Hair pulling and torispherical heads 4

[roadapple]

Concerning the rules in Sect. VIII, Div. 2: for torispherical heads.

Step 2 page 4-50

0.7 <=L/D<=1.0 (eq. 4.3.5) or go to Part 5.

I am designing a vessel out of 8 inch pipe with torispherical heads.
The D=7.625 in and the L=8.625.
Because the inside crown radius equals the outside diameter of the skirt (UG-32 (e)) and the D equals the inside diameter of the head (shcd. 80).

I know I am making a mistake and am willing to air it out in public - but by my numbers these will never satisfy the equation 4.3.5 as it will always be larger than 1.

Can someone please shine some light before I am completely bald?

roadapple

 

[SnTMan]

roadapple, I know that many head manufacturers sell both Code and non-Code torispherical heads and a Code head must be specified when needed. For pipe size work, why not just use a pipe cap?

Regards,

Mike

 

[roadapple]

Thank you first of all for your response.

If I use a pipe cap - questions:
1. Where can the "standard" cap dimensions be found?
2. How can the 0.7 <=L/D<=1.0 ever be met?
3. Does a pipe cap fall under torispherical or semi-elliptical?

If I am assume that the pipe cap is torispherical. The calculation for L/D - D is the inside diameter and L equals the outside diameter - so the L/D will always be greater than 1.

roadapple

 

[SnTMan]

roadapple, pipe cap overall dimensions are found in ASME B16.9 an in numerous hand books. They (and all B16.9 fittings) have the same pressure rating as a straight pipe of the same material and wall thickness. See UG-44.

Therefore you don't care whether they are torispherical, elliptical, or some other shape. No calculations for the cap are required. Makes life simple.

Regards,

Mike

 

[doct9960]

roadapple,

1. Where can the "standard" cap dimensions be found?

ASME B16.9

2. How can the 0.7 <=L/D<=1.0 ever be met?

See answer to #3 below.

3. Does a pipe cap fall under torispherical or semi-elliptical?

Ellipsoidal

 

[roadapple]

Thanks for the information.

I truly appreciate it.

roadapple

 

[eliebl]

See UG-44 when using standard pipe components. That will get you out of doing the calculations.

 

[Jcengman]

In the table 10-Ia in the ASME B19.9, we can found the dimensions for the "standard" caps, but anyone know where we can found the tolerances for these dimensions?

 

[JohnBreen]

DO NOT try to calculate the maximum allowable pressure for a pipe cap. You would have to do this calculation for all the different manufacturers of B16.9 welding fittings.

As stated above you should use standard (ASME standards B36.10M and B36.19M) pipe with standard (ASME B16.9) pipe caps. These are dimensional standards and you must also reference the ASTM standard for the material from which the fittings are made.

You must be aware that b16.9 DOES NOT standardize ALL the dimensions for all the welding fittings (e.g., pipe caps) that it covers. WALL THICKNESS AT ANY POINT ON THE FITTING IS NOT STANDARDIZED. The various radii of the curved parts of the caps vary from one manufacturer to another. Standard B16.9 welding fittings vary in shape and size and wall thickness from one manufacturer to another. HOWEVER, they are all capable of passing a proof test. These fittings are pressure rated. They must pass a proof test or must be shown to be able to pass a proof test. From a practical point of view this means that if you weld pipe caps to a straight piece of pipe (pipe and caps of the same standard schedule) and hydrostatically test the weldment (assembly) to destruction, the straight pipe MUST fail before the caps fail.

So if you calculate the required wall thickness for the pipe and round that thickness up to the next higher schedule and use B16.9 pipe caps of the same schedule you will be OK. Now you must think about the opening(s) in the pipe (cylinder or "shell") and do the area replacement calculations for that (those) opening(s).

Have fun, John

 

[doct9960]

Jcengman,

What edition of ASME B16.9 are you using?

For the 2003 edition, see Table 2 (Metric) or Table I1 (Imperial) for tolerances. For standard pipe caps, the following dimensions should meet the tolerances in the mentioned tables:
Overall length of cap (E)
Outside diameter at bevel (D)
Inside diameter at end

I don't have the 2007 edition at the moment. I guess the tolerances are still the same. Will check tomorrow

 

[gr2vessels]

As doct9960 said, E, D and ID are the tolerances, now they are in Table 13 and I-12 (Ed 2007).
cheers,
gr2vessels

 

[Jcengman]

Thanks for this information!

 

[GenB]

Gentlemen: I had had trouble in the pass finding approved ASME pipe caps since are mfd for
API work most caps only come w APi certs.
how can we address the problem?

 

[Thermodom]

Gentlemen,

Our ASME "U" stamp audit is coming up and our mock-up vessel is made with an 8" NPS pipe cap ASME B16.9.

The AI during his visit to prepare us for the audit has asks to insure the received pipe cap be checked for dimensions before being used.

Apparently, the dish radius and knuckle radius has to be verified that it meets the UG-81 tolerances, regardless of the markings.

Are the only official "shape" dimensions per ASME VIII Div.1 given as dish radius 0.9*D and knuckle radius 0.17*D?

How should the UG-81 tolerances be interpreted; "the inner surface ... should not deviate outside of the specified shape by more than 1.25% of D ...", does this mean the head should have a max inner dish radius of 1.0125*D*0.9? Similarly for the minimum dimensions? This is not a particularly big range of tolerance.

Personally, I find the word "surface" confusing in the UG-81 paragraph, a surface tolerance referenced by a lenght dimension.

Thank you

thermodom

 

[racookpe1978]

Well, consider that the specification (the 1.25% you quoted) is trying to limit the actual deviation (a linear distance) of the head radius from the "perfect" head radius of a "perfect" pressure vessel. So, the "surface tolerance" isn't measuring square area or number of pits per sq inch or depth of cracks or anything else except a distance.

Been about 15 years since I modeled PV heads for the old company's even older CAD system, but I don't recall the tolerance (radius) being only 1.25%. Double check that.

Otherwise, yes: knuckle radius is a precent of overall OD, and head radius is a different percent. The head tangent length (for large heads) is usually 3 inches, but tangent lengths for pipe caps differ between fabricators.

 

[GenB]

Thermodom,
wHAT'S THE MATERIAL SPECIFICATION:
ASTM OR ASME?
genblr

 

[Thermodom]

Thank you for the interest,

Material is SA-234-WPB, 8"NPS schedule STD.

I have clarified another thing, the tolerance of pipe schedule also applies. So in our case, I think my inner diameter can vary in this range:
a) largest inner diameter D : (8.625-2*0.875*0.322)=8.0615"
b) standard inner diameter D : (8.625-2*0.322) = 7.981"

THEN, I apply the tolerance of +1.25% and -5/8% on these two "D"s:

c) largest dish radius : 1.0125*8.0615*0.9 = 7.346"
d) smallest dish radius : (1-.00625)*7.981*0.9 = 7.138"

THUS, the tolerance is around 0.21" on the dish radius.

The knuckle radius has only a minimum value to be respected.

I may be tempted to run with this but I would appreciate a confirmation that the approach looks sound (in case I missed something)

A+

thermodom

PS.: Field checking these tolerances at receiving is not obvious, we are considering paper cut-outs...

 

[vesselfab]

i had always consider a cap to have a 2:1 ellipsoidal shape.

there is no real dish or knuckle radius per se.

There are some formulas that come close but no true radiuses that match every manufacture.

what I have used, right or wrong, is an ellipse with the od of the head with a 30* rotation, or a height of .25 the diameter.

draw it on cad. print it out cut it out and check it.

I can tell you that as long as you have an cut out of an ellipse that the height is 1/4 the diameter and it matches head, there is no A.I. that will question the head.

 

[racookpe1978]

No. Use the tried and true "old" calipers outside calipers as a go-no gage. The "threat" you face (most servere stress) is with the shallowest thickness head, so need a "go" gauge" set to the smallest allowed head thickness: anything thicker is satisfactory.

Starret makes them:

http://catalog.starrett.com/catalog/catalog/groups.asp?GroupID=516

Lots of comparable (curved leg) calipers as well that display the actual thickness - digitally or analog. (I no longer trust most hourly workers to reliably read an analog micrometer gauge, so the digital is problably a better choice today.)

 

[vesselfab]

I think what he is checking is the curvature to make sure it matches semi-ellipsoidal shape.

that's what i understood

if it's the thickness, well, you could use a simple D-meter or the calipers method.