Temperature Dependent Material Properties in HyperFEA

[bacjac]

I want to assess the effect of high temperature property degradation on mechanical load in a multi component beam model. I have Tinit at a high temperature , 1300F, and Tload at a slightly higher 1301F. This is to place the components at a high temperature state with minimal thermal load generated. I Run a nastran deck with Tinit above the subcases, subcase 1 with mechanical load sufficient to yield some of the components which are at a high temperature.  Subcase 2 with the thermal load. Then size HyperFEA. All components size to the room temperature allowable of IN718, though several components are overstressed at their actual Tinit, Tload temperature.   Does HyperSizer account for material property degradation for components individually?

In 718 is temperature dependent with about a 50% reduction in Fty RT to 1300F.

[James]

Yes, HyperSizer will read and average element and nodal temperatures from the input file, and use them as analysis temperatures for each component. You have to run the thermal subcase separately from the mechanical case and combine the thermal and mechanical sets together into 1 load case.

Here's a brief help topic on the subject.
http://hypersizer.com/help_7.3/#FE-Results-Import/results-thermal_loads.php

Also keep in mind thermal help/hurt factors. See:
http://hypersizer.com/help_7.3/#LoadSets/ls-factors.php

The temperatures should be defined as element or nodal. The 'initial temp' is not used, instead it's the applied temps. Assigning temperatures with at TEMPD card is supported.

I hope this is helpful.

-James

[bacjac]

I believe I tried that James, but I'd better try again to see if I did it right.  Thanks. Will let you know.

[bacjac]

Say, can the thermal and mechanical subcases have the same ID, since they are in separate run decks. I seem to remember something about that being an issue?

[bacjac]

Ok, found an error, missing an include file in one of my attempts. Got closer to what I think should happen.  Can you tell me James what HyperSizer uses for a temperature to get allowables, for a component of multiple elements, where each element is a different temperature. Is it an average or the max temperature. Appears in the "Stress" tab the max temperature is used. Ok, just good to know what the philosophy is. 

[James]

If you're running element-based or peak-element load processing metrics, the peak is used.
If you're running a n-sigma statistical method, the area-weighted average is used.
Take a look at the FEA loads tab, it should include some information about the temperature averaging.

-James

[bacjac]

that would be good. Somehow I'm not seeing the higher temperature in the Loads Tab. I size the component from that tab, either with a sigma or element based but both still show this component at 70F when I believe it should be 1300F from the Tload callout in the thermal rundeck.

[bacjac]

Still not sure if my conclusion is right James. Appears I don't get the expected behavior. Do you have an example that shows the process working ?

[bacjac]

OK James, thanks for the explanatory material I have finally got things sorted out right on my end and see everything working correctly.   Whew - sorry for the inconvenience. pdf attached shows the case checks that seem to be working correctly now.