Pub Date : 2026-01-19DOI: 10.1016/j.asej.2026.103985
G Divya Deepak, Pavan Hiremath, Subraya Krishna Bhat
Semantic segmentation is a critical perception task in autonomous vehicles, enabling pixel-wise classification of road scenes. In this study, we propose a systematic optimization of DeepLabV3+ semantic segmentation model using Taguchi Design of Experiments (DoE) technique to enhance its performance for real-time deployment in autonomous driving. We explore the influence of key hyperparameters. solver type (Adam, RMSProp, SGDM), learning rate (10−5, 10−4, 10−3) batch size (1, 2, 3), and L2 regularization (10−5, 10−4, 10−3), across three backbone networks: ResNet-18, ResNet-50, and MobileNetV2. Experiments were conducted on the Cambridge-driving Labeled Video Database (CamVid), a widely used benchmark for road scene understanding. The DoE approach efficiently reduced the number of training configurations while maximizing segmentation performance. The best-performing model, DeepLabV3+ with a ResNet-50 backbone, achieved a Mean Intersection over Union (IoU) of 76.23%, surpassing recent approaches. The proposed framework offers a practical strategy for deploying semantic segmentation models in autonomous vehicle systems.
语义分割是自动驾驶汽车的一项关键感知任务，可以实现道路场景的逐像素分类。本研究采用田口实验设计（Taguchi Design of Experiments, DoE）技术对DeepLabV3+语义分割模型进行了系统优化，以提高其在自动驾驶中实时部署的性能。我们探讨了关键超参数的影响。求解器类型（Adam, RMSProp， SGDM），学习率（10−5,10−4,10−3），批大小（1,2,3）和L2正则化（10−5,10−4,10−3），跨越三个骨干网：ResNet-18， ResNet-50和MobileNetV2。实验是在剑桥驾驶标记视频数据库（CamVid）上进行的，CamVid是一种广泛使用的道路场景理解基准。DoE方法有效地减少了训练配置的数量，同时最大限度地提高了分割性能。性能最好的DeepLabV3+模型采用ResNet-50骨干网，实现了76.23%的平均联交（IoU），超过了最近的方法。该框架为在自动驾驶汽车系统中部署语义分割模型提供了一种实用的策略。
{"title":"Taguchi-optimized DeepLabV3+ for semantic segmentation in autonomous driving applications","authors":"G Divya Deepak, Pavan Hiremath, Subraya Krishna Bhat","doi":"10.1016/j.asej.2026.103985","DOIUrl":"10.1016/j.asej.2026.103985","url":null,"abstract":"<div><div>Semantic segmentation is a critical perception task in autonomous vehicles, enabling pixel-wise classification of road scenes. In this study, we propose a systematic optimization of DeepLabV3+ semantic segmentation model using Taguchi Design of Experiments (DoE) technique to enhance its performance for real-time deployment in autonomous driving. We explore the influence of key hyperparameters. solver type (Adam, RMSProp, SGDM), learning rate (10<sup>−5</sup>, 10<sup>−4</sup>, 10<sup>−3</sup>) batch size (1, 2, 3), and L2 regularization (10<sup>−5</sup>, 10<sup>−4</sup>, 10<sup>−3</sup>), across three backbone networks: ResNet-18, ResNet-50, and MobileNetV2. Experiments were conducted on the Cambridge-driving Labeled Video Database (CamVid), a widely used benchmark for road scene understanding. The DoE approach efficiently reduced the number of training configurations while maximizing segmentation performance. The best-performing model, DeepLabV3+ with a ResNet-50 backbone, achieved a Mean Intersection over Union (IoU) of 76.23%, surpassing recent approaches. The proposed framework offers a practical strategy for deploying semantic segmentation models in autonomous vehicle systems.</div></div>","PeriodicalId":48648,"journal":{"name":"Ain Shams Engineering Journal","volume":"17 2","pages":"Article 103985"},"PeriodicalIF":5.9,"publicationDate":"2026-01-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146023356","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Predictive maintenance (PdM) relies on accurate estimation of the remaining useful life (RUL) to support efficient industrial maintenance. However, most RUL models overlook uncertainty quantification (UQ), which is essential for safety–critical decision-making. This study presents a hybrid uncertainty-aware framework that combines a Transformer backbone with Monte Carlo Dropout (MC Dropout) and Conformal Prediction (CP). The Transformer architecture effectively learns long-range temporal dependencies in sensor data, while MC Dropout approximates epistemic uncertainty arising from model limitations. CP complements this by producing prediction intervals that capture aleatoric variability caused by noise and operating conditions. The framework is validated using NASA’s C-MAPSS FD001 and FD003 datasets. It achieves strong performance on FD001, with MAE 8.11, RMSE 11.71, and a predictive score of 193.6, and on FD003, with MAE 7.21, RMSE 10.50, and R2 0.926. By jointly addressing both uncertainty types, the method yields well-calibrated confidence intervals, enhancing reliability and interpretability in PdM applications.
{"title":"Uncertainty aware predictive maintenance using a hybrid Transformer with Monte Carlo Dropout and conformal prediction","authors":"Chao-Lung Yang, Tamrat Yifter Meles, Atinkut Atinafu Yilma, Melkamu Mengstnew Teshome","doi":"10.1016/j.asej.2026.103992","DOIUrl":"10.1016/j.asej.2026.103992","url":null,"abstract":"<div><div>Predictive maintenance (PdM) relies on accurate estimation of the remaining useful life (RUL) to support efficient industrial maintenance. However, most RUL models overlook uncertainty quantification (UQ), which is essential for safety–critical decision-making. This study presents a hybrid uncertainty-aware framework that combines a Transformer backbone with Monte Carlo Dropout (MC Dropout) and Conformal Prediction (CP). The Transformer architecture effectively learns long-range temporal dependencies in sensor data, while MC Dropout approximates epistemic uncertainty arising from model limitations. CP complements this by producing prediction intervals that capture aleatoric variability caused by noise and operating conditions. The framework is validated using NASA’s C-MAPSS FD001 and FD003 datasets. It achieves strong performance on FD001, with MAE 8.11, RMSE 11.71, and a predictive score of 193.6, and on FD003, with MAE 7.21, RMSE 10.50, and R2 0.926. By jointly addressing both uncertainty types, the method yields well-calibrated confidence intervals, enhancing reliability and interpretability in PdM applications.</div></div>","PeriodicalId":48648,"journal":{"name":"Ain Shams Engineering Journal","volume":"17 2","pages":"Article 103992"},"PeriodicalIF":5.9,"publicationDate":"2026-01-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146023359","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"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 : 2026-01-17DOI: 10.1016/j.asej.2025.103953
Selman Ogras, Fevzi Onen
Hydropower structures have approved significant progress and innovation in the development of water resources over the last 30 years, leading to the construction of large hydroelectric projects. Dissipation the enormous energy generated is a significant zone of dam engineering. Effective project design, which addresses the hydraulic characteristics of dam discharge structures and the safe and economical distribution of the resulting energy, requires a comprehensive evaluation of physical modeling, prototype experiments, and numerical modeling results. In this study, the hydraulic characteristics of the Ilısu Dam spillway structure, determined by physical modeling studies, and the effectiveness of the energy dissipation structures were numerically investigated using Computational Fluid Dynamics (Flow3D). Evaluations were accomplished by comparing the 1/100 scale model of the spillway structure and the 1/30 scale of the discharge channel. The numerical analyses employed the RNG and standard k-ε turbulence models, separately. Thus, the effectiveness of turbulence models across the entire spillway structure was determined. Moreover,16 different thresholds were designed with different deflector angles and radii of the flip bucket, which is one of the effective structures in terms of energy dissipation, and these designs were numerically analyzed and compared with the results obtained both in our current study and previous studies in the literature.
{"title":"Numerical analysis of hydraulic characteristics of spillways and effectiveness of energy dissipation structures","authors":"Selman Ogras, Fevzi Onen","doi":"10.1016/j.asej.2025.103953","DOIUrl":"10.1016/j.asej.2025.103953","url":null,"abstract":"<div><div>Hydropower structures have approved significant progress and innovation in the development of water resources over the last 30 years, leading to the construction of large hydroelectric projects. Dissipation the enormous energy generated is a significant zone of dam engineering. Effective project design, which addresses the hydraulic characteristics of dam discharge structures and the safe and economical distribution of the resulting energy, requires a comprehensive evaluation of physical modeling, prototype experiments, and numerical modeling results. In this study, the hydraulic characteristics of the Ilısu Dam spillway structure, determined by physical modeling studies, and the effectiveness of the energy dissipation structures were numerically investigated using Computational Fluid Dynamics (Flow3D). Evaluations were accomplished by comparing the 1/100 scale model of the spillway structure and the 1/30 scale of the discharge channel. The numerical analyses employed the RNG and standard k-ε turbulence models, separately. Thus, the effectiveness of turbulence models across the entire spillway structure was determined. Moreover,16 different thresholds were designed with different deflector angles and radii of the flip bucket, which is one of the effective structures in terms of energy dissipation, and these designs were numerically analyzed and compared with the results obtained both in our current study and previous studies in the literature.</div></div>","PeriodicalId":48648,"journal":{"name":"Ain Shams Engineering Journal","volume":"17 2","pages":"Article 103953"},"PeriodicalIF":5.9,"publicationDate":"2026-01-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146023816","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"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 : 2026-01-14DOI: 10.1016/j.asej.2025.103966
Hao Su , Ling Yin , Chaochao Qiu , Lijuan Zhang , Weicheng Lin , Xinyong Mao
High-precision machine tools are vital in modern manufacturing, yet their accuracy is often degraded by thermal errors. Traditional models lack cross-machine generalization and rely heavily on large labeled data. This paper proposes a thermal error modeling approach combining an encoder–decoder temporal convolutional network (ED-TCN) with representation subspace distance (RSD) transfer learning for cross-machine prediction. The encoder–decoder structure captures multi-scale features via dilated causal convolutions and residual blocks, enhancing long-term dependency modeling. The RSD-based domain adaptation reduces inter-machine distribution discrepancies while preserving feature scales. Through semi-supervised transfer learning, high-precision prediction is achieved using only 20% of labeled target data, greatly reducing collection costs. Experimental results on two different machine tools under three operating conditions demonstrate outstanding performance, achieving an R2 of 99.5%, an RMSE of 1.201 µm, and an MAE of 1.008 µm, thereby confirming the practicality and robustness of the proposed method.
{"title":"A novel transfer learning model based on ED-TCN and RSD domain adaptation for thermal error prediction of multiple machine tools","authors":"Hao Su , Ling Yin , Chaochao Qiu , Lijuan Zhang , Weicheng Lin , Xinyong Mao","doi":"10.1016/j.asej.2025.103966","DOIUrl":"10.1016/j.asej.2025.103966","url":null,"abstract":"<div><div>High-precision machine tools are vital in modern manufacturing, yet their accuracy is often degraded by thermal errors. Traditional models lack cross-machine generalization and rely heavily on large labeled data. This paper proposes a thermal error modeling approach combining an encoder–decoder temporal convolutional network (ED-TCN) with representation subspace distance (RSD) transfer learning for cross-machine prediction. The encoder–decoder structure captures multi-scale features via dilated causal convolutions and residual blocks, enhancing long-term dependency modeling. The RSD-based domain adaptation reduces inter-machine distribution discrepancies while preserving feature scales. Through semi-supervised transfer learning, high-precision prediction is achieved using only 20% of labeled target data, greatly reducing collection costs. Experimental results on two different machine tools under three operating conditions demonstrate outstanding performance, achieving an R<sup>2</sup> of 99.5%, an RMSE of 1.201 µm, and an MAE of 1.008 µm, thereby confirming the practicality and robustness of the proposed method.</div></div>","PeriodicalId":48648,"journal":{"name":"Ain Shams Engineering Journal","volume":"17 2","pages":"Article 103966"},"PeriodicalIF":5.9,"publicationDate":"2026-01-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145979662","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
The assessment of water quality has become increasingly vital for maintaining the ecological balance and ensuring public safety across global water systems. This study examines the application of Quantum Machine Learning (QML) techniques in a real-world setting to predict water quality in the U20A region of the Umgeni Catchment, Durban, South Africa. We implemented the Quantum Support Vector Classifier (QSVC) and Quantum Neural Network (QNN) on a field-collected dataset. Our results demonstrate that the QSVC is more practical to implement and yields superior performance, achieving 75 % accuracy with polynomial and radial basis function kernels. In contrast, the QNN encountered persistent convergence issues, including the “dead neuron” problem, despite various optimization strategies. The findings provide a pragmatic framework for environmental monitoring applications, suggesting that QSVC offers a more viable near-term quantum approach for water quality classification tasks with imbalanced, real-world data.
{"title":"Predicting water quality using quantum machine learning: The case of the umgeni catchment (U20A) study region","authors":"Jamal Al-Karaki , Muhammad Al-Zafar Khan , Amjad Gawanmeh , Marwan Omar","doi":"10.1016/j.asej.2025.103925","DOIUrl":"10.1016/j.asej.2025.103925","url":null,"abstract":"<div><div>The assessment of water quality has become increasingly vital for maintaining the ecological balance and ensuring public safety across global water systems. This study examines the application of Quantum Machine Learning (QML) techniques in a real-world setting to predict water quality in the U20A region of the Umgeni Catchment, Durban, South Africa. We implemented the Quantum Support Vector Classifier (QSVC) and Quantum Neural Network (QNN) on a field-collected dataset. Our results demonstrate that the QSVC is more practical to implement and yields superior performance, achieving 75 % accuracy with polynomial and radial basis function kernels. In contrast, the QNN encountered persistent convergence issues, including the “dead neuron” problem, despite various optimization strategies. The findings provide a pragmatic framework for environmental monitoring applications, suggesting that QSVC offers a more viable near-term quantum approach for water quality classification tasks with imbalanced, real-world data.</div></div>","PeriodicalId":48648,"journal":{"name":"Ain Shams Engineering Journal","volume":"17 2","pages":"Article 103925"},"PeriodicalIF":5.9,"publicationDate":"2026-01-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145979665","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"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 : 2026-01-13DOI: 10.1016/j.asej.2025.103893
Umair Hussan , Sotdhipong Phichaisawat , Huaizhi Wang , Muhammad Ahsan Ayub , Muhammad Saqib Ali
The integration of intermittent renewable energy sources introduces significant operational uncertainties, challenging the economic efficiency and reliability of power systems. Centralized energy management strategies for multi-microgrid (MMG) networks face critical limitations in scalability, data privacy, and resilience to single-point failures. This study presents a scalable, privacy-preserving, and decentralized energy management framework for cooperative MMG networks to enhance operational efficiency and resilience. To achieve this, a novel consensus-based decentralized optimization algorithm is proposed, utilizing the Alternating Direction Method of Multipliers (ADMM), which decomposes the global optimal energy management problem into local subproblems that can be solved independently by each microgrid (MG). The method enables real-time coordination through iterative updates of local variables, consensus on power exchanges, and minimal information sharing—only power flows and dual variables between neighboring MGs. Simulation results on a modified 33-bus system with three interconnected MGs demonstrate that the proposed framework effectively balances supply and demand, optimizes energy storage utilization, and facilitates peer-to-peer energy trading, achieving lower operational costs and faster convergence compared to conventional ADMM, dual decomposition, consensus gradient, and proximal message passing methods. The proposed ADMM-based consensus framework provides a robust, scalable, and economically efficient solution for decentralized energy management in cooperative MMG systems.
{"title":"A novel consensus-based decentralized framework for optimal energy management in cooperative multi-microgrid networks using ADMM","authors":"Umair Hussan , Sotdhipong Phichaisawat , Huaizhi Wang , Muhammad Ahsan Ayub , Muhammad Saqib Ali","doi":"10.1016/j.asej.2025.103893","DOIUrl":"10.1016/j.asej.2025.103893","url":null,"abstract":"<div><div>The integration of intermittent renewable energy sources introduces significant operational uncertainties, challenging the economic efficiency and reliability of power systems. Centralized energy management strategies for multi-microgrid (MMG) networks face critical limitations in scalability, data privacy, and resilience to single-point failures. This study presents a scalable, privacy-preserving, and decentralized energy management framework for cooperative MMG networks to enhance operational efficiency and resilience. To achieve this, a novel consensus-based decentralized optimization algorithm is proposed, utilizing the Alternating Direction Method of Multipliers (ADMM), which decomposes the global optimal energy management problem into local subproblems that can be solved independently by each microgrid (MG). The method enables real-time coordination through iterative updates of local variables, consensus on power exchanges, and minimal information sharing—only power flows and dual variables between neighboring MGs. Simulation results on a modified 33-bus system with three interconnected MGs demonstrate that the proposed framework effectively balances supply and demand, optimizes energy storage utilization, and facilitates peer-to-peer energy trading, achieving lower operational costs and faster convergence compared to conventional ADMM, dual decomposition, consensus gradient, and proximal message passing methods. The proposed ADMM-based consensus framework provides a robust, scalable, and economically efficient solution for decentralized energy management in cooperative MMG systems.</div></div>","PeriodicalId":48648,"journal":{"name":"Ain Shams Engineering Journal","volume":"17 2","pages":"Article 103893"},"PeriodicalIF":5.9,"publicationDate":"2026-01-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145979666","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"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 : 2026-01-12DOI: 10.1016/j.asej.2025.103963
Chuan-qi Zhu, Zi-xuan Chen, Feng Lin, Jun Zhao
The intense dynamic loading induced by mechanical coal cutting and blasting pre-damage the bearing capacity of surrounding rock. To examine the influence of impact pressure on the evolution of damage to the coal and the effects on its static mechanical response, coal specimens were subjected to controlled multiple isobaric impact via a Split Hopkinson Pressure Bar (SHPB) system, generating specimens with progressive degrees of damage. Micro-focus CT scanning was used to characterize the morphology of fracture distribution, followed by uniaxial compression tests on damaged coal using an MTS-816 testing machine to investigate static mechanical properties. Key findings reveal: 1) The wave velocity progressively declined followed by accelerated reduction with increasing number of impacts, while the degree of damage exhibited initial gradual growth preceding rapid escalation; 2) The fracture porosity, fractal dimension, and fracture volume increased rapidly then stabilized, whereas the three-dimensional (3D) connectivity rose continuously. The volume of connected fractures ascended with the decelerating rate of growth, and the connected fracture ratio initially dropped then rose. Layer-wise porosity profiles indicated larger damage at specimen ends versus central regions; 3) The peak stress decreases rapidly − steadily − rapidly with the increase of impact times, while the elastic modulus shows a trend of gradually decreasing decline. Before the circumferential strain reaches the peak stress, the stress rises rapidly. The particle size distribution of the specimen after failure accumulates from more than 12.5 mm to less than 1 mm with the increase of the number of impacts; 4) Compare the correlation curves of the microstructural parameters and the macro-mechanics parameter, and compare the magnitudes of the correlation coefficients. By comprehensively comparing the relationship between the microstructural parameters and the peak stress and elastic modulus, it was found that the correlation coefficient between the fracture area and the peak stress and elastic modulus of the specimen was the highest, which were 0.976 and 0.990 respectively. These results provide theoretical and engineering foundations for mitigating instability hazards of coal mines.
{"title":"Experimental study of the static mechanical response of impact-damaged coal","authors":"Chuan-qi Zhu, Zi-xuan Chen, Feng Lin, Jun Zhao","doi":"10.1016/j.asej.2025.103963","DOIUrl":"10.1016/j.asej.2025.103963","url":null,"abstract":"<div><div>The intense dynamic loading induced by mechanical coal cutting and blasting pre-damage the bearing capacity of surrounding rock. To examine the influence of impact pressure on the evolution of damage to the coal and the effects on its static mechanical response, coal specimens were subjected to controlled multiple isobaric impact via a Split Hopkinson Pressure Bar (SHPB) system, generating specimens with progressive degrees of damage. Micro-focus CT scanning was used to characterize the morphology of fracture distribution, followed by uniaxial compression tests on damaged coal using an MTS-816 testing machine to investigate static mechanical properties. Key findings reveal: 1) The wave velocity progressively declined followed by accelerated reduction with increasing number of impacts, while the degree of damage exhibited initial gradual growth preceding rapid escalation; 2)<!--> <!-->The fracture porosity, fractal dimension, and fracture volume increased rapidly then stabilized, whereas the three-dimensional (3D) connectivity rose continuously. The volume of connected fractures ascended with the decelerating rate of growth, and the connected fracture ratio initially dropped then rose. Layer-wise porosity profiles indicated larger damage at specimen ends <em>versus</em> central regions; 3) The peak stress decreases rapidly − steadily − rapidly with the increase of impact times, while the elastic modulus shows a trend of gradually decreasing decline. Before the circumferential strain reaches the peak stress, the stress rises rapidly. The particle size distribution of the specimen after failure accumulates from more than 12.5 mm to less than 1 mm with the increase of the number of impacts; 4) Compare the correlation curves of the microstructural parameters and the macro-mechanics parameter, and compare the magnitudes of the correlation coefficients. By comprehensively comparing the relationship between the microstructural parameters and the peak stress and elastic modulus, it was found that the correlation coefficient between the fracture area and the peak stress and elastic modulus of the specimen was the highest, which were 0.976 and 0.990 respectively. These results provide theoretical and engineering foundations for mitigating instability hazards of coal mines.</div></div>","PeriodicalId":48648,"journal":{"name":"Ain Shams Engineering Journal","volume":"17 2","pages":"Article 103963"},"PeriodicalIF":5.9,"publicationDate":"2026-01-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145979663","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"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 : 2026-01-12DOI: 10.1016/j.asej.2025.103643
Muhammad Ali , Ayesha Zubair , Wasim Abbas , Zubair Masoud , Ali Aldrees
Diverse climatic conditions, ranging from hot deserts to highland climates, pose significant challenges in predicting the cost impact of green housing which is an essential for significant reduction in carbon emissions to mitigate the effects of climate change. This study employs an innovative approach using hybrid AI model to predict green housing costs, emphasizing stakeholder understanding within Pakistan’s unique socio-economic and climatic contexts. Data was collected from multiple climatic zones, focusing on eighteen key factors influencing green housing costs. The dataset underwent rigorous cleaning, preprocessing, and analysis, including density distribution, cumulative probability, and sensitivity assessments, with results visualized for better interpretation. A hybrid AI model was developed to enhance prediction accuracy by integrating algorithms like Support Vector Machine (SVM), Decision Tree and K-Nearest Neighbor (KNN). Machine learning models were trained, tested, and compared using metrics for model evaluation i.e., R-squared (R2), Root Mean Square Error (RMSE), Mean Absolute Error (MAE) and Mean Square Error (MSE). The hybrid model demonstrated superior performance with results having R2: 0.99, explaining 99.99 % of the dataset variance. Additionally, Pearson’s correlation matrix revealed that level of awareness (−0.97), climate-responsive design (−0.86), and material inventory (−0.83) exhibited the strongest negative correlations with cost impact, while site area temperature (0.72) had the most significant positive correlations. External validation using an independent dataset of 659 samples (R-squared: 0.81) and Taylor diagram analysis (standard deviation < 2.9 %, correlation > 0.82) further validated proposed model’s superiority over existing models. These findings provide a comprehensive cost prediction framework aiding stakeholders in making informed decisions on sustainable and cost-effective green housing.
{"title":"Integrating AI models for cost prediction in green housing in diverse climates − an innovative framework for stakeholder understanding in Pakistan","authors":"Muhammad Ali , Ayesha Zubair , Wasim Abbas , Zubair Masoud , Ali Aldrees","doi":"10.1016/j.asej.2025.103643","DOIUrl":"10.1016/j.asej.2025.103643","url":null,"abstract":"<div><div>Diverse climatic conditions, ranging from hot deserts to highland climates, pose significant challenges in predicting the cost impact of green housing which is an essential for significant reduction in carbon emissions to mitigate the effects of climate change. This study employs an innovative approach using hybrid AI model to predict green housing costs, emphasizing stakeholder understanding within Pakistan’s unique socio-economic and climatic contexts. Data was collected from multiple climatic zones, focusing on eighteen key factors influencing green housing costs. The dataset underwent rigorous cleaning, preprocessing, and analysis, including density distribution, cumulative probability, and sensitivity assessments, with results visualized for better interpretation. A hybrid AI model was developed to enhance prediction accuracy by integrating algorithms like Support Vector Machine (SVM), Decision Tree and K-Nearest Neighbor (KNN). Machine learning models were trained, tested, and compared using metrics for model evaluation i.e., R-squared (R2), Root Mean Square Error (RMSE), Mean Absolute Error (MAE) and Mean Square Error (MSE). The hybrid model demonstrated superior performance with results having R2: 0.99, explaining 99.99 % of the dataset variance. Additionally, Pearson’s correlation matrix revealed that level of awareness (−0.97), climate-responsive design (−0.86), and material inventory (−0.83) exhibited the strongest negative correlations with cost impact, while site area temperature (0.72) had the most significant positive correlations. External validation using an independent dataset of 659 samples (R-squared: 0.81) and Taylor diagram analysis (standard deviation < 2.9 %, correlation > 0.82) further validated proposed model’s superiority over existing models. These findings provide a comprehensive cost prediction framework aiding stakeholders in making informed decisions on sustainable and cost-effective green housing.</div></div>","PeriodicalId":48648,"journal":{"name":"Ain Shams Engineering Journal","volume":"17 2","pages":"Article 103643"},"PeriodicalIF":5.9,"publicationDate":"2026-01-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145979664","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"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 : 2026-01-12DOI: 10.1016/j.asej.2025.103923
Liguo Han , Feitao Dong , Hengfei Xiao , Fei Ding , Lijuan Zhao , Peng Li , Chuanzong Li , Yue Zhou
To ensure accurate identification and control of coal and gangue during top-coal caving mining, this study proposes a multimodal information fusion method integrating vibration data, infrared images, and RGB images. The vibration signals were transformed into time–frequency spectrograms using the Continuous Wavelet Transform (CWT), and a Wasserstein Generative Adversarial Network with Gradient Penalty (WGAN-GP) was employed for data augmentation to mitigate sample scarcity. Comparative experiments between early and late fusion strategies revealed that the late fusion approach based on the ResNet architecture yielded superior performance. With the optimal combination of vibration spectrograms (ResNet-18), infrared images (ResNet-50), and RGB images (ResNet-18), the model achieved a favorable balance between high accuracy and computational efficiency. Finally, a multi-domain co-simulation control system was developed for verification, demonstrating an average response time below 0.66 s under various rock-mixing ratio conditions. The proposed framework offers an effective technical solution for high-efficiency, clean coal production.
{"title":"Research on coal gangue identification based on multimodal fusion and multidomain collaborative simulation","authors":"Liguo Han , Feitao Dong , Hengfei Xiao , Fei Ding , Lijuan Zhao , Peng Li , Chuanzong Li , Yue Zhou","doi":"10.1016/j.asej.2025.103923","DOIUrl":"10.1016/j.asej.2025.103923","url":null,"abstract":"<div><div>To ensure accurate identification and control of coal and gangue during top-coal caving mining, this study proposes a multimodal information fusion method integrating vibration data, infrared images, and RGB images. The vibration signals were transformed into time–frequency spectrograms using the Continuous Wavelet Transform (CWT), and a Wasserstein Generative Adversarial Network with Gradient Penalty (WGAN-GP) was employed for data augmentation to mitigate sample scarcity. Comparative experiments between early and late fusion strategies revealed that the late fusion approach based on the ResNet architecture yielded superior performance. With the optimal combination of vibration spectrograms (ResNet-18), infrared images (ResNet-50), and RGB images (ResNet-18), the model achieved a favorable balance between high accuracy and computational efficiency. Finally, a multi-domain co-simulation control system was developed for verification, demonstrating an average response time below 0.66 s under various rock-mixing ratio conditions. The proposed framework offers an effective technical solution for high-efficiency, clean coal production.</div></div>","PeriodicalId":48648,"journal":{"name":"Ain Shams Engineering Journal","volume":"17 2","pages":"Article 103923"},"PeriodicalIF":5.9,"publicationDate":"2026-01-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145979667","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"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 : 2026-01-10DOI: 10.1016/j.asej.2025.103952
R. Juliana , Vignesh Janarthanan , M. Umamaheswari , S. Sasikanth
Wireless Sensor Networks (WSNs) require novel approaches to optimize localization accuracy and minimize energy usage due to their decentralized structure and the resource constraints of mobile sensor nodes. This research proposes LEE-NAM-SFS, a neural attention guided stochastic fractal search framework that equips sensor nodes with cognitive capabilities to adapt their movement patterns, sensing behavior, and communication strategies. The neuronal attention model selectively focuses on high-value measurements, while stochastic fractal search explores the high-dimensional search space of node trajectories and routing choices to jointly optimize localization and energy efficiency. Extensive simulations on a 100-node mobile WSN scenario show that, compared to FLA-RTWOA, DASUL, RDEANTN, MANAL, and QLAMSR, LEE-NAM-SFS improves localization accuracy by approximately 2–5 %, enhances energy efficiency by 5–10 %, increases data delivery rate by 2–5 %, expands effective coverage area by 5–10 %, and prolongs network lifetime by 5–10 %. These gains are achieved without compromising connectivity or data reliability.
{"title":"Neural attention guided stochastic fractal search for energy efficient localization in mobile wireless sensor networks","authors":"R. Juliana , Vignesh Janarthanan , M. Umamaheswari , S. Sasikanth","doi":"10.1016/j.asej.2025.103952","DOIUrl":"10.1016/j.asej.2025.103952","url":null,"abstract":"<div><div>Wireless Sensor Networks (WSNs) require novel approaches to optimize localization accuracy and minimize energy usage due to their decentralized structure and the resource constraints of mobile sensor nodes. This research proposes LEE-NAM-SFS, a neural attention guided stochastic fractal search framework that equips sensor nodes with cognitive capabilities to adapt their movement patterns, sensing behavior, and communication strategies. The neuronal attention model selectively focuses on high-value measurements, while stochastic fractal search explores the high-dimensional search space of node trajectories and routing choices to jointly optimize localization and energy efficiency. Extensive simulations on a 100-node mobile WSN scenario show that, compared to FLA-RTWOA, DASUL, RDEANTN, MANAL, and QLAMSR, LEE-NAM-SFS improves localization accuracy by approximately 2–5 %, enhances energy efficiency by 5–10 %, increases data delivery rate by 2–5 %, expands effective coverage area by 5–10 %, and prolongs network lifetime by 5–10 %. These gains are achieved without compromising connectivity or data reliability.</div></div>","PeriodicalId":48648,"journal":{"name":"Ain Shams Engineering Journal","volume":"17 2","pages":"Article 103952"},"PeriodicalIF":5.9,"publicationDate":"2026-01-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145928424","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"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