safety factor and linearized stresses 3

[LIAM_7]

Dear all,

I am new in ASME BV FEM based analysis.

I have a couple of questions:

1. Is it mandatory to go for elastic-plastic analysis if R/t<4?
2. I have carried out a elastic analysis in a situation of R/t=1,4. The results based on linearized stresses allows the tank to pass (according to Pm, Pb...rules specified in the norm). But when I check Von-Misses stress plot and Safety factor, this is below 0,5. Does it makes sense? I mena, is it possible for a case to pass the criteria of ASME but fail in terms of safety factor based on Von-Misses stress? Remark that maximun allowable stress of the material has been obtained from values provided by ASME BVPC.

Thanks in advance for the advice.

Liam

 

[DriveMeNuts]

Read Section 5.2.1.3 of ASME VIII Div 2. It answers your questions.

The elastic method relies on there being uniform linear through thickness membrane or bending stresses. For R/t<4, this doesn't happen and the result often becomes non-conservative, therefore ASME recommends using a different method.

 

[TGS4]

To answer your questions:

1) Yes - it is mandatory. The explanation is in 5.2.1.3
2) See the explanation in 5.2.1.3. Furthermore, your explanation of what you are doing does not make any sense.

Just a side-note - there is no such thing a a "safety factor" in Part 5. We have design margins that depend on the analysis method used and the failure mode under consideration. This is an important distinction.

 

[LIAM_7]

Thanks for the answers DrivemeNuts and TGS4.

I would like to explain better: when I talk about Safety Factor I do not refer to the norm. I am talking about the safety factor contours provided by ANSYS Mechanical. And for the loads applied (850 bar+hydrostatic pressure), the Safety factor is below 1 in a reasonable region of the tank shell (which, without a norm, I will consider a fail in the structure). But in the same analysis, if I define paths across the thickness of the structure (in the higher stresses area), I linearize the stress and criteria for Pm, Pl, Pl+Pb...is fulfiled (below the allowable stress specified in the norm, if I select it right it is 1.37e+08 for AISI 316 steel).

As a beginner probably I am not understanding something, but on one side intuition (safety factor) says that part fails and norm criteria says it will pass....

So TGS4, do you mean that I have to forget about the "typical" safety factor and only consider ASME criteria (plastic collapse, local failure...)? Or is there a relation between what you obtain in ANSYS Safety Factor and norm's criteria? If norm criteria is satisfied, should also occur that SF>1 on the entire domain?

Thanks for your support guys, really apprecciate it.

 

[TGS4]

If you are working on pressure equipment for which ASME Section VIII, Division 2, Part 5 is applicable, you can forget about everything else except what's in the Code. I have no idea what ANSYS does or what input you have to provide to get this so-called Safety Factor plot - but it has no relation to the Code, so cleanse it from your memory.

 

[LIAM_7]

Thanks for all the support.

Finally, I got a converged solution (this is supposed to be enough for one of the criteria specified in the code).

I realized that I have imposed loads for operating pressure, hydorstatic pressure, fixed support and gravity, but it seems that the code multiply this loads by a factor. If I do so and I get a converged solution using stress-strain curves, non lnear analysis I assume that the part still fulfill ASME criteria. It is hard to understand why a converged solution means is the only requirement. What will happen for example if I get a convergent solution with a lot of plastic strain? Criteria is still fulfiled?

What I am trying to say is that I do not understand the convergent criteria forgetting about numerical results provided by this convergent solution.

In any case (and despite it is not needed), my safety factor for this analysis is over 1 and also plastic strain is very small...

 

[victorpbr]

LIAM 7

I think other experienced members may give you better advice, but here is my 10 cents.

It's not about ONLY achieving convergence. You have to fulfill all the requirements of the part. When you multiply your loads for a factor and achieve convergence it means that your equipment maximum strain is within the material stress-strain curve, hence it did not reach the material limit specified, otherwise it would not converge. Make sure to get your material data correctly, with the correct stress strain curve obtained from ANNEX 3-D, and to select the proper material behavior in your software.

If this is a real equipment to be designed I would also advice you to seek a more experienced designer, you are dealing with a very high pressure (850 bar) and thick cylinders, there are lot of factors to be considered...


Victor
from Brazil

 

[DriveMeNuts]

Remember, you also need to comply with the 'local failure' requirements of the code.
It really does sound like you are winging it and need a mentor or need to attend an analysis course.
Do you even know how to classify stresses for an elastic analysis?

 

[LIAM_7]

Thanks for your comments and your advices to search a mentor. Unfortunately, I am not able to get one, si I need to learn this by myself, so I really appreaciate your comments. I have read about 20 papers on the topic just to get informed, but still have some doubts.

I am doing an assessment based on local failure, plastic collapse, fatigue and buckling. As far as I know, these are the 4 criteria established in the norm for a pressure vessel DBA.

I started with linear analysis (decomposing stresses in membrane, bending, peak along a SCL), but as I mentioned, the R/t makes the linear elastic useless for this case.

I created material S-S curve based on the rules specified in the norm. Loads in the simulation have been scaled accordingly. The simulation converges.

So:

1. Protection against plastic collpase seems to be OK.
2. Protection against local failure is assessed according to triaxial strain criteria. It passes. But for example, what will happen if the part is manufactured by machining instead of forming? cold forming strain is equal to 0? Or other norm should be used?
3. Protection against buckling collapse: eigen value buckling analysis has been carried out based on pre-stressed result from previous non linear analysis. With 4 modes computed, I obtain negative values. Are these values OK for comparison with 1.667/Bcr as the limit stated in the code for a elastic-plastic analysis?
4. Protection against fatigue will be carried out later.

Many thanks for your kind comments. As I remarked before, I am a beginner and probably everybody here has been a beginner in the past (and I am sure you had to learn a lot of things by yourself )