fatigue life

[autoguru]

I'd like to find out more about the fatigue-strain relationship of mechanical component.

Is it a linear relationship?

For example, if I increase the width of gear or chain, did I just doubled its fatigue life?

I remember faintly that it's not a linear relationship, but more of a curve.

Anyone know?
Thanks

 

[GregLocock]

Depends on many things. For a simple torsional fatigue test many steels have a fatigue life inversely proportional to strain^5

Approximately.



Cheers

Greg Locock

Please see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[autoguru]

What about the "endurance limit". Some steel alloys have a limit where if the load is below it, the fatigue life is infinite.

That means it's theoretically possible to make mechanical components with infinite fatigue life.

Is it often done in real life?

Thanks

 

[GregLocock]

Current thinking around here is that the endurance limit is an artefact produced by poor statistical analysis of the test results.

However, yes, in the past many people have designed to the endurance limit.





Cheers

Greg Locock

Please see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[patprimmer]

Steel bridges tend to have a very long life.

One would hope they are designed to infinite fatigue life.

Regards

eng-tips, by professional engineers for professional engineers
Please see FAQ731-376 for tips on how to make the best use of Eng-Tips Fora.

 

[autoguru]

My guess is most mechanical components are not designed to have infinite fatigue life even thought it's possible, due to size, cost, and necessity reasons.

 

[ivymike]

if you believe in "infinite fatigue life" and put the endurance limit of a ferrous material at the stress which can be tolerated for N>10^7 cycles, then you'd darn well better design your engine internals for "infinite life" in many cases. Do the math - how long does it take to accumulate 10^7 firing cycles at any given crankpin in a 4-stroke engine running at 2500rpm?

an infinite fatigue life is not really an infinite life in most cases anyway, due to corrosion, wear, and other factors which will change the strength and/or usefulness of the part as time goes on. ex: A piston ring may not snap in half, but if the face is worn off it's not much good anymore, is it?

 

[NormPeterson]

In the power plant piping business, the fatigue effects from postulated events having alternating stresses associated with 1E6 or more cycles may be ignored by Code stipulation (about 12500 psi for low carbon steels, BTW). Some work with steady state vibrations has been done out to 1E11 cycles and stresses around/slightly below half the above 1E6 value.

IIRC, Markl's fatigue testing involving piping tees and reversed bending also resulted in an inverse 5th power relation between alternating (bending) stress and fatigue life in cycles.

Norm

 

[NormPeterson]

Just to clarify, since editing existing posts does not work here.

The number of cycles for postulated events in power plants only range up into the ten thousand-ish range, hence there's no worry about hundreds of thousands or millions of cycles of some predictable event happening vs the 1E6 cycle 'limit'.

Norm

 

[GregLocock]

Wheels and engines both work to much higher number of cycles than that. The crankshaft on my car is at about 4*10^8 cycles, for example, and the wheels would be at about 10^8





Cheers

Greg Locock

Please see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[NickE]

Reed Valves in refrigeration/AC compressors are generally designed so that they see stresses well below the endurance limit.

 

[Tmoose]

Even parts designed for infinite life can suffer badly from (unexpected) stress concentrations. Wide journal bearings are likely to have higher edge loading due to misalignment. And "cylindrical" roller bearings often have a subtle crown to reduce edge loading to the overall benefit of service life despite the apparent reduction in contact. I'd expect the design of a 2X wide gear to have its stresses reduced by 1/2 only if supported on shafts and in housings designed, manufactured by genius wizards.

Current ball and roller bearing design encourages abandoning the infamous L10 life when the contact stresses are below some nominal value. But the fine print sets stringent requirements on cleanliness and the absolute attainment of EHD lubrication.

No comment on the relationship of endurance limit to stress levels. Especially when "good" stress inducing processes like shot-peening, nitiriding and prestrain can have such a remarkable effect on actual fatigue life.