A 320 Engine Relight

[Dave Therhino]

As said above the t<20 seconds refers to the time between the flameout (by switch movement or other cause) and moving the switch to ON or RUN to initiate a relight.

I was not involved in the testing for the A320/Leap, but based on my experience from other programs I'm pretty sure the reason for the 20 second limit on use of the quick relight procedure is the lack of adequate engine compressor stall margin after 20 seconds of rundown due to residual heat in compressor section. Within the first 20 seconds after flameout from high power, there is enough compressor speed (N2) to avoid compressor stall when the engine is relit and fuel to accelerate the engine is applied. After that many seconds of deceleration, the compressor speed is low enough that the heating of the air by the hot metal in the compressor introduces enough expansion of the air to, in combination with the backpressure from lighting the combustor and increasing fuel flow to accelerate the engine, cause the front stages of the compressor to stall and cause the engine to fail to accelerate. I don't know the Leap's specific characteristics, but an engine can potentially exceed the EGT limit in this condition.

After you reach that point in the deceleration of the N2 shaft, the only way to get a successful restart is to wait until the compressor cools a bit. By the time the engine has cooled enough, N2 is fairly stabilized and you would follow the stabilized windmill start procedure or starter assisted start procedure. I'm not familiar with the extent to which Airbus has automated selection of those processes, but motoring the engine or increasing airspeed would speed the cooling of the compressor.

The original Trent 800 relight logic dealt with this issue by keeping fuel flow on to hold the engine in a stalled condition but prevent further deceleration, ramping fuel up and back down every few seconds to see if the engine would accelerate. Holding the speed up kept the airflow higher and cooled the compressor faster. Once acceleration was sensed the engine control would proceed with a normal starting fuel flow schedule.

Tdracer was likely involved in control system changes to optimize the quick windmill relight characteristics of at least few Boeing engine installations. Maybe he'll chime in soon.