Air Receivers built right on the line

[crosby84]

I think this has been addressed previously, sort of, but I couldn't find my answer.

We evaluate pre-existing vessels (mechanical integrity, as they say) and frequently come across air receivers (ASME code stamped, NB#, etc.) that when we run thickness calculations, they are often at or below minimum required thickness. Even if the U-1 states a nominal shell thickness of 0.25", the min required thickness is 0.255" (exclusive of corrosion allowance), for example, and we have to tell the client that their vessel isn't good for the pressure on the nameplate, even if it hasn't corroded.

Why would manufacturers so frequently build these things right on the line with no CA, often even below the required thickness? Does anybody else run into this regularly?

Just curious to get others' experiences with this. Thanks!

 

[JStephen]

"Even if the U-1 states a nominal shell thickness of 0.25", the min required thickness is 0.255" "- Not sure I understand that statement.
But note that corrosion allowance is usually specified by the purchaser, not by the fabricator.

 

[SnTMan]

I'd say you are missing something...

Regards,

Mike

The problem with sloppy work is that the supply FAR EXCEEDS the demand

 

[crosby84]

Sorry, let me clarify.

The ASME calculations say that the design thickness for the shell is 0.255" when the nominal thickness listed on the U-1 is 0.250". So the calculations say that it was built too thin.

 

[A clark]

I suggest you should take a look at UG-16 (b)(4). The minimum thickness required by Section VIII for compressed air service is 3/32" (2.5 mm). What if any RT of UT was preformed?

Allan

 

[r6155]

See tolerances of plates in SA-20
¿Can you tell us the original measured thickness of new plate?

Regards
r6155

 

[mk3223]

I wonder what it means by the "frequently come across". Is the thinner vessel used by one Company or several companies? Were it designed by one Engineer or several different Engineers? Or,could it be resulted from the change of the operating conditions?

 

[A clark]

I think it is a misunderstanding or unawareness of UG-16 (b)(4. The MINIMUM thickness is 3/32 regardless of what is calculated, .250" meets the Code.

 

[SnTMan]

Well, not if tr is 0.255" as per the OP.

Regards,

Mike



The problem with sloppy work is that the supply FAR EXCEEDS the demand

 

[A clark]

I agree, but I don't think we have all the information, NDE etc and the program they are using may be applying a Ca where none exists.

 

[JStephen]

Crosby-
Normally, if thickness calculates as 0.255", I'd detail 0.3125" or whatever for vessel shells, etc.
If I remember right, the underrun tolerance is generally 0.01", and is rounded to 2 decimal places prior to applying the tolerance.
Maybe they rounded to 2 decimal places before specifying the thickness. Maybe they actually miked the plate to make sure it didn't underrun. There may be provisions dealing with one or both cases in the code, but I'd have to do some research to see. I know some of the tank standards say you can't order plate thinner than the minimum required. IE, the tolerance is to be applied to the as-received plate, you don't get to apply it to the calculated thickness.
If you're comparing your calculated thickness to their calculated thickness, there may be minor differences in how the numbers work out. Differences in static head or OD vs ID, etc.
I would anticipate that if calculated thickness came out to 0.249999", it would generally get built as 0.25" plate, provided there was not some other factor influencing the decision.
I assume that for fitness-for-service calculations, you're allowed to use a thinner in-place shell than the original as-constructed shell, even if there is no CA included. But I would have to research that to confirm.
Also, note that if you use the as-built thickness to calculate MAWP, you'll have this situation quite a bit, regardless of how they originally designed/detailed it.

 

[r6155]

If design thickness is 0,255” I can use any plate 0,25” with actual thickness between 0,24” to 0,28”

(0,25” - 0,01”) under tolerance to (0,25” +0,03”) over tolerance (SA-20)

Regards
r6155

 

[SnTMan]

It would seem this would be permitted under UG-16(c)(2), but...

What a horrible practice.

Regards,

Mike

The problem with sloppy work is that the supply FAR EXCEEDS the demand

 

[JStephen]

Well, right there in UG-16(c)(2), it says "Plate material shall be ordered not thinner than the design thickness". Taking that one statement by itself, it sounds like if calculated thickness is 0.255" or 0.250001", you're not allowed to order 0.250000" plate. I would be curious if that has been addressed elsewhere in the code or in any interpretations- it seems an obvious place for contractual disputes to arise. I'm not aware of any common practice of rounding required thickness to 2 decimal places- certainly, the CA isn't normally specified as 2 decimals.

 

[SnTMan]

...it seems an obvious place for contractual disputes to arise.

Boy howdy

The problem with sloppy work is that the supply FAR EXCEEDS the demand

 

[r6155]

@ crosby84
Have you the original calculations?, can you tell us?

Regards
r6155

 

[JStephen]

One other thing to point out, if you order 0.2500" plate, the underrun tolerance is 0.01", but it's rounded to 2 decimals before applying the tolerance, so actual plate can run 0.236" and still meet the 0.25" requirement.

 

[r6155]

EXAMPLE:
New Pressure vessel ASME VIII Div 1
Design thickness: 0,255”
In the shop there are several plates (remaining plates) with actual thickness 0,260” (ordered as 0,25”)
Can I use these plates for fabrication this new pressure vessel?
If not, ¿why not?

Regards
r6155

 

[david339933]

Standard Designation: BPV Section VIII Div 1
Edition/Addenda:
Para./Fig./Table No:
Subject Description: Section VIII, Division 1 (1998 Edition, 1998 Addenda); UG-16(c)
Date Issued: 10/25/1999
Record Number: BC99-222
Interpretation Number : VIII-1-98-84E
Question(s) and Reply(ies):
Question: May rounding rules provided in ASTM E 29 and referred by plate general requirements SA-6 and SA-20, be used when determining compliance with the undertolerance requirements of UG-16(c) in Section VIII, Division 1?
Reply: Yes.

SA-20: 10.3 Rounding Procedures
For purposes of determining conformance with the applicable product specification, a calculated value shall be rounded to the nearest 1 ksi for tensile and yield strengths, and to the nearest unit in the right hand place of figures used in expressing the limiting value for other values, in accordance with the rounding method given in Practice E29.

Any place I see plate thickness referred in SA-20 it is to three decimal places. I would take "the nearest unit in the right hand place" to be the third decimal.

 

[crosby84]

Thank you everyone for all the comments! This is great information.

r6155, we never know the original measured thickness, just what's put on the U-1 as the nominal thickness (for shells) and minimum thickness (for heads).

We never have the original design calculations for air receivers.

Thanks