Proceedings of the Eighth International Conference on Probabilistic Safety Assessment & Management (PSAM)
95 Modeling Accidents for Prioritising Prevention (PSAM-0102)
Download citation file:
- Ris (Zotero)
- Reference Manager
The Workgroup Occupational Risk Model (WORM) project in the Netherlands is developing a comprehensive set of scenarios to cover the full range of occupational accidents. The objective is to support companies in their risk analysis and prioritization of prevention. Its aim is to introduce risk-based thinking to the full range of accident prevention. The project is part of a larger initiative to reduce the burden of occupational accidents in the Netherlands by a significant percentage.
This paper describes the development of the modeling of the scenarios leading to accidents through a number of similar projects in the chemical industry, with the WORM project in general industry and in the future to air transport accidents in the CATS (Causal model for Air Transport Safety) project, which started in 2005. The modeling technique is based on the bowtie, adapted from major hazard modeling, with addition of more explicit modeling of the barriers needed for risk control, the tasks needed to ensure provision, use, monitoring and maintenance of the barriers, and the management resources and tasks required to ensure that these barrier life cycle tasks are carried out effectively. The way in which risk control measures are being modeled is moving from seeing them statically as barriers which can fail, to seeing them dynamically as (fallible) means for the system operation to stay within a safe envelope. The paper illustrates the issues arising from the modeling choices, with examples from the projects. These issues include the level of detail for linking causal factors with management influences; clear definitions of, and distinctions between the primary safety functions, their operationalization in the form of risk control measures and the management tasks needed to ensure that they are put in place and continue to work effectively; the conceptualization of risk control measures as complete entities made up of a set of elements fulfilling a generic set of functions; the problem of deciding on the depth of modeling of management tasks to resolve the so-called ‘Russian doll’ problem, and the incorporation of dynamic, on-line risk control alongside static barriers.
The paper illustrates how concepts develop slowly over a series of projects as a core team works continuously together and iteratively improves its approaches to a complex and difficult area of modeling.