Hi all,

I am supposed to analyse AFC inlet and outlet piping integrally on one system. Per Equipment drawing Header is free to move in transvers dir'n and anchored in longitudinal dir'n at the front end of the tube bundle.

I have combined both system by anchoring at the center on the inlet and outlet nozzle with imposed displacement per configuration.

My Problem is that if i model the Inlet/Outlet connection by RIGID(WEIGHTLESS) option, loads are comming in Nozzle( 24 no's 6" nozzle)is exceeding API-661 allowable. But if i change those rigid option( 325mm both side from anchor point)to flexible pipe, load is comming within allowable.

My question: Using the flexibility of AFC Front header( modeled as pipe)
is justified ?
Thanks in advance for any guidence.

It is not clear from your mail if the inlet and outlet header nozzles are on the same header box i.e if the arrangment is even pass or on different header boxes i.e odd pass. Normally the inlet header box is fixed i.e. fixed in the direction of the tubes and API 661 recommends 6 mm on each side or 13 mm on one side clearance between the header box and the C-Frame ( although modern Air Cooler designs are such that the header box is free to move axially).

Coming to the combined model, how do you plan to model the tubes, the tube supports etc? In my opinion a combined model is not only cumbersome but not realistic.

I would recommend that model the Air Cooler inlet and outlet separately and use rigid elements to model the Air Cooler.If the outlet nozzles are on the other header box. i.e. odd pass, use the tube expansions i.e. alpha Delta T Times tube length to find the "imposed displacement' on the outlet header box.This is a conservative estimate.For even pass, check if the "split header box" concept has been used.

Also use rotational restarints at both the inlet and outlet header box supports , this is just to simulate the stiffness of the Air cooler header boxes vis a vis the pipe stiffness i.e. the piping will find it very difficult to make it move in the rotational degree of freedoms ( I personally feel that it is very difficult for piping to make the Air Cooler move in the translational DOFs also but cannot prove it mathematically and till date no industry expert has answered this question to my satisfaction).

Some Air cooler arrangements include "thrust blocks" also which is basically an "interconnection" between the header boxes ( normally the header boxes are not connected to each other).This is to ensure minimum differential movement between the Air Cooler header box and the piping header.Again , as I mentioned in the last paragraph that I find it difficult to conceive that piping can make a header box move, so somehow this concept is not to my liking.

Regards

Do a search for air cooler and you'll find a lot of past discussion on this subject.

Dear Anindya,

This query is regarding the "Split header box" concept in AFC.

In "Split header box" concept in AFC, how to model it in C-II.

Which side of header in normally fixed ?

Regards,
Shamalia Bhavin