The answer is primarily "metal fatigue" and design margins.

For steel if the stress is below about 15% of ultimate yield then it
does not fatigue. However the lower fatigue limit for aluminum is
ZERO. Aluminum always fatigues no matter what the stress level is.
Every time the cylinder fires the steel barrel and the aluminum head
gets a stress cycle. The stress is proportional to the peak cylinder
pressure which varies with power output of the engine. The number of
cycles to failure verses stress level is not linear. Since it is non
linear if you design a part that is subject to cyclic stress then
half the stress it will last much more than two times as long before
it cracks.

If you buy a rebuilt cylinder you do not know if it has 500 hours or
5000 hours. You also can not always tell if it has already cracked
and been welded.

In my opinion flying through a rain storm would shock cool a cylinder
MUCH faster than a sudden power reduction.

(This is my opinion and I am sticking to it until I have test data to
indicate otherwise.)
John


On Sat, 2 Oct 2004 10:35:36 -0400, Peter R.
wrote:

What would be the most likely reason a cylinder in a high-time engine
would crack, old-age or poor temperature management on a descent?