alexandria engineering journal最新文献_第9页

Experimental investigation of wheel-rail lateral forces and dynamic running behavior in sharp curves of heavy-haul railway systems 重载铁路急弯道轮轨侧向力及动态运行特性试验研究

IF 6.8 2区工程技术 Q1 ENGINEERING, MULTIDISCIPLINARY

alexandria engineering journal

Pub Date : 2025-12-09 DOI: 10.1016/j.aej.2025.12.012

Junrui Zhang , Wengang Fan , Daolin Si , Jiang Li , Dongyu Zhao

For the heavy-haul railway curve section with a 500 m radius, dynamic whole-vehicle passing tests were conducted by setting three different curve superelevation values of 60 mm, 90 mm, and 120 mm, achieving comprehensive dynamic testing of all parameters in wheel-rail interaction. Strain gauges were attached to the rails to measure the wheel-rail lateral forces when the train passed at different speeds, and the variation patterns of lateral forces on the guiding and trailing axles under different superelevation conditions were analyzed. A real-time angle-of-attack (AoA) measurement method based on synchronous wheel-rail force sampling was proposed, overcoming the limitations of traditional techniques in dynamically capturing wheelset attitudes, which can be used to analyze the operational characteristics of guiding and trailing axles under varying superelevation conditions. Results demonstrate that unbalanced superelevation (-5 to −25 mm) effectively reduces rail wear and enhances curve negotiation for 25 t axle-load trains. Increased deficient superelevation decreases the steering axle’s AoA but increases the trailing axle’s AoA. Lateral forces exhibit distinct patterns: for steering axles, inner rail inward forces decrease, while outer rail outward forces increase and inward forces diminish; trailing axles experience amplified lateral forces on both rails in opposing directions. These findings reveal superelevation’s dual role in governing lateral force distribution and wheelset alignment. The proposed AoA measurement technique provides critical insights into dynamic wheelset behavior during sharp curve transitions. The study establishes an optimal superelevation range to balance lateral force mitigation and operational stability, offering actionable guidelines for track maintenance and superelevation design in heavy-haul corridors. This work advances the understanding of vehicle-track interaction mechanics in tight curves, with direct implications for extending rail service life and improving safety in challenging geometric conditions.

针对半径为500 m的重载铁路曲线段，通过设置60 mm、90 mm和120 mm三种不同的曲线超标高值，进行整车动力通过试验，实现轮轨相互作用各参数的综合动力试验。在钢轨上安装应变片，测量列车以不同速度通过时的轮轨侧向力，分析了不同超高工况下导向轴和尾轴侧向力的变化规律。提出了一种基于轮轨同步力采样的实时攻角测量方法，克服了传统技术在动态捕获轮对姿态方面的局限性，可用于分析导向轴和尾轴在不同超仰角条件下的工作特性。结果表明，对于25 t轴载列车，非平衡超仰角（-5 ~ - 25 mm）可有效减少钢轨磨损，提高曲线协调能力。缺陷超高程的增加降低了转向轴的AoA，但增加了后桥的AoA。横向力表现出明显的模式：对于转向轴，内轨向内力减小，外轨向外力增大，向内力减小；拖尾轴在相反方向上的两个轨道上经历了放大的横向力。这些发现揭示了超高在控制侧向力分布和轮对对齐中的双重作用。提出的AoA测量技术提供了关键的见解动态轮对行为在急剧曲线过渡。该研究建立了一个最佳的超标高范围，以平衡横向力缓解和运行稳定性，为重载走廊的轨道维护和超标高设计提供可操作的指导方针。这项工作促进了对紧密曲线中车辆-轨道相互作用力学的理解，对延长铁路使用寿命和提高具有挑战性的几何条件下的安全性具有直接意义。

{"title":"Experimental investigation of wheel-rail lateral forces and dynamic running behavior in sharp curves of heavy-haul railway systems","authors":"Junrui Zhang , Wengang Fan , Daolin Si , Jiang Li , Dongyu Zhao","doi":"10.1016/j.aej.2025.12.012","DOIUrl":"10.1016/j.aej.2025.12.012","url":null,"abstract":"<div><div>For the heavy-haul railway curve section with a 500 m radius, dynamic whole-vehicle passing tests were conducted by setting three different curve superelevation values of 60 mm, 90 mm, and 120 mm, achieving comprehensive dynamic testing of all parameters in wheel-rail interaction. Strain gauges were attached to the rails to measure the wheel-rail lateral forces when the train passed at different speeds, and the variation patterns of lateral forces on the guiding and trailing axles under different superelevation conditions were analyzed. A real-time angle-of-attack (AoA) measurement method based on synchronous wheel-rail force sampling was proposed, overcoming the limitations of traditional techniques in dynamically capturing wheelset attitudes, which can be used to analyze the operational characteristics of guiding and trailing axles under varying superelevation conditions. Results demonstrate that unbalanced superelevation (-5 to −25 mm) effectively reduces rail wear and enhances curve negotiation for 25 t axle-load trains. Increased deficient superelevation decreases the steering axle’s AoA but increases the trailing axle’s AoA. Lateral forces exhibit distinct patterns: for steering axles, inner rail inward forces decrease, while outer rail outward forces increase and inward forces diminish; trailing axles experience amplified lateral forces on both rails in opposing directions. These findings reveal superelevation’s dual role in governing lateral force distribution and wheelset alignment. The proposed AoA measurement technique provides critical insights into dynamic wheelset behavior during sharp curve transitions. The study establishes an optimal superelevation range to balance lateral force mitigation and operational stability, offering actionable guidelines for track maintenance and superelevation design in heavy-haul corridors. This work advances the understanding of vehicle-track interaction mechanics in tight curves, with direct implications for extending rail service life and improving safety in challenging geometric conditions.</div></div>","PeriodicalId":7484,"journal":{"name":"alexandria engineering journal","volume":"134 ","pages":"Pages 98-109"},"PeriodicalIF":6.8,"publicationDate":"2025-12-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145735322","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}

引用次数: 0

Autonomous charging system via manipulator-UGV docking using zero-shot 6-DoF pose estimation 基于零射击六自由度姿态估计的机器人- ugv对接自主充电系统

IF 6.8 2区工程技术 Q1 ENGINEERING, MULTIDISCIPLINARY

alexandria engineering journal

Pub Date : 2025-12-09 DOI: 10.1016/j.aej.2025.12.015

Minkyu Jung, Andrew Jaeyong Choi

To ensure long-term field operation of Unmanned Ground Vehicles (UGVs) without human intervention, autonomous charging capability is essential. This paper presents a fully autonomous charging system that enables a UGV to perform path planning, align itself with the charging station, and a 6-DoF manipulator to locate and connect to the charging port without relying on external markers. By combining vision-language detection with geometric reasoning on RGB-D data, the system estimates a complete 6-DoF pose without relying on fiducial markers or task-specific retraining, thereby enhancing adaptability across different docking environments. Specifically, it employs the open-vocabulary object detector YOLO-World to identify the docking port in the RGB image and selects the corresponding 3D position from the depth map. To estimate orientation, the surrounding point cloud is clustered to extract the object shape, from which the without relying on fiducial markers or task-specific retraining, thereby enhancing adaptability across different docking environments. principal axis is computed, resulting in a fully generalizable and marker-free docking pipeline. This 6-DoF pose is used to compute a feasible manipulator configuration through inverse kinematics. The proposed pipeline is integrated into a ROS-based framework and enables precise docking in unstructured outdoor environments without manual supervision. Experimental results demonstrate the robustness and accuracy of the system under real-world conditions, achieving a docking success rate of 90% and an average total operation time of 34.4 s.

为了确保无人地面车辆（ugv）在没有人为干预的情况下长期野外作业，自主充电能力至关重要。本文提出了一种完全自主的充电系统，使UGV能够进行路径规划，并与充电站对齐，六自由度机械手能够在不依赖外部标记的情况下定位并连接到充电端口。通过将视觉语言检测与基于RGB-D数据的几何推理相结合，系统在不依赖基准标记或特定任务再训练的情况下估计出完整的6自由度姿态，从而增强了对不同对接环境的适应性。具体来说，它使用开放词汇对象检测器YOLO-World识别RGB图像中的对接端口，并从深度图中选择相应的3D位置。为了估计方向，对周围的点云进行聚类提取物体形状，从而不依赖于基准标记或特定任务的再训练，从而增强了跨不同对接环境的适应性。计算主轴，得到一个完全通用的无标记对接管道。利用该六自由度位姿通过运动学逆解计算出可行的机械臂构型。拟议的管道集成到基于ros的框架中，可以在无需人工监督的非结构化室外环境中进行精确对接。实验结果证明了该系统在实际条件下的鲁棒性和准确性，实现了90%的对接成功率，平均总操作时间为34.4 s。

{"title":"Autonomous charging system via manipulator-UGV docking using zero-shot 6-DoF pose estimation","authors":"Minkyu Jung, Andrew Jaeyong Choi","doi":"10.1016/j.aej.2025.12.015","DOIUrl":"10.1016/j.aej.2025.12.015","url":null,"abstract":"<div><div>To ensure long-term field operation of Unmanned Ground Vehicles (UGVs) without human intervention, autonomous charging capability is essential. This paper presents a fully autonomous charging system that enables a UGV to perform path planning, align itself with the charging station, and a 6-DoF manipulator to locate and connect to the charging port without relying on external markers. By combining vision-language detection with geometric reasoning on RGB-D data, the system estimates a complete 6-DoF pose without relying on fiducial markers or task-specific retraining, thereby enhancing adaptability across different docking environments. Specifically, it employs the open-vocabulary object detector YOLO-World to identify the docking port in the RGB image and selects the corresponding 3D position from the depth map. To estimate orientation, the surrounding point cloud is clustered to extract the object shape, from which the without relying on fiducial markers or task-specific retraining, thereby enhancing adaptability across different docking environments. principal axis is computed, resulting in a fully generalizable and marker-free docking pipeline. This 6-DoF pose is used to compute a feasible manipulator configuration through inverse kinematics. The proposed pipeline is integrated into a ROS-based framework and enables precise docking in unstructured outdoor environments without manual supervision. Experimental results demonstrate the robustness and accuracy of the system under real-world conditions, achieving a docking success rate of 90% and an average total operation time of 34.4 s.</div></div>","PeriodicalId":7484,"journal":{"name":"alexandria engineering journal","volume":"134 ","pages":"Pages 122-134"},"PeriodicalIF":6.8,"publicationDate":"2025-12-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145735269","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}

引用次数: 0

An improved chaotic whale optimization algorithm for MHA-MLP stock trend forecasting 一种改进的混沌鲸优化算法用于MHA-MLP股票走势预测

IF 6.8 2区工程技术 Q1 ENGINEERING, MULTIDISCIPLINARY

alexandria engineering journal

Pub Date : 2025-12-06 DOI: 10.1016/j.aej.2025.12.005

Ziru Li

Stock markets, characterized by high noise, non-stationarity, and stochastic fluctuations, pose significant challenges for accurate trend prediction. While deep learning models show promise, their training is often inefficient due to tedious parameter optimization. This paper proposes a novel hybrid prediction model that synergistically integrates an improved chaotic whale optimization algorithm with a multi-head attention mechanism and a multilayer perceptron neural network. The overall architecture consists of three core components: data preprocessing, an attention-enhanced multilayer perceptron module, and an optimization module based on the improved whale optimization algorithm. The attention-enhanced multi-layer perceptron employs residual connections between its attention and hidden layers to mitigate the vanishing gradient problem and effectively capture global temporal dependencies in the input features. The chaotic whale optimization module innovatively utilizes T-distribution wavelet mutation and polynomial differential evolution strategies to automatically optimize the model's architecture, specifically the number of attention heads and hidden layers. Extensive experiments on A-share and U.S. stock market datasets validate the model's efficacy. The results demonstrate that the chaotic whale optimization algorithm not only reduces the number of hyperparameter optimization iterations by 32 % but also improves fitness value convergence accuracy by 19 %, offering a robust and efficient solution for modeling complex financial data.

股票市场具有高噪声、非平稳性和随机波动的特点，对准确的趋势预测提出了重大挑战。虽然深度学习模型显示出前景，但由于繁琐的参数优化，它们的训练往往效率低下。本文提出了一种新的混合预测模型，该模型将改进的混沌鲸优化算法与多头注意机制和多层感知器神经网络协同集成。整体架构包括三个核心组件：数据预处理、注意力增强多层感知器模块和基于改进鲸鱼优化算法的优化模块。注意增强多层感知器利用其注意层和隐藏层之间的残差连接来缓解梯度消失问题，并有效捕获输入特征中的全局时间依赖性。混沌鲸鱼优化模块创新地利用t分布小波突变和多项式差分进化策略自动优化模型的结构，特别是注意头和隐藏层的数量。在a股和美国股市数据集上的大量实验验证了该模型的有效性。结果表明，混沌鲸优化算法不仅将超参数优化迭代次数减少了32% %，而且将适应度值收敛精度提高了19% %，为复杂金融数据建模提供了一种鲁棒高效的解决方案。

{"title":"An improved chaotic whale optimization algorithm for MHA-MLP stock trend forecasting","authors":"Ziru Li","doi":"10.1016/j.aej.2025.12.005","DOIUrl":"10.1016/j.aej.2025.12.005","url":null,"abstract":"<div><div>Stock markets, characterized by high noise, non-stationarity, and stochastic fluctuations, pose significant challenges for accurate trend prediction. While deep learning models show promise, their training is often inefficient due to tedious parameter optimization. This paper proposes a novel hybrid prediction model that synergistically integrates an improved chaotic whale optimization algorithm with a multi-head attention mechanism and a multilayer perceptron neural network. The overall architecture consists of three core components: data preprocessing, an attention-enhanced multilayer perceptron module, and an optimization module based on the improved whale optimization algorithm. The attention-enhanced multi-layer perceptron employs residual connections between its attention and hidden layers to mitigate the vanishing gradient problem and effectively capture global temporal dependencies in the input features. The chaotic whale optimization module innovatively utilizes T-distribution wavelet mutation and polynomial differential evolution strategies to automatically optimize the model's architecture, specifically the number of attention heads and hidden layers. Extensive experiments on A-share and U.S. stock market datasets validate the model's efficacy. The results demonstrate that the chaotic whale optimization algorithm not only reduces the number of hyperparameter optimization iterations by 32 % but also improves fitness value convergence accuracy by 19 %, offering a robust and efficient solution for modeling complex financial data.</div></div>","PeriodicalId":7484,"journal":{"name":"alexandria engineering journal","volume":"134 ","pages":"Pages 69-80"},"PeriodicalIF":6.8,"publicationDate":"2025-12-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145683520","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}

引用次数: 0

QuantumMedKD: A hybrid quantum–classical knowledge distillation framework for medical image analysis QuantumMedKD：用于医学图像分析的混合量子-经典知识蒸馏框架

IF 6.8 2区工程技术 Q1 ENGINEERING, MULTIDISCIPLINARY

alexandria engineering journal

Pub Date : 2025-12-05 DOI: 10.1016/j.aej.2025.12.007

MD Nahid Hassan Nishan , Mohammad Junayed Hasan , M.R.C. Mahdy

While hybrid quantum–classical architectures and knowledge distillation have each been explored independently in medical imaging and model compression, no prior work has unified these domains within a single framework. This study addresses this gap by investigating the integration of quantum–classical neural networks with knowledge distillation for medical image analysis, with a focus on circuit behavior, parameter efficiency, and diagnostic performance. The framework (QuantumMedKD) employs classical CNN architectures (ResNet-50, EfficientNet-B0, Xception) as teacher models, transferring learned representations to parameter-efficient quantum student networks via parameterized quantum circuits across qubit configurations (3-8), demonstrated through pneumonia detection from pediatric chest radiographs. Experimental validation reveals remarkable parameter efficiency, requiring only 24-36 trainable parameters compared to millions for classical counterparts, achieving compression ratios exceeding

1 0^{5}

while maintaining competitive diagnostic performance. Optimal configurations achieve 84.00% accuracy (EfficientNet-B0, 6-qubit) and 73.33% (Xception, 4-qubit), with knowledge distillation providing statistically significant improvements (

p

-vallues

<

0.001, Cohen’s d

>

2.0). Medical evaluation confirms diagnostic capability with sensitivity 81.4%, specificity 73.9%, and AUC 0.89, establishing quantum–classical knowledge transfer viability for resource-constrained healthcare deployment. QuantumMedKD reveals essential principles and proof-of-concept for quantum-enhanced healthcare AI systems within and beyond the noisy intermediate-scale quantum (NISQ) era for high efficiency and deployability, paving the way for future advancements in the field.

虽然混合量子经典架构和知识蒸馏各自在医学成像和模型压缩中进行了独立的探索，但之前的工作没有将这些领域统一在一个框架内。本研究通过研究量子经典神经网络与医学图像分析知识蒸馏的集成来解决这一差距，重点关注电路行为、参数效率和诊断性能。该框架（QuantumMedKD）采用经典的CNN架构（ResNet-50、EfficientNet-B0、Xception）作为教师模型，通过跨量子位配置的参数化量子电路（3-8）将学习到的表征转移到参数化的量子学生网络中，并通过儿童胸部x光片的肺炎检测进行了演示。实验验证显示了显著的参数效率，只需要24-36个可训练参数，而传统的同类方法需要数百万个，在保持有竞争力的诊断性能的同时实现超过105的压缩比。最优配置的准确率达到84.00% （EfficientNet-B0, 6量子位）和73.33% （Xception， 4量子位），知识蒸馏提供了统计学上显著的改进（p值<； 0.001, Cohen 's d > 2.0）。医学评估确认诊断能力的灵敏度为81.4%，特异性为73.9%，AUC为0.89，为资源受限的医疗部署建立了量子经典知识转移可行性。QuantumMedKD揭示了量子增强医疗保健人工智能系统在嘈杂的中等规模量子（NISQ）时代内外的基本原则和概念验证，以实现高效率和可部署性，为该领域的未来发展铺平了道路。

{"title":"QuantumMedKD: A hybrid quantum–classical knowledge distillation framework for medical image analysis","authors":"MD Nahid Hassan Nishan , Mohammad Junayed Hasan , M.R.C. Mahdy","doi":"10.1016/j.aej.2025.12.007","DOIUrl":"10.1016/j.aej.2025.12.007","url":null,"abstract":"<div><div>While hybrid quantum–classical architectures and knowledge distillation have each been explored independently in medical imaging and model compression, no prior work has unified these domains within a single framework. This study addresses this gap by investigating the integration of quantum–classical neural networks with knowledge distillation for medical image analysis, with a focus on circuit behavior, parameter efficiency, and diagnostic performance. The framework (QuantumMedKD) employs classical CNN architectures (ResNet-50, EfficientNet-B0, Xception) as teacher models, transferring learned representations to parameter-efficient quantum student networks via parameterized quantum circuits across qubit configurations (3-8), demonstrated through pneumonia detection from pediatric chest radiographs. Experimental validation reveals remarkable parameter efficiency, requiring only 24-36 trainable parameters compared to millions for classical counterparts, achieving compression ratios exceeding <span><math><mrow><mn>1</mn><msup><mrow><mn>0</mn></mrow><mrow><mn>5</mn></mrow></msup></mrow></math></span> while maintaining competitive diagnostic performance. Optimal configurations achieve 84.00% accuracy (EfficientNet-B0, 6-qubit) and 73.33% (Xception, 4-qubit), with knowledge distillation providing statistically significant improvements (<span><math><mi>p</mi></math></span>-vallues <span><math><mo><</mo></math></span> 0.001, Cohen’s d <span><math><mo>></mo></math></span> 2.0). Medical evaluation confirms diagnostic capability with sensitivity 81.4%, specificity 73.9%, and AUC 0.89, establishing quantum–classical knowledge transfer viability for resource-constrained healthcare deployment. QuantumMedKD reveals essential principles and proof-of-concept for quantum-enhanced healthcare AI systems within and beyond the noisy intermediate-scale quantum (NISQ) era for high efficiency and deployability, paving the way for future advancements in the field.</div></div>","PeriodicalId":7484,"journal":{"name":"alexandria engineering journal","volume":"134 ","pages":"Pages 49-68"},"PeriodicalIF":6.8,"publicationDate":"2025-12-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145684093","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}

引用次数: 0

EEG-based multi-scale perturbed brain cognitive pattern recognition network for gamer level classification 基于脑电图的多尺度摄动脑认知模式识别网络用于玩家等级分类

IF 6.8 2区工程技术 Q1 ENGINEERING, MULTIDISCIPLINARY

alexandria engineering journal

Pub Date : 2025-12-05 DOI: 10.1016/j.aej.2025.12.004

Lijun Jiang , Yanping Chen , Kexuan Liu , Xingyuan Chen , Li Dong , Weiyi Ma , Diankun Gong , Dezhong Yao

Electroencephalogram (EEG)-based classification of video game experts versus amateurs reveals cognitive brain patterns underlying complex abilities, with applications in attention modulation, medical rehabilitation, and health monitoring. While deep learning has advanced EEG-based cognitive state detection, challenges remain in extracting meaningful patterns from noisy data and preventing reverse inference attacks on user privacy. Here, we propose a novel multi-scale perturbed brain cognitive pattern recognition network (MsPE). Its key contributions are: (1) a multi-scale weak encryption method with attention mechanisms that protects privacy by perturbing EEG signals in temporal and frequency domains; (2) ConvFormer modules with adaptive channel sizes (3, 5, 15) and attention fusion to generate personalized perturbations while preserving task-relevant information; (3) a Denoise Feature Extraction Block (DFEB) using deep separable CNNs with skip connections to extract spatio-temporal features and reduce noise. Validated on a gaming EEG dataset, MsPE achieves 88.75% accuracy, 90.27% recall, 85.11% specificity, an F1 score of 0.8923, and a Kappa coefficient of 0.6957, outperforming existing methods. Interpretability analysis reveals distinct cognitive patterns between experts and amateurs in the temporal, occipital, and frontal lobes, with the most pronounced differences in the frontal lobe. This study advances an effective, secure, and accurate EEG-based cognitive pattern analysis solution.

基于脑电图（EEG）的电子游戏专家与业余爱好者的分类揭示了复杂能力背后的认知大脑模式，并在注意力调节、医疗康复和健康监测方面得到了应用。虽然深度学习具有先进的基于脑电图的认知状态检测，但在从噪声数据中提取有意义的模式和防止对用户隐私的反向推理攻击方面仍然存在挑战。本文提出了一种新的多尺度扰动脑认知模式识别网络（MsPE）。主要贡献有：(1)提出了一种基于注意机制的多尺度弱加密方法，该方法通过对脑电图信号进行时域和频域扰动来保护隐私；(2)具有自适应通道大小（3,5,15）和注意力融合的ConvFormer模块，在保留任务相关信息的同时产生个性化扰动；(3)利用带跳跃连接的深度可分离cnn提取时空特征并降低噪声的降噪特征提取块（DFEB）。在游戏脑电数据集上验证，MsPE的准确率为88.75%，召回率为90.27%，特异性为85.11%，F1得分为0.8923，Kappa系数为0.6957，优于现有方法。可解释性分析揭示了专家和业余爱好者在颞叶、枕叶和额叶的不同认知模式，其中额叶的差异最明显。本研究提出了一种有效、安全、准确的基于脑电图的认知模式分析解决方案。

{"title":"EEG-based multi-scale perturbed brain cognitive pattern recognition network for gamer level classification","authors":"Lijun Jiang , Yanping Chen , Kexuan Liu , Xingyuan Chen , Li Dong , Weiyi Ma , Diankun Gong , Dezhong Yao","doi":"10.1016/j.aej.2025.12.004","DOIUrl":"10.1016/j.aej.2025.12.004","url":null,"abstract":"<div><div>Electroencephalogram (EEG)-based classification of video game experts versus amateurs reveals cognitive brain patterns underlying complex abilities, with applications in attention modulation, medical rehabilitation, and health monitoring. While deep learning has advanced EEG-based cognitive state detection, challenges remain in extracting meaningful patterns from noisy data and preventing reverse inference attacks on user privacy. Here, we propose a novel multi-scale perturbed brain cognitive pattern recognition network (MsPE). Its key contributions are: (1) a multi-scale weak encryption method with attention mechanisms that protects privacy by perturbing EEG signals in temporal and frequency domains; (2) ConvFormer modules with adaptive channel sizes (3, 5, 15) and attention fusion to generate personalized perturbations while preserving task-relevant information; (3) a Denoise Feature Extraction Block (DFEB) using deep separable CNNs with skip connections to extract spatio-temporal features and reduce noise. Validated on a gaming EEG dataset, MsPE achieves 88.75% accuracy, 90.27% recall, 85.11% specificity, an F1 score of 0.8923, and a Kappa coefficient of 0.6957, outperforming existing methods. Interpretability analysis reveals distinct cognitive patterns between experts and amateurs in the temporal, occipital, and frontal lobes, with the most pronounced differences in the frontal lobe. This study advances an effective, secure, and accurate EEG-based cognitive pattern analysis solution.</div></div>","PeriodicalId":7484,"journal":{"name":"alexandria engineering journal","volume":"134 ","pages":"Pages 37-48"},"PeriodicalIF":6.8,"publicationDate":"2025-12-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145684092","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}

引用次数: 0

Edge-guided Multi-scale Attention Fusion Network for gastrointestinal tumor image classification 边缘引导多尺度注意力融合网络用于胃肠道肿瘤图像分类

IF 6.8 2区工程技术 Q1 ENGINEERING, MULTIDISCIPLINARY

alexandria engineering journal

Pub Date : 2025-12-05 DOI: 10.1016/j.aej.2025.11.042

Bin Xu , Qiaoli Lv , Chengyan Bian , Kangpeng Yan , Wenjie Fang , Jiao Cai , Sunmin Chen , Qi Wang , Yiming Zhao , Xingchen Wu , Hengrui Liu , Hua Li

Automated classification of gastrointestinal tumor images is a pivotal technology in computer-aided diagnostic systems. However, medical images are often affected by high-frequency noise due to limitations of acquisition devices and intrinsic tissue characteristics. In addition, lesion regions frequently exhibit blurred edges and low contrast, posing challenges for accurate extraction of discriminative features. To address these challenges, we propose a novel Edge-guided Multi-scale Attention Fusion Network (EdgeMAF-Net) for gastrointestinal tumor image classification. Specifically, we introduce a cross-stage partial fusion module that dynamically allocates features to both CNN and Transformer branches, enabling simultaneous modeling of local details and global context. This is complemented by a high-frequency attenuation and noise suppression mechanism, as well as a multi-scale edge attention calibration module, which integrates a three-stage enhancement strategy to capture features at different scales and delineate blurred boundaries. The module leverages a feature enhancement attention block to weight multi-source features, combined with a multi-scale edge enhancement block employing multi-scale pooling and edge extraction, and an adaptive gated fusion block to dynamically adjust feature fusion. EdgeMAF-Net outperforms existing methods in terms of accuracy, sensitivity, and boundary localization on the Chaoyang, Kvasir, and GasHisSDB datasets. Our code is available at https://github.com/Bambi-lab/EdgeMAF-Net.

胃肠道肿瘤图像的自动分类是计算机辅助诊断系统中的一项关键技术。然而，由于采集设备和固有组织特性的限制，医学图像经常受到高频噪声的影响。此外，病变区域经常出现边缘模糊和对比度低的情况，这给准确提取判别特征带来了挑战。为了解决这些挑战，我们提出了一种新的边缘引导的多尺度注意力融合网络（EdgeMAF-Net）用于胃肠道肿瘤图像分类。具体来说，我们引入了一个跨阶段部分融合模块，该模块可以动态地将特征分配给CNN和Transformer分支，从而能够同时对局部细节和全局上下文进行建模。此外还有高频衰减和噪声抑制机制，以及多尺度边缘注意力校准模块，该模块集成了三阶段增强策略，以捕获不同尺度的特征并描绘模糊的边界。该模块利用特征增强关注块对多源特征进行加权，结合采用多尺度池化和边缘提取的多尺度边缘增强块和自适应门控融合块对特征融合进行动态调整。在朝阳、Kvasir和GasHisSDB数据集上，EdgeMAF-Net在精度、灵敏度和边界定位方面优于现有方法。我们的代码可在https://github.com/Bambi-lab/EdgeMAF-Net上获得。

{"title":"Edge-guided Multi-scale Attention Fusion Network for gastrointestinal tumor image classification","authors":"Bin Xu , Qiaoli Lv , Chengyan Bian , Kangpeng Yan , Wenjie Fang , Jiao Cai , Sunmin Chen , Qi Wang , Yiming Zhao , Xingchen Wu , Hengrui Liu , Hua Li","doi":"10.1016/j.aej.2025.11.042","DOIUrl":"10.1016/j.aej.2025.11.042","url":null,"abstract":"<div><div>Automated classification of gastrointestinal tumor images is a pivotal technology in computer-aided diagnostic systems. However, medical images are often affected by high-frequency noise due to limitations of acquisition devices and intrinsic tissue characteristics. In addition, lesion regions frequently exhibit blurred edges and low contrast, posing challenges for accurate extraction of discriminative features. To address these challenges, we propose a novel Edge-guided Multi-scale Attention Fusion Network (EdgeMAF-Net) for gastrointestinal tumor image classification. Specifically, we introduce a cross-stage partial fusion module that dynamically allocates features to both CNN and Transformer branches, enabling simultaneous modeling of local details and global context. This is complemented by a high-frequency attenuation and noise suppression mechanism, as well as a multi-scale edge attention calibration module, which integrates a three-stage enhancement strategy to capture features at different scales and delineate blurred boundaries. The module leverages a feature enhancement attention block to weight multi-source features, combined with a multi-scale edge enhancement block employing multi-scale pooling and edge extraction, and an adaptive gated fusion block to dynamically adjust feature fusion. EdgeMAF-Net outperforms existing methods in terms of accuracy, sensitivity, and boundary localization on the Chaoyang, Kvasir, and GasHisSDB datasets. Our code is available at <span><span>https://github.com/Bambi-lab/EdgeMAF-Net</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":7484,"journal":{"name":"alexandria engineering journal","volume":"134 ","pages":"Pages 21-36"},"PeriodicalIF":6.8,"publicationDate":"2025-12-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145665623","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}

引用次数: 0

A dynamic mesh refinement in Physic-Informed Neural Networks for handling Stiff Partial Differential Equations 在物理信息神经网络中处理刚性偏微分方程的动态网格细化

IF 6.8 2区工程技术 Q1 ENGINEERING, MULTIDISCIPLINARY

alexandria engineering journal

Pub Date : 2025-12-04 DOI: 10.1016/j.aej.2025.11.041

Nursyiva Irsalinda , Maharani A. Bakar , Fatimah Noor Harun , Ruwaidiah Idris , Sugiyarto Surono

Stiff Partial Differential Equations (PDEs) pose considerable challenges due to the presence of steep gradients, which often lead to large residuals when employing coarse training mesh grids, while excessively refined grids substantially increase computational costs. To address this trade off, we propose an Augmented Mesh Refinement Physics Informed Neural Network (Aug-MR PINN), a novel framework that dynamically allocates mesh points in regions with sharp gradients. Unlike existing approaches, Aug-MR PINN adaptively adjusts the mesh distribution during training by evaluating residual based criteria, thereby maintaining a balance between accuracy and efficiency. The method was validated on three representative stiff PDEs, namely Burgers’, Telegraph, and Flow Mixing equations, under two scenarios which are percentile and threshold based refinement. Our key findings show that Aug-MR PINN consistently achieves lower loss values and improved convergence compared to state of the art methods such as Adaptive Mesh Refinement with Finite Element Method (AMR-FEM), basic PINN, Restarting PINN (R-PINN), and Extended PINN (X-PINN). Moreover, both refinement strategies yielded comparable performance, confirming the robustness of the proposed approach. Beyond stiff PDEs, this framework provides a promising foundation for broader applications in computational physics, such as fluid dynamics, wave propagation, and multiphase systems.

由于存在陡峭的梯度，刚性偏微分方程（PDEs）带来了相当大的挑战，当使用粗糙的训练网格网格时，往往会导致较大的残差，而过度精细的网格大大增加了计算成本。为了解决这个问题，我们提出了一种增强网格细化物理通知神经网络（augr - mr PINN），这是一种新的框架，可以在具有尖锐梯度的区域中动态分配网格点。与现有方法不同，augr - mr PINN在训练过程中通过评估残差标准自适应调整网格分布，从而保持准确性和效率之间的平衡。在基于百分位数和阈值的两种改进方案下，对三个具有代表性的刚性偏微分方程（Burgers’s、Telegraph’s和Flow Mixing equations）进行了验证。我们的主要研究结果表明，与有限元法（AMR-FEM）、基本PINN、重新启动PINN （R-PINN）和扩展PINN （X-PINN）等最先进的方法相比，augr - mr PINN始终具有更低的损失值和更好的收敛性。此外，两种优化策略产生了相当的性能，证实了所提出方法的鲁棒性。除了刚性偏微分方程之外，该框架还为计算物理（如流体动力学、波传播和多相系统）的更广泛应用提供了有希望的基础。

{"title":"A dynamic mesh refinement in Physic-Informed Neural Networks for handling Stiff Partial Differential Equations","authors":"Nursyiva Irsalinda , Maharani A. Bakar , Fatimah Noor Harun , Ruwaidiah Idris , Sugiyarto Surono","doi":"10.1016/j.aej.2025.11.041","DOIUrl":"10.1016/j.aej.2025.11.041","url":null,"abstract":"<div><div>Stiff Partial Differential Equations (PDEs) pose considerable challenges due to the presence of steep gradients, which often lead to large residuals when employing coarse training mesh grids, while excessively refined grids substantially increase computational costs. To address this trade off, we propose an Augmented Mesh Refinement Physics Informed Neural Network (Aug-MR PINN), a novel framework that dynamically allocates mesh points in regions with sharp gradients. Unlike existing approaches, Aug-MR PINN adaptively adjusts the mesh distribution during training by evaluating residual based criteria, thereby maintaining a balance between accuracy and efficiency. The method was validated on three representative stiff PDEs, namely Burgers’, Telegraph, and Flow Mixing equations, under two scenarios which are percentile and threshold based refinement. Our key findings show that Aug-MR PINN consistently achieves lower loss values and improved convergence compared to state of the art methods such as Adaptive Mesh Refinement with Finite Element Method (AMR-FEM), basic PINN, Restarting PINN (R-PINN), and Extended PINN (X-PINN). Moreover, both refinement strategies yielded comparable performance, confirming the robustness of the proposed approach. Beyond stiff PDEs, this framework provides a promising foundation for broader applications in computational physics, such as fluid dynamics, wave propagation, and multiphase systems.</div></div>","PeriodicalId":7484,"journal":{"name":"alexandria engineering journal","volume":"134 ","pages":"Pages 1-20"},"PeriodicalIF":6.8,"publicationDate":"2025-12-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145665622","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}

引用次数: 0

Developing a volume delay function (VDF) for the urban roads of the city of Alexandria 为亚历山大市的城市道路开发一个体积延迟函数（VDF）

IF 6.8 2区工程技术 Q1 ENGINEERING, MULTIDISCIPLINARY

alexandria engineering journal

Pub Date : 2025-12-01 DOI: 10.1016/j.aej.2025.10.027

Mounir M. Abdelaal, Sherif Ebeido, Wael Bekheet, Hassan El Ghazoly

The Volume Delay Function (VDF) is crucial in trip assignment for calculating congested travel time on road networks. The BPR equation, widely used for VDF, requires calibration of parameters to accurately estimate travel time. Inaccurate travel time estimation leads to unreliable trip assignments and travel demand forecasting. Thus, calibrating BPR parameters (α, β and c) for local conditions is essential. This research used Alexandria as a case study, collecting data of link travel times and traffic volumes were collected. The calibration process utilizes nonlinear optimization techniques to optimize the BPR parameters. Data collection covers 66 sites with a total of 198 h of video taping which comprise 632 observatoins (15 min intervals). Three steps were used to calibrate the parameters, visual fitting helped getting rid of outliers. Later, relaxed linearized parameter estimation was performed. The resulted parameters (without capacity) were less roboust, yet, with better goodness of fit due to the relaxation conditions. The final calibration incorporated the capacity resulted in a more comparable parameter values with lowers goodness of fit. The values of parameters (α, β, c) were estimated for different road types and all types combined. The parameter values showed excellent trending confirming the estimation validity.

体积延迟函数（VDF）是计算路网拥挤时间的重要方法。广泛用于VDF的BPR方程需要对参数进行校准才能准确估计行程时间。不准确的出行时间估计导致不可靠的出行分配和出行需求预测。因此，根据当地条件校准BPR参数（α， β和c）是必不可少的。本研究以亚历山德里亚市为个案研究对象，收集了各路段行车时间和交通量的数据。校准过程采用非线性优化技术来优化BPR参数。数据收集覆盖66个站点，共198 h录像，包括632次观测（15 分钟间隔）。采用三个步骤对参数进行校正，视觉拟合有助于去除异常值。然后进行松弛线性化参数估计。得到的参数（无容量）鲁棒性较差，但由于松弛条件的影响，拟合优度较好。最终校正纳入容量导致参数值更具可比性，但拟合优度较低。对不同道路类型和所有道路类型的组合进行了参数（α, β, c）的估计。参数值呈现良好的趋势，证实了估计的有效性。

{"title":"Developing a volume delay function (VDF) for the urban roads of the city of Alexandria","authors":"Mounir M. Abdelaal, Sherif Ebeido, Wael Bekheet, Hassan El Ghazoly","doi":"10.1016/j.aej.2025.10.027","DOIUrl":"10.1016/j.aej.2025.10.027","url":null,"abstract":"<div><div>The Volume Delay Function (VDF) is crucial in trip assignment for calculating congested travel time on road networks. The BPR equation, widely used for VDF, requires calibration of parameters to accurately estimate travel time. Inaccurate travel time estimation leads to unreliable trip assignments and travel demand forecasting. Thus, calibrating BPR parameters (α, β and c) for local conditions is essential. This research used Alexandria as a case study, collecting data of link travel times and traffic volumes were collected. The calibration process utilizes nonlinear optimization techniques to optimize the BPR parameters. Data collection covers 66 sites with a total of 198 h of video taping which comprise 632 observatoins (15 min intervals). Three steps were used to calibrate the parameters, visual fitting helped getting rid of outliers. Later, relaxed linearized parameter estimation was performed. The resulted parameters (without capacity) were less roboust, yet, with better goodness of fit due to the relaxation conditions. The final calibration incorporated the capacity resulted in a more comparable parameter values with lowers goodness of fit. The values of parameters (α, β, c) were estimated for different road types and all types combined. The parameter values showed excellent trending confirming the estimation validity.</div></div>","PeriodicalId":7484,"journal":{"name":"alexandria engineering journal","volume":"133 ","pages":"Pages 528-537"},"PeriodicalIF":6.8,"publicationDate":"2025-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145615818","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}

引用次数: 0

A novel framework for detection of noncommunicable diseases via prompt engineering and domain knowledge integration 基于快速工程和领域知识整合的非传染性疾病检测新框架

IF 6.8 2区工程技术 Q1 ENGINEERING, MULTIDISCIPLINARY

alexandria engineering journal

Pub Date : 2025-12-01 DOI: 10.1016/j.aej.2025.11.050

Mohammad Junayed Hasan , Suhra Noor , Sifat Momen

Noncommunicable diseases (NCDs) are the leading cause of global morbidity, yet existing AI models rely predominantly on clinical data, neglecting rich information available in population surveys. This study introduces a novel framework integrating prompt-based domain knowledge extraction with traditional machine learning for NCD detection from survey data. Unlike knowledge graph construction or retrieval-augmented generation approaches requiring extensive infrastructure, our method employs prompt engineering to systematically elicit epidemiological expertise from large language models for feature selection, validated against independent medical expert rankings with strong concordance (Kendall’s tau of 0.71). The dual-track approach fuses prompt-guided feature screening with Recursive Feature Elimination and Cross-Validation, reconciling domain priors with statistical rigor. Applied to Danish population survey data with 18,957 respondents across 243 variables, the framework achieves F1-score of 91.35%, precision of 98.59%, and accuracy of 85.25%, outperforming existing methods by 2%–3% in F1-score and 3%–10% in accuracy. Synthetic data validation confirms robustness across diverse data distributions. Local Interpretable Model-agnostic Explanations reveal that medical expenditure, physical mobility, and self-rated health constitute primary predictors, enabling clinically actionable risk stratification. A deployed web interface demonstrates real-world applicability with sub-100ms inference. The work establishes prompt-mediated knowledge integration as a scalable pathway for healthcare prediction from population-level determinants.

非传染性疾病（NCDs）是全球发病率的主要原因，但现有的人工智能模型主要依赖临床数据，忽视了人口调查中提供的丰富信息。本文提出了一种新的框架，将基于提示的领域知识提取与传统的机器学习相结合，用于从调查数据中检测非传染性疾病。与需要大量基础设施的知识图谱构建或检索增强生成方法不同，我们的方法采用快速工程，系统地从大型语言模型中提取流行病学专业知识进行特征选择，并根据具有强一致性的独立医学专家排名进行验证（Kendall 's tau为0.71）。双轨方法融合了即时引导特征筛选、递归特征消除和交叉验证，调和了领域先验和统计严谨性。将该框架应用于丹麦人口调查数据，共有18957名受访者，涉及243个变量，f1得分为91.35%，精度为98.59%，准确度为85.25%，比现有方法的f1得分和准确度分别提高了2%-3%和3%-10%。合成数据验证确认了跨不同数据分布的稳健性。本地可解释的模型不可知论解释表明，医疗支出、身体活动能力和自评健康构成主要预测因素，从而实现临床可操作的风险分层。一个已部署的web界面演示了低于100ms推理的实际适用性。这项工作建立了即时介导的知识整合作为一个可扩展的途径，从人口水平的决定因素的医疗保健预测。

{"title":"A novel framework for detection of noncommunicable diseases via prompt engineering and domain knowledge integration","authors":"Mohammad Junayed Hasan , Suhra Noor , Sifat Momen","doi":"10.1016/j.aej.2025.11.050","DOIUrl":"10.1016/j.aej.2025.11.050","url":null,"abstract":"<div><div>Noncommunicable diseases (NCDs) are the leading cause of global morbidity, yet existing AI models rely predominantly on clinical data, neglecting rich information available in population surveys. This study introduces a novel framework integrating prompt-based domain knowledge extraction with traditional machine learning for NCD detection from survey data. Unlike knowledge graph construction or retrieval-augmented generation approaches requiring extensive infrastructure, our method employs prompt engineering to systematically elicit epidemiological expertise from large language models for feature selection, validated against independent medical expert rankings with strong concordance (Kendall’s tau of 0.71). The dual-track approach fuses prompt-guided feature screening with Recursive Feature Elimination and Cross-Validation, reconciling domain priors with statistical rigor. Applied to Danish population survey data with 18,957 respondents across 243 variables, the framework achieves F1-score of 91.35%, precision of 98.59%, and accuracy of 85.25%, outperforming existing methods by 2%–3% in F1-score and 3%–10% in accuracy. Synthetic data validation confirms robustness across diverse data distributions. Local Interpretable Model-agnostic Explanations reveal that medical expenditure, physical mobility, and self-rated health constitute primary predictors, enabling clinically actionable risk stratification. A deployed web interface demonstrates real-world applicability with sub-100ms inference. The work establishes prompt-mediated knowledge integration as a scalable pathway for healthcare prediction from population-level determinants.</div></div>","PeriodicalId":7484,"journal":{"name":"alexandria engineering journal","volume":"133 ","pages":"Pages 586-614"},"PeriodicalIF":6.8,"publicationDate":"2025-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145681625","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}

引用次数: 0

Multi-objective optimization of enhanced stator permanent magnet hybrid stepping motor based on Taguchi and respond surface method 基于田口法和响应面法的增强型定子永磁混合式步进电机多目标优化

IF 6.8 2区工程技术 Q1 ENGINEERING, MULTIDISCIPLINARY

alexandria engineering journal

Pub Date : 2025-12-01 DOI: 10.1016/j.aej.2025.11.046

Xiaobao Chai, Jinglin Liu

This paper presents an enhanced stator permanent magnet hybrid stepping motor (ESPMHSM) as a promising HSM. However, the motor involves multiple structural parameters, and the interactions among these parameters substantially influence motor performance during the optimization process. In order to address this issue, this paper proposes a multi-objective optimization method integrating the Taguchi method and the Response Surface method (RSM). The detent torque, torque density, torque ripple and PM weight are selected as the optimization objectives. Firstly, the Taguchi method is used to construct an orthogonal experimental matrix. Subsequently, the trends and proportions of the influence of the optimization variables on the optimization objectives are analyzed, and the optimal ranges for each optimization variable are determined. Secondly, the optimization functions are formulated using the least-squares method and the solutions of these functions are derived through the RSM. Then, the electromagnetic performance between the optimized and the initial motors is analyzed. Finally, a prototype is manufactured and tested. The results validate the feasibility of the optimization method and the correctness of the analyses.

本文提出了一种增强型定子永磁混合式步进电机（ESPMHSM）。然而，电机涉及多个结构参数，在优化过程中，这些参数之间的相互作用对电机性能有很大影响。为了解决这一问题，本文提出了一种将田口法与响应面法相结合的多目标优化方法。选取制动力矩、转矩密度、转矩脉动和PM权重作为优化目标。首先，采用田口法构造正交实验矩阵。随后，分析各优化变量对优化目标影响的趋势和比例，确定各优化变量的最优范围。其次，利用最小二乘法构造了优化函数，并通过最小二乘法求出了优化函数的解。然后，对优化后的电机与初始电机的电磁性能进行了分析。最后，制作了样机并进行了测试。结果验证了优化方法的可行性和分析的正确性。

{"title":"Multi-objective optimization of enhanced stator permanent magnet hybrid stepping motor based on Taguchi and respond surface method","authors":"Xiaobao Chai, Jinglin Liu","doi":"10.1016/j.aej.2025.11.046","DOIUrl":"10.1016/j.aej.2025.11.046","url":null,"abstract":"<div><div>This paper presents an enhanced stator permanent magnet hybrid stepping motor (ESPMHSM) as a promising HSM. However, the motor involves multiple structural parameters, and the interactions among these parameters substantially influence motor performance during the optimization process. In order to address this issue, this paper proposes a multi-objective optimization method integrating the Taguchi method and the Response Surface method (RSM). The detent torque, torque density, torque ripple and PM weight are selected as the optimization objectives. Firstly, the Taguchi method is used to construct an orthogonal experimental matrix. Subsequently, the trends and proportions of the influence of the optimization variables on the optimization objectives are analyzed, and the optimal ranges for each optimization variable are determined. Secondly, the optimization functions are formulated using the least-squares method and the solutions of these functions are derived through the RSM. Then, the electromagnetic performance between the optimized and the initial motors is analyzed. Finally, a prototype is manufactured and tested. The results validate the feasibility of the optimization method and the correctness of the analyses.</div></div>","PeriodicalId":7484,"journal":{"name":"alexandria engineering journal","volume":"133 ","pages":"Pages 564-576"},"PeriodicalIF":6.8,"publicationDate":"2025-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145681632","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}

引用次数: 0