accelerated creep testing in Ultem

[subsearobot]

Hello folks,
I am reviewing a design that uses unfilled ultem as a structural component with a sustained load.
Cheap FEA (Nastran, linear) predicts low strains in the bulk at around .2%.

This part, however, requires very high physical precision, so if it creeps at all, could knock out alignments of components.

I would like to find a thermal schedule that we can use to run an accelerated test (5 year or perhaps 10 year simulation). (TSSP)?

any assistance would be appreciated. I'm sure you can tell i am not a polymer scientist!

thanks

 

[rickfischer51]

I assume you are talking about a test on a finished part with typical service loads. The rate of thermally activated processes in polymers (and a lot of other stuff) typically follows an Arrhenius distribution for temperatures below the glass transition temperature, so in theory this could be used to scale a creep test on your component at an elevated temperature. The trick is to determine the activation energy and scale factor. I found this creep curve for Ultem 1000 in 30 seconds by googling. Note that the curves are at different temperatures and stresses. You could try plotting natural log of the creep rates (the slope of the curve after the knee) vs 1/T which is the absolute temp in Kelvin, to get the activation energy and scale factor, then use those to predict the test test duration at your desired test temp. This would likely give a reasonable first order approximation. Keep in mind that you want to keep your test temp below the glass transition temp. Polymer behavior changes drastically above T[sub]g[/sub].

Rick Fischer
Principal Engineer
Argonne National Laboratory