Suppression of secondary explosions in transformer rooms

Abstract
Large-scale explosion tests in a 50m ³ room have been performed to simulate real transformer room explosions. The scale and the room layout were chosen as realistic as possible, with a moderate vent area. Prior to ignition hydrogen gas and transformer oil was sprayed into the room. Reference tests were carried out to investigate how gas concentration (i.e. energy deposited in the electric arc) and amount of oil mist would influence the explosion strength. Thereafter 3 manufacturers were invited to mitigate and suppress the explosions. The manufacturers were Atex (superheated water), Fike (powder) and Kidde (water suppression with salts added). In the initial tests showed that the amount of gas is a very important parameter for the explosion load, an implication of this is that a good circuit breaker in the transformer is important. The amount of oil mist was found to influence the explosion loads less. The suppression tests showed that for weak explosions, with average gas concentrations about lower explosion limit, all concepts reduced the explosion loads significantly. For higher gas concentrations (twice as high, still far from worst-case), only a moderate pressure reduction, or no reduction at all, was seen. Some concepts reduced the pressure impulse and flame exposure time significantly. For the strong explosions, it is assumed that performance of most systems could be improved, and suggestions for improvements for all concepts are proposed.