Gestion des risques pour les laboratoires de recherche et d'enseignement

The academic research and education risk problem as well as the diversification of risk types have increased concurrently with research development. Moreover, academia induces important worsening factors affecting risk such as high turnover of collaborators, students in education being not salaried and not trained to lab work and high concentration of research labs with diverse hazards. A problem that frequently faces safety personnel in research/teaching labs is to determine how serious each known hazard is? And to decide to what extent it should concentrate his resources to correct the situation. Questions like which risks must be first assessed and how to prioritize them are commonly asked. In response to that preoccupation, we developed a new risk ranking index namely Laboratory Criticity Index-LCI under the so called Laboratory Assessment and Risk Analysis-LARA methodology. The proposed methodology is based on multi-criteria modeling, combining two approaches: the Risk Priority Number-Failure Mode Effect and Critically Analysis (RPN-FMECA) and the Analytic Hierarchy Process (AHP). The common project, between the two polytechnic schools (EPFL and ETHZ), has consisted to formulate and build up the following methodology for assessing risks in research environment. It concentrates on two major sides playing a role in safety: Research and education in chemistry and physics. In research/teaching labs and specifically for sciences involving chemistry, physics or biology, risk could be defined, in a first step, as a subtle relationship between hazard probability and severity induced by this hazard. This first step is not sufficient, other actors such as, worsening factors, exposure factors, cost factors for example, will also play an important role in defining risk. To calculate the risk level, hence the LCI index, several parameters characterizing the site and processes are identified. This include description of process, hazard identification, risk perception (RP), measurement of hazard impact (Ih), probability of occurrence of an accident (POA), research specificities (RS), hazard detectability (HD), implementation of corrective actions (CM). It follows that the LCI, based on these independent variables, becomes a single and unique index, and one-dimensional quantification and prioritization of risks to rapidly locate risk in an area acceptable, tolerable or unacceptable. Through this project, a database is developed, inventorying all hazards presented in laboratories, the related processes in which they are involved, the sources of hazards, their causes and consequences, and worsening factors. Parameters for LCI calculation of each hazard are also stored. LCI is implemented in the database and a web application is create to systemize the used of information. As a unified and interactive tool built on a multidisciplinary approach, it provides a way to perform a preliminary and rapid analysis of hazards and risks, helping and guiding safety teams in research/teaching labs and education. The developed methodology is simple, easy to perform, robust, user-friendly, intuitive and delivers a hierarchisation of risks for correctives actions. The ranking helps to clearly identify critical areas in research/teaching labs. The association between risk level and geographical zones on site is primordial to have a good safety management and risks in over 800 research labs will be analyzed using this methodology.