), Whale Optimization Algorithm (WOA) and Great Trevally Optimization (GTO), the experimental results show that the optimized model performs better and predicts battery capacity with increased accuracy.</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

Capacity prediction model for lithium-ion batteries based on bi-directional LSTM neural network optimized by adaptive convergence factor gold rush optimizer

  • 期刊论文

  • 发表日期：

    2025年2月21日

查看全文
  • Xiao-Tian Wang,
  • Jie-Sheng Wang,
  • Song-Bo Zhang,
  • Xun Liu,
  • Yong-Cheng Sun,
  • Yi-Peng Shang-Guan

摘要

Lithium-ion batteries play a crucial role in our daily lives. Accurately predicting the State of Health (SOH) and Remaining Useful Life of lithium-ion batteries is essential. To solve the limitations in prediction accuracy of traditional methods, a new approach is proposed that uses an adaptive convergence factor improved Gold Rush Optimizer (GRO) to optimize a Bi-directional Long Short-Term Memory (BiLSTM) neural network for predicting lithium-ion battery capacity. Since the capacity degradation of lithium-ion batteries follows time series patterns, BiLSTM has been chosen to solve this issue effectively. At the same time, to enhance the learning capability of BiLSTM, GRO is used to optimize the network's weights and bias parameters, resulting in higher prediction accuracy. In addition, an adaptive convergence factor is proposed to balance GRO's local and global search capabilities. Compared to other neural networks such as Feed-forward Neural Network (FNN), Long Short-Term Memory (LSTM) neural network, and BiLSTM, as well as different meta-heuristic optimization algorithms to optimize BiLSTM like Differential Evolution (DE), Harris Hawk Optimization (HHO), Whale Optimization Algorithm (WOA) and Great Trevally Optimization (GTO), the experimental results show that the optimized model performs better and predicts battery capacity with increased accuracy.