Knowledge management for industrial safety, generic resource platform combined with an ontology based approach

International audienceThe capacity to manage risks and maintain industrial safety is largely based on the capacity of various actors to acquire, maintain and share knowledge on a large variety of subjects. The actors are, of course, the plant operator but also the employees, the competent authorities, the external maintenance teams or internal or external experts in charge of risk assessment and design of risk management. The knowledge ranges from the regulatory framework to the details of a machine or a process but also includes the general knowledge about industrial safety, the hazardous phenomena, the properties of the substances. Part of this knowledge is also largely tacit. It lies in the brain of the scientific experts or the employees who are able to make the connection between apparently disconnected pieces of knowledge. Detecting, extracting, maintaining and communicating this knowledge are typical knowledge management activities. This is the reason why INERIS has initiated a research program dedicated to knowledge management with several initial outcomes that will be presented in this paper together with the general objectives of the program and the perspectives for the coming years. The first part of the paper is dedicated to a recall of the knowledge management principles based mainly on the typology of knowledge introduced by Nonaka. It is put in relation with the typology of knowledge used in safety management activities. Two examples of knowledge management systems are presented and related to the typology of knowledge previously described. In both cases, the added value resides in the capacity to establish relations between different knowledge elements. This is partly realised by documents indexing and the use of efficient information retrieval tools. The development of an ontology of industrial safety contributes to achieving these goals. It constitutes a reference for indexing and will offer a structure for future developments of semantic web-based tools. Such an ontology is presently under development. Its use to improve the indexing and information retrieval in the knowledge management systems is illustrated