A Natech accident is the impact of a NAtural event on an industrial plant (such as an ICPE- Classified Installation for the Environmental Protection). The impact can trigger a TECHnological accident, damaging the vicinity of the industrial facility (inhabitants, their properties and the environment). The final objective of this work is to develop a risk analysis methodology for the prevention of technological accidents triggered by flooding of industrial facilities. Therefore, the work begins with an analysis of past technological accidents triggered by floods. Thus, it was identified that Natech accidents are the consequences of hazardous materials releases from critical equipment, which in turn are damaged directly and indirectly by flood waters. Depending on the type of hazardous substances released, three categories of technological accidents are observed (fire, pollution and explosion). Afterwards, a risk analysis methodology was developed. It is mainly based on the MADS-MOSAR method, which was modified in order to be adapted to the Natech issue. The analysis resulted in the identification of five generic scenarios of technological accidents triggered by floods that can occur within an industrial installation. These scenarios were represented using the bow-tie tool. Then, around sixty preventive and protective measures were identified. The validation of the elaborated scenarios was done, on one hand, in the surface treatment industrial sector and, on the other hand, in two specific industrial facilities. The validation phase helped to test the completeness and plausibility of the generic scenarios, as well as the relevance of the measures previously identified. Furthermore, the validated scenarios were used to develop a checklist helping operators to decrease the vulnerability of their industrial facilities to technological accidents triggered by floods, along with a catalog of preventive and protective measures. Finally, the tool was applied to an industrial installation in order to enhance it and test its relevance by operators.
