Hello,

I have a interpretation problem with the scope for the calculation. UHX-13.2 say "...and edge conditions".
Does it mean both tubesheets must have the same configuration acc. Fig UHX-13.1?
All other code of practise calculations (TEMA, AD 2000) I have seen ignore this and calc. the tubesheets separate.
Note, from 10 HX I found 6 with different "edge conditions".
If we replace now TEMA with UHX and AD 2000 with EN 13445-3 (which is nearly the same as UHX), what must I do? FEM?.....
We have done som HX FEM-Calculations. That's not funny.

Regards

Frank

Hi,
UHX assumes that tube bundle is symmetric. A very low-class approach, isn't it? But on the other side, it may be an advantage, because when you have both tubesheets fixed, and they ar not identical (damnit, my tubesheets never are!)- you don't use the UHX, full of errors, by the way. You follow U-2(g) and use whatever methods you like. 13445 or AD, or TEMA or FEA...

Regards
Konrad

Hi Konrad,

what I want to say is: AD 2000 is officially replaced by the EN 13445-3 and TEMA by UHX. Both are only tolerated. OK, AD 2000 has get the stamp "same as EN 13445-3..." in the PED. But this is policy.
Official we have in Europe two CoP's for HX: EN 13445-3 for internal use and ASME for export. So we have to build symmetric HX's or use FEM.
Nice job....

The other thing is, AD 2000 has only symmtric fixed tubesheets (with and without expansion joint). But I have never heard that's an AI demanded a FEM for an HX with leftside: "extended as a flange" and rightside: "integral". I think the practice experience and the time are the reasons for this handling. But from the conservative few point....

For the practicability use of the UHX we should know:
Can we calculate an HX with leftside Config. a and rightside Config. b with UHX?
Get we different stresses from right to left in the asymmetric case?
If not, which CoP should we use? U-2g is silly if the other CoP have the same problem.

Regards
Frank

Frank, I would use TEMA. See RCB-7.166.
TEMA is not phased out nor forbidden. ASME just converted nonmandatory App.AA to a mandatory UHX, but U-2(g) is still in use, luckilly.
Konrad

Hi there,

In my experience as an AI we always appreciate Manufacturers with the expertise and standing to make a clear decision. These are the cases we have seen:

- UHX although the edges are different, but justified by the length of the shell and proven by the similar stress levels into the tubes and similar displacements of the bundle, ...
- U-2(g) and TEMA because the edges are different, but justified by...
- U-2(g) and AD-Merkblaetter because the edges are different, but justified by...
- U-2(g) and EN-13445 because the edges are different, but justified by...
- U-2(g) and FEM, justified by...

Any of these cases can be accepted, if the calculation is done in a professional way and documentation meets Code. The manufacturer shall approve the design calculation.

The worst are Manufacturer who do not have a clue of the calculation and want to push their responsibility to the AI, the TÜV, an Engineering company, the program engineer, or the cleaning woman.

When they contract for a design calculation on the market and the cheapest bidder wins, the trouble is there.

In the fancy cases the HX is "ASME but not stamped", so there is no inspector at all.

Hi,

I agree with Frank: If the HX is not symmetric, and the influence of the missing symmetry can be neglected, UHX can be used.

At this point nobody volunteers to miss the chance to use some other Code. Most ASME HX are designed not to UHX in real life.
Maybe that will change when the mistakes are eliminated.

Dirk

Many thanks for the discussion.
It was helpfully.

Regards

Frank

Let me add one conclusion:
Try to estimate bending moments acting upon the tubesheet.
One moment is from expanding tubes, acting with an arm of distance of tubes to the shell (only which ones- the outer, or some mean distance...)
Second moment is from the bolts, acting on a bolt-to-shell distance.

If second moment is much smaller than the first one- then stresses are more or less symmetrical. Shapes are not, but stresses are your concern.

Konrad