Tom Jenkins said:
I don't know the theoretical background, but I absolutely guarantee that if you have a motor free wheeling close to full load rpm and connect it across the line you will on occasion get very high torques. I have seen it snap shafts and I have seen it shear couplings.
I have seen some similar things. In this situation the motor is not nor ever will be a generator because it has no form of excitation to produce electricity.
I assume (at least in this case) the wind is making the fan turn backwards. The wind can easily turn a fan in speeds equal to or greater than those obtained thru the motor.
You know there is relationship between speed and torque; therefore if it is rotationg (for whatever reason) in reverse, when you apply power you will have toque in direct opposition to each other. If the applied power (torque) is significantly greater then the rotatotional force the sudden change could shear or at least bend parts.
In the 60's I was driving a Chevy with an automatic tansmission and floor mounted shifter, doing about 80 when someone hit the stick and knocked it in reverse. The driveshaft did not survive, thank goodness it was the driveshaft though.
JUST IN CASE: I know many will think "You can not do that with an automatic." Most of the cars, with automatics, before the mid 60's could be pushed or roll started AND could be placed in reverse while in motion. I had a 56 Crown Victoria that had a 6v electrical system and a bad starter, I just rolled started it, never did get a starter for it.
As suggested, obtain all the data you can on the fan i.e. speed of rotation (in reverse) etc then call RA to see what they say.
OR
You could add a brake to prevent it from turning when not powered.
OR
Install louvers that are only open when the fan motor is powered or air blows the proper direction.