Dear sirs:

just last week, a accident happened to one expansion joint on one plant. the four tie bars all were broken during piping running. please see enclosed picture.
Now we are study the reason.
According to my experience, the force (read in CII) always is very bigger on the rigid element in expansion joint.Before I always neglect this force since it is so big.

Here I want to know how to read the force correctly on tie bars in CII calculation model.

thanks!

stress river,

The tie rod forces would be shown in the CaesarII restraints report, if the tie rod ends to lugs are modeled as restraints with Cnode to connect the tie rods. If the tie rods are modeled simply as elements connected to lugs, then the local elements forces report would show the loads for the tie rods.

The results of the CaesarII analysis could be compared to the simple calculation for pressure thrust, Force = Area * pressure, = I.D.* 3.14 * P where the tie rod load = F / 4 , given 4 tie rods and just the pressure for load on the tie rods. There are also weight and thermal movements to add to the tie rod loads, but the large diameter of the pipe would seem to result in pressure thrust causing largest component of tie rod loads.

The pipe apears to be 900 to 1000 mm diameter. The pressure might not be too high - the miter elbows would not be used for high pressures. The tie rods appear to be about same diameter as the rungs on the adjacent ladder. The tie rods might 12 mm diameter. The tie rod diameter would be suspected to be under-size with the tensile failure of the tie rods.

A comparison to another expansion joint in a 10 bar steam service would be useful. The size 300 mm expansion joint in 10 bar steam service has tie rods of 38 mm diameter. You can do the pressure thrust calculations for the 300 mm diameter 10 bar steam expansion joint, and the pressure thrust of the failed larger expansion joint with its diameter and pressure, to see perhaps how much larger diameter the tie rods might be expected.

stress river,

A correction to the described 300 mm expansion joint in steam service. The design pressure was 4 bar, and not 10 bar steam. The tie rods have even more design factor in the 38 mm diameter -for working pressure or hydrotest.

Dear Richard:
I made a calculation for your example :
1. Firstly, I assume the effective area of bellow is Pi*(300*300)/4;
2. the design pressure is 4bar;
3. the pressure thrust: 0.4*Pi*(300*300)/4=28260N;
4. for each tie rod: F=28260/4=7065N;
5. for the DN38 tie bar, if material is A36,
.yield stress : 250Mpa
.effective area: 817mm2(I can't find effective area for DN38, This one is for DN36);
.the maximun force for each DN36 tie bar: 250*817=204250;
So whetehr DN38 is too conservetive for your example?

For my example:
1. the effective area of bellow is Pi*(1066*1066)/4;
2. the design pressure is 2.25bar;
3. the pressure thrust: 0.225*Pi*(1066*1066)/4=200708N;
4. for each tie rod: F=200708/4=50177N;
5. for the tie bar, the diameter is DN27,
.yield stress : 315Mpa
.effective area: 459mm2
.the maximun force for each DN27 tie bar: 315*459=144585N;
So the DN27 is enough.

Richard, Am I right?
if right, it seem I must find the original reason from installation.

thanks!

I wouldn't assume that the thrust load is shared equally between 4 tie rods. That would be a perfect world.

Hi,

From the picture, it does not appear that size of pipe will be DN 300, it will be more than that.

Also, check the convolutions are designed as per EJMA standard. The failure may occur if the inplane and column stability fails.

stress river,

The comment of Dave Diehl for the tie rods not sharing equally the load is very true. If the CaesarII results report is available for the tie rod end restraints, the individual tie rod loads would show how the deflection and rotation of the piping may cause some of the tie rods to be zero load. The loads on the tie rods are developed from all of the pressure, weight, and thermal movement components for forces to be restrained. That is part of the reason for the safety factor appears to be at least 3 times. If there is thermal expansion upward at the upper end of the expansion joint in the photo, then the tie rods could be pulled by huge thermal loads.

If one tie rod is loaded to yield, then another tie rod might be assumed to then carry some of the forces, but there could occur a 'domino' or cascading failure where each tie rod fails in sequence.

Another comparison for tie rod design would be to make the tie rod area of 459 mm2 x 4 (of your expansion joint) at least 50% of the pipe wall area,
1100 x Pi x wall thickness, for adequate tie rod area.

The design pressure of 2.25 bar might easily be exceeded. Some higher design pressure of even 6 bar or more could be required, depending on the process variations. If there are pressure pulses from valve opening, or two phase vapor-liquid slugs, then the pressure peak could be much more.

Dear richard and DAVE:
please permit me to continue this issue.
I try to simulate the expansion joint exactlay and I can get a exact force on tie-bar.
1.I set the axial stiffness and lateral stiffness for expansion joint with operation temperature;
2.I use simple way to simulate tie bar with ambient temperature;
So when I check the force on tie-bar, I found the force from thermal displacement is very small and main force is from pressure thrust.

It is confusing, my understanding is all thermal force should be burden by tie-bar. when one DN1000 is warmed from ambient temperature to more than 100 degree, I think the force should be big.

can you give some advice to how to simulate universal expansion joint to get exact tie-bar force?
thanks!

Tied joints are built to accept lateral deflection. The tie rods hold the pressure thrust. If you are expecting a larger axial load on those rods from thermal strain, then you are not properly using those tie rods.
If your tie rods are not double-nutted, then, should the compressive axial load due to thermal strain exceed the tension due to pressure, the tie rods would relax and the joint compress. But again, that's not the standard use of tie rods.