Stress Classification at Cutwater of Sect III Class 3 Pump 2

[KiddT]

I'll lay out the useful information as best I can. The code that must be followed for this is Sect III, Div. 1, Subsection ND (Class 3). This is for a single volute casing pump (similar in theory to Fig. ND-3423-1). Let me also start by saying the functionality of these pumps isn't really in question (they already exist in service) only their compliance to the safety factors required by ASME code need to be proven. Nozzle loads exist but are a small component of the resulting stress which comes almost entirely from the 250 psi of internal pressure in the pump. This results in a level of stress that is acceptable through most of the pump except for a very high stress at the cutwater. I've attached a picture of the stresses at the cutwater to give an idea of whats going on (maximum principal stress as I believe thats the important one per ND-3416 note 17).

Now clearly looking at the stresses in that region they are well above S and S * 1.5 for local membrane and bending (Material is SA 351 CF 8M so S = 20 ksi). Firstly, the definitions of the general membrane, local membrane, and bending given in section ND-3416 all discount concentration stresses and only local membrane accounts for discontinuities. Because of the complex nature of the cutwater (two radii meeting at one point where the volute and discharge passage separate) I would think a significant component of the resulting stresses in that region would in fact be due to concentration stresses (peak stresses?) especially since the stresses are mainly high at the surface. So I'm not sure how much of these stresses even fall within the stress requirements laid out by ND-3416.

I tried to perform linearization of the stresses from the maximum point in ANSYS but how to pick a meaningful SCL in a complex 3D geometry like that was a bit difficult and the results ANSYS gave me indicated that the stresses were almost entirely bending and peak stresses were small, which surprised me. I'm not confident that those results were really meaningful though.

Basically my question is whether anyone has advice on how I can classify these stresses to confirm if this pump meets the code or not? I tried reading previous forum posts on the subjects and I can tell there's lots of great discussions that have gone on but a lot of it seemed to stem from Sect VIII Div 2 type analysis using plastic analysis or the categorization of secondary/primary stresses and peak stresses and I'm not sure that really applies for Class 3. Thanks very much for any thoughts.

 

[TGS4]

You're right that the responses will deal with plastic analysis. First, a quote from Section VIII, Division 2

For components with a complex geometry and/or complex loading, the categorization of stresses requires significant knowledge and judgment. This is especially true for three-dimensional stress fields. Application of the limit load or elastic-plastic analysis methods is recommended for cases where the categorization process may produce ambiguous results.

You have a situation where the elastic analysis will give ambiguous results. I would recommend following NB-3228 - it should be applicable to pumps as well.

There's no way around it - the geometry that you have is not well suited to elastic analysis ans stress categorization. Sorry.

 

[KiddT]

I sort of figured linearization may not be ideal for this case based on the geometry. Thanks for the quote that sums it up fairly well. I do still have a question though. You're mentioning a plastic analysis following NB-3228 but if I'm not mistaken for a Class 3 Component following Subsection ND, there is no equivalent to 3228 (section 3200 is not in the ND code at all I don't think). So would an analysis of this type really still be in following with a Class 3 component and thus be registerable as such or are you leaving class 3 criteria at that point? That's why I wasn't sure how much of the high stresses in that region fell into the three stresses mentioned in class 3 because they don't include secondary or peak stresses.

 

[TGS4]

Unfortunately, that's a little out of my area of understanding. Is there someone withe the NRC that you could discuss this with?

 

[prex]

Those local stresses are due to the mutual constraint of adjacent structures (as are for example the stresses at the junction of a head with a shell), so they are secondary in nature.
In fact the peak of stresses is so localized, that a plastic analysis would likely show a redistribution of stresses (strains), so that you would face a less steep peak.
However I'm not suggesting to perform such an analysis, also because your code doesn't allow for it; also by not doing that, you stay on the safe side.
I also seem to recall (from maaany years back) that Class 3 components are calculated much like ASME VIII Div.1 vessels, that is without consideration (nor limitation of course) of secondary stresses.
If you can confirm the last point, then you should simply disregard those localized stresses and satisfy yourself with checking the limits for general membrane and for local membrane. The latter may be obtained from your model: if you can have the membrane component of stress at the peak region <1.5S, then you are OK (otherwise I would say that you have a problem, and only a plastic analysis could come at rescue).
As there is no primary bending in a vessel, except at flat covers or faces, you can discard the bending component, as it is secondary.
If you can't maintain such a position, then you can check the primary+secondary stresses against the limit 3S, and if you can satisfy it, I would say you are OK.
And of course the secondary stress would be obtained by stress linearization along the most stressed segment through the thickness in your model. Don't worry if there is not much difference between the local stress at the surface and the linearized stress: we generally forget that a linearized stress may well be higher than the maximum stress at the same location.

prex

http://www.xcalcs.com

: Online engineering calculations

http://www.megamag.it

: Magnetic brakes and launchers for fun rides

http://www.levitans.com

: Air bearing pads

 

[KiddT]

My understanding is also that Class 3 components are similar in nature to the design basis of VIII Div. 1 and I also believe from going through the codes that secondary stress are not considered for those codes. I'm a little curious how it can be asserted that except at a flat plate or face bending is always secondary. I'm certainly not saying that's wrong but I wouldn't have known that for sure myself. There's no case where bending stress would occur in a non flat feature where that stress would not be self limiting and thus primary? Also I'm not sure I could use the 3S limit (or be tied to it) because while that is mentioned in VIII Div. 1 I did not see it as a criteria for Class 3 pumps.

 

[prex]

Div.1 explicitly recognizes (though I can't recall where) that high secondary stresses may exist in a vessel and that those are not limited. And the limit 3S is not in Div.1 (nor is the limit 1.5S for local membrane), but it is common practice to borrow it from Div.2 when a detailed analysis is performed.
Concerning the primary bending, you must recall that primary stresses are those that are necessary to satisfy the laws of equilibrium. Now generally a curved closed surface will be capable of retaining pressure by membrane stresses only, just like a balloon. However I can't affirm that this is rigorously true for any type of shape: an example may be an expansion joint that will only be capable of resisting an axial thrust by bending (but of course a joint will not be left to resist pressure axially). If you have doubts on the ability of your shape to resist pressure as a membrane, you should calculate a shell model using elements resisting as membranes only.

prex

http://www.xcalcs.com

: Online engineering calculations

http://www.megamag.it

: Magnetic brakes and launchers for fun rides

http://www.levitans.com

: Air bearing pads