Application of state of health estimation and remaining useful life prediction for lithium-ion batteries based on AT-CNN-BiLSTM DOI Creative Commons

Fengming Zhao,

D. Gao,

Yuan-Ming Cheng

et al.

Scientific Reports, Journal Year: 2024, Volume and Issue: 14(1)

Published: Nov. 23, 2024

Ensuring the long-term safe usage of lithium-ion batteries hinges on accurately estimating State Health $$(\textrm{SOH})$$ and predicting Remaining Useful Life (RUL). This study proposes a novel prediction method based AT-CNN-BiLSTM architecture. Initially, key parameters such as voltage, current, temperature, SOH are extracted averaged for each cycle to ensure uniformity reliability input data. The CNN is utilized extract deep features from data, followed by BiLSTM analyze temporal dependencies in data sequences. Since multidimensional parameter used predict trend batteries, an attention mechanism employed enhance weight highly relevant vectors, improving model's analytical capabilities. Experimental results demonstrate that CNN-BiLSTM-Attention model achieves absolute error 0 RUL prediction, $$R^{2}$$ value greater than 0.9910 , MAPE less 0.9003 . Comparative analysis with hybrid neural network algorithms LSTM, BiLSTM, CNN-LSTM confirms proposed high accuracy stability estimation prediction.

Language: Английский

Multi-level information identification for civil aviation safety risks: A hierarchical multi-branch deep learning approach DOI

Minglan Xiong,

Huawei Wang,

Zhaoguo Hou

et al.

Information Sciences, Journal Year: 2025, Volume and Issue: unknown, P. 121888 - 121888

Published: Jan. 1, 2025

Language: Английский

Citations

0
Log-Cumulative feature alignment for enhanced Prognosis of Aero-Engine remaining Useful life DOI
Yang Ge, Xingxing Jiang, Benlian Xu

et al.

Expert Systems with Applications, Journal Year: 2025, Volume and Issue: unknown, P. 127277 - 127277

Published: March 1, 2025

Language: Английский

Citations

0
Remaining Useful Life Prediction Method Based on Dual-Path Interaction Network with Multiscale Feature Fusion and Dynamic Weight Adaptation DOI Creative Commons
Zhe Lu, Bing Li,

C. D. Fu

et al.

Actuators, Journal Year: 2024, Volume and Issue: 13(10), P. 413 - 413

Published: Oct. 13, 2024

In fields such as manufacturing and aerospace, remaining useful life (RUL) prediction estimates the failure time of high-value assets like industrial equipment aircraft engines by analyzing series data collected from various sensors, enabling more effective predictive maintenance. However, significant temporal diversity operational complexity during operation make it difficult for traditional single-scale, single-dimensional feature extraction methods to effectively capture complex dependencies multi-dimensional interactions. To address this issue, we propose a Dual-Path Interaction Network, integrating Multiscale Temporal-Feature Convolution Fusion Module (MTF-CFM) Dynamic Weight Adaptation (DWAM). This approach adaptively extracts information across different scales, interaction information. Using Commercial Modular Aero-Propulsion System Simulation (C-MAPSS) dataset comprehensive performance evaluation, our method achieved RMSE values 0.0969, 0.1316, 0.086, 0.1148; MAPE 9.72%, 14.51%, 8.04%, 11.27%; Score results 59.93, 209.39, 67.56, 215.35 four categories. Furthermore, MTF-CFM module demonstrated an average improvement 7.12%, 10.62%, 7.21% in RMSE, MAPE, multiple baseline models. These validate effectiveness potential proposed model improving accuracy robustness RUL prediction.

Language: Английский

Citations

1
Application of state of health estimation and remaining useful life prediction for lithium-ion batteries based on AT-CNN-BiLSTM DOI Creative Commons

Fengming Zhao,

D. Gao,

Yuan-Ming Cheng

et al.

Scientific Reports, Journal Year: 2024, Volume and Issue: 14(1)

Published: Nov. 23, 2024

Ensuring the long-term safe usage of lithium-ion batteries hinges on accurately estimating State Health $$(\textrm{SOH})$$ and predicting Remaining Useful Life (RUL). This study proposes a novel prediction method based AT-CNN-BiLSTM architecture. Initially, key parameters such as voltage, current, temperature, SOH are extracted averaged for each cycle to ensure uniformity reliability input data. The CNN is utilized extract deep features from data, followed by BiLSTM analyze temporal dependencies in data sequences. Since multidimensional parameter used predict trend batteries, an attention mechanism employed enhance weight highly relevant vectors, improving model's analytical capabilities. Experimental results demonstrate that CNN-BiLSTM-Attention model achieves absolute error 0 RUL prediction, $$R^{2}$$ value greater than 0.9910 , MAPE less 0.9003 . Comparative analysis with hybrid neural network algorithms LSTM, BiLSTM, CNN-LSTM confirms proposed high accuracy stability estimation prediction.

Language: Английский

Citations

1