Author ORCID Identifier
https://orcid.org/0009-0007-3109-5865
Document Type
Theses, Ph.D
Disciplines
1.1 MATHEMATICS, Automation and control systems, Chemical engineering (plants, products), Chemical process engineering, *human – machine relations, Interdisciplinary
Abstract
This thesis addresses a operational challenge in modern industrial operations: the increasing complexity of systems and the consequent cognitive burden on operators. As industrial technologies advance, the human-computer interface has become the primary conduit for information flow, playing a pivotal role in operational decision-making. However, the proliferation of data often leads to information overload, potentially compromising rather than enhancing operator performance. This research explores an approach to this pressing issue through the application of Bayesian networks as decision support systems in safety- critical scenarios. Our study employs a multi-faceted approach, combining theoretical modeling with empirical testing. Through collaboration with industry partners, we obtained and analyzed alarm logs from an offshore gas platform, providing crucial real-world context to our research. This industry insight informed the development of a control room simulation, serving as a testbed for our proposed decision support system. The core of our research lies in the application of Bayesian networks to process complex, multi-variable data streams in real-time. These networks, capable of incorporating both empirical data and expert knowledge, offer a robust framework for modeling intricate systems and supporting informed decision-making under uncertainty. To evaluate the efficacy of our developed decision support systems, we conducted a series of controlled experiments using the developed simulation. Participants were recruited and given training to perform a process control task in three scenarios. In each scenario, we collected performance metrics alongside subjective participant feedback and physiological signals to assess the mental workload and situational awareness of participant with or without the AI-enhanced decision support. The experimental design was aimed at providing a comprehensive evaluation of the overall effect of the tools on the likelihood of optimal performance in the face of safety-critical alarms. The results of our study indicate improvements in operator performance and reduced cognitive workload among participants using the decision support system. Intriguingly, we observed a reduction in situational awareness, a finding that warrants further investigation and highlights the complex interplay between technological support and human cognition in high-stress environments. Furthermore, using the experimental data, we developed a human reliability assessment model. This model identified the key operational variables, such as the number of alarms, iv along with variables related to the human factor, such as task load, as effective predictors of human error. This thesis not only contributes to the theoretical understanding of human- computer interaction in safety-critical systems but also provides practical insights for industry application. We present a detailed methodology for implementing Bayesian network-based decision support systems, offering a valuable resource for practitioners and researchers alike. In conclusion, this research represents a contribution in the field of industrial safety and human factors engineering. By addressing the challenges of information overload and system complexity, we pave the way for more resilient, efficient, and safer industrial operations across various sectors.
DOI
https://doi.org/10.21427/4sj0-je38
Recommended Citation
Mietkiewicz, Joseph, "Bayesian networks for safety-critical systems" (2025). Theses. 18.
https://arrow.tudublin.ie/sfehthes/18
Funder
European Union
Creative Commons License
This work is licensed under a Creative Commons Attribution-NonCommercial-Share Alike 4.0 International License.
Included in
Applied Mathematics Commons, Chemical Engineering Commons, Mathematics Commons, Statistical Models Commons
Publication Details
This dissertation is submitted for the degree of Doctor of Philosophy, School of Food Science and Environmental Health, Technological University Dublin, 2026.
doi:10.21427/4sj0-je38