Stress evaluation in a shell model

[Anonymous]

Hi Autodesk community

This might be the wrong place to post this, but I'll give it a shot.

I am trying to simulate an unreinforced tee with an in-plane stress condition. From this extract the appropriate stresses and determine the SIF. The maximum stress should be used to determine this. However since it is a sharp corner this will generate a singularity.

Is there a way to extract the correct maximum stress according to EN-13445-3 Annex C or ASME sec VIII div2 part 5? I think they are roughly the same.

As seen in the figure below there is a sharp corner present. This would create a singularity. The setup is similar to A.R.C Markl test which he made in 1950's. What's bothering me is that I have seen some articles about determine SIF from similar cases with a shell model and a sharp edge, but they do not clarify how they extracted the stresses and how they treated the singularity.

I have been considering using “hot spot stress” method but I understood it is highly approximate, but I might be wrong here.

Stress linearization is not relevant since it is plate elements and thus it is not necessary to extract any Pm ad Pb since they can be seen directly.

I have been considering using solid elements and employing a stress classification line (SCL). I will cite from their report. 

“However, an SCL or SCP may originate or end at a singularity, because the integration of the loads along the line or on the plane mitigates the effect of the singularity.” - Greg Hollinger and John Hechmer

But solid elements are quite demanding in comparison to shell.

Anyone bumped in to any similar problem and found a way to evaluate these stresses correctly?

/Alexander

[dharhay]

Wow, this is an interesting post for a change!

First, can you inform us of this 'Hot spot stress' method?

My 2 cents are that you need to make a solid model with sufficient mesh refinement to ensure that you are seeing the proper results.  Then you need to compare with your shell model's results.  There must be a work around for the concerns of Hollinger and Heckmer somewhere.

[dharhay]

I found this document from Solid Works.  Looks very much what you have described.

http://www.eng.uwo.ca/designcentre/FEA%20resources/116_Welds.pdf

Also, back a few decades the work of Bijaard had hand methods to determine moments for tees.  This may be a WRC Bulletin.

[Anonymous]

Hi Dave,  I'm glad that I have intrigued you with this post.

As for Hot spot stress method this is kind of the best I could find, Fatigue life assessment of welded bridge details using structural hot spot stress method, in chapter 2.3 you can read a bit about the method.

That link from Solidworks was very interesting. I will dig deeper into that.

I didn't want to mention it in my first post, that I am trying to determine a SIF without a weld. Meaning I would like it to be an arbitrary case in terms of welding giving the freedom for the welder/customer to choose whichever weld fits them the best.
Since the many weld assement methods are dealing with a kind of singularity in terms of a crack and sharp corners I though something similar can be applied in my case and "pretend" that my sharp edge has a weld.

Since there are some minimum weld requirements in ASME 3/4'' or 0.7t where t is the minimum thickness. I could use that, but in PVD in Autodesk does such a nice mesh and as far as I know you are not able to employ a weld in PVD.
As I have seen so far it feels like a matter of interpeting the codes than the FEA. Since some of the stresses do converge and some don't depending on the load case and which stresses you looking at.

PVD is a nice tool, really good for determine Pm and Pb for code design. But can you go deep enough to determine the SIF from PVD model?

//Alexander