Significance of fluid-structure interaction phenomena for containment response to ex-vessel steam explosions

When studying the structural response of a containment building to ex-vessel steam explosion loads, a two-step procedure is often used. In the first step of this procedure the structures are treated as rigid and the pressure-time history generated by the explosion, at the rigid wall, is calculated. In the second step the calculated pressure is applied to the structures. The obvious weakness of the two-step procedure is that it does not correspond to the real dynamic behaviour of the fluid-structure system. The purpose of this paper is to identify and evaluate the relevant fluid-structure interaction phenomena. This is achieved through direct treatment of the explosion process and the structural response. The predictions of a direct and two-step treatment are compared for a BWR Mark II containment design, consisting of two concentric walls interacting with water masses in the central and annular pools. It is shown that the two-step approach leads to unrealistic energy transfer in the containment system studied and to significant overestimation of the deflection of the containment wall. As regards the pedestal wall, the direct method analysis shows that the flexibility of this wall affects the pressure-time history considerably. Three load types have been identified for this wall namely shock load, water blow as a result of water cavitation, and hydrodynamic load. Reloading impulse due to cavitation phenomena plays an important role as it amounts to ≈40% of the total impulse load. Investigation of the generality of the cavitation phenomena in the context of ex-vessel steam explosion loads was outside the scope of this work.