Yes - I believe that these critical fixings (crank assembly and big end) should never be reused.
If done properly they will have been torqued to proof and, as such, their life is spent.
In fatigue-critical applications, bolts are pre-tensioned to 70-80% of their yield stress, which may be well over 1000 MPa for high strength fasteners.
Crank cheek flanges and flywheel are thus under compression, and a simple bolt/flange analysis reveals that the alternating stress occuring in the bolts is low and in the order of 10-20% of the bolt material's YS, if sizing is done correctly.
Should bolts be torqued to proof (=yield) stress in a fatigue-critical application? Never! If this is done (in error), the abused bolt and nut need to be replaced.
A petrol engine in a car has a design life of 150 000 km usually - we may assume the design life of a motorcycle to be 100 000 km. While the average speed of a car thoughout its design life is 70 km/hr, a motorcycle may fetch 80 km/hr.
This corresponds to 1250 running hours. Further assuming the average tachometer reading is 3000 rpm, we derive at 2.85E8 load cycles for the crankshaft bolts.
Bolt load depends on the revs to the square, but for now let's assume that the bolts were designed for an average of n=3000 rpm. A proper analysis would require a so-called rainfall analysis.
Fatigue curves (stress level vs. load cycles to rupture) for high-strength steels have a kink at 1.0E7 load cycles usually, above which their fatigue life is considered "infinite". My simple calculation shows that the crankshaft bolts were designed for an "infinite" life.
So, does this mean they will never fail at high load cycle figures? No! But the occurence is so low, it is never recorded.
Should I replace my crankshaft bolts? If you suspect that your bike has covered 1250 running hours (your crankshaft grind may give a clue - I'd be wary below 0.020"), as a matter of caution I suggest you replace the bolts. It's a cheap insurance.
- Knut