Happened before to ATR72. Exactly the same problem. The ATPCS in the 72 is triggered by the same source as the needle on the torque gauge. The digital readout comes from another source. (I can't remember now what units they come from) In this case the digital readout remained, although the needle dropped to zero, therefore a FDR reading of the torque was retained. The engine did fully auto-feather. I believe the torque eventually reached something like 170-180%, despite the fact that the crew realised it just as it happened and made an attempt to pull the power lever back. All props removed and scrapped, but stayed together. Surprisingly, the theory is that as the RPM reduces, the torque should peak and then drop off rapidly despite running at full power, so it is unlikely that the props would disintegrate. Torque is really the force resisting rotation, so both blade angle and RPM combine to produce it. I have been told that the RPM would decrease extremely rapidly and limit the torque achieved. The props are the most limiting factor and the engine should be fine unless the ITT limit was exceeded.
ATR maintenance manual lists the limits for overtorques and I believe about 145% torque or above requires immediate removal and replacement of all affected propeller blades. Below 145% (or so) various inspections are required to ascertain if any significant damage had been done. They are very strict inspections and it can be thought of as - below 145% the props must come off unless the inspection proves that they are still intact. Even at that I believe only a ferry flight may be allowed.
If you ever see a fluctuating needle on a 72 (particularly one that momentarily drops below 20%) think twice about going! I presume the 500's (42 or 72) are the same. Or, turn the ATPCS off and then work out the RTOW and realise you will have to leave all the PAX behind!
Very interesting to hear that it happened somewhere else.
Keep the eyes on that UPTRIM light!!!!
CP