Stress due to thermal expansion connecting nozzles

[Ehiman1]

Hello, I noticed a client request that we never had before regarding calculation of stacked heat exchangers.
We have two AES type in stacked and client asks to calculate the stress due to Thermal expansion of connecting nozzles, does someone can help me? In other projects clients never asked to us to consider this condition

 

[NRP99]

Can you simply use lαΔT? Probably yes if you need just rough estimation of expansion or probably not if you need to submit detailed calculation. You need to do thermal analysis and find out temperature distribution and then go for the structural analysis with the temperature as load input to find the stress distribution and expansion.

 

[Christine74]

In order to minimize thermal stresses, for stacked TEMA "E" shells it is preferable to route the flow through the shells so that the interconnecting nozzles between the two shells are located at the channel end rather than the shell cover end, if possible.

I would probably take the average of the shell fluid inlet temperature and the shell fluid outlet temperature and call that the mean metal temperature of the shell. From that you can determine the differential expansion between the two shells, then determine the stresses using FEA. ASME B31.3 might have guidelines as well.

If needed, additional flexibility can be added by increasing the distance between the shells or by installing an expansion joint in one of the shells.


-Christine

 

[KevinNZ]

A piping designer could model the HEXs and piping and provide the nozzle loads at the vessel walls. Then apply usual nozzle calculations.

 

[Some Curious Guy]

I would suggest that you speak to your client about his expectation. Is he expecting an FEA analysis. That would be significant cost adder!!. Normally client does not ask for this unless there is some huge temperature difference or other valid concerns. This is also clearly specified in the Purchase order or in some form pre order discussions.

1) FEA would be the most appropriate method to do this. Hope you have inhouse capabilities to do it. Otherwise sub contract it.
2) Another method as @KevinNZ suggested is to have your Piping Engineer model the heat exchanger in Caesar/Stadd/Autopipe etc and give you the loads. You can then find out the stresses by using usual methods ( WRC-107/297 / Noz Pro etc ). Assuming you are vessel vendor and do not have a piping stress expert its wise to directly sub contract it to some FEA vendor instead of a Piping stress vendor.
3) An easier but rough method is to estimate the mean temperature of the upper and lower shell and find Delta T. Use this Delta T to find differnetial thermal expansion Delta L Use Delta L to find corresponding Force due to differential thermal expansion between the two shells. By inspecting the geometry and location of interconnecting nozzle and saddle arrive at some estimate of how this load would be shared by each nozzle. Once you have these force you can do simple WRC-107/297 / Noz Pro Analysis to find out nozzle stresses. You would have to do some reading about stress classification and allowable limits per Section VIII Div 2 Part 5.

Bottom line is. Ask the client about his expectations. Agree on an FEA analysis only if the client had clearly stated in their preorder communication. An experienced Engineer would not insist on an FEA analysis for small temperature difference.

All the best.