Impact Testing Reduction Figure UCS-66.1 note1

[Dutch-EGR]

Hello,

We are facing a problem with a thicker component in a ASME VIII div 1 pressure vessel with a minimal design temperature of -46°C.

The vessels required thickness is determined by its external pressure giving it a min required thickness tr of 43mm. The internal pressure gives a tr of 32mm. tn- corrosion is 47mm.

Looking at Figure UCS 66.1:
1) If I would use the alternative ratio based on internal stress it would give me 0.67 it would give me a huge reduction in impact temperature requirement.

2) If I would use the tr external it would only give me a reduction ratio of 0.91 so a far higher impact testing requirement.

When I put on my layers hat I would say go for option 1.
My engineering hat says option 2 and never take risks with these issues.

What is your opinion about how to interpret note 1 of Figure UCS-66.1?

Kind regards

 

[Geoff13]

In Figure UCS-66.1 the definition of tr says "Required thickness of the component under consideration in the corroded condition for all applicable loadings [Figure UCS-66.2, Note (2)], ..."

Figure UCS-66.2 Note 2 defines applicable loadings as "Loadings, including those listed in UG-22, which result in general primary membrane tensile stress at the coincident MDMT." The key is the reference to "general primary membrane tensile stress", which external pressure wouldn't typically cause. This phrase is used a number of times in UCS-66. Plus it makes sense that you need tensile stress for brittle fracture to occur.

I would read these two sentences to say your Option 1 is the proper interpretation.

 

[SnTMan]

Dutch-EGR, reading thru the various notes and so forth of Figs UCS-66.1 and 66.2 it seems justified to use the Alternative Ratio based on "general primary membrane tensile stress", which of course discounts the external pressure condition. As well, it seems brittle fracture is more likely under tensile stress than compressive stress. So, you could (probably) reasonably justify your option 1.

However, you'd need to get your AI and any client(s) to buy in as well. And then there is the "sleep at night" factor also.

Unless there were serious penalties (cost, delivery, etc) for doing so I'd most likely just impact test and be done with it. I'm a KISS kinda guy

Regards,

Mike



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