(80f) Runaway Reaction ; Validating a Less Overestimating Vent Sizing Method

Vent sizing methods were based for a long time on the assumption of one-phase venting (vapor/gas). Under the impulse of the DIERS (Design Institute for Emergency System Relief), these methods evolved to the use of simple analytical models (and complex computer codes) based on data obtained from adiabatic calorimetry. These vent sizing methods take the occurrence of two-phase flow into account. Nevertheless, because of the simplifying assumptions used (mainly conservation of the initial reactive mass and homogeneous vessel venting), DIERS methods often lead to unrealistically large vent area. This is especially the case of untempered systems (gas-generating and hybrid systems, peroxide decomposition for example). H. Fauske [Properly size vents for non reactive and reactive chemicals ; Fauske, Chem. Engng. Prog. 2000 ; pp 17-29] proposed a new vent sizing method, based on large scale experiments, which refers to one phase flow venting. It thus leads to a much smaller vent area. The objective of this presentation is twofold. The first objective is to propose our understanding of Fauske's method bases : when changing from one phase to two phase venting, decrease of mass inventory in the reactor could compensate for the venting velocity decrease. The second objective is to verify whether this idea can always apply. We tested this approach with a gassy system example by a computer simulation, which takes the vented mass loss and the venting velocity decrease into account. The calculation, which was built with some assumptions, allowed us to identify type of systems for which this approach could potentially be not conservative. One is when the reaction kinetics are already rapid at vent opening and there is not much time for mass to vent before turnaround. Another is when high quality vent flow occurs (gas with small fraction of liquid) so that vented mass is relatively small, the vent quality being however severely decreased.