Journal of Industrial Information Integration最新文献_第3页

Cognitive collaboration: A design methodology for future-oriented intelligent industrial systems 认知协作：面向未来的智能工业系统的设计方法

IF 10.4 1区计算机科学 Q1 COMPUTER SCIENCE, INTERDISCIPLINARY APPLICATIONS

Journal of Industrial Information Integration

Pub Date : 2026-01-20 DOI: 10.1016/j.jii.2026.101071

Hongwei Jiang , Jiapeng You , Zhiyang Chen , Xinguo Ming , Poly Z.H. Sun

Industrial systems are undergoing a paradigm shift from digitalization to intelligentization. However, current practices often suffer from systemic fragmentation and a lack of cognitive capability, creating a gap between high-level intelligence and physical execution. Existing research either focuses on specific algorithms or remains at the level of abstract concepts like cognitive digital twins, lacking a systematic design methodology to bridge this gap. To address this, this paper proposes cognitive collaboration, a novel design methodology guided by the dual-process theory of cognitive science. This paper redefines industrial systems as proactive partners possessing four synergistic capabilities: Intent Insight, Cognitive Evolution, Autonomous Planning, and Embodied Reconfiguration. Distinct from traditional approaches, a rigorous quantitative framework is introduced, including a dual-dimensional design space and a capability priority index (CPI) calibrated by expert-weighted analytic hierarchy process (AHP). Furthermore, to resolve the conflict between generative artificial intelligence (AI)’s creativity and industrial safety, a generator–verifier dual-system architecture is constructed to constrain stochastic outputs within verifiable safety envelopes. The methodology is systematically validated through the reconfiguration of a production line operations and maintenance (O&M) system and a comparative analysis of a complex ship design case. Preliminary quantitative results from a prototype experiment further confirm that the proposed framework effectively bridges the gap between generative flexibility and industrial reliability, offering a verifiable, operable pathway for building the next generation of intelligent industrial systems.

工业系统正经历着从数字化向智能化的范式转变。然而，目前的实践经常受到系统分裂和缺乏认知能力的影响，在高水平的智力和实际执行之间造成了差距。现有的研究要么侧重于具体的算法，要么停留在抽象概念的层面，如认知数字双胞胎，缺乏系统的设计方法来弥补这一差距。为了解决这一问题，本文提出了一种以认知科学的双过程理论为指导的新型设计方法——认知协作。本文将工业系统重新定义为具有四种协同能力的主动合作伙伴：意图洞察、认知进化、自主规划和具体化重构。与传统方法不同，引入了一个严格的定量框架，包括一个二维设计空间和一个由专家加权层次分析法（AHP）校准的能力优先指数（CPI）。此外，为了解决生成式人工智能（AI）的创造力与工业安全之间的冲突，构建了一个生成器-验证者双系统架构，将随机输出约束在可验证的安全信封内。通过生产线操作和维护（O&；M）系统的重新配置和复杂船舶设计案例的比较分析，系统地验证了该方法。原型实验的初步定量结果进一步证实，所提出的框架有效地弥合了生成灵活性和工业可靠性之间的差距，为构建下一代智能工业系统提供了可验证、可操作的途径。

{"title":"Cognitive collaboration: A design methodology for future-oriented intelligent industrial systems","authors":"Hongwei Jiang , Jiapeng You , Zhiyang Chen , Xinguo Ming , Poly Z.H. Sun","doi":"10.1016/j.jii.2026.101071","DOIUrl":"10.1016/j.jii.2026.101071","url":null,"abstract":"<div><div>Industrial systems are undergoing a paradigm shift from digitalization to intelligentization. However, current practices often suffer from systemic fragmentation and a lack of cognitive capability, creating a gap between high-level intelligence and physical execution. Existing research either focuses on specific algorithms or remains at the level of abstract concepts like cognitive digital twins, lacking a systematic design methodology to bridge this gap. To address this, this paper proposes cognitive collaboration, a novel design methodology guided by the dual-process theory of cognitive science. This paper redefines industrial systems as proactive partners possessing four synergistic capabilities: Intent Insight, Cognitive Evolution, Autonomous Planning, and Embodied Reconfiguration. Distinct from traditional approaches, a rigorous quantitative framework is introduced, including a dual-dimensional design space and a capability priority index (CPI) calibrated by expert-weighted analytic hierarchy process (AHP). Furthermore, to resolve the conflict between generative artificial intelligence (AI)’s creativity and industrial safety, a generator–verifier dual-system architecture is constructed to constrain stochastic outputs within verifiable safety envelopes. The methodology is systematically validated through the reconfiguration of a production line operations and maintenance (O&M) system and a comparative analysis of a complex ship design case. Preliminary quantitative results from a prototype experiment further confirm that the proposed framework effectively bridges the gap between generative flexibility and industrial reliability, offering a verifiable, operable pathway for building the next generation of intelligent industrial systems.</div></div>","PeriodicalId":55975,"journal":{"name":"Journal of Industrial Information Integration","volume":"50 ","pages":"Article 101071"},"PeriodicalIF":10.4,"publicationDate":"2026-01-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146014987","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

Semi-supervised cross-domain fault diagnosis via contrastive pre-training and annotation-efficient alignment strategy 基于对比预训练和高效标注对齐策略的半监督跨域故障诊断

IF 10.4 1区计算机科学 Q1 COMPUTER SCIENCE, INTERDISCIPLINARY APPLICATIONS

Journal of Industrial Information Integration

Pub Date : 2026-01-20 DOI: 10.1016/j.jii.2026.101076

Lechang Yang , Xianghao Zhang , Feng Zhu , Zhe Wang , Xiaoge Zhang

Cross-domain fault diagnosis is a critical task in predictive maintenance for fleet management. However, existing transfer learning and distribution matching methods are often impractical in real-world scenarios, especially under few-shot conditions, where their diagnostic performance cannot be consistently guaranteed. To address this issue, this study proposes a novel semi-supervised framework for cross-domain fault diagnosis, based on the pre-training–fine-tuning paradigm. In our approach, self-supervised contrastive learning is employed for centralized multi-domain pre-training, followed by supervised fine-tuning and contrastive re-learning to achieve robust model alignment across different machines and operating conditions. To effectively capture temporal dependencies in structured sensor data and improve sample efficiency, we incorporate a time-series contrastive learning method, Time-Series Representation Learning via Temporal and Contextual Contrasting (TS-TCC), as the core component of the pre-training stage. Furthermore, we introduce a two-stage sample selection strategy that enables annotation-efficient model alignment. This design ensures consistently reliable diagnostic performance on the target domain while minimizing labeling effort. We validate our framework using two benchmark datasets: the Prognostics and Health Management Data Challenge 2022 dataset for Hydraulic Rock Drill (HRD) fault classification and the Paderborn University (PU) Bearing dataset. Experimental results demonstrate substantial improvements over existing methods. For the HRD dataset, our approach achieves 96.62% accuracy under Condition 1, representing a 45.79% improvement over the best baseline method. Similarly, for the PU Bearing dataset, we achieve 90.93% accuracy under Condition 1, exceeding the best baseline by 62.62%. Comparable performances are observed across other experimental conditions in both datasets.

跨域故障诊断是机队管理预测性维护中的一项重要任务。然而，现有的迁移学习和分布匹配方法在现实场景中往往是不切实际的，特别是在少射条件下，它们的诊断性能不能得到一致的保证。为了解决这一问题，本研究提出了一种基于预训练-微调范式的跨域故障诊断半监督框架。在我们的方法中，采用自监督对比学习进行集中的多域预训练，然后进行监督微调和对比再学习，以实现跨不同机器和操作条件的鲁棒模型对齐。为了有效捕获结构化传感器数据中的时间依赖性并提高样本效率，我们采用了一种时间序列对比学习方法，即通过时间和上下文对比的时间序列表示学习（TS-TCC），作为预训练阶段的核心组成部分。此外，我们还引入了一种两阶段的样本选择策略，该策略可以实现高效注释的模型对齐。这种设计确保了在目标域上始终可靠的诊断性能，同时最大限度地减少了标记工作。我们使用两个基准数据集验证了我们的框架：用于水力岩石钻机（HRD）故障分类的预测和健康管理数据挑战2022数据集和帕德博恩大学（PU）轴承数据集。实验结果表明，与现有方法相比，有了很大的改进。对于HRD数据集，我们的方法在条件1下达到96.62%的准确率，比最佳基线方法提高了45.79%。同样，对于PU轴承数据集，我们在条件1下实现了90.93%的精度，比最佳基线高出62.62%。在两个数据集的其他实验条件下观察到可比较的性能。

{"title":"Semi-supervised cross-domain fault diagnosis via contrastive pre-training and annotation-efficient alignment strategy","authors":"Lechang Yang , Xianghao Zhang , Feng Zhu , Zhe Wang , Xiaoge Zhang","doi":"10.1016/j.jii.2026.101076","DOIUrl":"10.1016/j.jii.2026.101076","url":null,"abstract":"<div><div>Cross-domain fault diagnosis is a critical task in predictive maintenance for fleet management. However, existing transfer learning and distribution matching methods are often impractical in real-world scenarios, especially under few-shot conditions, where their diagnostic performance cannot be consistently guaranteed. To address this issue, this study proposes a novel semi-supervised framework for cross-domain fault diagnosis, based on the pre-training–fine-tuning paradigm. In our approach, self-supervised contrastive learning is employed for centralized multi-domain pre-training, followed by supervised fine-tuning and contrastive re-learning to achieve robust model alignment across different machines and operating conditions. To effectively capture temporal dependencies in structured sensor data and improve sample efficiency, we incorporate a time-series contrastive learning method, Time-Series Representation Learning via Temporal and Contextual Contrasting (TS-TCC), as the core component of the pre-training stage. Furthermore, we introduce a two-stage sample selection strategy that enables annotation-efficient model alignment. This design ensures consistently reliable diagnostic performance on the target domain while minimizing labeling effort. We validate our framework using two benchmark datasets: the Prognostics and Health Management Data Challenge 2022 dataset for Hydraulic Rock Drill (HRD) fault classification and the Paderborn University (PU) Bearing dataset. Experimental results demonstrate substantial improvements over existing methods. For the HRD dataset, our approach achieves 96.62% accuracy under Condition 1, representing a 45.79% improvement over the best baseline method. Similarly, for the PU Bearing dataset, we achieve 90.93% accuracy under Condition 1, exceeding the best baseline by 62.62%. Comparable performances are observed across other experimental conditions in both datasets.</div></div>","PeriodicalId":55975,"journal":{"name":"Journal of Industrial Information Integration","volume":"50 ","pages":"Article 101076"},"PeriodicalIF":10.4,"publicationDate":"2026-01-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146014988","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

Ensemble-based ship weather multi-objective route optimization 基于集成的船舶天气多目标航路优化

IF 10.4 1区计算机科学 Q1 COMPUTER SCIENCE, INTERDISCIPLINARY APPLICATIONS

Journal of Industrial Information Integration

Pub Date : 2026-01-17 DOI: 10.1016/j.jii.2026.101075

Kumars Mahmoodi , Jari Böling , Roberto Vettor

Many traditional and state-of-the-art ship routing methods rely on single-objective formulations, deterministic weather inputs, or fixed operational assumptions, which may lead to suboptimal or impractical routing decisions under realistic and uncertain marine environments. This study presents an ensemble-based multi-objective optimization framework for ship route planning under uncertain weather conditions. The framework integrates a neural network model, trained on real onboard ship performance data and tuned using Bayesian hyperparameter optimization, to predict fuel consumption based on ship speed and marine weather parameters. An ensemble of weather forecasts is assigned to route waypoints using a bootstrapping method, enabling the evaluation of multiple cost functions reflecting trade-offs between voyage time, fuel consumption, and safety. Four optimization objective strategies — ensemble mean, worst-case, risk-aware, and Conditional Value-at-Risk (CVaR) — are implemented within a Grey Wolf Optimizer (GWO) to derive optimal routes across various voyages. The results demonstrate notable variations in route performance based on the selected strategy. For example, the CVaR approach achieves a balance between robustness and efficiency, with voyage fuel consumption for the longest journey (Voyage 3) reaching 490,475 kg, while the worst-case strategy prioritizes risk-averse paths, resulting in the highest fuel usage at 505,308 kg. Conversely, the ensemble mean strategy offers the lowest average fuel consumption (474,078 kg) but may expose the voyage to higher uncertainty. Furthermore, the proposed GWO demonstrates high precision in schedule adherence, maintaining arrival time deviations within a 30-minute margin across all optimized voyages, thereby justifying its effectiveness in handling complex multi-objective constraints. The developed framework is applicable to real-time voyage optimization and can support ship operators in achieving fuel efficiency and safety under varying ocean conditions.

许多传统和最先进的船舶路由方法依赖于单目标公式、确定性天气输入或固定的操作假设，这可能导致在现实和不确定的海洋环境下的次优或不切实际的路由决策。针对不确定天气条件下船舶航路规划问题，提出了一种基于集成的多目标优化框架。该框架集成了一个神经网络模型，该模型接受了真实船载性能数据的训练，并使用贝叶斯超参数优化进行了调整，可以根据船速和海洋天气参数预测燃油消耗。使用自举方法将天气预报集合分配给航线航路点，从而能够评估反映航行时间、燃料消耗和安全之间权衡的多个成本函数。在灰狼优化器（GWO）中实现了四种优化目标策略——集合均值、最坏情况、风险意识和条件风险值（CVaR），以获得跨不同航程的最佳路线。结果表明，根据所选择的策略，路由性能会发生显著变化。例如，CVaR方法实现了鲁棒性和效率之间的平衡，最长航程（航程3）的航次燃油消耗达到490,475 kg，而最坏情况策略优先考虑风险规避路径，导致最高的燃油消耗为505,308 kg。相反，整体平均策略提供最低的平均燃料消耗（474,078公斤），但可能使航行面临更高的不确定性。此外，所提出的GWO在计划遵守方面具有很高的精度，在所有优化的航程中保持30分钟的到达时间偏差，从而证明了其在处理复杂的多目标约束方面的有效性。所开发的框架适用于实时航行优化，可以支持船舶运营商在不同的海洋条件下实现燃油效率和安全性。

{"title":"Ensemble-based ship weather multi-objective route optimization","authors":"Kumars Mahmoodi , Jari Böling , Roberto Vettor","doi":"10.1016/j.jii.2026.101075","DOIUrl":"10.1016/j.jii.2026.101075","url":null,"abstract":"<div><div>Many traditional and state-of-the-art ship routing methods rely on single-objective formulations, deterministic weather inputs, or fixed operational assumptions, which may lead to suboptimal or impractical routing decisions under realistic and uncertain marine environments. This study presents an ensemble-based multi-objective optimization framework for ship route planning under uncertain weather conditions. The framework integrates a neural network model, trained on real onboard ship performance data and tuned using Bayesian hyperparameter optimization, to predict fuel consumption based on ship speed and marine weather parameters. An ensemble of weather forecasts is assigned to route waypoints using a bootstrapping method, enabling the evaluation of multiple cost functions reflecting trade-offs between voyage time, fuel consumption, and safety. Four optimization objective strategies — ensemble mean, worst-case, risk-aware, and Conditional Value-at-Risk (CVaR) — are implemented within a Grey Wolf Optimizer (GWO) to derive optimal routes across various voyages. The results demonstrate notable variations in route performance based on the selected strategy. For example, the CVaR approach achieves a balance between robustness and efficiency, with voyage fuel consumption for the longest journey (Voyage 3) reaching 490,475 kg, while the worst-case strategy prioritizes risk-averse paths, resulting in the highest fuel usage at 505,308 kg. Conversely, the ensemble mean strategy offers the lowest average fuel consumption (474,078 kg) but may expose the voyage to higher uncertainty. Furthermore, the proposed GWO demonstrates high precision in schedule adherence, maintaining arrival time deviations within a 30-minute margin across all optimized voyages, thereby justifying its effectiveness in handling complex multi-objective constraints. The developed framework is applicable to real-time voyage optimization and can support ship operators in achieving fuel efficiency and safety under varying ocean conditions.</div></div>","PeriodicalId":55975,"journal":{"name":"Journal of Industrial Information Integration","volume":"50 ","pages":"Article 101075"},"PeriodicalIF":10.4,"publicationDate":"2026-01-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145995691","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

Compound fault diagnosis of diesel engines by combining CDCGAN and multistage transfer learning 结合CDCGAN和多级迁移学习的柴油机复合故障诊断

IF 10.4 1区计算机科学 Q1 COMPUTER SCIENCE, INTERDISCIPLINARY APPLICATIONS

Journal of Industrial Information Integration

Pub Date : 2026-01-17 DOI: 10.1016/j.jii.2026.101073

Xin Zhao , Wenjie Liu , Jianhua Shi , Yangyu Zhao , Zikang Li

Long-term operation of mining diesel engines with high power density within a complex working environment of open-pit mines causes them to suffer from compound faults and difficult diagnosis. Therefore, a compound fault diagnosis method that combines a Conditional Deep Convolutional Generative Adversarial Network (CDCGAN) and Multistage Transfer Learning (MTL) is proposed in this paper. This method overcomes different issues arising in compound fault detection, such as sample scarcity, insufficient single signal characterization, and low distinguishability of one-dimensional vibration signal features. The Continuous Wavelet Transform (CWT) and CDCGAN are introduced to process the one-dimensional raw data. An improved Transfer Learning (TL) algorithm based on an MTL strategy is also proposed by incorporating pretraining, fine-tuning, and feature fusion techniques. A ResNetCBAM model integrating the Residual Neural Network (ResNet) with the Convolutional Block Attention Module (CBAM) is trained based on the algorithm. Validation experiments are performed on real diesel engine fault data to evaluate the method’s performance. It is shown that the proposed method’s accuracy improves by 13.75%, 8.75%, 6.37%, and 4.58%, compared with four baseline methods including a one-dimensional convolutional neural network (1D-CNN) with raw one-dimensional vibration signals, a two-dimensional convolutional neural network (2D-CNN) with time-frequency images obtained via the CWT, ResNetCBAM with CDCGAN-augmented data, and ResNetCBAM with conventional TL, respectively. The proposed method achieves 100% diagnostic accuracy on the test data, thus establishing a reliable theoretical basis for the intelligent compound fault diagnosis in diesel engines.

高功率密度矿用柴油机长期在复杂的露天矿工作环境中运行，导致其故障复杂，诊断困难。为此，本文提出了一种结合条件深度卷积生成对抗网络（CDCGAN）和多阶段迁移学习（MTL）的复合故障诊断方法。该方法克服了复合故障检测中存在的样本稀缺性、单信号表征不足、一维振动信号特征可辨性低等问题。引入连续小波变换（CWT）和CDCGAN对一维原始数据进行处理。基于迁移学习策略，结合预训练、微调和特征融合技术，提出了一种改进的迁移学习算法。在此基础上训练了残差神经网络（ResNet）与卷积块注意模块（CBAM）相结合的ResNetCBAM模型。在柴油机实际故障数据上进行了验证实验，以评价该方法的性能。结果表明，与基于原始一维振动信号的一维卷积神经网络（1D-CNN）、基于CWT的时频图像的二维卷积神经网络（2D-CNN）、基于cdcgan增强数据的ResNetCBAM和基于传统TL的ResNetCBAM四种基线方法相比，该方法的准确率分别提高了13.75%、8.75%、6.37%和4.58%。该方法对试验数据的诊断准确率达到100%，为柴油机智能复合故障诊断奠定了可靠的理论基础。

{"title":"Compound fault diagnosis of diesel engines by combining CDCGAN and multistage transfer learning","authors":"Xin Zhao , Wenjie Liu , Jianhua Shi , Yangyu Zhao , Zikang Li","doi":"10.1016/j.jii.2026.101073","DOIUrl":"10.1016/j.jii.2026.101073","url":null,"abstract":"<div><div>Long-term operation of mining diesel engines with high power density within a complex working environment of open-pit mines causes them to suffer from compound faults and difficult diagnosis. Therefore, a compound fault diagnosis method that combines a Conditional Deep Convolutional Generative Adversarial Network (CDCGAN) and Multistage Transfer Learning (MTL) is proposed in this paper. This method overcomes different issues arising in compound fault detection, such as sample scarcity, insufficient single signal characterization, and low distinguishability of one-dimensional vibration signal features. The Continuous Wavelet Transform (CWT) and CDCGAN are introduced to process the one-dimensional raw data. An improved Transfer Learning (TL) algorithm based on an MTL strategy is also proposed by incorporating pretraining, fine-tuning, and feature fusion techniques. A ResNetCBAM model integrating the Residual Neural Network (ResNet) with the Convolutional Block Attention Module (CBAM) is trained based on the algorithm. Validation experiments are performed on real diesel engine fault data to evaluate the method’s performance. It is shown that the proposed method’s accuracy improves by 13.75%, 8.75%, 6.37%, and 4.58%, compared with four baseline methods including a one-dimensional convolutional neural network (1D-CNN) with raw one-dimensional vibration signals, a two-dimensional convolutional neural network (2D-CNN) with time-frequency images obtained via the CWT, ResNetCBAM with CDCGAN-augmented data, and ResNetCBAM with conventional TL, respectively. The proposed method achieves 100% diagnostic accuracy on the test data, thus establishing a reliable theoretical basis for the intelligent compound fault diagnosis in diesel engines.</div></div>","PeriodicalId":55975,"journal":{"name":"Journal of Industrial Information Integration","volume":"50 ","pages":"Article 101073"},"PeriodicalIF":10.4,"publicationDate":"2026-01-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145995692","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

Optimization of connection models in digital twin systems: Efficient merging and assembly strategy for enhanced scalability and resource optimization 数字孪生系统中连接模型的优化：增强可扩展性和资源优化的有效合并和装配策略

IF 10.4 1区计算机科学 Q1 COMPUTER SCIENCE, INTERDISCIPLINARY APPLICATIONS

Journal of Industrial Information Integration

Pub Date : 2026-01-15 DOI: 10.1016/j.jii.2026.101070

Chongxin Wang , Xiaojun Liu , Feixiang Wang , Fengyi Feng , Lv Feng

This paper presents an innovative approach to optimizing connection models in complex digital twin systems. Traditional digital twin systems are often hindered by inefficient connection models, resulting in excessive thread and memory consumption and conflicts during functional expansion. To address these challenges, we propose a Virtual Sensor-based dual strategy combining merging and assembly techniques within the framework of the five-dimensional digital twin model. The merging strategy groups and merges similar models to eliminate redundancies, reducing model complexity and resource consumption. The assembly strategy integrates multiple sub-connection models into a more complex, scalable model. This ensures dynamic adjustment and synchronization of information across various system dimensions. A case study in a packaging production line demonstrates an over 40% reduction in connection models. Due to the deployed stateless singleton architecture, this structural simplification directly translates into a proportional decrease in resource consumption, specifically reducing active thread occupation by approximately 40% and substantially lowering memory usage. These results confirm the proposed method's effectiveness in enhancing scalability and resource efficiency, highlighting its significant industrial applicability.

本文提出了一种优化复杂数字孪生系统连接模型的创新方法。传统的数字孪生系统经常受到低效的连接模型的阻碍，导致线程和内存消耗过多，在功能扩展过程中产生冲突。为了解决这些挑战，我们提出了一种基于虚拟传感器的双策略，在五维数字孪生模型框架内结合合并和装配技术。合并策略对相似的模型进行分组和合并，以消除冗余，降低模型复杂性和资源消耗。装配策略将多个子连接模型集成到一个更复杂、可扩展的模型中。这确保了跨不同系统维度的动态调整和信息同步。一个包装生产线的案例研究表明，连接模型减少了40%以上。由于部署了无状态单例架构，这种结构简化直接转化为资源消耗的比例减少，特别是将活动线程占用减少了大约40%，并大幅降低了内存使用。这些结果证实了该方法在提高可扩展性和资源效率方面的有效性，突出了其显著的工业适用性。

{"title":"Optimization of connection models in digital twin systems: Efficient merging and assembly strategy for enhanced scalability and resource optimization","authors":"Chongxin Wang , Xiaojun Liu , Feixiang Wang , Fengyi Feng , Lv Feng","doi":"10.1016/j.jii.2026.101070","DOIUrl":"10.1016/j.jii.2026.101070","url":null,"abstract":"<div><div>This paper presents an innovative approach to optimizing connection models in complex digital twin systems. Traditional digital twin systems are often hindered by inefficient connection models, resulting in excessive thread and memory consumption and conflicts during functional expansion. To address these challenges, we propose a Virtual Sensor-based dual strategy combining merging and assembly techniques within the framework of the five-dimensional digital twin model. The merging strategy groups and merges similar models to eliminate redundancies, reducing model complexity and resource consumption. The assembly strategy integrates multiple sub-connection models into a more complex, scalable model. This ensures dynamic adjustment and synchronization of information across various system dimensions. A case study in a packaging production line demonstrates an over 40% reduction in connection models. Due to the deployed stateless singleton architecture, this structural simplification directly translates into a proportional decrease in resource consumption, specifically reducing active thread occupation by approximately 40% and substantially lowering memory usage. These results confirm the proposed method's effectiveness in enhancing scalability and resource efficiency, highlighting its significant industrial applicability.</div></div>","PeriodicalId":55975,"journal":{"name":"Journal of Industrial Information Integration","volume":"50 ","pages":"Article 101070"},"PeriodicalIF":10.4,"publicationDate":"2026-01-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145995696","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

PV-EIR: An embodied intelligent robot for solar panel cleaning in desert regions PV-EIR：沙漠地区太阳能电池板清洁的具身智能机器人

IF 10.4 1区计算机科学 Q1 COMPUTER SCIENCE, INTERDISCIPLINARY APPLICATIONS

Journal of Industrial Information Integration

Pub Date : 2026-01-13 DOI: 10.1016/j.jii.2026.101069

Xinhao Wang , Lin Yang , Daqiang Zhang

To address the cleaning challenges faced by robots in the extreme conditions of desert photovoltaic power plants, this study proposes a photovoltaic embodied intelligence (EI) robot based on bionic brain-cerebellum coordination control. First, domain knowledge integration is used for multi-objective task planning and interpretable strategy generation, and a dedicated decision-making brain model (PV-LLM) is constructed for desert PV scenarios. Then, a PV-Cerebellum module is designed through simulation-reality dataset learning, thereby ensuring efficient mapping and robust control from abstract tasks to specific skill library invocations. Finally, the two core skills of the proposed robot, real-time obstacle avoidance navigation and panel alignment cleaning, are validated via simulation tests in the Gazebo simulation environment under desert conditions. Experimental results demonstrate that the proposed robotic system can achieve obstacle-avoidance response times as low as 0.36 s and cleaning coverage rates up to 96.67 %, validating its engineering feasibility. Overall, this work provides a new practical solution for reliable photovoltaic panel cleaning in harsh desert environments.

针对沙漠光伏电站极端条件下机器人的清洁挑战，本研究提出了一种基于仿生脑-小脑协调控制的光伏具身智能（EI）机器人。首先，将领域知识集成用于多目标任务规划和可解释策略生成，构建了沙漠光伏场景专用决策脑模型（PV- llm）；然后，通过模拟-现实数据集学习设计pv -小脑模块，从而确保从抽象任务到特定技能库调用的高效映射和鲁棒控制。最后，通过Gazebo模拟环境在沙漠条件下的仿真测试，验证了机器人的两项核心技能——实时避障导航和面板对齐清理。实验结果表明，该机器人系统的避障响应时间低至0.36 s，清洁覆盖率高达96.67%，验证了其工程可行性。总的来说，这项工作为恶劣沙漠环境下光伏板的可靠清洁提供了一种新的实用解决方案。

{"title":"PV-EIR: An embodied intelligent robot for solar panel cleaning in desert regions","authors":"Xinhao Wang , Lin Yang , Daqiang Zhang","doi":"10.1016/j.jii.2026.101069","DOIUrl":"10.1016/j.jii.2026.101069","url":null,"abstract":"<div><div>To address the cleaning challenges faced by robots in the extreme conditions of desert photovoltaic power plants, this study proposes a photovoltaic embodied intelligence (EI) robot based on bionic brain-cerebellum coordination control. First, domain knowledge integration is used for multi-objective task planning and interpretable strategy generation, and a dedicated decision-making brain model (PV-LLM) is constructed for desert PV scenarios. Then, a PV-Cerebellum module is designed through simulation-reality dataset learning, thereby ensuring efficient mapping and robust control from abstract tasks to specific skill library invocations. Finally, the two core skills of the proposed robot, real-time obstacle avoidance navigation and panel alignment cleaning, are validated via simulation tests in the Gazebo simulation environment under desert conditions. Experimental results demonstrate that the proposed robotic system can achieve obstacle-avoidance response times as low as 0.36 s and cleaning coverage rates up to 96.67 %, validating its engineering feasibility. Overall, this work provides a new practical solution for reliable photovoltaic panel cleaning in harsh desert environments.</div></div>","PeriodicalId":55975,"journal":{"name":"Journal of Industrial Information Integration","volume":"50 ","pages":"Article 101069"},"PeriodicalIF":10.4,"publicationDate":"2026-01-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145962597","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

Multi-dimensional framework for assessing digital twin maturity in construction machinery 工程机械数字孪生成熟度评估的多维框架

IF 10.4 1区计算机科学 Q1 COMPUTER SCIENCE, INTERDISCIPLINARY APPLICATIONS

Journal of Industrial Information Integration

Pub Date : 2026-01-13 DOI: 10.1016/j.jii.2026.101067

Ruibo Hu, Wenting Gong, Ke Chen, Hanbin Luo

Smart construction machinery enabled by digital twin (DT) technology has significant potential to enhance construction safety and efficiency. However, the lack of a dedicated maturity assessment model for construction machinery DT (CMDT) reveals a critical gap in existing research. This study proposes a structured framework for assessing CMDT maturity, designed to evaluate the readiness and developmental stage of DT implementations in construction machinery. The framework was developed based on a comprehensive literature review and expert interviews, yielding a maturity assessment model comprising five dimensions and 24 indicators across five maturity levels. The model adopts a two-stage assessment approach that integrates expert competency-based grouping and weighting with an improved bi-objective optimization model. To enhance the robustness of expert opinion aggregation, a penalty-weight mechanism is embedded in the objective function, effectively balancing consensus and confidence. The proposed framework was validated through a real-world case study involving an automated construction system (ACS). The results demonstrate the capability of the framework to identify CMDT maturity levels and inform improvement pathways. Overall, this study provides an evidence-based tool to accelerate DT adoption in the construction sector.

数字孪生（DT）技术支持的智能施工机械在提高施工安全和效率方面具有巨大潜力。然而，缺乏一个专门的工程机械成熟度评估模型（CMDT）显示了现有研究的一个重要空白。本研究提出了一个评估CMDT成熟度的结构化框架，旨在评估工程机械中DT实施的准备情况和发展阶段。该框架是在综合文献综述和专家访谈的基础上开发的，产生了一个成熟度评估模型，该模型包括五个成熟度级别的五个维度和24个指标。该模型采用两阶段评价方法，将基于专家能力的分组和加权与改进的双目标优化模型相结合。为了增强专家意见聚合的鲁棒性，在目标函数中嵌入了惩罚权重机制，有效地平衡了共识和置信度。提出的框架通过涉及自动化施工系统（ACS）的实际案例研究进行了验证。结果证明了框架识别CMDT成熟度级别和告知改进途径的能力。总体而言，本研究提供了一个以证据为基础的工具，以加速建筑行业对DT的采用。

{"title":"Multi-dimensional framework for assessing digital twin maturity in construction machinery","authors":"Ruibo Hu, Wenting Gong, Ke Chen, Hanbin Luo","doi":"10.1016/j.jii.2026.101067","DOIUrl":"10.1016/j.jii.2026.101067","url":null,"abstract":"<div><div>Smart construction machinery enabled by digital twin (DT) technology has significant potential to enhance construction safety and efficiency. However, the lack of a dedicated maturity assessment model for construction machinery DT (CMDT) reveals a critical gap in existing research. This study proposes a structured framework for assessing CMDT maturity, designed to evaluate the readiness and developmental stage of DT implementations in construction machinery. The framework was developed based on a comprehensive literature review and expert interviews, yielding a maturity assessment model comprising five dimensions and 24 indicators across five maturity levels. The model adopts a two-stage assessment approach that integrates expert competency-based grouping and weighting with an improved bi-objective optimization model. To enhance the robustness of expert opinion aggregation, a penalty-weight mechanism is embedded in the objective function, effectively balancing consensus and confidence. The proposed framework was validated through a real-world case study involving an automated construction system (ACS). The results demonstrate the capability of the framework to identify CMDT maturity levels and inform improvement pathways. Overall, this study provides an evidence-based tool to accelerate DT adoption in the construction sector.</div></div>","PeriodicalId":55975,"journal":{"name":"Journal of Industrial Information Integration","volume":"50 ","pages":"Article 101067"},"PeriodicalIF":10.4,"publicationDate":"2026-01-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145962599","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

Implementing TinyML in Internet of Things devices: A systematic literature review 在物联网设备中实现TinyML：系统的文献综述

IF 10.4 1区计算机科学 Q1 COMPUTER SCIENCE, INTERDISCIPLINARY APPLICATIONS

Journal of Industrial Information Integration

Pub Date : 2026-01-09 DOI: 10.1016/j.jii.2026.101065

Andrés Felipe Solis Pino , Daniel Steven Moran Pizarro , Pablo H. Ruiz , Vanessa Agredo-Delgado , Cesar Alberto Collazos , Fernando Moreira

The Internet of Things is at the heart of society and is experiencing rapid expansion. Its integration with Artificial Intelligence and Machine Learning has led to the emergence of Tiny Machine Learning (TinyML), which enables data processing directly on the device, improving efficiency, reducing latency, and increasing data privacy. Despite the growing relevance of TinyML in the Internet of Things, there is a lack of systematic literature reviews providing a holistic understanding of its implementation, advances, and challenges, which hinders a clear understanding of the available empirical evidence and best practices. To bridge this gap, this study presents a systematic literature review, adhering to the PRISMA protocol and employing a multi-database search strategy, identifying 114 primary studies. The review reveals that TinyML is consolidating as a transformative paradigm for the Internet of Things, experiencing significant research growth since 2020. Applications are diverse, with healthcare and environmental monitoring being the most notable examples. Deep learning models, particularly convolutional neural networks, are frequently employed in this context. The main challenges identified include security vulnerabilities, the need to address ethical considerations like algorithmic bias, and hardware limitations related to memory and processing power. Ultimately, this review offers valuable insights into the current state and prospects of TinyML in the Internet of Things, providing a valuable resource for researchers, developers, and decision-makers in this rapidly evolving field.

物联网是社会的核心，正在迅速发展。它与人工智能和机器学习的结合导致了微型机器学习（TinyML）的出现，它可以直接在设备上处理数据，提高效率，减少延迟，并增加数据隐私。尽管TinyML在物联网中的相关性越来越大，但缺乏系统的文献综述，无法全面了解其实现、进展和挑战，这阻碍了对现有经验证据和最佳实践的清晰理解。为了弥补这一差距，本研究提出了一项系统的文献综述，遵循PRISMA协议并采用多数据库搜索策略，确定了114项主要研究。回顾显示，TinyML正在巩固其作为物联网变革范例的地位，自2020年以来，其研究成果显著增长。应用程序多种多样，医疗保健和环境监测是最显著的例子。深度学习模型，特别是卷积神经网络，在这种情况下经常被使用。确定的主要挑战包括安全漏洞，需要解决算法偏见等道德问题，以及与内存和处理能力相关的硬件限制。最后，本综述对TinyML在物联网领域的现状和前景提供了有价值的见解，为这个快速发展的领域的研究人员、开发人员和决策者提供了宝贵的资源。

{"title":"Implementing TinyML in Internet of Things devices: A systematic literature review","authors":"Andrés Felipe Solis Pino , Daniel Steven Moran Pizarro , Pablo H. Ruiz , Vanessa Agredo-Delgado , Cesar Alberto Collazos , Fernando Moreira","doi":"10.1016/j.jii.2026.101065","DOIUrl":"10.1016/j.jii.2026.101065","url":null,"abstract":"<div><div>The Internet of Things is at the heart of society and is experiencing rapid expansion. Its integration with Artificial Intelligence and Machine Learning has led to the emergence of Tiny Machine Learning (TinyML), which enables data processing directly on the device, improving efficiency, reducing latency, and increasing data privacy. Despite the growing relevance of TinyML in the Internet of Things, there is a lack of systematic literature reviews providing a holistic understanding of its implementation, advances, and challenges, which hinders a clear understanding of the available empirical evidence and best practices. To bridge this gap, this study presents a systematic literature review, adhering to the PRISMA protocol and employing a multi-database search strategy, identifying 114 primary studies. The review reveals that TinyML is consolidating as a transformative paradigm for the Internet of Things, experiencing significant research growth since 2020. Applications are diverse, with healthcare and environmental monitoring being the most notable examples. Deep learning models, particularly convolutional neural networks, are frequently employed in this context. The main challenges identified include security vulnerabilities, the need to address ethical considerations like algorithmic bias, and hardware limitations related to memory and processing power. Ultimately, this review offers valuable insights into the current state and prospects of TinyML in the Internet of Things, providing a valuable resource for researchers, developers, and decision-makers in this rapidly evolving field.</div></div>","PeriodicalId":55975,"journal":{"name":"Journal of Industrial Information Integration","volume":"50 ","pages":"Article 101065"},"PeriodicalIF":10.4,"publicationDate":"2026-01-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145957253","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

Multimodal data-enabled large model for machine fault diagnosis towards intelligent operation and maintenance 支持多模态数据的大型机器故障诊断模型，实现智能运维

IF 10.4 1区计算机科学 Q1 COMPUTER SCIENCE, INTERDISCIPLINARY APPLICATIONS

Journal of Industrial Information Integration

Pub Date : 2026-01-09 DOI: 10.1016/j.jii.2026.101061

Shupeng Yu , Xiang Li , Yaguo Lei , Bin Yang , Naipeng Li , Ke Feng

Large language models (LLMs) have been showing growing potential in the field of intelligent operation and maintenance, due to their strong capabilities in understanding and generating knowledge across data in multiple modalities. However, in operation and maintenance, time-series signals are among the most critical monitoring data, and their unique formats and high dimensionality pose significant challenges for direct application of LLMs. To address this limitation, we propose a novel large multimodal model for fault diagnosis (LMM-FD), which is a key problem in operation and maintenance. The proposed large multimodal framework effectively aligns time-series vibration signals with textual fault diagnosis knowledge, enabling interpretable and generalized fault diagnosis. The framework includes signal preprocessing, cross-modal alignment through a knowledge graph and graph neural networks, and automated generation of textual diagnostic reports. Extensive experiments on machinery condition monitoring datasets demonstrate that LMM-FD consistently outperforms existing baselines by leveraging multimodal data and constructed triplet-based knowledge graph. The proposed model obtains fairly high accuracy on multiple fault diagnosis scenarios, while achieving strong zero-shot generalization capabilities to unseen compound faults. Furthermore, by bridging numerical sensor data with textual knowledge, LMM-FD provides interpretable fault descriptions, highlighting its potential for practical industrial applications.

大型语言模型（llm）在智能运维领域显示出越来越大的潜力，因为它们具有强大的跨数据理解和生成知识的能力。然而，在运维中，时间序列信号是最关键的监测数据之一，其独特的格式和高维度给llm的直接应用带来了重大挑战。为了解决这一问题，本文提出了一种新的大型多模态故障诊断模型（LMM-FD），这是运维中的一个关键问题。提出的大型多模态框架有效地将时间序列振动信号与文本故障诊断知识对齐，实现可解释和广义故障诊断。该框架包括信号预处理，通过知识图和图形神经网络进行跨模态对齐，以及文本诊断报告的自动生成。在机械状态监测数据集上进行的大量实验表明，LMM-FD通过利用多模态数据和构建基于三元组的知识图，始终优于现有基线。该模型在多种故障诊断场景下具有较高的诊断准确率，同时对未见的复合故障具有较强的零点泛化能力。此外，通过将数字传感器数据与文本知识连接起来，LMM-FD提供可解释的故障描述，突出了其在实际工业应用中的潜力。

{"title":"Multimodal data-enabled large model for machine fault diagnosis towards intelligent operation and maintenance","authors":"Shupeng Yu , Xiang Li , Yaguo Lei , Bin Yang , Naipeng Li , Ke Feng","doi":"10.1016/j.jii.2026.101061","DOIUrl":"10.1016/j.jii.2026.101061","url":null,"abstract":"<div><div>Large language models (LLMs) have been showing growing potential in the field of intelligent operation and maintenance, due to their strong capabilities in understanding and generating knowledge across data in multiple modalities. However, in operation and maintenance, time-series signals are among the most critical monitoring data, and their unique formats and high dimensionality pose significant challenges for direct application of LLMs. To address this limitation, we propose a novel large multimodal model for fault diagnosis (LMM-FD), which is a key problem in operation and maintenance. The proposed large multimodal framework effectively aligns time-series vibration signals with textual fault diagnosis knowledge, enabling interpretable and generalized fault diagnosis. The framework includes signal preprocessing, cross-modal alignment through a knowledge graph and graph neural networks, and automated generation of textual diagnostic reports. Extensive experiments on machinery condition monitoring datasets demonstrate that LMM-FD consistently outperforms existing baselines by leveraging multimodal data and constructed triplet-based knowledge graph. The proposed model obtains fairly high accuracy on multiple fault diagnosis scenarios, while achieving strong zero-shot generalization capabilities to unseen compound faults. Furthermore, by bridging numerical sensor data with textual knowledge, LMM-FD provides interpretable fault descriptions, highlighting its potential for practical industrial applications.</div></div>","PeriodicalId":55975,"journal":{"name":"Journal of Industrial Information Integration","volume":"50 ","pages":"Article 101061"},"PeriodicalIF":10.4,"publicationDate":"2026-01-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145957254","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

AI for Information Integration and Processing in Digital Twins (AI4IIP-DT) 面向数字孪生信息集成与处理的人工智能（ai4ip - dt）

IF 15.7 1区计算机科学 Q1 COMPUTER SCIENCE, INTERDISCIPLINARY APPLICATIONS

Journal of Industrial Information Integration

Pub Date : 2026-01-09 DOI: 10.1016/j.jii.2026.101066

Prof. Hervé Panetto, Prof. Michele Dassisti, Prof. Qing Li, Dr. Yannick Naudet

引用次数: 0