How can Rupture Allowable Stress be Less than Creep Allowable?

[PSSC]

Hi guys,

I hope the collective knowledge around here can raise me up a little.
I have come across some cut sheets for different high temp casting materials (specifically A351 HK40) where the rupture strength is listed as less than the creep strength.
I thought that rupture testing is to failure at a constant temp, and creep was a measurement of strain (elongation) at a controlled temp for a controlled time.
If this is true, then the specimen would break before it bends.
Which is fine if it is very brittle.
But the literature does provide a higher creep strength at the same temp that the piece has supposedly already failed at a lower stress value.
So I am confused and would greatly appreciate some clarity on this.

Thanks

 

[metengr]

For stress rupture testing the material will creep at some accelerated rate (x%/hr) and is taken to failure because this the specified test. The stress at failure is only reported because some designers want this known stress level to cause rupture, and not worry about creep rate.

The rate of creep deformation is a defined test method where strain is measured as a function of time based on a pre-selected temperature and stress (which could be higher or lower than the stress to rupture in x hours). Typically, stress to failure is not reported, only x% of deformation and or x%/hr of creep rate.

 

[PSSC]

Thank you for your reply.
I think I understand the testing procedures, what I had thought happened is inline with what you explained.

Here is an example of my confusion.
Per Kubota Metals for A351 HK40

Rupture at 1700F for 10,000 hours occurs at 2.67ksi
1% Creep at 1700F for 10,000 hours occurs at 3.35ksi.

So in a test for 10,000 hours the material went to rupture (failure) at 2.67 ksi, but only elongated 1% at 3.35ksi?
This seems to me to be like saying the Tensile strength of a material is less than the Yield strength.

I am sorry if I am missing something obvious, but I am definetely missing something.

 

[metengr]

The actual test data you posted does seem rather odd in behavior. It is possible there was an error in the published data - seen it before and it will happen on occasion.

I would contact Kubota and inquire as to their published test values. Normally, the stress rupture value will be much higher at the same temperature in comparison to creep rate. The stress to rupture may be a misprint because of unit conversions - you never know.

Here some information I have on HK40 from another publication

Cast Heat Resistant Alloys from NiDI

1800 deg F

Stress (psi) Rupture Time (hr)
4800 100
2800 1,000
1700 10,000

 

[metengr]

Part 2;

At 1800 Deg F, the creep rate for HK alloys is

2500 psi for 0.0001%/hr. This does not mean failure, only an established creep rate between 1,000 and 10,000 hours as indicated in the post above.

 

[blacksmith37]

They measure two different things; deformation and fracture. For some applications ,deformation limits service life. For others like furnace tubes ,fracture/leakage defines life. Over the decades more design is based on rupture tests/data as they are cheaper to run, and if rupture occurs before reaching the creep design life ( as may be possible in the furnace tubes your are working with) it is very dissapointing.
And in Cr:Mo type materials , the amount of strain at rupture decreases at longer times.

 

[PSSC]

I guess my question is this:

If you have a sample piece of material, and you hold it to 1800F for 10,000 hours and the most stress you can put on it before it fractures is 1700psi, then how could you ever put 2500psi on it?
It would break at 1700psi.
You can't bend/elongate that which is already broken.
It broke at 1700psi, or is what you guys are saying is rupture does not equal "broken".

It may not seem like it, but I am really trying to wrap my head around this.

 

[metengr]

PSSC;
Look at the data I provided again

You have rupture at 10,000 hours at 1800 deg F using a fixed stress of 1700 psi. During this time the material has been subjected to increased deformation rate because of accelerated creep damage in this type of test.

You now increase the stress level, knowing the rupture time will be lower than 10,000 hours, but instead you decide to measure the actual rate of strain (deformation) over a defined period of time prior to rupture (which in the above case would be between 1,000 and 10,000 hours.

As mentioned above, some designers use creep deformation rate to decide on failure versus stress rupture.

This is all it is.

 

[PSSC]

Ok, I think I get it.
After the strain is recorded at a lower time span, then the stress is extrapolated out to 10,000 hours?

 

[metengr]

After the strain is recorded at a lower time span, then the stress is extrapolated out to 10,000 hours?

Perhaps this is where the confusion lies. No, these are two independent tests (not combined). Also, you are not dealing with linear functions, these are power functions used for predicting creep behavior regarding time, as a function of stress and temperature.

 

[Compositepro]

There are many materials that have a stress/strain curve with a peak that then decreases to failure. often this is due to necking and the stress not being corrected to the true cross-sectional area. There are uses for this data but you do have to fully understand what is being measured, and how.

 

[blacksmith37]

Before 1970, some creep data was developed with constant stress. The beam of the creep machine had a "tail" that reduced load as the test bar stretched- I was told it wasn't worth the effort.
For extrapolating, The Larsen-Miller parameter was pretty good; with one stress rupture point , you could make reasonable estimate for other conditions.

 

[PSSC]

Ok, taking the rupture stress measurement after necking has occurred, and this reducing the stress level makes sense to me.
This I can visualize.
The creep stress level is recorded on a test specimen that is kept in the second phase of creep.
The final rupture tension is recorded right before it breaks, necking has occurred so it takes less tension to break it than it took to get it to the point of failure.
I remember this from college, but completely forgot about it when looking at this data.
I just knew it looked odd to me.

If I understand, I would like to thank you all for helping me.
If I don't please don't give up on me.

 

[davefitz]

Don't forget the add'l safety factors applied due to melt-melt variability , test data scatter, and other affects that are addressed in the code allowable formulation. Not to mention constant lowering of allowables based on long term teting as oppposed to the initial short term creep test.

 

[metman]

The stress/strain curve for mild steel has a temporary reduction in stress after the yield point. So I imagine the curve for the given material at temperature to be similar except failure at a stress below yield stress. Or maybe this is a poor analogy because it can only be understood as two separate tests?