EN13445 Ch. 13.4.5 Allowable Stresses in Bolting - Plate Heat Exchangers

[Ollie42]

Hi

Have come across PHE (Plate Heat Exchaners) from a well known competitor, supplying for semi-welded PHE for heat pump applications (NH3) in the high pressure segment PN63.

The immediate impression is that it looks somewhat fragile for 63 Bars, and if We make reverse engineering on the units, the design code EN13445 would limit these units to appr. 28 Bars.

It is specifically the quantity and size of Tie Bars which differs greatly from what would be designed following requested rules from EN13445.
In the competitors design the axial stresses exceeds allowable values , from ch. 11.4.3...where you have to use the lesser value of Rpo2/3 at design temp, or Rm/4 at room temperature.

Tie Bar material are identical to whats normally used (42CrMo4) steel.

EN13445 is a certified design code approved for PED.

We have contacted our 3'rd party to scrutinize the design , and they conform that the design we are following are the correct method.

Any shortcuts, or workarounds in the design codes that we have overlooked??
Other methods for 3'rd party approval than calcs.....excessive testing or what to do ???
Mystery ?

Best Regards Ole

 

[XL83NL]

Sounds somewhat familiar to me. Have had the same doubts with an EU manufacturer. Is this EU based as well? You mentioned the PED and I je harmonized standard EN 13445…

Though not European, have you looked at ASME VIII-1 mandatory appendix 45 on PHE?

Huub
- You never get what you expect, you only get what you inspect.

 

[Vikko]

If I understood correctly your case: the test is always the final best approach and it confirms the design.

If you are talking of a typical flange joint (see the drawings in the standards):
- EN 13445-3 ch 11 it is just the same as ASME VIII-1 apx 2, and many other different design codes, all based on the Taylor-Forge Method for flange design. This method will always oversize the dimensions of the flange joint, as pressure gets higher (for example 63bar). You cannot expect to have a small joint with this method. So you have to check tab 6-1 and compare it to ch. 11.4.3.1 and you will see much difference in the nominal design stress of the same material, if it is considered as a bolt than when it is considered as another material: this is because the Taylor Forge method for the bolts takes a very low nominal design stress so it will always oversize the required bolt area.

Otherwise you shall use EN 13445-3 Anx G (or also very similar EN 1591 but more complicate), which are more advanced technich or AMSE VIII-2.

All the previous methods I mentioned, are Design By Formula, so you shall follow strictly the formulas and refer to interpretations and calculation softwares.

The work around, especially in complext geometry (for example not round, a square flange, where these methods are all not applicable, or an o-ring working as piston/rod sealing, or otjher seal mechanism where these methods are not applicable), you shall do the Design By Analysis (example: Finite Element Analysis). EN 13445-3 provides 2 ways: anx B (plastic limit) and anx C (elastic), then you need to use a simulation calculator and do your own assumptions.