TEMA RCB-8 Flexible Shell Element

[MikeG7]

Good day all.
I am puzzled by the applied loadings used to check the stress in the flexible shell element (FSE) of a thick joint per "TEMA 9th Ed. 2007".

I have followed the procedure to determine the equivalent pressure due to expansion, equivalent shell-side pressure and tube-side pressure.

These are combined to arrive at displacements that are to be used for a finite element analysis per the TEMA 9th Ed. guidelines at various locations.

There is very little "guidance" in TEMA 9th edition how to arrive at the displacements but Tema 8th edition has Table RCB-8.7 which provides the method do work out the equivalent force on the joint (and from the known stiffness of the FSE, one can work out the applied displacements).

In my case, I have differential expansion leading to a compression of the FSE (shell wants to grow longer than cooler tubes of same material).

So far, no problem. However, when considering the condition of "Shell side pressure only" , the arrived displacement has a negative sign, leading to a compressive force (or displacement) to be applied to the FSE. This is surprising as shell pressure would surely want to cause an elongation or expansion. I did my own calcs and compared to software output and both are the same in terms of the sign.

Anyone care to share some light or opinion on the matter?

Also any comment if the method I am using for displacement (by using force and spring rate) is acceptable/logical.

Thanks

 

[SnTMan]

MikeG7, I have been puzzling over this. If the condition is pressure only, I don't see the joint being compressed. Per UHX13.5.5 differential expansion gamma (γ) is zero for pressure cases.

Are you perhaps doing the calculation per UHX-17(c)(1)(b)? Is it possible this is the quantity of interest? I am frankly having trouble interpreting the meaning of the same, "shell axial displacement", but I suppose if it is positive, that would lead to compression of the joint.

For your second question (displacement, force, spring rate) I'd say, yes that is reasonable, and likely the only way to get there.

Regards,

Mike

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

 

[MikeG7]

Hi Mike (SnTman).
I have been doing the calculation per TEMA 9th Ed.
I would also now like to compare with ASME UHX but I am aware of the differences accounted for in ASME UHX (being a more rigorous and updated method) taking into account the stiffness of the tube bundle and the resistance of the attached shell in determining the tubesheet thickness required for bending.
But there are many similarities with TMA and ASME UHX.
You can "check" me if you want:-

TEMA inputs as follows for reference
Shell ID = 914 mm 35.984 "
Tubesheet thickness = 50 mm 1.968 "
Shell side pressure 75 Psi
Tube side design pressure = 75 psi
No vacuum
Shell side temp design = 135 C / 275 F
Tube side design temp = 93 C / 199.4 F
tube length = 9000 mm / 354 "
No tubes = 725
Dia. tubes = 19 mm / 0.75 "
Tube wall = 1.65 mm / 0.0650 "
expansion joint ID = 1170 mm / 46.06"
expansion joint outer knuckle radius = inner radius = 50mm / 2"
No straight part on the expansion joint top
shell/channel thickness 6 mm / 0.236"
no corrosion allowance
Mean temps are 90 C / 194 F for shell, 64 C/ 147 F for tubes (colder than shell)
Material all 316L, tubes to SA-249 shell and tubesheets to SA-240.

 

[SnTMan]

Well, gee, it's not that I don't *want* to, it's just that I am not sure what to check

Seriously, which calculation is giving you the result that the joint is compressed?

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

 

[MikeG7]

No issue - thought I'd just share the details with you.
The issue arises when the shell side prime pressure (P's per A.153 of Tema 9th Ed.) comes out negative sign (indicating compression in the shell). This is used in determining the force (and hence displacement) to be applied on the flexible joint for the FEA.
As a note, the following trends were found:
1)increasing internal pressure tends towards more compression in the shell
2) decreasing number of tubes tends towards more tension in the shell (can be explained by the reduction in stiffness of the tube bundle)
3) increasing the joint stiffness tends towards more tension in the shell

 

[SnTMan]

MikeG7, sorry to say I don't have software to run this. Still trying to understand it.

More later perhaps....

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

 

[MikeG7]

Sure Mike and thanks for your input (not just on this thread but generally).

 

[SnTMan]

Kind words indeed. Thank you

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

 

[SnTMan]

To revisit: The only understanding I could come to of the situation of shell pressure only resulting in compressive stress in the shell cylinder is the result of extension of the expansion joint due to pressure. This could (would?) occur if the joint were more flexible than the array of tubes.

To try out the idea I calculated the variable J when Sp' is equal to zero, using the data provided. In turn this results in an expansion joint spring rate of roughly 8 * 10^5 lb / in.

(BTW to me this seems kind of light for a flanged and flued joint, but who knows?)

I calculated Stiffness of the tube array (AE / L) as roughly 8 * 10^6 lb / in, or 10x the expansion joint stiffness.

A lesser value Sj means a more negetive Sp', so the idea seems feasible.

I guess

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