Enhanced SOC estimation method for lithium-ion batteries using Bayesian-optimized TCN–LSTM neural networks
Lithium-ion battery state-of-charge (SOC) estimation is crucial for enhancing the safety and performance of battery management systems and prolonging battery lifespan. However, the nonlinear dynamics of battery charging and discharging processes pose significant challenges for achieving accurate SOC...
Saved in:
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
AIP Publishing LLC
2025-01-01
|
| Series: | AIP Advances |
| Online Access: | http://dx.doi.org/10.1063/5.0243811 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1832542775105552384 |
|---|---|
| author | Taotao Hu Xiting Zhu Maokai Tian |
| author_facet | Taotao Hu Xiting Zhu Maokai Tian |
| author_sort | Taotao Hu |
| collection | DOAJ |
| description | Lithium-ion battery state-of-charge (SOC) estimation is crucial for enhancing the safety and performance of battery management systems and prolonging battery lifespan. However, the nonlinear dynamics of battery charging and discharging processes pose significant challenges for achieving accurate SOC estimation. To address this, a novel Bayesian-optimized temporal convolution network (TCN)–long short-term memory (LSTM) network is proposed, combining temporal convolution network (TCN) and long short-term memory (LSTM) to enhance SOC estimation accuracy. This study leverages the CALCE dataset and employs the random forest algorithm for dimensionality reduction, selecting voltage, discharge capacity, and discharge energy as key input features. The TCN is utilized to extract temporal features of the battery parameters, while the LSTM network captures long-term dependencies in sequential data. This study uses Bayesian optimization to fine-tune the hyperparameters of the TCN–LSTM model, ensuring optimal performance. Experimental results demonstrate the effectiveness of the proposed method, achieving a coefficient of determination (R2) of 0.996, a mean absolute error of 0.146%, and a root mean square error of 0.207%. Comparative analyses with other SOC estimation methods and additional datasets further validate the robustness, accuracy, and generalizability of the Bayesian-optimized TCN–LSTM network. |
| format | Article |
| id | doaj-art-31430a5f1e7f4ed285efeea052a628dc |
| institution | Kabale University |
| issn | 2158-3226 |
| language | English |
| publishDate | 2025-01-01 |
| publisher | AIP Publishing LLC |
| record_format | Article |
| series | AIP Advances |
| spelling | doaj-art-31430a5f1e7f4ed285efeea052a628dc2025-02-03T16:40:42ZengAIP Publishing LLCAIP Advances2158-32262025-01-01151015038015038-1410.1063/5.0243811Enhanced SOC estimation method for lithium-ion batteries using Bayesian-optimized TCN–LSTM neural networksTaotao Hu0Xiting Zhu1Maokai Tian2School of Commerce and Circulation, Anhui Institute of International Business, Hefei 231100, ChinaChina University of Mining and Technology (Beijing), Beijing 100083, ChinaSchool of Commerce and Circulation, Anhui Institute of International Business, Hefei 231100, ChinaLithium-ion battery state-of-charge (SOC) estimation is crucial for enhancing the safety and performance of battery management systems and prolonging battery lifespan. However, the nonlinear dynamics of battery charging and discharging processes pose significant challenges for achieving accurate SOC estimation. To address this, a novel Bayesian-optimized temporal convolution network (TCN)–long short-term memory (LSTM) network is proposed, combining temporal convolution network (TCN) and long short-term memory (LSTM) to enhance SOC estimation accuracy. This study leverages the CALCE dataset and employs the random forest algorithm for dimensionality reduction, selecting voltage, discharge capacity, and discharge energy as key input features. The TCN is utilized to extract temporal features of the battery parameters, while the LSTM network captures long-term dependencies in sequential data. This study uses Bayesian optimization to fine-tune the hyperparameters of the TCN–LSTM model, ensuring optimal performance. Experimental results demonstrate the effectiveness of the proposed method, achieving a coefficient of determination (R2) of 0.996, a mean absolute error of 0.146%, and a root mean square error of 0.207%. Comparative analyses with other SOC estimation methods and additional datasets further validate the robustness, accuracy, and generalizability of the Bayesian-optimized TCN–LSTM network.http://dx.doi.org/10.1063/5.0243811 |
| spellingShingle | Taotao Hu Xiting Zhu Maokai Tian Enhanced SOC estimation method for lithium-ion batteries using Bayesian-optimized TCN–LSTM neural networks AIP Advances |
| title | Enhanced SOC estimation method for lithium-ion batteries using Bayesian-optimized TCN–LSTM neural networks |
| title_full | Enhanced SOC estimation method for lithium-ion batteries using Bayesian-optimized TCN–LSTM neural networks |
| title_fullStr | Enhanced SOC estimation method for lithium-ion batteries using Bayesian-optimized TCN–LSTM neural networks |
| title_full_unstemmed | Enhanced SOC estimation method for lithium-ion batteries using Bayesian-optimized TCN–LSTM neural networks |
| title_short | Enhanced SOC estimation method for lithium-ion batteries using Bayesian-optimized TCN–LSTM neural networks |
| title_sort | enhanced soc estimation method for lithium ion batteries using bayesian optimized tcn lstm neural networks |
| url | http://dx.doi.org/10.1063/5.0243811 |
| work_keys_str_mv | AT taotaohu enhancedsocestimationmethodforlithiumionbatteriesusingbayesianoptimizedtcnlstmneuralnetworks AT xitingzhu enhancedsocestimationmethodforlithiumionbatteriesusingbayesianoptimizedtcnlstmneuralnetworks AT maokaitian enhancedsocestimationmethodforlithiumionbatteriesusingbayesianoptimizedtcnlstmneuralnetworks |