IEEE Transactions on Industrial Informatics最新文献_第9页

Digital Twin for Multitimescale Economic Dispatch in Cyber-Physical Energy Systems 信息物理能源系统中多时间尺度经济调度的数字孪生

IF 9.8 1区计算机科学

IEEE Transactions on Industrial Informatics Pub Date : 2026-10-01 Epub Date: 2026-06-11 DOI: 10.1109/TII.2026.3699131

Jiaqi Tian;Bonan Huang;Tianyi Li;David Wenzhong Gao;Zhongming Yao;Yumeng Song;Yushuai Li

{"title":"Digital Twin for Multitimescale Economic Dispatch in Cyber-Physical Energy Systems","authors":"Jiaqi Tian;Bonan Huang;Tianyi Li;David Wenzhong Gao;Zhongming Yao;Yumeng Song;Yushuai Li","doi":"10.1109/TII.2026.3699131","DOIUrl":"https://doi.org/10.1109/TII.2026.3699131","url":null,"abstract":"Cyber-physical energy systems (CPESs) involve multitimescale decision-making to balance short- and long-term dispatch requirements. However, existing distributed economic dispatch methods typically emphasize either short-term optimality or long-term predictive capability but rarely achieve both simultaneously. To address this issue, we propose a digital twin (DT)-empowered framework, model, and method for multitimescale economic dispatch (MTED) in CPESs. First, we design a two-layer MTED framework that integrates a physical layer and a DT layer. The DT layer consists of individual DTs, which serve as virtual replicas of the physical components in the physical layer, and a federated DT that coordinates the interactions among these individual DTs. Second, we formulate the MTED problem in CPESs as a complex optimization problem that involves both nonconvex and convex decision variables, related to long and short timescales, respectively. Third, we propose a federated MuZero-Newton-descent method to solve the MTED problem. By using this method, we can make the short-term optimal decision while ensuring long-term predictive capability for planning. Finally, simulations demonstrate the effectiveness of the proposed method.","PeriodicalId":13301,"journal":{"name":"IEEE Transactions on Industrial Informatics","volume":"22 10","pages":"9325-9336"},"PeriodicalIF":9.8,"publicationDate":"2026-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149097127","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

PurifyFD: Mining Latent Domains for Trustworthy Single-Domain Fault Diagnosis Under Open-World Shifts PurifyFD：开放世界变换下可信赖单域故障诊断的潜在域挖掘

IF 9.8 1区计算机科学

IEEE Transactions on Industrial Informatics Pub Date : 2026-10-01 Epub Date: 2026-06-12 DOI: 10.1109/TII.2026.3696400

Bowen Yu;Zhonglei Shen;Donghai Han;Shaocheng Kong;Zhibin Zhao;Liuyang Zhang;Ruqiang Yan;Xuefeng Chen

{"title":"PurifyFD: Mining Latent Domains for Trustworthy Single-Domain Fault Diagnosis Under Open-World Shifts","authors":"Bowen Yu;Zhonglei Shen;Donghai Han;Shaocheng Kong;Zhibin Zhao;Liuyang Zhang;Ruqiang Yan;Xuefeng Chen","doi":"10.1109/TII.2026.3696400","DOIUrl":"https://doi.org/10.1109/TII.2026.3696400","url":null,"abstract":"Single-domain generalization (SDG) is critical for Industrial Internet of Things edge scenarios where only limited single-source data are available. Existing SDG approaches predominantly rely on generative models or heavy augmentation to synthesize auxiliary domains, which often introduce unrealistic artifacts and training instability. In contrast, we articulate that single-source industrial data inherently contain latent nonstationarities (e.g., temporal drifts) that can be mined rather than synthesized. Based on this insight, we propose <italic>PurifyFD</i>, a trustworthy and edge-efficient framework that reframes SDG as a feature purification problem. Specifically, we devise a closed-loop max–min adversarial game: the maximization step automatically amplifies the worst-case discrepancies of mined latent domains, while the minimization step suppresses feature sensitivity to these shifts via spatiotemporal disentanglement. This process yields purified representations with high intra-class compactness and shift-invariance. Rather than relying on heavy auxiliary out-of-distribution modules, the resulting compact representation naturally supports lightweight parameter-free energy scoring for more reliable open-set rejection in edge deployment. Extensive experiments on case western reserve university (CWRU), JNU, and a dual-testbed dataset demonstrate that PurifyFD achieves a favorable trade-off among cross-domain accuracy, open-set reliability, and inference efficiency compared to state-of-the-art representative methods.","PeriodicalId":13301,"journal":{"name":"IEEE Transactions on Industrial Informatics","volume":"22 10","pages":"8895-8906"},"PeriodicalIF":9.8,"publicationDate":"2026-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149097249","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

CE-MAC: Cloud–Edge Multiagent Collaboration for Adaptive Model Pruning CE-MAC：自适应模型修剪的云边缘多代理协作

IF 9.8 1区计算机科学

IEEE Transactions on Industrial Informatics Pub Date : 2026-10-01 Epub Date: 2026-06-12 DOI: 10.1109/TII.2026.3697192

Qiyu Sun;Yangcheng Wang;Mengjia Gong;Jing Xu;Yang Tang

{"title":"CE-MAC: Cloud–Edge Multiagent Collaboration for Adaptive Model Pruning","authors":"Qiyu Sun;Yangcheng Wang;Mengjia Gong;Jing Xu;Yang Tang","doi":"10.1109/TII.2026.3697192","DOIUrl":"https://doi.org/10.1109/TII.2026.3697192","url":null,"abstract":"Deploying high accuracy models on resource-limited edge is essential for industrial applications, yet challenging due to the diverse and dynamic real-world environment. Cloud–edge collaboration helps by training high-capacity models in the cloud and deploying compressed models to edge nodes. However, traditional pipelines assume a fixed setting and lack the flexibility to adapt compression strategies to the heterogeneous and evolving environment. To overcome this, we propose CE-MAC, a cloud–edge collaborative large language model (LLM)-based multiagent system (LLM-MAS) for adaptive edge model compression. We adopt model pruning for its support of customizable and progressive compression, enabling flexible adaptation in dynamic edge environments. The LLM-MAS acts as the decision hub, generating device-specific pruning policies based on real-time feedback from both cloud and edge, enabling better cloud–edge co-evolution. We evaluate CE-MAC on image classification, keyword spotting, bearing fault diagnosis, and electricity theft detection tasks, achieving state-of-the-art performance across heterogeneous data and varying compression levels.","PeriodicalId":13301,"journal":{"name":"IEEE Transactions on Industrial Informatics","volume":"22 10","pages":"8990-9000"},"PeriodicalIF":9.8,"publicationDate":"2026-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149097714","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Semantic Block Reconstruction for Refactoring Industrial Control Systems in Flexible Manufacturing 基于语义块重构的柔性制造工业控制系统

IF 9.8 1区计算机科学

IEEE Transactions on Industrial Informatics Pub Date : 2026-10-01 Epub Date: 2026-07-07 DOI: 10.1109/TII.2026.3701792

Maodong Lin;Qinyun Hu;Wenchao Zou;Klaus Wendelberger;Wenbin Dai

引用次数: 0

MST-Net: A Multiscale Spatio-Temporal Learning Framework for Short-Term Wind Resource Forecasting MST-Net：用于短期风资源预报的多尺度时空学习框架

IF 9.8 1区计算机科学

IEEE Transactions on Industrial Informatics Pub Date : 2026-10-01 Epub Date: 2026-06-18 DOI: 10.1109/TII.2026.3698570

Yimin Lou;Gary G. Yen;Can Zhang;Zhenan He

{"title":"MST-Net: A Multiscale Spatio-Temporal Learning Framework for Short-Term Wind Resource Forecasting","authors":"Yimin Lou;Gary G. Yen;Can Zhang;Zhenan He","doi":"10.1109/TII.2026.3698570","DOIUrl":"10.1109/TII.2026.3698570","url":null,"abstract":"Ultra-short-term and short-term wind resource forecasting is crucial for optimizing wind farm dispatch efficiency and ensuring grid stability. However, the high volatility of wind resources makes accurate predictions challenging, but not completely infeasible. Conventional methods often fail to effectively capture both temporal dynamics of wind resources and spatial dependencies among turbines, leading to limited accuracy. This article proposes MST-Net, a multiscale spatio-temporal learning network, aiming at enhancing short-term wind speed and direction forecasting for multiple turbines within a wind farm. Regarding temporal feature extraction, the multitime-scale reshape and intrainter periodic time series prediction modules are employed. They capture short-term local dependencies and long-term periodic patterns by extracting multiscale features across adjacent and phase-aligned time points, respectively. For spatial feature extraction, the channel leading indicator refinement module innovatively incorporates geographic location and wind direction to identify upstream turbines, quantifying the lead-lag relationships between upstream and target turbines by calculating correlation coefficients. In addition, this is accomplished in a two-stage training strategy to balance channel independence and dependence: the first stage focuses on learning temporal patterns independently for each channel, while the second stage incorporates spatial correlation between channels. On two real-world wind farm datasets with distinct periodic characteristics and turbine density distributions, MST-Net achieves consistent improvements. Specifically, it reduces wind speed mean squared error by 15.1% on the Shanxi dataset and 13.3% on the Penmanshiel dataset compared with state-of-the-art baselines, demonstrating robust generalization under complex spatio-temporal coupling conditions.","PeriodicalId":13301,"journal":{"name":"IEEE Transactions on Industrial Informatics","volume":"22 10","pages":"9245-9256"},"PeriodicalIF":9.8,"publicationDate":"2026-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148289088","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

ROSOMP: Coordinated Scheduling of OpenMP Parallel Regions in ROS 2 Executor ROSOMP: ROS 2执行器中OpenMP并行区域的协调调度

IF 9.8 1区计算机科学

IEEE Transactions on Industrial Informatics Pub Date : 2026-10-01 Epub Date: 2026-06-09 DOI: 10.1109/TII.2026.3695353

Zilong Wang;Songran Liu;Bochao Cao;Yang Wang;Minghe Yu;Decai Li;Wang Yi

{"title":"ROSOMP: Coordinated Scheduling of OpenMP Parallel Regions in ROS 2 Executor","authors":"Zilong Wang;Songran Liu;Bochao Cao;Yang Wang;Minghe Yu;Decai Li;Wang Yi","doi":"10.1109/TII.2026.3695353","DOIUrl":"10.1109/TII.2026.3695353","url":null,"abstract":"The Robot Operating System (ROS 2) is widely adopted for modular robotic applications. OpenMP provides an approach for parallelizing CPU-bound workloads, and is commonly used in ROS-based applications for perception and control. However, integrating OpenMP into the ROS 2 executor introduces semantic mismatches: OpenMP threads are managed independently by the OS scheduler, breaking ROS-level priority semantics and leading to priority inversion, thread oversubscription, and timing unpredictability. This article presents ROSOMP, a unified framework that enables priority-driven scheduling and bounded execution of OpenMP parallel regions within the ROS 2 runtime. By converting OpenMP constructs into schedulable units and executing them within the existing ROS 2 executor thread pool, ROSOMP enforces global priority consistency while preserving the default programming model of ROS 2 and OpenMP. Furthermore, we develop a response time analysis to formally establish the predictability of ROSOMP. Experimental results demonstrate that ROSOMP significantly improves timing determinism and reduces system overhead compared to conventional OpenMP integration in ROS 2.","PeriodicalId":13301,"journal":{"name":"IEEE Transactions on Industrial Informatics","volume":"22 10","pages":"8789-8800"},"PeriodicalIF":9.8,"publicationDate":"2026-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148289265","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Multihead Gating and Self-Attention Network for Predicting Hydro-Wind-Geothermal Renewable Energy System 水力-风力-地热可再生能源系统预测的多头闸门自关注网络

IF 9.8 1区计算机科学

IEEE Transactions on Industrial Informatics Pub Date : 2026-10-01 Epub Date: 2026-06-23 DOI: 10.1109/TII.2026.3701071

Han Wu;Haidong Shao;Xiao-Zhi Gao;Jiani Heng;Zhe Chen

{"title":"Multihead Gating and Self-Attention Network for Predicting Hydro-Wind-Geothermal Renewable Energy System","authors":"Han Wu;Haidong Shao;Xiao-Zhi Gao;Jiani Heng;Zhe Chen","doi":"10.1109/TII.2026.3701071","DOIUrl":"10.1109/TII.2026.3701071","url":null,"abstract":"The precise prediction of hybrid hydro-wind-geothermal renewable energy (HRE) generation is essential for the stability, optimization, and intelligent management of modern integrated energy systems. However, existing HRE prediction methods face three key challenges: first, unified feature extraction causes interference among linear, slow-varying nonlinear, and abrupt temporal changes, second, the lack of structured feature-level learning limits the ability to handle complex fluctuations, and third, intrinsic correlations among coupled renewable sources are often overlooked. To tackle these challenges, this article introduces a multihead gating and self-attention network, named MGS-Net, which includes four components: a data denoising part, a multihead gating part, a self-attention part, and a prediction and training part. Specifically, to address the first challenge, the multihead gating part employs three parallel gated convolutional branches with a “divide-and-conquer” strategy to separate the extraction of linear trends, slow-varying nonlinearities, and abrupt changes. To overcome the second challenge, the self-attention part leverages the query-key-value paradigm to enable structured feature-level learning and reinforce temporal memory. To tackle the third challenge, the overall hierarchical framework systematically models intrinsic correlations among hydro, wind, and geothermal sources. In addition, the data denoising part employs complete ensemble empirical mode decomposition with adaptive noise to mitigate noise, while the prediction and training part incorporates a Gaussian kernel-based loss function to enhance training stability. Experimental results demonstrate that MGS-Net outperforms deep learning benchmarks with recursive, parallel, and hybrid structures, and achieves an overall mean absolute percentage error reduction of 19.7% compared to the best-performing benchmark (convolutional neural network-gated recurrent unit).","PeriodicalId":13301,"journal":{"name":"IEEE Transactions on Industrial Informatics","volume":"22 10","pages":"9432-9443"},"PeriodicalIF":9.8,"publicationDate":"2026-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148314444","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Vulnerability Analysis for Virtual Power Plant Voltage Support Service Under Informed DoS Attacks 知情DoS攻击下虚拟电厂电压保障服务的漏洞分析

IF 9.8 1区计算机科学

IEEE Transactions on Industrial Informatics Pub Date : 2026-10-01 Epub Date: 2026-07-02 DOI: 10.1109/TII.2026.3702910

Juanwei Chen;Jun Yan;Marthe Kassouf;Mourad Debbabi

{"title":"Vulnerability Analysis for Virtual Power Plant Voltage Support Service Under Informed DoS Attacks","authors":"Juanwei Chen;Jun Yan;Marthe Kassouf;Mourad Debbabi","doi":"10.1109/TII.2026.3702910","DOIUrl":"10.1109/TII.2026.3702910","url":null,"abstract":"Virtual power plants (VPPs) enable scalable coordination of distributed energy resources (DERs) for services such as energy trading and voltage support. However, the extensive communication required introduces significant cybersecurity risks by expanding the attack surface. This article presents an informed denial-of-service attack (IDA) that exploits observable DER-VPP communications to estimate voltage sensitivity and strategically degrade voltage regulation without access to grid topology or high attack intensity. Two estimation methods are developed for daily operations and voltage support data, enabling attackers to cluster and prioritize DER targets. Evaluated on the IEEE 123-bus feeder with 28 DERs, results show that as little as 12 h of operational data or ten voltage support events can yield over 95% cosine similarity with true voltage sensitivity, enabling impactful DER selection and disrupting voltage control. Mitigation strategies are also proposed and validated, including dynamic switching between centralized and decentralized control modes to maintain grid stability under DoS conditions.","PeriodicalId":13301,"journal":{"name":"IEEE Transactions on Industrial Informatics","volume":"22 10","pages":"9594-9605"},"PeriodicalIF":9.8,"publicationDate":"2026-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148372471","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Novel Identification Method for Particles in Transformer Oil Based on Microflow Imaging Technology and Machine Learning 基于微流成像技术和机器学习的变压器油颗粒识别新方法

IF 9.8 1区计算机科学

IEEE Transactions on Industrial Informatics Pub Date : 2026-10-01 Epub Date: 2026-07-31 DOI: 10.1109/TII.2026.3699570

Chen Chen;Ruijin Liao;Jianyi Wang;Lijun Yang;Zhengyu Chen;Shangquan Feng;Xiang Wang

{"title":"Novel Identification Method for Particles in Transformer Oil Based on Microflow Imaging Technology and Machine Learning","authors":"Chen Chen;Ruijin Liao;Jianyi Wang;Lijun Yang;Zhengyu Chen;Shangquan Feng;Xiang Wang","doi":"10.1109/TII.2026.3699570","DOIUrl":"10.1109/TII.2026.3699570","url":null,"abstract":"During the operation of transformer, insulation oil is inevitably contaminated by particles, and different particles may influence transformer performance to varying degrees. The accurate and timely identification of the particle types in oil is important for assessing the operational risks faced by equipment in service. In this article, microflow imaging (MFI) technology was employed to analyze oil containing typical contaminants, such as bubbles, carbon, copper, and cellulose particles, thereby establishing a dataset of representative morphological characteristics of particles in oil. To address feature mixing during particle characterization, the local outlier factor (LOF) method was applied to filter abnormal feature points in the raw dataset and reconstruct a new dataset, thereby reducing the influence of outliers on classification accuracy. A sparse oblique decision tree (SODT) model was combined with a bagging-based voting mechanism to classify the reconstructed dataset. The proposed particle classification method, which integrates MFI with LOF-SODT-Bagging, achieved an average recognition accuracy of 95.78% for typical particles in transformer oil. Compared with that of the SODT model, the average accuracy of the proposed method was increased by 10.39%. Furthermore, validation with feature datasets obtained from parallel experiments indicated that the proposed method maintained an accuracy of 77.99% even under bubble interference, confirming its robustness and general applicability. Finally, the method was applied to classify particles in the oil of 15 transformers in service; the recognition results were consistent with the actual transformer operating conditions. Thus, the proposed approach effectively provides particle feature information from transformer oil and accurately identifies different particle types.","PeriodicalId":13301,"journal":{"name":"IEEE Transactions on Industrial Informatics","volume":"22 10","pages":"9361-9372"},"PeriodicalIF":9.8,"publicationDate":"2026-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148633677","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

QuantEIT: Ultra-Lightweight Quantum-Inspired Inference for Electrical Impedance Tomography QuantEIT：电阻抗断层成像的超轻量级量子启发推理

IF 12.3 1区计算机科学

IEEE Transactions on Industrial Informatics Pub Date : 2026-10-01 DOI: 10.1109/tii.2026.3731100

Hao Fang, Sihao Teng, Hao Yu, Zhe Liu, Siyi Yuan, Huaiwu He, Yunjie Yang

引用次数: 0