Estimating Endurance Limit For Stress Ratios Other Than -1

[rcl5011]

I was wondering if anyone has any feedback on how to go about estimating endurance limits for steel in particular, when the stress fluctuations are other than fully reversed (R=-1). Fully reversed fatigue endurance limit is a fairly straight forward process following standard protocols on relating it to tensile strength of the material, and then adding various de-rating factors to it to arrive at an estimated endurance limit. This is only valid for fully reversed loading though. When stress ratios are other than -1, how do you go about coming up with the estimates of endurance limit? Obvious safest way to go about it would be to assume allowable stress would be at R=-1 regardless of whether or not the stress ratio is other than this. This would result in a safe design, but potentially overly expensive/larger than would be actually required. Assume I cannot fatigue test the actual material, and I need to come up with an estimate.

Thanks

 

[rb1957]

there are several equivalent fatigue damage rules in my business (aerospace).
R = -1 is a cycle +ve stress to -ve stress
R = 0 is a cycle +ve stress to zero
+ve stress is the same for both cycles … fatigue damage should be very similar (as compression has little impact for fatigue).
R > 0 … you need to go a long way (stress min almost equal to stress max) before the assumption of R = 0 is too conservative.

another day in paradise, or is paradise one day closer ?

 

[rcl5011]

I understand what you are saying about stress ratios greater than 0 are not all that conservative relative to R=0 unless stress min approaches stress max. What I am mainly concerned with is estimating these fatigue endurance limits for ratios other than R=-1. I can calculate what the R=-1 endurance limit is like I mentioned before. I do not have data on R=0, and I also do not have a way of calculating it. If someone has a method of estimating endurance limits of steel at R=0 then I would greatly appreciate understanding how to go about doing this.

Besides being handed a S-N diagram of the material with R=-1, R=0, and R=.5 for instance, I cannot determine with a reasonable accuracy what my endurance limit at R=.4 is for example, if all I have is the tensile strength of the material, knowing it is steel, knowing it's surface finish, operating temp, geometry, etc. Knowing these properties I can very easily calculate endurance limit for R=-1 (fully reversed loading) using the standard method of if Sut<200ksi, use .5*Sut for endurance limit, multiplied by de-rating factors. How do I get to the other ratios? Is testing the only method, or is there a correlation between all steel materials relating other stress ratios to R=-1?

 

[rb1957]

you can get some fatigue curves for aerospace steels in MIL HDBK 5 or AR-MMPDS-01.

I've attached a couple of pages that show that R = -1 is a very conservative model … surprisingly so.

another day in paradise, or is paradise one day closer ?

 

[rcl5011]

Thanks. I'm going to look and see if I can find some sources for fatigue curves of the material I'm interested in. It's looking like there's no way to correlate other ratios to fully reversed.

 

[rb1957]

I wouldn't say that (no way to correlate other stress ratios). There are equivalent stress models, that need to be tuned to the material.

But I would be confident in saying that R = -1 is the most conservative. So using your extreme stress and assuming R = -1 is most conservative. Using your stress range and assuming R = -1 is mostly likely (less) conservative.

another day in paradise, or is paradise one day closer ?

 

[mskds545]

The Goodman diagram is often used to account for mean stress greater than zero.

https://en.wikipedia.org/wiki/Goodman_relation.

There are many other variations on the Goodman approach. Of course, Shigley warns that "As more data were generated it became clear that a fatigue criterion, rather than being a 'fence,' was more like a zone or band wherein the probability of failure could be estimated...We also caution that it is deterministic and the phenomenon is not. It is biased and we cannot quantify the bias. It is not conservative. It is a stepping-stone to understanding." So keep that in mind too.

-mskds545

 

[rb1957]

and so we apply a high safe life factor (typically between 3 and 5; 10 if we can afford it)

another day in paradise, or is paradise one day closer ?