Now that I have seen this photograph the problem is very obvious - as highlighted in the above quote.Quote:
Originally Posted by Bush65
https://www.aulro.com/afvb/images/im...013/11/840.jpg
The splines look (from where I am) like they are rolled, not cut, which is good. However they finish close to a shoulder, which is bad.
The pics don't show the important end of the hub, nor where the end of the hub finishes relative to the end of the spline and the shoulder. Both of those are important details when considering the stress raisers.
Splines are bad stress raisers, but their affect can be reduced with appropriated detailing. Rolling is more beneficial than cutting. If cut they should have a long run-out. By far the best option is to increase the diameter where the spline is so the root diameter of the spline is much more than the diameter of the shaft and in conjunction use a very large transition radius (you will see this kind of detail in many torsion bars used in some vehicle suspensions).
Steps in shaft diameter are bad stress raisers and the larger the size difference the greater the stress concentration. Using a large fillet radius reduces the stress concentration. Note that there are two steps in the diameter, one step up to the seal journal and another to the gear hub. The step to the gear hub is a large, and the intermediate step up to the seal journal will go some way to reducing the stress concentration that would have occurred if it wasn't present.
When a component such as a bearing or hub is fitted on a shaft, it will be a stress raiser. This is why I mentioned the fitting of the splined hub before - the amount of stress concentration depends upon a number of details.
When there are more that one stress raiser within close proximity their affects are compounded. As the separation distance increases the compound affect reduces and can be dismissed when the distance exceeds about one shaft diameter.
In this case we have several stress raisers, a spline, two shoulders, and a fitted hub, in close proximity, such that the stress concentration from each will increase the overall stress and reduce the fatigue strength of the shaft.
Attention to detail for reducing stress concentration is very important when designing components that are subject to fatigue. That is components subjected to fluctuating, cyclic or repeated load events. Loads that reverse direction are worse than those applied in one direction.
Fatigue strength/life is complex, but is characterised by failure after a number of load cycles each of which are less than the static strength of the component. Each load application resulting in a stress magnitude above the Endurance Strength (BTW only possessed by steels) causes damage which accumulates (called cumulative damage - see Miner's Rule).
Stress raisers have little affect on components under static loads.
When evaluating fatigue strength, we reduce the static strength of the component using reduction factors that account for the stress raisers (which cause stress concentration), but also the material strength, and the size of the component. Greater reduction factors must be used for higher strength steels, and for larger diameter components.
BTW the Australian Standard for design of rotating shafts (sorry can't recall the AS number (1410 rings a bell)) is only applicable for tensile strength less than 900MPa, which seems to be related to the difficulty of evaluating a reliable/appropriate reduction factor for material strength over 900MPa.
The reduction factor for diameter is due to the affect of heat treatment on larger components.
So we have here a stub shaft that has stress raisers that I have pointed out. These come before designers regularly, nowhere near as frequent as design for static loads, but every mechanical design engineer knows about them. Applied correctly, the design process should have ensured a shaft diameter that will give an appropriate finite life, if not an infinite life (stress needs to be below the endurance strength to achieve infinite life).
My opinion is that proper consideration has not been given for shock/impact loads that occur in much four wheel driving. Designing for shock or impact loads is not an every day process, far from it, and many don't naturally anticipate it or are familiar with how to design for it.
In this case one of the most important design details is resilience the ability to absorb impact energy by transforming it into strain energy, without exceeding the allowable stress. The amount of strain energy is a function of the distribution of stress over the volume of material. If you have a small volume of material the stress will be higher for the equivalent strain energy.
BTW I should have pointed out that there is a direct relation between stress and strain. Modulus of Elasticity (E is approximately constant for all steels) = Stress divided by Strain, which can be rearranged to Stress = E x Strain.
I will bet my bottom dollar that the failure occurred because the shaft lacks enough resilience, thus very high stresses were concentrated in the area at the end of the splines. It wouldn't have necessarily had to fail during a shock load application, but any and all shock loads during its life will have contributed to the accumulation of fatigue damage.
With that design the volume of material that can absorb the impact energy and transform it into strain energy is the miniscule amount of axle from the inner end of the spline in the flange, to the shoulder of the seal journal. It is no wonder it failed there and the plane of failure supports this (it is not remotely consistent with torque overload, which occurs at an angle where torsional shear stress occurs).
