I am modeling a section of pipe that has a large rectangular duct teeing off it. The duct consists of 2 different sizes with a reducing section between them. My plan is to get the duct stiffnesses using FEA and then input that into my Caesar pipe model.

The question I have is how should I best input the stiffnesses in the Caesar model? Maybe use a rigid element for the duct and attach it via cnode stiffnesses to the pipe...

One word comes to mind ewww! Rectangular sections properties are not symetric like a nice round cyclinder.

Then you throw in localized buckling perhaps and more ewww....

If your FEA will give you a spring rate tha defines the localized krates of all 6DOF at the intersection between the pipe and the duct then you should define them in the beam element model at the intersection i.e., the pipe O.D. see the tech reference manual for nozzle flexibity modeling.... and good luck!


Remember you will end up with an approximation!

My FEA model will give me spring rates for all 6 DOF and I can use it to analyse the stresses at the pipe to duct connection. I was trying to figure out exactly how to get the stiffnesses into the Caesar model so the pipe system behaves correctly.

I was thinking of putting a rigid element from the pipe centre to the OD and then cnoding a rigid element representing the rectangular duct to it using the stiffnesses from the FEA model. Is that how you would do it?

yes exactly see pg 129-132 of the applications guide....

But the rectangular section will not behave as a circular tube... maybe define a structural user defined element and include it into the job ????

Yes, you do want to use the Structural Modeler to define the rectangular elements you need. This would be better than trying to get element stiffnesses from FEA, and then specifying these at a point in CAESAR II.

In the Structural Modeler you would:
1) Define a material for the duct
2) Define a cross section (or sections) for the duct. This will require the "user defined" cross section so that you can specify both inertias and the area.
3) Define the duct element (or elements). They should all have distinct node numbers, so you can include them in your piping model where appropriate.

"ewwwww."

Will that be defined in the upcoming release of B31.3?

Or did you mean "ewe". Here in Texas, that gets the Aggies all riled up.

"ewwwww" as in the situation is stinky! What you may say when you drive by an aging road killed varmint.

This model when done 100% correctly will still have shortcomings due to the use of a beam element model. If the loads and stresses are critical in nature its definitely ewwww.

Ewww is not and probably never will be a code defined word... but I would vote for its addition and definition.