Although I am an engineer and hopefully possess the tools to do an analysis of the Woodruff key in the crankset application, Mark Bulgier is correct that I'm not willing to do the work. I will say that it is probably not the key that is doing the power transmission, but the friction between the arm and the axle when the fixing bolt is tightened and draws the taper together. The use of tapered shafts to transmit power is very common (almost every machine tool is driven by a tapered coupling.) The keys are there to guarantee the correct phasing during assembly. This could have been a simple solution to eliminating cotters while still using the same types of machine tools that were used to make cottered cranks. The machining of square tapers (especially internal ones) is considerable more difficult. The integrity of the tapered connection is dependent on the appropriate installation force (bolt torque in this case), and on not using any low friction or extreme pressure lubricants on the taper. I've seen many sheared keys on tapered connections where the installer thought that this was a good place for anti-seize compound. It wasn't. If the tapered joint is allowed to move, eventually the key or the axle will be destroyed. If the joint is tight, the key will show essentially no wear when it is disassembled. This also applies to cottered crank assembly, as a properly tightened, (or better yet, pressed in) cotter applies a force that results in a frictional drive around the whole circumference of the axle, and not just the shearing force of the cotter section. Most of you have probably seen the result of a loosely installed cotter, where the axle has cut through the cotter. The improperly tightened cotter wasn't any softer than the properly installed one, and as long as it was more than finger tight, it probably made contact along the whole of the axle flat. But because it didn't preload the entire circumference of the axle enough, it was allowed to move and destroy the cotter.
Davis Jensen
Lomita, CA