Hydrogen Explosions

DOE-STD-6003-96

6.1.3.10 Hydrogen Explosions

For a potential hydrogen explosion in a safety-class SSC, DOE 1989 specifies design

requirements that require clarification. This is provided in some detail below.

A hydrogen detonation is a potential hazard that may be a design-basis event (typical

probability >106/yr). If it is within the design basis and the SSC under evaluation is a confine-

ment barrier, then the required integrity of the barrier must be maintained during and after this

event, although the non-safety-related functions of the SSC (such as ability to maintain high

vacuum) can be compromised. If the SSC is not safety-class and a hydrogen detonation is

credible, it must be shown that no failure due to this event can degrade the function of an adja-

cent safety-class SSC.

To determine if a potential hydrogen detonation is a design-basis event, it is important to

evaluate the likelihood of having the three ingredients for detonation at the same time: hydrogen

and oxygen in the appropriate mixtures and an ignition source (NRC 1989). Generally, the

energy required to ignite hydrogen-air mixtures is modest (NRC 1989). Since the plasma typi-

cally contains much higher levels of stored energy, for analysis of the vacuum vessel it should

be assumed that a point ignition source is always present during normal operations and wall

conditioning. The factors determining the likelihood of a detonation are then the availability of

hydrogen isotopes and air. Hydrogen isotopes are present in the solid matrix of the plasma-

facing components at substantial levels. This is not ordinarily available for combustion or deto-

nation although a portion (including tritium) may be released if a detonation occurs. If hot

plasma-facing components or the vacuum vessel are cooled with water, a leak could result in

the generation of hydrogen from water (steam) and beryllium (or carbon or tungsten) reactions

(Smolik 1991, 1992). The precise amount of hydrogen generated depends on the first wall

material and temperature and the size and duration of the water leak, but typical conditions in a

D-T fusion plasma could generate sufficient quantities of hydrogen for a detonation. Air also has

to be present for a detonation. If air is adjacent to the SSC under evaluation, the in-leakage of

air is possible due to the same event that generated the hydrogen. For example, beryllium-

steam reactions from a water leak during wall conditioning can result in internal pressures of

several bar or more (NET 1993), which may be beyond the design value of the SSC under eval-

uation. This air source can be eliminated in the device design by incorporating an inert gas vol-

ume in the region between the SSC under evaluation and the next confinement barrier. To

determine the probability of a hydrogen detonation, an analysis of the above factors must be

performed for a particular design. The likelihood of a loss-of-coolant event cannot be generally

excluded given performance of actively cooled systems to date and the anticipated in-vessel

service conditions in a D-T fusion facility.

To preclude a hydrogen detonation for consideration as a design-basis event, it will typi-

cally be necessary to demonstrate a low event probability by

a. minimizing hydrogen generation by careful design, including material selection of the

plasma-facing components or the fluids used for active in-vessel component cooling,