Översikt
- Datum:Startar 22 januari 2026, 10:00Slutar 22 januari 2026, 13:00
- Plats:IMS Room VDL - Virtual lab
- Opponent:Professor J. Ordieres Meré, Universidad Politécnica de Madrid, Spain
- AvhandlingLäs avhandlingen (Öppnas i ny flik)
Manufacturing industries recognize the transformative potential of artificial intelligence solutions within manufacturing maintenance, yet successful implementation remains limited despite widespread acknowledgment of their benefits. Although theoretical advantages of using AI such as improved efficiency, cost reduction, and equipment reliability are well-documented, a persistent gap exists between anticipated outcomes and realized value in industrial settings. Many AI initiatives that begin as successful proof-of-concepts struggle to achieve full-scale production deployment, highlighting gaps in understanding the specific socio-technical challenges practitioners encounter.
This thesis addresses this gap by investigating the interconnected challenges of AI integration in manufacturing maintenance and systematically evaluating operational frameworks to identify the most suitable approach for successful deployment. Through a comprehensive multi-method approach within the Design Research Methodology framework, this research provides both theoretical insights and practical solutions for bridging the gap between AI potential and industrial implementation reality.
The research identifies interconnected challenge domains that collectively constrain AI implementation: infrastructure limitations, scalability constraints, workforce skill gaps, and inadequate maintenance strategies for deployed AI systems. The theoretical process model reveals these challenges as an interconnected system rather than isolated barriers. Through systematic evaluation of alternative operational frameworks, MLOps emerges as a particularly suitable approach, with fundamental characteristics that address integration obstacles. In Addition, the research demonstrates how containerized monitoring infrastructure combined with human-centric methodology creates a powerful foundation for MLOps implementation. This work presents a network map that guides practitioners by linking identified challenges to suitable MLOps architectural components. By establishing MLOps as the enabling operational framework and providing evidence based architectural guidance, this thesis transforms AI solutions from experimental technology into a reliable support tool for manufacturing maintenance, enabling organizations to benefit from data driven maintenance solutions while contributing to the ongoing evolution toward Industry 4.0 and beyond.
This thesis addresses this gap by investigating the interconnected challenges of AI integration in manufacturing maintenance and systematically evaluating operational frameworks to identify the most suitable approach for successful deployment. Through a comprehensive multi-method approach within the Design Research Methodology framework, this research provides both theoretical insights and practical solutions for bridging the gap between AI potential and industrial implementation reality.
The research identifies interconnected challenge domains that collectively constrain AI implementation: infrastructure limitations, scalability constraints, workforce skill gaps, and inadequate maintenance strategies for deployed AI systems. The theoretical process model reveals these challenges as an interconnected system rather than isolated barriers. Through systematic evaluation of alternative operational frameworks, MLOps emerges as a particularly suitable approach, with fundamental characteristics that address integration obstacles. In Addition, the research demonstrates how containerized monitoring infrastructure combined with human-centric methodology creates a powerful foundation for MLOps implementation. This work presents a network map that guides practitioners by linking identified challenges to suitable MLOps architectural components. By establishing MLOps as the enabling operational framework and providing evidence based architectural guidance, this thesis transforms AI solutions from experimental technology into a reliable support tool for manufacturing maintenance, enabling organizations to benefit from data driven maintenance solutions while contributing to the ongoing evolution toward Industry 4.0 and beyond.
