Uncertainty Propagation in Dynamic Event Trees - Initial Results for a Modified Tank Problem

The coupling of plant simulation models and stochastic models representing failure events in Dynamic Event Trees (DET) is a framework to model the dynamic interactions among physical processes, equipment failures, and operator responses. The benefits of the framework, as a number of applications show, include, for instance, the capability to account for the aleatory timing of equipment failures or operator actions on sequence outcomes and to consider the impact of the number of available trains (rather than having to identify the bounding cases). The integration of physical and stochastic models may additionally enhance the treatment of uncertainties. Probabilistic Safety Assessments as currently implemented, e.g. for Level 1, propagate the (epistemic) uncertainties in the probability distributions for the failure probabilities or frequencies; this approach does not consider propagate uncertainties in the physical model (parameters). The coupling of deterministic (physical) and probabilistic models in integrated simulations such as the DET allows both types of uncertainties to be considered. The starting point in this work is to consider wrapping an epistemic loop, in which the epistemic distributions are sampled, around the DET simulation. To examine the adequacy of this approach, and to allow different approaches and approximations (for uncertainty propagation) to be compared, a simple problem is proposed as a basis for comparisons. This paper presents initial results on uncertainty propagation in DETs, obtained for a tank problem that is derived from a similar one defined for control system failures and dynamic reliability. An operator response has been added to consider stochastic timing.