8. Conclusion

Although manufacturers are free to use the methodologies of their choice, experience shows that, depending on the criticality of the accident, the use of certain approaches is virtually unavoidable in hazard studies. For example, the bow-tie approach is virtually unavoidable for the analysis of accidents causing hazardous phenomena with lethal effects outside the site. Since 2003, the risk analyses carried out in hazard studies have become much more complex, largely due to the introduction of the probabilistic approach, but also to a "refinement" in the assumptions and models used to estimate the effects of hazardous phenomena. Despite the increasing complexity of hazard studies, it is important not to confine them to the sole purpose of meeting the ever-increasing number of regulatory requirements. It is therefore important to always ensure that the elements reported in the hazard study...