Modelling Tee in Duct

Posted by: Goodsalt

Modelling Tee in Duct - 11/08/07 10:21 AM

We're running a 42" duct analysis that includes a single tee and branch. We neglected to specify the type (reinforced, etc.) in other words we left the field blank, and it ran. An overstress condition at the joint disappeared because of this. What's the program doing when the tee type isn't specified?
Posted by: Richard Ay

Re: Modelling Tee in Duct - 11/08/07 10:45 AM

If you leave the "tee type" blank, c2 assumes you are going to enter your own SIF value. If you leave this blank also, c2 uses an SIF of 1.0.

This is typically not what you want to do, which is why when you go through Error Checking this situation produces "Warning 33E". This warning states:

There is a GEOMETRIC INTERSECTION at node "x" and there was NO user defined INTERSECTION TYPE. Standard pipe flexibilities and SIF's = 1.0 will be used unless overridden by the user on each element framing into the intersection.
Posted by: Goodsalt

Re: Modelling Tee in Duct - 11/12/07 04:40 PM

By the way this duct is 42" diameter, 0.035" wall, so it doesn't meet the criteria for D/t <= 100. Neglecting to specify the type intersection allowed the run to pass code compliance. This is the first time it did so, since in earlier runs we had specified the type. Calculating SIF's per the code yields extremely high numbers, since the flexibility characteristic is so high. Anyone know how to handle this?
Posted by: John C. Luf

Re: Modelling Tee in Duct - 11/12/07 05:05 PM

Its called 3 dimensional analysis via FEA! c2 as well as the code assumes classic beam element behaviour of the piping your section will behave more like a pop or beer can and not a piece of 4"NPS standard weight pipe.

The fact that you are discovering this is a good thing the fact that you are doing it after the fact of building a model is a bad thing!






Posted by: CraigB

Re: Modelling Tee in Duct - 11/13/07 09:31 AM

Why are you wasting your time applying B31 rules to a construct that is surely not included within the B31 family of codes?