“I’m wondering if this is ever a concern for manufacturers, particularly aluminium frames, or if a frame that’s designed stiff enough will have a very long fatigue life anyway. Anyone know? “
That is the aim of trying to make an aluminium bike frame; a stiffer frame will flex less, the flexing of aluminium is where the problems start with aluminium fatigue life. Although a stiffer frame may be heavier to avoid flexing; an aluminium frame that does flex (a lot) will have a shorter fatigue life than one that does not flex as much.
A steel (or titanium frame) can flex with less of a problem for the fatigue life of the frame.
Car springs are made of steel instead of aluminium for the same reasons.
Don Ferris of Anvil has some thoughts on the subject:
http://www.sportandme.com/docs/sports/cycling/frame_weight_wars.html
“When discussing aluminium, someone always brings up airplanes.
Airplane design showcases what aluminium does best: acceptable strength and a low relative weight. But, aluminium’s lack of a fatigue limit is one very good reason why there is stringent monitoring of dynamically or cyclically stressed aluminium structures. Airplanes are also designed to allow sub-assembly replacement as they approach the end of their life cycle, which is an option you really don’t have on a bicycle frame. The inverse example along these lines may be why springs and paperclips aren’t made of aluminium. “
http://www.frameforum.org/forum2/index.php?showtopic=2317
“Even the largest and most efficient cargo aircraft have a payload ratio of less than 4 to 1, with most being less than 3 to 1. In other words, for one pound of aircraft weight, they can carry 3 pounds of payload. A 20 pound road bike with only a 150 pound rider has a load carrying ratio of 8.5 to 1. Put a 200-pounder on that bike and it grows to 11 to 1. Put that 200 pounder on a 40-pound DH rig designed for 9-g loads, and you’re still at 6 to 1. No the best way to compare stresses obviously, but it does put into perspective how bicycle frames are exceptionally stressed for their mass.”
Similar story applies to CFRP, if used on aeroplanes; there is a system of classification of parts on planes eg critical parts, sensitive part. Sensitive parts, if they fail, will cause a drop in aircraft performance in flight eg speed, fuel consumption etc, Critical parts, if they fail, will lead to the plane falling out of the sky. CFRP is not first choice for critical parts.