https://www.mdu.se/

In response to increasing demands for reliability and uptime, organizations are progressively monitoring more of their mission-critical assets through various sensing and data collection devices. The accumulated data enables several emerging technologies, particularly data-driven approaches such as machine learning, which are becoming more viable in industrial contexts. These technologies have the potential to enhance the effectiveness and efficiency of asset management and maintenance. A key framework for realizing this potential is prognostics and health management, an engineering approach that deals with the identification and prognostication of system degradation. A major aspect of prognostics and health management is remaining useful life prediction, which develops models to forecast the remaining operational time of systems. Accurate prediction of future system state provides useful insight that aids maintenance planning. This thesis addresses challenges and aspects of data-driven remaining useful life prediction with a focus on deep learning-based approaches. The research proposes solutions to key challenges in remaining useful life prediction, including limited access to complete run-to-failure trajectories, data sharing constraints, and decentralized training requirements. Additionally, this thesis investigates remaining useful life predictions for discrete power electronics, components used in safety-critical high-power applications such as automotive systems -- an area that remains understudied within prognostics and health management. The findings demonstrate that remaining useful life prediction is a viable technology in these domains, with models benefiting from self-supervised pretraining and decentralized training through federated learning. Furthermore, the research establishes that discrete power electronics can be effectively monitored using data-driven remaining useful life prediction methods.

Systems of systems (SoS) are interconnected systems that include human, social, organizational, and technological elements. These complex systems have increased in today’s digital society; the elements in a SoS are called constituent systems (CS) and are characterized by managerial and operationally independent interactions to achieve a common goal. Such a goal could be lower cost, robustness, or greater capability. Examples of SoS in modern society include systems in the health, aviation, transport, crisis management, and energy sectors and smart cities. SoS are dynamic and evolve. This dynamic and evolving nature of SoS, coupled with interactions between CS and their managerial and operational independence, poses challenges for risk analysis. Traditional risk analysis methods tend to focus on individual components and known probabilities. This renders them inadequate for the complexities of risk that could arise during CS interactions and can emerge as SoS evolves. The main goal of the thesis is to contribute towards a general process for systematic risk analysis of interconnected societal systems, where modeling methods and tools support each process step. Using a non-traditional approach to safety, the System-Theoretic Accident Model and Processes (STAMP), this thesis validates its suitability for risk analysis in SoS. Results indicated that the approach supported risk analysis in SoS, however, various risk sources with SoS characteristics, such as dynamic structure and latent risks, are not sufficiently handled in STAMP. This thesis demonstrates and proposes a process broadening approach for SoS risk identification, analysis, and modeling adapted for SoS. The Thesis validated this approach by analyzing and modeling risk for a wildfire and COVID-19 management SoS. Further, the thesis contributes to the body of knowledge of a process for identifying risks in SoS. This process demonstrates how broader sources could be identified in SoS and the thesis contributes to knowledge of methods for SoS risk analysis. This systematic identification and analysis of risk in SoS contributes to a conceptual framework for risk analysis in SoS. Furthermore, through a systematic literature review, the thesis contributes to the state of the art in SoS risk analysis, management, and governance providing indications to researchers with areas for research in SoS risk analysis. The thesis enhances the understanding of risk in SoS by bridging the gap between the literature and practice, for risk analysis in non-traditional systems such as SoS. Future work aims to create a general process for systematic risk analysis of interconnected societal systems.