Advanced Engineering Informatics最新文献

Bridge substructure damage morphology identification based on the underwater sonar point cloud data 基于水下声纳点云数据的桥梁下部结构损伤形态识别

IF 8 1区工程技术 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE

Advanced Engineering Informatics

Pub Date : 2024-10-01 DOI: 10.1016/j.aei.2024.102936

Shuaihui Zhang , Yanjie Zhu , Wen Xiong , C.S. Cai , Jinquan Zhang

Bridge underwater foundation inspection is always a prominent and challenging issue due to an unknown and unsafe underwater environment. Effective identification of bridge foundations is significant for the safety assessment of water-related bridges. However, due to the interference of numerous objective factors in the water environment (e.g., water quality, flow velocity, water depth, etc.), reliable and valid data are often difficult to obtain, and the inspection of bridge substructures remains a major challenge, especially for deep water bridge foundations. To solve this problem, a damage morphology identification method based on underwater sonar point cloud data (USPCD) is proposed in this paper for underwater bridge structures. The method is divided into two stages, including potential damage region attention and fine damage morphology identification. The former considers the regional connectivity properties of the damage, focusing on potential damage regions employing a curve fitting method based on iterative median absolute deviation. The latter gives a significant density difference between intact and damaged regions based on the density-based spatial clustering of applications with the noise clustering method to separate damaged data points from normal data points while preserving fine damage morphology features. Based on the swept sonar point cloud of underwater piles from a cross-Yangtze River bridge, we simulated spalling and cavity damage at different scales to comprehensively evaluate our proposed method. The results show that the method can detect damage at different scales and can identify most of the damaged regions. For larger-scale damage, four evaluation indicators are kept at a high level, in which the maximum GTOR and IOU can reach 95.8 % and 85.9 %, respectively. For small-scale damage, based on the synthesized high-resolution point cloud, the method can accurately identify even the damage as small as 12 cm with GTOR above 94 % and IOU over 85 %.

由于未知和不安全的水下环境，桥梁水下地基检测一直是一个突出和具有挑战性的问题。桥梁地基的有效识别对于涉水桥梁的安全评估意义重大。然而，由于水环境中众多客观因素（如水质、流速、水深等）的干扰，往往难以获得可靠有效的数据，桥梁下部结构检测仍是一大难题，尤其是深水桥梁地基。为解决这一问题，本文提出了一种基于水下声纳点云数据（USPCD）的水下桥梁结构损伤形态识别方法。该方法分为两个阶段，包括潜在损伤区域关注和精细损伤形态识别。前者考虑了损伤的区域连通性，采用基于迭代绝对偏差中值的曲线拟合方法重点关注潜在损伤区域。后者在基于密度的空间聚类应用的基础上，利用噪声聚类方法给出了完整区域和受损区域之间的显著密度差异，从而将受损数据点从正常数据点中分离出来，同时保留了精细的损伤形态特征。基于某跨长江大桥水下桩基的声纳扫频点云，我们模拟了不同尺度的剥落和空洞损伤，以全面评估我们提出的方法。结果表明，该方法可以检测到不同尺度的损坏，并能识别大部分损坏区域。对于较大尺度的损伤，四项评价指标均保持在较高水平，其中最大 GTOR 和 IOU 分别可达 95.8 % 和 85.9 %。对于小尺度损伤，基于合成的高分辨率点云，该方法甚至可以准确识别小至 12 厘米的损伤，GTOR 超过 94%，IOU 超过 85%。

{"title":"Bridge substructure damage morphology identification based on the underwater sonar point cloud data","authors":"Shuaihui Zhang , Yanjie Zhu , Wen Xiong , C.S. Cai , Jinquan Zhang","doi":"10.1016/j.aei.2024.102936","DOIUrl":"10.1016/j.aei.2024.102936","url":null,"abstract":"<div><div>Bridge underwater foundation inspection is always a prominent and challenging issue due to an unknown and unsafe underwater environment. Effective identification of bridge foundations is significant for the safety assessment of water-related bridges. However, due to the interference of numerous objective factors in the water environment (e.g., water quality, flow velocity, water depth, etc.), reliable and valid data are often difficult to obtain, and the inspection of bridge substructures remains a major challenge, especially for deep water bridge foundations. To solve this problem, a damage morphology identification method based on underwater sonar point cloud data (USPCD) is proposed in this paper for underwater bridge structures. The method is divided into two stages, including potential damage region attention and fine damage morphology identification. The former considers the regional connectivity properties of the damage, focusing on potential damage regions employing a curve fitting method based on iterative median absolute deviation. The latter gives a significant density difference between intact and damaged regions based on the density-based spatial clustering of applications with the noise clustering method to separate damaged data points from normal data points while preserving fine damage morphology features. Based on the swept sonar point cloud of underwater piles from a cross-Yangtze River bridge, we simulated spalling and cavity damage at different scales to comprehensively evaluate our proposed method. The results show that the method can detect damage at different scales and can identify most of the damaged regions. For larger-scale damage, four evaluation indicators are kept at a high level, in which the maximum <em>GTOR</em> and <em>IOU</em> can reach 95.8 % and 85.9 %, respectively. For small-scale damage, based on the synthesized high-resolution point cloud, the method can accurately identify even the damage as small as 12 cm with <em>GTOR</em> above 94 % and <em>IOU</em> over 85 %.</div></div>","PeriodicalId":50941,"journal":{"name":"Advanced Engineering Informatics","volume":"62 ","pages":"Article 102936"},"PeriodicalIF":8.0,"publicationDate":"2024-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142659115","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"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

Fault diagnosis of blast furnace based on incomplete multi-source domain adaptation with feature fusion 基于特征融合的不完全多源域适应的高炉故障诊断

IF 8 1区工程技术 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE

Advanced Engineering Informatics

Pub Date : 2024-10-01 DOI: 10.1016/j.aei.2024.102946

Dali Gao, Chunjie Yang, Xiao-Yu Tang, Xiongzhuo Zhu, Xiaoke Huang

Aiming at the model mismatch caused by changes in data distribution, transfer learning (TL) has been introduced to fault diagnosis of the blast furnace (BF) ironmaking process. However, most existing TL methods require that the category space of each source and target domain be identical, and ignore the semantic information of multi-source data under domain adaptation. To address these issues, we propose a novel method based on incomplete multi-source domain adaptation with feature fusion for fault diagnosis of BF. Firstly, a multi-scale convolutional network is set to effectively extract diverse features while enabling information interaction through point-wise convolution. Secondly, Transfer Vision Transformer is constructed for each source domain to fuse global and local features, and extract domain-specific knowledge with more semantic information. Finally, the model weights each source classifier based on the inter-domain similarity to obtain the result. Experiments on actual BF data validate the effectiveness of the proposed method.

针对数据分布变化引起的模型不匹配问题，迁移学习（TL）被引入高炉（BF）炼铁过程的故障诊断中。然而，现有的迁移学习方法大多要求每个源域和目标域的类别空间完全相同，而忽略了多源数据在域适应下的语义信息。针对这些问题，我们提出了一种基于不完全多源域适应与特征融合的新方法，用于 BF 故障诊断。首先，我们设置了一个多尺度卷积网络，以有效提取各种特征，同时通过点向卷积实现信息交互。其次，为每个源域构建转移视觉变换器，以融合全局和局部特征，并提取具有更多语义信息的特定域知识。最后，该模型根据域间相似性对每个源分类器进行加权，从而得出结果。实际 BF 数据实验验证了所提方法的有效性。

引用次数: 0

3D guiding assisted augmented assembly technology with rapid object detection in dynamic environment 在动态环境中快速检测物体的 3D 导向辅助增强装配技术

IF 8 1区工程技术 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE

Advanced Engineering Informatics

Pub Date : 2024-10-01 DOI: 10.1016/j.aei.2024.102857

Chengshun Li, Xiaonan Yang, Yaoguang Hu, Shangsi Wu, Jingfei Wang, Peng Wang

In the field of industrial assembly, augmented reality (AR) technology has played an important role and demonstrated its enormous development potential in the future. With the current development of product assembly towards customization and diversification, it is difficult to meet the requirements of augmented assembly (AA) by relying on static instructions registered with markers. However, most augmented assembly guidance systems used for dynamic environments are complex, cumbersome, and exhibit high latency, significantly impacting the user experience. In addition, the narrow field of view (Fov) of AR glasses also limits its further application in industrial scene. In response to the above issues, this article proposes an improved 3D guiding assisted augmented assembly technology. Firstly, a lightweight model Yolov7-Slim is proposed to achieve object detection on 2D images, which reduces File size by 26.7 % and improves running speed by 15.3 % compared to the Yolov7-tiny model. Secondly, a 3D positioning algorithm is proposed to achieve the rapid conversion of 2D coordinates to 3D coordinates. Finally, a user-oriented two-stage guidance mechanism is designed to compensate for the limitation of the narrow Fov of AR glasses. To quantify the performance of proposed technology, a 3D guiding assisted augmented assembly system (3DG3AS) was developed and validated in a reducer assembly experiment.

在工业装配领域，增强现实（AR）技术发挥了重要作用，并展示了其未来巨大的发展潜力。当前，产品装配正朝着定制化和多样化方向发展，仅靠标记注册的静态指令已难以满足增强装配（AA）的要求。然而，大多数用于动态环境的增强装配引导系统都非常复杂、繁琐，而且延迟较高，严重影响了用户体验。此外，AR 眼镜的视野（Fov）狭窄也限制了其在工业场景中的进一步应用。针对上述问题，本文提出了一种改进的 3D 导向辅助增强装配技术。首先，提出了一种轻量级模型 Yolov7-Slim 来实现二维图像上的物体检测，与 Yolov7-tiny 模型相比，文件大小减少了 26.7%，运行速度提高了 15.3%。其次，提出了一种三维定位算法，以实现二维坐标到三维坐标的快速转换。最后，设计了一种面向用户的两阶段引导机制，以弥补 AR 眼镜视场角窄的限制。为了量化所提技术的性能，我们开发了三维导向辅助增强装配系统（3DG3AS），并在减速器装配实验中进行了验证。

{"title":"3D guiding assisted augmented assembly technology with rapid object detection in dynamic environment","authors":"Chengshun Li, Xiaonan Yang, Yaoguang Hu, Shangsi Wu, Jingfei Wang, Peng Wang","doi":"10.1016/j.aei.2024.102857","DOIUrl":"10.1016/j.aei.2024.102857","url":null,"abstract":"<div><div>In the field of industrial assembly, augmented reality (AR) technology has played an important role and demonstrated its enormous development potential in the future. With the current development of product assembly towards customization and diversification, it is difficult to meet the requirements of augmented assembly (AA) by relying on static instructions registered with markers. However, most augmented assembly guidance systems used for dynamic environments are complex, cumbersome, and exhibit high latency, significantly impacting the user experience. In addition, the narrow field of view (Fov) of AR glasses also limits its further application in industrial scene. In response to the above issues, this article proposes an improved 3D guiding assisted augmented assembly technology. Firstly, a lightweight model Yolov7-Slim is proposed to achieve object detection on 2D images, which reduces File size by 26.7 % and improves running speed by 15.3 % compared to the Yolov7-tiny model. Secondly, a 3D positioning algorithm is proposed to achieve the rapid conversion of 2D coordinates to 3D coordinates. Finally, a user-oriented two-stage guidance mechanism is designed to compensate for the limitation of the narrow Fov of AR glasses. To quantify the performance of proposed technology, a 3D guiding assisted augmented assembly system (3DG3AS) was developed and validated in a reducer assembly experiment.</div></div>","PeriodicalId":50941,"journal":{"name":"Advanced Engineering Informatics","volume":"62 ","pages":"Article 102857"},"PeriodicalIF":8.0,"publicationDate":"2024-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142417045","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"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

Generating TRIZ-inspired guidelines for eco-design using Generative Artificial Intelligence 利用生成式人工智能生成受 TRIZ 启发的生态设计指南

IF 8 1区工程技术 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE

Advanced Engineering Informatics

Pub Date : 2024-10-01 DOI: 10.1016/j.aei.2024.102846

C.K.M. Lee , Jingying Liang , K.L. Yung , K.L. Keung

Environmental considerations are emerging as stimuli for innovation during the eco-design ideation process. Integrating TRIZ (Teoriya Resheniya Izobretatelskikh Zadatch─Theory of Inventive Problem Solving) methodology into eco-design offers a structured problem-solving approach to address sustainability challenges. However, developing innovative designs requires expertise in TRIZ concepts and access to resources, which makes it a time-consuming process and can limit its application for eco-design innovation quickly. This study leverages the analytical and generative capabilities of large language models (LLMs) to enhance the TRIZ methodology and automate the ideation process in eco-design. An intelligent tool, “Eco-innovate Assistant,” is designed to provide users with eco-innovative solutions with design sketches. Its effectiveness is validated and evaluated through comparative studies. The findings demonstrate the potential of LLMs in automating design processes, catalyzing a transformation in AI-driven innovation and ideation in eco-design.

在生态设计构思过程中，环境因素正在成为创新的刺激因素。将 TRIZ（Teoriya Resheniya Izobretatelskikh Zadatch──发明性问题解决理论）方法融入生态设计，为应对可持续性挑战提供了一种结构化的问题解决方法。然而，开发创新设计需要 TRIZ 概念方面的专业知识和资源，这使其成为一个耗时的过程，并可能限制其在生态设计创新中的快速应用。本研究利用大型语言模型（LLMs）的分析和生成能力来增强TRIZ方法，并使生态设计中的构思过程自动化。研究设计了一个智能工具 "生态创新助手"，通过设计草图为用户提供生态创新解决方案。通过比较研究对其有效性进行了验证和评估。研究结果表明了 LLM 在设计流程自动化方面的潜力，催化了人工智能驱动的创新和生态设计构思的变革。

引用次数: 0

An ensembled multilabel classification method for the short-circuit detection of electrolytic refining 用于电解精炼短路检测的集合多标签分类法

IF 8 1区工程技术 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE

Advanced Engineering Informatics

Pub Date : 2024-10-01 DOI: 10.1016/j.aei.2024.102919

Yusi Dai , Chunhua Yang , Hongqiu Zhu , Can Zhou

Short-circuits occurring in the electrolytic refining process of non-ferrous smelting are a main factor that consumes extra energy and affects the metal quality. This paper proposes an ensembled multilabel classification method for short-circuit detection based on infrared images and makes up for the defect of previous methods using object-detection neural networks being hard to directly apply in industrial sites. Different from the object-detection methods, the multilabel classification method does not output the imaging positions but directly obtains the realistic positions, i.e. plate numbers, of the faulty plates. By introducing a new convolutional neural network named FlatNet, no extra work is required to get the realistic positions of the faulty plates. To address the data imbalance inherent to multilabel classification, dynamic weights that pay more attention both to the minority class and difficult samples are presented, forming a bilateral constraint on the missed and the false detections. At the end of the method, we design a greedy ensemble approach driven by validation F1-scores for the promotion of detection performance and stability. Experiments conducted with real-world data verify the effectiveness of the proposed fault detection method.

有色金属冶炼的电解精炼过程中发生的短路是消耗额外能源和影响金属质量的主要因素。本文提出了一种基于红外图像的短路检测集合多标签分类方法，弥补了以往使用对象检测神经网络的方法难以直接应用于工业现场的缺陷。与物体检测方法不同的是，多标签分类方法不输出成像位置，而是直接获取故障车牌的实际位置，即车牌号码。通过引入名为 FlatNet 的新卷积神经网络，无需额外工作即可获得故障车牌的实际位置。为了解决多标签分类中固有的数据不平衡问题，我们提出了同时关注少数类和困难样本的动态权重，从而对漏检和误检形成了双边约束。在方法的最后，我们设计了一种由验证 F1 分数驱动的贪婪集合方法，以提高检测性能和稳定性。利用真实世界数据进行的实验验证了所提出的故障检测方法的有效性。

{"title":"An ensembled multilabel classification method for the short-circuit detection of electrolytic refining","authors":"Yusi Dai , Chunhua Yang , Hongqiu Zhu , Can Zhou","doi":"10.1016/j.aei.2024.102919","DOIUrl":"10.1016/j.aei.2024.102919","url":null,"abstract":"<div><div>Short-circuits occurring in the electrolytic refining process of non-ferrous smelting are a main factor that consumes extra energy and affects the metal quality. This paper proposes an ensembled multilabel classification method for short-circuit detection based on infrared images and makes up for the defect of previous methods using object-detection neural networks being hard to directly apply in industrial sites. Different from the object-detection methods, the multilabel classification method does not output the imaging positions but directly obtains the realistic positions, i.e. plate numbers, of the faulty plates. By introducing a new convolutional neural network named FlatNet, no extra work is required to get the realistic positions of the faulty plates. To address the data imbalance inherent to multilabel classification, dynamic weights that pay more attention both to the minority class and difficult samples are presented, forming a bilateral constraint on the missed and the false detections. At the end of the method, we design a greedy ensemble approach driven by validation F1-scores for the promotion of detection performance and stability. Experiments conducted with real-world data verify the effectiveness of the proposed fault detection method.</div></div>","PeriodicalId":50941,"journal":{"name":"Advanced Engineering Informatics","volume":"62 ","pages":"Article 102919"},"PeriodicalIF":8.0,"publicationDate":"2024-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142578041","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"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

Research on click enhancement strategy of hand-eye dual-channel human-computer interaction system: Trade-off between sensing area and area cursor 手眼双通道人机交互系统的点击增强策略研究：感应区域与区域光标之间的权衡

IF 8 1区工程技术 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE

Advanced Engineering Informatics

Pub Date : 2024-10-01 DOI: 10.1016/j.aei.2024.102880

Ya-Feng Niu, Rui Chen, Yi-Yan Wang, Xue-Ying Yao, Yun Feng

This study aims to explore the application of click enhancement strategies in a “Sight Line Localization + Hand Triggered” eye-control human–computer interaction system, proposes a click enhancement strategy to solve two critical problems in eye-control human–computer interaction: Midas touch and low spatial accuracy. By conducting ergonomics experiments, we verify that the proposed click enhancement strategy can effectively improve the operational performance of hand-eye dual-channel HCI systems. The experimental results show that the accuracy of the operation can be significantly improved by using a cursor size equal to the diameter of the interaction control and a sensing area size of 1.8 times the diameter of the control. Based on the comprehensive consideration of operation efficiency and comfort, 0.75 times the control diameter of the cursor and 1.8 times the control diameter of the sensing area are the optimal parameter configurations. The results not only solve the problems of Midas touch and low spatial accuracy but also significantly reduce visual fatigue, thus improving the ease of use and robustness of the hand-eye dual-channel human–computer interaction system.

本研究旨在探索点击增强策略在 "视线定位+手触发 "眼控人机交互系统中的应用，提出了一种点击增强策略，以解决眼控人机交互中的两个关键问题：Midas touch 和低空间精度。通过人机工程学实验，我们验证了所提出的点击增强策略能有效改善手眼双通道人机交互系统的操作性能。实验结果表明，光标大小等于交互控制器直径，感应区域大小为控制器直径的 1.8 倍，可以显著提高操作精度。在综合考虑操作效率和舒适度的基础上，光标控制直径的 0.75 倍和传感区域控制直径的 1.8 倍是最佳参数配置。该结果不仅解决了 Midas 触摸和空间精度低的问题，还显著降低了视觉疲劳，从而提高了手眼双通道人机交互系统的易用性和鲁棒性。

{"title":"Research on click enhancement strategy of hand-eye dual-channel human-computer interaction system: Trade-off between sensing area and area cursor","authors":"Ya-Feng Niu, Rui Chen, Yi-Yan Wang, Xue-Ying Yao, Yun Feng","doi":"10.1016/j.aei.2024.102880","DOIUrl":"10.1016/j.aei.2024.102880","url":null,"abstract":"<div><div>This study aims to explore the application of click enhancement strategies in a “Sight Line Localization + Hand Triggered” eye-control human–computer interaction system, proposes a click enhancement strategy to solve two critical problems in eye-control human–computer interaction: Midas touch and low spatial accuracy. By conducting ergonomics experiments, we verify that the proposed click enhancement strategy can effectively improve the operational performance of hand-eye dual-channel HCI systems. The experimental results show that the accuracy of the operation can be significantly improved by using a cursor size equal to the diameter of the interaction control and a sensing area size of 1.8 times the diameter of the control. Based on the comprehensive consideration of operation efficiency and comfort, 0.75 times the control diameter of the cursor and 1.8 times the control diameter of the sensing area are the optimal parameter configurations. The results not only solve the problems of Midas touch and low spatial accuracy but also significantly reduce visual fatigue, thus improving the ease of use and robustness of the hand-eye dual-channel human–computer interaction system.</div></div>","PeriodicalId":50941,"journal":{"name":"Advanced Engineering Informatics","volume":"62 ","pages":"Article 102880"},"PeriodicalIF":8.0,"publicationDate":"2024-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142532085","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"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

Optimized machine learning methods for identifying the stiffness loss of CRTS-II slab track based on vehicle vibration signals 基于车辆振动信号识别 CRTS-II 板式轨道刚度损失的优化机器学习方法

IF 8 1区工程技术 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE

Advanced Engineering Informatics

Pub Date : 2024-10-01 DOI: 10.1016/j.aei.2024.102886

Tao Shi , Ping Lou , T.Y. Yang

Vehicle loads and environmental actions inevitably cause stiffness loss in CRTS-II slab track. Accurately identifying the stiffness loss of the slab track has been a crucial issue to the operation safety of vehicle-CRTS-II slab track coupled system (VSCS). However, existing identification methods for the slab track conditions which often focus on a single damage condition of track service status are inefficient and laborious. This study proposes optimized machine learning (ML) methods for automatically identifying the stiffness loss of the CRTS-II slab track utilizing vehicle vibration signals. The proposed methods achieve the intelligent identification of fastener and interface damage in the slab track, with high identification efficiency and low manpower cost. Four selected ML methods, i.e., support vector machine (SVM), random forest (RF), light gradient boosting machine (LGBM), and artificial neural network (ANN) with optimized hyperparameters are developed to identify the stiffness loss of the slab track. 2200 cases from the dynamic model of VSCS under different conditions of fastener and interface failure are generated to train and test the ML models. The proposed ML methods perform well in the training and testing process, demonstrating that the presented ML methods can accurately identify the stiffness loss of the slab track. Furthermore, the stacking-ensemble learning framework is presented to optimize the performance of the above four ML methods for identifying the stiffness loss of the slab track. The maximum improvement in accuracy for the four selected ML models, utilizing the acceleration of vehicle body and bogie, is 109.09 % and 31.58 %, respectively. The stacking generation has strong anti-noise robustness and generalization ability, proving the excellent reliability and stability of the proposed optimized ML methods. The feature importance of the ML method based on the vehicle acceleration is also analyzed. The proposed efficient and capable optimized ML methods are expected to be widely adopted to intelligently identify the complex service status of track structure utilizing vehicle vibration signals.

车辆荷载和环境作用不可避免地会造成 CRTS-II 板式轨道的刚度损失。准确识别板轨刚度损失一直是车辆-CRTS-II 板轨耦合系统（VSCS）运行安全的关键问题。然而，现有的板式轨道状况识别方法通常只关注轨道服务状态的单一损坏状况，效率低且费力。本研究提出了利用车辆振动信号自动识别 CRTS-II 板式轨道刚度损失的优化机器学习（ML）方法。所提方法实现了板式轨道扣件和接口损伤的智能识别，识别效率高，人力成本低。所选的四种 ML 方法，即支持向量机 (SVM)、随机森林 (RF)、轻梯度提升机 (LGBM) 和优化超参数的人工神经网络 (ANN) 被用来识别板式轨道的刚度损失。在紧固件和界面失效的不同条件下，从 VSCS 动态模型中生成了 2200 个案例，用于训练和测试 ML 模型。所提出的 ML 方法在训练和测试过程中表现良好，表明所提出的 ML 方法可以准确识别板轨的刚度损失。此外，还提出了堆叠-集合学习框架，以优化上述四种识别板轨刚度损失的 ML 方法的性能。利用车体和转向架的加速度，所选的四种 ML 模型的最大精度分别提高了 109.09 % 和 31.58 %。堆叠生成具有很强的抗噪声鲁棒性和泛化能力，证明了所提出的优化 ML 方法具有出色的可靠性和稳定性。此外，还分析了基于车辆加速度的 ML 方法的特征重要性。所提出的高效、实用的优化 ML 方法有望被广泛应用于利用车辆振动信号智能识别轨道结构的复杂服务状态。

{"title":"Optimized machine learning methods for identifying the stiffness loss of CRTS-II slab track based on vehicle vibration signals","authors":"Tao Shi , Ping Lou , T.Y. Yang","doi":"10.1016/j.aei.2024.102886","DOIUrl":"10.1016/j.aei.2024.102886","url":null,"abstract":"<div><div>Vehicle loads and environmental actions inevitably cause stiffness loss in CRTS-II slab track. Accurately identifying the stiffness loss of the slab track has been a crucial issue to the operation safety of vehicle-CRTS-II slab track coupled system (VSCS). However, existing identification methods for the slab track conditions which often focus on a single damage condition of track service status are inefficient and laborious. This study proposes optimized machine learning (ML) methods for automatically identifying the stiffness loss of the CRTS-II slab track utilizing vehicle vibration signals. The proposed methods achieve the intelligent identification of fastener and interface damage in the slab track, with high identification efficiency and low manpower cost. Four selected ML methods, i.e., support vector machine (SVM), random forest (RF), light gradient boosting machine (LGBM), and artificial neural network (ANN) with optimized hyperparameters are developed to identify the stiffness loss of the slab track. 2200 cases from the dynamic model of VSCS under different conditions of fastener and interface failure are generated to train and test the ML models. The proposed ML methods perform well in the training and testing process, demonstrating that the presented ML methods can accurately identify the stiffness loss of the slab track. Furthermore, the stacking-ensemble learning framework is presented to optimize the performance of the above four ML methods for identifying the stiffness loss of the slab track. The maximum improvement in accuracy for the four selected ML models, utilizing the acceleration of vehicle body and bogie, is 109.09 % and 31.58 %, respectively. The stacking generation has strong anti-noise robustness and generalization ability, proving the excellent reliability and stability of the proposed optimized ML methods. The feature importance of the ML method based on the vehicle acceleration is also analyzed. The proposed efficient and capable optimized ML methods are expected to be widely adopted to intelligently identify the complex service status of track structure utilizing vehicle vibration signals.</div></div>","PeriodicalId":50941,"journal":{"name":"Advanced Engineering Informatics","volume":"62 ","pages":"Article 102886"},"PeriodicalIF":8.0,"publicationDate":"2024-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142532087","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"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

Platform service portfolio management (PSPM) of social digitalization platform for cloud-based collaborative product development ecosystem: A structural approach 基于云的协作产品开发生态系统的社会数字化平台的平台服务组合管理（PSPM）：一种结构性方法

IF 8 1区工程技术 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE

Advanced Engineering Informatics

Pub Date : 2024-10-01 DOI: 10.1016/j.aei.2024.102854

Yuguang Bao , Xianyu Zhang , Zhihua Chen , Tongtong Zhou , Xinguo Ming

In the context of digital transformation, social digitalization platform (SDP) emerges and play an important and irreplaceable role in the whole industrial ecosystems. The main value creation purpose of SDP is to develop general technical service portfolios and customized solutions. For SDPs, platform service portfolio management (PSPM) is a really difficult problem. The decision must dynamically respond to the complex external environments of business requirement, technology evolution, and system innovation. The traditional experience-driven decision-making process confuses various factors together, and it is difficult to produce scientific decisions continuously and effectively. To help these platform-kind firms make wise decision, we put forward a novel and practical decision-making methodological framework for PSPM issues. Firstly, a general system architecture model, namely Business-Digitalization-System (BDS) model, is proposed based on system engineering, which helps platform service component identification more refined. Next, the PSPM problem is formalized and modelled as a causal relation graph-based intertwined system analysis procedure. A novel platform service portfolio evaluation method is developed to obtain the priority of the identified service alternatives. The approach can effectively handle the component system importance and its heterogeneous causal interdependencies simultaneously. Meanwhile, the combinatorial manipulation of subjective judgements and objective measure improves the reasonability and efficiency within a multi-stakeholder group. Finally, a case study in the constructive industry is presented and the results of method comparisons show the feasibility and advantages of the methodology.

在数字化转型的背景下，社会数字化平台（SDP）应运而生，并在整个工业生态系统中发挥着不可替代的重要作用。SDP 创造价值的主要目的是开发通用技术服务组合和定制解决方案。对于 SDP 来说，平台服务组合管理（PSPM）确实是一个难题。决策必须动态响应业务需求、技术演进和系统创新等复杂的外部环境。传统的经验驱动型决策过程会将各种因素混淆在一起，难以持续有效地做出科学决策。为了帮助这些平台型企业做出明智的决策，我们针对 PSPM 问题提出了一个新颖实用的决策方法框架。首先，提出了基于系统工程的通用系统架构模型，即业务-数字化-系统（BDS）模型，帮助平台服务组件的识别更加精细化。其次，将 PSPM 问题形式化，并将其建模为基于因果关系图的交织系统分析程序。开发了一种新颖的平台服务组合评估方法，以获得已识别服务备选方案的优先级。该方法可同时有效处理组件系统的重要性及其异构因果相互依存关系。同时，主观判断和客观测量的组合操作提高了多利益相关者群体的合理性和效率。最后，介绍了一个建筑行业的案例研究，方法比较结果表明了该方法的可行性和优势。

{"title":"Platform service portfolio management (PSPM) of social digitalization platform for cloud-based collaborative product development ecosystem: A structural approach","authors":"Yuguang Bao , Xianyu Zhang , Zhihua Chen , Tongtong Zhou , Xinguo Ming","doi":"10.1016/j.aei.2024.102854","DOIUrl":"10.1016/j.aei.2024.102854","url":null,"abstract":"<div><div>In the context of digital transformation, social digitalization platform (SDP) emerges and play an important and irreplaceable role in the whole industrial ecosystems. The main value creation purpose of SDP is to develop general technical service portfolios and customized solutions. For SDPs, platform service portfolio management (PSPM) is a really difficult problem. The decision must dynamically respond to the complex external environments of business requirement, technology evolution, and system innovation. The traditional experience-driven decision-making process confuses various factors together, and it is difficult to produce scientific decisions continuously and effectively. To help these platform-kind firms make wise decision, we put forward a novel and practical decision-making methodological framework for PSPM issues. Firstly, a general system architecture model, namely <em>Business-Digitalization-System</em> (BDS) model, is proposed based on system engineering, which helps platform service component identification more refined. Next, the PSPM problem is formalized and modelled as a causal relation graph-based intertwined system analysis procedure. A novel platform service portfolio evaluation method is developed to obtain the priority of the identified service alternatives. The approach can effectively handle the component system importance and its heterogeneous causal interdependencies simultaneously. Meanwhile, the combinatorial manipulation of subjective judgements and objective measure improves the reasonability and efficiency within a multi-stakeholder group. Finally, a case study in the constructive industry is presented and the results of method comparisons show the feasibility and advantages of the methodology.</div></div>","PeriodicalId":50941,"journal":{"name":"Advanced Engineering Informatics","volume":"62 ","pages":"Article 102854"},"PeriodicalIF":8.0,"publicationDate":"2024-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142532005","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"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 interpretable data-driven approach for process flowsheet convergence troubleshooting 一种可解释的数据驱动方法，用于工艺流程表衔接故障排除

IF 8 1区工程技术 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE

Advanced Engineering Informatics

Pub Date : 2024-10-01 DOI: 10.1016/j.aei.2024.102873

Shifeng Qu, Xinjie Wang, Wenli Du, Feng Qian

Practitioners typically alleviate the convergence problem of process flowsheet models through manual adjustment of the convergence-related flowsheet inputs, which is labor-intensive and relies heavily on expert experience. This paper aims to realize fast troubleshooting for process flowsheets with convergence problems and proposes an interpretable approach for the adjustment of the flowsheets to liberate the manpower for process model maintenance. Specifically, the flowsheet convergence problem is addressed from a data-driven perspective for the first time. The correlation between flowsheet inputs selected according to expert knowledge and convergence status is modeled utilizing the tree-based framework to capture the flowsheet convergence behavior. In addition, a novel interpretable adjustment procedure based on an adaptive minimum mean strategy is constructed to automatically identify strongly convergence-related flowsheet inputs and provide them with quantitative adjustment suggestions. The proposed approach shows effectiveness on non-convergence flowsheets with a success rate of up to 92.5%.

实践者通常通过人工调整与收敛相关的流程表输入来缓解流程表模型的收敛问题，这种方法劳动密集且严重依赖专家经验。本文旨在实现对存在收敛问题的工艺流程表的快速故障诊断，并提出一种可解释的流程表调整方法，以解放工艺模型维护的人力。具体来说，这是首次从数据驱动的角度来解决流程表收敛问题。根据专家知识选择的流程表输入与收敛状态之间的相关性，利用基于树的框架进行建模，以捕捉流程表收敛行为。此外，还构建了基于自适应最小均值策略的新型可解释调整程序，以自动识别与收敛密切相关的流程表输入，并为其提供定量调整建议。所提出的方法在非收敛流程表上显示出了有效性，成功率高达 92.5%。

{"title":"An interpretable data-driven approach for process flowsheet convergence troubleshooting","authors":"Shifeng Qu, Xinjie Wang, Wenli Du, Feng Qian","doi":"10.1016/j.aei.2024.102873","DOIUrl":"10.1016/j.aei.2024.102873","url":null,"abstract":"<div><div>Practitioners typically alleviate the convergence problem of process flowsheet models through manual adjustment of the convergence-related flowsheet inputs, which is labor-intensive and relies heavily on expert experience. This paper aims to realize fast troubleshooting for process flowsheets with convergence problems and proposes an interpretable approach for the adjustment of the flowsheets to liberate the manpower for process model maintenance. Specifically, the flowsheet convergence problem is addressed from a data-driven perspective for the first time. The correlation between flowsheet inputs selected according to expert knowledge and convergence status is modeled utilizing the tree-based framework to capture the flowsheet convergence behavior. In addition, a novel interpretable adjustment procedure based on an adaptive minimum mean strategy is constructed to automatically identify strongly convergence-related flowsheet inputs and provide them with quantitative adjustment suggestions. The proposed approach shows effectiveness on non-convergence flowsheets with a success rate of up to 92.5%.</div></div>","PeriodicalId":50941,"journal":{"name":"Advanced Engineering Informatics","volume":"62 ","pages":"Article 102873"},"PeriodicalIF":8.0,"publicationDate":"2024-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142529671","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"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

Multidomain neural process model based on source attention for industrial robot anomaly detection 基于源关注的多域神经过程模型用于工业机器人异常检测

IF 8 1区工程技术 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE

Advanced Engineering Informatics

Pub Date : 2024-10-01 DOI: 10.1016/j.aei.2024.102910

Bo Yang , Yuhang Huang , Jian Jiao , Wenlong Xu , Lei Liu , Keqiang Xie , Nan Dong

Industrial robots are vital intelligent equipment in modern industries. Periodic maintenance, which is costly and cannot prevent unexpected failures, is necessary to reduce the probability of failure and extend their service life. Therefore, this study pioneers the application of neural processes in industrial robot anomaly detection. On the basis of the attentive neural process framework, a multidomain fusion neural process (MNP) model based on source attention (SA), which introduces a multidomain path that improves the ability of the model to decouple latent distributions of observed data in industrial environments, is proposed. The multidomain path consists of the following parts: First, a time–frequency domain feature extraction module (TFDFEM) is proposed to extract rich time–frequency domain features from raw signals. Second, a dual-purpose SA module is designed to calibrate the temporal and spectral features within the signal, enabling the model to prioritize relevant features. Last, an SA-based multidomain fusion strategy (MDFS) is developed to fuse and complement features from different domains. Numerous experiments based on robots in an automotive welding and bolt fastening lines show that the MNP achieves an average accuracy of 90.8%, outperforming existing models by at least 6.2%. The average F1 is 94.7%, which outperforms existing models by 4.2%. Therefore, our model provides a promising tool for the state-based maintenance of industrial robots. The code for this project is available at https://github.com/hyh7323/Multi-domain-Neural-Process.

工业机器人是现代工业的重要智能设备。定期维护成本高昂，且无法避免意外故障的发生，因此有必要定期维护，以降低故障发生概率，延长使用寿命。因此，本研究开创性地将神经过程应用于工业机器人异常检测。在注意神经过程框架的基础上，提出了基于源注意（SA）的多域融合神经过程（MNP）模型，该模型引入了多域路径，提高了模型对工业环境中观测数据潜在分布的解耦能力。多域路径由以下部分组成：首先，提出了一个时频域特征提取模块（TFDFEM），用于从原始信号中提取丰富的时频域特征。其次，设计了一个两用 SA 模块，用于校准信号中的时域和频域特征，使模型能够优先处理相关特征。最后，开发了一种基于 SA 的多域融合策略 (MDFS)，用于融合和补充来自不同域的特征。基于汽车焊接和螺栓紧固生产线机器人的大量实验表明，MNP 的平均准确率达到 90.8%，比现有模型至少高出 6.2%。平均 F1 为 94.7%，比现有模型高出 4.2%。因此，我们的模型为基于状态的工业机器人维护提供了一个前景广阔的工具。该项目的代码可在 https://github.com/hyh7323/Multi-domain-Neural-Process 上获取。

{"title":"Multidomain neural process model based on source attention for industrial robot anomaly detection","authors":"Bo Yang , Yuhang Huang , Jian Jiao , Wenlong Xu , Lei Liu , Keqiang Xie , Nan Dong","doi":"10.1016/j.aei.2024.102910","DOIUrl":"10.1016/j.aei.2024.102910","url":null,"abstract":"<div><div>Industrial robots are vital intelligent equipment in modern industries. Periodic maintenance, which is costly and cannot prevent unexpected failures, is necessary to reduce the probability of failure and extend their service life. Therefore, this study pioneers the application of neural processes in industrial robot anomaly detection. On the basis of the attentive neural process framework, a multidomain fusion neural process (MNP) model based on source attention (SA), which introduces a multidomain path that improves the ability of the model to decouple latent distributions of observed data in industrial environments, is proposed. The multidomain path consists of the following parts: First, a time–frequency domain feature extraction module (TFDFEM) is proposed to extract rich time–frequency domain features from raw signals. Second, a dual-purpose SA module is designed to calibrate the temporal and spectral features within the signal, enabling the model to prioritize relevant features. Last, an SA-based multidomain fusion strategy (MDFS) is developed to fuse and complement features from different domains. Numerous experiments based on robots in an automotive welding and bolt fastening lines show that the MNP achieves an average accuracy of 90.8%, outperforming existing models by at least 6.2%. The average F1 is 94.7%, which outperforms existing models by 4.2%. Therefore, our model provides a promising tool for the state-based maintenance of industrial robots. The code for this project is available at <span><span>https://github.com/hyh7323/Multi-domain-Neural-Process</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":50941,"journal":{"name":"Advanced Engineering Informatics","volume":"62 ","pages":"Article 102910"},"PeriodicalIF":8.0,"publicationDate":"2024-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142532247","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"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