Page 267 - Failure Analysis Case Studies II

252
of section location (causing a stress raiser) and premature failure results. The position of failure,
interestingly, is not exactly where the maximum nominal bending stress occurs (which is at the step
change of diameters) but nearby at the end of keyway region-where the stress concentration is
greater. At the fillet radius at the step change in diameters, though, the shaft is nearly as vulnerable.
Calculations show that the bending stress magnitude is much greater than the shear stress caused
by torsion due to accelerating and braking, and this combined with the much lower frequency of
occurrence of fluctuation of shear stresses effectively eliminates torsion effects as a significant
contribution to fatigue damage. This finding has important implications upon the possible improve-
ment made to the design, as discussed in Section 6.

6. PROPOSED DESIGN MODIFICATIONS

As the current shaft assembly is prone to fatigue failure after a relatively short operational life,
modifications to the detail design are essential if long life is to be obtained. A number of relatively
simple changes to the detail design would substantially improve the resulting safe life of future drive
shafts in service. These can be achieved in a number of ways and some of the most effective and
simple alterations are as follows:
(I) Shortening the length of the keyway groove. The current length is far greater than that required
to transmit the torque from shaft to the wheel. The length may readily be reduced and located
at the outside ends of the shaft as shown in Fig. 5. This will ensure that the end of the keyway
groove (hence location of stress raiser) will be kept clear of the more highly stressed inner
portion of the shaft.
(2) Increasing the diameter of the inner portion of the shaft hub. This may readily be increased
from 12 mm to, say, 16 mm diameter along a 30 mm length and the drive wheel sleeve bore
correspondingly increased to accommodate it. There should remain sufficient depth of thread
left for the M6 locking screw but a smaller screw may readily be used (an M4 would be
adequate).
(3) Increasing the shoulder fillet radius. Although the shaft did not fail at the change of diameter
shoulder, an increase from R3 to R5 would nevertheless be desirable in lowering the stress
concentration.
(4) Reducing the drive wheel hub/sleeve length. If this were reduced even slightly (and the shaft
shoulder increased accordingly) this would be desirable from fatigue considerations and still
leave adequate space for the locking screw.
A drawing showing the proposed four modifications incorporated into the design is given in Fig.

Fig. 5. Wheel shaft with reduced length keyway groove.