From data to knowledge: modelling the imperfect fit for safety management in socio-technical systems

The complexity of socio-technical systems is threatening system safety, and safety science acknowledged the need to rely on approaches that go beyond deterministic cause-effect relationships. However, acknowledging this complexity is nothing but easily knowledgeable, nor fully manageable. This thesis explores how to leverage the impossibility of imperfect knowledge to strengthen safety management in socio-technical systems. This aim is achieved by transforming available data into useful knowledge. To this end, this thesis searches for a model that is capable of enabling such transition, namely, the “imperfect fit”. A set of design criteria for a modelling approach to characterise socio-technical complexity is developed at first. Then, the design criteria are used to design methodologies for analysing relevant complex socio-technical systems’ models. Eventually, the thesis suggests how a model based on knowledge graphs shows compelling signs of the imperfect fit, ultimately strengthening safety management in complex systems.