To whom it may concern

I mainly use Caesar II for the flexibility analysis of FRP/GRP pipelines. I have been using the software for the past year and I am still not 100% sure whether I am modelling tee's in the correct manner. Initially I modelled tee's in the same manner as one would model a steel tee, ie. I modelled the branch and header pipes having the same thickness as the actual tee, I then modified the SIF's to accommodate the increased thickness of the tee. One of the problems with this method is that the pressure stress multiplier cannot (unless I am doing something incorrectly) be modified and this results in high stresses at the tee's.

I then adopted a more realistic approach (personally, this approach seems more accurate when modelling moulded tee's, but once again I am not 100% sure whether it is correct) where I model the tee as it is manufactured. Initially I obtain the manufactures specifications for tees' and other fittings. Each tee in Caesar II is then modelled using 3 separate elements which are in turn connected to the adjacent pipe elements. The 3 tee elements are then given the correct wall thicknesses, OD's and material properties as opposed to the pipe parameters. This method has the effect of reducing the SIF's and pressure stress multipliers due to the increased wall thickness. Is this method correct for a moulded tee's?

My next question concerns the modelling of "set-in" tee's. If the diameters of the connecting pipes are different (ie. a 40NB pipe connected to a 200NB pipe), then a different manufacturing procedure is followed. If the pipe is for example filament wound, then a hole would be machined into the larger diameter pipe, the insulation material would then be welded and the pipes would be reinforced with a suitable compensation pad (usually a combination of CSM and WR) surrounding the smaller diameter pipe. Therefore, the method which I described above cannot really be used as the comp-pad does not necessarily surround the whole of the larger diameter pipe. How should this type of connection be accurately modelled?

If somebody knows the correct procedures, please could you respond to my message as I cannot find any relevant information / documentation concerning this matter. I really appreciate your time.

Best Regards

Jason

In answer to your first question:
Yes this method is fine.

For your second question:
the answer is "You can't". Note that CAESAR utilizes the 3D Beam element. This is essentially an infinitely thin stick - with properties such as diameter, thickness and modulus. The concepts of volume, surfaces, thickness variations and so forth are beyond the capabilities of this element. You are limited to uniform cross sections, and overall system behavior.