UHX Applicability 1

[stanrick]

UHX-10(a) says that for the rules of Part UHX to apply, a tubesheet must be "uniformly perforated over a nominally circular area". I conclude, then, that if a tubesheet is not "uniformly perforated over a nominally circular area", the Tube Requirements of UHX-4(e) are not mandatory. In that case, I am free to analyze the tubesheets as welded flat heads (UG-34) with the tubes acting as stay-rods having strength determined by Appendix A and the cross-sectional characteristics of the tubes.

My question is this: Are there guidelines for what constitutes a "nominally circular area"? Aspect ratio perhaps?

Consider a tubesheet on a 10" ID Shell. If the overall tube pattern was 6" square with a diagonal of roughly 8.5", I might reasonably call this a "nominally circular area". But I would not consider a rectangular area 6" wide x 3" tall to be circular.

Is this something for which I must find an agreement with my AI (and Canadian Inspectors for CRNs)?

 

[SnTMan]

stanrick, this subject has been discussed before, in the context of "no tube in window" designs, which resembles your last example. Do a search.

For lack of alternatives, Part UHX is routinely applied to these designs. TEMA rules can be used also, but have the same limitation, nominally circular area.

My company builds S&T exchangers for which we use Part UHX (whatever the tubefield looks like), and we also build rectangular air coolers, for which we use UG-34 for tubesheet design, although we do not take staying by the tubes into account.

Never had any kicks about the design method used, however for your case, a visit with your AI and the Canadian juridiction may pay off. Nova Scotia is one thing, Alberta something else altogether.

Regards,

Mike

 

[stanrick]

Mike,

Thanks for guiding me to the No-Tube-in-Window reference. Your feedback is consistent with what my AI says.

I've used UHX on non-circular tube distributions in the past. But lately have found that the tube column bending results (UHX-13.5.9) are driving me to what seems like over-conservative results on larger shell diameters.

I appreciate your input.

Rick

 

[SnTMan]

stanrick, you are welcome.

Regular readers will know I am not big fan of Part UHX tubesheet design

Regards

Mike

 

[MrBTU]

I'm getting responses from European PED reviewers stating that for no-tube-in-window designs neither UHX nor TEMA are applicable and they are demanding FEA's. FEA's for tubesheets are very complex - we were recently quoted $5K for one.
Who said PED wasn't a trade barrier???

 

[SnTMan]

FEA by people you don't know. Scary

Mike

 

[TGS4]

FEA for $5k. That's crazy cheap! And a complex tubesheet FEA for $5k. Either you were sold ocean-front property in Arizona or they really didn't know what they were doing.

 

[curtis2004]

As far as I know FEA pricing should be ~$50-$100/hour. You can astimate how much does it take to make a model, meshing, run time, results, making a report.

 

[MrBTU]

Actually we do know the people who performed the FEA, they did a great job, it was accepted by the Notified Body the first time through, and they have extensive experience doing tubesheets, so it was more efficient than us doing it in house.

 

[TGS4]

curtis2004 - not to take this thread off-topic, but do you really think that is reasonable? Assuming that the FEA people are working for a consulting company, where billing rate mark-ups are in the 2.5-3 range. That would mean that you would expect that your competent FEA engineer would be paid $20-$33/hr? You _might_ get a new grad for that. Throw in software costs, and ALL you could get for that is a co-op student or new grad. That doesn't smell like competent to me...

Think more in the $150-$200/hr range and you're a little better.

 

[SnTMan]

MrBTU, no disrespect to your FEA vendor, I was speaking more from past experience

Regards,

Mike