This is not a question. More like a tip.

I have discussed this issue in the past but there was a case just now that showed how important it is to model universal expansion joint with two tie rods correctly. The case is a replacement job on a line that is some 40 years old connecting to a turbine. Original design has one hinge and one universal unit. The arrangement was decided to be maintained although one hinged and two gimbals could be better (more expensive for sure).

Original design has tie rods in vertical position. Replacement expansion joints were designed to be about the same as before and the line was analyzed using CAESAR II. As there were no maximum forces and moments available recommended values were used. One force and one moment were above the allowed. To confirm two people modeled the expansion joints using different modelling technique. Same result.

Tie rods were rotated 90 degree to horizontal and the analysis shows all forces and moments to be well under the allowable limits.

Why the difference in the results? Simple answer. End rotation. Universal expansion joint ends can rotate around one axis if there are only two rods. To understand why take two pens and a plate and try to balance the plate on the two upright pens. Any universal expansion joint with internal pressure works the same.

To get the work right model using complex modelling. Get the rod orientation right. For this modelling you need to know nipple (center pipe) length and element spring rates. These rates are not unit rates but individual element rates. Using the pressure thrust design the tie rod diameter and check EJMA recommended diameter. Use these diameters and correct temperature and you get accurate results.

The last item what you need to do is to make sure that the unit is installed in correct orientation. That is most probably the most difficult part. Recommended solution is to make a note to the expansion joint specification sheet: 2 rods, orientation is.... manufacturer to include label giving installation orientation.

Safer to use three or more rods but then the forces and moments can be too high as no rotations are possible. Note that due to high forces involved two rods may not work always!

CAESAR II is very good in analyzing systems with expansion joints. You just have to get the modelling right.

I do expansion joint designs. Unfortunately there seem to be increasing number of failures. I would say the main reason is not by the manufacturers but by the pipe designers. Poor expansion joint specifications and/or simply wrong designs. Expansion joints are not easy items. Ask help from specialists.

And remember. You cannot have tie rods if you have axial movement

Thanks a lot for sharing!!