Umeå universitets logga

This thesis introduces an ontology-based approach to proactive risk assessment inhigh-risk industrial environments, specifically in molten metal handling processes at smelting plants. The ontology is structured based on distributed cognition theory, which is highly relevant in this setting due to the complex interactions between human operators, equipment, and the operational environment. Distributed cognition allows for a holistic representation of how knowledge and responsibilities are shared among stakeholders, making it well-suited for capturing the dynamic and collaborative nature of risk management in such systems. A grounded theory methodology is employed for transparent knowledge elicitation, capturing expert insights from professionals working at the Rönnskär smelting plant. Through a series of structured knowledge engineering steps, this expert knowledge is transformed into an ontology for proactive risk assessment in this type of safety-critical setting. An evaluation of an implementation integrating the ontology-based back end with a dynamic front end demonstrated strong alignment with expert decision-making and understanding of the environment, with notable reductions in communication gaps. Importantly, users reported high satisfaction, particularly regarding safety assessments, with the system proactively recognizing risks such as equipment containing hot molten metal, which can not be manually inspected. This highlights the system’s potential to prevent critical safety incidents in smelting environments. While developed for a specific industrial context, the methodology of creating a shared cognitive environment to bridge knowledge gaps across various professions shows promise for broader applications. This approach could potentially enhance decision-making, collaboration, and safety across diverse sectors, offering opportunities for improvement in complex operational environments beyond industrial smelting.