Inconel 718 creep equation from MILHDBK5H to FE form

[jas12345]

hi
i've asked this question a few places & had no reply, so
maybe someone on this forum can help.
i want to perform a creep analysis in Ansys.

i have the following creep equation for Inconel 718 taken from MILHDBK5H (Table 6.3.5.1.7(a)-(e)),
Log t = B1 + B2.T + B3.X + B4.X^2 + B5.X^3

where B1,B2,B3,B4,B5 are constants, T = deg R and X = Log y = Log (stress, ksi), so that

Log t = B1 + B2.T + B3.(Log y) + B4.(Log y)(Log y) + B3.(Log y)(Log y)(Log y)

the curves in MILHDBK are for isothermal creep for 0.1%, 0.2%, 0.5%, 5.0% and creep rupture.

how do i get this in the form for one of the standard implicit creep equations in Ansys such as

cr_dot = C1 (y^C2) (cr^C3) (e^(-C4/T))
where C1,C2,C3,C4 are constants, cr_dot is creep strain rate, cr is creep strain and e is natural log base.

can this be done?
Should I set up a User Creep Subroutine (which I have never done before & I’m no expert in creep analysis!)?
Does anyone have any creep laws for Inconel 718 that they can share or recommended?

thanks
jas

 

[btrueblood]

There is no easy, direct method to invert the fitted equations from the handbook into the form your solver wants for a creep model. You have time to a known strain as multiple functions (at various known strains), the multiple functions are polynomial wrt stress...bleah.

There are several plots in the handbook, and (in the version H hdbk copy that I have) they are pretty clear, so that one could pretty easily interpolate the data points, and re-fit them into the model form your solver wants. In doing so, you may find that a single model poorly fits the data across the entire temperature and/or stress regime in the data, and that you will have better fit to the data by breaking the curve fits into piecewise continuous functions over temperature and/or stress or some combination of the two. Something else to note and be careful of, is that I believe the T in the Ansys model is absolute temperature (i.e. deg. R or Kelvins), not deg. F like in the data, so be careful you get the conversion right when fitting.

The good thing about the ANSYS creep model is that it is unlikely to "blow up" and give you ridiculuous results when the temperature drops below the level shown in the test data. Try plotting the MIL-HDBK-5 curve fits at low temperatures and/or stresses below the levels of the curves shown plotted, and see if they make sense when the conditions are out-of-bounds.