Journal of Loss Prevention in The Process Industries最新文献

{"title":"An intelligent monitoring approach for urban natural gas pipeline leak using semi-supervised learning generative adversarial networks","authors":"Xinhong Li ,&nbsp;Runquan Li ,&nbsp;Ziyue Han ,&nbsp;Xin'an Yuan ,&nbsp;Xiuquan Liu","doi":"10.1016/j.jlp.2024.105476","DOIUrl":"10.1016/j.jlp.2024.105476","url":null,"abstract":"<div><div>Traditional gas pipeline leak monitoring methods are subjected to the long response times and high false alarm rates. Deep learning can enhance the accuracy and real-time performance of pipeline leak monitoring. This paper develops an intelligent monitoring approach for urban gas pipeline leaks based on a semi-supervised learning Generative Adversarial Network (SGAN). First, the Isolation Forest algorithm is used to classify anomalies in the collected process parameter data of urban natural gas pipelines. One-Hot Encoding is used to label a small amount of sample data of pipeline leak. Second, both the labeled and unlabeled data are input into SGAN model for semi-supervised learning and classification to monitor the state of urban gas pipeline leak. The methodology addresses the imbalance between pipeline leak status data and normal data. The comparison with GAN and MLP shows that the methodology reaches the highest values in all evaluation metrics (precision = 94.1%, accuracy = 95.63%, recall = 93.93%, F1 score = 94.32%). The superior performance and accuracy make it more effective for urban natural gas pipeline leak monitoring.</div></div>","PeriodicalId":16291,"journal":{"name":"Journal of Loss Prevention in The Process Industries","volume":"92 ","pages":"Article 105476"},"PeriodicalIF":3.6,"publicationDate":"2024-11-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142593642","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}

{"title":"Multi-fidelity transfer learning for complex bund overtopping prediction with varying input dimensions","authors":"Xiaoyang Luan ,&nbsp;Bin Zhang","doi":"10.1016/j.jlp.2024.105477","DOIUrl":"10.1016/j.jlp.2024.105477","url":null,"abstract":"<div><div>This study introduces a novel multi-fidelity transfer learning framework designed to predict the complex bund overtopping fraction in catastrophic tank failure scenarios. The framework addresses the challenge of varying input dimensions between low-fidelity and high-fidelity datasets by effectively integrating these disparate sources of data. In this case, low-fidelity data, generated from empirical formulas based on simple bund configurations, is first used for initial model pre-training. The model is then fine-tuned using a smaller, high-fidelity dataset obtained through computational fluid dynamics simulations, which account for more complex bund configurations, including additional breakwater parameters. This approach enhances the model's predictive accuracy and generalization capability, particularly in scenarios with limited high-fidelity data. Case studies demonstrate that the transfer learning model outperforms traditional models trained solely on high-fidelity data, offering significant reductions in computational cost while maintaining robust predictive performance. The proposed framework not only advances the understanding and prediction of bund effectiveness but also provides a versatile tool applicable to a wide range of engineering problems involving multi-fidelity data.</div></div>","PeriodicalId":16291,"journal":{"name":"Journal of Loss Prevention in The Process Industries","volume":"92 ","pages":"Article 105477"},"PeriodicalIF":3.6,"publicationDate":"2024-11-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142578135","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}

{"title":"Laminar burning velocity of ethanol-ethyl acetate mixtures in air","authors":"Ernesto Salzano ,&nbsp;Benedetta Anna De Liso ,&nbsp;Francesco Cammarota ,&nbsp;Valeria Di Sarli","doi":"10.1016/j.jlp.2024.105479","DOIUrl":"10.1016/j.jlp.2024.105479","url":null,"abstract":"<div><div>Alcohols, esters, and their mixtures are ubiquitous in the process industry. However, safety data remain limited, especially for mixtures. In this work, the laminar burning velocity, S_l, of ethanol-ethyl acetate mixtures in air was investigated (at 90 °C and atmospheric pressure) through computations performed using three different detailed chemical kinetic mechanisms, and measurements obtained from pressure time histories recorded during closed vessel explosion experiments. Computations were run varying the equivalence ratio, φ, from 0.6 to 1.7, and the mole fraction of ethanol in the fuel mixture from 0 (only ethyl acetate) to 1 (only ethanol). For all systems, explosion experiments were carried out at φ = 1.1, i.e., the composition at which, according to calculations, S_l achieves its maximum value. Regardless of the kinetic mechanism, reasonable agreement is found between computed and experimental data, including experimental data retrieved from the literature for ethanol and ethyl acetate. Results show that the behavior of ethanol-ethyl acetate is bounded between the behaviors of ethanol and ethyl acetate, approaching the former/latter as the mixture is enriched in ethanol/ethyl acetate. Over the whole range of equivalence ratios explored, the values of S_l for ethanol-ethyl acetate are smaller than those obtained by averaging the corresponding values of ethanol and ethyl acetate according to their mole proportions in the fuel mixture. This is also confirmed experimentally. A simple Le Chatelier’s mixing rule-like formula is proved to predict values of S_l that closely match both computed and experimental data, suggesting that the nature of the interaction between ethanol and ethyl acetate is predominantly thermal rather than chemical.</div></div>","PeriodicalId":16291,"journal":{"name":"Journal of Loss Prevention in The Process Industries","volume":"92 ","pages":"Article 105479"},"PeriodicalIF":3.6,"publicationDate":"2024-11-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142661651","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}

{"title":"Large eddy simulations of electrification of liquid dielectrics in channel flow","authors":"Mathieu Calero ,&nbsp;Holger Grosshans ,&nbsp;Miltiadis V. Papalexandris","doi":"10.1016/j.jlp.2024.105465","DOIUrl":"10.1016/j.jlp.2024.105465","url":null,"abstract":"<div><div>In this paper we present a computational study of electrification of liquid dielectrics in channel flow and at moderate turbulence intensities. For purposes of computational savings this study consists of large eddy simulations, according to which the large turbulent scales are directly resolved by the computational grid while the effect of the smaller unresolved scales is suitably modeled. First we examine different subgrid-scale models for the electric charge density. Their efficacy is then assessed via comparisons with results from direct numerical simulations at low turbulence intensities. According to our comparisons and analysis, the most efficient model is the one based on the eddy diffusivity approach and the introduction of the turbulent electric Schmidt number. Subsequently, we present numerical results of electrification of benzene in channel flow and at different Reynolds numbers. Herein we discuss the various stages of the electrification process and the variation of the total charge and streaming current with the turbulence intensity. According to our findings, the increase of the turbulence intensity rises dramatically not only the amount of charge transported in the bulk of the fluid but also the rate at which this transport takes place. Finally, we present results for the statistics of the charge density, which provide additional insight about the distribution of the electric charges in the liquid.</div></div>","PeriodicalId":16291,"journal":{"name":"Journal of Loss Prevention in The Process Industries","volume":"92 ","pages":"Article 105465"},"PeriodicalIF":3.6,"publicationDate":"2024-10-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142586404","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}

{"title":"CFD simulation and experimental study on the flow field and typical energetic dust dispersion in the 5 L column vessel","authors":"Haoyang Guo,&nbsp;Mengli Yin,&nbsp;Shengnan Shi,&nbsp;Wenhui Wang,&nbsp;Erhai An,&nbsp;Xiong Cao","doi":"10.1016/j.jlp.2024.105473","DOIUrl":"10.1016/j.jlp.2024.105473","url":null,"abstract":"<div><div>The explosion of energetic dust is more hazardous than general industrial dust, but there was a dearth of suitable equipment for measuring the explosion characteristics of energetic dust. At present, a 5 L column vessel can be used to measure the explosion characteristics of energetic dust has been created. This study examined the internal flow field and dispersion pattens of RDX dust with different particle sizes within the 5L vessel by combining CFD simulation and experiments. The results demonstrated that the flow field underwent a transformation following the introduction of RDX dust, accompanied by a notable decline in turbulence level and velocity within the vessel, and the comparison of simulation and experimental results showed that the valve weakened the airflow. The large-sized (Class-II) dust showed significant settling and more independent motion trajectory, whereas the small-sized (Class-V) dust formed a dust column in the central line of the vessel and showed better suspension, but the flow field exhibited a lack of symmetry. The optimal ignition delay time was 70 ms, and in the 70 ± 10 ms interval, the distribution of Class-II dust was more uniform. Compared with the existing 20 L spherical vessel, the 5 L vessel exhibited a markedly lower but more uniform turbulence level, and the dust distribution was significantly improved, meanwhile solved the problems of dust accumulation in the pipeline and particle size change. This study provides an important reference for the further research of 5 L column vessel and proves that it can achieve higher safety and accuracy in energetic dust explosion experiment.</div></div>","PeriodicalId":16291,"journal":{"name":"Journal of Loss Prevention in The Process Industries","volume":"92 ","pages":"Article 105473"},"PeriodicalIF":3.6,"publicationDate":"2024-10-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142553867","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}

{"title":"Machine learning-based literature review on the concept of safety barriers against hazardous events","authors":"Elena Stefana ,&nbsp;Marilia Ramos ,&nbsp;Nicola Paltrinieri","doi":"10.1016/j.jlp.2024.105470","DOIUrl":"10.1016/j.jlp.2024.105470","url":null,"abstract":"<div><div>The identification and implementation of effective safety barriers represent the core aim of any safety management process. Safety barriers, whether physical and/or non-physical, serve to prevent or mitigate hazardous events. The interest towards this topic has stimulated a wide spectrum of research activities that is difficult to investigate and summarise by only human analysts. In such a context, data-driven methods could assist the screening and review of the extensive number of contributions on safety barriers. However, to the best of our knowledge, this has not been addressed yet. For this reason, we propose a systematic literature review of the concept of safety barriers by employing an automated unsupervised Machine Learning-based clustering technique. To ensure the effectiveness of the automated clustering, we also performed a manual cleansing. As a result, 769 articles published until 2023 were retrieved from the Scopus database, and were grouped into 21 relevant cluster. Such clusters characterise the main research streams on the safety barriers, including risk assessment approaches, quantitative methods estimating event probabilities, design and management principles, and applications of the barriers in critical and industrial domains. This paper thus advocates for adopting a different perspective to investigate the safety barrier knowledge, by leveraging on the potentialities offered by the scientometric analysis and mapping in the safety science.</div></div>","PeriodicalId":16291,"journal":{"name":"Journal of Loss Prevention in The Process Industries","volume":"92 ","pages":"Article 105470"},"PeriodicalIF":3.6,"publicationDate":"2024-10-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142572051","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Machine learning-based precise monitoring of aluminium-magnesium alloy dust","authors":"Fengyu Zhao,&nbsp;Wei Gao,&nbsp;Jianxin Lu,&nbsp;Haipeng Jiang","doi":"10.1016/j.jlp.2024.105471","DOIUrl":"10.1016/j.jlp.2024.105471","url":null,"abstract":"<div><div>Al-Mg alloys are widely used in industrial production, which can lead to occupational health issues and explosion hazards. The study focuses on applying a machine learning-enhanced Kalman filtering algorithm to detect the concentration of Al-Mg alloy dust, significantly reducing dust hazards and constructing an efficient and safe dust reduction and removal system. A machine learning-based Kalman filter algorithm is proposed for fast and accurate detection of high Al-Mg dust concentrations (200–1200 g/m³). The results show that the KFGRU approach outperforms the traditional line filter method, achieving answer times between 2.6 s and 6 s—an improvement of 62.5% over the traditional method. As far as the forecast accuracy is concerned, the KFGRU method yields a minimal curve deviation value, reaching as low as 0.097, which represents a significant improvement compared to the 0.151 of the Kalman filter algorithm, the 0.217 of the sliding average method, and the 0.177 of the median filter methods.</div></div>","PeriodicalId":16291,"journal":{"name":"Journal of Loss Prevention in The Process Industries","volume":"92 ","pages":"Article 105471"},"PeriodicalIF":3.6,"publicationDate":"2024-10-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142532532","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}

{"title":"Study on the inhibition of hydrogen explosion using gas-solid two-phase inhibitors","authors":"Yanan Yu ,&nbsp;Haowen Qu ,&nbsp;Ping Ping ,&nbsp;Yi Liu","doi":"10.1016/j.jlp.2024.105466","DOIUrl":"10.1016/j.jlp.2024.105466","url":null,"abstract":"<div><div>Hydrogen, as a renewable energy source, is widely used in energy supply and industrial production. However, due to frequent hydrogen explosion accidents posing significant threats to people and property, exploring effective inhibition methods and mechanisms becomes crucial. This study employs a 20L spherical device to evaluate the inhibitory effects of different inert gases (such as N₂ and CO₂) combined with various inert powders (including SiO₂, NH₄H₂PO₄, and NaH₂PO₄) as two-phase inhibitors on hydrogen explosions. The results show that single-phase CO₂ outperforms N₂ in explosion suppression, achieving complete prevention of hydrogen explosions at a 60% CO₂ concentration. N₂ shows minimal inhibition at concentrations below 40%, and even at 60%, it fails to provide complete inhibition. Two-phase inhibitors significantly enhance the inhibitory effects of inert gases, and reduce their required amounts. Both under N₂ and CO₂ conditions, the effectiveness of inhibitors is ranked as follows: NH₄H₂PO₄ &gt; SiO₂ &gt; NaH₂PO₄, with NH₄H₂PO₄ exhibiting excellent inhibitory capabilities. Mechanistic analysis, conducted using HSC and CHEMKIN molecular dynamics software, reveals that the basic reaction H + O₂ = O + OH possesses the highest sensitivity coefficient in hydrogen combustion, underscoring its vital function in elevating explosion temperatures. Adding NH₄H₂PO₄ lowers the maximum combustion temperature of hydrogen and prolongs the time required to reach this temperature. The primary function of NH₄H₂PO₄ is to lower the levels of H and O radicals by creating NH₃ intermediates and merging them with O and H atoms, which in turn diminishes the intensity of the reaction and effectively inhibits explosions.</div></div>","PeriodicalId":16291,"journal":{"name":"Journal of Loss Prevention in The Process Industries","volume":"92 ","pages":"Article 105466"},"PeriodicalIF":3.6,"publicationDate":"2024-10-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142571659","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}

{"title":"Experimental investigation on the vented flame and pressure behaviour of hydrogen-air mixtures","authors":"Yuhuai Sheng ,&nbsp;Zhenmin Luo ,&nbsp;Litao Liu ,&nbsp;Zhe Yang ,&nbsp;Fan Meng ,&nbsp;Zhe Dong ,&nbsp;Yanni Zhang ,&nbsp;Jiao Qu ,&nbsp;Jun Deng ,&nbsp;Tao Wang","doi":"10.1016/j.jlp.2024.105469","DOIUrl":"10.1016/j.jlp.2024.105469","url":null,"abstract":"<div><div>To explore the changes in pressure and flame propagation inside and outside the vessel when the pipe diameter does not match the vent diameter, and to check the conservatism of standards NFPA 68 and EN 14491 for this case, an experimental apparatus was utilized to investigate the influences of static activation pressure (0.7–1.75 bar) and vent diameter (30–70 mm) on the deflagration behavior of hydrogen–air mixtures (Φ: 0.6–1.4). The vented pressure, flame propagation, and pressure–flame interaction characteristics of hydrogen–air mixtures were observed and analyzed. Additionally, the conservatism of the calculated venting area under the experimental conditions of NPFA 68 and EN 14491 was verified. The results indicated that under an equivalence ratio of 0.6, the pressure-time curve inside the container exhibited only one peak (P_max11). In the stoichiometric and fuel-rich states, the pressure-time curve inside the container exhibited two peaks attributed to the decrease in venting efficiency due to the secondary explosion inside the pipe, increasing the turbulence intensity within the container. When the static activation pressure is 0.7 bar, the pressures of the three vent diameters of 30, 50 and 70 dropped by 11.34%, 24.38% and 26.86% respectively. And the P_max11 at Φ = 1.0 and Φ = 1.4 are similar, while this phenomenon is not observed for other vent diameters. When the vent diameter was inconsistent with the duct diameter. the calculations of both standards were conservative. However, under these testing conditions (vent diameter: 30–70 mm, static activation pressure: 0.7–1.75 bar), the NPFA 68 calculations yield a conservatism range of 3.3–11, whereas EN14491 ranges from 1.6 to 15.7. The NPFA 68 results were more stable and concentrated, making these conditions more suitable for industry safety design in hydrogen venting.</div></div>","PeriodicalId":16291,"journal":{"name":"Journal of Loss Prevention in The Process Industries","volume":"92 ","pages":"Article 105469"},"PeriodicalIF":3.6,"publicationDate":"2024-10-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142553868","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}

{"title":"The study of flame length for the horizontal jet flame of carbon dioxide and propane mixed gas","authors":"Yang Gao ,&nbsp;Zhengkang Lu ,&nbsp;Yuke Gao ,&nbsp;Peixiang He ,&nbsp;Chunmei Wang ,&nbsp;Changfa Tao","doi":"10.1016/j.jlp.2024.105467","DOIUrl":"10.1016/j.jlp.2024.105467","url":null,"abstract":"<div><div>Inert gases, such as carbon dioxide (CO₂), are often presented in fuel gases and influence their combustion characteristics significantly. This paper investigates the horizontal length of buoyant turbulent jet flame under different boundary conditions, including different nozzle diameters, heat release rates, and CO₂ volume fractions. According to the experiments, it is found that the horizontal length of jet flame increases with the heat release rate under the same nozzle diameter. Additionally, the flame length initially increases with CO₂ concentration but then decreases. Based on an analysis of the buoyancy and momentum flux of the turbulent jet flame, a new correlation has been developed, which relates the horizontal flame length to the heat release rate and CO₂ volume fraction. This correlation can be served as a valuable reference for fire prevention and control in gas leakage fire scenarios.</div></div>","PeriodicalId":16291,"journal":{"name":"Journal of Loss Prevention in The Process Industries","volume":"92 ","pages":"Article 105467"},"PeriodicalIF":3.6,"publicationDate":"2024-10-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142571658","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}