2:1 Elliptical Vs 2:1 Torispherical Heads 3

[darrenyee88]

Hi all,

I understand that 2:1 Ellipsoidal head has crown radius (L) = 0.9D and Knuckle radius (r) = 0.17D. So is it true that 2:1 Torispherical head has the same crown radius and knuckle radius as 2:1 Ellipsoidal head ? If yes, Can someone guide me where to find this information.

Or someone can guide me where to find information regarding above to let me have a good understanding.

Thanks and cheers.

 

[VPbris]

Hi,

I know this post is a few months old but I did not think I needed to create a new one since it is exactly about this topic.

First of all, thank you all for sharing your thoughts on this, really great discussion.

I understand the "inconsistency" of this approximation and I share the thoughts of TGS4 on this. However I still have one question, on designing using the formula provided on UG-32(d) and by the aproximation of 90/17 using 1-4(d) I end up with a thickness 1.19 times higher when calculating as a torispherical head. I also plotted the ratio between these thicknesss varying diameter,pressure,allowable stresses and when comparing, this ratio remains always the same.

So I ask, when designing a vessel if I specify a semi eliptical head say of minimum thickness X, and my head manufacturer supply me with a approximated torispherical head of minimum thickness X, is this vessel code compliant? I ask because this 20% thickness difference, that on one hand helps saving money, but on the other hand can lead to safety issues. What is your opinion on that.

Kind Regards,

 

[TGS4]

That inconsistency is definitely an issue that is well recognized by the ASME Code Committee. As far as I am concerned, it all depends on whether you're getting a true semi-ellipsoidal or torispherical head.

Btw, I am putting forward a proposal at the August Code Committee meetings to delete the equivalency in UG-36(c), per r6155's suggestion above. The proposal will also require head manufacturers too explicitly state which shape they are using. And, it will impose the external pressure tolerances on thin heads where knuckle buckling is a governing condition. Ideally, I would like to completely replace the VIII-1 head rules with what's in VIII-2, but that may be a tougher sell. Comments from the community?

 

[victorpbr]

I agree on deleting the equivalency in UG-36(c). However, in Divison 2, this equivalency is the basis of the design for the elipsoidal head in 4.3.7.1, will be there any change on that also?

The way I see, if it goes with only deleting the UG-36(c), the head manufacturers that supply "false" elipsoidal heads will either have to adapt their technology (high costs for them), or the designers will have to stop specifying "true" elipsoidal heads, hence adding more cost to the project as the thickness calculated by the "false" elipsoidal head is about 20% higher.

The scenario today is designers specifying "true" heads and getting "false" heads. Good for the investor (less cost on head plates), good for the manufacturer (no need to invest in different dies). However, it is indeed an engineering problem, one has to define if this approximation can or cannot be made and the rest will be consequence for that.

In my humble opinion, I think the best way would be as you described, completely replacing the rules of division 1 by the rules on division 2. This way we would be getting always "false" elipsoidal heads, but that is what we would be designing for. But as you said, a tough sell to be made.

 

[BJI]

Out of interest, have there been any vessel failures cause by this discrepancy between SE and TS heads initiating the investigation of these methods, or is it more for completeness of the S8 codes?

 

[TGS4]

No failures that I'm aware of. Possibly hydrotest failures. The margin against failure is high enough that 18% shouldn't matter. Internal self-consistency is important, however.

 

[BenStewart]

In one respect the equivalency comment and discrepancy between UG-32(c) and 1-4 is good because it triggers useful discussions between engineers and manufacturers. Conversely, as already stated, although geometrically similar the circumferential stress can be quite different and the discrepancy can lead to different opinions and inconsistencies when doing design and/or in-service assessments. Going forward, for consistency above all else, I think the deletion of the equivalency comment and mandating manufacturers explicitly state the shape and external pressure limits is an excellent idea.

I'm certain this wouldn't fall under the ASME B&PV Working Group remit but what are the thoughts to adding detail within API 579 on what considerations should be made if assessing wall thinning damage - or other damage mechanisms - in the region of the head where there is a compressive circumferential stress. Perhaps a sentence or two within the 2C.3.5 or 2C.3.6 on the geometric sensitivity and taking actual measurements using a template, plumb line and/or laser etc.? Possibly something within Part 5 of ASME VIII-2 also to cover instances where features are added to this region.

 

[JStephen]

From reading above, it sounds like a head produced to the 90:17 meets the tolerances for a 2:1 elliptical?
IE, if you have the physical head in front of you, is there any way to confirm which profile it was intended to represent?
So far as changing up the code, I'm not seeing what problem that solves.
I'd be curious if those two profiles get closer together when pressurized.