Pub Date : 2026-03-01Epub Date: 2025-07-04DOI: 10.1016/j.jer.2025.06.005
Wei Jiang , Yajing Liu , Haoyang Zhang
The sparsity and irregular data distribution in graph convolutional network (GCN) challenge efficient inference. Previous FPGA-based GCN accelerators have implemented targeted designs for efficient memory access and matrix computation. However, they often overlook the symmetric sparse matrices (SSM) involved in GCN calculations. Specifically, the adjacency matrices in GCNs lead to numerous symmetric sparse matrix multiplications, presenting significant potential for data reuse, including data storage format, memory access strategy, and computation method. To address these challenges, this work proposes a novel GCN accelerator that improves both data storage and access efficiency while enhancing computational performance. First, this paper introduces a compression format compatible with both regular sparse matrices and symmetric sparse matrices, called Special Packet-level Column-only Coordinate-list (SPCOO). SPCOO reduces memory consumption while enhancing data reuse. Additionally, this work proposes a specialized processing element (PE) to handle both symmetric sparse matrix multiplication (SSpMM) and regular sparse matrix multiplication (SpMM) simultaneously. All computations are executed on the unified PE to improve computational efficiency. Finally, SSM-GCN was deployed and validated on the Alveo U50 accelerator card. Experimental results show that the proposed SPCOO format is compatible with symmetric sparse matrices and achieves the lowest storage overhead compared to previous compression formats. SSM-GCN demonstrates an average inference speedup of 230.75 over the CPU and 10.37 over the GPU. In terms of energy efficiency, SSM-GCN achieves an average improvement of 2748.27 compared to the GPU. Furthermore, compared to state-of-the-art FPGA-based GCN accelerators, SSM-GCN improves DSP efficiency by 3.81.
{"title":"SSM-GCN: An FPGA-based efficient GCN accelerator for symmetric sparse matrices","authors":"Wei Jiang , Yajing Liu , Haoyang Zhang","doi":"10.1016/j.jer.2025.06.005","DOIUrl":"10.1016/j.jer.2025.06.005","url":null,"abstract":"<div><div>The sparsity and irregular data distribution in graph convolutional network (GCN) challenge efficient inference. Previous FPGA-based GCN accelerators have implemented targeted designs for efficient memory access and matrix computation. However, they often overlook the symmetric sparse matrices (SSM) involved in GCN calculations. Specifically, the adjacency matrices in GCNs lead to numerous symmetric sparse matrix multiplications, presenting significant potential for data reuse, including data storage format, memory access strategy, and computation method. To address these challenges, this work proposes a novel GCN accelerator that improves both data storage and access efficiency while enhancing computational performance. First, this paper introduces a compression format compatible with both regular sparse matrices and symmetric sparse matrices, called Special Packet-level Column-only Coordinate-list (SPCOO). SPCOO reduces memory consumption while enhancing data reuse. Additionally, this work proposes a specialized processing element (PE) to handle both symmetric sparse matrix multiplication (SSpMM) and regular sparse matrix multiplication (SpMM) simultaneously. All computations are executed on the unified PE to improve computational efficiency. Finally, SSM-GCN was deployed and validated on the Alveo U50 accelerator card. Experimental results show that the proposed SPCOO format is compatible with symmetric sparse matrices and achieves the lowest storage overhead compared to previous compression formats. SSM-GCN demonstrates an average inference speedup of 230.75<span><math><mo>×</mo></math></span> over the CPU and 10.37<span><math><mo>×</mo></math></span> over the GPU. In terms of energy efficiency, SSM-GCN achieves an average improvement of 2748.27<span><math><mo>×</mo></math></span> compared to the GPU. Furthermore, compared to state-of-the-art FPGA-based GCN accelerators, SSM-GCN improves DSP efficiency by 3.81<span><math><mo>×</mo></math></span>.</div></div>","PeriodicalId":48803,"journal":{"name":"Journal of Engineering Research","volume":"14 1","pages":"Pages 712-721"},"PeriodicalIF":2.2,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147453958","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-03-01Epub Date: 2025-10-18DOI: 10.1016/j.jer.2025.10.011
Amir Hosein Riazi , Seyed mohammad ali Boutorabi , Mohammadreza Aboutalebi , Mohsen Ostad Shabani
Foundry is one of the most commonly used and least expensive methods for producing metallic parts. Many defects can occur during the casting of parts; however, most of these defects can be avoided by designing an effective running/gating system. In this regard, the velocity of the melt entering the mold is a critical parameter in forming a flawless and defect-free component. The gating system is responsible for controlling this velocity. If the entry rate is too high, causing the metal to fountain or splash inside the mold cavity, the quality of the casting is compromised. In this paper, methods developed to reduce melt velocity were investigated using simulation to avoid defects such as oxide cracks. These methods include an extended runner, filtering, a fan ingate, and a diffuser to reduce velocity by increasing the cross-sectional area. By applying these methods simultaneously in a single design, the velocity of the melt entering the mold was controlled and maintained at 0.5 m/s. The importance of the critical value of 0.5 m/s was demonstrated through simulation. The findings of this paper could help foundry engineers design gating systems that produce high-quality, defect-free steel castings, such as those used in marine and automotive applications.
{"title":"Applicable methods for reducing melt velocity in gating systems: Simulation and comparative evaluation of usable techniques","authors":"Amir Hosein Riazi , Seyed mohammad ali Boutorabi , Mohammadreza Aboutalebi , Mohsen Ostad Shabani","doi":"10.1016/j.jer.2025.10.011","DOIUrl":"10.1016/j.jer.2025.10.011","url":null,"abstract":"<div><div>Foundry is one of the most commonly used and least expensive methods for producing metallic parts. Many defects can occur during the casting of parts; however, most of these defects can be avoided by designing an effective running/gating system. In this regard, the velocity of the melt entering the mold is a critical parameter in forming a flawless and defect-free component. The gating system is responsible for controlling this velocity. If the entry rate is too high, causing the metal to fountain or splash inside the mold cavity, the quality of the casting is compromised. In this paper, methods developed to reduce melt velocity were investigated using simulation to avoid defects such as oxide cracks. These methods include an extended runner, filtering, a fan ingate, and a diffuser to reduce velocity by increasing the cross-sectional area. By applying these methods simultaneously in a single design, the velocity of the melt entering the mold was controlled and maintained at 0.5 m/s. The importance of the critical value of 0.5 m/s was demonstrated through simulation. The findings of this paper could help foundry engineers design gating systems that produce high-quality, defect-free steel castings, such as those used in marine and automotive applications.</div></div>","PeriodicalId":48803,"journal":{"name":"Journal of Engineering Research","volume":"14 1","pages":"Pages 606-613"},"PeriodicalIF":2.2,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147453978","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-03-01Epub Date: 2025-05-15DOI: 10.1016/j.jer.2025.05.003
Chia Wei Khor, Nur Syazreen Ahmad
Bluetooth Low Energy (BLE) is a promising technology for indoor localization due to its strong signal penetration and low power consumption, especially in environments where Wi-Fi signals are unreliable. BLE beacons are compact and battery-powered, enabling flexible deployment without the need for wired infrastructure. In this study, we propose a low-cost indoor localization approach based on a Temporal Convolutional Network (TCN), and perform a comparative analysis using two signal filtering techniques: a median filter (MF) and a Recursive Moving Average Filter (RMAF). The method was evaluated in a cluttered indoor environment measuring 10.7 m × 6.3 m using only three BLE beacons. Its performance was compared with state-of-the-art machine learning methods, including weighted k-nearest neighbors, support vector machine, and bagged trees. Experimental results demonstrate that the proposed TCN consistently achieved high accuracy, with at least 98 % accuracy across all input conditions (raw RSSI, MF, and RMAF), showcasing robust baseline reliability. However, the F1-score, which provides a more balanced measure of precision and recall, revealed the most significant improvements in performance. Compared to the second-best machine learning method in each condition, the TCN achieved F1-score improvements of 27.1 % with raw input, 22.9 % with MF, and 9.7 % with RMAF. The best overall result was achieved with the TCN + RMAF configuration, which reached an outstanding accuracy of 99 % and an F1-score of 84.8 %. These results emphasize that while all configurations of the proposed method perform reliably in terms of accuracy, the use of filtering-particularly RMAF further enhances the model’s ability to correctly detect user positions, making it a robust and practical solution for BLE-based indoor localization.
{"title":"BLE-based indoor localization with temporal convolutional network","authors":"Chia Wei Khor, Nur Syazreen Ahmad","doi":"10.1016/j.jer.2025.05.003","DOIUrl":"10.1016/j.jer.2025.05.003","url":null,"abstract":"<div><div>Bluetooth Low Energy (BLE) is a promising technology for indoor localization due to its strong signal penetration and low power consumption, especially in environments where Wi-Fi signals are unreliable. BLE beacons are compact and battery-powered, enabling flexible deployment without the need for wired infrastructure. In this study, we propose a low-cost indoor localization approach based on a Temporal Convolutional Network (TCN), and perform a comparative analysis using two signal filtering techniques: a median filter (MF) and a Recursive Moving Average Filter (RMAF). The method was evaluated in a cluttered indoor environment measuring 10.7 m × 6.3 m using only three BLE beacons. Its performance was compared with state-of-the-art machine learning methods, including weighted k-nearest neighbors, support vector machine, and bagged trees. Experimental results demonstrate that the proposed TCN consistently achieved high accuracy, with at least 98 % accuracy across all input conditions (raw RSSI, MF, and RMAF), showcasing robust baseline reliability. However, the F1-score, which provides a more balanced measure of precision and recall, revealed the most significant improvements in performance. Compared to the second-best machine learning method in each condition, the TCN achieved F1-score improvements of 27.1 % with raw input, 22.9 % with MF, and 9.7 % with RMAF. The best overall result was achieved with the TCN + RMAF configuration, which reached an outstanding accuracy of 99 % and an F1-score of 84.8 %. These results emphasize that while all configurations of the proposed method perform reliably in terms of accuracy, the use of filtering-particularly RMAF further enhances the model’s ability to correctly detect user positions, making it a robust and practical solution for BLE-based indoor localization.</div></div>","PeriodicalId":48803,"journal":{"name":"Journal of Engineering Research","volume":"14 1","pages":"Pages 966-976"},"PeriodicalIF":2.2,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147454058","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-03-01Epub Date: 2025-09-19DOI: 10.1016/j.jer.2025.09.009
M.M. Mosallem , Zienab A. Ahmed
<div><div>Flow around bluff bodies induces vortex shedding, leading to fluctuating forces and vortex-induced vibrations. In many engineering fields, this vibration response can lead to structure stress and even damage. This study numerically investigates the effect of a small control rod with diameter ratios <span><math><mrow><mi>d</mi><mo>/</mo><mi>D</mi><mo>=</mo><mn>0.06</mn></mrow></math></span> and 0.07 on the wake and aerodynamic forces of two side-by-side circular cylinders. The cylinders are spaced at a center-to-center spacing ratio (<span><math><mrow><mi>T</mi><mo>/</mo><mi>D</mi><mspace></mspace></mrow></math></span>) of 1.25, with a Reynolds number, <span><math><mrow><msub><mrow><mi>Re</mi></mrow><mrow><mi>D</mi></mrow></msub><mo>=</mo><mn>1.2</mn><mo>×</mo><msup><mrow><mn>10</mn></mrow><mrow><mn>4</mn></mrow></msup></mrow></math></span>. The rod is placed in the center line behind the cylinders. The unsteady, incompressible two-dimensional turbulent flow is modeled using Reynolds-Averaged Navier-Stokes equations with the <span><math><mrow><mfenced><mrow><mtext>SST</mtext></mrow></mfenced></mrow><mi>k</mi><mo>−</mo><mi>ω</mi></math></span> turbulence model. Numerical simulations are conducted using Ansys Fluent software. In the uncontrolled case, the flow exhibits strong flip-flopping behavior and large fluctuations in drag and lift, with distinct shedding frequencies of <span><math><mrow><mi>St</mi><mo>=</mo><mn>0.173</mn></mrow></math></span> for the upper cylinder and <span><math><mrow><mi>St</mi><mo>=</mo><mn>0.128</mn></mrow></math></span> for the lower cylinder. In addition, a low-frequency flip-flopping mode appears at <span><math><mrow><mi>St</mi><mo>=</mo><mn>0.013</mn></mrow></math></span>, indicates wake instability. For <span><math><mrow><mi>d</mi><mo>/</mo><mi>D</mi><mo>=</mo><mn>0.06</mn></mrow></math></span>, the drag coefficient fluctuation amplitude decreases by 38 % for the upper cylinder and 31 % for the lower cylinder, while the mean drag decreases by 10.2 % and 6.5 %, respectively. The corresponding lift fluctuation ranges are reduced by 60 % and 18 %, and the root mean square lift coefficient for both cylinders is almost doubled. Strouhal numbers of both cylinders converge to a common value of <span><math><mrow><mi>St</mi><mo>=</mo><mn>0.13</mn></mrow></math></span>. and the flip-flopping Strouhal number shifts to <span><math><mrow><mi>St</mi><mo>=</mo><mn>0.022</mn></mrow></math></span>, indicating that the rod modifies the stability of the flip-flopping motion. For <span><math><mrow><mi>d</mi><mo>/</mo><mi>D</mi><mo>=</mo><mn>0.07</mn></mrow></math></span>, the vortex shedding is significantly suppressed, and stable recirculation bubbles are formed behind the cylinders, eliminating the flip-flopping phenomenon. The drag and lift coefficients exhibit periodic fluctuations. These fluctuations are reduced by approximately 90 % compared to the uncontrolled case. The mean drag coefficient decreases by 9.7 % for the
钝体周围的流动引起旋涡脱落，导致波动力和旋涡引起的振动。在许多工程领域中，这种振动响应会导致结构受力甚至损伤。数值研究了直径比为d/ d =0.06和0.07的小控制棒对两个并排圆柱体尾流和气动力的影响。圆柱体中心间距比（T/D）为1.25，雷诺数ReD=1.2×104。杆被放置在汽缸后面的中心线上。采用reynolds - average Navier-Stokes方程和SSTk−ω湍流模型对非定常不可压缩二维湍流进行了建模。采用Ansys Fluent软件进行了数值模拟。在不受控制的情况下，流动表现出强烈的翻转行为，阻力和升力波动较大，有明显的脱落频率，上缸St=0.173，下缸St=0.128。此外，在St=0.013处出现低频倒转模式，表明尾流不稳定。当d/ d =0.06时，上缸阻力系数波动幅值减小38%，下缸阻力系数波动幅值减小31%，平均阻力分别减小10.2%和6.5%。相应的升力波动范围分别减小了60%和18%，两个气缸的均方升力系数几乎增加了一倍。两个柱体的斯特劳哈尔数收敛于一个公共值St=0.13。翻转斯特罗哈尔数变为St=0.022，说明杆改变了翻转运动的稳定性。当d/ d =0.07时，旋涡脱落得到明显抑制，气缸后形成稳定的再循环气泡，消除了翻转现象。阻力系数和升力系数呈现周期性波动。与不受控制的情况相比，这些波动减少了约90%。上气缸的平均阻力系数降低了9.7%，下气缸的平均阻力系数降低了7%。升力系数的均方根大约是两个气缸的两倍。光谱分析证实，次级圆柱的相互作用频率完全消除。结果表明，d/ d =0.07的控制棒是一种有效的被动流动控制策略，可促进节能和提高结构安全性。
{"title":"Suppression of flip-flopping wake instability behind two side-by-side cylinders using a passive control rod at subcritical Reynolds number","authors":"M.M. Mosallem , Zienab A. Ahmed","doi":"10.1016/j.jer.2025.09.009","DOIUrl":"10.1016/j.jer.2025.09.009","url":null,"abstract":"<div><div>Flow around bluff bodies induces vortex shedding, leading to fluctuating forces and vortex-induced vibrations. In many engineering fields, this vibration response can lead to structure stress and even damage. This study numerically investigates the effect of a small control rod with diameter ratios <span><math><mrow><mi>d</mi><mo>/</mo><mi>D</mi><mo>=</mo><mn>0.06</mn></mrow></math></span> and 0.07 on the wake and aerodynamic forces of two side-by-side circular cylinders. The cylinders are spaced at a center-to-center spacing ratio (<span><math><mrow><mi>T</mi><mo>/</mo><mi>D</mi><mspace></mspace></mrow></math></span>) of 1.25, with a Reynolds number, <span><math><mrow><msub><mrow><mi>Re</mi></mrow><mrow><mi>D</mi></mrow></msub><mo>=</mo><mn>1.2</mn><mo>×</mo><msup><mrow><mn>10</mn></mrow><mrow><mn>4</mn></mrow></msup></mrow></math></span>. The rod is placed in the center line behind the cylinders. The unsteady, incompressible two-dimensional turbulent flow is modeled using Reynolds-Averaged Navier-Stokes equations with the <span><math><mrow><mfenced><mrow><mtext>SST</mtext></mrow></mfenced></mrow><mi>k</mi><mo>−</mo><mi>ω</mi></math></span> turbulence model. Numerical simulations are conducted using Ansys Fluent software. In the uncontrolled case, the flow exhibits strong flip-flopping behavior and large fluctuations in drag and lift, with distinct shedding frequencies of <span><math><mrow><mi>St</mi><mo>=</mo><mn>0.173</mn></mrow></math></span> for the upper cylinder and <span><math><mrow><mi>St</mi><mo>=</mo><mn>0.128</mn></mrow></math></span> for the lower cylinder. In addition, a low-frequency flip-flopping mode appears at <span><math><mrow><mi>St</mi><mo>=</mo><mn>0.013</mn></mrow></math></span>, indicates wake instability. For <span><math><mrow><mi>d</mi><mo>/</mo><mi>D</mi><mo>=</mo><mn>0.06</mn></mrow></math></span>, the drag coefficient fluctuation amplitude decreases by 38 % for the upper cylinder and 31 % for the lower cylinder, while the mean drag decreases by 10.2 % and 6.5 %, respectively. The corresponding lift fluctuation ranges are reduced by 60 % and 18 %, and the root mean square lift coefficient for both cylinders is almost doubled. Strouhal numbers of both cylinders converge to a common value of <span><math><mrow><mi>St</mi><mo>=</mo><mn>0.13</mn></mrow></math></span>. and the flip-flopping Strouhal number shifts to <span><math><mrow><mi>St</mi><mo>=</mo><mn>0.022</mn></mrow></math></span>, indicating that the rod modifies the stability of the flip-flopping motion. For <span><math><mrow><mi>d</mi><mo>/</mo><mi>D</mi><mo>=</mo><mn>0.07</mn></mrow></math></span>, the vortex shedding is significantly suppressed, and stable recirculation bubbles are formed behind the cylinders, eliminating the flip-flopping phenomenon. The drag and lift coefficients exhibit periodic fluctuations. These fluctuations are reduced by approximately 90 % compared to the uncontrolled case. The mean drag coefficient decreases by 9.7 % for the ","PeriodicalId":48803,"journal":{"name":"Journal of Engineering Research","volume":"14 1","pages":"Pages 491-502"},"PeriodicalIF":2.2,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147454109","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-03-01Epub Date: 2024-05-04DOI: 10.1016/j.jer.2024.04.023
Farhan Lafta Rashid , Hakim S. Aljibori , Hayder I. Mohammed , Arman Ameen , Shabbir Ahmad , Mohamed Bechir Ben Hamida , Ameer H. Al-Rubaye
This study addresses challenges in enhancing the thermal efficiency of parabolic solar collector energy systems using hybrid nanofluids, focusing on issues like nanoparticle clumping and decreased effectiveness. The objective is to optimize design parameters for improved energy absorption and efficiency by evaluating the thermal performance of hybrid nanofluids through theoretical and experimental analyses, aiming to enhance the overall efficiency of solar collector systems. The thermal performance of solar collector systems was evaluated by conducting numerical simulations and experimental analyses to investigate the effects of various nanoparticle compositions and concentrations. The findings suggest that hybrid nanofluids, specifically Au-Cu/EO and Cu-Al2O3, demonstrate enhanced heat transfer properties in comparison to conventional fluids, resulting in efficiency enhancements ranging from 22.44% to 35.01%. Compared to water, Al2O3/water (0.04%), and MWCNT/water (0.04%), the solar collector's thermal efficiency improves by 197.1%, 69.2%, and 6.1%, respectively. Furthermore, the research emphasizes the potential advantages of integrating precise nanoparticle concentrations to improve thermal efficiency while reducing the adverse effects of friction factors. The results emphasize the significance of tackling primary obstacles such as the clumping together of nanoparticles, heightened energy demands for pumping, and elevated expenses in the manufacture of hybrid nanofluids. The study enhances the advancement of cost-effective and efficient solar collector systems by identifying limits and suggesting alternative solutions. The research highlights the necessity for additional investigation into innovative combinations of nanomaterials, fine-tuning of fluid characteristics, and thorough evaluations of long-term stability in order to forward the practical use of hybrid nanofluids in solar energy systems.
{"title":"Recent advances and developments of the application of hybrid nanofluids in parabolic solar collector energy systems and guidelines for future prospects","authors":"Farhan Lafta Rashid , Hakim S. Aljibori , Hayder I. Mohammed , Arman Ameen , Shabbir Ahmad , Mohamed Bechir Ben Hamida , Ameer H. Al-Rubaye","doi":"10.1016/j.jer.2024.04.023","DOIUrl":"10.1016/j.jer.2024.04.023","url":null,"abstract":"<div><div>This study addresses challenges in enhancing the thermal efficiency of parabolic solar collector energy systems using hybrid nanofluids, focusing on issues like nanoparticle clumping and decreased effectiveness. The objective is to optimize design parameters for improved energy absorption and efficiency by evaluating the thermal performance of hybrid nanofluids through theoretical and experimental analyses, aiming to enhance the overall efficiency of solar collector systems. The thermal performance of solar collector systems was evaluated by conducting numerical simulations and experimental analyses to investigate the effects of various nanoparticle compositions and concentrations. The findings suggest that hybrid nanofluids, specifically Au-Cu/EO and Cu-Al<sub>2</sub>O<sub>3</sub>, demonstrate enhanced heat transfer properties in comparison to conventional fluids, resulting in efficiency enhancements ranging from 22.44% to 35.01%. Compared to water, Al<sub>2</sub>O<sub>3</sub>/water (0.04%), and MWCNT/water (0.04%), the solar collector's thermal efficiency improves by 197.1%, 69.2%, and 6.1%, respectively. Furthermore, the research emphasizes the potential advantages of integrating precise nanoparticle concentrations to improve thermal efficiency while reducing the adverse effects of friction factors. The results emphasize the significance of tackling primary obstacles such as the clumping together of nanoparticles, heightened energy demands for pumping, and elevated expenses in the manufacture of hybrid nanofluids. The study enhances the advancement of cost-effective and efficient solar collector systems by identifying limits and suggesting alternative solutions. The research highlights the necessity for additional investigation into innovative combinations of nanomaterials, fine-tuning of fluid characteristics, and thorough evaluations of long-term stability in order to forward the practical use of hybrid nanofluids in solar energy systems.</div></div>","PeriodicalId":48803,"journal":{"name":"Journal of Engineering Research","volume":"14 1","pages":"Pages 246-265"},"PeriodicalIF":2.2,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141027003","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-03-01Epub Date: 2025-05-26DOI: 10.1016/j.jer.2025.05.009
Salah Almudhhi, Haitham M.S. Lababidi, Ali A. Garrouch
The viscosity of heavy oil is a critical parameter in designing surface and well production equipment, reservoir simulations, and field development projects. However, direct measurement of heavy crude oil viscosity is often costly and requires representative fluid samples, which can be challenging to obtain. This strong dependence of viscosity on temperature further complicates the process. Consequently, the industry relies on predictive correlations to estimate crude oil viscosity. These correlations, however, have significant limitations, as they tend to oversimplify complex relationships, reducing their practical utility. These models often fall short in capturing the full complexity of dependencies, especially those arising from compositional variations in heavy oil. This in return, limits their accuracy and applicability across a wider range of pressures, temperatures, and fluid compositions. Over the past two decades, machine learning (ML) has gained substantial interest in modeling heavy oil viscosity. However, the results from ML models often show high variability and inherent bias. This work develops and compares the performance of two ML models, specifically, backpropagation neural networks (BPNN) and general regression neural networks (GRNN). The proposed models were trained to predict the dead heavy-oil viscosity using a dataset consisting of 371 records and six input variables: temperature, density, , and mole fractions, and their respective molecular weights. Descriptive statistics and correlation analysis identified temperature and density as the most influential factors in predicting heavy oil viscosity. To enhance model performance, log-transformation of viscosity and temperature were applied. The BPNN, with 15 neurons in the hidden layer, achieved an average absolute relative error (AARE) of 1.27 % on the blind testing dataset, significantly outperformed the GRNN, which had an AARE of 15 %. Despite this disparity in performance, both neural networks significantly outperformed several empirical correlations reported in the literature by a considerable margin. The BPNN model demonstrated superior prediction accuracy due to several key factors. Its iterative weight optimization via backpropagation allows minimizing errors and improving predictions effectively. The hidden layers of the BPNN automatically learn complex data patterns, making it particularly effective with large, high-dimensional datasets. In contrast, the GRNN, while computationally efficient, faces challenges with memory limitations and generalization. Additionally, the BPNN allows greater flexibility through fine-tuning hyperparameters such as learning rate and activation functions, making it the preferred model for predicting heavy oil viscosity.
{"title":"Application of machine learning for modeling heavy oil viscosity","authors":"Salah Almudhhi, Haitham M.S. Lababidi, Ali A. Garrouch","doi":"10.1016/j.jer.2025.05.009","DOIUrl":"10.1016/j.jer.2025.05.009","url":null,"abstract":"<div><div>The viscosity of heavy oil is a critical parameter in designing surface and well production equipment, reservoir simulations, and field development projects. However, direct measurement of heavy crude oil viscosity is often costly and requires representative fluid samples, which can be challenging to obtain. This strong dependence of viscosity on temperature further complicates the process. Consequently, the industry relies on predictive correlations to estimate crude oil viscosity. These correlations, however, have significant limitations, as they tend to oversimplify complex relationships, reducing their practical utility. These models often fall short in capturing the full complexity of dependencies, especially those arising from compositional variations in heavy oil. This in return, limits their accuracy and applicability across a wider range of pressures, temperatures, and fluid compositions. Over the past two decades, machine learning (ML) has gained substantial interest in modeling heavy oil viscosity. However, the results from ML models often show high variability and inherent bias. This work develops and compares the performance of two ML models, specifically, backpropagation neural networks (BPNN) and general regression neural networks (GRNN). The proposed models were trained to predict the dead heavy-oil viscosity using a dataset consisting of 371 records and six input variables: temperature, density, <span><math><msub><mrow><mi>C</mi></mrow><mrow><mn>7</mn><mo>+</mo></mrow></msub></math></span>, and <span><math><msub><mrow><mi>C</mi></mrow><mrow><mn>26</mn></mrow></msub></math></span> mole fractions, and their respective molecular weights. Descriptive statistics and correlation analysis identified temperature and density as the most influential factors in predicting heavy oil viscosity. To enhance model performance, log-transformation of viscosity and temperature were applied. The BPNN, with 15 neurons in the hidden layer, achieved an average absolute relative error (AARE) of 1.27 % on the blind testing dataset, significantly outperformed the GRNN, which had an AARE of 15 %. Despite this disparity in performance, both neural networks significantly outperformed several empirical correlations reported in the literature by a considerable margin. The BPNN model demonstrated superior prediction accuracy due to several key factors. Its iterative weight optimization via backpropagation allows minimizing errors and improving predictions effectively. The hidden layers of the BPNN automatically learn complex data patterns, making it particularly effective with large, high-dimensional datasets. In contrast, the GRNN, while computationally efficient, faces challenges with memory limitations and generalization. Additionally, the BPNN allows greater flexibility through fine-tuning hyperparameters such as learning rate and activation functions, making it the preferred model for predicting heavy oil viscosity.</div></div>","PeriodicalId":48803,"journal":{"name":"Journal of Engineering Research","volume":"14 1","pages":"Pages 1204-1214"},"PeriodicalIF":2.2,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147454236","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-03-01Epub Date: 2025-10-10DOI: 10.1016/j.jer.2025.10.003
Md Mehedi Hasan Emon , Tahsina Khan
This study aims to investigate the mediating role of green logistics in enhancing green supply chain performance in the context of Bangladesh, focusing on the relationships among green procurement, eco-friendly transportation, sustainable manufacturing, and reverse logistics. A quantitative research methodology was employed, utilizing a structured questionnaire distributed to 425 professionals in various industries. A total of 276 valid responses were analyzed using Smart PLS to assess the relationships and mediating effects of green logistics. The findings indicate that green procurement and sustainable manufacturing significantly enhance green supply chain performance. In contrast, eco-friendly transportation and reverse logistics do not exhibit direct effects; however, their impact is mediated by green logistics. The study reveals that green logistics serves as a crucial enabler for integrating sustainability practices within supply chains. The results provide valuable insights for businesses and policymakers in Bangladesh, highlighting the importance of investing in green logistics strategies to improve operational efficiency and sustainability outcomes. By fostering sustainable supply chain practices, organizations can contribute to environmental preservation and social responsibility in the Bangladeshi context. This research fills a critical gap in the literature by examining the mediating role of green logistics in a developing country, offering a comprehensive framework that integrates multiple sustainability dimensions. The study is limited by its focus on a single country and the use of convenience sampling, which may impact the generalizability of the findings.
{"title":"Exploring the mediating role of green logistics in enhancing green supply chain performance: Evidence from Bangladesh","authors":"Md Mehedi Hasan Emon , Tahsina Khan","doi":"10.1016/j.jer.2025.10.003","DOIUrl":"10.1016/j.jer.2025.10.003","url":null,"abstract":"<div><div>This study aims to investigate the mediating role of green logistics in enhancing green supply chain performance in the context of Bangladesh, focusing on the relationships among green procurement, eco-friendly transportation, sustainable manufacturing, and reverse logistics. A quantitative research methodology was employed, utilizing a structured questionnaire distributed to 425 professionals in various industries. A total of 276 valid responses were analyzed using Smart PLS to assess the relationships and mediating effects of green logistics. The findings indicate that green procurement and sustainable manufacturing significantly enhance green supply chain performance. In contrast, eco-friendly transportation and reverse logistics do not exhibit direct effects; however, their impact is mediated by green logistics. The study reveals that green logistics serves as a crucial enabler for integrating sustainability practices within supply chains. The results provide valuable insights for businesses and policymakers in Bangladesh, highlighting the importance of investing in green logistics strategies to improve operational efficiency and sustainability outcomes. By fostering sustainable supply chain practices, organizations can contribute to environmental preservation and social responsibility in the Bangladeshi context. This research fills a critical gap in the literature by examining the mediating role of green logistics in a developing country, offering a comprehensive framework that integrates multiple sustainability dimensions. The study is limited by its focus on a single country and the use of convenience sampling, which may impact the generalizability of the findings.</div></div>","PeriodicalId":48803,"journal":{"name":"Journal of Engineering Research","volume":"14 1","pages":"Pages 1116-1129"},"PeriodicalIF":2.2,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147454238","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-03-01Epub Date: 2025-09-24DOI: 10.1016/j.jer.2025.09.012
Chia-Nan Wang, Thanh-Tam Truong, Tu-Uyen Doan
Vietnam’s strategic location in the Asia-Pacific region and its long coastline provide strong potential for port operations and international freight forwarding. Nevertheless, the logistics industry faces declining trade volumes, rising costs, and operational pressures, complicating financial performance evaluation. This study develops a hybrid Multi-Criteria Decision-Making (MCDM) model integrating Modigliani-Miller and Regret theories with STPF Entropy, STPF CRITIC, and R-VIKOR to assess and rank the financial performance of 13 listed Vietnamese logistics firms from 2020 to 2024. Key financial criteria include EBIT, corporate tax, marginal dividend tax, cost of debt and equity, debt size, and tax rate, with EBIT carrying the highest weight (23.5 %). Results show that SGP ranks highest (Qi = 0.001), while MVN, GMD, and PHP rank lowest. Sensitivity and comparative analyses with six other MCDM methods (VIKOR, TOPSIS, MOORA, MICRA, COPRAS, CODAS) confirm the proposed model's robustness, stability, and reliability. Minor fluctuations observed in mid-range firms highlight that operational efficiency and internal capacity are more influential than behavioral factors such as regret avoidance. The standardized Q scale facilitates consistent comparison under uncertainty, providing managers and investors with actionable insights. The model offers a comprehensive and reliable framework for evaluating financial performance in Vietnam’s logistics sector, supporting informed strategic and investment decisions.
{"title":"Evaluating port and logistics companies using integrated fuzzy MCDM with MM–regret framework","authors":"Chia-Nan Wang, Thanh-Tam Truong, Tu-Uyen Doan","doi":"10.1016/j.jer.2025.09.012","DOIUrl":"10.1016/j.jer.2025.09.012","url":null,"abstract":"<div><div>Vietnam’s strategic location in the Asia-Pacific region and its long coastline provide strong potential for port operations and international freight forwarding. Nevertheless, the logistics industry faces declining trade volumes, rising costs, and operational pressures, complicating financial performance evaluation. This study develops a hybrid Multi-Criteria Decision-Making (MCDM) model integrating Modigliani-Miller and Regret theories with STPF Entropy, STPF CRITIC, and R-VIKOR to assess and rank the financial performance of 13 listed Vietnamese logistics firms from 2020 to 2024. Key financial criteria include EBIT, corporate tax, marginal dividend tax, cost of debt and equity, debt size, and tax rate, with EBIT carrying the highest weight (23.5 %). Results show that SGP ranks highest (Q<sub>i</sub> = 0.001), while MVN, GMD, and PHP rank lowest. Sensitivity and comparative analyses with six other MCDM methods (VIKOR, TOPSIS, MOORA, MICRA, COPRAS, CODAS) confirm the proposed model's robustness, stability, and reliability. Minor fluctuations observed in mid-range firms highlight that operational efficiency and internal capacity are more influential than behavioral factors such as regret avoidance. The standardized Q scale facilitates consistent comparison under uncertainty, providing managers and investors with actionable insights. The model offers a comprehensive and reliable framework for evaluating financial performance in Vietnam’s logistics sector, supporting informed strategic and investment decisions.</div></div>","PeriodicalId":48803,"journal":{"name":"Journal of Engineering Research","volume":"14 1","pages":"Pages 1088-1103"},"PeriodicalIF":2.2,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147453730","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-03-01Epub Date: 2025-10-10DOI: 10.1016/j.jer.2025.10.002
Qiuming Luo, Kunzhong Wu
This paper presents an image acquisition system based on a Contact Image Sensor (CIS) for industrial print quality inspection and designs a specialized System-on-Chip (SoC) integrating control, image data acquisition, and transmission functionalities, leveraging the open-source Reduced Instruction Set Computer-V (RISC-V) architecture. The system combines the high integration and low-cost advantages of the CIS module with the flexibility and customizability of the RISC-V architecture, providing an efficient, low-cost, and real-time image acquisition solution. By precisely controlling a dual-axis motion platform, the system drives the CIS to perform real-time scanning of printed outputs, while transmitting the acquired data to a host computer for defect detection. A data mapping algorithm, together with a real-time data transfer mechanism and system framework based on multiple Direct Memory Access (DMA) buffering and interrupts, improves storage space utilization while ensuring data integrity, reliability, and real-time performance. Experimental results demonstrate that the system can efficiently and reliably acquire and transmit high-resolution images, meeting both image acquisition and real-time system requirements while significantly reducing hardware costs. In tests, the proposed solution improves storage utilization by approximately 25.7% and operational speed by 26.5% compared to traditional single-buffer serial transmission, with further gains as the size of the scanned images increases. This research not only provides an innovative, cost-effective solution for industrial print quality inspection but also enriches the RISC-V ecosystem and expands its application in industrial control domains.
{"title":"A RISC-V based image acquisition system using CIS for industrial print quality inspection","authors":"Qiuming Luo, Kunzhong Wu","doi":"10.1016/j.jer.2025.10.002","DOIUrl":"10.1016/j.jer.2025.10.002","url":null,"abstract":"<div><div>This paper presents an image acquisition system based on a Contact Image Sensor (CIS) for industrial print quality inspection and designs a specialized System-on-Chip (SoC) integrating control, image data acquisition, and transmission functionalities, leveraging the open-source Reduced Instruction Set Computer-V (RISC-V) architecture. The system combines the high integration and low-cost advantages of the CIS module with the flexibility and customizability of the RISC-V architecture, providing an efficient, low-cost, and real-time image acquisition solution. By precisely controlling a dual-axis motion platform, the system drives the CIS to perform real-time scanning of printed outputs, while transmitting the acquired data to a host computer for defect detection. A data mapping algorithm, together with a real-time data transfer mechanism and system framework based on multiple Direct Memory Access (DMA) buffering and interrupts, improves storage space utilization while ensuring data integrity, reliability, and real-time performance. Experimental results demonstrate that the system can efficiently and reliably acquire and transmit high-resolution images, meeting both image acquisition and real-time system requirements while significantly reducing hardware costs. In tests, the proposed solution improves storage utilization by approximately 25.7% and operational speed by 26.5% compared to traditional single-buffer serial transmission, with further gains as the size of the scanned images increases. This research not only provides an innovative, cost-effective solution for industrial print quality inspection but also enriches the RISC-V ecosystem and expands its application in industrial control domains.</div></div>","PeriodicalId":48803,"journal":{"name":"Journal of Engineering Research","volume":"14 1","pages":"Pages 1104-1115"},"PeriodicalIF":2.2,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147453735","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-03-01Epub Date: 2025-05-27DOI: 10.1016/j.jer.2025.05.010
Mohamed Hammad , Samia A. Chelloug , Samah AlShathri , Ahmed A. Abd El-Latif
Scene classification is a fundamental challenge in computer vision. Recognizing the complexity of this task is the aim of our study that addresses the need for accurate and robust scene classification by leveraging the capabilities of two widely recognized databases. The motivation behind this research lies in enhancing the accuracy and efficiency of scene classification systems. Therefore, our primary goal is to explore and implement a comprehensive methodology that combines transfer learning and automated machine learning techniques to achieve superior classification results. Our approach commences with a meticulous data loading process, followed by preprocessing steps to ensure the optimal representation of information. We have conducted class distribution analysis to understand the dataset's nuances. Subsequently, we have employed two key models: MobileNetV2 for transfer learning and a custom convolutional neural network (CNN) model featuring batch normalization. This diverse methodology aims to capture intricate patterns within the data. An innovative step of our approach involves employing Tree-based Pipeline Optimization Tool (TPOT), an automated machine learning tool, for model selection and hyperparameter tuning. The results underscore the effectiveness of our methodology, achieving impressive classification accuracy across diverse scenes. This research contributes valuable insights into the integration of transfer learning and automated machine learning for robust and accurate scene recognition, offering a comprehensive approach to address the complexities of scene classification.
{"title":"Optimizing scene classification: A robust approach with transfer learning and automated machine learning integration","authors":"Mohamed Hammad , Samia A. Chelloug , Samah AlShathri , Ahmed A. Abd El-Latif","doi":"10.1016/j.jer.2025.05.010","DOIUrl":"10.1016/j.jer.2025.05.010","url":null,"abstract":"<div><div>Scene classification is a fundamental challenge in computer vision. Recognizing the complexity of this task is the aim of our study that addresses the need for accurate and robust scene classification by leveraging the capabilities of two widely recognized databases. The motivation behind this research lies in enhancing the accuracy and efficiency of scene classification systems. Therefore, our primary goal is to explore and implement a comprehensive methodology that combines transfer learning and automated machine learning techniques to achieve superior classification results. Our approach commences with a meticulous data loading process, followed by preprocessing steps to ensure the optimal representation of information. We have conducted class distribution analysis to understand the dataset's nuances. Subsequently, we have employed two key models: MobileNetV2 for transfer learning and a custom convolutional neural network (CNN) model featuring batch normalization. This diverse methodology aims to capture intricate patterns within the data. An innovative step of our approach involves employing Tree-based Pipeline Optimization Tool (TPOT), an automated machine learning tool, for model selection and hyperparameter tuning. The results underscore the effectiveness of our methodology, achieving impressive classification accuracy across diverse scenes. This research contributes valuable insights into the integration of transfer learning and automated machine learning for robust and accurate scene recognition, offering a comprehensive approach to address the complexities of scene classification.</div></div>","PeriodicalId":48803,"journal":{"name":"Journal of Engineering Research","volume":"14 1","pages":"Pages 688-711"},"PeriodicalIF":2.2,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147453804","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
扫码关注我们
求助内容：
应助结果提醒方式：
京公网安备 11010802042870号