I am looking for the most realistic way to model Air Coolers in CAESAR II, especially the restraint of the header box.

In this case, I have a double pass air cooler with split header boxes for inlet and outlet, and two nozzles per box.

One way to model the header box and connection to nozzles, is as a rigid without weight and a restraint in the center of the header box.

In terms on the header box's restraint, which is my concern, it may be indicated with X, Y & Z restraints, and no rotation restraint. Is this the best approach?

Please advice.



------------------
Pablo Guerra

I think your modeling method is correct but it is not the best way. You need to check the installation type of your tube bundle assembly for more accuracy.
There are many modeling techniques in Air-cooler piping. I think the modeling of header box depends on the design type of head box and tube bundle assembly. Your Air-Cooler is double pass. Normally single pass air-coolers are not as critical as double pass with center baffle. This means the conservative methods do not effect your calculation results.
Your design is double pass with split header boxes, not with center baffle. The differential thermal expansion between inlet header and outlet header is also not a problem for the separated header boxes with thrust block device. However you have to check the lateral movement for multi-bundle Air-cooler. If I were you,I will model as belows ;

In case of using thrust block between header boxes ;
If I assume center header box is anchored among multi-tube bundles, X(lateral direction) is free( Z and Y are fixed), and you have to control the maximum movement of the last end tube box or pipe header. RX,RY,and RZ are depend on the flexibility of tube bundle. However, I think these three DOF may be modeled as free restraint but this does not mean the nozzle rotations are free.

In case of no thrust block between header boxes ;
If your design do not use thrust block, you need to consider the friction effect due to the weight of header box and tube bundle assembly. However, it is uneasy to estimate this friction effect, even though sliding device is employed. If I assume each header box is restrained at Z(axial direction of tube bundle) only, X should be gap restraint within +-6mm( Z and Y are fixed). RX,RY and RZ are same as above.
Normally Air-cooler consists of header box connected to tube bundle and side frame. There is small clearance(+-6mm or 13 mm based on API 661) between tube edge and side frame. This gap allows header piping to move heade box laterally with the loads in proportion to the flexibility of tube bundle and friction load due to tube bundle weight.

The most realistic way is to incorporate the actual design of header box in your model such as thrust block, anchor bolt and slot hole, clearance between tube and side frame, sliding device, etc.


[This message has been edited by Soon Ryang, WEE (edited June 27, 2000).]