Generative and Predictive AI for digital twin systems in manufacturing

Dan Dai; Baixiang Zhao; Zhiwen Yu; Pasquale Franciosa; Dariusz Ceglarek

doi:10.3389/frai.2025.1655470

Generative and Predictive AI for digital twin systems in manufacturing

Dan Dai, Baixiang Zhao, Zhiwen Yu, Pasquale Franciosa, Dariusz Ceglarek^*

^*Corresponding author for this work

Research output: Contribution to journal › Article › peer-review

Abstract

The integration of Artificial Intelligence (AI) and Digital Twin (DT) technology is reshaping modern manufacturing by enabling real-time monitoring, predictive maintenance, and intelligent process optimisation. This paper presents the design and partial implementation of an AI-enabled Digital Twin System (AI-DT) for manufacturing, focusing on the deployment of Generative AI (GAI) and Predictive AI (PAI) modules. The GAI component is used to augment training data, perform geometric inspection, and generate 3D virtual testing environments from multiview video input. Meanwhile, PAI leverages sensor data to enable proactive defect detection and predictive quality analysis in welding processes. These integrated capabilities significantly enhance the system's ability to anticipate issues and support decision-making. While the framework also envisions incorporating Explainable AI (EAI), Context-Aware AI (CAI), and Agentic AI (AAI) for future extensions, the current work establishes a robust foundation for scalable, intelligent digital twin systems in smart manufacturing. Our findings contribute toward improving operational efficiency, quality assurance, and early-stage digital-physical convergence.

Original language	English
Article number	1655470
Number of pages	13
Journal	Frontiers in Artificial Intelligence
Volume	8
Early online date	17 Dec 2025
DOIs	https://doi.org/10.3389/frai.2025.1655470
Publication status	E-pub ahead of print - 17 Dec 2025

Bibliographical note

© 2025 Dai, Zhao, Yu, Franciosa and Ceglarek. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

Data Access Statement

The datasets presented in this article are not readily available because as those data are used for the BMW, we should apply from BMW. Requests to access the datasets should be directed to [email protected].

Funding

This was supported by the EPSRC MSI: Made Smarter Innovation Research Centre for Smart, Collaborative Industrial Robotics (EP/V062158/1).

Keywords

quality assurance
Generative AI
smart manufacturing
AI-enabled digital twin
Predictive AI

Access to Document

10.3389/frai.2025.1655470Licence: CC BY 4.0

Generative and Predictive AI for digital twin systems in manufacturing
© 2025 Dai, Zhao, Yu, Franciosa and Ceglarek. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
Final published version, 4.09 MBLicence: CC BY 4.0

Cite this

@article{c024b650fbd74b1386dec35d3f367fbb,

title = "Generative and Predictive AI for digital twin systems in manufacturing",

abstract = "The integration of Artificial Intelligence (AI) and Digital Twin (DT) technology is reshaping modern manufacturing by enabling real-time monitoring, predictive maintenance, and intelligent process optimisation. This paper presents the design and partial implementation of an AI-enabled Digital Twin System (AI-DT) for manufacturing, focusing on the deployment of Generative AI (GAI) and Predictive AI (PAI) modules. The GAI component is used to augment training data, perform geometric inspection, and generate 3D virtual testing environments from multiview video input. Meanwhile, PAI leverages sensor data to enable proactive defect detection and predictive quality analysis in welding processes. These integrated capabilities significantly enhance the system's ability to anticipate issues and support decision-making. While the framework also envisions incorporating Explainable AI (EAI), Context-Aware AI (CAI), and Agentic AI (AAI) for future extensions, the current work establishes a robust foundation for scalable, intelligent digital twin systems in smart manufacturing. Our findings contribute toward improving operational efficiency, quality assurance, and early-stage digital-physical convergence.",

keywords = "quality assurance, Generative AI, smart manufacturing, AI-enabled digital twin, Predictive AI",

author = "Dan Dai and Baixiang Zhao and Zhiwen Yu and Pasquale Franciosa and Dariusz Ceglarek",

note = "{\textcopyright} 2025 Dai, Zhao, Yu, Franciosa and Ceglarek. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.",

year = "2025",

month = dec,

day = "17",

doi = "10.3389/frai.2025.1655470",

language = "English",

volume = "8",

journal = "Frontiers in Artificial Intelligence",

issn = "2624-8212",

publisher = "Frontiers Media S.A.",

}

TY - JOUR

T1 - Generative and Predictive AI for digital twin systems in manufacturing

AU - Dai, Dan

AU - Zhao, Baixiang

AU - Yu, Zhiwen

AU - Franciosa, Pasquale

AU - Ceglarek, Dariusz

N1 - © 2025 Dai, Zhao, Yu, Franciosa and Ceglarek. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

PY - 2025/12/17

Y1 - 2025/12/17

N2 - The integration of Artificial Intelligence (AI) and Digital Twin (DT) technology is reshaping modern manufacturing by enabling real-time monitoring, predictive maintenance, and intelligent process optimisation. This paper presents the design and partial implementation of an AI-enabled Digital Twin System (AI-DT) for manufacturing, focusing on the deployment of Generative AI (GAI) and Predictive AI (PAI) modules. The GAI component is used to augment training data, perform geometric inspection, and generate 3D virtual testing environments from multiview video input. Meanwhile, PAI leverages sensor data to enable proactive defect detection and predictive quality analysis in welding processes. These integrated capabilities significantly enhance the system's ability to anticipate issues and support decision-making. While the framework also envisions incorporating Explainable AI (EAI), Context-Aware AI (CAI), and Agentic AI (AAI) for future extensions, the current work establishes a robust foundation for scalable, intelligent digital twin systems in smart manufacturing. Our findings contribute toward improving operational efficiency, quality assurance, and early-stage digital-physical convergence.

AB - The integration of Artificial Intelligence (AI) and Digital Twin (DT) technology is reshaping modern manufacturing by enabling real-time monitoring, predictive maintenance, and intelligent process optimisation. This paper presents the design and partial implementation of an AI-enabled Digital Twin System (AI-DT) for manufacturing, focusing on the deployment of Generative AI (GAI) and Predictive AI (PAI) modules. The GAI component is used to augment training data, perform geometric inspection, and generate 3D virtual testing environments from multiview video input. Meanwhile, PAI leverages sensor data to enable proactive defect detection and predictive quality analysis in welding processes. These integrated capabilities significantly enhance the system's ability to anticipate issues and support decision-making. While the framework also envisions incorporating Explainable AI (EAI), Context-Aware AI (CAI), and Agentic AI (AAI) for future extensions, the current work establishes a robust foundation for scalable, intelligent digital twin systems in smart manufacturing. Our findings contribute toward improving operational efficiency, quality assurance, and early-stage digital-physical convergence.

KW - quality assurance

KW - Generative AI

KW - smart manufacturing

KW - AI-enabled digital twin

KW - Predictive AI

UR - https://www.frontiersin.org/journals/artificial-intelligence/articles/10.3389/frai.2025.1655470/full

UR - http://www.scopus.com/inward/record.url?scp=105026246102&partnerID=8YFLogxK

U2 - 10.3389/frai.2025.1655470

DO - 10.3389/frai.2025.1655470

M3 - Article

SN - 2624-8212

VL - 8

JO - Frontiers in Artificial Intelligence

JF - Frontiers in Artificial Intelligence

M1 - 1655470

ER -

Generative and Predictive AI for digital twin systems in manufacturing

Abstract

Bibliographical note

Data Access Statement

Funding

Keywords

Access to Document

Other files and links

Fingerprint

Cite this