I modelled stack type of heat exchanger with the saddles as rigid. In Seismic case the top nozzle of heat exchanger sways 50 mm(laterally).
I can see the saddle swaying due to modelling of saddle as single element(assumed diameter same as heat exchanger) whereas saddle has quite susbstantial length and 16 bolts to hold.
There will be some sway but 50 mm looks quiet high.
How to model this saddle accurately to get correct results for sway.

Thanks,

Where beam elements are not good enough shell elements are the answer ( or 3D elements where details of through thickness stresses are very important).However, first check the accurary of your model ( take a second opinion ie show it to someone from your organization to check for correctness of model) or send it to techsupport@coade.com .

Regards

Thanks Anindya for your suggestion, 3D element is good solution but would take some time for me to explore. I will start working on that one.
But other way I was thinking If we can use RX and RY rotational restraints on sliding saddle. Would it be correct?
X is axial direction of the HE, Y is vertical.

Regards,

Hi Ajay
Do you have one anchor support & one sliding support for exchanger?
If so I think you should not have that kind of problem as you mention.
If your both supports are sliding support in my opinion you can Put Y Rx Rx support in caesar model to avoid strange result on nozzle behavior for seismic & wind case. Note Rx & Rz here is not a physical support & just model in caesar to get approx result of exchanger nozzle behavior.
I seek expert comments also on my opinion in this case.

The input was done by another stress engineer, I am checking the system as
senior stress engineer. Fixed end is modelled as XYZ restraint, sliding as YZ restraint. The model is correct, the seismic values are very high as it is D class site with the upper end of stack swaying largely.
the heat exchanger dia is 1500 where as pipe dia is DN 600m, so you can imagine how piping can dominate the heat exchanger.

The nozzle loads are high hence trying to get the heat exchanger stiffness as correct as possible.

Regards,

Hi Ajay
Do not model Fixed end as XYZ restraint but model as "ANC" I mean fully anchor support & you will get the desired result.