International Journal of Thermofluids最新文献

Corrosion effects on a leakage steam control valve: analysis and risk assessment on the biodiesel plant case study 泄漏蒸汽控制阀的腐蚀效应：生物柴油装置案例分析与风险评估

Q1 Chemical Engineering

International Journal of Thermofluids

Pub Date : 2026-01-08 DOI: 10.1016/j.ijft.2026.101554

Anggara Dwita Burmana , Rondang Tambun , Taslim , Barbara Ernst , Yacine Benguerba , Iriany

General surface erosion can be caused by the continuous exposure of metal surfaces to corrosive materials, resulting in further weight loss. This happened in the steam control valve, where continuous exposure to the steam flow rate resulted in erosion of the control valve wall. The purpose of this study is to discuss the causes of steam control valve failure due to corrosion in a biodiesel plant using the gravimetric and salt spray methods, based on the effects caused, and to review the safety risks and implication in more detail. Corrosion which occurred in the first year was 0.7321 mm/year, while corrosion which occurred in years 12 to 15 was relatively more stable at about 0.005 mm/year. Weight and metal losses during the first year were 45 g/year and 0.6052 mm/year, while those during the 14th to 15th years were relatively stable at about 4 g/years and 0.0630 mm/years. The results of the control valve salt spray chamber measurements are value obtained shows that higher NaCl concentration causes higher corrosion. In this observation, the highest value obtained was 0.3181 mm/year at 20% NaCl concentration, while the lowest value was 0.1548 mm/year at 5% NaCl concentration. A robust maintenance schedule helps identify early signs of wear and corrosion, allowing for timely intervention. Implementing a comprehensive maintenance and predictive programme will save time and reduce operating costs for a biodiesel plant.

由于金属表面持续暴露于腐蚀性材料中，导致进一步的重量损失，通常会引起表面侵蚀。这种情况发生在蒸汽控制阀中，连续暴露在蒸汽流量下导致控制阀壁的侵蚀。本研究的目的是利用重量法和盐雾法探讨生物柴油装置蒸汽控制阀因腐蚀而失效的原因，并根据所造成的影响对其安全风险和影响进行更详细的审查。第一年发生的腐蚀为0.7321 mm/年，而12至15年发生的腐蚀相对稳定，约为0.005 mm/年。第1年的失重和金属损失量分别为45 g/年和0.6052 mm/年，第14 ~ 15年的失重和金属损失量相对稳定，分别为4 g/年和0.0630 mm/年左右。控制阀盐雾室测量结果表明，NaCl浓度越高，腐蚀越严重。结果表明，在20% NaCl浓度下，其最大值为0.3181 mm/年，在5% NaCl浓度下，其最小值为0.1548 mm/年。强有力的维护计划有助于识别磨损和腐蚀的早期迹象，从而及时进行干预。实施全面的维护和预测计划将节省时间并降低生物柴油厂的运营成本。

{"title":"Corrosion effects on a leakage steam control valve: analysis and risk assessment on the biodiesel plant case study","authors":"Anggara Dwita Burmana , Rondang Tambun , Taslim , Barbara Ernst , Yacine Benguerba , Iriany","doi":"10.1016/j.ijft.2026.101554","DOIUrl":"10.1016/j.ijft.2026.101554","url":null,"abstract":"<div><div>General surface erosion can be caused by the continuous exposure of metal surfaces to corrosive materials, resulting in further weight loss. This happened in the steam control valve, where continuous exposure to the steam flow rate resulted in erosion of the control valve wall. The purpose of this study is to discuss the causes of steam control valve failure due to corrosion in a biodiesel plant using the gravimetric and salt spray methods, based on the effects caused, and to review the safety risks and implication in more detail. Corrosion which occurred in the first year was 0.7321 mm/year, while corrosion which occurred in years 12 to 15 was relatively more stable at about 0.005 mm/year. Weight and metal losses during the first year were 45 g/year and 0.6052 mm/year, while those during the 14th to 15th years were relatively stable at about 4 g/years and 0.0630 mm/years. The results of the control valve salt spray chamber measurements are value obtained shows that higher NaCl concentration causes higher corrosion. In this observation, the highest value obtained was 0.3181 mm/year at 20% NaCl concentration, while the lowest value was 0.1548 mm/year at 5% NaCl concentration. A robust maintenance schedule helps identify early signs of wear and corrosion, allowing for timely intervention. Implementing a comprehensive maintenance and predictive programme will save time and reduce operating costs for a biodiesel plant.</div></div>","PeriodicalId":36341,"journal":{"name":"International Journal of Thermofluids","volume":"32 ","pages":"Article 101554"},"PeriodicalIF":0.0,"publicationDate":"2026-01-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145981466","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Study of aspect ratio effect in combined shear-driven and pressure-difference driven flow characterization of a third grade fluid in a rectangular channel 矩形通道中三级流体剪切-压差联合流动特性的展弦比效应研究

Q1 Chemical Engineering

International Journal of Thermofluids

Pub Date : 2026-01-07 DOI: 10.1016/j.ijft.2026.101552

Ayusman Nayak, Rajiva Lochan Mohanty, Sumanta Chaudhuri

Combined shear-driven and pressure-difference driven flow of a third grade fluid through a rectangular channel is studied as it has wide applications in extrusion, glass-fibre drawing, blood flow in arteries etc.. Pressure-difference driven flow case can be retrieved by setting the upper walls velocity to be zero. First, governing equation is formulated from the constitutive equation of third grade fluid considering hydro-dynamically fully developed flow, where the axial velocity is independent of the axial coordinate. The dimensionless governing equation is solved using the least square method using Symbolic Computation in MATLAB. As no experimental result or other numerical or analytical results are reported in the literature, the equation is again solved by least square homotopy perturbation method and standard numerical technique. Results from both these methods are compared for validation. In addition, velocity distribution in the limit of vanishing aspect ratio is compared with that of the previously reported results of third grade fluid flow between parallel plates. In both the cases, results are in close match which validate the results of the present study. Influence of aspect ratio, third grade fluid parameter on velocity distribution are analyzed. In case of combined shear- driven and pressure-difference driven flow, the effect of favourable and adverse pressure gradient is discussed. It is noted that increase in aspect ratio, arrests the flow separation and back flow in the channel. Increase in third-grade fluid parameter also increases the flow resistance with consequent prevention of flow separation and back flow.

研究了剪切驱动和压差驱动三级流体在矩形通道中的联合流动，该通道在挤压、玻璃纤维拉伸、动脉血流等方面有广泛的应用。通过将上壁速度设置为零，可以恢复压差驱动的流动情况。首先，考虑流动力充分发展，轴向速度与轴向坐标无关，由三级流体本构方程推导控制方程；利用MATLAB中的符号计算，采用最小二乘法求解无量纲控制方程。由于文献中没有实验结果，也没有其他数值或解析结果的报道，方程再次采用最小二乘同伦摄动法和标准数值技术求解。对两种方法的结果进行了比较验证。此外，还将消失长径比极限下的速度分布与已有报道的平行板间三级流体流动结果进行了比较。在这两种情况下，结果是密切匹配的，验证了本研究的结果。分析了展弦比、三级流体参数对速度分布的影响。在剪切驱动和压差驱动联合流动的情况下，讨论了有利压力梯度和不利压力梯度的影响。宽高比的增大抑制了流道内的流动分离和回流。三级流体参数的增加也增加了流动阻力，从而防止了流动分离和回流。

{"title":"Study of aspect ratio effect in combined shear-driven and pressure-difference driven flow characterization of a third grade fluid in a rectangular channel","authors":"Ayusman Nayak, Rajiva Lochan Mohanty, Sumanta Chaudhuri","doi":"10.1016/j.ijft.2026.101552","DOIUrl":"10.1016/j.ijft.2026.101552","url":null,"abstract":"<div><div>Combined shear-driven and pressure-difference driven flow of a third grade fluid through a rectangular channel is studied as it has wide applications in extrusion, glass-fibre drawing, blood flow in arteries etc.. Pressure-difference driven flow case can be retrieved by setting the upper walls velocity to be zero. First, governing equation is formulated from the constitutive equation of third grade fluid considering hydro-dynamically fully developed flow, where the axial velocity is independent of the axial coordinate. The dimensionless governing equation is solved using the least square method using Symbolic Computation in MATLAB. As no experimental result or other numerical or analytical results are reported in the literature, the equation is again solved by least square homotopy perturbation method and standard numerical technique. Results from both these methods are compared for validation. In addition, velocity distribution in the limit of vanishing aspect ratio is compared with that of the previously reported results of third grade fluid flow between parallel plates. In both the cases, results are in close match which validate the results of the present study. Influence of aspect ratio, third grade fluid parameter on velocity distribution are analyzed. In case of combined shear- driven and pressure-difference driven flow, the effect of favourable and adverse pressure gradient is discussed. It is noted that increase in aspect ratio, arrests the flow separation and back flow in the channel. Increase in third-grade fluid parameter also increases the flow resistance with consequent prevention of flow separation and back flow.</div></div>","PeriodicalId":36341,"journal":{"name":"International Journal of Thermofluids","volume":"32 ","pages":"Article 101552"},"PeriodicalIF":0.0,"publicationDate":"2026-01-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145981355","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","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 preparation and thermal performance of carbon sphere-modified PCM perlite-paraffin composite 碳球改性PCM珍珠岩-石蜡复合材料的制备及热性能研究

Q1 Chemical Engineering

International Journal of Thermofluids

Pub Date : 2026-01-06 DOI: 10.1016/j.ijft.2026.101553

Qi Chen , Wei Na , Quanhe Gai , Jie Yang , Yan Wang , Xucheng Chen , Zhao Cao , Shuguang Li

A carbon sphere modification approach was developed to construct a perlite/paraffin composite phase change material (CS-PCM) with both high photothermal conversion efficiency and stable thermal energy storage capability. The surface of the perlite matrix was modified using sodium carboxymethyl cellulose (CMCNa), increasing the water contact angle from 0° to 67° Micro/nanoscale carbon spheres (0.1–20 μm) derived from tea waste were embedded into a resin matrix, demonstrating superior performance compared to conventional carbon-based additives. The CS-PCM was incorporated into thermal storage walls through a scalable manufacturing process, with phase change materials comprising 30 % of the system. Experimental application in high-latitude greenhouses during winter confirmed that the integrated system maintained an internal crop environment at 13.6 ± 4 °C. This composite material features both low cost and reduced carbon emissions. The results offer a promising solution for thermal management in protected agriculture in cold regions and provide a theoretical and technical reference for the design of multifunctional phase change materials.

采用碳球改性的方法制备了光热转换效率高、储热性能稳定的珍珠岩/石蜡复合相变材料（CS-PCM）。采用羧甲基纤维素钠（CMCNa）对珍珠岩基体表面进行改性，将水接触角从0°增加到67°，将茶渣制备的微纳碳球（0.1 ~ 20 μm）嵌入树脂基体中，表现出优于传统碳基添加剂的性能。CS-PCM通过可扩展的制造工艺集成到储热壁中，相变材料占系统的30%。冬季在高纬度温室的试验应用证实，该综合系统将作物内部环境维持在13.6±4°C。这种复合材料具有成本低、碳排放少的特点。研究结果为寒冷地区保护农业的热管理提供了一种有希望的解决方案，并为多功能相变材料的设计提供了理论和技术参考。

{"title":"Research on preparation and thermal performance of carbon sphere-modified PCM perlite-paraffin composite","authors":"Qi Chen , Wei Na , Quanhe Gai , Jie Yang , Yan Wang , Xucheng Chen , Zhao Cao , Shuguang Li","doi":"10.1016/j.ijft.2026.101553","DOIUrl":"10.1016/j.ijft.2026.101553","url":null,"abstract":"<div><div>A carbon sphere modification approach was developed to construct a perlite/paraffin composite phase change material (CS-PCM) with both high photothermal conversion efficiency and stable thermal energy storage capability. The surface of the perlite matrix was modified using sodium carboxymethyl cellulose (CMC<img>Na), increasing the water contact angle from 0° to 67° Micro/nanoscale carbon spheres (0.1–20 μm) derived from tea waste were embedded into a resin matrix, demonstrating superior performance compared to conventional carbon-based additives. The CS-PCM was incorporated into thermal storage walls through a scalable manufacturing process, with phase change materials comprising 30 % of the system. Experimental application in high-latitude greenhouses during winter confirmed that the integrated system maintained an internal crop environment at 13.6 ± 4 °C. This composite material features both low cost and reduced carbon emissions. The results offer a promising solution for thermal management in protected agriculture in cold regions and provide a theoretical and technical reference for the design of multifunctional phase change materials.</div></div>","PeriodicalId":36341,"journal":{"name":"International Journal of Thermofluids","volume":"32 ","pages":"Article 101553"},"PeriodicalIF":0.0,"publicationDate":"2026-01-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145981465","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

A stable high mixing performance of Koch fractal array obstacle-based micromixer 基于Koch分形阵列障碍物的微混合器具有稳定的高混合性能

Q1 Chemical Engineering

International Journal of Thermofluids

Pub Date : 2026-01-05 DOI: 10.1016/j.ijft.2026.101550

Lilik Hasanah , Fasya Nur Afifah , Roer Eka Pawinanto , Muhammad Iqbal , Gilang Gumilar , Muhammad Yusuf , Widyaningrum Indrasari , Ida Hamidah , Jumril Yunas , Budi Mulyanti

Passive micromixers are essential components in microfluidic systems, enabling efficient fluid mixing under laminar flow conditions without external energy input. This study numerically investigates the mixing performance and flow mechanisms of passive micromixers incorporating Koch fractal obstacle arrays. Four fractal geometries—Secondary Snowflakes Fractal (SSF), Tertiary Snowflakes Fractal (TSF), Rounded Secondary Snowflakes Fractal (RSSF), and Rounded Tertiary Snowflakes Fractal (RTSF) are analyzed in both same-side and different-side configurations using COMSOL Multiphysics over a wide Reynolds number range (Re = 0.1–100). Model validation against benchmark obstacle-based micromixers from the literature shows good agreement in mixing efficiency and pressure drop, confirming the reliability of the numerical framework. The results demonstrate that Koch fractal obstacles enhance mixing through flow splitting, stretching, and chaotic advection. Among all configurations, the different-side rounded tertiary snowflake fractal (DSRTSF) exhibits the most stable and consistently high mixing performance, achieving a maximum mixing efficiency of 97.70% at a Reynolds number (Re) of 0.1. Performance index analysis further reveals that rounded fractal geometries provide a favorable balance between mixing efficiency and pressure drop. These findings offer practical design guidelines for high-performance fractal obstacle-based passive micromixers in lab-on-a-chip and microfluidic applications.

无源微混合器是微流体系统的重要组成部分，它可以在层流条件下实现高效的流体混合，而无需外部能量输入。本文对含Koch分形障碍物阵列的无源微混合器的混合性能和流动机理进行了数值研究。使用COMSOL Multiphysics在宽雷诺数范围（Re = 0.1-100）下分析了四种分形几何形状——二次雪花分形（SSF）、三级雪花分形（TSF）、圆形二级雪花分形（RSSF）和圆形三级雪花分形（RTSF）。针对文献中基于障碍物的基准微混合器的模型验证表明，混合效率和压降具有良好的一致性，证实了数值框架的可靠性。结果表明，科赫分形障碍物通过气流分裂、拉伸和混沌平流增强混合。不同侧面圆形三级雪花分形（DSRTSF）的混合性能最稳定，在雷诺数（Re）为0.1时，混合效率最高，达到97.70%。性能指标分析进一步表明，圆形分形几何形状在混合效率和压降之间提供了良好的平衡。这些发现为芯片实验室和微流控应用中基于分形障碍物的高性能无源微混合器提供了实用的设计指南。

{"title":"A stable high mixing performance of Koch fractal array obstacle-based micromixer","authors":"Lilik Hasanah , Fasya Nur Afifah , Roer Eka Pawinanto , Muhammad Iqbal , Gilang Gumilar , Muhammad Yusuf , Widyaningrum Indrasari , Ida Hamidah , Jumril Yunas , Budi Mulyanti","doi":"10.1016/j.ijft.2026.101550","DOIUrl":"10.1016/j.ijft.2026.101550","url":null,"abstract":"<div><div>Passive micromixers are essential components in microfluidic systems, enabling efficient fluid mixing under laminar flow conditions without external energy input. This study numerically investigates the mixing performance and flow mechanisms of passive micromixers incorporating Koch fractal obstacle arrays. Four fractal geometries—Secondary Snowflakes Fractal (SSF), Tertiary Snowflakes Fractal (TSF), Rounded Secondary Snowflakes Fractal (RSSF), and Rounded Tertiary Snowflakes Fractal (RTSF) are analyzed in both same-side and different-side configurations using COMSOL Multiphysics over a wide Reynolds number range (<em>Re</em> = 0.1–100). Model validation against benchmark obstacle-based micromixers from the literature shows good agreement in mixing efficiency and pressure drop, confirming the reliability of the numerical framework. The results demonstrate that Koch fractal obstacles enhance mixing through flow splitting, stretching, and chaotic advection. Among all configurations, the different-side rounded tertiary snowflake fractal (DSRTSF) exhibits the most stable and consistently high mixing performance, achieving a maximum mixing efficiency of 97.70% at a Reynolds number (<em>Re</em>) of 0.1. Performance index analysis further reveals that rounded fractal geometries provide a favorable balance between mixing efficiency and pressure drop. These findings offer practical design guidelines for high-performance fractal obstacle-based passive micromixers in lab-on-a-chip and microfluidic applications.</div></div>","PeriodicalId":36341,"journal":{"name":"International Journal of Thermofluids","volume":"32 ","pages":"Article 101550"},"PeriodicalIF":0.0,"publicationDate":"2026-01-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145981356","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

CFD analysis of temperature-driven flow for passive HVAC applications in Irish homes 爱尔兰家庭被动式暖通空调应用中温度驱动流量的CFD分析

Q1 Chemical Engineering

International Journal of Thermofluids

Pub Date : 2026-01-04 DOI: 10.1016/j.ijft.2026.101551

Rahul D , Anjaneya G , Manjunatha N K , Sunil S , Mohan Kumar G R , Girish Kumar G S , C.Durga Prasad , Nithesh Kumble Gokuldas , Nilesh H. Khandare , Nimona Hailu

Buoyancy-driven ventilation is a sustainable alternative to conventional HVAC systems, leveraging temperature differences to induce airflow. This study analyzes buoyancy-driven flow using Computational Fluid Dynamics (CFD) in ANSYS Fluent to evaluate its feasibility for natural ventilation. The simulation models a contracting pipe with internal heating fins, made of copper, which gets heated by sunlight up to 60 °C, inducing airflow through natural convection. A pressure inlet at 0 gauge pressure and a pressure outlet at −20 Pa were applied. The steady-time simulation used the realizable k-ε turbulence model with buoyancy effects enabled and the energy equation enabled. Results show flow from the inlet to the outlet domain, which was confirmed by evaluating volumetric flow rate at the outlet and observing velocity streamlines, confirming that the design effectively enhances natural ventilation. A grid independence study was conducted by evaluating the Grashof number (Gr) on one of the heated walls. Grid independence was achieved with 693635 elements with a Grashof number value of 3*e¹⁰. A parametric study was conducted for varying wall temperatures (30 °C, 40 °C, 50 °C, and 60 °C), and the Volumetric flow rate at the outlet was monitored. Results indicate a linear trend between temperature and volumetric flow at the outlet, proving that more air can be expelled with higher blade temperatures.

浮力驱动通风是传统暖通空调系统的可持续替代方案，利用温差诱导气流。本研究利用ANSYS Fluent中的计算流体动力学（Computational Fluid Dynamics， CFD）对浮力驱动的流动进行分析，以评估其用于自然通风的可行性。该模拟模拟了一个内部带有铜制发热片的收缩管道，该管道被阳光加热到60°C，通过自然对流诱导气流。施加0表压力的压力入口和−20 Pa的压力出口。稳态时间模拟采用可实现的k-ε湍流模型，启用浮力效应并启用能量方程。结果表明，通过对出口体积流量的评估和对速度流线的观察，证实了从进口到出口的流动，证实了该设计有效地增强了自然通风。通过对其中一个加热壁面的Grashof数（Gr）进行评估，进行了网格独立性研究。使用693635个元素实现网格独立性，Grashof数值为3*e10。在不同的壁面温度（30°C、40°C、50°C和60°C）下进行了参数化研究，并监测了出口的体积流量。结果表明，出口温度与体积流量呈线性关系，表明叶片温度越高，排风量越大。

{"title":"CFD analysis of temperature-driven flow for passive HVAC applications in Irish homes","authors":"Rahul D , Anjaneya G , Manjunatha N K , Sunil S , Mohan Kumar G R , Girish Kumar G S , C.Durga Prasad , Nithesh Kumble Gokuldas , Nilesh H. Khandare , Nimona Hailu","doi":"10.1016/j.ijft.2026.101551","DOIUrl":"10.1016/j.ijft.2026.101551","url":null,"abstract":"<div><div>Buoyancy-driven ventilation is a sustainable alternative to conventional HVAC systems, leveraging temperature differences to induce airflow. This study analyzes buoyancy-driven flow using Computational Fluid Dynamics (CFD) in ANSYS Fluent to evaluate its feasibility for natural ventilation. The simulation models a contracting pipe with internal heating fins, made of copper, which gets heated by sunlight up to 60 °C, inducing airflow through natural convection. A pressure inlet at 0 gauge pressure and a pressure outlet at −20 Pa were applied. The steady-time simulation used the realizable k-ε turbulence model with buoyancy effects enabled and the energy equation enabled. Results show flow from the inlet to the outlet domain, which was confirmed by evaluating volumetric flow rate at the outlet and observing velocity streamlines, confirming that the design effectively enhances natural ventilation. A grid independence study was conducted by evaluating the Grashof number (Gr) on one of the heated walls. Grid independence was achieved with 693635 elements with a Grashof number value of 3*e<sup>10</sup>. A parametric study was conducted for varying wall temperatures (30 °C, 40 °C, 50 °C, and 60 °C), and the Volumetric flow rate at the outlet was monitored. Results indicate a linear trend between temperature and volumetric flow at the outlet, proving that more air can be expelled with higher blade temperatures.</div></div>","PeriodicalId":36341,"journal":{"name":"International Journal of Thermofluids","volume":"32 ","pages":"Article 101551"},"PeriodicalIF":0.0,"publicationDate":"2026-01-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145981357","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Heat and mass transfer and irreversibilities of humid air flow over a flat plate operating under vacuum 真空条件下平板上湿空气流动的传热传质和不可逆性

Q1 Chemical Engineering

International Journal of Thermofluids

Pub Date : 2026-01-01 DOI: 10.1016/j.ijft.2025.101540

Vo Long Hai , Nguyen Duc Nam , Nguyen Minh Phu

This study presents a numerical investigation into the coupled heat and mass transfer phenomena and the associated thermodynamic irreversibilities (entropy generation) of humid air flowing over a flat plate under vacuum conditions (40 kPa to 100 kPa). A comprehensive mathematical model, incorporating the continuity, momentum, energy, and species transport equations, was established and solved using the finite difference method. The model's accuracy was validated by comparing the numerical local heat transfer coefficient results, which demonstrated excellent agreement with theoretical predictions. Key findings indicate that lower operating pressures significantly enhance mass transfer efficiency: the local mass transfer coefficient approached its maximum value of approximately 4.5 × 10⁻⁴ m/s at 40 kPa. This enhancement translates directly to water production, with the hourly condensation rate increasing from about 0.11 kg/m²-h at 100 kPa to approximately 0.19 kg/m²-h at 40 kPa. Examination of irreversibilities showed that thermal irreversibility is the predominant contributor to entropy generation. At 100 kPa, the local thermal irreversibility peaked at approximately 50 W/m³-K (compared to concentration irreversibility up to 10 W/m³-K and viscous irreversibility around 1.6 × 10⁻³ W/m³-K. On average, the total irreversibility increased with operating pressure, ranging from approximately 0.6 to 0.85 W/m³-K as the pressure rose from 40 kPa to 100 kPa. These quantitative insights offer valuable understanding for optimizing the design and improving the performance of humid air systems operating under vacuum by focusing on reducing thermal energy dissipation.

本文对真空条件下（40kpa至100kpa）平板上流动的湿空气的传热传质耦合现象和相关的热力学不可逆性（熵产）进行了数值研究。建立了包含连续性、动量、能量和物质输运方程的综合数学模型，并用有限差分法求解。通过对局部换热系数数值计算结果的比较，验证了模型的准确性，结果与理论预测结果非常吻合。关键研究结果表明，较低的操作压力显著提高了传质效率：在40 kPa时，局部传质系数接近最大值，约为4.5 × 10−4 m/s。这种增强直接转化为产水，每小时冷凝速率从100 kPa时的约0.11 kg/m2-h增加到40 kPa时的约0.19 kg/m2-h。对不可逆性的考察表明，热不可逆性是熵产生的主要因素。在100 kPa时，局部热不可逆性峰值约为50 W/m3-K（相比之下，浓度不可逆性高达10 W/m3-K，粘性不可逆性约为1.6 × 10−3 W/m3-K）。总的不可逆性随着操作压力的增加而增加，从40 kPa增加到100 kPa，其变化范围约为0.6 ~ 0.85 W/m3-K。这些定量的见解为优化设计和提高真空下湿空气系统的性能提供了有价值的理解，重点是减少热能耗散。

{"title":"Heat and mass transfer and irreversibilities of humid air flow over a flat plate operating under vacuum","authors":"Vo Long Hai , Nguyen Duc Nam , Nguyen Minh Phu","doi":"10.1016/j.ijft.2025.101540","DOIUrl":"10.1016/j.ijft.2025.101540","url":null,"abstract":"<div><div>This study presents a numerical investigation into the coupled heat and mass transfer phenomena and the associated thermodynamic irreversibilities (entropy generation) of humid air flowing over a flat plate under vacuum conditions (40 kPa to 100 kPa). A comprehensive mathematical model, incorporating the continuity, momentum, energy, and species transport equations, was established and solved using the finite difference method. The model's accuracy was validated by comparing the numerical local heat transfer coefficient results, which demonstrated excellent agreement with theoretical predictions. Key findings indicate that lower operating pressures significantly enhance mass transfer efficiency: the local mass transfer coefficient approached its maximum value of approximately 4.5 × 10<sup>−4</sup> m/s at 40 kPa. This enhancement translates directly to water production, with the hourly condensation rate increasing from about 0.11 kg/m<sup>2</sup>-h at 100 kPa to approximately 0.19 kg/m<sup>2</sup>-h at 40 kPa. Examination of irreversibilities showed that thermal irreversibility is the predominant contributor to entropy generation. At 100 kPa, the local thermal irreversibility peaked at approximately 50 W/m<sup>3</sup>-K (compared to concentration irreversibility up to 10 W/m<sup>3</sup>-K and viscous irreversibility around 1.6 × 10<sup>−3</sup> W/m<sup>3</sup>-K. On average, the total irreversibility increased with operating pressure, ranging from approximately 0.6 to 0.85 W/m<sup>3</sup>-K as the pressure rose from 40 kPa to 100 kPa. These quantitative insights offer valuable understanding for optimizing the design and improving the performance of humid air systems operating under vacuum by focusing on reducing thermal energy dissipation.</div></div>","PeriodicalId":36341,"journal":{"name":"International Journal of Thermofluids","volume":"31 ","pages":"Article 101540"},"PeriodicalIF":0.0,"publicationDate":"2026-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145926542","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Comparative data-driven modeling of thermal energy storage using artificial neural networks and multiple linear regression 基于人工神经网络和多元线性回归的热能存储数据驱动模型比较

Q1 Chemical Engineering

International Journal of Thermofluids

Pub Date : 2026-01-01 DOI: 10.1016/j.ijft.2025.101539

Eman Abdelhafez , Nabeel Abu Shaban , Mohammad Hamdan , Maher Al-Maghalseh

Efficient thermal energy storage (TES) is essential for enhancing the reliability and sustainability of solar thermal systems, particularly under fluctuating solar radiation conditions. This study investigates the predictive performance of different data-driven approaches—Multiple Linear Regression (MLR), Multilayer Perceptron (MLP), and Radial Basis Function (RBF) neural networks—for estimating stored thermal energy in a solar thermal tank. Experimental data were obtained from a controlled solar simulator setup that incorporated halogen-lamp irradiation, a flat-plate collector, and nanoparticle-enhanced water as the storage medium. Eight independent variables—including collector inlet and outlet temperatures, tank and ambient temperatures, flow rate, solar radiation, nanoparticle concentration, and specific heat capacity—were used as model inputs. Results show that the MLP model significantly outperformed both MLR and RBF, achieving the highest correlation coefficient (R = 0.647), and the lowest RMSE (346.35) and MBE (152.49), demonstrating superior accuracy and generalization. By contrast, MLR exhibited limited predictive power due to its linear assumptions, while RBF suffered from high testing error and poor generalization. These findings underscore the suitability of neural network models, particularly MLP, for capturing the nonlinear dynamics of TES systems, providing a robust framework for system optimization and improved energy management strategies.

高效的热能储存（TES）对于提高太阳能热系统的可靠性和可持续性至关重要，特别是在波动的太阳辐射条件下。本研究探讨了不同数据驱动方法的预测性能——多元线性回归（MLR）、多层感知器（MLP）和径向基函数（RBF）神经网络——用于估计太阳能热储箱中储存的热能。实验数据是从一个受控的太阳模拟器装置中获得的，该装置包括卤素灯照射，平板收集器和纳米粒子增强水作为存储介质。八个独立变量——包括集热器进出口温度、水箱和环境温度、流速、太阳辐射、纳米颗粒浓度和比热容——被用作模型输入。结果表明，MLP模型的相关系数最高（R = 0.647）， RMSE最低（346.35），MBE最低（152.49），显著优于MLR和RBF，具有较好的准确率和泛化能力。相比之下，MLR由于其线性假设而具有有限的预测能力，而RBF则存在较高的测试误差和较差的泛化能力。这些发现强调了神经网络模型（尤其是MLP）在捕获TES系统非线性动力学方面的适用性，为系统优化和改进的能源管理策略提供了一个强大的框架。

{"title":"Comparative data-driven modeling of thermal energy storage using artificial neural networks and multiple linear regression","authors":"Eman Abdelhafez , Nabeel Abu Shaban , Mohammad Hamdan , Maher Al-Maghalseh","doi":"10.1016/j.ijft.2025.101539","DOIUrl":"10.1016/j.ijft.2025.101539","url":null,"abstract":"<div><div>Efficient thermal energy storage (TES) is essential for enhancing the reliability and sustainability of solar thermal systems, particularly under fluctuating solar radiation conditions. This study investigates the predictive performance of different data-driven approaches—Multiple Linear Regression (MLR), Multilayer Perceptron (MLP), and Radial Basis Function (RBF) neural networks—for estimating stored thermal energy in a solar thermal tank. Experimental data were obtained from a controlled solar simulator setup that incorporated halogen-lamp irradiation, a flat-plate collector, and nanoparticle-enhanced water as the storage medium. Eight independent variables—including collector inlet and outlet temperatures, tank and ambient temperatures, flow rate, solar radiation, nanoparticle concentration, and specific heat capacity—were used as model inputs. Results show that the MLP model significantly outperformed both MLR and RBF, achieving the highest correlation coefficient (<em>R</em> = 0.647), and the lowest RMSE (346.35) and MBE (152.49), demonstrating superior accuracy and generalization. By contrast, MLR exhibited limited predictive power due to its linear assumptions, while RBF suffered from high testing error and poor generalization. These findings underscore the suitability of neural network models, particularly MLP, for capturing the nonlinear dynamics of TES systems, providing a robust framework for system optimization and improved energy management strategies.</div></div>","PeriodicalId":36341,"journal":{"name":"International Journal of Thermofluids","volume":"31 ","pages":"Article 101539"},"PeriodicalIF":0.0,"publicationDate":"2026-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145926538","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Analysis of Plane Poiseuille flow of non-isothermal couple stress fluid between two parallel inclined plates using two reliable methods 用两种可靠的方法分析平行斜板间非等温偶应力流体的平面泊泽维尔流

Q1 Chemical Engineering

International Journal of Thermofluids

Pub Date : 2026-01-01 DOI: 10.1016/j.ijft.2025.101520

Muhammad Farooq , Faisal Zia , Rashid Nawaz , Alamgeer Khan , Ilker Ozsahin , Hijaz Ahmad , Waleed Mohammed Abdelfattah

This study is motivated by the need to understand complex thermal and hydrodynamic behaviors of couple stress fluids, which commonly occur in lubrication systems, microfluidic devices, and polymeric material processing. Its significance lies in modeling non-isothermal couple stress fluid flow through an inclined Poiseuille channel bounded by two heated parallel plates, a configuration relevant to advanced heat and mass transfer applications. The aim is to determine the velocity profile, temperature distribution, volumetric flow rate, average velocity, and shear stress for the incompressible fluid. To achieve this, the highly nonlinear coupled ordinary differential equations governing the system are solved using the Optimal Homotopy Asymptotic Method and the Homotopy Perturbation Method, which provide accurate approximate solutions without linearization. The major findings show excellent agreement between the two approaches, confirming their validity, while parametric studies reveal how physical factors such as couple stress effects, plate inclination, and thermal gradients influence the flow. The specific applications of this work include lubrication processes, thermal energy devices, and fluid transport systems requiring precise control of flow and heat transfer.

这项研究的动机是需要了解耦合应力流体的复杂热和流体动力学行为，这通常发生在润滑系统，微流体装置和聚合物材料加工中。其意义在于模拟非等温偶应力流体通过由两个受热平行板包围的倾斜泊泽维尔通道的流动，这是一种与高级传热传质应用相关的构型。目的是确定不可压缩流体的速度分布、温度分布、体积流量、平均速度和剪切应力。为此，采用最优同伦渐近法和同伦摄动法对控制系统的高度非线性耦合常微分方程进行了求解，得到了不需要线性化的精确近似解。主要研究结果表明，两种方法之间的一致性非常好，证实了它们的有效性，而参数研究揭示了诸如耦合应力效应、板倾角和热梯度等物理因素如何影响流动。这项工作的具体应用包括润滑过程，热能装置和流体输送系统，需要精确控制流动和传热。

{"title":"Analysis of Plane Poiseuille flow of non-isothermal couple stress fluid between two parallel inclined plates using two reliable methods","authors":"Muhammad Farooq , Faisal Zia , Rashid Nawaz , Alamgeer Khan , Ilker Ozsahin , Hijaz Ahmad , Waleed Mohammed Abdelfattah","doi":"10.1016/j.ijft.2025.101520","DOIUrl":"10.1016/j.ijft.2025.101520","url":null,"abstract":"<div><div>This study is motivated by the need to understand complex thermal and hydrodynamic behaviors of couple stress fluids, which commonly occur in lubrication systems, microfluidic devices, and polymeric material processing. Its significance lies in modeling non-isothermal couple stress fluid flow through an inclined Poiseuille channel bounded by two heated parallel plates, a configuration relevant to advanced heat and mass transfer applications. The aim is to determine the velocity profile, temperature distribution, volumetric flow rate, average velocity, and shear stress for the incompressible fluid. To achieve this, the highly nonlinear coupled ordinary differential equations governing the system are solved using the Optimal Homotopy Asymptotic Method and the Homotopy Perturbation Method, which provide accurate approximate solutions without linearization. The major findings show excellent agreement between the two approaches, confirming their validity, while parametric studies reveal how physical factors such as couple stress effects, plate inclination, and thermal gradients influence the flow. The specific applications of this work include lubrication processes, thermal energy devices, and fluid transport systems requiring precise control of flow and heat transfer.</div></div>","PeriodicalId":36341,"journal":{"name":"International Journal of Thermofluids","volume":"31 ","pages":"Article 101520"},"PeriodicalIF":0.0,"publicationDate":"2026-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145926541","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Base pressure control through micro jets at supersonic mach numbers using experimental and machine learning approach 基于实验和机器学习方法的超音速马赫数微射流基压控制

Q1 Chemical Engineering

International Journal of Thermofluids

Pub Date : 2026-01-01 DOI: 10.1016/j.ijft.2025.101533

Abdul Aabid , Sher Afghan Khan , Yasir Javed

Sudden expansion phenomena are prevalent in defense and automotive applications, where flow separation at the blunt base of structures such as fuselages, missiles, and rockets leads to low-pressure recirculation zones, significantly reducing base pressure and increasing drag. This study presents active control methods using microjets to regulate base pressure, employing experimental and machine learning approaches. Experiments were conducted using duct diameters of 16 mm, 18 mm, 22 mm, and 25 mm, level of expansion, the Nozzle pressure ratio ranging from 3 to 11, Mach numbers (1.25, 1.3, 1.48, 1.6, 2.0, and 3.0), and length-to-diameter ratios (10–1) were varied to evaluate their impact on flow evolution and base pressure. Active control was achieved using micro-jets of 0.5 mm radius, positioned at 90° intervals along a pitch circle with a radius of 0.65 times the nozzle exit diameter. Micro-jets significantly increased base pressure under favorable pressure-gradient conditions for the Mach numbers 1.25, 1.3, 1.48, 1.6, and 2.0. At Mach M = 3, the control is ineffective as the NPRs are such that the flow from the nozzle remained over-expanded. Furthermore, machine learning (ML) algorithms were utilized to predict base pressure outcomes and optimize control strategies. These algorithms demonstrated high predictive accuracy, as evidenced by low error rates, indicating their reliability in high-speed flow-control applications. The findings reveal that base pressure is strongly influenced by nozzle pressure ratio, Mach number, L/D ratio, and duct area ratio. The study presents cost-effective, energy-efficient methods to enhance base pressure, offering critical insights into the aerodynamic optimization of high-speed systems. This comprehensive approach integrates experimental techniques and ML–based predictions to achieve optimal results in flow control.

突然膨胀现象在国防和汽车应用中很普遍，其中在诸如机身，导弹和火箭等结构的钝基座处的流动分离导致低压再循环区，显着降低基座压力并增加阻力。本研究采用实验和机器学习方法，提出了利用微射流调节基压的主动控制方法。实验采用直径分别为16 mm、18 mm、22 mm和25 mm、膨胀水平、喷管压力比3 ~ 11、马赫数（1.25、1.3、1.48、1.6、2.0和3.0）和长径比（10-1）进行，以评估它们对流动演化和基压的影响。采用半径为0.5 mm的微射流，沿螺距圆以90°间隔定位，半径为喷嘴出口直径的0.65倍，实现了主动控制。在马赫数为1.25、1.3、1.48、1.6和2.0的有利压力梯度条件下，微射流显著提高了基压。在马赫数M = 3时，控制是无效的，因为npr是这样的，从喷嘴流出的流仍然过度膨胀。此外，利用机器学习（ML）算法预测基压结果并优化控制策略。这些算法具有较高的预测精度和较低的错误率，表明了它们在高速流量控制应用中的可靠性。结果表明，喷管压力比、马赫数、L/D比和风管面积比对基压影响较大。该研究提出了经济高效的方法来提高基压，为高速系统的气动优化提供了重要的见解。这种综合方法将实验技术和基于ml的预测相结合，以实现流量控制的最佳结果。

{"title":"Base pressure control through micro jets at supersonic mach numbers using experimental and machine learning approach","authors":"Abdul Aabid , Sher Afghan Khan , Yasir Javed","doi":"10.1016/j.ijft.2025.101533","DOIUrl":"10.1016/j.ijft.2025.101533","url":null,"abstract":"<div><div>Sudden expansion phenomena are prevalent in defense and automotive applications, where flow separation at the blunt base of structures such as fuselages, missiles, and rockets leads to low-pressure recirculation zones, significantly reducing base pressure and increasing drag. This study presents active control methods using microjets to regulate base pressure, employing experimental and machine learning approaches. Experiments were conducted using duct diameters of 16 mm, 18 mm, 22 mm, and 25 mm, level of expansion, the Nozzle pressure ratio ranging from 3 to 11, Mach numbers (1.25, 1.3, 1.48, 1.6, 2.0, and 3.0), and length-to-diameter ratios (10–1) were varied to evaluate their impact on flow evolution and base pressure. Active control was achieved using micro-jets of 0.5 mm radius, positioned at 90° intervals along a pitch circle with a radius of 0.65 times the nozzle exit diameter. Micro-jets significantly increased base pressure under favorable pressure-gradient conditions for the Mach numbers 1.25, 1.3, 1.48, 1.6, and 2.0. At Mach <em>M</em> = 3, the control is ineffective as the NPRs are such that the flow from the nozzle remained over-expanded. Furthermore, machine learning (ML) algorithms were utilized to predict base pressure outcomes and optimize control strategies. These algorithms demonstrated high predictive accuracy, as evidenced by low error rates, indicating their reliability in high-speed flow-control applications. The findings reveal that base pressure is strongly influenced by nozzle pressure ratio, Mach number, L/D ratio, and duct area ratio. The study presents cost-effective, energy-efficient methods to enhance base pressure, offering critical insights into the aerodynamic optimization of high-speed systems. This comprehensive approach integrates experimental techniques and ML–based predictions to achieve optimal results in flow control.</div></div>","PeriodicalId":36341,"journal":{"name":"International Journal of Thermofluids","volume":"31 ","pages":"Article 101533"},"PeriodicalIF":0.0,"publicationDate":"2026-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145926540","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Hydrogen absorption in metal hydrides: A dual-head tree-based framework for material class classification 金属氢化物中的氢吸收：材料类别分类的双头树框架

Q1 Chemical Engineering

International Journal of Thermofluids

Pub Date : 2026-01-01 DOI: 10.1016/j.ijft.2025.101521

Zaid Allal , Hassan N. Noura , Flavien Vernier , Ola Salman , Khaled Chahine

Metal hydrides represent a highly promising alternative for hydrogen storage due to their favorable absorption properties and relatively moderate requirements in terms of temperature and pressure during the uptake stage. Furthermore, their ability to undergo multiple absorption–desorption cycles without significant degradation makes them particularly well-suited for repeated charge–discharge operations. In this work, we propose a two-headed machine learning framework for the classification of metal hydrides. The first head determines whether a material belongs to the AB family of hydrides. If not, the second head classifies it into one of the remaining categories: MIS, SS, MG, or complex hydrides. This hierarchical modeling strategy, inspired by fault detection and diagnosis in biphasic systems, proved effective in improving classification accuracy. The framework was trained using six tree-based estimators: Random Forest, LightGBM, XGBoost, CatBoost, Gradient Boosting, and Extra Trees. Strong imputation techniques, including the KNN imputer, were also employed to address missing data and ensure robustness. The results demonstrate that for the first head, the Random Forest model achieved an accuracy of 88.6% in identifying the AB family class. For the second head, Gradient Boosting reached 91.61% accuracy, which was further improved to 92.56% after hyperparameter tuning. When compared to previous studies using the same dataset, our framework exhibits significantly stronger performance. Moreover, its predictions were validated through the integration of explainable artificial intelligence (XAI) methods, ensuring both interpretability and reliability. This framework will be further optimized to extend beyond class-level identification toward the precise prediction of hydride composition formulas, with the aim of better supporting the design of materials with enhanced hydrogen absorption capacities.

由于金属氢化物具有良好的吸收特性，并且在吸收阶段对温度和压力的要求相对适中，因此金属氢化物是一种非常有前途的储氢方法。此外，它们经历多次吸收-解吸循环而没有明显降解的能力使它们特别适合重复充放电操作。在这项工作中，我们提出了一个用于金属氢化物分类的双头机器学习框架。第一个头确定材料是否属于AB族氢化物。如果不是，第二个头部将其分类为其余类别之一：MIS， SS， MG或复合氢化物。这种分层建模策略受到双相系统故障检测和诊断的启发，有效地提高了分类精度。该框架使用六个基于树的估计器进行训练：Random Forest， LightGBM, XGBoost, CatBoost， Gradient Boosting和Extra Trees。强输入技术，包括KNN输入器，也被用来解决缺失的数据和确保鲁棒性。结果表明，对于第一个头部，随机森林模型在识别AB族类方面达到了88.6%的准确率。对于第二个头部，Gradient Boosting的准确率达到了91.61%，经过超参数调优后进一步提高到92.56%。与之前使用相同数据集的研究相比，我们的框架表现出明显更强的性能。此外，通过可解释人工智能（XAI）方法的集成验证了其预测，确保了可解释性和可靠性。该框架将进一步优化，以超越类别级别的识别，扩展到氢化物组成公式的精确预测，以更好地支持具有增强吸氢能力的材料的设计。

{"title":"Hydrogen absorption in metal hydrides: A dual-head tree-based framework for material class classification","authors":"Zaid Allal , Hassan N. Noura , Flavien Vernier , Ola Salman , Khaled Chahine","doi":"10.1016/j.ijft.2025.101521","DOIUrl":"10.1016/j.ijft.2025.101521","url":null,"abstract":"<div><div>Metal hydrides represent a highly promising alternative for hydrogen storage due to their favorable absorption properties and relatively moderate requirements in terms of temperature and pressure during the uptake stage. Furthermore, their ability to undergo multiple absorption–desorption cycles without significant degradation makes them particularly well-suited for repeated charge–discharge operations. In this work, we propose a two-headed machine learning framework for the classification of metal hydrides. The first head determines whether a material belongs to the AB family of hydrides. If not, the second head classifies it into one of the remaining categories: MIS, SS, MG, or complex hydrides. This hierarchical modeling strategy, inspired by fault detection and diagnosis in biphasic systems, proved effective in improving classification accuracy. The framework was trained using six tree-based estimators: Random Forest, LightGBM, XGBoost, CatBoost, Gradient Boosting, and Extra Trees. Strong imputation techniques, including the KNN imputer, were also employed to address missing data and ensure robustness. The results demonstrate that for the first head, the Random Forest model achieved an accuracy of 88.6% in identifying the AB family class. For the second head, Gradient Boosting reached 91.61% accuracy, which was further improved to 92.56% after hyperparameter tuning. When compared to previous studies using the same dataset, our framework exhibits significantly stronger performance. Moreover, its predictions were validated through the integration of explainable artificial intelligence (XAI) methods, ensuring both interpretability and reliability. This framework will be further optimized to extend beyond class-level identification toward the precise prediction of hydride composition formulas, with the aim of better supporting the design of materials with enhanced hydrogen absorption capacities.</div></div>","PeriodicalId":36341,"journal":{"name":"International Journal of Thermofluids","volume":"31 ","pages":"Article 101521"},"PeriodicalIF":0.0,"publicationDate":"2026-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145926454","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0