[en] This report describes research on flame acceleration and deflagration-to-detonation transition (DDT) for hydrogen-air mixtures carried out in the FLAME facility, and describes its relevance to nuclear reactor safety. Flame acceleration and DDT can generate high peak pressures that may cause failure of containment. FLAME is a large rectangular channel 30.5 m long, 2.44 m high, and 1.83 m wide. It is closed on the ignition end and open on the far end. The three test variables were hydrogen mole fraction (12--30%), degree of transverse venting (by moving steel top plates---0%, 13%, and 50%), and the absence or presence of certain obstacles in the channel (zero or 33% blockage ratio). The most important variable was the hydrogen mole fraction. The presence of the obstacles tested greatly increased the flame speeds, overpressures, and tendency for DDT compared to similar tests without obstacles. Different obstacle configurations could have greater or lesser effects on flame acceleration and DDT. Large degrees of transverse venting reduced the flame speeds, overpressures, and possibility of DDT. For small degrees of transverse venting (13% top venting), the flame speeds and overpressures were higher than for no transverse venting with reactive mixtures (>18% H₂), but they were lower with leaner mixtures. The effect of the turbulence generated by the flow out the vents on increasing flame speed can be larger than the effect of venting gas out of the channel and hence reducing the overpressure. With no obstacles and 50% top venting, the flame speeds and overpressures were low, and there was no DDT. For all other cases, DDT was observed above some threshold hydrogen concentration. DDT was obtained at 15% H₂ with obstacles and no transverse venting. 67 refs., 62 figs