首页 > 最新文献

Advanced Engineering Informatics最新文献

英文 中文
PA-WSDIS: A prior-aware weakly supervised defect instance segmentation model for car body surface
IF 8 1区 工程技术 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE Pub Date : 2025-03-23 DOI: 10.1016/j.aei.2025.103254
Yike He , Yueming Wang , Weiwei Jiang , Songyu Hu , Jianzhong Fu
Car body surface defect instance segmentation is essential for ensuring product quality and setting precise size thresholds for defects during product inspection process. However, few defect instance segmentation applications has been found in industrial scenarios until now. This is due in large part to the fact that the pixel-level annotation of defects is cumbersome and labor-intensive. Although various weakly supervised methods have shown promising results, they usually lack the ability to fully explore prior information and the awareness of hierarchical semantic correlations, thereby limiting the defect instance segmentation performance. To address this issue, we propose a novel prior-aware weakly supervised defect instance segmentation (PA-WSDIS) model for car body surface, removing the need for pixel-level labeling. First, we design a box-driven coarse mask generator to obtain coarse masks, which serve as potential proposals for the subsequent refinement process. Then, we propose a boundary guided prior constraint loss, consisting of boundary alignment and pixel-pair similarity mining losses, to fully leverage prior information to enhance the discriminative ability and provide reliable refinement guidance for the model. Finally, we propose a correlative semantic calibration loss, which comprehensively perceives the rich semantic features of different dimensions from both local and global perspectives. With the collaborative constraints of these meticulously designed loss functions, precise instance segmentation results are achieved. Experimental results showcase the outstanding performance of the PA-WSDIS model with an impressive 87.4% mAP50mask, which is considerably superior to state-of-the-art methods. As far as we know, our proposed method is the first weakly supervised instance segmentation model based on bounding box labels for industrial defect detection tasks.
{"title":"PA-WSDIS: A prior-aware weakly supervised defect instance segmentation model for car body surface","authors":"Yike He ,&nbsp;Yueming Wang ,&nbsp;Weiwei Jiang ,&nbsp;Songyu Hu ,&nbsp;Jianzhong Fu","doi":"10.1016/j.aei.2025.103254","DOIUrl":"10.1016/j.aei.2025.103254","url":null,"abstract":"<div><div>Car body surface defect instance segmentation is essential for ensuring product quality and setting precise size thresholds for defects during product inspection process. However, few defect instance segmentation applications has been found in industrial scenarios until now. This is due in large part to the fact that the pixel-level annotation of defects is cumbersome and labor-intensive. Although various weakly supervised methods have shown promising results, they usually lack the ability to fully explore prior information and the awareness of hierarchical semantic correlations, thereby limiting the defect instance segmentation performance. To address this issue, we propose a novel prior-aware weakly supervised defect instance segmentation (PA-WSDIS) model for car body surface, removing the need for pixel-level labeling. First, we design a box-driven coarse mask generator to obtain coarse masks, which serve as potential proposals for the subsequent refinement process. Then, we propose a boundary guided prior constraint loss, consisting of boundary alignment and pixel-pair similarity mining losses, to fully leverage prior information to enhance the discriminative ability and provide reliable refinement guidance for the model. Finally, we propose a correlative semantic calibration loss, which comprehensively perceives the rich semantic features of different dimensions from both local and global perspectives. With the collaborative constraints of these meticulously designed loss functions, precise instance segmentation results are achieved. Experimental results showcase the outstanding performance of the PA-WSDIS model with an impressive 87.4% <span><math><msubsup><mrow><mi>mAP</mi></mrow><mrow><mn>50</mn></mrow><mrow><mi>mask</mi></mrow></msubsup></math></span>, which is considerably superior to state-of-the-art methods. As far as we know, our proposed method is the first weakly supervised instance segmentation model based on bounding box labels for industrial defect detection tasks.</div></div>","PeriodicalId":50941,"journal":{"name":"Advanced Engineering Informatics","volume":"65 ","pages":"Article 103254"},"PeriodicalIF":8.0,"publicationDate":"2025-03-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143682410","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
A transfer learning method: Universal domain adaptation with noisy samples for bearing fault diagnosis
IF 8 1区 工程技术 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE Pub Date : 2025-03-22 DOI: 10.1016/j.aei.2025.103243
Yi Sun, Hongliang Song, Liang Guo, Hongli Gao, Ao Cao
Under the influence of frequent start-stop driving and rail launching during the service of urban rail vehicles, the source domain samples contain a large number of noise labels and noise samples. Moreover, the feature distribution and sample categories of the target domain and source domain are different because the urban rail vehicles are affected by the fluctuation of passenger flow and long-term service. This paper summarizes this real task in rail transportation as universal domain adaptation with noisy samples (UDANS). A novel multibranch convolutional neural network is proposed to solve the above problem. By optimizing the divergence of the two classifier outputs, the following objectives can be achieved: detecting noisy source samples, finding private classes in the target domain, and aligning the distribution of the source domain and the target domain. Finally, the results of the wheelset bearing dataset show that the method has advantages in rail transportation fault diagnosis.
{"title":"A transfer learning method: Universal domain adaptation with noisy samples for bearing fault diagnosis","authors":"Yi Sun,&nbsp;Hongliang Song,&nbsp;Liang Guo,&nbsp;Hongli Gao,&nbsp;Ao Cao","doi":"10.1016/j.aei.2025.103243","DOIUrl":"10.1016/j.aei.2025.103243","url":null,"abstract":"<div><div>Under the influence of frequent start-stop driving and rail launching during the service of urban rail vehicles, the source domain samples contain a large number of noise labels and noise samples. Moreover, the feature distribution and sample categories of the target domain and source domain are different because the urban rail vehicles are affected by the fluctuation of passenger flow and long-term service. This paper summarizes this real task in rail transportation as universal domain adaptation with noisy samples (UDANS). A novel multibranch convolutional neural network is proposed to solve the above problem. By optimizing the divergence of the two classifier outputs, the following objectives can be achieved: detecting noisy source samples, finding private classes in the target domain, and aligning the distribution of the source domain and the target domain. Finally, the results of the wheelset bearing dataset show that the method has advantages in rail transportation fault diagnosis.</div></div>","PeriodicalId":50941,"journal":{"name":"Advanced Engineering Informatics","volume":"65 ","pages":"Article 103243"},"PeriodicalIF":8.0,"publicationDate":"2025-03-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143682406","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
Towards a self-cognitive complex product design system: A fine-grained multi-modal feature recognition and semantic understanding approach using large language models in mechanical engineering
IF 8 1区 工程技术 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE Pub Date : 2025-03-22 DOI: 10.1016/j.aei.2025.103265
Xinxin Liang, Zuoxu Wang, Jihong Liu
Facing the promising tendency of human-artificial intelligence (AI) collaborative product design, fine-grained and multi-modal mechanical part recognition and semantic understanding have become a basic task for achieving a self-cognitive product design system. However, traditional semantic understanding approaches for mechanical parts can only handle single-modal data, which is either textual or image data, resulting in the following limitations 1) insufficient mining on fine-grained part’s functional/behavioral/structural information, and 2) ineffectiveness on multi-modal part information alignment, therefore restricting the intelligence level of the previous product design assistants. To mitigate these challenges, this paper proposes a fine-grained multimodal reasoning approach for mechanical part semantic understanding. The proposed approach utilizes a pre-trained Convolutional Neural Network (CNN) for visual feature extraction, a large language model (LLM) called LLaMA3 for advanced textual analysis, and a Unified Feature Fusion Module (UFFM) to facilitate robust cross-modal interactions. A positive and negative sample generation mechanism is implemented to refine the model’s ability to discern subtle variations in complex components. Experimental evaluations on the Industrial Part Multimodal Dataset (IPMD) demonstrate a significant improvement in classification accuracy, providing a more precise and intelligent solution for the semantic understanding in complex product design systems.
{"title":"Towards a self-cognitive complex product design system: A fine-grained multi-modal feature recognition and semantic understanding approach using large language models in mechanical engineering","authors":"Xinxin Liang,&nbsp;Zuoxu Wang,&nbsp;Jihong Liu","doi":"10.1016/j.aei.2025.103265","DOIUrl":"10.1016/j.aei.2025.103265","url":null,"abstract":"<div><div>Facing the promising tendency of human-artificial intelligence (AI) collaborative product design, fine-grained and multi-modal mechanical part recognition and semantic understanding have become a basic task for achieving a self-cognitive product design system. However, traditional semantic understanding approaches for mechanical parts can only handle single-modal data, which is either textual or image data, resulting in the following limitations 1) insufficient mining on fine-grained part’s functional/behavioral/structural information, and 2) ineffectiveness on multi-modal part information alignment, therefore restricting the intelligence level of the previous product design assistants. To mitigate these challenges, this paper proposes a fine-grained multimodal reasoning approach for mechanical part semantic understanding. The proposed approach utilizes a pre-trained Convolutional Neural Network (CNN) for visual feature extraction, a large language model (LLM) called LLaMA3 for advanced textual analysis, and a Unified Feature Fusion Module (UFFM) to facilitate robust cross-modal interactions. A positive and negative sample generation mechanism is implemented to refine the model’s ability to discern subtle variations in complex components. Experimental evaluations on the Industrial Part Multimodal Dataset (IPMD) demonstrate a significant improvement in classification accuracy, providing a more precise and intelligent solution for the semantic understanding in complex product design systems.</div></div>","PeriodicalId":50941,"journal":{"name":"Advanced Engineering Informatics","volume":"65 ","pages":"Article 103265"},"PeriodicalIF":8.0,"publicationDate":"2025-03-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143682407","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
Using sociotechnical network modeling to analyze the impact of blockchain for supply chain on the risk of procuring counterfeit electronic parts
IF 8 1区 工程技术 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE Pub Date : 2025-03-22 DOI: 10.1016/j.aei.2025.103272
Hirbod Akhavantaheri , Peter Sandborn , Diganta Das
Safety-critical, mission-critical, and infrastructure-critical systems (e.g., aerospace, transportation, defense, and power generation) are forced to source parts over exceptionally long periods of time from a supply chain that they do not control. Such systems are exposed to the dual risks of the impacts of system failure and the exposure to an unauthorized electronics marketplace over decades. Therefore, critical systems operators, manufacturers, and sustainers, must implement policies and technologies to reduce the risk of obtaining counterfeit parts.
Blockchain technology, as a distributed ledger platform, has shown promise for resolving the issues associated with a lack of trust, transparency in peer-to-peer transactional networks, and compromised supply chains. There are opportunities to apply blockchain for supply chain concepts to mitigate the risks associated with part authenticity in the electronic part supply chain.
This paper introduces a supply-chain blockchain framework resilient to aging (e.g., the loss of involvement of the original component manufacture and its authorized distributors, and loss of part transaction history). An agent-based model is introduced as a novel platform to test the impact of the proposed blockchain framework on supply-chain parties as well as the prevalence of counterfeits in the electronics supply chain. The model can validate the proposed protocol over the entire life cycle of a part (i.e., from active production to discontinuance and beyond) and predict the parties’ adoption rates, and changes in the prevalence of counterfeit parts.
Application of the model to a public participation blockchain based on Ethereum ERC- 721 protocols indicates that the participation level of independent distributors directly affects the efficacy of blockchain in the prevention of transactions containing counterfeit parts. A proposed certification-based blockchain participation approach can be effective if certifications require large enough test accuracy limits and high previous owner certification thresholds.
{"title":"Using sociotechnical network modeling to analyze the impact of blockchain for supply chain on the risk of procuring counterfeit electronic parts","authors":"Hirbod Akhavantaheri ,&nbsp;Peter Sandborn ,&nbsp;Diganta Das","doi":"10.1016/j.aei.2025.103272","DOIUrl":"10.1016/j.aei.2025.103272","url":null,"abstract":"<div><div>Safety-critical, mission-critical, and infrastructure-critical systems (e.g., aerospace, transportation, defense, and power generation) are forced to source parts over exceptionally long periods of time from a supply chain that they do not control. Such systems are exposed to the dual risks of the impacts of system failure and the exposure to an unauthorized electronics marketplace over decades. Therefore, critical systems operators, manufacturers, and sustainers, must implement policies and technologies to reduce the risk of obtaining counterfeit parts.</div><div>Blockchain technology, as a distributed ledger platform, has shown promise for resolving the issues associated with a lack of trust, transparency in peer-to-peer transactional networks, and compromised supply chains. There are opportunities to apply blockchain for supply chain concepts to mitigate the risks associated with part authenticity in the electronic part supply chain.</div><div>This paper introduces a supply-chain blockchain framework resilient to aging (e.g., the loss of involvement of the original component manufacture and its authorized distributors, and loss of part transaction history). An agent-based model is introduced as a novel platform to test the impact of the proposed blockchain framework on supply-chain parties as well as the prevalence of counterfeits in the electronics supply chain. The model can validate the proposed protocol over the entire life cycle of a part (i.e., from active production to discontinuance and beyond) and predict the parties’ adoption rates, and changes in the prevalence of counterfeit parts.</div><div>Application of the model to a public participation blockchain based on Ethereum ERC- 721 protocols indicates that the participation level of independent distributors directly affects the efficacy of blockchain in the prevention of transactions containing counterfeit parts. A proposed certification-based blockchain participation approach can be effective if certifications require large enough test accuracy limits and high previous owner certification thresholds.</div></div>","PeriodicalId":50941,"journal":{"name":"Advanced Engineering Informatics","volume":"65 ","pages":"Article 103272"},"PeriodicalIF":8.0,"publicationDate":"2025-03-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143682409","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
引用次数: 0
A new perspective on non-ferrous metal price forecasting: An interpretable two-stage ensemble learning-based interval-valued forecasting system
IF 8 1区 工程技术 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE Pub Date : 2025-03-21 DOI: 10.1016/j.aei.2025.103267
Wendong Yang , Hao Zhang , Jianzhou Wang , Yan Hao
An accurate non-ferrous metal price prediction model is critical for formulating national economic policies, planning company production, and mitigating risk. Existing research improves the performance of prediction models based on point data but neglects the value of interval data and model interpretability, resulting in suboptimal predictions. Hence, this study proposes an interval-valued forecasting system for non-ferrous metal prices via interpretable two-stage ensemble learning. An interval-valued data preprocessing module is designed to improve predictive ability and enhance modeling diversity in terms of data by introducing various interval-valued mapping strategies. To enhance the modeling diversity of the predictors, a meta-predictor module that incorporates four advanced deep-learning models that produce various sub-predictors is proposed. A two-stage ensemble learning module is developed to obtain final interval-valued non-ferrous metal prices based on all sub-predictors. In the first stage, based on temporal fusion transformers, different deep-learning models are combined to reduce the bias in individual predictors. In the second stage, based on an attention mechanism, different interval-valued mapping strategies are combined to improve forecasting performance. Multiple comparative experiments and analyses are conducted using real non-ferrous metal market data. In an empirical study, the proposed system achieved the best results. Taking a copper dataset as an example, the system results for the IMAPE, IRMSE, IARV, and UI were 0.57826 %, 62.51197, 0.02147, and 0.14651, respectively. The results show that the proposed system not only outperforms both individual and advanced ensemble models in terms of accuracy and robustness but also offers valuable interpretable insights for improving interval-valued forecasting power.
{"title":"A new perspective on non-ferrous metal price forecasting: An interpretable two-stage ensemble learning-based interval-valued forecasting system","authors":"Wendong Yang ,&nbsp;Hao Zhang ,&nbsp;Jianzhou Wang ,&nbsp;Yan Hao","doi":"10.1016/j.aei.2025.103267","DOIUrl":"10.1016/j.aei.2025.103267","url":null,"abstract":"<div><div>An accurate non-ferrous metal price prediction model is critical for formulating national economic policies, planning company production, and mitigating risk. Existing research improves the performance of prediction models based on point data but neglects the value of interval data and model interpretability, resulting in suboptimal predictions. Hence, this study proposes an interval-valued forecasting system for non-ferrous metal prices via interpretable two-stage ensemble learning. An interval-valued data preprocessing module is designed to improve predictive ability and enhance modeling diversity in terms of data by introducing various interval-valued mapping strategies. To enhance the modeling diversity of the predictors, a meta-predictor module that incorporates four advanced deep-learning models that produce various sub-predictors is proposed. A two-stage ensemble learning module is developed to obtain final interval-valued non-ferrous metal prices based on all sub-predictors. In the first stage, based on temporal fusion transformers, different deep-learning models are combined to reduce the bias in individual predictors. In the second stage, based on an attention mechanism, different interval-valued mapping strategies are combined to improve forecasting performance. Multiple comparative experiments and analyses are conducted using real non-ferrous metal market data. In an empirical study, the proposed system achieved the best results. Taking a copper dataset as an example, the system results for the IMAPE, IRMSE, IARV, and UI were 0.57826 %, 62.51197, 0.02147, and 0.14651, respectively. The results show that the proposed system not only outperforms both individual and advanced ensemble models in terms of accuracy and robustness but also offers valuable interpretable insights for improving interval-valued forecasting power.</div></div>","PeriodicalId":50941,"journal":{"name":"Advanced Engineering Informatics","volume":"65 ","pages":"Article 103267"},"PeriodicalIF":8.0,"publicationDate":"2025-03-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143682403","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
A semi-Markov process based digital twin for safety evaluation of cable-stayed bridges with cable corrosion
IF 8 1区 工程技术 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE Pub Date : 2025-03-21 DOI: 10.1016/j.aei.2025.103270
Xin-Yu Guo , Sheng-En Fang , Xinqun Zhu , Jianchun Li
The cables of a cable-stayed bridge are susceptible to structural degradation due to environmental corrosion and fatigue, which directly affects the safety and operational performance of the bridge. As the process of the degradation in practice is very slow, it is difficult to be monitored during the bridge service life. Hence, this study aims to develop a novel semi-Markov process based digital twin (DT) framework for safety evaluation of cable-stayed bridges considering cable corrosion. The framework encompasses a physical twin layer, a DT layer and the information interaction medium. The physical twin layer mainly comprises the bridge physical entity and its associated monitoring system that provides a variety of perceptual data for DT modeling. In the DT layer, the DT model acts as a virtual counterpart of the physical bridge for mirroring and forecasting the bridge’s mechanical behaviors. The information interaction medium plays a crucial role in the bidirectional information communication between the physical and digital twin layers. Two types of information interaction media have been utilized including a cable force influence matrix and a semi-Markov process. The former enables updating the DT model to precisely match the data measured from the physical bridge. Meanwhile, the semi-Markov process depicts the probability of the bridge’s condition considering the cable corrosion during the different service periods. The proposed procedure can predict the bridge state and evaluate the safety by comparing the predicted state with the monitored values. The proposed framework has been successfully validated on a real-world cable-stayed bridge. The results showed the proposed DT framework was reliable and effective for evaluating the bridge condition.
{"title":"A semi-Markov process based digital twin for safety evaluation of cable-stayed bridges with cable corrosion","authors":"Xin-Yu Guo ,&nbsp;Sheng-En Fang ,&nbsp;Xinqun Zhu ,&nbsp;Jianchun Li","doi":"10.1016/j.aei.2025.103270","DOIUrl":"10.1016/j.aei.2025.103270","url":null,"abstract":"<div><div>The cables of a cable-stayed bridge are susceptible to structural degradation due to environmental corrosion and fatigue, which directly affects the safety and operational performance of the bridge. As the process of the degradation in practice is very slow, it is difficult to be monitored during the bridge service life. Hence, this study aims to develop a novel semi-Markov process based digital twin (DT) framework for safety evaluation of cable-stayed bridges considering cable corrosion. The framework encompasses a physical twin layer, a DT layer and the information interaction medium. The physical twin layer mainly comprises the bridge physical entity and its associated monitoring system that provides a variety of perceptual data for DT modeling. In the DT layer, the DT model acts as a virtual counterpart of the physical bridge for mirroring and forecasting the bridge’s mechanical behaviors. The information interaction medium plays a crucial role in the bidirectional information communication between the physical and digital twin layers. Two types of information interaction media have been utilized including a cable force influence matrix and a semi-Markov process. The former enables updating the DT model to precisely match the data measured from the physical bridge. Meanwhile, the semi-Markov process depicts the probability of the bridge’s condition considering the cable corrosion during the different service periods. The proposed procedure can predict the bridge state and evaluate the safety by comparing the predicted state with the monitored values. The proposed framework has been successfully validated on a real-world cable-stayed bridge. The results showed the proposed DT framework was reliable and effective for evaluating the bridge condition.</div></div>","PeriodicalId":50941,"journal":{"name":"Advanced Engineering Informatics","volume":"65 ","pages":"Article 103270"},"PeriodicalIF":8.0,"publicationDate":"2025-03-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143682404","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
Dual graph driven-consistent representation learning method for semi-supervised fault diagnosis of rotating machinery
IF 8 1区 工程技术 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE Pub Date : 2025-03-21 DOI: 10.1016/j.aei.2025.103274
Zhichao Jiang , Dongdong Liu , Huaqing Wang , Lingli Cui
Graph neural network (GNN) is an effective tool for semi-supervised fault diagnosis of rotating machinery. However, existing GNN based-semi-supervised methods only rely on single graph structure to learn feature representation under limited labeled samples, while the information of different topology graph structures cannot be directly fused due to the large difference of feature extracting, leading to insufficient node relationships and label information mining. Besides, static or limited dynamic feature extraction of neighbor nodes will hinder the expressiveness of semi-supervised GNN models. To overcome these limitations, a dual graph driven-consistent representation learning method (DGDCRL) is proposed in this paper. First, a dual graph structure with two different topology graphs is conducted using graph label passing method, in which limited labeled sample information are fully leveraged and richer topology structure information among nodes can be captured. Second, a consistent representation learning method with gated-dynamic enhanced graph attention module (GDEGAT) is proposed to extract the common embeddings from two topology graphs, where a DEGAT layer is developed to aggregate neighbor information more dynamically and expressively. Besides, to enhance the alignment between the embeddings of the same nodes across two topology graphs, we design a consistent representation loss. Two datasets are used to validate the performance of the proposed method, indicating that the proposed DGDCRL method with GDEGAT module can achieve the effective diagnosis results of rotating machinery under both constant and variable speed conditions, and the DGDCRL method can effectively enhance the semi-supervised diagnostic ability of baseline GNNs under low labeled rates.
{"title":"Dual graph driven-consistent representation learning method for semi-supervised fault diagnosis of rotating machinery","authors":"Zhichao Jiang ,&nbsp;Dongdong Liu ,&nbsp;Huaqing Wang ,&nbsp;Lingli Cui","doi":"10.1016/j.aei.2025.103274","DOIUrl":"10.1016/j.aei.2025.103274","url":null,"abstract":"<div><div>Graph neural network (GNN) is an effective tool for semi-supervised fault diagnosis of rotating machinery. However, existing GNN based-semi-supervised methods only rely on single graph structure to learn feature representation under limited labeled samples, while the information of different topology graph structures cannot be directly fused due to the large difference of feature extracting, leading to insufficient node relationships and label information mining. Besides, static or limited dynamic feature extraction of neighbor nodes will hinder the expressiveness of semi-supervised GNN models. To overcome these limitations, a dual graph driven-consistent representation learning method (DGDCRL) is proposed in this paper. First, a dual graph structure with two different topology graphs is conducted using graph label passing method, in which limited labeled sample information are fully leveraged and richer topology structure information among nodes can be captured. Second, a consistent representation learning method with gated-dynamic enhanced graph attention module (GDEGAT) is proposed to extract the common embeddings from two topology graphs, where a DEGAT layer is developed to aggregate neighbor information more dynamically and expressively. Besides, to enhance the alignment between the embeddings of the same nodes across two topology graphs, we design a consistent representation loss. Two datasets are used to validate the performance of the proposed method, indicating that the proposed DGDCRL method with GDEGAT module can achieve the effective diagnosis results of rotating machinery under both constant and variable speed conditions, and the DGDCRL method can effectively enhance the semi-supervised diagnostic ability of baseline GNNs under low labeled rates.</div></div>","PeriodicalId":50941,"journal":{"name":"Advanced Engineering Informatics","volume":"65 ","pages":"Article 103274"},"PeriodicalIF":8.0,"publicationDate":"2025-03-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143682405","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
Multi-source information fusion for dynamic safety risk prediction of aerial building machine using spatial–temporal multi-graph convolution network
IF 8 1区 工程技术 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE Pub Date : 2025-03-19 DOI: 10.1016/j.aei.2025.103261
Jiaqi Wang , Yuqing Fan , Xi Pan , Jun Sun , Limao Zhang
Aerial Building Machine (ABM) is an innovative and comprehensive construction equipment employed in high-rise building construction, capable of climbing upwards through the use of multiple hydraulic cylinders. The lifting operation of ABM is a critical phase for construction safety, yet there is limited research on forecasting and warning for the ABM lifting process. This study proposes a novel spatial–temporal forecasting model that combines the Graph Neural Network (GNN) and Temporal Convolutional Network (TCN), along with a computational framework for modeling multi-source data fusion and multi-graph construction. Specifically, fast-DTW and direct-LiNGAM are used to capture and analyze the internal relationships within monitoring data from diverse sensors. The proposed Multi-Graph TCN (MGTCN) fuses spatial and temporal features to make multi-step ahead predictions and provides explanations for dynamic safety risks. An ABM case in China is employed to illustrate the feasibility and effectiveness of the proposed framework. The results indicate that: (1) MGTCN exhibits high accuracy in multi-step prediction, with an average R2 of 0.917 at 10-step prediction; (2) MGTCN demonstrates strong robustness, maintaining the R2 change rate within 3% for every two steps up to the 20-step prediction; (3) The proposed model outperforms others in terms of generalization and stability when compared to models using single-source data or single-graph construction. The contribution of this research lies in transforming large-scale engineering data into graph-based prior knowledge, which is input into a spatial–temporal fusion GNN prediction model, improving accuracy and robustness. The significance of this study is in addressing a gap in ABM risk prediction and advancing time series prediction research in construction engineering. The utilization of intelligent prediction techniques not only ensures the safety management of lifting operations with ABM but also promises to improve the informatization of construction processes for high-rise buildings.
{"title":"Multi-source information fusion for dynamic safety risk prediction of aerial building machine using spatial–temporal multi-graph convolution network","authors":"Jiaqi Wang ,&nbsp;Yuqing Fan ,&nbsp;Xi Pan ,&nbsp;Jun Sun ,&nbsp;Limao Zhang","doi":"10.1016/j.aei.2025.103261","DOIUrl":"10.1016/j.aei.2025.103261","url":null,"abstract":"<div><div>Aerial Building Machine (ABM) is an innovative and comprehensive construction equipment employed in high-rise building construction, capable of climbing upwards through the use of multiple hydraulic cylinders. The lifting operation of ABM is a critical phase for construction safety, yet there is limited research on forecasting and warning for the ABM lifting process. This study proposes a novel spatial–temporal forecasting model that combines the Graph Neural Network (GNN) and Temporal Convolutional Network (TCN), along with a computational framework for modeling multi-source data fusion and multi-graph construction. Specifically, fast-DTW and direct-LiNGAM are used to capture and analyze the internal relationships within monitoring data from diverse sensors. The proposed Multi-Graph TCN (MGTCN) fuses spatial and temporal features to make multi-step ahead predictions and provides explanations for dynamic safety risks. An ABM case in China is employed to illustrate the feasibility and effectiveness of the proposed framework. The results indicate that: (1) MGTCN exhibits high accuracy in multi-step prediction, with an average R<sup>2</sup> of 0.917 at 10-step prediction; (2) MGTCN demonstrates strong robustness, maintaining the R<sup>2</sup> change rate within 3% for every two steps up to the 20-step prediction; (3) The proposed model outperforms others in terms of generalization and stability when compared to models using single-source data or single-graph construction. The contribution of this research lies in transforming large-scale engineering data into graph-based prior knowledge, which is input into a spatial–temporal fusion GNN prediction model, improving accuracy and robustness. The significance of this study is in addressing a gap in ABM risk prediction and advancing time series prediction research in construction engineering. The utilization of intelligent prediction techniques not only ensures the safety management of lifting operations with ABM but also promises to improve the informatization of construction processes for high-rise buildings.</div></div>","PeriodicalId":50941,"journal":{"name":"Advanced Engineering Informatics","volume":"65 ","pages":"Article 103261"},"PeriodicalIF":8.0,"publicationDate":"2025-03-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143654839","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
Segmentation refinement of thin cracks with Minimum Strip Cuts
IF 8 1区 工程技术 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE Pub Date : 2025-03-19 DOI: 10.1016/j.aei.2025.103249
Wanchen Hou , Jingyuan He , Chenghao Cui , Fan Zhong , Xinbo Jiang , Lin Lu , Jizhe Zhang , Changhe Tu
Accurate segmentation is crucial for dataset labeling and morphological analysis of thin cracks, for which small segmentation error may take great effect to the results. However, due to the complex morphology and thin structures, so far it is still very hard to accurately segment thin cracks either by manual or by segmentation algorithms. In this paper we propose an approach for accurate and efficient segmentation of thin cracks. The core of our approach is an optimization-based image segmentation method designed specifically for extracting optimal thin strip regions from images. Our method can be considered as an extension of previous minimum path search problem, so we call it as Minimum Strip Cuts, or StripCuts for short. An effective objective function for segmentation refinement of thin cracks is proposed, whose global optimal solution can be obtained efficiently based on the proposed volumetric dynamic programming and crack linearization methods. Our method is robust to low-contrast cracks and complex background, and can run in real-time. Therefore, it can be used for refining exist datasets, and also for the post-processing of crack segmentation methods. Based on the proposed refinement method, we introduce a new crack segmentation dataset RefinedCracks, which provides accurate refined annotations for previous main crack segmentation datasets. The importance of refinement to the training and evaluation of crack segmentation methods is also verified by both quantitative and qualitative evaluations.
{"title":"Segmentation refinement of thin cracks with Minimum Strip Cuts","authors":"Wanchen Hou ,&nbsp;Jingyuan He ,&nbsp;Chenghao Cui ,&nbsp;Fan Zhong ,&nbsp;Xinbo Jiang ,&nbsp;Lin Lu ,&nbsp;Jizhe Zhang ,&nbsp;Changhe Tu","doi":"10.1016/j.aei.2025.103249","DOIUrl":"10.1016/j.aei.2025.103249","url":null,"abstract":"<div><div>Accurate segmentation is crucial for dataset labeling and morphological analysis of thin cracks, for which small segmentation error may take great effect to the results. However, due to the complex morphology and thin structures, so far it is still very hard to accurately segment thin cracks either by manual or by segmentation algorithms. In this paper we propose an approach for accurate and efficient segmentation of thin cracks. The core of our approach is an optimization-based image segmentation method designed specifically for extracting optimal thin strip regions from images. Our method can be considered as an extension of previous minimum path search problem, so we call it as Minimum Strip Cuts, or StripCuts for short. An effective objective function for segmentation refinement of thin cracks is proposed, whose global optimal solution can be obtained efficiently based on the proposed volumetric dynamic programming and crack linearization methods. Our method is robust to low-contrast cracks and complex background, and can run in real-time. Therefore, it can be used for refining exist datasets, and also for the post-processing of crack segmentation methods. Based on the proposed refinement method, we introduce a new crack segmentation dataset RefinedCracks, which provides accurate refined annotations for previous main crack segmentation datasets. The importance of refinement to the training and evaluation of crack segmentation methods is also verified by both quantitative and qualitative evaluations.</div></div>","PeriodicalId":50941,"journal":{"name":"Advanced Engineering Informatics","volume":"65 ","pages":"Article 103249"},"PeriodicalIF":8.0,"publicationDate":"2025-03-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143654842","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
A two-stage multisource heterogeneous information fusion framework for operating condition identification of industrial rotary kilns 用于工业回转窑运行状况识别的两阶段多源异构信息融合框架
IF 8 1区 工程技术 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE Pub Date : 2025-03-19 DOI: 10.1016/j.aei.2025.103251
Fengrun Tang , Yonggang Li , Fan Mo , Chunhua Yang , Bei Sun
The operating condition identification plays an irreplaceable role for the low-carbon and high-efficiency operation of industrial rotary kilns. However, existing single-stage multisource heterogeneous information fusion methods lack a unified framework to simultaneously fuse the complementary properties among visible images, infrared images, and process data, thus limiting the condition recognition accuracy. Moreover, smoke and dust interference make it challenging to extract critical image features such as flame brightness and blast pipe position, increasing the difficulty of condition recognition. To this end, this paper proposes a two-stage multisource heterogeneous information fusion (TSMHIF) framework for operating condition identification of industrial rotary kilns. First, in the initial fusion stage, a condition-aware visible and infrared image fusion network (CAVIF) is designed to generate fused images containing complementary properties of source images. In this network, a self-developed novel industrial system is utilized to collect aligned visible-infrared images of industrial rotary kilns. Next, an interpretable feature engineering is constructed by incorporating extracted shallow features based on mechanism knowledge and mined deep features with an autoencoder, and the blast pipe position in the shallow features is quantified by a keypoint detection algorithm based on a cascaded pyramid network (CPN). Then, in the comprehensive fusion stage, a multiplication operation is employed to fuse multisource heterogeneous deep features from fused images and process data to recognize the operating conditions. Finally, a joint training strategy is developed to balance the image fusion and condition classification networks. The classification loss, i.e., condition-aware loss, guides the training of the visible-infrared image fusion network to improve the visual quality of the fused images. The industrial experiments show that our proposed method exhibits superior performance in terms of identification accuracy, condition prediction deviation, and visual quality of fused images compared to other competitors.
{"title":"A two-stage multisource heterogeneous information fusion framework for operating condition identification of industrial rotary kilns","authors":"Fengrun Tang ,&nbsp;Yonggang Li ,&nbsp;Fan Mo ,&nbsp;Chunhua Yang ,&nbsp;Bei Sun","doi":"10.1016/j.aei.2025.103251","DOIUrl":"10.1016/j.aei.2025.103251","url":null,"abstract":"<div><div>The operating condition identification plays an irreplaceable role for the low-carbon and high-efficiency operation of industrial rotary kilns. However, existing single-stage multisource heterogeneous information fusion methods lack a unified framework to simultaneously fuse the complementary properties among visible images, infrared images, and process data, thus limiting the condition recognition accuracy. Moreover, smoke and dust interference make it challenging to extract critical image features such as flame brightness and blast pipe position, increasing the difficulty of condition recognition. To this end, this paper proposes a two-stage multisource heterogeneous information fusion (TSMHIF) framework for operating condition identification of industrial rotary kilns. First, in the initial fusion stage, a condition-aware visible and infrared image fusion network (CAVIF) is designed to generate fused images containing complementary properties of source images. In this network, a self-developed novel industrial system is utilized to collect aligned visible-infrared images of industrial rotary kilns. Next, an interpretable feature engineering is constructed by incorporating extracted shallow features based on mechanism knowledge and mined deep features with an autoencoder, and the blast pipe position in the shallow features is quantified by a keypoint detection algorithm based on a cascaded pyramid network (CPN). Then, in the comprehensive fusion stage, a multiplication operation is employed to fuse multisource heterogeneous deep features from fused images and process data to recognize the operating conditions. Finally, a joint training strategy is developed to balance the image fusion and condition classification networks. The classification loss, i.e., condition-aware loss, guides the training of the visible-infrared image fusion network to improve the visual quality of the fused images. The industrial experiments show that our proposed method exhibits superior performance in terms of identification accuracy, condition prediction deviation, and visual quality of fused images compared to other competitors.</div></div>","PeriodicalId":50941,"journal":{"name":"Advanced Engineering Informatics","volume":"65 ","pages":"Article 103251"},"PeriodicalIF":8.0,"publicationDate":"2025-03-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143654515","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
期刊
Advanced Engineering Informatics
全部 Acc. Chem. Res. ACS Applied Bio Materials ACS Appl. Electron. Mater. ACS Appl. Energy Mater. ACS Appl. Mater. Interfaces ACS Appl. Nano Mater. ACS Appl. Polym. Mater. ACS BIOMATER-SCI ENG ACS Catal. ACS Cent. Sci. ACS Chem. Biol. ACS Chemical Health & Safety ACS Chem. Neurosci. ACS Comb. Sci. ACS Earth Space Chem. ACS Energy Lett. ACS Infect. Dis. ACS Macro Lett. ACS Mater. Lett. ACS Med. Chem. Lett. ACS Nano ACS Omega ACS Photonics ACS Sens. ACS Sustainable Chem. Eng. ACS Synth. Biol. Anal. Chem. BIOCHEMISTRY-US Bioconjugate Chem. BIOMACROMOLECULES Chem. Res. Toxicol. Chem. Rev. Chem. Mater. CRYST GROWTH DES ENERG FUEL Environ. Sci. Technol. Environ. Sci. Technol. Lett. Eur. J. Inorg. Chem. IND ENG CHEM RES Inorg. Chem. J. Agric. Food. Chem. J. Chem. Eng. Data J. Chem. Educ. J. Chem. Inf. Model. J. Chem. Theory Comput. J. Med. Chem. J. Nat. Prod. J PROTEOME RES J. Am. Chem. Soc. LANGMUIR MACROMOLECULES Mol. Pharmaceutics Nano Lett. Org. Lett. ORG PROCESS RES DEV ORGANOMETALLICS J. Org. Chem. J. Phys. Chem. J. Phys. Chem. A J. Phys. Chem. B J. Phys. Chem. C J. Phys. Chem. Lett. Analyst Anal. Methods Biomater. Sci. Catal. Sci. Technol. Chem. Commun. Chem. Soc. Rev. CHEM EDUC RES PRACT CRYSTENGCOMM Dalton Trans. Energy Environ. Sci. ENVIRON SCI-NANO ENVIRON SCI-PROC IMP ENVIRON SCI-WAT RES Faraday Discuss. Food Funct. Green Chem. Inorg. Chem. Front. Integr. Biol. J. Anal. At. Spectrom. J. Mater. Chem. A J. Mater. Chem. B J. Mater. Chem. C Lab Chip Mater. Chem. Front. Mater. Horiz. MEDCHEMCOMM Metallomics Mol. Biosyst. Mol. Syst. Des. Eng. Nanoscale Nanoscale Horiz. Nat. Prod. Rep. New J. Chem. Org. Biomol. Chem. Org. Chem. Front. PHOTOCH PHOTOBIO SCI PCCP Polym. Chem.
×
引用
GB/T 7714-2015
复制
MLA
复制
APA
复制
导出至
BibTeX EndNote RefMan NoteFirst NoteExpress
×
0
微信
客服QQ
Book学术公众号 扫码关注我们
反馈
×
意见反馈
请填写您的意见或建议
请填写您的手机或邮箱
×
提示
您的信息不完整,为了账户安全,请先补充。
现在去补充
×
提示
您因"违规操作"
具体请查看互助需知
我知道了
×
提示
现在去查看 取消
×
提示
确定
Book学术官方微信
Book学术文献互助
Book学术文献互助群
群 号:481959085
Book学术
文献互助 智能选刊 最新文献 互助须知 联系我们:info@booksci.cn
Book学术提供免费学术资源搜索服务,方便国内外学者检索中英文文献。致力于提供最便捷和优质的服务体验。
Copyright © 2023 Book学术 All rights reserved.
ghs 京公网安备 11010802042870号 京ICP备2023020795号-1