SCL obsolete?

[TxAg78]

I've been away from ASME certification for "a few years". Things have certainly changed with the 2007 code! After reviewing related threads in this forum, I'm wondering what's the common practice for applying FEA in design by analysis Section III and Section VIII, Division 2.

There are essentially 2 methods:
1) Linear analysis with linearization along SCL
2) Limit load or elastic-plastic analysis methods

I am most familiar with linearization using a SCL. This is the "traditional" method I learned. It's deceptively simple, as it requires proper classification of each stress category (primary, secondary, etc).

Load limit is nice for those comfortable with nonlinear analysis. Create the model, define an elastic-plastic material, apply the required load(s), and let the computer do the work! Easy-breezy! (said a bit tongue in cheek)

So, having been away from the industry, what do most folks do these days? Is SCL now "obsolete"? If not, how does one choose between a linear or nonlinear approach?

I'm interested in a broad set of insights. Thx.

 

[TGS4]

Personally, I prefer the elastic-plastic approach. The limit load is OK for sizing-style calculations, but you can't use it for anything else. The elastic-plastic can be used for the local failure check, buckling, and now even fatigue and ratcheting.

I use elastic where it's dead-simple easy to use an SCL, or where non-linearities such as contact may come in to play (contact and factored load don't play well together).

Elastic-plastic is so nice because it does not require classification and categorization. It's more up-front work, additional computational time, and a huge savings on the post-processing.

My 2 cents.

 

[TxAg78]

TGS4,
Thanks for the feedback. I kinda figured that's what you'd say (based on previous posts).

I guess I'm a little confused by your nonlinear comments. I had the impression that load limit and elastic plastic were the same. Apparently not.

Is there a short explanation of the differences?

Also, I'm still curious to hear from "the rest of you" out there regarding SCL usage. I suspect it's still prevelant, but have no basis for my assumption. Thanks!

 

[TGS4]

Limit Load: Elastic-perfectly-plastic material with linear geometry

Elastic-plastic: elastic-plastic true stress-true strain curve with hardening and non-linear geometry.

You can't use the limit load for the local failure check or the buckling failure check. in fact, other than the plastic collapse failure mode, limit load doesn't have any other uses. Not to mention that the strain results (and deflections) from a limit load analysis have no physical meaning...

 

[TxAg78]

TGS4-
Thanks for the clarification. I assume by "nonlinear geometry" you mean "large deformation/strain" (and "linear" == "small").

I see why you prefer Elastic-Plastic. From a FEA standpoint, these are pratically the same. By that I mean an industrial strengh nonlinear solver should be able to solve either problem. There's only a little more data input required for the analyst to run as elastic-plastic. In fact, perfectly plastic material props sometimes run into stability problems you don't see when there's strain hardening.

Thanks!