Article
Thermodynamics
Yufeng Huang, Jun Tao, Gang Sun, Tengyun Wu, Liling Yu, Xinbin Zhao
Summary: In this paper, a novel digital twin approach based on deep multimodal information fusion is proposed, which integrates information from physical-based models and data-driven models for real-time fault detection and isolation. The experimental results show that this approach improves the accuracy of fault diagnosis and reduces the error of parameter prediction.
Review
Chemistry, Analytical
Narjes Davari, Bruno Veloso, Gustavo de Assis Costa, Pedro Mota Pereira, Rita P. Ribeiro, Joao Gama
Summary: This article surveys existing ML and DL techniques for handling PdM in the railway industry, highlighting the challenges and the importance of choosing the appropriate method for optimal performance.
Article
Computer Science, Hardware & Architecture
Jing Liang, Li Ma, Shan Liang, Hao Zhang, Zhonglin Zuo, Juan Dai
Summary: Leak detection in natural gas pipelines is a significant and ongoing problem in the oil and gas industry. The complexity of pipeline construction, lack of closed-form solutions, and the need for experienced personnel limit the creation of accurate physical models. Additionally, industrial automation faces challenges with data overload and limited information. This paper proposes a data-driven digital twin (DT) method as a new solution to address these challenges. The proposed approach includes a DT pipeline learning and updating scheme based on operational data and a DT-driven leak detection method that utilizes data interaction and fusion. The effectiveness of the approach is demonstrated through a simulated leak scenario in a real running natural gas pipeline.
COMPUTERS & ELECTRICAL ENGINEERING
(2023)
Article
Automation & Control Systems
Dong Liu, Yu Du, Wenjie Chai, ChangQi Lu, Ming Cong
Summary: This paper proposes a digital twin and data-driven quality prediction architecture, which realizes real-time quality prediction and appearance defect quality prediction in die-casting manufacturing through the virtual-real interaction of digital twin and the collaborative working mode. The real-time quality prediction accuracy of die-casting process is improved through data preprocessing and XGBoost-based learning method. The detection problem of complex appearance defects in castings is solved by a deep learning based neural network.
IEEE TRANSACTIONS ON INDUSTRIAL INFORMATICS
(2022)
Article
Automation & Control Systems
Cunsong Wang, Ningyun Lu, Yuehua Cheng, Bin Jiang
Summary: The proposed data-driven degradation prognostic strategy effectively addresses the challenges posed by unlabeled, unbalanced condition monitoring data and uncertainties of the prognostics process. The strategy fully considers uncertainties, and introduces multivariate health estimation and degradation prediction models to estimate the remaining useful life of aero-engines, showing effectiveness and feasibility through verification with NASA data.
IEEE TRANSACTIONS ON CYBERNETICS
(2021)
Article
Engineering, Industrial
Qiang Feng, Yue Zhang, Bo Sun, Xing Guo, Donming Fan, Yi Ren, Yanjie Song, Zili Wang
Summary: Digital twin technology is applied to smart manufacturing systems to provide valuable information for predictive maintenance, but there is a lack of research in this area. This paper proposes a predictive maintenance decision-making framework driven by digital twin, considering component dependencies and comprehensive maintenance resources. An optimal maintenance schedule can be obtained in real time, and an integer linear programming model is formulated to minimize maintenance costs while meeting production capacity. A matheuristics algorithm is introduced for various maintenance decision scenarios, and a case study is conducted on an offshore oil and gas production system.
JOURNAL OF MANUFACTURING SYSTEMS
(2023)
Article
Engineering, Mechanical
Liang Zhou, Huawei Wang, Shanshan Xu
Summary: This paper proposes a health assessment method based on depth digital twin for real-time monitoring and assessment of the health status of aircraft engines. The method combines mechanism modeling and data-driven modeling, and achieves high accuracy in identifying failure modes and assessing failure levels.
ENGINEERING FAILURE ANALYSIS
(2022)
Article
Engineering, Electrical & Electronic
Yongchao Zhang, Jia Hu, Geyong Min
Summary: The paper proposes a digital twin-driven intelligent task offloading framework for collaborative mobile edge computing (MEC). By mapping the MEC system into a virtual space using digital twin and optimizing task offloading decisions with deep reinforcement learning, the proposed framework effectively adapts to dynamic environments and significantly improves the MEC system's income.
IEEE JOURNAL ON SELECTED AREAS IN COMMUNICATIONS
(2023)
Article
Computer Science, Artificial Intelligence
Xiaoqiao Wang, Mingzhou Liu, Conghu Liu, Lin Ling, Xi Zhang
Summary: The service stability of industrial robots is crucial for intelligent manufacturing operations. Knowledge-based work plays a central role in intelligent manufacturing, and the expression and construction of robot data and knowledge are important for predictive maintenance (PdM). This study proposes a PdM method based on data and knowledge, which automatically formulates PdM strategies using a running-state feature-recognition model and fault prediction. The effectiveness of the method is verified through application to welding robots.
EXPERT SYSTEMS WITH APPLICATIONS
(2023)
Article
Chemistry, Physical
Safa Meraghni, Labib Sadek Terrissa, Meiling Yue, Jian Ma, Samir Jemei, Noureddine Zerhouni
Summary: This study utilizes digital twin technology to establish a health monitoring and remaining life prediction system for PEMFC systems, demonstrating high prediction accuracy and robustness through experimental data validation.
INTERNATIONAL JOURNAL OF HYDROGEN ENERGY
(2021)
Article
Computer Science, Interdisciplinary Applications
Raymon van Dinter, Bedir Tekinerdogan, Cagatay Catal
Summary: The main objective of this study is to develop and evaluate a Reference Architecture for Digital Twin-based predictive maintenance systems. The research involved domain analysis, UML diagram design, and evaluation through case studies. The results showed the effectiveness of the Reference Architecture in designing Application Architectures for different scenarios.
COMPUTERS & INDUSTRIAL ENGINEERING
(2023)
Article
Engineering, Industrial
Jinyue Li, Gang Zhao, Pengfei Zhang, Maocheng Xu, He Cheng, Pengfei Han
Summary: This paper proposes an aero-engine assembly quality assessment method based on cumulative block information modeling and process-oriented assembly twinning, and verifies the effectiveness of this method through experiments.
JOURNAL OF MANUFACTURING SYSTEMS
(2023)
Review
Engineering, Industrial
Chong Chen, Huibin Fu, Yu Zheng, Fei Tao, Ying Liu
Summary: The recent advance of digital twin (DT) has facilitated the development of predictive maintenance (PdM) by enabling accurate equipment status recognition and proactive fault prediction. However, the research and application of DT for PdM are still in their infancy, as the role of machine learning (ML) in this area has not been fully investigated.
JOURNAL OF MANUFACTURING SYSTEMS
(2023)
Article
Engineering, Industrial
Liangliang Zhuang, Ancha Xu, Xiao-Lin Wang
Summary: Recent years have seen significant advances in predictive maintenance (PdM) for complex industrial systems. However, existing literature primarily focuses on either prognostics or maintenance decision making, without integrating the two stages. In this paper, we propose a dynamic PdM framework that integrates prognostics and maintenance decision making, using a Bayesian deep learning model to characterize the relationship between degradation features and remaining useful life (RUL). The framework can generate a predictive RUL distribution that effectively represents prognostic uncertainties, and update maintenance decisions based on the latest predictive RUL information.
RELIABILITY ENGINEERING & SYSTEM SAFETY
(2023)
Article
Chemistry, Analytical
Jung-Sing Jwo, Cheng-Hsiung Lee, Ching-Sheng Lin
Summary: Introducing Industry 4.0 into the manufacturing processes of aircraft composite materials is inevitable due to the complexity of the aerospace and defense industry. This study proposes the concept of Data Twin to simplify high-fidelity virtual models and uses machine learning approaches to achieve it. A microservice software architecture, Cyber-Physical Factory (CPF), is also proposed to simulate the shop floor environment, along with two war rooms for establishing a collaborative platform.