Continue to Site

Eng-Tips is the largest engineering community on the Internet

Intelligent Work Forums for Engineering Professionals

  • Congratulations GregLocock on being selected by the Eng-Tips community for having the most helpful posts in the forums last week. Way to Go!

SHELL MESH FEA FOR STRESS LINEARIZATION

Status
Not open for further replies.

Doodler3D

Mechanical
Jan 20, 2020
188
All,

I was looking into papers to get more information on ASME VIII DIV 2, 5-A.4.2.2 Stress Linearization Procedure from a learning perspective. However, there are very few cases of the method being used, both in papers and relevant forums. Are there fundamental flaws in using shells or is it a standard procedure to perform stress classification/linearization using brick elements or tets?

Thank you.





 
Replies continue below

Recommended for you

Shells are inherently linear through-thickness. So the results from a shell analysis already have the linearization completed. They are perfectly acceptable.
 
Thank you TGS4. I was wondering why shell mesh linearization is not popular. Most reports have 4-8 elements thru thickness for SCLs. Early career FEA naivety.
 
Software like NozzlePro uses FEM using shell elements. Like TGS4 says, the linearised result is taken directly from the FEM result. There is no long-winded linearisation procedure.

With respect to choosing between brick and shell elements, I prefer to use bricks wherever possible because I find it difficult to identify the t/D where shell elements result becomes poor quality. With a simple shell of 2 metres diameter and 40mm thick, I tend to use bricks because I can get 3-4 elements through-thickness without needing to have millions of elements (resulting in a short solve time). For a shell of 2 metres diameter and 5mm thick, using shell elements is the better choice. It will give a good quality result and you won't have to wait forever for a million element result.

Another way I look at it, is you can do a sensitivity study using bricks to get a result which converges (by doing multiple runs with increased through-thickness elements). Using shell elements, I'm not sure how you do a sensitivity study. Simply increasing the density of elements around a discontinuity will lead to some sort of convergence, however it doesn't necessarily result in a good quality result, especially if there is a component with a larger t/D ratio. Quality is difficult to judge.
 
@DriveMeNuts

Shell analysis gives us membrane, bending but not peak stresses. However, ASME VIII DIV 2 2019, 5-A.4.2.2.(c) Stress Linearization Procedure describes the peak stress tensor. What does that imply?
 
A linearization procedure could be used with shells if, for example, you wanted to extract the membrane and bending stress across a stiffener / gusset width, not through thickness obviously.

@DriveMeNuts, you didn't classify what type of analysis for the justification for using solid elements over shells, which seems to be the most important factor IMO. Regardless your brick model will still be far more inefficient that running shells. You check convergence in the same way, so for any method using 'structural stresses' where an elastic analysis is still valid, you can get good results for most geometries.
 
@BJI.

I'm going by previous consultant's reports. They predominantly used shells and classified ALL stresses (Pm, Pl, Pb, Q, F) through their own methods (I can't figure this out). This was in way back in 2004-2006 so things have come a long way. However, the consultants were on the ASME research committee at some point in their careers so one can't refute their engineering judgement.
 
@DoodlerDaru, I wasn't aware I was refuting anything, what in particular do you disagree with? I didn't mention anything about checking code limits. However, if I disagree with any engineering methods or judgement, I am entitled to refute it (and obligated to in many cases), ASME committee member or not.
 
As a member of the ASME Code Committee responsible for the DBA rules, I say refute away. The Committee members are just regular engineers, with a passion for this stuff. There's nothing special about that group (myself included).
 
@BJI the latter part of my comment was not directed at you. I apologize if that sounded harsh.

 
Status
Not open for further replies.

Part and Inventory Search

Sponsor