Protection against impacts of hazardous events is an essential part of our everyday life. The challenge is to minimise the risk at much as possible at costs that are acceptable to society. Optimising risk under cost constraints is a typical example of ‘decision making under uncertainties’. The classical engineering approach to deal with uncertainties is to introduce safety factors in the design or safety assessments of structures, as such making the system more “robust”. A potential disadvantage of safety factors is that they can result in unnecessarily expensive designs or unnecessarily short life spans of structures. This may be mitigated by adopting a fully probabilistic approach, in which all relevant uncertainties are explicitly taken into account. A fully probabilistic method is likely to result in less expensive designs that still satisfy the safety requirement(s). The additional advantage of the probabilistic approach is that in principle all potential events that may lead to “system failure” are taken into account and as such, no potential harmful events are overlooked or underestimated.
Probabilistic methods are increasingly popular, also due to the increase in computation power over the past decades. Nevertheless, most practitioners still have the tendency to opt for the more traditional safety factor approach. A probable reason is that a shift towards a probabilistic approach would require practitioners getting out of their comfort zone. Another may be that some unaware of the potential benefits of a fully probabilistic approach. Deltares has developed a probabilistic Toolkit to facilitate practitioners in carrying out probabilistic analyses. This Toolkit is easy to use and as such bridges the gap between experts and non-experts in probabilistic analyses. The Toolkit has a set of ‘built-in’ probabilistic computation techniques which can be linked to the users own models describing the structure or system under consideration. As such, the Toolkit can be applied to virtually any model or system.
We present the setup of the Toolkit and the theory behind the probabilistic computation techniques. The added value of the Toolkit is demonstrated with three case studies.