ASME Local Membrane Stress Distance

[MSL68]

ASME BPVC (Section III Appendix XIII-3100, or in older Code versions, this is described in the Definitions section) says "A stressed region may be considered local if the distance over which the membrane stress intensity exceeds 1.1S does not extend in the meridional direction more than 1.0*sqrt(Rt), where R is the minimum midsurface radius of curvature and t is the minimum thickness of the region considered.

The only example the Code gives is based on an axisymmetric shell (Figure XIII-1300-1), and this requirement makes sense for this type of geometry. However, what is supposed to be considered R and t in the case of a nozzle in a shell? Is the "region considered" the entire shell/nozzle junction (circled in red in the figure), where the R and t values come from MIN(r.shell, r.nozzle), and similar for the thickness? If that is the case, should that distance be considered for the entire 'region considered', as in sqrt(Rt) is the length from 'a' to 'b'?

Or, would the shell and nozzle be considered separately, where a sqrt(Rt) is calculated individually with their respective geometries, and that length for the shell is arrow 'a', and 'b' for the nozzle?

My thinking is the first interpretation is more reasonable, because when you think about a infinitely large plate with a small hole, you would expect the attenuation length of the local membrane stress to be quite small and the size/thickness of the shell has little bearing on how it attenuates.

 

[TGS4]

That distance is measured away from the SCL, using the value of R and t for the respective components. You cannot draw a valid SCL in the junction ring (transition region) - see WRC 429 for the definitions, so the only possible option is to measure away from the SCL that is located at the intersection of the structural region and the transition region.

So, your second option.

 

[RadiateurFou]

I think you have to exclude first the weld region, as stresses there would be considered peak. Then near the weld you will consider stresses as local.
I don't think having a defined distance of local membrane is useful, i think it is best to ask your friend engineering judgement here

 

[TGS4]

@RadiateurFou - whether you do or don't think that a defined distance is useful, it's in the Code, and therefore a Mandatory requirement.

 

[RadiateurFou]

Doesn't the code say "may"?

 

[MSL68]

@TGS4, thanks for the clarification. I believe I was thinking too much about it. Perhaps I'm wrong here, but it could be more thought of as "the shell's ability to withstand local stresses", in which case, it makes sense the dimensions should be of the shell in consideration, and not the junction as I initially thought.

@RadiateurFou, the Code says "may consider local if...". I guess if you would rather consider it Pm instead of PL, then sure. But, this 'may' is one where it's letting you classify the stress as PL, if you meet the criteria given.

 

[TGS4]

@MSL68 - you are correct that Section III, XIII-3120(b)'s use of the verb "may" is a permissive in the context that you state.

For additional clarity, VIII-2, 5.2.2.2(b)(1) removes the may and simply states that "a region of stress is considered as local if the distance...".

This locality definition has been in VIII-2 and III since day one. It is a consequence of the fundamental limit analysis approach and thin-shell theory.

In III, you can bypass all of this stress classification effort by using XIII-3200. In VIII-2, you can use 5.2.3 or 5.2.4.