FEA Stress Results At Notches For Grey Iron Castings

[sdra2]

I have a number of questions regarding stress analysis of irregular geometry in grey iron castings.

(1) How can I evaluate the theoretical stress concentration factor (using FEA techniques in my case) for irregular geometry? I have read that for, say, a simple bar in tension with a notch, comparison of the notch root stress with the "core" figures gives the theoretical stress concentration. This makes sense as an illustrative example, but can such a method be applied to more complex shapes and, if so, how? How can a "good" figure for the nominal section stress levels be obtained? A "mid-plane" figure or something similar? Is this a dangerous general approach because it is very geometry dependent?

(2) As an alternative to (1) above, is there any value in running analyses with varying notch sizes? This will plainly give rise to a trend of stress concentration data, but without a base case to compare this against, what purpose would such data serve?

(3) Assuming that somehow I obtain a figure for the theoretical stress concentration factor (Kt), can I correct for statically loaded grey iron using Kf=1+q(Kt-1) like I might for a fatigue loading situation?

(4) If (3) above is valid, at what point does it become invalid? As Kt is based on elasticity theory, does this idea become invalid at stress levels beyond the proportional limit of the material?


What I am basically trying to understand is: what is it that constitutes a design stress limit in a Finite Element model for a grey iron casting?

This could be an as-cast surface, or a machined surface.

If the FE model is reporting a high stress in a notch or radius, can I factor this reported stress figure down because of the "notch insensitivity" of grey iron, or is it simply the case that as there are many discontinuities within the material matrix anyway, it is not valid to use Finite Element or indeed any stress evaluation techniques to evaluate the suitability of grey iron where (macroscopic) notches are present?

Is there a sound design methodology for evaluating the suitability of grey iron castings for a given duty at all, or is it that there are so many uncertainties due to the nature of the material that only physical testing can suffice?

 

[GregLocock]

Well, why not use FEA to generate a stress field in the vicinity of the notch (without modelling the notch itself), then apply Kt to those stresses and the expected notch size?



Cheers

Greg Locock

Please see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[sdra2]

Thanks for the reply Greg.

One of the problems that I have is that the irregularity of the geometry I am looking at does not lend itself readily to comparison with any tabulated figures for Kt: another is that I am uncertain when, and to what extent, Kt is valid for grey iron anyway. As far as I am aware, a reasonably constructed FEA model will report stress figures at notches that approximate use of full Kt values for elastic materials - is this correct?

What I am looking at is analagous, say, to an engine block with ribs inside the coolant passageways, and I am trying to ascertain the importance of reported high stresses at radii where these ribs link to other walls.

If I remove the radii, will I not be looking at a stress field generated at a "singularity", thus generating a false stress field?

Similarly, if I remove the rib entirely, then surely I won't capture the structure's stiffness correctly within the model.

Please pardon any idiotic statements or questions as I am fairly new to using FEA.

 

[GregLocock]

"If I remove the radii, will I not be looking at a stress field generated at a "singularity", thus generating a false stress field?"

Step back a bit. Don't worry about the stress actually at the singularity, model the stresses at a boundary say 30 mm (for a small engine) away from the singularity. Then go and look up an analagous case in a stress analysis book, using that geometry. However you've said you can't do that, so that's that idea blow out of the water. St Venant's principle says (roughly) that the local errors in the stress field caused by bad modelling near the notch won't affect the loads 'remote' from the notch. Goodness knows what distance 'remote' is, several elements and several wall thicknesses at a rough guess.

Does the high stress at the rib matter? Yes. Probably. If it really exists and is cyclic then almost certainly.


Cheers

Greg Locock

Please see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[TGS4]

What is the ultimate use of the casting? If it is for pressure vessels/piping ten there are rules in the appropriate design Codes and Standards that can guide you.

 

[sdra2]

TGS4

The casting is not for a pressure vessel, but it is certainly possible that pressurised grey iron castings with bracing may be covered by ASME VIII or something - will take a look.

Thanks.