FT coolers are very fragile and it is VERY important to be certain that the pipe loadings are minimized. You should not model the boundary condition as one (single acting) restraint, and certainly not in the +Y direction. The pipe to cooler interface is going to be either a weld or a flanged connection, both of which will transfer moments.

I suggest a model/boundary condition setup similar to the "vessel" examples we have in the Applications Guide. I would model to the nozzle, connect to the cooler with a CNODE, and then continue the model down to the base plates. (The CNODE will give you loads in the restraint report, and if you want to implement WRC297, introduce some flexibility here.) At the base plates you can apply friction and model gaps to simulate your slotted bolt holes.

[ August 20, 2001: Message edited by: rich_ay ]