I don't think we have enough details of the set up to really know why it failed.
Wind turbines are severe life testing machines and things will fail on a wind turbine that would run under other conditions.
Unless a coupling is a press fit on a shaft there is likely to be movement during acceleration and deceleration. Even a keyway has to be perfectly cut to avoid this forwards and backwards fretting that will eventually loosen things. I have had satisfactory results with a parallel keyed shaft that is well fitted and locked axially with a good big nut on the end of the thread.
The taperlock scheme is by far the most satisfactory and normally will work perfectly well without a key but a key backup helps.
For small motors and shafts the fit between coupling and shaft is very critical and anything less than a press fit will likely be a source of trouble. If keyed and two grub screws are used fitted at 90 deg then I would expect it to be ok as long as though it is not a very loose fit. For some reason the rest of the world uses one grub screw or two diametrically opposite one another and this will guarantee failure.
A loose pin through the hub and shaft gives short term protection but it will eventually fret through and fail. If the fit is good a roll pin or a well fitted taper pin will be ok, but if the hub is not a really good slide fit or better on the shaft then again these things will eventually fail.
Drilling holes through the shaft can seriously weaken some of the small motor shafts so you have to watch that you don't cause a shaft failure.
Parallel splines can be just as bad unless you have a really decent nut on the end of the shaft and lock the thing endways so tight that it can't fret on the splines ( think of car drive shafts and the big nut with huge tightening torque).
Any form of hub with a taper fitting is fine ( with or without a key) but it will need pullers to remove it and if you work up a tower you will not appreciate having to handle a set of pullers. If you can lower the whole thing then there is no problem.
Even decently engineered things fail and I sometimes fear dreadful things when I see some of the low grade engineering with some of the newcomers with small motors, and limited facilities, they are disasters waiting to happen.
Surprisingly nearly all these things actually fall off when the speed drops and the thing just lands underneath on the ground. When something departs at high speed then there is real potential for damage.
Flux