Journal of Loss Prevention in The Process Industries最新文献_第2页

Influence of cathode chemistry and state of charge on flammability and explosion parameters of lithium-ion battery vent gas 阴极化学和电荷状态对锂离子电池排气可燃性和爆炸参数的影响

IF 4.2 3区工程技术 Q2 ENGINEERING, CHEMICAL

Journal of Loss Prevention in The Process Industries

Pub Date : 2026-06-01 Epub Date: 2026-02-01 DOI: 10.1016/j.jlp.2026.105939

Domenico Enicchiaro, Almerinda Di Benedetto

The growing demand for energy storage systems highlights the need for safe and efficient lithium-ion batteries. A major safety concern is thermal runaway, which leads to the release of flammable battery vent gases (BVGs). This study systematically investigates the flammability characteristics of BVGs from LIBs with different cathode materials and state of charge levels through numerical simulations. Key combustion properties, including laminar flame speed, radical production and ignition delay time were evaluated under varying temperatures and equivalence ratios. Sensitivity analyses were also conducted to gain deeper insight into the roles of radical interactions in the combustion behaviours. Simulations were performed using CHEMKIN's Sandia PREMIX module for 1-D freely propagating flames and 0-D closed batch reactors using the San Diego mechanism. Results show that cathode chemistry significantly affects BVG flammability. NCA and NCM90505 produce vent gases with high H₂ and CO content, leading to higher S_u (65–70 cm/s) and shorter IDT. In contrast, NCM811 and NCM622 generate fewer flammable gases with lower S_u (35–45 cm/s) and longer IDT due to their higher concentrations of CO₂, CH₄, and C₂H₄. LFP, despite its high H₂ production, exhibits a balanced combustion profile. Laminar flame speed data for all BVG compositions are well described by a simple power-law correlation. SOC levels also strongly influence flammability. For SOC above 50 %, S_u, IDT, and radical production remain relatively stable, as vent gas composition changes minimally impact combustion properties. However, below 50 % SOC, S_u decreases, IDT increases, and radical production declines, with the strongest suppression at 0 % SOC. This effect is primarily due to the higher CO₂ concentration, which acts as a thermal sink, absorbing heat and slowing flame propagation rather than affecting combustion kinetics.

能源存储系统需求的增长凸显了对安全高效锂离子电池的需求。一个主要的安全问题是热失控，它会导致易燃电池排气气体（bvg）的释放。本研究通过数值模拟系统地研究了不同正极材料和电荷水平的锂离子电池的燃烧特性。在不同的温度和当量比下，评估了关键的燃烧性能，包括层流火焰速度、自由基产生和点火延迟时间。敏感性分析也进行了深入了解自由基相互作用在燃烧行为中的作用。使用CHEMKIN的Sandia预混模块对一维自由传播火焰和使用San Diego机制的0-D封闭批式反应器进行了模拟。结果表明，阴极化学对BVG的可燃性有显著影响。NCA和NCM90505产生高H2和CO含量的排气，导致更高的Su （65-70 cm/s）和更短的IDT。相比之下，NCM811和NCM622由于其较高的CO2、CH4和C2H4浓度，产生的可燃气体较少，Su较低（35-45 cm/s）， IDT较长。LFP虽然产氢量高，但其燃烧曲线平衡。所有BVG成分的层流火焰速度数据都可以用一个简单的幂律相关来描述。SOC水平也强烈影响可燃性。当SOC高于50%时，Su、IDT和自由基产量保持相对稳定，因为排气成分的变化对燃烧性能的影响最小。然而，在SOC低于50%时，Su减少，IDT增加，自由基产量下降，在SOC为0%时抑制最强。这种影响主要是由于较高的二氧化碳浓度，它作为一个热汇，吸收热量和减缓火焰传播，而不是影响燃烧动力学。

{"title":"Influence of cathode chemistry and state of charge on flammability and explosion parameters of lithium-ion battery vent gas","authors":"Domenico Enicchiaro, Almerinda Di Benedetto","doi":"10.1016/j.jlp.2026.105939","DOIUrl":"10.1016/j.jlp.2026.105939","url":null,"abstract":"<div><div>The growing demand for energy storage systems highlights the need for safe and efficient lithium-ion batteries. A major safety concern is thermal runaway, which leads to the release of flammable battery vent gases (BVGs). This study systematically investigates the flammability characteristics of BVGs from LIBs with different cathode materials and state of charge levels through numerical simulations. Key combustion properties, including laminar flame speed, radical production and ignition delay time were evaluated under varying temperatures and equivalence ratios. Sensitivity analyses were also conducted to gain deeper insight into the roles of radical interactions in the combustion behaviours. Simulations were performed using CHEMKIN's Sandia PREMIX module for 1-D freely propagating flames and 0-D closed batch reactors using the San Diego mechanism. Results show that cathode chemistry significantly affects BVG flammability. NCA and NCM90505 produce vent gases with high H<sub>2</sub> and CO content, leading to higher S<sub>u</sub> (65–70 cm/s) and shorter IDT. In contrast, NCM811 and NCM622 generate fewer flammable gases with lower S<sub>u</sub> (35–45 cm/s) and longer IDT due to their higher concentrations of CO<sub>2</sub>, CH<sub>4</sub>, and C<sub>2</sub>H<sub>4</sub>. LFP, despite its high H<sub>2</sub> production, exhibits a balanced combustion profile. Laminar flame speed data for all BVG compositions are well described by a simple power-law correlation. SOC levels also strongly influence flammability. For SOC above 50 %, S<sub>u</sub>, IDT, and radical production remain relatively stable, as vent gas composition changes minimally impact combustion properties. However, below 50 % SOC, S<sub>u</sub> decreases, IDT increases, and radical production declines, with the strongest suppression at 0 % SOC. This effect is primarily due to the higher CO<sub>2</sub> concentration, which acts as a thermal sink, absorbing heat and slowing flame propagation rather than affecting combustion kinetics.</div></div>","PeriodicalId":16291,"journal":{"name":"Journal of Loss Prevention in The Process Industries","volume":"101 ","pages":"Article 105939"},"PeriodicalIF":4.2,"publicationDate":"2026-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146191363","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"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

A dynamic decision framework for improving tank fire emergency decision-making using influence diagrams and micro-economics 运用影响图和微观经济学改进坦克火灾应急决策的动态决策框架

IF 4.2 3区工程技术 Q2 ENGINEERING, CHEMICAL

Journal of Loss Prevention in The Process Industries

Pub Date : 2026-06-01 Epub Date: 2026-02-02 DOI: 10.1016/j.jlp.2026.105954

Shile He , Jianfeng Zhou , Genserik Reniers

This study proposes a dynamic decision-making framework for fire emergency management in hazardous chemical storage tank areas, based on Bayesian networks and influence diagrams. The framework comprises two core components: the generation of initial response strategies and the evaluation of them. By analyzing the thermal radiation coupling effect between tanks, this study introduces the closeness centrality metric to identify key tank nodes, for which multiple emergency response strategies are further formulated. To evaluate these emergency response strategies, this study employs influence diagrams to evaluate the utility of each strategy and proposes a simplified utility function. The study further evaluates the impact of domino effects by dynamically updating the influence diagram, recalculating the closeness centrality, and adjusting the key nodes based on the assumption that a fire spreads to adjacent tanks. A case study illustrates that, across a range of firefighting and cooling efficiency assumptions, the framework successfully adapts to changing scenarios, optimizing strategies to improve the timeliness and accuracy of emergency responses. This provides a viable optimization method for complex tank fire emergency decision-making.

本研究提出一种基于贝叶斯网络和影响图的危险化学品储罐区火灾应急管理动态决策框架。该框架包括两个核心部分：制定初步应对战略和对其进行评价。通过分析储罐之间的热辐射耦合效应，引入紧密度中心性度量来识别储罐关键节点，并进一步制定多种应急响应策略。为了评估这些应急响应策略，本研究采用影响图来评估每种策略的效用，并提出了一个简化的效用函数。该研究通过动态更新影响图，重新计算接近中心性，并根据火势蔓延到相邻油箱的假设调整关键节点，进一步评估多米诺效应的影响。一个案例研究表明，在一系列消防和冷却效率假设中，该框架成功地适应了不断变化的情景，优化了策略，提高了应急响应的及时性和准确性。为复杂坦克火灾应急决策提供了一种可行的优化方法。

{"title":"A dynamic decision framework for improving tank fire emergency decision-making using influence diagrams and micro-economics","authors":"Shile He , Jianfeng Zhou , Genserik Reniers","doi":"10.1016/j.jlp.2026.105954","DOIUrl":"10.1016/j.jlp.2026.105954","url":null,"abstract":"<div><div>This study proposes a dynamic decision-making framework for fire emergency management in hazardous chemical storage tank areas, based on Bayesian networks and influence diagrams. The framework comprises two core components: the generation of initial response strategies and the evaluation of them. By analyzing the thermal radiation coupling effect between tanks, this study introduces the closeness centrality metric to identify key tank nodes, for which multiple emergency response strategies are further formulated. To evaluate these emergency response strategies, this study employs influence diagrams to evaluate the utility of each strategy and proposes a simplified utility function. The study further evaluates the impact of domino effects by dynamically updating the influence diagram, recalculating the closeness centrality, and adjusting the key nodes based on the assumption that a fire spreads to adjacent tanks. A case study illustrates that, across a range of firefighting and cooling efficiency assumptions, the framework successfully adapts to changing scenarios, optimizing strategies to improve the timeliness and accuracy of emergency responses. This provides a viable optimization method for complex tank fire emergency decision-making.</div></div>","PeriodicalId":16291,"journal":{"name":"Journal of Loss Prevention in The Process Industries","volume":"101 ","pages":"Article 105954"},"PeriodicalIF":4.2,"publicationDate":"2026-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146191259","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"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

Study on the effect of safety barriers on the diffusion behavior of crude oil in flood 安全屏障对原油在洪水中扩散行为的影响研究

IF 4.2 3区工程技术 Q2 ENGINEERING, CHEMICAL

Journal of Loss Prevention in The Process Industries

Pub Date : 2026-06-01 Epub Date: 2026-02-03 DOI: 10.1016/j.jlp.2026.105941

Hao Sun , Fatong Gong , Shengzhu Zhang , Rongxue Kang , Ruipeng Tong , Zongzhi Wu

As climate change intensifies, the impact of natural disasters on chemical facilities and infrastructure — along with the resulting threat of hazardous material leaks — has been recognized as an emerging risk. Flood-induced Natech events often result in the release of hazardous substances, which can lead to fires and explosions during dispersion. This study focuses on crude oil storage tanks to investigate the patterns of hazardous chemical leakage and dispersion in flood-induced Natech events, employing a scaled-down model method to establish a tank farm model. Crude oil diffusion similarity experiments were designed based on the Froude similarity criterion. Through experimental observations, we qualitatively described the diffusion patterns of crude oil under different flood conditions. The dispersion process was categorized into four phases: initial leakage, rapid overflow, dynamic equilibrium, and leakage termination. Based on the flow state of the crude oil, each phase was subdivided into specific zones with qualitative descriptions of the flow characteristics in each zone. This work helps emergency responders identify the phase of crude oil dispersion during Natech events and develop effective emergency response plans.

随着气候变化的加剧，自然灾害对化学设施和基础设施的影响——以及由此产生的危险物质泄漏的威胁——已被认为是一个新兴的风险。洪水引发的Natech事件通常会导致有害物质的释放，在扩散过程中可能导致火灾和爆炸。本研究以原油储罐为研究对象，采用缩尺模型方法建立了储罐库模型，探讨了洪水引发的Natech事件中危险化学品泄漏和扩散的规律。基于Froude相似准则设计了原油扩散相似实验。通过实验观察，定性描述了不同洪水条件下原油的扩散规律。将扩散过程分为初始泄漏、快速溢流、动态平衡和泄漏终止四个阶段。根据原油的流动状态，将每个相细分为特定的区域，并对每个区域的流动特性进行定性描述。这项工作有助于应急响应人员确定Natech事件期间原油分散的阶段，并制定有效的应急响应计划。

{"title":"Study on the effect of safety barriers on the diffusion behavior of crude oil in flood","authors":"Hao Sun , Fatong Gong , Shengzhu Zhang , Rongxue Kang , Ruipeng Tong , Zongzhi Wu","doi":"10.1016/j.jlp.2026.105941","DOIUrl":"10.1016/j.jlp.2026.105941","url":null,"abstract":"<div><div>As climate change intensifies, the impact of natural disasters on chemical facilities and infrastructure — along with the resulting threat of hazardous material leaks — has been recognized as an emerging risk. Flood-induced Natech events often result in the release of hazardous substances, which can lead to fires and explosions during dispersion. This study focuses on crude oil storage tanks to investigate the patterns of hazardous chemical leakage and dispersion in flood-induced Natech events, employing a scaled-down model method to establish a tank farm model. Crude oil diffusion similarity experiments were designed based on the Froude similarity criterion. Through experimental observations, we qualitatively described the diffusion patterns of crude oil under different flood conditions. The dispersion process was categorized into four phases: initial leakage, rapid overflow, dynamic equilibrium, and leakage termination. Based on the flow state of the crude oil, each phase was subdivided into specific zones with qualitative descriptions of the flow characteristics in each zone. This work helps emergency responders identify the phase of crude oil dispersion during Natech events and develop effective emergency response plans.</div></div>","PeriodicalId":16291,"journal":{"name":"Journal of Loss Prevention in The Process Industries","volume":"101 ","pages":"Article 105941"},"PeriodicalIF":4.2,"publicationDate":"2026-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146191255","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"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

Research on accident evolution analysis of hazardous gas leak based on graph neural networks 基于图神经网络的危险气体泄漏事故演化分析研究

IF 4.2 3区工程技术 Q2 ENGINEERING, CHEMICAL

Journal of Loss Prevention in The Process Industries

Pub Date : 2026-06-01 Epub Date: 2026-01-30 DOI: 10.1016/j.jlp.2026.105938

Xinqi Zhang , Yujie Lin , Shubo Yang , Yi Liu , Anfeng Yu

Understanding the propagation underlying leakage risk evolution is essential for enhancing emergency response in process safety management. However, traditional risk evolution modeling approaches remain rely on expert judgment, which introduces analytical uncertainties and inherent temporal delays in capturing accident progression dynamics, thereby undermining the efficacy of emergency response interventions. To address these limitations, this study proposes a novel accident evolution analysis method named GraphSAGE-DEMATEL-ISM. This hybrid model integrates knowledge graph inference mechanisms with deep learning architectures to achieve autonomous discovery of risk factor interdependencies. Specifically, the framework employs Graph Sample and Aggregate neural networks (GraphSAGE) to systematically extract and model complex interaction relationships among risk factors. Subsequently, the Decision-Making Trial and Evaluation Laboratory-Interpretive Structural Modeling (DEMATEL-ISM) is applied to elucidate the hierarchical structure and identify critical factors governing risk evolution dynamics. The framework's efficacy is validated through evolution analysis of a real gas leakage explosion, providing direct insights into the critical risk factors and propagation structures. The proposed method demonstrates significant advantages through its capacity for efficient risk evolution analysis without dependence on expert. The integration of graph neural networks with risk evolution analysis method offers a paradigm shift toward proactive risk management strategies in process industries.

了解泄漏风险演变的传播过程对于提高过程安全管理中的应急响应能力至关重要。然而，传统的风险演化建模方法仍然依赖于专家判断，这在捕获事故进展动态时引入了分析不确定性和固有的时间延迟，从而削弱了应急响应干预的有效性。为了解决这些局限性，本研究提出了一种新的事故演变分析方法GraphSAGE-DEMATEL-ISM。该混合模型将知识图推理机制与深度学习体系结构相结合，实现了风险因素相互依赖关系的自主发现。具体而言，该框架采用图样本和聚合神经网络（GraphSAGE）系统地提取和建模风险因素之间复杂的相互作用关系。随后，应用决策试验与评估实验室-解释结构模型（DEMATEL-ISM）来阐明层次结构，识别控制风险演化动态的关键因素。通过对真实气体泄漏爆炸的演化分析，验证了该框架的有效性，从而直接了解了关键风险因素和传播结构。该方法能够在不依赖专家的情况下进行有效的风险演化分析，具有显著的优势。图神经网络与风险演化分析方法的集成为过程工业的主动风险管理策略提供了一种范式转换。

{"title":"Research on accident evolution analysis of hazardous gas leak based on graph neural networks","authors":"Xinqi Zhang , Yujie Lin , Shubo Yang , Yi Liu , Anfeng Yu","doi":"10.1016/j.jlp.2026.105938","DOIUrl":"10.1016/j.jlp.2026.105938","url":null,"abstract":"<div><div>Understanding the propagation underlying leakage risk evolution is essential for enhancing emergency response in process safety management. However, traditional risk evolution modeling approaches remain rely on expert judgment, which introduces analytical uncertainties and inherent temporal delays in capturing accident progression dynamics, thereby undermining the efficacy of emergency response interventions. To address these limitations, this study proposes a novel accident evolution analysis method named GraphSAGE-DEMATEL-ISM. This hybrid model integrates knowledge graph inference mechanisms with deep learning architectures to achieve autonomous discovery of risk factor interdependencies. Specifically, the framework employs Graph Sample and Aggregate neural networks (GraphSAGE) to systematically extract and model complex interaction relationships among risk factors. Subsequently, the Decision-Making Trial and Evaluation Laboratory-Interpretive Structural Modeling (DEMATEL-ISM) is applied to elucidate the hierarchical structure and identify critical factors governing risk evolution dynamics. The framework's efficacy is validated through evolution analysis of a real gas leakage explosion, providing direct insights into the critical risk factors and propagation structures. The proposed method demonstrates significant advantages through its capacity for efficient risk evolution analysis without dependence on expert. The integration of graph neural networks with risk evolution analysis method offers a paradigm shift toward proactive risk management strategies in process industries.</div></div>","PeriodicalId":16291,"journal":{"name":"Journal of Loss Prevention in The Process Industries","volume":"101 ","pages":"Article 105938"},"PeriodicalIF":4.2,"publicationDate":"2026-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146191371","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"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

Fault detection method for complex chemical processes based on MSCNN-DRSN-Transformer 基于MSCNN-DRSN-Transformer的复杂化工过程故障检测方法

IF 4.2 3区工程技术 Q2 ENGINEERING, CHEMICAL

Journal of Loss Prevention in The Process Industries

Pub Date : 2026-06-01 Epub Date: 2026-02-06 DOI: 10.1016/j.jlp.2026.105962

Fubo Zhang, Qibin Huang, Haifan Liao, Yisheng Liu, Aihua Liu

Addressing the key challenges of strong nonlinearity, multi-scale dynamic characteristics, and noise interference in fault detection of complex chemical processes, this paper proposes a hybrid fault detection model based on MSCNN-DRSN-Transformer. The model is built upon a “local feature perception – adaptive noise suppression – global coupling modeling” framework, enabling effective fusion of multi-scale features and long-range temporal dependencies. Initially, a Multi-Scale Convolutional Neural Network (MSCNN) extracts local features across different time scales in parallel. Subsequently, a Deep Residual Shrinkage Network (DRSN) enhances the feature signal-to-noise ratio via an adaptive soft-thresholding mechanism. Finally, a Transformer encoder captures global temporal dependencies. Experimental validation on the Tennessee Eastman (TE) standard process dataset demonstrates that the proposed model achieves an average fault detection rate (FDR) of 99.12% and an average false alarm rate (FAR) of only 0.19% across six typical fault types. Compared with state-of-the-art methods such as DPCA, OSCAE-CNN, and MFA-Transformer, it improves the average FDR by 13.91%, 1.72%, and 1.75%, respectively. These results verify the model's superior performance and engineering applicability for fault detection in complex chemical processes, offering an effective technical solution for intelligent monitoring in the process industries.

针对复杂化工过程故障检测中存在的强非线性、多尺度动态特性和噪声干扰等关键问题，提出了一种基于MSCNN-DRSN-Transformer的混合故障检测模型。该模型建立在“局部特征感知-自适应噪声抑制-全局耦合建模”框架之上，能够有效融合多尺度特征和长时间依赖关系。首先，多尺度卷积神经网络（MSCNN）在不同时间尺度上并行提取局部特征。随后，深度残余收缩网络（DRSN）通过自适应软阈值机制提高特征信噪比。最后，Transformer编码器捕获全局时间依赖性。在Tennessee Eastman （TE）标准过程数据集上的实验验证表明，该模型在6种典型故障类型上的平均故障检测率（FDR）为99.12%，平均虚警率（FAR）仅为0.19%。与DPCA、OSCAE-CNN和MFA-Transformer等最先进的方法相比，该方法的平均FDR分别提高了13.91%、1.72%和1.75%。这些结果验证了该模型在复杂化工过程故障检测中的优越性能和工程适用性，为过程工业智能监控提供了有效的技术解决方案。

{"title":"Fault detection method for complex chemical processes based on MSCNN-DRSN-Transformer","authors":"Fubo Zhang, Qibin Huang, Haifan Liao, Yisheng Liu, Aihua Liu","doi":"10.1016/j.jlp.2026.105962","DOIUrl":"10.1016/j.jlp.2026.105962","url":null,"abstract":"<div><div>Addressing the key challenges of strong nonlinearity, multi-scale dynamic characteristics, and noise interference in fault detection of complex chemical processes, this paper proposes a hybrid fault detection model based on MSCNN-DRSN-Transformer. The model is built upon a “local feature perception – adaptive noise suppression – global coupling modeling” framework, enabling effective fusion of multi-scale features and long-range temporal dependencies. Initially, a Multi-Scale Convolutional Neural Network (MSCNN) extracts local features across different time scales in parallel. Subsequently, a Deep Residual Shrinkage Network (DRSN) enhances the feature signal-to-noise ratio via an adaptive soft-thresholding mechanism. Finally, a Transformer encoder captures global temporal dependencies. Experimental validation on the Tennessee Eastman (TE) standard process dataset demonstrates that the proposed model achieves an average fault detection rate (FDR) of 99.12% and an average false alarm rate (FAR) of only 0.19% across six typical fault types. Compared with state-of-the-art methods such as DPCA, OSCAE-CNN, and MFA-Transformer, it improves the average FDR by 13.91%, 1.72%, and 1.75%, respectively. These results verify the model's superior performance and engineering applicability for fault detection in complex chemical processes, offering an effective technical solution for intelligent monitoring in the process industries.</div></div>","PeriodicalId":16291,"journal":{"name":"Journal of Loss Prevention in The Process Industries","volume":"101 ","pages":"Article 105962"},"PeriodicalIF":4.2,"publicationDate":"2026-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146191367","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"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

Assessing the degree of pyrophoricity for gaseous silanes 评价气态硅烷的致焰性程度

IF 4.2 3区工程技术 Q2 ENGINEERING, CHEMICAL

Journal of Loss Prevention in The Process Industries

Pub Date : 2026-06-01 Epub Date: 2026-02-02 DOI: 10.1016/j.jlp.2026.105935

Trung Thanh Nguyen , Hsiao-Yun Tsai , Jenq-Renn Chen , Eugene Y. Ngai

A pyrophoric gas is defined as a flammable gas having an autoignition temperature (AIT) below 54 °C. However, determining AIT below ambient temperature is difficult. On the other hand, pyrophoric gases, such as silane, when releasing steadily into ambient air with a velocity above a critical flow velocity may lead to delayed ignition in which the ignition is delayed indefinitely. The V_c is found to be a function of the vent size (D). A simple critical shear rate, defined as 8 V_c/D, is considered a potential useful indicator for assessing the relative degree of pyrophoricity. Monochlorosilane (SiH₃Cl), a pyrophoric gas with reported AIT of <20 °C is used as an example and results of steady release tests are compared with those of silane and disilane. The results are surprising. Degree of pyrophoricity for the three gases are: disilane > silane > monochlorosilane. Lower degree of pyrophoricity implies lower reactivity towards air and thus delayed ignitions and vapor cloud explosions are favored. The critical shear rate required to quench the autoignition kernel will be a useful indicator for assessing the hazards of pyrophoric gases.

可燃气体是指自燃温度（AIT）低于54℃的可燃气体。然而，确定低于环境温度的AIT是困难的。另一方面，焦性气体，如硅烷，当以高于临界流速的速度稳定释放到环境空气中时，可能导致延迟点火，即无限期延迟点火。Vc是排气口大小(D)的函数。一个简单的临界剪切速率，定义为8 Vc/D，被认为是评估相对焦性程度的潜在有用指标。以一氯硅烷（SiH3Cl）为例，将其稳定释放试验结果与硅烷和二硅烷的稳定释放试验结果进行了比较。结果令人惊讶。三种气体的发光度分别为：二硅烷；硅烷；单氯硅烷。较低的致火性意味着对空气的反应性较低，因此有利于延迟点火和蒸气云爆炸。扑灭自燃核所需的临界剪切速率将是评估燃烧性气体危害的有用指标。

{"title":"Assessing the degree of pyrophoricity for gaseous silanes","authors":"Trung Thanh Nguyen , Hsiao-Yun Tsai , Jenq-Renn Chen , Eugene Y. Ngai","doi":"10.1016/j.jlp.2026.105935","DOIUrl":"10.1016/j.jlp.2026.105935","url":null,"abstract":"<div><div>A pyrophoric gas is defined as a flammable gas having an autoignition temperature (AIT) below 54 °C. However, determining AIT below ambient temperature is difficult. On the other hand, pyrophoric gases, such as silane, when releasing steadily into ambient air with a velocity above a critical flow velocity may lead to delayed ignition in which the ignition is delayed indefinitely. The <em>V</em><sub><em>c</em></sub> is found to be a function of the vent size (D). A simple critical shear rate, defined as 8 <em>V</em><sub><em>c</em></sub>/D, is considered a potential useful indicator for assessing the relative degree of pyrophoricity. Monochlorosilane (SiH<sub>3</sub>Cl), a pyrophoric gas with reported AIT of <20 °C is used as an example and results of steady release tests are compared with those of silane and disilane. The results are surprising. Degree of pyrophoricity for the three gases are: disilane > silane > monochlorosilane. Lower degree of pyrophoricity implies lower reactivity towards air and thus delayed ignitions and vapor cloud explosions are favored. The critical shear rate required to quench the autoignition kernel will be a useful indicator for assessing the hazards of pyrophoric gases.</div></div>","PeriodicalId":16291,"journal":{"name":"Journal of Loss Prevention in The Process Industries","volume":"101 ","pages":"Article 105935"},"PeriodicalIF":4.2,"publicationDate":"2026-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146191364","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"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

Gas leakage and diffusion patterns in buried natural gas pipelines: Effect of soil properties and burial depth 埋地天然气管道中气体泄漏和扩散模式：土壤性质和埋深的影响

IF 4.2 3区工程技术 Q2 ENGINEERING, CHEMICAL

Journal of Loss Prevention in The Process Industries

Pub Date : 2026-06-01 Epub Date: 2026-01-21 DOI: 10.1016/j.jlp.2026.105930

Zeyuan Ding , Kun Chen , Hongyuan Li , Hongfu Mi , Chi-Min Shu

Buried natural gas pipelines, as the primary mode of gas transportation, have exhibited pronounced risks due to complex soil environments that hinder accurate prediction of gas leakage diffusion patterns, posing severe threats to life and property. This study employed COMSOL Multiphysics to create a three-dimensional numerical model, systematically investigating the combined effects of soil porosity (0.2–0.6), moisture content (0.01–0.6), permeability (0.5–50 Darcy, 1 Darcy = 10⁻¹² m²). Furthermore, pipeline burial depth (0.3–3 m) affects gas leakage dynamics. Key findings revealed that under low moisture conditions (1% water content), elevated soil porosity accelerates vertical gas migration by 35%–48%, enabling expeditious surface accumulation with methane concentrations exceeding 15% LEL (lower explosive limit). Conversely, at typical moisture levels (20% water content), porosity variations showed a negligible impact on gas distribution. Soil moisture emerges as a dominant inhibitory factor: Increasing moisture from 0.05 to 0.6 lessened high-concentration zones (≥5% methane) by 40%–62% through improved capillary resistance. Permeability escalation amplifies hazardous boundaries exponentially, with 50D permeability scenarios showing a 2.5–fold expansion compared with 0.5D cases. Shallow burial (0.3–1 m) prioritises vertical diffusion, elevating surface concentrations to 8%–12% LEL within 100 min, while deeper burial (>2 m) redirects 70%–85% of gas laterally, creating expansive subsurface plumes (>4 m radius) with delayed surface arrival (>300 min). By integrating multi-physics simulations, this study clarified the mechanistic interactions between soil parameters and gas leakage behaviour, offering scientific insights for optimising leak detection, risk assessment, and emergency management in buried pipelines. These findings rendered vital engineering guidance for ameliorating pipeline loss prevention and mitigating environmental hazards.

埋地天然气管道作为天然气输送的主要方式，由于土壤环境复杂，难以准确预测天然气泄漏扩散模式，存在明显的风险，对生命财产造成严重威胁。本研究利用COMSOL Multiphysics软件建立三维数值模型，系统研究了土壤孔隙度（0.2-0.6）、含水率（0.01-0.6）、渗透率（0.5-50 Darcy, 1 Darcy = 10−12 m2）的综合效应。管道埋深（0.3-3 m）影响气体泄漏动态。主要研究结果表明，在低水分条件下（含水率为1%），土壤孔隙度升高会加速35%-48%的垂直气体运移，使甲烷浓度超过15% LEL（爆炸下限）时能够迅速在地表聚集。相反，在典型水分水平（20%含水量）下，孔隙度变化对气体分布的影响可以忽略不计。土壤湿度是主要的抑制因素：土壤湿度从0.05增加到0.6，通过提高毛细阻力，高浓度区（≥5%甲烷）减少40%-62%。渗透率增加会成倍地放大危险边界，在渗透率为50D的情况下，与渗透率为0.5D的情况相比，危险边界扩大了2.5倍。浅埋层（0.3-1米）优先进行垂直扩散，在100分钟内将地表浓度提升至8%-12% LEL，而深埋层（2米）将70%-85%的天然气转向横向，形成膨胀的地下羽流（半径为4米），延迟到达地表（300分钟）。通过整合多物理场模拟，本研究阐明了土壤参数与天然气泄漏行为之间的机制相互作用，为优化地埋管道的泄漏检测、风险评估和应急管理提供了科学见解。这些发现为改善管道损失预防和减轻环境危害提供了重要的工程指导。

{"title":"Gas leakage and diffusion patterns in buried natural gas pipelines: Effect of soil properties and burial depth","authors":"Zeyuan Ding , Kun Chen , Hongyuan Li , Hongfu Mi , Chi-Min Shu","doi":"10.1016/j.jlp.2026.105930","DOIUrl":"10.1016/j.jlp.2026.105930","url":null,"abstract":"<div><div>Buried natural gas pipelines, as the primary mode of gas transportation, have exhibited pronounced risks due to complex soil environments that hinder accurate prediction of gas leakage diffusion patterns, posing severe threats to life and property. This study employed COMSOL Multiphysics to create a three-dimensional numerical model, systematically investigating the combined effects of soil porosity (0.2–0.6), moisture content (0.01–0.6), permeability (0.5–50 Darcy, 1 Darcy = 10<sup>−12</sup> m<sup>2</sup>). Furthermore, pipeline burial depth (0.3–3 m) affects gas leakage dynamics. Key findings revealed that under low moisture conditions (1% water content), elevated soil porosity accelerates vertical gas migration by 35%–48%, enabling expeditious surface accumulation with methane concentrations exceeding 15% LEL (lower explosive limit). Conversely, at typical moisture levels (20% water content), porosity variations showed a negligible impact on gas distribution. Soil moisture emerges as a dominant inhibitory factor: Increasing moisture from 0.05 to 0.6 lessened high-concentration zones (≥5% methane) by 40%–62% through improved capillary resistance. Permeability escalation amplifies hazardous boundaries exponentially, with 50D permeability scenarios showing a 2.5–fold expansion compared with 0.5D cases. Shallow burial (0.3–1 m) prioritises vertical diffusion, elevating surface concentrations to 8%–12% LEL within 100 min, while deeper burial (>2 m) redirects 70%–85% of gas laterally, creating expansive subsurface plumes (>4 m radius) with delayed surface arrival (>300 min). By integrating multi-physics simulations, this study clarified the mechanistic interactions between soil parameters and gas leakage behaviour, offering scientific insights for optimising leak detection, risk assessment, and emergency management in buried pipelines. These findings rendered vital engineering guidance for ameliorating pipeline loss prevention and mitigating environmental hazards.</div></div>","PeriodicalId":16291,"journal":{"name":"Journal of Loss Prevention in The Process Industries","volume":"101 ","pages":"Article 105930"},"PeriodicalIF":4.2,"publicationDate":"2026-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146081617","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"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

Reliable dynamic causality analysis for efficient prescriptive maintenance of degraded industrial equipment 可靠的动态因果分析，为有效的规范维护退化工业设备

IF 4.2 3区工程技术 Q2 ENGINEERING, CHEMICAL

Journal of Loss Prevention in The Process Industries

Pub Date : 2026-06-01 Epub Date: 2026-01-20 DOI: 10.1016/j.jlp.2026.105922

Karim Nadim , Ahmed Ragab , Hakim Ghezzaz , Mohamed-Salah Ouali

Developing efficient prescriptive maintenance strategies is essential in mitigating the performance degradation of equipment in energy-intensive process industries. This paper proposes an innovative approach that integrates data-driven causality analysis and reinforcement learning (RL) to reduce/slow-down the equipment degradation. The degradation is first modeled by constructing representative health indicators (HIs) using deep learning autoencoders. The HIs are then exploited using clustering and interpretable machine learning techniques to identify the degradation's root causes. Afterward, a dynamic causal model is discovered in the form of a Petri net (PN) using process mining techniques. The causal model incorporates the temporal information and sequential relationships between the identified root causes. Finally, an RL agent is integrated with the PN model to recommend the optimal sequence of events that diminishes the performance degradation rate. The proposed approach is tested successfully on a complex case study of a black liquor concentrator in a Kraft pulp mill that is subjected to a decline in operational performance and capacity due to the fouling degradation phenomenon. The obtained results show that the fouling rate was minimized, leading to an approximate 30 % saving in maintenance costs and a reduction of around 4.8 kt CO2/year in greenhouse gas emissions.

在能源密集型加工工业中，制定有效的规范维护战略对于减轻设备性能退化至关重要。本文提出了一种集成数据驱动的因果关系分析和强化学习（RL）的创新方法，以减少/减缓设备退化。首先通过使用深度学习自编码器构建具有代表性的健康指标（HIs）来对退化进行建模。然后利用聚类和可解释的机器学习技术来识别退化的根本原因。然后，使用过程挖掘技术以Petri网（PN）的形式发现了动态因果模型。因果模型结合了时间信息和确定的根本原因之间的顺序关系。最后，RL代理与PN模型集成，以推荐减少性能退化率的最佳事件序列。该方法在卡夫纸浆厂黑液浓缩厂的复杂案例研究中得到了成功的验证，该浓缩厂由于污染降解现象而导致运行性能和容量下降。结果表明，污垢率最低，导致维护成本节省约30%，温室气体排放量减少约4.8 kt CO2/年。

{"title":"Reliable dynamic causality analysis for efficient prescriptive maintenance of degraded industrial equipment","authors":"Karim Nadim , Ahmed Ragab , Hakim Ghezzaz , Mohamed-Salah Ouali","doi":"10.1016/j.jlp.2026.105922","DOIUrl":"10.1016/j.jlp.2026.105922","url":null,"abstract":"<div><div>Developing efficient prescriptive maintenance strategies is essential in mitigating the performance degradation of equipment in energy-intensive process industries. This paper proposes an innovative approach that integrates data-driven causality analysis and reinforcement learning (RL) to reduce/slow-down the equipment degradation. The degradation is first modeled by constructing representative health indicators (HIs) using deep learning autoencoders. The HIs are then exploited using clustering and interpretable machine learning techniques to identify the degradation's root causes. Afterward, a dynamic causal model is discovered in the form of a Petri net (PN) using process mining techniques. The causal model incorporates the temporal information and sequential relationships between the identified root causes. Finally, an RL agent is integrated with the PN model to recommend the optimal sequence of events that diminishes the performance degradation rate. The proposed approach is tested successfully on a complex case study of a black liquor concentrator in a Kraft pulp mill that is subjected to a decline in operational performance and capacity due to the fouling degradation phenomenon. The obtained results show that the fouling rate was minimized, leading to an approximate 30 % saving in maintenance costs and a reduction of around 4.8 kt CO2/year in greenhouse gas emissions.</div></div>","PeriodicalId":16291,"journal":{"name":"Journal of Loss Prevention in The Process Industries","volume":"101 ","pages":"Article 105922"},"PeriodicalIF":4.2,"publicationDate":"2026-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146191260","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"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

Cryogenic storage safety: Experimental evaluation of insulation under extreme conditions 低温储存安全性：极端条件下绝缘的实验评价

IF 4.2 3区工程技术 Q2 ENGINEERING, CHEMICAL

Journal of Loss Prevention in The Process Industries

Pub Date : 2026-06-01 Epub Date: 2026-02-06 DOI: 10.1016/j.jlp.2026.105961

Aliasghar Hajhariri , Robert Eberwein , Davide Camplese , Giordano Emrys Scarponi , Valerio Cozzani , Holger Seidlitz

Hydrogen is recognized as a keystone of the global energy transition, offering a clean, high-energy-density energy carrier ideal for storage and transportation. Among various storage options, liquid hydrogen (LH2) is especially advantageous for both mobile and stationary applications. However, ensuring the safety and performance of LH2 storage systems under extreme thermal conditions, such as fire exposure, remains an engineering challenge.

This study introduces an experimental framework, called the Cryogenic High-Temperature Thermal Vacuum Chamber (CHTTVC), designed to investigate the thermal-hydraulic response of vacuum-insulated cryogenic tanks under fire-like conditions. The apparatus enables evaluation of insulation performance, such as perlite and multilayer insulation (MLI), with a focus on thermal degradation, heat ingress, and vacuum stability.

Results indicate that combustible MLIs undergo substantial thermal degradation, leading to heat ingress rates of up to 6.5 kW and the formation of hazardous combustion by-products. In contrast, non-combustible MLIs and bulk insulation materials restrict heat ingress to approximately 3 kW while more effectively preserving vacuum integrity. Combustible MLIs also exhibit pronounced pressure increases in the evacuated section, reaching ∼6 × 10⁴ Pa, nearly six times higher than those observed for non-combustible counterparts. Analysis of effective emissivity further reveals an enhancement in radiative heat transfer, approximately five times, for combustible MLIs following degradation. Additionally, marked thermal stratification develops under both nominal and extreme heat loads, with temperature gradients approaching 10 °C per 100 mm during sustained thermal exposure.

氢被认为是全球能源转型的基石，提供了一种清洁、高能量密度的能源载体，是储存和运输的理想选择。在各种存储选择中，液氢（LH2）对于移动和固定应用都特别有利。然而，确保LH2存储系统在极端热条件下（如火灾暴露）的安全性和性能仍然是一个工程挑战。本研究引入了一个实验框架，称为低温高温热真空室（CHTTVC），旨在研究在类似火的条件下真空绝热低温罐的热-液压响应。该设备能够评估绝缘性能，例如珍珠岩和多层绝缘（MLI），重点是热降解，热进入和真空稳定性。结果表明，可燃性mli经历了大量的热降解，导致热进入率高达6.5 kW，并形成有害的燃烧副产物。相比之下，不可燃的mli和大块绝缘材料将热进入限制在约3kw，同时更有效地保持真空完整性。可燃性mli在抽真空段也表现出明显的压力增加，达到~ 6 × 104 Pa，几乎是非可燃性mli的6倍。对有效发射率的分析进一步表明，可燃性mli在降解后的辐射传热增强了约5倍。此外，在名义热负荷和极端热负荷下，明显的热分层都会发生，在持续热暴露期间，温度梯度接近每100毫米10°C。

{"title":"Cryogenic storage safety: Experimental evaluation of insulation under extreme conditions","authors":"Aliasghar Hajhariri , Robert Eberwein , Davide Camplese , Giordano Emrys Scarponi , Valerio Cozzani , Holger Seidlitz","doi":"10.1016/j.jlp.2026.105961","DOIUrl":"10.1016/j.jlp.2026.105961","url":null,"abstract":"<div><div>Hydrogen is recognized as a keystone of the global energy transition, offering a clean, high-energy-density energy carrier ideal for storage and transportation. Among various storage options, liquid hydrogen (LH2) is especially advantageous for both mobile and stationary applications. However, ensuring the safety and performance of LH2 storage systems under extreme thermal conditions, such as fire exposure, remains an engineering challenge.</div><div>This study introduces an experimental framework, called the Cryogenic High-Temperature Thermal Vacuum Chamber (CHTTVC), designed to investigate the thermal-hydraulic response of vacuum-insulated cryogenic tanks under fire-like conditions. The apparatus enables evaluation of insulation performance, such as perlite and multilayer insulation (MLI), with a focus on thermal degradation, heat ingress, and vacuum stability.</div><div>Results indicate that combustible MLIs undergo substantial thermal degradation, leading to heat ingress rates of up to 6.5 kW and the formation of hazardous combustion by-products. In contrast, non-combustible MLIs and bulk insulation materials restrict heat ingress to approximately 3 kW while more effectively preserving vacuum integrity. Combustible MLIs also exhibit pronounced pressure increases in the evacuated section, reaching ∼6 × 10<sup>4</sup> Pa, nearly six times higher than those observed for non-combustible counterparts. Analysis of effective emissivity further reveals an enhancement in radiative heat transfer, approximately five times, for combustible MLIs following degradation. Additionally, marked thermal stratification develops under both nominal and extreme heat loads, with temperature gradients approaching 10 °C per 100 mm during sustained thermal exposure.</div></div>","PeriodicalId":16291,"journal":{"name":"Journal of Loss Prevention in The Process Industries","volume":"101 ","pages":"Article 105961"},"PeriodicalIF":4.2,"publicationDate":"2026-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146191365","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"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

A comprehensive review of AI-enabled predictive maintenance for hydrogen-powered transport: Advancing process safety and sustainability 氢动力运输人工智能预测性维护的全面回顾：提高过程安全性和可持续性

IF 4.2 3区工程技术 Q2 ENGINEERING, CHEMICAL

Journal of Loss Prevention in The Process Industries

Pub Date : 2026-06-01 Epub Date: 2026-02-02 DOI: 10.1016/j.jlp.2026.105953

Suresh Vellaiyan , Khalid Aljohani , Bassam S. Aljohani , Shanmugavel Kuppusamy , Nguyen Van Minh

Hydrogen-powered transportation represents a critical pathway toward decarbonized mobility. However, the reliability and safety of key subsystems, such as high-pressure hydrogen storage, fuel cells, and Balance of Plant (BoP) components, remain major barriers to large-scale deployment. This review evaluates the emerging role of Artificial Intelligence (AI)-driven Predictive Maintenance (PdM) in overcoming these challenges and advancing the sustainability of hydrogen mobility systems. Integrating Internet of Things sensing, digital twin modeling, and advanced data analytics, AI-based PdM enables early fault detection, anomaly diagnosis, and accurate estimation of component remaining useful life, thereby minimizing unplanned downtime and extending service life. Moreover, prescriptive maintenance approaches, which provide actionable recommendations in addition to failure prediction, are discussed as a next step toward autonomous maintenance management. The review also examines enabling technologies such as edge computing, federated learning, and human–machine interfaces that enhance data security, responsiveness, and operator interaction. Persistent challenges, including data heterogeneity, algorithm interpretability, cybersecurity threats, and the absence of regulatory frameworks, are critically analyzed. Finally, future perspectives highlight the need for standardized assessment protocols and cross-disciplinary collaboration to ensure safe, scalable, and intelligent adoption of AI-based PdM in hydrogen-powered transport. Overall, this review provides a comprehensive assessment framework connecting predictive intelligence, system resilience, and sustainability in next-generation hydrogen mobility.

氢动力交通是实现脱碳交通的关键途径。然而，关键子系统的可靠性和安全性，如高压储氢、燃料电池和工厂平衡（BoP）组件，仍然是大规模部署的主要障碍。本文评估了人工智能（AI）驱动的预测性维护（PdM）在克服这些挑战和推进氢动力系统可持续性方面的新兴作用。基于ai的PdM集成了物联网传感、数字孪生建模和高级数据分析，可以实现早期故障检测、异常诊断和准确估计组件剩余使用寿命，从而最大限度地减少意外停机时间，延长使用寿命。此外，规范性维护方法除了故障预测外，还提供了可操作的建议，作为自主维护管理的下一步进行了讨论。该报告还研究了诸如边缘计算、联邦学习和人机界面等增强数据安全性、响应能力和操作员交互的使能技术。持续的挑战，包括数据异构、算法可解释性、网络安全威胁和缺乏监管框架，都进行了批判性分析。最后，未来的展望强调了标准化评估协议和跨学科合作的必要性，以确保在氢动力交通中安全、可扩展和智能地采用基于人工智能的PdM。总体而言，本综述提供了一个综合评估框架，将下一代氢动力的预测智能、系统弹性和可持续性联系起来。

{"title":"A comprehensive review of AI-enabled predictive maintenance for hydrogen-powered transport: Advancing process safety and sustainability","authors":"Suresh Vellaiyan , Khalid Aljohani , Bassam S. Aljohani , Shanmugavel Kuppusamy , Nguyen Van Minh","doi":"10.1016/j.jlp.2026.105953","DOIUrl":"10.1016/j.jlp.2026.105953","url":null,"abstract":"<div><div>Hydrogen-powered transportation represents a critical pathway toward decarbonized mobility. However, the reliability and safety of key subsystems, such as high-pressure hydrogen storage, fuel cells, and Balance of Plant (BoP) components, remain major barriers to large-scale deployment. This review evaluates the emerging role of Artificial Intelligence (AI)-driven Predictive Maintenance (PdM) in overcoming these challenges and advancing the sustainability of hydrogen mobility systems. Integrating Internet of Things sensing, digital twin modeling, and advanced data analytics, AI-based PdM enables early fault detection, anomaly diagnosis, and accurate estimation of component remaining useful life, thereby minimizing unplanned downtime and extending service life. Moreover, prescriptive maintenance approaches, which provide actionable recommendations in addition to failure prediction, are discussed as a next step toward autonomous maintenance management. The review also examines enabling technologies such as edge computing, federated learning, and human–machine interfaces that enhance data security, responsiveness, and operator interaction. Persistent challenges, including data heterogeneity, algorithm interpretability, cybersecurity threats, and the absence of regulatory frameworks, are critically analyzed. Finally, future perspectives highlight the need for standardized assessment protocols and cross-disciplinary collaboration to ensure safe, scalable, and intelligent adoption of AI-based PdM in hydrogen-powered transport. Overall, this review provides a comprehensive assessment framework connecting predictive intelligence, system resilience, and sustainability in next-generation hydrogen mobility.</div></div>","PeriodicalId":16291,"journal":{"name":"Journal of Loss Prevention in The Process Industries","volume":"101 ","pages":"Article 105953"},"PeriodicalIF":4.2,"publicationDate":"2026-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146191258","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"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