Materials Data for ASME BPVC UG-101 Burst Test Calculation

[gvc99]

I'm trying to determine the MAWP of a vessel based on burst testing using UG-101 P = B/4 X Smu X E /Smur. I have the burst test data, and I can find the minimum tensile strength of the vessel material, Smu(316L bar), but I do not have actual tensile data for the material that the burst test specimens were fabricated from. UG-101 allows an alternate calculation that uses Smur, the "maximum tensile strength of range of specification at room temperature". The specification for the 316L bar only lists the minimum tensile strength. Where do I find the "maximum tensile strength of range of specification at room temperature"?

 

[metengr]

Most material specifications require only a minimum reported value for ultimate tensile and yield strengths. You should use the reported minimum property values for your calculations.

 

[gvc99]

Thanks for the reply, but I understand the thinking behind the UG-101 requirement to ratio the min tensile to max tensile in the formula. In lieu of actual tensile data for the component burst tested, you want the highest expected tensile value for the specficed material, just in case you happened to get lucky and have a heat of material with the highest properties in your test specimen. Otherwise you would be non-conservative in your prediction of MAWP for a vessel made from min spec material.

 

[metengr]

gvc99;
I just reviewed PG-101, and from my understanding of paragraph a) 2 look at (a). Tests based on yielding of the part can be used to establish internal MAWP, with a ratio of minimum specified yield strength to minimum specified ultimate tensile strength of 0.625 or less.

(b) refers to actual burst testing.

 

[gvc99]

We did the burst test, not the instrumented yield test, so I guess I'm stuck looking for the "maximum tensile strength of range of specification at room temperature".

I checked SA-276 which is what the vessel material was purchased to, but only the minimum tensile strength is specified.

 

[metengr]

gvc99;
Ah, after some additional reading, I believe you need to remove actual samples from the burst tested component (in an unaffected region) and have them tested. The samples should not be torch cut. These values are used in the calculations.