Etransportation最新文献

Exploiting physics-knowledge from unlabeled data to enhance battery lifetime prediction 利用未标记数据中的物理知识来增强电池寿命预测

IF 17 1区工程技术 Q1 ENERGY & FUELS

Etransportation

Pub Date : 2026-01-29 DOI: 10.1016/j.etran.2026.100560

Aihua Tang , Yuehan Li , Jinpeng Tian , Quanqing Yu , Ning Yu , Yuchen Xu

Accurately predicting battery lifetime is essential for ensuring the long-term operation of electrochemical energy storage systems. While machine learning has provided promising solutions, its performance degrades significantly in the absence of sufficient full-life degradation data on which it heavily depends. In this study, although direct acquisition of remaining useful life and cycles to knee-point labels from battery degradation data without reaching end-of-life is infeasible, valuable physics-related degradation knowledge can still be extracted from such incomplete data to enhance lifetime prediction. Accordingly, a physics-knowledge guided lifetime prediction method is proposed to utilize one-cycle constant-current curve to jointly predict remaining useful life and cycles to knee-point. More critically, this method can implicitly guide convolutional neural network training with incremental capacity knowledge obtained from incomplete-lifespan degradation data. This yields a pre-trained model that can be rapidly adapted using only a few remaining useful life and cycles to knee-point labels. The validity of the proposed method has been extensively validated on three full-lifespan degradation datasets comprising over 40,000 samples. The validation results show that by using only 10 % of the lifetime labels from the samples, the proposed method can achieve prediction with an error of less than 21 cycles on cells with the end-of-life distribution of 100–500 cycles, which reduces the error by more than 50 % compared with the traditional method. In conclusion, this study emphasizes the prospect of enhancing battery lifetime prediction through physics-knowledge in rare-label cases.

准确预测电池寿命是保证电化学储能系统长期运行的关键。虽然机器学习提供了很有前途的解决方案，但在缺乏足够的全寿命退化数据的情况下，它的性能会显著下降，而这正是机器学习所依赖的。在本研究中，虽然在未达到寿命终止的情况下，从电池退化数据中直接获取剩余使用寿命和循环到膝点标签是不可行的，但仍然可以从这些不完整的数据中提取有价值的与物理相关的退化知识，以增强寿命预测。据此，提出了一种物理知识指导下的寿命预测方法，利用单周期恒流曲线联合预测剩余使用寿命和到膝点的周期。更关键的是，该方法可以隐式地指导卷积神经网络训练，使用从不完全寿命退化数据中获得的增量容量知识。这就产生了一个预训练的模型，该模型可以使用少量剩余的使用寿命和周期来快速适应膝点标签。所提出方法的有效性已在包含超过40,000个样本的三个全寿命退化数据集上得到广泛验证。验证结果表明，该方法仅使用样本中10%的寿命标签，就能对寿命终止分布在100-500个周期的细胞实现误差小于21个周期的预测，与传统方法相比，误差降低了50%以上。总之，这项研究强调了在罕见情况下通过物理知识增强电池寿命预测的前景。

{"title":"Exploiting physics-knowledge from unlabeled data to enhance battery lifetime prediction","authors":"Aihua Tang , Yuehan Li , Jinpeng Tian , Quanqing Yu , Ning Yu , Yuchen Xu","doi":"10.1016/j.etran.2026.100560","DOIUrl":"10.1016/j.etran.2026.100560","url":null,"abstract":"<div><div>Accurately predicting battery lifetime is essential for ensuring the long-term operation of electrochemical energy storage systems. While machine learning has provided promising solutions, its performance degrades significantly in the absence of sufficient full-life degradation data on which it heavily depends. In this study, although direct acquisition of remaining useful life and cycles to knee-point labels from battery degradation data without reaching end-of-life is infeasible, valuable physics-related degradation knowledge can still be extracted from such incomplete data to enhance lifetime prediction. Accordingly, a physics-knowledge guided lifetime prediction method is proposed to utilize one-cycle constant-current curve to jointly predict remaining useful life and cycles to knee-point. More critically, this method can implicitly guide convolutional neural network training with incremental capacity knowledge obtained from incomplete-lifespan degradation data. This yields a pre-trained model that can be rapidly adapted using only a few remaining useful life and cycles to knee-point labels. The validity of the proposed method has been extensively validated on three full-lifespan degradation datasets comprising over 40,000 samples. The validation results show that by using only 10 % of the lifetime labels from the samples, the proposed method can achieve prediction with an error of less than 21 cycles on cells with the end-of-life distribution of 100–500 cycles, which reduces the error by more than 50 % compared with the traditional method. In conclusion, this study emphasizes the prospect of enhancing battery lifetime prediction through physics-knowledge in rare-label cases.</div></div>","PeriodicalId":36355,"journal":{"name":"Etransportation","volume":"28 ","pages":"Article 100560"},"PeriodicalIF":17.0,"publicationDate":"2026-01-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146080680","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Facilitating battery quality classification: Early life prediction with sequence-sampling data augmentation 促进电池质量分类：使用序列采样数据增强的早期寿命预测

IF 17 1区工程技术 Q1 ENERGY & FUELS

Etransportation

Pub Date : 2026-01-27 DOI: 10.1016/j.etran.2026.100553

Dongxu Guo , Tianpeng Lu , Tao Sun , Xin Lai , Xuebing Han , Yuejiu Zheng

With the rapid development of electric transportation systems, early-stage quality classification of lithium-ion batteries (LIBs) is crucial for improving the overall performance of battery systems throughout their life-cycle. However, the complex degradation mechanisms of LIBs lead to significant differences in the aging rates of individual cells under identical conditions, which directly affects the accuracy of early-stage quality classification. To address this challenge, this paper proposes a novel framework for predicting the full life-cycle end of life (EOL) of LIBs, combining a sequence sampling-based virtual battery construction scheme with semi-supervised learning. The framework achieves high-precision EOL prediction by augmenting early-cycle data and leveraging the automated feature extraction capabilities of a masked autoencoder (MAE), using only minimal labeled data. Experimental validation demonstrates that the mean absolute percentage error (MAPE) on the validation set can be reduced to 2.6%. This research not only provides a new approach for early-stage battery quality classification utilizing minimal labeled data but also offers robust support for enhancing pack efficiency and enabling pre-screening of abnormal cells, through efficient data utilization and precise predictive capabilities.

随着电动交通系统的快速发展，锂离子电池的早期质量分级对于提高电池系统全生命周期的整体性能至关重要。然而，由于LIBs复杂的降解机制，导致相同条件下单个细胞的衰老速率存在显著差异，这直接影响了早期质量分类的准确性。为了解决这一挑战，本文提出了一个新的框架来预测lib的全生命周期寿命结束（EOL），将基于序列采样的虚拟电池构建方案与半监督学习相结合。该框架通过增加早期周期数据和利用掩码自动编码器（MAE）的自动特征提取功能，仅使用最小的标记数据，实现高精度的EOL预测。实验验证表明，该方法可以将验证集的平均绝对百分比误差（MAPE）降低到2.6%。这项研究不仅为早期电池质量分类提供了新的方法，利用最小的标记数据，而且通过有效的数据利用和精确的预测能力，为提高电池组效率和实现异常电池的预筛选提供了强有力的支持。

{"title":"Facilitating battery quality classification: Early life prediction with sequence-sampling data augmentation","authors":"Dongxu Guo , Tianpeng Lu , Tao Sun , Xin Lai , Xuebing Han , Yuejiu Zheng","doi":"10.1016/j.etran.2026.100553","DOIUrl":"10.1016/j.etran.2026.100553","url":null,"abstract":"<div><div>With the rapid development of electric transportation systems, early-stage quality classification of lithium-ion batteries (LIBs) is crucial for improving the overall performance of battery systems throughout their life-cycle. However, the complex degradation mechanisms of LIBs lead to significant differences in the aging rates of individual cells under identical conditions, which directly affects the accuracy of early-stage quality classification. To address this challenge, this paper proposes a novel framework for predicting the full life-cycle end of life (EOL) of LIBs, combining a sequence sampling-based virtual battery construction scheme with semi-supervised learning. The framework achieves high-precision EOL prediction by augmenting early-cycle data and leveraging the automated feature extraction capabilities of a masked autoencoder (MAE), using only minimal labeled data. Experimental validation demonstrates that the mean absolute percentage error (MAPE) on the validation set can be reduced to 2.6%. This research not only provides a new approach for early-stage battery quality classification utilizing minimal labeled data but also offers robust support for enhancing pack efficiency and enabling pre-screening of abnormal cells, through efficient data utilization and precise predictive capabilities.</div></div>","PeriodicalId":36355,"journal":{"name":"Etransportation","volume":"28 ","pages":"Article 100553"},"PeriodicalIF":17.0,"publicationDate":"2026-01-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146080681","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Optimization of mass transport in PEM electrolysis cell via Triply Periodic Minimal Surfaces (TPMS) based integrated transport layer 基于三周期最小表面（TPMS）的集成传输层优化PEM电解池的质量传输

IF 17 1区工程技术 Q1 ENERGY & FUELS

Etransportation

Pub Date : 2026-01-23 DOI: 10.1016/j.etran.2026.100551

Dachen Tao , Yudong Zhang , Jun Li , Xun Zhu , Dingding Ye , Yang Yang , Masrur Khodiev , Qiang Liao

The decarbonization of heavy-duty transport depends critically on affordable green hydrogen, with proton exchange membrane electrolysis cell (PEMEC) serving as a key green-hydrogen production technology due to its high efficiency and dynamic response to renewable power. However, severe mass transfer limitations at the anode—primarily caused by oxygen bubble accumulation—restrict PEMEC performance at high current densities (>2 A cm⁻²), thereby elevating hydrogen production cost and hindering its competitiveness for mobility applications. In the study, an innovative integrated transport layer (ITL) is proposed by inspiring from the triply periodic minimal surface (TPMS) structure. The TPMS structure is optimized for mass transfer through gas-liquid two-phase flow simulations. Guided by the results, the TPMS-based flow field is fabricated via 3D printing and evaluated in an electrolyzer. The simulations reveal that the TPMS structure significantly enhances gas-liquid distribution uniformity. Specifically, it increases water saturation at the catalytic layer interface by 110 %, and improves the oxygen distribution uniformity index by 78 % over conventional flow fields. The TPMS flow field reduces the cell voltage by 50 mV at 2 A cm⁻² and decreases mass transfer loss by 44.6 %, compared to conventional serpentine flow fields. This work provides a critical theoretical foundation for designing high-performance mass transport structures in PEMEC.

重型运输的脱碳关键取决于价格合理的绿色氢，而质子交换膜电解电池（PEMEC）因其高效率和对可再生能源的动态响应而成为关键的绿色氢生产技术。然而，阳极处严重的传质限制（主要是由氧泡积累引起的）限制了PEMEC在高电流密度（>2 A cm - 2）下的性能，从而提高了制氢成本，阻碍了其在迁移应用中的竞争力。本文从三周期最小表面（TPMS）结构出发，提出了一种创新的集成传输层（ITL）。通过气液两相流模拟，优化了TPMS结构的传质性能。在实验结果的指导下，利用3D打印技术制作了基于tpms的流场，并在电解槽中进行了评估。仿真结果表明，TPMS结构显著提高了气液分布均匀性。与常规流场相比，催化层界面水饱和度提高了110%，氧分布均匀性指数提高了78%。与传统的蛇形流场相比，TPMS流场在2 A cm−2时可使电池电压降低50 mV，传质损失降低44.6%。这项工作为设计高性能的质量传输结构提供了重要的理论基础。

{"title":"Optimization of mass transport in PEM electrolysis cell via Triply Periodic Minimal Surfaces (TPMS) based integrated transport layer","authors":"Dachen Tao , Yudong Zhang , Jun Li , Xun Zhu , Dingding Ye , Yang Yang , Masrur Khodiev , Qiang Liao","doi":"10.1016/j.etran.2026.100551","DOIUrl":"10.1016/j.etran.2026.100551","url":null,"abstract":"<div><div>The decarbonization of heavy-duty transport depends critically on affordable green hydrogen, with proton exchange membrane electrolysis cell (PEMEC) serving as a key green-hydrogen production technology due to its high efficiency and dynamic response to renewable power. However, severe mass transfer limitations at the anode—primarily caused by oxygen bubble accumulation—restrict PEMEC performance at high current densities (>2 A cm<sup>−2</sup>), thereby elevating hydrogen production cost and hindering its competitiveness for mobility applications. In the study, an innovative integrated transport layer (ITL) is proposed by inspiring from the triply periodic minimal surface (TPMS) structure. The TPMS structure is optimized for mass transfer through gas-liquid two-phase flow simulations. Guided by the results, the TPMS-based flow field is fabricated via 3D printing and evaluated in an electrolyzer. The simulations reveal that the TPMS structure significantly enhances gas-liquid distribution uniformity. Specifically, it increases water saturation at the catalytic layer interface by 110 %, and improves the oxygen distribution uniformity index by 78 % over conventional flow fields. The TPMS flow field reduces the cell voltage by 50 mV at 2 A cm<sup>−2</sup> and decreases mass transfer loss by 44.6 %, compared to conventional serpentine flow fields. This work provides a critical theoretical foundation for designing high-performance mass transport structures in PEMEC.</div></div>","PeriodicalId":36355,"journal":{"name":"Etransportation","volume":"28 ","pages":"Article 100551"},"PeriodicalIF":17.0,"publicationDate":"2026-01-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146080682","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

STL-LLM: A seasonal-trend decomposition-enhanced large language model for battery capacity aging trajectory prediction 基于季节趋势分解的电池容量老化轨迹预测大语言模型

IF 17 1区工程技术 Q1 ENERGY & FUELS

Etransportation

Pub Date : 2026-01-19 DOI: 10.1016/j.etran.2026.100549

Xuan Liu , Yan Lyu , Jie Gao , Cunfu He , Mengmeng Geng , Maosong Fan

Reliable monitoring of lithium-ion battery health is critical for electric vehicles and energy storage systems. Accurate prediction of the remaining capacity aging trajectory remains essential for battery management, yet current machine learning approaches often fail to capture long-term temporal dependencies in degradation data or leverage heterogeneous datasets effectively. In particular, while Pre-trained Large Language Models (LLMs) exhibit powerful reasoning abilities, their application to time-series-based capacity aging trajectory prediction is hindered by a fundamental modality mismatch. To address this, we propose STL-LLM, a novel framework integrating Seasonal-Trend decomposition using LOESS (STL) with frozen LLMs. STL-LLM disentangles battery health sequences into seasonal and trend components, reprograms these temporal features into text-aligned prompts, and employs prefix-based prompting to enhance temporal reasoning. The LLM's output is projected to generate a capacity aging trajectory prediction. Evaluations demonstrate STL-LLM's state-of-the-art accuracy across three public battery datasets, with consistent superiority in ablation and sensitivity studies. From a methodological perspective, STL-LLM offers a principled cross-modal representation learning solution for time-series forecasting, enabling frozen LLM deployment in non-text domains with minimal tuning. Practically, the framework provides a scalable and generalizable approach for battery prognostics, with potential applications in predictive maintenance and cloud-based battery management systems. More broadly, this work bridges the modality gap between structured time-series signals and pre-trained language models. It introduces a transferable paradigm for leveraging LLMs, which holds significant potential for advancing scientific time-series analysis and sequence modeling. While the direct application lies in battery health monitoring for new energy vehicles, this framework creates a pathway for broader impacts across energy systems.

对锂离子电池健康状况的可靠监测对电动汽车和储能系统至关重要。准确预测剩余容量老化轨迹对于电池管理至关重要，但目前的机器学习方法往往无法捕获退化数据中的长期时间依赖性或有效利用异构数据集。特别是，虽然预训练的大型语言模型（llm）具有强大的推理能力，但它们在基于时间序列的容量老化轨迹预测中的应用受到基本模态不匹配的阻碍。为了解决这个问题，我们提出了STL- llm，这是一个利用黄土（STL）和冷冻llm结合季节趋势分解的新框架。STL-LLM将电池健康序列分解为季节和趋势组件，将这些时间特征重新编程为与文本对齐的提示，并使用基于前缀的提示来增强时间推理。预计LLM的输出将生成产能老化轨迹预测。评估表明，STL-LLM在三个公共电池数据集上具有最先进的准确性，在烧蚀和灵敏度研究中具有一贯的优势。从方法学的角度来看，STL-LLM为时间序列预测提供了原则性的跨模态表示学习解决方案，使LLM能够以最小的调优在非文本域中进行冻结部署。实际上，该框架为电池预测提供了一种可扩展和通用的方法，在预测性维护和基于云的电池管理系统中具有潜在的应用前景。更广泛地说，这项工作弥合了结构化时间序列信号和预训练语言模型之间的模态差距。它为利用llm引入了一个可转移的范例，这对于推进科学的时间序列分析和序列建模具有重要的潜力。虽然直接应用于新能源汽车的电池健康监测，但该框架为整个能源系统的更广泛影响创造了途径。

{"title":"STL-LLM: A seasonal-trend decomposition-enhanced large language model for battery capacity aging trajectory prediction","authors":"Xuan Liu , Yan Lyu , Jie Gao , Cunfu He , Mengmeng Geng , Maosong Fan","doi":"10.1016/j.etran.2026.100549","DOIUrl":"10.1016/j.etran.2026.100549","url":null,"abstract":"<div><div>Reliable monitoring of lithium-ion battery health is critical for electric vehicles and energy storage systems. Accurate prediction of the remaining capacity aging trajectory remains essential for battery management, yet current machine learning approaches often fail to capture long-term temporal dependencies in degradation data or leverage heterogeneous datasets effectively. In particular, while Pre-trained Large Language Models (LLMs) exhibit powerful reasoning abilities, their application to time-series-based capacity aging trajectory prediction is hindered by a fundamental modality mismatch. To address this, we propose STL-LLM, a novel framework integrating Seasonal-Trend decomposition using LOESS (STL) with frozen LLMs. STL-LLM disentangles battery health sequences into seasonal and trend components, reprograms these temporal features into text-aligned prompts, and employs prefix-based prompting to enhance temporal reasoning. The LLM's output is projected to generate a capacity aging trajectory prediction. Evaluations demonstrate STL-LLM's state-of-the-art accuracy across three public battery datasets, with consistent superiority in ablation and sensitivity studies. From a methodological perspective, STL-LLM offers a principled cross-modal representation learning solution for time-series forecasting, enabling frozen LLM deployment in non-text domains with minimal tuning. Practically, the framework provides a scalable and generalizable approach for battery prognostics, with potential applications in predictive maintenance and cloud-based battery management systems. More broadly, this work bridges the modality gap between structured time-series signals and pre-trained language models. It introduces a transferable paradigm for leveraging LLMs, which holds significant potential for advancing scientific time-series analysis and sequence modeling. While the direct application lies in battery health monitoring for new energy vehicles, this framework creates a pathway for broader impacts across energy systems.</div></div>","PeriodicalId":36355,"journal":{"name":"Etransportation","volume":"28 ","pages":"Article 100549"},"PeriodicalIF":17.0,"publicationDate":"2026-01-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146006571","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Layered electro-thermal modeling and self-heating optimization for large-capacity Li-ion batteries 大容量锂离子电池分层电热建模及自热优化

IF 17 1区工程技术 Q1 ENERGY & FUELS

Etransportation

Pub Date : 2026-01-17 DOI: 10.1016/j.etran.2026.100544

Shenghao Li , Cheng Lin , Yu Tian , Zhenyi Tao , Peng Xie

Integrated internal/external heating at low temperatures is an important approach to improving the environmental adaptability of lithium-ion batteries. However, for large-capacity batteries, it faces the problem of temperature non-uniformity caused by inhomogeneous heat production and slow heat diffusion. Due to the lack of effective modeling of internal non-uniformity, the impact of temperature gradients during heating on battery degradation remains unclear, and there is a lack of theoretical constraints on temperature non-uniformity. In this study, a layered one-dimensional electro-thermal coupled model with 6 sections is proposed to analyze electro-thermal non-uniformity during battery heating, followed by experimental validation. Based on the model, a multi-stage variable duty cycle heating strategy is obtained through multi-objective optimization and constraints considering aging. Subsequently, the characteristics of internal non-uniformity are further analyzed to reveal the theoretically based control patterns of temperature non-uniformity. The results show that under various operating conditions, the relative error of the model is less than 5 %, and the calculation time for a single heating is less than 10 s. The proposed strategy can increase the heating rate by up to 12.5 % without increasing degradation. It is found that a control strategy with dynamically increasing heating power can ensure rapid heating while improving electro-thermal uniformity and reducing battery degradation. This work solves a critical challenge for electric vehicles, enabling rapid cold-start without accelerating degradation in large-format power batteries. The proposed model and method have broad applicability in the field of battery thermal management.

低温内外一体化加热是提高锂离子电池环境适应性的重要途径。但对于大容量电池来说，由于产热不均匀、热扩散缓慢，存在温度不均匀的问题。由于缺乏对内部不均匀性的有效建模，加热过程中温度梯度对电池退化的影响尚不清楚，并且缺乏对温度不均匀性的理论约束。本研究提出了一种分层的6段一维电热耦合模型来分析电池加热过程中的电热不均匀性，并进行了实验验证。在此基础上，通过多目标优化和考虑老化约束，得到了多阶段变占空比加热策略。随后，进一步分析了内部不均匀性的特性，揭示了基于理论的温度不均匀性控制模式。结果表明，在各种工况下，该模型的相对误差小于5%，单次加热的计算时间小于10 s。所提出的策略可以在不增加降解的情况下将加热速率提高12.5%。研究发现，采用动态增加加热功率的控制策略，既能保证快速加热，又能提高电热均匀性，减少电池退化。这项工作解决了电动汽车的一个关键挑战，实现了快速冷启动，而不会加速大型动力电池的退化。该模型和方法在电池热管理领域具有广泛的适用性。

{"title":"Layered electro-thermal modeling and self-heating optimization for large-capacity Li-ion batteries","authors":"Shenghao Li , Cheng Lin , Yu Tian , Zhenyi Tao , Peng Xie","doi":"10.1016/j.etran.2026.100544","DOIUrl":"10.1016/j.etran.2026.100544","url":null,"abstract":"<div><div>Integrated internal/external heating at low temperatures is an important approach to improving the environmental adaptability of lithium-ion batteries. However, for large-capacity batteries, it faces the problem of temperature non-uniformity caused by inhomogeneous heat production and slow heat diffusion. Due to the lack of effective modeling of internal non-uniformity, the impact of temperature gradients during heating on battery degradation remains unclear, and there is a lack of theoretical constraints on temperature non-uniformity. In this study, a layered one-dimensional electro-thermal coupled model with 6 sections is proposed to analyze electro-thermal non-uniformity during battery heating, followed by experimental validation. Based on the model, a multi-stage variable duty cycle heating strategy is obtained through multi-objective optimization and constraints considering aging. Subsequently, the characteristics of internal non-uniformity are further analyzed to reveal the theoretically based control patterns of temperature non-uniformity. The results show that under various operating conditions, the relative error of the model is less than 5 %, and the calculation time for a single heating is less than 10 s. The proposed strategy can increase the heating rate by up to 12.5 % without increasing degradation. It is found that a control strategy with dynamically increasing heating power can ensure rapid heating while improving electro-thermal uniformity and reducing battery degradation. This work solves a critical challenge for electric vehicles, enabling rapid cold-start without accelerating degradation in large-format power batteries. The proposed model and method have broad applicability in the field of battery thermal management.</div></div>","PeriodicalId":36355,"journal":{"name":"Etransportation","volume":"28 ","pages":"Article 100544"},"PeriodicalIF":17.0,"publicationDate":"2026-01-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146025729","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Joint prediction of polarization losses and internal states in fuel cell via time–frequency feature fusion and machine learning 基于时频特征融合和机器学习的燃料电池极化损耗和内部状态联合预测

IF 17 1区工程技术 Q1 ENERGY & FUELS

Etransportation

Pub Date : 2026-01-16 DOI: 10.1016/j.etran.2026.100548

Jiaping Xie , Kunyi Feng , Hao Yuan , Zhaoming Liu , Chao Wang , Wei Tang , Yabo Wang , Penglong Bao , Xuezhe Wei , Haifeng Dai

The real-time decoupling of polarization losses and internal states is fundamental for extending the lifespan of proton exchange membrane fuel cells (PEMFCs), yet existing methods struggle with the trade-off between measurement speed and information depth. This study proposes a novel synergistic time–frequency fusion framework for the joint prediction of polarization losses and internal state distributions. By leveraging a two-dimensional multi-scale agglomerate model, we construct a high-fidelity dataset that captures the intricate mapping between frequency-domain signatures and microscopic reaction distributions. A comprehensive sensitivity analysis identifies impedance amplitude and phase angle at 79.43 Hz and 10 Hz as optimal features, capturing critical information about reaction interfaces and mass transport that are often neglected in traditional time-domain analysis. These identified features, integrated with macro-level operating conditions, are fed into a Gaussian Process Regression (GPR) model. Results demonstrate superior predictive accuracy with a Mean Absolute Percentage Error (MAPE) below 4% for all key variables. Furthermore, the model exhibits exceptional robustness under 30 dB noise levels and dynamic New European Driving Cycle (NEDC) conditions, successfully tracking transient concentration fluctuations. This work offers a highly efficient and cost-effective approach for online health management by extracting physical insight from less on-board measurement information.

极化损失和内部状态的实时解耦是延长质子交换膜燃料电池（pemfc）寿命的基础，但现有的方法在测量速度和信息深度之间进行权衡。本研究提出了一种新的时频协同融合框架，用于联合预测极化损失和内态分布。通过利用二维多尺度凝聚体模型，我们构建了一个高保真的数据集，该数据集捕获了频域特征和微观反应分布之间的复杂映射。综合灵敏度分析确定了79.43 Hz和10 Hz的阻抗幅值和相位角为最佳特征，捕获了传统时域分析中经常忽略的反应界面和质量输运的关键信息。这些识别的特征，与宏观层面的操作条件相结合，被输入到高斯过程回归（GPR）模型中。结果表明，所有关键变量的平均绝对百分比误差（MAPE）低于4%，具有优越的预测准确性。此外，该模型在30 dB噪声水平和动态新欧洲驾驶循环（NEDC）条件下表现出出色的鲁棒性，成功跟踪瞬态浓度波动。这项工作通过从较少的机载测量信息中提取物理洞察，为在线健康管理提供了一种高效且具有成本效益的方法。

{"title":"Joint prediction of polarization losses and internal states in fuel cell via time–frequency feature fusion and machine learning","authors":"Jiaping Xie , Kunyi Feng , Hao Yuan , Zhaoming Liu , Chao Wang , Wei Tang , Yabo Wang , Penglong Bao , Xuezhe Wei , Haifeng Dai","doi":"10.1016/j.etran.2026.100548","DOIUrl":"10.1016/j.etran.2026.100548","url":null,"abstract":"<div><div>The real-time decoupling of polarization losses and internal states is fundamental for extending the lifespan of proton exchange membrane fuel cells (PEMFCs), yet existing methods struggle with the trade-off between measurement speed and information depth. This study proposes a novel synergistic time–frequency fusion framework for the joint prediction of polarization losses and internal state distributions. By leveraging a two-dimensional multi-scale agglomerate model, we construct a high-fidelity dataset that captures the intricate mapping between frequency-domain signatures and microscopic reaction distributions. A comprehensive sensitivity analysis identifies impedance amplitude and phase angle at 79.43 Hz and 10 Hz as optimal features, capturing critical information about reaction interfaces and mass transport that are often neglected in traditional time-domain analysis. These identified features, integrated with macro-level operating conditions, are fed into a Gaussian Process Regression (GPR) model. Results demonstrate superior predictive accuracy with a Mean Absolute Percentage Error (MAPE) below 4% for all key variables. Furthermore, the model exhibits exceptional robustness under 30 dB noise levels and dynamic New European Driving Cycle (NEDC) conditions, successfully tracking transient concentration fluctuations. This work offers a highly efficient and cost-effective approach for online health management by extracting physical insight from less on-board measurement information.</div></div>","PeriodicalId":36355,"journal":{"name":"Etransportation","volume":"28 ","pages":"Article 100548"},"PeriodicalIF":17.0,"publicationDate":"2026-01-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146006523","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

From hype to impact: A roadmap for trustworthy battery AI 从炒作到影响：值得信赖的电池AI路线图

IF 17 1区工程技术 Q1 ENERGY & FUELS

Etransportation

Pub Date : 2026-01-13 DOI: 10.1016/j.etran.2026.100546

Jingyuan Zhao , Yunhong Che , Yuqi Li , Stephen Harris

Artificial intelligence is increasingly used across the battery lifecycle, including materials screening, manufacturing quality control, diagnostics, and second-life assessment, yet its real-world impact remains limited by fragmented data, constrained interpretability, and the absence of deployment-ready governance. This Commentary proposes a roadmap for trustworthy, field-ready battery AI shaped by three structural priorities. First, open and standardized data ecosystems, supported by interoperable metadata and benchmark tasks, are essential for overcoming heterogeneous and siloed datasets. Second, privacy-preserving industrial collaboration can be enabled through federated learning, encrypted inference, synthetic data, and auditable governance frameworks aligned with safety-critical expectations. Third, physically grounded and interpretable models that embed electrochemical priors, enforce physical constraints, and quantify uncertainty are required to ensure robustness across chemistries, formats, and operating regimes. This roadmap reframes battery AI from isolated performance gains toward trustworthy, system-level intelligence capable of delivering sustained scientific and industrial impact.

人工智能在电池生命周期中的应用越来越广泛，包括材料筛选、制造质量控制、诊断和二次使用评估，但其对现实世界的影响仍然受到数据碎片化、可解释性受限以及缺乏部署就绪治理的限制。本评论提出了一个值得信赖的、现场就绪的电池AI路线图，该路线图由三个结构优先事项构成。首先，开放和标准化的数据生态系统，由可互操作的元数据和基准任务支持，对于克服异构和孤立的数据集至关重要。其次，可以通过联邦学习、加密推理、合成数据和符合安全关键期望的可审计治理框架来实现保护隐私的工业协作。第三，物理基础和可解释的模型需要嵌入电化学先验、强制物理约束和量化不确定性，以确保跨化学、格式和操作制度的稳健性。该路线图将电池AI从孤立的性能提升重新定义为可信赖的系统级智能，能够提供持续的科学和工业影响。

{"title":"From hype to impact: A roadmap for trustworthy battery AI","authors":"Jingyuan Zhao , Yunhong Che , Yuqi Li , Stephen Harris","doi":"10.1016/j.etran.2026.100546","DOIUrl":"10.1016/j.etran.2026.100546","url":null,"abstract":"<div><div>Artificial intelligence is increasingly used across the battery lifecycle, including materials screening, manufacturing quality control, diagnostics, and second-life assessment, yet its real-world impact remains limited by fragmented data, constrained interpretability, and the absence of deployment-ready governance. This Commentary proposes a roadmap for trustworthy, field-ready battery AI shaped by three structural priorities. First, open and standardized data ecosystems, supported by interoperable metadata and benchmark tasks, are essential for overcoming heterogeneous and siloed datasets. Second, privacy-preserving industrial collaboration can be enabled through federated learning, encrypted inference, synthetic data, and auditable governance frameworks aligned with safety-critical expectations. Third, physically grounded and interpretable models that embed electrochemical priors, enforce physical constraints, and quantify uncertainty are required to ensure robustness across chemistries, formats, and operating regimes. This roadmap reframes battery AI from isolated performance gains toward trustworthy, system-level intelligence capable of delivering sustained scientific and industrial impact.</div></div>","PeriodicalId":36355,"journal":{"name":"Etransportation","volume":"28 ","pages":"Article 100546"},"PeriodicalIF":17.0,"publicationDate":"2026-01-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146006572","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Optimal economic integrated thermal management of battery and cabin for connected electric vehicles considering battery degradation 考虑电池退化的网联电动汽车电池与驾驶室综合热管理经济优化

IF 17 1区工程技术 Q1 ENERGY & FUELS

Etransportation

Pub Date : 2026-01-01 DOI: 10.1016/j.etran.2025.100540

Qian Ma , Yan Ma , Jinwu Gao , Hong Chen

The integrated thermal management system (ITMS) for the battery and cabin is essential to improve thermal safety, energy efficiency, battery lifespan, and passenger comfort in connected electric vehicle (CEV). The ITMS consumes considerable energy to maintain battery and cabin temperatures in the optimal range, which severely reduces the CEV’s driving range. To solve the ITMS optimization problem for CEV and achieve eco-cooling, this article proposes a two-stage optimization strategy for ITMS based on multi-horizon economic nonlinear model predictive control (TS-MH-ENMPC), which considers the total economic cost of cooling system energy consumption and battery degradation. Firstly, a control-oriented nonlinear ITMS model is developed to predict the battery and cabin temperature changes. Then, a two-stage cooling optimization strategy based on economic nonlinear model predictive control (MPC) is proposed to achieve optimal driving economy, which divides the ITMS into fast cooling stage and temperature maintenance stage with different cooling objectives. Finally, to address the multi-timescale problem of slow dynamic response in thermal system and fast response in power transfer, a multi-prediction horizon MPC framework is introduced to fully utilize the intelligent transportation system (ITS) information to achieve optimal economic performance over long prediction horizon, which solves the optimization problem of the integrated system with dynamic responses at different time scales and reduces the computational burden. The simulation results under various conditions show that the proposed method reduces the total economic cost of energy consumption and battery degradation. And a sensitivity analysis is conducted on ambient temperatures, battery prices, and electricity prices. Compared to the traditional MPC, rule-based, the total economic cost of the TS-MH-ENMPC is reduced by 5.24% and 7.09%, and the driving distance is increased by 3.03% and 6.65%. The co-simulation results on real-world traffic data show that the proposed method improves driving economy and thermal performance under preview information uncertainty and model mismatch.

用于电池和驾驶室的集成热管理系统（ITMS）对于提高互联电动汽车（CEV）的热安全性、能效、电池寿命和乘客舒适度至关重要。ITMS消耗了大量的能量来保持电池和座舱温度在最佳范围内，这严重降低了CEV的行驶里程。为了解决电动汽车ITMS优化问题，实现生态冷却，本文提出了一种考虑冷却系统能耗和电池退化总经济成本的基于多水平经济非线性模型预测控制（TS-MH-ENMPC）的两阶段ITMS优化策略。首先，建立了面向控制的非线性ITMS模型来预测电池和舱室温度的变化。然后，提出了一种基于经济非线性模型预测控制（MPC）的两阶段冷却优化策略，将ITMS分为快速冷却阶段和温度维持阶段，并根据不同的冷却目标实现最优的驾驶经济性。最后，针对热力系统动态响应慢、输电系统动态响应快的多时间尺度问题，引入多预测层MPC框架，充分利用智能交通系统（ITS）信息实现长预测层的最优经济性能，解决了不同时间尺度下综合系统动态响应的优化问题，减少了计算量。各种条件下的仿真结果表明，该方法降低了能量消耗和电池退化的总经济成本。并对环境温度、电池价格、电价进行敏感性分析。与基于规则的传统MPC相比，TS-MH-ENMPC的总经济成本分别降低了5.24%和7.09%，行驶距离分别增加了3.03%和6.65%。实际交通数据的联合仿真结果表明，该方法在预览信息不确定和模型不匹配的情况下提高了驾驶经济性和热性能。

{"title":"Optimal economic integrated thermal management of battery and cabin for connected electric vehicles considering battery degradation","authors":"Qian Ma , Yan Ma , Jinwu Gao , Hong Chen","doi":"10.1016/j.etran.2025.100540","DOIUrl":"10.1016/j.etran.2025.100540","url":null,"abstract":"<div><div>The integrated thermal management system (ITMS) for the battery and cabin is essential to improve thermal safety, energy efficiency, battery lifespan, and passenger comfort in connected electric vehicle (CEV). The ITMS consumes considerable energy to maintain battery and cabin temperatures in the optimal range, which severely reduces the CEV’s driving range. To solve the ITMS optimization problem for CEV and achieve eco-cooling, this article proposes a two-stage optimization strategy for ITMS based on multi-horizon economic nonlinear model predictive control (TS-MH-ENMPC), which considers the total economic cost of cooling system energy consumption and battery degradation. Firstly, a control-oriented nonlinear ITMS model is developed to predict the battery and cabin temperature changes. Then, a two-stage cooling optimization strategy based on economic nonlinear model predictive control (MPC) is proposed to achieve optimal driving economy, which divides the ITMS into fast cooling stage and temperature maintenance stage with different cooling objectives. Finally, to address the multi-timescale problem of slow dynamic response in thermal system and fast response in power transfer, a multi-prediction horizon MPC framework is introduced to fully utilize the intelligent transportation system (ITS) information to achieve optimal economic performance over long prediction horizon, which solves the optimization problem of the integrated system with dynamic responses at different time scales and reduces the computational burden. The simulation results under various conditions show that the proposed method reduces the total economic cost of energy consumption and battery degradation. And a sensitivity analysis is conducted on ambient temperatures, battery prices, and electricity prices. Compared to the traditional MPC, rule-based, the total economic cost of the TS-MH-ENMPC is reduced by 5.24% and 7.09%, and the driving distance is increased by 3.03% and 6.65%. The co-simulation results on real-world traffic data show that the proposed method improves driving economy and thermal performance under preview information uncertainty and model mismatch.</div></div>","PeriodicalId":36355,"journal":{"name":"Etransportation","volume":"27 ","pages":"Article 100540"},"PeriodicalIF":17.0,"publicationDate":"2026-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145925129","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Unveiling thermal risks of presumed safe lithium iron phosphate batteries 揭示假定安全的磷酸铁锂电池的热风险

IF 17 1区工程技术 Q1 ENERGY & FUELS

Etransportation

Pub Date : 2026-01-01 DOI: 10.1016/j.etran.2025.100531

Dian Zhang , Kai Chen , Xin Shen , Tao Wang , Yuan Ma , Yiren Zhong , Xuning Feng , Yuping Wu , Xin-Bing Cheng

LiFePO₄ batteries underpin global decarbonization efforts due to their intrinsic safety and low cost. However, emerging fire incidents in grid-scale storage demand reevaluation of their thermal stability. In contrast to the extensive focus on the stability P-O bond in PO₄, we reveal a previously overlooked gas-phase failure mechanism: delithiated LiFePO₄ cathodes undergo reductive decomposition (>600 °C) under H₂ generated during thermal runaway, forming FeP/Fe₂P and accelerating energy release. Crucially, this reaction is absent in inert atmospheres and intensifies with delithiation and battery capacity. By replacing carbonate electrolytes (H₂ source) and implementing ceramic separators (mimicking solid-state barriers), we suppress LiFePO₄ decomposition even at 800 °C. This work redefines LiFePO₄ safety paradigms, emphasizing large-system risks driven by gas-cathode interactions, and provides actionable strategies to enhance grid-storage resilience.

LiFePO4电池因其固有的安全性和低成本而支撑着全球的脱碳努力。然而，电网规模储能系统中出现的火灾事件要求对其热稳定性进行重新评估。与广泛关注PO4中P-O键的稳定性相反，我们揭示了一个以前被忽视的气相破坏机制：在热失控过程中产生的H2作用下，稀薄的LiFePO4阴极发生还原性分解（>600℃），形成FeP/Fe2P并加速能量释放。关键是，这种反应在惰性气氛中不存在，并随着电池容量的减少而加剧。通过取代碳酸盐电解质（H2源）和采用陶瓷分离器（模拟固态屏障），即使在800°C下，我们也能抑制LiFePO4的分解。这项工作重新定义了LiFePO4的安全范式，强调了气阴极相互作用驱动的大系统风险，并提供了增强电网存储弹性的可行策略。

{"title":"Unveiling thermal risks of presumed safe lithium iron phosphate batteries","authors":"Dian Zhang , Kai Chen , Xin Shen , Tao Wang , Yuan Ma , Yiren Zhong , Xuning Feng , Yuping Wu , Xin-Bing Cheng","doi":"10.1016/j.etran.2025.100531","DOIUrl":"10.1016/j.etran.2025.100531","url":null,"abstract":"<div><div>LiFePO<sub>4</sub> batteries underpin global decarbonization efforts due to their intrinsic safety and low cost. However, emerging fire incidents in grid-scale storage demand reevaluation of their thermal stability. In contrast to the extensive focus on the stability P-O bond in PO<sub>4</sub>, we reveal a previously overlooked gas-phase failure mechanism: delithiated LiFePO<sub>4</sub> cathodes undergo reductive decomposition (>600 °C) under H<sub>2</sub> generated during thermal runaway, forming FeP/Fe<sub>2</sub>P and accelerating energy release. Crucially, this reaction is absent in inert atmospheres and intensifies with delithiation and battery capacity. By replacing carbonate electrolytes (H<sub>2</sub> source) and implementing ceramic separators (mimicking solid-state barriers), we suppress LiFePO<sub>4</sub> decomposition even at 800 °C. This work redefines LiFePO<sub>4</sub> safety paradigms, emphasizing large-system risks driven by gas-cathode interactions, and provides actionable strategies to enhance grid-storage resilience.</div></div>","PeriodicalId":36355,"journal":{"name":"Etransportation","volume":"27 ","pages":"Article 100531"},"PeriodicalIF":17.0,"publicationDate":"2026-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145883814","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Passenger-aware reinforcement learning for efficient and robust energy management of fuel cell buses 基于乘客感知强化学习的燃料电池客车高效鲁棒能量管理

IF 17 1区工程技术 Q1 ENERGY & FUELS

Etransportation

Pub Date : 2026-01-01 DOI: 10.1016/j.etran.2025.100537

Chunchun Jia , Wei Liu , K.T. Chau , Hongwen He , Jiaming Zhou , Songyan Niu

Energy management strategies (EMSs) are essential for enhancing the efficiency, durability, and economic viability of fuel cell buses (FCBs). However, existing EMSs typically rely on fixed vehicle loads or idealized passenger assumptions, while neglecting the dynamic variations in passenger number and composition. This simplification introduces biased power demand distributions, underestimates the impact of human-occupancy heat loads under hot-weather conditions on air-conditioning system (ACS) energy use, and ultimately hinders the reproducibility of reported energy savings in real-world operation. To address these limitations, this study proposes a passenger-aware collaborative EMS aimed at enhancing the driving economy of FCBs under hot-weather conditions. Distinct from prior approaches, this study leverages a dual-source passenger perception framework that fuses video recognition with electronic card swiping data to obtain reliable real-time estimates of both passenger count and gender distribution. Gender-dependent body mass differences and heterogeneous metabolic heat generation are systematically integrated into the EMS framework, ensuring accurate modeling of passenger-induced variations in vehicle mass and cabin thermal load. Within this framework, the twin delayed deep deterministic policy gradient algorithm achieves the coordinated control of the fuel cell output power and the ACS cooling capacity. Extensive evaluations under real-world driving cycles and surveyed passenger datasets demonstrate the superiority of the proposed EMS. Compared with state-of-the-art baselines, the proposed method achieves at least a 0.62 % reduction in ACS energy consumption and a 2.11 % reduction in overall operational costs, without compromising cabin comfort. Importantly, in a representative scenario with 40 passengers, this method improves driving economy by 0.92–1.87 % over a gender-agnostic baseline at male passenger proportions of 0 %, 50 %, or 100 %, confirming the practical significance of incorporating passenger information. Given that urban buses operate continuously and costs scale near-linearly with energy and degradation, even modest percentage improvements over fleet-scale deployments and vehicle lifetimes can yield meaningful economic benefits.

能源管理策略（ems）对于提高燃料电池客车（fcb）的效率、耐久性和经济可行性至关重要。然而，现有的EMSs通常依赖于固定的车辆负载或理想化的乘客假设，而忽略了乘客数量和组成的动态变化。这种简化引入了有偏差的电力需求分布，低估了炎热天气条件下人类居住热负荷对空调系统（ACS）能源使用的影响，并最终阻碍了实际运行中报告的节能的可重复性。为了解决这些限制，本研究提出了一种乘客意识的协同EMS，旨在提高高温天气条件下fcb的驱动经济性。与之前的方法不同，本研究利用了一个双源乘客感知框架，将视频识别与电子刷卡数据融合在一起，以获得可靠的乘客数量和性别分布的实时估计。性别相关的身体质量差异和异质性代谢热产生被系统地整合到EMS框架中，确保准确建模乘客引起的车辆质量和客舱热负荷变化。在此框架下，双延迟深度确定性策略梯度算法实现了燃料电池输出功率与ACS制冷量的协调控制。在真实驾驶循环和调查乘客数据集下的广泛评估证明了所提出的EMS的优越性。与最先进的基线相比，所提出的方法在不影响客舱舒适度的情况下，至少减少了0.62%的ACS能耗和2.11%的总体运营成本。重要的是，在40名乘客的代表性场景中，该方法在男性乘客比例为0%、50%或100%的情况下，比性别不可知的基线提高了0.92 - 1.87%的驾驶经济性，证实了纳入乘客信息的实际意义。考虑到城市公交车持续运行，成本与能源和环境退化几乎成线性关系，即使是在车队规模部署和车辆使用寿命上的适度改进，也能产生有意义的经济效益。

{"title":"Passenger-aware reinforcement learning for efficient and robust energy management of fuel cell buses","authors":"Chunchun Jia , Wei Liu , K.T. Chau , Hongwen He , Jiaming Zhou , Songyan Niu","doi":"10.1016/j.etran.2025.100537","DOIUrl":"10.1016/j.etran.2025.100537","url":null,"abstract":"<div><div>Energy management strategies (EMSs) are essential for enhancing the efficiency, durability, and economic viability of fuel cell buses (FCBs). However, existing EMSs typically rely on fixed vehicle loads or idealized passenger assumptions, while neglecting the dynamic variations in passenger number and composition. This simplification introduces biased power demand distributions, underestimates the impact of human-occupancy heat loads under hot-weather conditions on air-conditioning system (ACS) energy use, and ultimately hinders the reproducibility of reported energy savings in real-world operation. To address these limitations, this study proposes a passenger-aware collaborative EMS aimed at enhancing the driving economy of FCBs under hot-weather conditions. Distinct from prior approaches, this study leverages a dual-source passenger perception framework that fuses video recognition with electronic card swiping data to obtain reliable real-time estimates of both passenger count and gender distribution. Gender-dependent body mass differences and heterogeneous metabolic heat generation are systematically integrated into the EMS framework, ensuring accurate modeling of passenger-induced variations in vehicle mass and cabin thermal load. Within this framework, the twin delayed deep deterministic policy gradient algorithm achieves the coordinated control of the fuel cell output power and the ACS cooling capacity. Extensive evaluations under real-world driving cycles and surveyed passenger datasets demonstrate the superiority of the proposed EMS. Compared with state-of-the-art baselines, the proposed method achieves at least a 0.62 % reduction in ACS energy consumption and a 2.11 % reduction in overall operational costs, without compromising cabin comfort. Importantly, in a representative scenario with 40 passengers, this method improves driving economy by 0.92–1.87 % over a gender-agnostic baseline at male passenger proportions of 0 %, 50 %, or 100 %, confirming the practical significance of incorporating passenger information. Given that urban buses operate continuously and costs scale near-linearly with energy and degradation, even modest percentage improvements over fleet-scale deployments and vehicle lifetimes can yield meaningful economic benefits.</div></div>","PeriodicalId":36355,"journal":{"name":"Etransportation","volume":"27 ","pages":"Article 100537"},"PeriodicalIF":17.0,"publicationDate":"2026-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145883818","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0