A vapor explosion—sometimes called an energetic fuel-coolant interaction (FCI)—is a process in which a hot liquid (fuel) transfers its internal energy to a colder, more volatile liquid (coolant); thus the coolant vaporizes at high pressures and expands and does work on its surroundings. When the two liquids come into contact, the coolant begins to vaporize at the fuel-coolant liquid interface as a vapor film separates the two liquids. The system remains in this nonexplosive, metastable state for a delayed period that ranges from a few milliseconds to a few seconds. During this time the fuel and coolant liquid intermix as the result of density and velocity differences as well as vapor production (Figure 1).
Next, vapor film destabilization occurs and triggers fuel fragmentation; this rapidly increases the fuel surface area and thus vaporizes more coolant liquid and increases the local vapor pressure. This "explosive" vapor formation spatially propagates throughout the fuel-coolant mixture and thus causes the macroscopic region to become pressurized by the coolant vapor. Subsequently the high-pressure coolant vapor expands against the inertial constraint of the surroundings and the mixture. The high-pressure vapor that is produced, the dynamic liquid-phase shock waves, and the slug kinetic energy can all be destructive to the surroundings. In addition, if the fuel is metallic, this explosive dispersal may cause exothermic metal-water chemical reactions that can enhance the work output and produce hydrogen.
The concept of mixing (sometimes called premixing) is vague and has not been well-defined regarding energetic FCIs, In combustion engines, the fuel and oxidant are considered well-mixed when the liquid fuel has atomized into small droplets and they have dispersed in the oxidant and vaporized into it; that is, the fuel and oxidant have become a homogeneous fluid under the proper stoichiometric conditions. In vapor explosions, the interacting species are liquids; thus gaseous interpenetration is impossible. However, the qualitative attribute of attaining a homogeneous geometry is a reasonable concept for mixing in the FCI. Qualitatively, it could be described as a condition in which the fuel and coolant liquids disperse within one another. Past work [Henry, Fauske (1981) and Corradini (1988), (1991)] suggests the following criteria for "good" mixing:
Stable film boiling between fuel and coolant liquids;
The discontinuous liquid must be dispersed in a continuum of the other liquid (void fraction <50%);
Local length scale of the discontinuous liquid much smaller than the system length scale.
In the presence of a trigger, a vapor explosion could spatially propagate through the fuel-coolant mixture. In the past, a number of large-scale experiments were conducted to investigate this explosion phase of the energetic FCI [e.g., Hohmann et al. (1993)]. The first attempt to estimate the maximum work potential from the explosion was provided by Hicks and Menzies (1965) and involved a thermodynamic analysis of the fuel-coolant interaction. The path assumed for this ideal explosion involved a constant volume thermal equilibration of the fuel and coolant (1-2), followed by an isentropic expansion of the fuel-coolant mixture (2-3). Recently, Bang (1992) used the Hicks-Menzies approach with a complete fuel and coolant equation of state to estimate the upper-bound work potential for a variety of fuel-coolant pairs (Figure 2). The peak explosion conversion ratio (work/fuel thermal energy) was again found to be maximized at nearly equal fuel-coolant volumes, and the magnitude depended on the initial void fraction. Also Bang showed equivalence between this and the Board-Hall (1975) detonation model.
A number of theories have been advanced to explain the vapor explosion process, [Corradini (1988), (1991)] in which models were developed to be used in reactor safety analyses and probabilistic studies. For the validation of such models, comparisons must be made with large-scale explosion data. The focus of these comparisons is on observations which can be used to help validate the explosion models. Particular attention should be paid to the qualitative trends, what is quantitatively measured, and the uncertainty of these measurements. Such continuing work helps identify any future experiments that may be useful in basic understanding and model development.
Bang, K. H. and Corradini, M. L. (1992) Therrnodynamic analyses of vapor explosions, NURETH-5, Salt Lake City UT. Sept 1992.
Board, S.J. and Hall, W. B. (1975) Detonation of fuel-coolant explosions, Nature, 254, 319 March 1975.
Corradini, M. L., Kim, B. J. and Oh, M. D. (1988) Vapor explosion in light water reactors: A Review of Theory and Modelling, Progress in Nuclear Energy, 22, 117. DOI: 10.1016/0149-1970(88)90004-2
Corradini, M. L. (1991) Vapor explosions: A review of experiments for accident analysis, J. Nuclear Safety, 32, 3, 337, Sept 1991.
Fauske, H. K. (1981) Required initial conditions for energetic steam explosions, J. Heat Transfer, 19, 99
Hicks, E. P. and Menzies, D. C. (1965) Theoretical studies on the fast reactor maximum accident, Argonne Lab Report, ANL-7120, Oct 1965
Hohmann, H., et al. (1992) KROTOS 26-30 Experimental data report, JRC Ispra Technical Note No. 1.92.115, Nov 1992.