linearization and stress classification 1

[freschif]

Dear All,
I have some few questions about linearization and stress classification according ASME VIII Div 2 (2013)- Part 5 - Protection against plastic collapse. I kindly ask your help to solve my doubts.

Premise:
At paragraph 5.2.2.4 of the Code - Step 1 is written that: "... separate load cases are analyzed to evaluate "load-controlled" loads ... and "strain controlled" loads..."
At paragraph 5.2.2.4 of the Code - Step 2 is written that: "At the point that is being investigated, calculate the stress tensor (...) FOR EACH TYPE OF LOAD. ASSIGN EACH OF THE COMPUTED STRESS TENSORS TO ONE OR GROUP OF THE CATEGORIES DEFINED BELOW. ..."
At paragraph 5.2.2.4 of the Code - Step 3 is written that: "Sum the stress tensors (stresses are added on a component basis) ASSIGNED TO EACH EQUIVALENT STRESS CATEGORY..."
In table 5.6 is shown that different source of stresses could lead to different classification

Questions:
1) On FEA software should I keep separate cases and results for each load (e.g.: 1 case for pressure; 1 case for external loads; 1 case for thermal loads; 1 case for liquid weight; ...; and then perform LINEARIZATION SEPARATELY and CLASSIFICATION SEPARATELY and only after sum for all loads (on component basis) to proceed with Von Mises stresses and comparison with allowables?

2) alternatively to question 1 above: On FEA software should I keep separate cases and results for "load-controlled" loads and "strain controlled" loads", that means that I can group in a single case loads as pressure, external loads, liquid weight, ... and in a separate group loads as thermal loads; imposed displacements, ...; and then perform LINEARIZATION SEPARATELY and CLASSIFICATION SEPARATELY and only after sum for all 2 CASES(on component basis) to proceed with Von Mises stresses and comparison with allowables?

3) in the second case how could I make classification according to table 5.6, if this requires to separate stresses from different origin?

4) in all 2 approach is correct to calculate Von Mises equiv. stress relative to the sums Pl+Pb; Pl+Pb+Q; Pl+Pb+Q+F starting from tensor components (sum on component basis of Pl with Pb, etc) and not just as sum of Von Mises Equiv. stresses of Pl + Von Mises Equiv. stresses of Pb + Von Mises Equiv. stresses of Q, etc?


Additional question:
I normally use for FEA UGS-Siemens NX with Nastran.
5) Do you know if exist some tool/routine to help in linearization for this software?

I know that some of these points have allready been discussed in previous topics, but apparently I didn't find precise answer to my questions.

These are the topics that I read:
thread794-115532
thread794-280606
thread794-254439
thread794-279233
thread794-342622
thread794-349076
thread794-358058
thread794-357511
thread794-348684

Maybe you can suggest me other interesting topics I could have lost.

Federico Freschi

 

[TGS4]

1) No. The results for the load case combination is sufficient.

2) That is not necessary, although some people do it that way.

3) You don't make classifications to Table 5.6. Those are some examples (note the title of the Table), and it is the responsibility of the analyst to make the final judgement for the classification. However, different loss do cause stresses of different classification. Doing as you suggest with 2) simplifies that process.

4) You are correct. However, you will note that for Protection Against Plastic Collapse, only Pl and Pl+Pb are required. The other combinations are covered under different failure modes which have different loading scenarios.

5) I am not familiar with that software. I generally don't trust the FEA software to perform the linearization in accordance with 5-A.4.1.2, Step 2 (a). I wrote a MathCad sheet to perform the linearization for me.