Pub Date : 2025-04-27DOI: 10.1016/j.aei.2025.103385
Jie Wang , Yaqian You , Zhijie Zhou , Peng Zhang
In engineering scenarios, the performance of industrial systems varies continuously, making it necessary to develop a prediction model to track system performance. Recently, a modeling approach known as the concise belief rule base (CBRB) has provided an effective reference for performance prediction. However, CBRB ignores the decay phenomenon of information credibility during the prediction process, leading to suboptimal output accuracy. To address this limitation, a novel performance prediction model based on the concise belief rule base with credibility decay (CBRB-CD) is put forward. The proposed model incorporates a decay factor to reflect the property that the credibility of belief rules decays over time. Meanwhile, the decay factor is aggregated into the fusion process of belief rules, from which the prediction results are generated. Furthermore, a stability analysis of the prediction model is carried out by introducing external perturbations to validate the prediction results. The analysis results quantitatively reveal the changing patterns of prediction results under perturbed environments. Finally, real-world experiments on aerospace relays demonstrate the feasibility of the proposed model.
{"title":"Concise belief rule base with credibility decay for system performance prediction","authors":"Jie Wang , Yaqian You , Zhijie Zhou , Peng Zhang","doi":"10.1016/j.aei.2025.103385","DOIUrl":"10.1016/j.aei.2025.103385","url":null,"abstract":"<div><div>In engineering scenarios, the performance of industrial systems varies continuously, making it necessary to develop a prediction model to track system performance. Recently, a modeling approach known as the concise belief rule base (CBRB) has provided an effective reference for performance prediction. However, CBRB ignores the decay phenomenon of information credibility during the prediction process, leading to suboptimal output accuracy. To address this limitation, a novel performance prediction model based on the concise belief rule base with credibility decay (CBRB-CD) is put forward. The proposed model incorporates a decay factor to reflect the property that the credibility of belief rules decays over time. Meanwhile, the decay factor is aggregated into the fusion process of belief rules, from which the prediction results are generated. Furthermore, a stability analysis of the prediction model is carried out by introducing external perturbations to validate the prediction results. The analysis results quantitatively reveal the changing patterns of prediction results under perturbed environments. Finally, real-world experiments on aerospace relays demonstrate the feasibility of the proposed model.</div></div>","PeriodicalId":50941,"journal":{"name":"Advanced Engineering Informatics","volume":"65 ","pages":"Article 103385"},"PeriodicalIF":8.0,"publicationDate":"2025-04-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143877322","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}
Pub Date : 2025-04-27DOI: 10.1016/j.aei.2025.103366
Arjun Rachana Harish , Zhaolin Yuan , Ming Li , Hongxia Yang , George Q. Huang
The collaborative garment designing lifecycle involves stages such as designing, styling, and patterning. Some of these stages can be partially or fully automated using industrial large models (LMs), such as generative and large language models. The key to quick and cost-effective order fulfillment is the orchestration of group interactions, or a group chat, between the stakeholders and LMs in garment design. However, certain unaddressed aspects, such as knowledge retention, generalization, and complexity of group interaction, are critical to realizing group chat for garment design. This study proposes a framework called ChatFashion for group chat in garment design. Transformer, a core construct of the proposed framework, orchestrates interaction among stakeholders and industrial LMs. It undergoes an evolution with the intelligence it picks up from its interaction with diverse stakeholders and industrial LMs, allowing it to act as a one-stop solution for multidisciplinary design needs. This study contributes to theory in the following aspects. First, it proposes transformers to eliminate concerns about knowledge retention by industrial LMs. Second, while other studies focus on the benefits of industrial LMs to simplify individual stages in garment design, this study introduces the design and demonstration of a ChatFashion framework for collaborative garment designing using industrial LMs. Finally, this study advances the literature on prompt engineering of industrial LMs by utilizing collaborative learning (or models learning from each other) to capture and orchestrate the group chat among stakeholders, signifying its practicality and value for research in garment design.
{"title":"Collaborative garment design through group chatting with generative industrial large models","authors":"Arjun Rachana Harish , Zhaolin Yuan , Ming Li , Hongxia Yang , George Q. Huang","doi":"10.1016/j.aei.2025.103366","DOIUrl":"10.1016/j.aei.2025.103366","url":null,"abstract":"<div><div>The collaborative garment designing lifecycle involves stages such as designing, styling, and patterning. Some of these stages can be partially or fully automated using industrial large models (LMs), such as generative and large language models. The key to quick and cost-effective order fulfillment is the orchestration of group interactions, or a group chat, between the stakeholders and LMs in garment design. However, certain unaddressed aspects, such as knowledge retention, generalization, and complexity of group interaction, are critical to realizing group chat for garment design. This study proposes a framework called ChatFashion for group chat in garment design. Transformer, a core construct of the proposed framework, orchestrates interaction among stakeholders and industrial LMs. It undergoes an evolution with the intelligence it picks up from its interaction with diverse stakeholders and industrial LMs, allowing it to act as a one-stop solution for multidisciplinary design needs. This study contributes to theory in the following aspects. First, it proposes transformers to eliminate concerns about knowledge retention by industrial LMs. Second, while other studies focus on the benefits of industrial LMs to simplify individual stages in garment design, this study introduces the design and demonstration of a ChatFashion framework for collaborative garment designing using industrial LMs. Finally, this study advances the literature on prompt engineering of industrial LMs by utilizing collaborative learning (or models learning from each other) to capture and orchestrate the group chat among stakeholders, signifying its practicality and value for research in garment design.</div></div>","PeriodicalId":50941,"journal":{"name":"Advanced Engineering Informatics","volume":"65 ","pages":"Article 103366"},"PeriodicalIF":8.0,"publicationDate":"2025-04-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143877324","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}
The utility of Unmanned Aerial Vehicles (UAVs) for routine pavement distresses inspection has been increasingly recognized due to their efficiency, flexibility, safety, and low-cost automation. However, UAV-acquired high-altitude images present unique challenges for deep learning-based semantic segmentation models, such as minute crack details, blurred boundaries, and high levels of environmental noise. We propose GLoU-MiT, a lightweight segmentation model designed to address the difficulties of UAV-based pavement crack segmentation. Our model integrates a U-shaped Mix Transformer architecture for efficient hierarchical feature extraction, a Global-Local Mamba-Guided Skip Connection for improved feature alignment and computational efficiency, and a Boundary / Semantic Deep Supervision Refinement Module to enhance segmentation precision in complex scenarios. Extensive experiments on UAV-Crack500, CrackSC and Crack500 datasets demonstrate that GLoU-MiT effectively improves segmentation accuracy, particularly in low-contrast and complex background environments, making it a robust solution for UAV-based pavement crack inspection tasks. Furthermore, inference speed and energy consumption evaluations conducted on the Jetson Orin Nano (8 GB) show that our model achieves an excellent balance between accuracy, energy efficiency, and speed. The code will be released at: https://github.com/SHAN-JH/GLoU-MiT.
{"title":"GLoU-MiT: Lightweight Global-Local Mamba-Guided U-mix transformer for UAV-based pavement crack segmentation","authors":"Jinhuan Shan , Yue Huang , Wei Jiang , Dongdong Yuan , Feiyang Guo","doi":"10.1016/j.aei.2025.103384","DOIUrl":"10.1016/j.aei.2025.103384","url":null,"abstract":"<div><div>The utility of Unmanned Aerial Vehicles (UAVs) for routine pavement distresses inspection has been increasingly recognized due to their efficiency, flexibility, safety, and low-cost automation. However, UAV-acquired high-altitude images present unique challenges for deep learning-based semantic segmentation models, such as minute crack details, blurred boundaries, and high levels of environmental noise. We propose GLoU-MiT, a lightweight segmentation model designed to address the difficulties of UAV-based pavement crack segmentation. Our model integrates a U-shaped Mix Transformer architecture for efficient hierarchical feature extraction, a Global-Local Mamba-Guided Skip Connection for improved feature alignment and computational efficiency, and a Boundary / Semantic Deep Supervision Refinement Module to enhance segmentation precision in complex scenarios. Extensive experiments on UAV-Crack500, CrackSC and Crack500 datasets demonstrate that GLoU-MiT effectively improves segmentation accuracy, particularly in low-contrast and complex background environments, making it a robust solution for UAV-based pavement crack inspection tasks. Furthermore, inference speed and energy consumption evaluations conducted on the Jetson Orin Nano (8 GB) show that our model achieves an excellent balance between accuracy, energy efficiency, and speed. The code will be released at: <span><span>https://github.com/SHAN-JH/GLoU-MiT</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":50941,"journal":{"name":"Advanced Engineering Informatics","volume":"65 ","pages":"Article 103384"},"PeriodicalIF":8.0,"publicationDate":"2025-04-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143877323","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}
Thermal scene reconstruction holds great potential for various applications, such as building energy analysis and non-destructive infrastructure testing. However, existing methods rely on dense scene measurements and use RGB images for 3D reconstruction, incorporating thermal data only through a post-hoc projection. Due to the lower resolution of thermal cameras and the challenges of RGB/Thermal camera calibration, this post-hoc projection often results in spatial discrepancies between temperatures projected onto the 3D model and real temperatures at the surface. We propose ThermoNeRF, a novel multimodal Neural Radiance Fields (NerF) that renders new RGB and thermal views of a scene with joint optimization of the geometry and thermal information while preventing cross-modal interference. To compensate for the lack of texture in thermal images, ThermoNeRF leverages paired RGB and thermal images to learn scene geometry while maintaining separate networks for reconstructing RGB color and temperature values, ensuring accurate and modality-specific representations. We also introduce ThermoScenes, a dataset of paired RGB+thermal images comprising 8 scenes of building facades and 8 scenes of everyday objects enabling evaluation in diverse scenarios. On ThermoScenes, ThermoNeRF achieves an average mean absolute error of 1.13 °C for buildings and 0.41 °C for other scenes when predicting temperatures of previously unobserved views. This improves accuracy by over 50% compared to concatenated RGB+thermal input in standard NeRF. While ThermoNeRF performs well on aligned RGB-thermal images, future work could address misaligned or unpaired data for better generalization. Code and dataset are available online.
{"title":"ThermoNeRF: A multimodal Neural Radiance Field for joint RGB-thermal novel view synthesis of building facades","authors":"Mariam Hassan , Florent Forest , Olga Fink , Malcolm Mielle","doi":"10.1016/j.aei.2025.103345","DOIUrl":"10.1016/j.aei.2025.103345","url":null,"abstract":"<div><div>Thermal scene reconstruction holds great potential for various applications, such as building energy analysis and non-destructive infrastructure testing. However, existing methods rely on dense scene measurements and use RGB images for 3D reconstruction, incorporating thermal data only through a post-hoc projection. Due to the lower resolution of thermal cameras and the challenges of RGB/Thermal camera calibration, this post-hoc projection often results in spatial discrepancies between temperatures projected onto the 3D model and real temperatures at the surface. We propose ThermoNeRF, a novel multimodal Neural Radiance Fields (NerF) that renders new RGB and thermal views of a scene with joint optimization of the geometry and thermal information while preventing cross-modal interference. To compensate for the lack of texture in thermal images, ThermoNeRF leverages paired RGB and thermal images to learn scene geometry while maintaining separate networks for reconstructing RGB color and temperature values, ensuring accurate and modality-specific representations. We also introduce ThermoScenes, a dataset of paired RGB+thermal images comprising 8 scenes of building facades and 8 scenes of everyday objects enabling evaluation in diverse scenarios. On ThermoScenes, ThermoNeRF achieves an average mean absolute error of 1.13 °C for buildings and 0.41 °C for other scenes when predicting temperatures of previously unobserved views. This improves accuracy by over 50% compared to concatenated RGB+thermal input in standard NeRF. While ThermoNeRF performs well on aligned RGB-thermal images, future work could address misaligned or unpaired data for better generalization. <span><span>Code</span><svg><path></path></svg></span> and <span><span>dataset</span><svg><path></path></svg></span> are available online.</div></div>","PeriodicalId":50941,"journal":{"name":"Advanced Engineering Informatics","volume":"65 ","pages":"Article 103345"},"PeriodicalIF":8.0,"publicationDate":"2025-04-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143874436","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}
Pub Date : 2025-04-26DOI: 10.1016/j.aei.2025.103393
Behnam M. Tehrani, Aladdin Alwisy
Robotics hold immense potential to transform the industrialized construction (IC) sector by enhancing productivity, drawing on its similarities with the manufacturing industry. However, the unique nature of construction projects and the demand for customized building designs present substantial technological hurdles for incorporating robotics into IC activities. The time and cost burdens of the project- and task-driven reprogramming of robotic systems can outweigh the perceived benefits of robotics-based automation. Those burdens are further compounded by the notable lack of research addressing the technical collaboration issues between virtual design and construction software, specifically Building Information Modeling (BIM) tools, and commercially available robotic software suites. This research aims to promote the adoption of construction robotics by streamlining the design-to-manufacturing processes of framing tasks in IC. The proposed framework is integrated into a software platform that establishes a collaborative environment for BIM and assembly-based robotics in panel framing tasks of IC, referred to as BIM-to-Bot (B2B). This platform uses architectural BIM models from Revit and automatically converts building design information into shop drawings and robotics-based manufacturing procedures for a multi-robot system designed specifically for framing tasks in IC. The proposed framework was evaluated through a case study revealing significant efficiency improvements. The drafting process improved by 93.33%, and the robotic programming process improved by 88.66% when using the proposed method compared to manual drafting and programming, resulting in an 89.34% time savings in BIM-to-Robot processes. By enabling the integration of robotics into panel framing tasks, this research bridges the crucial gap between building designs and robotic assembly. The enhanced efficiency associated with the proposed streamlined design-to-manufacturing processes is expected to pave the way for the broader adoption of robotics in the construction industry, heralding a new era of IC.
{"title":"Streamlining design-to-manufacturing for assembly-based robotics in wood panel framing tasks of industrialized construction: Introducing a BIM-to-BoT (B2B) framework","authors":"Behnam M. Tehrani, Aladdin Alwisy","doi":"10.1016/j.aei.2025.103393","DOIUrl":"10.1016/j.aei.2025.103393","url":null,"abstract":"<div><div>Robotics hold immense potential to transform the industrialized construction (IC) sector by enhancing productivity, drawing on its similarities with the manufacturing industry. However, the unique nature of construction projects and the demand for customized building designs present substantial technological hurdles for incorporating robotics into IC activities. The time and cost burdens of the project- and task-driven reprogramming of robotic systems can outweigh the perceived benefits of robotics-based automation. Those burdens are further compounded by the notable lack of research addressing the technical collaboration issues between virtual design and construction software, specifically Building Information Modeling (BIM) tools, and commercially available robotic software suites. This research aims to promote the adoption of construction robotics by streamlining the design-to-manufacturing processes of framing tasks in IC. The proposed framework is integrated into a software platform that establishes a collaborative environment for BIM and assembly-based robotics in panel framing tasks of IC, referred to as BIM-to-Bot (B2B). This platform uses architectural BIM models from Revit and automatically converts building design information into shop drawings and robotics-based manufacturing procedures for a multi-robot system designed specifically for framing tasks in IC. The proposed framework was evaluated through a case study revealing significant efficiency improvements. The drafting process improved by 93.33%, and the robotic programming process improved by 88.66% when using the proposed method compared to manual drafting and programming, resulting in an 89.34% time savings in BIM-to-Robot processes. By enabling the integration of robotics into panel framing tasks, this research bridges the crucial gap between building designs and robotic assembly. The enhanced efficiency associated with the proposed streamlined design-to-manufacturing processes is expected to pave the way for the broader adoption of robotics in the construction industry, heralding a new era of IC.</div></div>","PeriodicalId":50941,"journal":{"name":"Advanced Engineering Informatics","volume":"65 ","pages":"Article 103393"},"PeriodicalIF":8.0,"publicationDate":"2025-04-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143874877","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}
Pub Date : 2025-04-26DOI: 10.1016/j.aei.2025.103351
Zhe Huang , Yongcai Wang , Deying Li , Yunjun Han , Lei Wang
Collaborative perception can significantly enhance perception performance through information sharing among multiple smart vehicles and roadside perception systems. Existing methods typically require sharing and aggregating all information from other collaborators, which can significantly reduce collaborative performance in detecting distant or occluded objects due to the large amount of redundant information in the final perception. To address this issue, we propose a novel collaborative framework, named the Semantic Aware Heterogeneous Network (SAHNet), which extracts, shares and fuses perceptually crucial and useful information among heterogeneous collaborators to improve the performance of 3D object detection. Specifically, we first design a Foreground and Boundary Feature Selection (FBFS) to enhance meaningful feature extraction. Then a Heterogeneous Feature Transfer module (HFF) is then proposed to account for collaborators’ heterogeneity to better transfer perception-critical features. Finally, we introduce a Semantic Feature Fusion module (SFF) that effectively aggregates features using semantic information. The proposed framework has been extensively compared and evaluated on two simulation datasets and one real-world dataset. The experimental results demonstrate that SAHNet consistently outperforms existing methods in collaborative object detection, demonstrating strong robustness even under conditions with localization noise and time delays. Additionally, we have provided a comprehensive ablation study to illustrate the effectiveness of each module within our framework.
{"title":"Collaborative 3D object detection by smart vehicles considering semantic information and agent heterogeneity","authors":"Zhe Huang , Yongcai Wang , Deying Li , Yunjun Han , Lei Wang","doi":"10.1016/j.aei.2025.103351","DOIUrl":"10.1016/j.aei.2025.103351","url":null,"abstract":"<div><div>Collaborative perception can significantly enhance perception performance through information sharing among multiple smart vehicles and roadside perception systems. Existing methods typically require sharing and aggregating all information from other collaborators, which can significantly reduce collaborative performance in detecting distant or occluded objects due to the large amount of redundant information in the final perception. To address this issue, we propose a novel collaborative framework, named the Semantic Aware Heterogeneous Network (SAHNet), which extracts, shares and fuses perceptually crucial and useful information among heterogeneous collaborators to improve the performance of 3D object detection. Specifically, we first design a Foreground and Boundary Feature Selection (FBFS) to enhance meaningful feature extraction. Then a Heterogeneous Feature Transfer module (HFF) is then proposed to account for collaborators’ heterogeneity to better transfer perception-critical features. Finally, we introduce a Semantic Feature Fusion module (SFF) that effectively aggregates features using semantic information. The proposed framework has been extensively compared and evaluated on two simulation datasets and one real-world dataset. The experimental results demonstrate that SAHNet consistently outperforms existing methods in collaborative object detection, demonstrating strong robustness even under conditions with localization noise and time delays. Additionally, we have provided a comprehensive ablation study to illustrate the effectiveness of each module within our framework.</div></div>","PeriodicalId":50941,"journal":{"name":"Advanced Engineering Informatics","volume":"65 ","pages":""},"PeriodicalIF":8.0,"publicationDate":"2025-04-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143873669","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}
Pub Date : 2025-04-26DOI: 10.1016/j.aei.2025.103382
Boyu Wang , Fangzhou Lin , Mingkai Li , Zhenyu Liang , Zhengyi Chen , Mingzhu Wang , Jack C.P. Cheng
Mechanical, electrical, and plumbing (MEP) systems are critical for delivering essential services and ensuring comfortable environments. To improve the management efficiency of these complex systems, digital twins (DTs) that reflect the as-is conditions of facilities are increasingly being adopted. To generate DT models, laser scanners are widely used to capture as-built environments in the form of high-resolution images and dense 3D measurements. However, existing scan-to-BIM methods primarily produce basic geometric models, lacking detailed descriptive attributes of the components. To address this limitation, this paper proposes an informative DT model generation method for MEP systems based on fine-grained object recognition and object-aware scan-vs-BIM. The proposed method adopts a few-shot learning strategy to detect target objects in complex 3D environments and identify their family types based on vision foundation models. Following this, the association between as-designed components and as-built installations is formulated as a bipartite graph matching problem, which is solved using the Hungarian algorithm. This enables the automated updating of as-designed models into as-built DT models. Notably, the proposed association method is robust and applicable to components with significant installation deviations, a common challenge in MEP systems. The feasibility of the proposed approach was validated through experiments conducted on two construction sites in Hong Kong. Results demonstrated that the proposed approach significantly enhanced the accuracy of the scan-vs-BIM of MEP systems, thereby enabling informative DT model generation.
{"title":"Informative As-Built Modeling as a Foundation for Digital Twins Based on Fine-Grained Object Recognition and Object-Aware Scan-vs-BIM for MEP Scenes","authors":"Boyu Wang , Fangzhou Lin , Mingkai Li , Zhenyu Liang , Zhengyi Chen , Mingzhu Wang , Jack C.P. Cheng","doi":"10.1016/j.aei.2025.103382","DOIUrl":"10.1016/j.aei.2025.103382","url":null,"abstract":"<div><div>Mechanical, electrical, and plumbing (MEP) systems are critical for delivering essential services and ensuring comfortable environments. To improve the management efficiency of these complex systems, digital twins (DTs) that reflect the as-is conditions of facilities are increasingly being adopted. To generate DT models, laser scanners are widely used to capture as-built environments in the form of high-resolution images and dense 3D measurements. However, existing scan-to-BIM methods primarily produce basic geometric models, lacking detailed descriptive attributes of the components. To address this limitation, this paper proposes an informative DT model generation method for MEP systems based on fine-grained object recognition and object-aware scan-vs-BIM. The proposed method adopts a few-shot learning strategy to detect target objects in complex 3D environments and identify their family types based on vision foundation models. Following this, the association between as-designed components and as-built installations is formulated as a bipartite graph matching problem, which is solved using the Hungarian algorithm. This enables the automated updating of as-designed models into as-built DT models. Notably, the proposed association method is robust and applicable to components with significant installation deviations, a common challenge in MEP systems. The feasibility of the proposed approach was validated through experiments conducted on two construction sites in Hong Kong. Results demonstrated that the proposed approach significantly enhanced the accuracy of the scan-vs-BIM of MEP systems, thereby enabling informative DT model generation.</div></div>","PeriodicalId":50941,"journal":{"name":"Advanced Engineering Informatics","volume":"65 ","pages":"Article 103382"},"PeriodicalIF":8.0,"publicationDate":"2025-04-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143877359","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}
Pub Date : 2025-04-26DOI: 10.1016/j.aei.2025.103392
Zhixin Liu , Shanhe Lou , Yixiong Feng , Wenhui Huang , Bingtao Hu , Chengyu Lu , Jianrong Tan
Computer-aided conceptual design (CACD) is a core means for the development of new products, as it can materialize designers’ inherent thinking. However, when designers encounter stagnation during CACD, they need to consult third-party design knowledge to seek inspiration, which frequently disrupts their design thinking process. Deep learning-empowered design methods and design knowledge management can be a potential solution to address these issues. This study proposes a multimodal design data-driven interactive design method. Multimodal data are utilized to identify the designer’s implicit intentions while design attention is abstracted to match relevant knowledge as computer feedback. It achieves the “designer-computer-designer” closed-loop interactive design through the mediation of design attention. The multimodal design data (design images and design descriptions) is obtained through sketch modeling and verbal protocol analysis experiments. A multimodal Transformer based on T2T-ViT and Bert (TB-Multiformer) is constructed to capture multimodal features to identify conceptual design intentions by utilizing cross-modal design attention modules and self-design attention modules. Since the identified attention can be used to match the knowledge that designers are more concerned about, an attention-based design knowledge recommendation method (AbDKR) is proposed to provide proactive knowledge feedback. It can prevent designers from spending time searching for design knowledge and helps them maintain sufficient inspiration. A case study on the conceptual design of two types of mechanical structure is conducted to illustrate the feasibility and practicability of the proposed approach.
{"title":"More attention for computer-aided conceptual design: A multimodal data-driven interactive design method","authors":"Zhixin Liu , Shanhe Lou , Yixiong Feng , Wenhui Huang , Bingtao Hu , Chengyu Lu , Jianrong Tan","doi":"10.1016/j.aei.2025.103392","DOIUrl":"10.1016/j.aei.2025.103392","url":null,"abstract":"<div><div>Computer-aided conceptual design (CACD) is a core means for the development of new products, as it can materialize designers’ inherent thinking. However, when designers encounter stagnation during CACD, they need to consult third-party design knowledge to seek inspiration, which frequently disrupts their design thinking process. Deep learning-empowered design methods and design knowledge management can be a potential solution to address these issues. This study proposes a multimodal design data-driven interactive design method. Multimodal data are utilized to identify the designer’s implicit intentions while design attention is abstracted to match relevant knowledge as computer feedback. It achieves the “designer-computer-designer” closed-loop interactive design through the mediation of design attention. The multimodal design data (design images and design descriptions) is obtained through sketch modeling and verbal protocol analysis experiments. A multimodal Transformer based on T2T-ViT and Bert (TB-Multiformer) is constructed to capture multimodal features to identify conceptual design intentions by utilizing cross-modal design attention modules and self-design attention modules. Since the identified attention can be used to match the knowledge that designers are more concerned about, an attention-based design knowledge recommendation method (AbDKR) is proposed to provide proactive knowledge feedback. It can prevent designers from spending time searching for design knowledge and helps them maintain sufficient inspiration. A case study on the conceptual design of two types of mechanical structure is conducted to illustrate the feasibility and practicability of the proposed approach.</div></div>","PeriodicalId":50941,"journal":{"name":"Advanced Engineering Informatics","volume":"65 ","pages":"Article 103392"},"PeriodicalIF":8.0,"publicationDate":"2025-04-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143874878","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}
Pub Date : 2025-04-24DOI: 10.1016/j.aei.2025.103357
Diwang Ruan , Yan Wang , Yiliang Qian , Jianping Yan , Zhaorong Li
System dynamics modeling holds significant importance in engineering, especially for high-dimensional, non-linear, and time-varying systems. Traditional methods often encounter challenges such as poor interpretability, low computational efficiency, and limited generalization capabilities. To address these issues, this paper proposes a novel framework for dynamics modeling, Deep Stacked State-observer based Neural Network (DSSO-NN). This framework integrates the Extended State Observer (ESO) with state–space equations, leveraging the efficient state estimation of ESO and the powerful fitting capabilities of neural networks. Firstly, based on the system’s state–space equations, an ESO is constructed and then discretized to obtain neurons tailored for system modeling. Subsequently, serial and parallel structures are explored and compared to determine the optimal structure for validation, culminating in the construction of the DSSO network. Furthermore, critical factors influencing DSSO-NN performance, including ESO hyperparameter (), system order, and the number of layers, are optimized. Experimental results on Case Western Reserve University and FEMTO datasets demonstrate that DSSO-NN effectively captures system dynamics and achieves superior performance. This study showcases the robust performance and broad application potential of DSSO-NN in bearing dynamics modeling, providing a novel approach for complex dynamics modeling.
{"title":"Deep stacked state-observer based neural network (DSSO-NN): A new network for system dynamics modeling and application in bearing","authors":"Diwang Ruan , Yan Wang , Yiliang Qian , Jianping Yan , Zhaorong Li","doi":"10.1016/j.aei.2025.103357","DOIUrl":"10.1016/j.aei.2025.103357","url":null,"abstract":"<div><div>System dynamics modeling holds significant importance in engineering, especially for high-dimensional, non-linear, and time-varying systems. Traditional methods often encounter challenges such as poor interpretability, low computational efficiency, and limited generalization capabilities. To address these issues, this paper proposes a novel framework for dynamics modeling, Deep Stacked State-observer based Neural Network (DSSO-NN). This framework integrates the Extended State Observer (ESO) with state–space equations, leveraging the efficient state estimation of ESO and the powerful fitting capabilities of neural networks. Firstly, based on the system’s state–space equations, an ESO is constructed and then discretized to obtain neurons tailored for system modeling. Subsequently, serial and parallel structures are explored and compared to determine the optimal structure for validation, culminating in the construction of the DSSO network. Furthermore, critical factors influencing DSSO-NN performance, including ESO hyperparameter (<span><math><mi>δ</mi></math></span>), system order, and the number of layers, are optimized. Experimental results on Case Western Reserve University and FEMTO datasets demonstrate that DSSO-NN effectively captures system dynamics and achieves superior performance. This study showcases the robust performance and broad application potential of DSSO-NN in bearing dynamics modeling, providing a novel approach for complex dynamics modeling.</div></div>","PeriodicalId":50941,"journal":{"name":"Advanced Engineering Informatics","volume":"65 ","pages":"Article 103357"},"PeriodicalIF":8.0,"publicationDate":"2025-04-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143863824","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}
Pub Date : 2025-04-24DOI: 10.1016/j.aei.2025.103387
Lin Lin, Lingyu Yue, Dan Liu, Jinlei Wu, Sihao Zhang, Yikun Liu, Shiwei Suo
Due to the aircraft wing’s topological structure and lightweight design requirements, strain sensors installed on the wing are very limited. Traditional methods, relying on limited sensor data as a single information source, are insufficient for full-stress field monitoring, leading to a high prediction error. To address this issue, a novel wing stress field reconstruction method with limited measurement points is developed via multi-source heterogeneous information fusion. To be specific, two information fusion modules are designed to jointly overcome the challenges of limited measurement data and high non-linearity during full-stress field reconstruction. On one hand, the finite element mechanism-based information fusion module (FEMIFM) is proposed to derive and establish a mechanical model that relates the wing stress to positional parameter, in order to introduce physical information and reduce the non-linearity of the reconstruction mapping. On the other hand, the simulation stress expectation-based information fusion module (SSEIFM) leverages stress expectations derived from simulated stress fields under various operating conditions to incorporate statistical information, thereby enhancing the robustness and reasonableness of reconstruction results. Moreover, a soft-threshold loss function is proposed, which ignores zero-drift errors of strain sensors, improving the reconstruction accuracy of critical stress points. Finally, the developed method can be seamlessly integrated with popular neural networks (i.e., Transformer, convolutional neural networks, multilayer perceptron, etc.). Extensive experiments are conducted to validate the effectiveness of the developed method on an actual aircraft wing stress dataset.
{"title":"Reconstruction method of aircraft wing stress field under limited measurement points via multi-source heterogeneous information fusion","authors":"Lin Lin, Lingyu Yue, Dan Liu, Jinlei Wu, Sihao Zhang, Yikun Liu, Shiwei Suo","doi":"10.1016/j.aei.2025.103387","DOIUrl":"10.1016/j.aei.2025.103387","url":null,"abstract":"<div><div>Due to the aircraft wing’s topological structure and lightweight design requirements, strain sensors installed on the wing are very limited. Traditional methods, relying on limited sensor data as a single information source, are insufficient for full-stress field monitoring, leading to a high prediction error. To address this issue, a novel wing stress field reconstruction method with limited measurement points is developed via multi-source heterogeneous information fusion. To be specific, two information fusion modules are designed to jointly overcome the challenges of limited measurement data and high non-linearity during full-stress field reconstruction. On one hand, the finite element mechanism-based information fusion module (FEMIFM) is proposed to derive and establish a mechanical model that relates the wing stress to positional parameter, in order to introduce physical information and reduce the non-linearity of the reconstruction mapping. On the other hand, the simulation stress expectation-based information fusion module (SSEIFM) leverages stress expectations derived from simulated stress fields under various operating conditions to incorporate statistical information, thereby enhancing the robustness and reasonableness of reconstruction results. Moreover, a soft-threshold loss function is proposed, which ignores zero-drift errors of strain sensors, improving the reconstruction accuracy of critical stress points. Finally, the developed method can be seamlessly integrated with popular neural networks (i.e., Transformer, convolutional neural networks, multilayer perceptron, etc.). Extensive experiments are conducted to validate the effectiveness of the developed method on an actual aircraft wing stress dataset.</div></div>","PeriodicalId":50941,"journal":{"name":"Advanced Engineering Informatics","volume":"65 ","pages":"Article 103387"},"PeriodicalIF":8.0,"publicationDate":"2025-04-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143867887","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}
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