I am new to modeling expansion joints and have reviewed a handful of the post. What is confusing me is that when I input the data for the stiffness the loads on my suction nozzle sky rocket. I am sure this is a user error but I am not sure where. Also, I have noticed that the expansive cases all pass when I use the Expansion joint but the other cases all fail.

Any help is appreciated.

It's probably your expansion joint thrust load. If you enter an effective ID for your XJ and you have pressure on your XJ spreadsheet, CAESAR II will apply a pressure thrust load on either end of the XJ (pressure times effective bellows area). If you have tie rids in your model, that thrust load would put your rod(s) in tension and little pressure thrust load will bleed into your nozzle. But without any tie rod in the model, that thrust load simply pushes on your suction nozzle. This load, in the real world, does not find itself on the nozzle, it is on the impeller and casing behind it.
If your XJ is intended to compensate for axial growth, then tie rods are not part of your design. If, however, the XJ is installed for its lateral compensation, then tie rods can be used to carry the pressure thrust load.
If you want pressure thrust in your untied model, place a force on the first elbow upstream of the pump and pointing away from the pump and assume the counteracting pressure thrust (pointing towards the pump) is carried by the pump base. More sophisticated modelling would have you break the thrust load into two components - 1) based on the inside pipe area and located at the upstream bend 2) based on the effective area minus the inside pipe area and located on the upstream end of the XJ.

Thanks Dave,

You are correct I am trying to reduce the axial growth. However, my understanding of these is much less than this. Currently I have the pump as a rigid anchor followed by a support than the bellow followed by a 90 going below grade.

Why does entering the effective ID cause the axial loads to grow so much? Is this as you said that "CAESAR II will apply a pressure thrust load on either end"? If that is the case then what I am seeing when I add the corresponding pressure thrust on the elbow will lower this.


Am I understanding this correctly?

Sorry for not being more clear.
Entering the effective ID in the CAESAR II input for an XJ will automatically add the pressure*area to either end of the XJ. If you do not specify this effective ID, CAESAR II will not add those pressure loads which pull the joint apart.
My suggestion is to NOT enter the effective ID; INSTEAD, place the pressure thrust loads, by hand, at the more-correct locations in the line. In doing so, most of that "automated" pressure load that you see pushing on the nozzle will be removed.
There are several posts on this Forum regarding this pressure load on pump nozzles. Take a look.

Dave,

Thanks for the all the help. I have been looking at many of the different threads on expansion joints as well as the application modeler. I found the "Expansion Joint Manual Thrust Force Input" thread to be very insightful. Where I am still having trouble is when to enter an Affective ID and when not to. I have modeled the XJ as is shown in the Application modeler on page 83 as the "Equivalent Single Bellows Lateral". I have applied restraints the zero weight rigid as shown on page 84 "Universal Expansion Joints - Simple Model".

Please correct any misunderstandings that I have.

I understand that by entering the Effective ID that Caesar will calculate the Pressure thrust. I also understand that if I do not enter the Effective ID than I need to manually apply the Pressure Thrust at the bend and the nozzle.

What I do not understand is why do the loads increase in all directions when either of the above is done compared to a model that does not have a bellow in it.

I also know to subtract the applied Pressure Thrust from the Nozzle load should it be manually applied.

If there is no pressure-retaining hardware (tie rods or hinge plates) it is important to consider the pressure effects around the XJ. You can include the pressure thrust in the analysis - either by entering an Effective ID or by placing the pressure forces by hand. If you model the pressure-retaining hardware, that pressure load should simply load up that hardware rather than bleed out into the system. If you are NOT modeling that pressure-retaining hardware and the XJ is tied or hinged, do not apply a pressure thrust force in your model (this is that simple model).
Regarding questionable results with and without the XJ... You should be able to "deconstruct" your model to further identify the source of those large loads. I would suggest looking at individual load components to identify the cause. For example, does P1 cause the load or, if defining your own thrust load locations, F1; or is it a thermal issue (T1). Try it with and without the effective ID specified. Narrow down the source and you will eventually identify it.