Filled PPS Creep Modeling

[Mike Guido]

Hello to all...it occurs to me that I tried to post this to what was probably the wrong forum ("Downloads"). My apologies to the moderator...

I am an experienced (>20 yrs) FEA user, but I have never performed a creep analysis. I have a consulting client who wants to see the long-term behavior of their product under load; i.e. they want us to quote them on doing a creep analysis of their component. The component is molded from 40% glass- or mica-filled polyphenylene sulfide (PPS) under long-term exposure to water, at temperatures (obviously!) below 100C.

I have started to peruse the literature, including the ANSYS Documentation about creep modeling. Before I start purchasing reference materials, does anyone have some suggestions about which creep model might be appropriate for this type of material? The projected lifetime of the component is in the 10-to-20-year range. The manufacturer has some creep data on this material, but the only published information is for times less that 10,000 hours--they may be able to provide more extensive data upon request.

I know that the standard caveats apply when applying material data over a range greater than the range where the data was acquired. Other than comments/answers re: the questions above, any other information would be welcome.

Thank you,
--Mike

[Muzialis]

Mike,

it is difficult to give you an advice without knowing more.I have performed quite creep analysis of elastomeric components and in first approximation results are very often decent using linear viscoelasticity.

I personally would give it a go with the data you have assuming the materials in question have a decaying creep as rubber: otherwise (power law creep or similar) I would not extrapolate and would ask for further info.

Then there is the temperature issue, and I sincerely do not know much about the materials you mention to claim they are thermorheologically simple.

Have a nice weekend

Muzialis

[Mike Guido]

Thank you, Muzialis.

In this case, the material is a thermoplastic with a high percentage of filler, so it does not act much like an elastomer at all. Its stress/strain curve is very linear up to the point of yielding, and there is very little elongation exhibited between yielding and failure. The material does of course show greater creep deformation over time at higher temperatures than at lower temperatures.

As someone with no experience in the characteristics of the different standard creep models, I was hoping to get some advice as to which of the creep models I might try when fitting the data from this generic type of material.

Thanks again,

Mike

[Jorgen]

I agree with Muzialis that the first model to try is linear viscoelasticity (LVE).
The accuracy of LVE will depend on the magnitude of the stress that is applied, and the range of temperatures that the plastic is exposed to.

How large is the maximum stress relative to the yield stress?
What range of temperatures is the component exposed to?
What is the glass transition temperature of the filled PPS?

- Jorgen

[Mike Guido]

Jorgen, thank you for chiming in. Some additional material information:

I don't see any data for glass transition temperature for the material, but the manufacturer gives a Heat Deflection Temperature in excess of 500F and a crystalline melting point of about 285°C (545°F).

They also have some creep data for temperatures of 75F, 150F and 250F at the manufacturer's website (page 9, 10) (http://www.cpchem.com/enu/docs_ryton/RytonMechanicalProperties.pdf). These curves only extend out to 1000-2000 hours, whereas the customer wants to know what happens at service lifetimes between 10 and 20 *years*. Based on a static model of a preliminary design, it appears that some parts of the component could experience stresses between 50% and 80% of yielding. This material exhibits very little elongation, so the difference between the yield point and failure is probably insignificant.

The component will be subjected to continuous exposure to hot (not boiling) water, so the curve for 150F is probably the most appropriate one.

Thanks again and best regards,
--Mike

[Muzialis]

Hello Mike,

I do not have any absolute advice for you but I will gladly share my opinion.

The reason for which I would be reluctant to perform a creep analysis using data for 2000 years and projecting the results over 20 years is the change in material properties due to ageing at such a high temperature. This is likely to be much more appreciable than the error the extrapolation would lead to, I think.

Personally I would prefer to have accelerated ageing testing performed: in the most fortunate case the material ages according reasonably to Arrhenius type relationships testing over 1, 3 and 5 months would give you a very good estimate of the elastic properties over 20 years. Having these you can evaluate the meaningfulness of any analysis over the long term.

All the best

Muzialis