API 660 loads on BEM Heat Exchanger Channel with Concentric Cone

[FPPE]

Hi,

I'm designing a BEM heat exchanger with concentric cone and I have a doubt on evaluation of API 660 loads.
Design code is ASME VIII-1 2023 and equipment shall be marked with U designator.
Design conditions tube side: 1.77 MPa + FV @ 100 °C
Tube side MOC: SS316
Please consider the following sketch:



As you can see, I have to consider API 660 values for external loads on nozzles.
I need the following clarification:

1) API loads are applied at the intersection of shell to nozzle junction. In this case we do not have a shell, but a girth flange connected to the extended portion of tubesheet.
This seems to be a case which should be evaluated only by FEA, but API loads should be applied on flange resulting critical than the ones applied at the junction as required by standard itself.
2) Gasket of flange-tubesheet joint is a GMGC kammprofile type and there are No. 36 (0.875"). Radial load on nozzle could be a problem for flange leakege?

Is there a way to consider those loads applied on flange by formulas for which we can avoid the FEA? Is it possible in software such us PV Elite?
Do you have any other advice for these design?

Thanks in advance

 

[Trestala]

The API 660 nozzle load values are just a standard table used as reference in the design. When used, it acts as a starting point where you can design your nozzle connections (e.g. local stress at the nozzle-shell junction) and also acts as a limit for the Piping team for the loading that their piping will impose to equipment's nozzle.

Even those API 660 values are specified as acting on the nozzle-to-shell junction, those values have to be aligned with your Piping team on what is the actual loading being applied to your flange connection. You can align with your piping team that the API 660 standard load values will be considered to act on the 20" flange connection since you don't have a shell. Or have a different set of design load values acting on the 20" flange.

You don't necessarily need to use FEA to evaluate the girth flange-to-tubesheet connection. You can just account for the super-imposed axial load and moment coming from the 20" flange to the girth flange-to-tubesheet. There are different approaches, and you could probably convert those nozzle load values into an equivalent pressure that can be included in the flange and tubesheet calculations.

 

[FPPE]

Hi Trestala,

Thank you very much for your answer.
Could you kindly provide a reference for the equivalent pressure calculation?

Thank you

 

[Si-bo]

Talking about ASME div.1 u-stamped heat exchangers, I've always calculated those kind of nozzles as below:

- FEA analysis of the nozzle considering loads applied at the flange-cone junction
- Appendix 2 of ASME VIII div.1 calculation of the flange considering additional equivalent pressure due to axial load and total resulting bending moment

For the equivalent pressure consider:
4 * Force / (pi * G^2) as equivalent pressure due to axial force
16 * Moment / (pi * G^3) as equivalent pressure due to bending moment
with G being the gasket reaction diameter.

Also referring to @Trestala 's comment, it's not a matter of the numerical value of the loads, but it's a matter of how to consider these loads.


Hope this helps,

S.

 

[FPPE]

Hello Si-bo,

Thank you for your answer.
I need some clarifications:

1. With "Calculation of the flange", do you mean the girth flange connected to tubesheet right, not the standard flange (which is only verified accordint to B16.5 rating (we do not have neither corrosion allowance so rating is applicable)?
2. Regarding the equivalent pressure: assuming that we apply the loads on the flange face, how do we take into account the distance between the point of load application and the flange girth when determining the formulas for equivalent pressure? I guess this is an important parameter that would change the stresses in the flange.
3. The equivalent pressure mentioned has two formulas, how are these computed in practice to verify the girth flange according to Appendix 2?
4. Considering API loads (Fx, Fy, Fz, Mx, My, Mz) how do you determine “Force” and “Moment.” If the approach is similar to UG-44(b), Force should be Fy (radial to STD flange shown) and Moment should be the square root of (Mx^2+Mz^2), but in doing so we will be ignoring Fx, Fz forces and torque My.
5. Regarding FEA: I agree that in this case FEA should be necessary. Do you also usually model the bolts and flange with which you couple the girth flange in question? So in this case I should model the tubesheet as well? Analyzing the channel to tubesheet flange by equivalent pressure, I would model for FEA up to the circumferential weld cone to flange, excluding the girth flange, already checked by design by formulas.

Thanks in advance

​

 

[Si-bo]

With "Calculation of the flange", do you mean the girth flange connected to tubesheet right, not the standard flange (which is only verified accordint to B16.5 rating (we do not have neither corrosion allowance so rating is applicable)?

First of all, B16.5 pressure-rating is based only on material of construction, temperature and pressure, corrosion allowance is not rilevant. and B16.5 does not takeinto account of external loads.
The nozzle flange is subjected to external loads, so you need to include those in your calculation, that's why I consider app.2 formulae + equivalent pressure.
I agree that nozzle loads on the 20" flange have an effect also on the girth flange, but I usually don't consider those because you never know what are the values there and you should also consider the effect of all the other nozzles loads.

Regarding the equivalent pressure: assuming that we apply the loads on the flange face, how do we take into account the distance between the point of load application and the flange girth when determining the formulas for equivalent pressure? I guess this is an important parameter that would change the stresses in the flange.

You don't know the effect of the distance from the point of the load application because you are considering tabular loads of API660 (unless you perform a stress analysis). In this case, just consider API660 loads both at flange facing and nozzle-shell junction to be more conservative.

The equivalent pressure mentioned has two formulas, how are these computed in practice to verify the girth flange according to Appendix 2?

Simply make appendix 2 calc. considering design pressure + force equivalent pressure + bending moment equivalent pressure

Considering API loads (Fx, Fy, Fz, Mx, My, Mz) how do you determine “Force” and “Moment.” If the approach is similar to UG-44(b), Force should be Fy (radial to STD flange shown) and Moment should be the square root of (Mx^2+Mz^2), but in doing so we will be ignoring Fx, Fz forces and torque My.

Yes you are ignoring 2 forces and torque moment, but that's the formulae approach. You should perform FEA analysis in order to include those.

Regarding FEA: I agree that in this case FEA should be necessary. Do you also usually model the bolts and flange with which you couple the girth flange in question? So in this case I should model the tubesheet as well? Analyzing the channel to tubesheet flange by equivalent pressure, I would model for FEA up to the circumferential weld cone to flange, excluding the girth flange, already checked by design by formulas.

As 1st point, I'm talking about 20" nozzle flange. And since i'm considering flange-cone junction, just model cone and nozzle flange hub with loads applied at their junction.

 

[r6155]

Please sketch OP to scale and your opinion may differ.

 

[FPPE]

Si-bo, thank you for all the clarifications!

r6155, could you please clarify your answer? Thanks

 

[CuMo]

In PV Elite you have the option to subject flanges to external loads.
UG-44 b)
Design both the small and large flanges of the channel by adding the maximum Axial load and Equivalent Bending moment in the Torque; Loads; Partition Gasket Information module.
This will allow PV Elite to transform the loads into equivalent pressure loads. Make sure you investigate both positive and negative load scenarios /All combinations/.
The reason you want to check both flanges (Big and small) is because you might find the bolting to the tubesheet is actually governing the design.

Now, the situation with the actual concentric cone isn't so easy.
You would need to satisfy UG-22 c) External loads from piping.
Such external loads can lead to buckling of the cones.
To the best of my knowledge - there is no facility in DIV.1 which would help you here, so you'd need to go to DIV.2 Class 2, which means FEA. /NozzlePro, Ansys, etc./

 

[FPPE]

Hello CuMo,

Thank you very much, I have applied the loads in the Torque; Loads; Partition Gasket Information module.
In order to be sure, I will perform anyway the finite element analysis in accordance to VIII-2 Class 2.

 

[r6155]

See the sketch with a CONICAL SHELL

Regards

 

[FPPE]

Thanks for the sketch r6155.
Could you please explain what we need to pay attention to?
Thanks in advance

 

[r6155]

You can avoid the knuckle.

Regards

 

[GD2]

Hi,

I'm designing a BEM heat exchanger with concentric cone and I have a doubt on evaluation of API 660 loads.
Design code is ASME VIII-1 2023 and equipment shall be marked with U designator.
Design conditions tube side: 1.77 MPa + FV @ 100 °C
Tube side MOC: SS316
Please consider the following sketch:

View attachment 1532

As you can see, I have to consider API 660 values for external loads on nozzles.
I need the following clarification:

1) API loads are applied at the intersection of shell to nozzle junction. In this case we do not have a shell, but a girth flange connected to the extended portion of tubesheet.
This seems to be a case which should be evaluated only by FEA, but API loads should be applied on flange resulting critical than the ones applied at the junction as required by standard itself.
2) Gasket of flange-tubesheet joint is a GMGC kammprofile type and there are No. 36 (0.875"). Radial load on nozzle could be a problem for flange leakege?

Is there a way to consider those loads applied on flange by formulas for which we can avoid the FEA? Is it possible in software such us PV Elite?
Do you have any other advice for these design?

Thanks in advance

Unfortunately, you can't validate the allowable nozzle loads by any WRC Bulletins (107/537, 297) due to the geometry and given limitations.
The best way will be to go for FEA.

 

[GD2]

Hello Si-bo,

Thank you for your answer.
I need some clarifications:

1. With "Calculation of the flange", do you mean the girth flange connected to tubesheet right, not the standard flange (which is only verified accordint to B16.5 rating (we do not have neither corrosion allowance so rating is applicable)?
2. Regarding the equivalent pressure: assuming that we apply the loads on the flange face, how do we take into account the distance between the point of load application and the flange girth when determining the formulas for equivalent pressure? I guess this is an important parameter that would change the stresses in the flange.
3. The equivalent pressure mentioned has two formulas, how are these computed in practice to verify the girth flange according to Appendix 2?
4. Considering API loads (Fx, Fy, Fz, Mx, My, Mz) how do you determine “Force” and “Moment.” If the approach is similar to UG-44(b), Force should be Fy (radial to STD flange shown) and Moment should be the square root of (Mx^2+Mz^2), but in doing so we will be ignoring Fx, Fz forces and torque My.
5. Regarding FEA: I agree that in this case FEA should be necessary. Do you also usually model the bolts and flange with which you couple the girth flange in question? So in this case I should model the tubesheet as well? Analyzing the channel to tubesheet flange by equivalent pressure, I would model for FEA up to the circumferential weld cone to flange, excluding the girth flange, already checked by design by formulas.

Thanks in advance

​

FPPE,
You can validate the adequacy of the 20 inch flange by couple of options:
1. Kellog's Equivalent Pressure Method - Very traditional where you convert the external loads (Forces and Moments) into an equivalent pressure and then compare the total pressure load to the ASME B16.5 pressure rating. Very simple. Few of my colleagues have already highlighted this.
2. Use ASME Sec VIII Div 1 UG-44(b) - M will be the resultant Moment (SRSS) and F will be the axial force component.

Typically the above two checks are for validity of the flange (if the flange is adequate), not stress check at the flange-cone joint. If you read carefully, API 660 Table 2 values are at Nozzle/shell junction, typical for a standard vessel design. WRC are common to do these stress checks but you can't apply it due to your geometry and design.

3. The other way to go around it, will be as Trestala had advised, you need to talk to the piping team and give them the nozzle loads to make sure that when they do the piping design and layout, the loads are not exceeded. Typically, most of the external load comes from thermal (67%) and rest from weight (33%).

4. API 660 doesn't specifically say if the external load table 2 is for ASME B16.5 flanges. I would assume yes.

The Girth Flange is independent and you need to design per ASME Sec VIII Div 1 Appendix 2.