This fault tree analysis of a total flooding Halon 1301 extinguishing system in a computer room illustrates a methodology developed to estimate the probability that vaporizing liquid suppression systems may fail to protect against fire or smoke damage. The methodology follows four steps: 1) specify a prototype system; 2) identify failure scenarios; 3) determine the probability of occurrence of the failure scenarios; and 4) perform uncertainty calculations.
The prototype used here is a total flooding Halon 1301 system installed in a computer room in 1977. Three fire scenarios are analyzed: (1) a fire of electrical origin inside an equipment cabinet; (2) a paper trash fire; and (3) smoke entering from a fire outside the room (this last scenario is a measure of the detection, control system, and smoke damper effectiveness).
The results show that, of the three scenarios, the halon system is most effective against the fire of electrical origin providing the electrical ignition source is de-energized before the halon supply is depleted. The results also show that the system provides reliable protection against smoke damage from a fire outside the room, but is less effective against the paper trash fire because of the reignition hazard if the supply is depleted. Other analyses performed include the effect of different inspection intervals, human intervention, and acceptance tests on overall system reliability.