Hello everyone,

I’m working on the piping system connected to a 20” pump nozzle and the designer put an expansion joint after the pump nozzle and a line stopper after the expansion joint. The expansion joint that is being used is an untied single type bellow. Initially, they calculated this with no effective ID or no pressure thrust was considered and the nozzle is within the allowable. When I tried to input the effective ID, the nozzle failed. The anticipated maximum pressure is at 8.5 kg/sq.cm and it will produce a pressure thrust of around 15300 kgf, this force is to be carried by the restraint (stopper) and the pump nozzle so this value is very disastrous to both restraints.

1. Is there any problem on the way the piping is supported?
2. Is the pressure is too high?
3. Is there any problem with the type of expansion joint?
4. What other solution in order to overcome large pressure thrust?

1. Potentially big problems. 15.3 tonnes is a large force to contain.
2. Nothing wrong with the pressure ?
3. All EXJ's are potentially weak points, but fine if properly design and implimented in a system based design.
4. Tie rods / hinge / gimbal / stand clear.

Hello,

please remember,
the full force from pressure thrust don't act on the nozzle of the pump. The force from pressure acts on the inside area of the pump and indicates a horizontal force and a moment on the anchor of the pump.
The nozzle of the pump is nearly free from forces because of the expansion joint in front of the nozzle.
At the nozzle acts only a horizontal force from the inside pressure and the "ring-area" of the bellows (Outsidediameter of bellow - Insidediameter of bellow). In the other direction the forces act on the end of the pipe (bend) and so on the Fixpoint besides the expansion joint.

If you use an expansion joint with tie rods, then the forces from pressure stay in the expansion joint and the tie rods.
But ... if you have some thermal expansion between fixpoint and pump then the tie rods dont't fit closely and the full pressure thrust acts again on the body of the pump.

The problem of expansion joints with fixpoints in front of a pump nozzle should be coordinated with the manufacturer of the pump.

Hi,

There is no problem with your model, the results are logic, The only way to fix it is using tie Rods. You should speak to the designer who put a simple 20" expansion joint without Tie Rods near a pump Nozzle, this is a big mistake and keep in mind that more your ND is big more your pressure thrust effect is destructive even with low pressures.

Best Regards,

@ Mr. Oertel
Thanks for the reply

If I will use tie rod, is the pressure thrust will act on the nozzle of the pump when the bellow compressed due to the expansion of the piping?
You said that the pressure thrust will act only on the inside area of the pump and the nozzle will only absorb the force from the difference of Outside and inside bellows area.

@ abdelkader
Thanks also.

I tried to model an expansion with tie rod on the nozzle, the pressure thrust was transferred and absorbed by the tie rod..the thrust force is no longer acting on the nozzle. But prior to that, I anticipated that the pressure thrust will act on the nozzle because the nozzle and piping will tend to compress the bellows and it will opposed by the pressure thrust. Can you please explain why this happened?

Regards,

Clyde

Hi,

First of all, the tie rods are connected to the expansion joints flanges not to the pump Nozzle, that's how you have to model your tie rods because the pressure thrust force is transferred to the expansion joint flange before passing to the pump nozzle.

I think that the pressure thrust forces applied on the expansion joints flanges is greater then the nozzle and pipe forces applied on the same flanges thats why when you use tie rods, the rods absorbs the pressure thrust effect until this forces becomes equilibrated.

This is my personnel analysis, may be someone else have a better explication of this issue.

Regards,

That is when the tie rod is in tension and resisting the pressure thrust, maybe the pressure thrust will be eliminated..

1.How about when the piping will compress the bellows? what are the forces that will act on the nozzles? how do we threat the pressure thrust if that is the case?

2.If I will design the pipe with an untied single bellow with a directional anchor on the adjacent side to resist the thrust force, I think it will be the same as with the tie rods.

Now my other concern is how to interpret this results in CAESAR II? since the pressure thrust will not be absorbed by the nozzle but the restraints and the equipment anchor. When I entered the Effective ID, there is a remarkable amount of load at restraints and nozzle.

That is when the tie rod is in tension and resisting the pressure thrust, maybe the pressure thrust will be eliminated..

1.How about when the piping will compress the bellows? what are the forces that will act on the nozzles? how do we threat the pressure thrust if that is the case?

2.If I will design the pipe with an untied single bellow with a directional anchor on the adjacent side to resist the thrust force, I think it will be the same as with the tie rods.

Now my other concern is how to interpret this results in CAESAR II? since the pressure thrust will not be absorbed by the nozzle but the restraints and the equipment anchor. When I entered the Effective ID, there is a remarkable amount of load at restraints and nozzle.

That is when the tie rod is in tension and resisting the pressure thrust, maybe the pressure thrust will be eliminated..

1.How about when the piping will compress the bellows? what are the forces that will act on the nozzles? how do we threat the pressure thrust if that is the case?

2.If I will design the pipe with an untied single bellow with a directional anchor on the adjacent side to resist the thrust force, I think it will be the same as with the tie rods.

Now my other concern is how to interpret this results in CAESAR II? since the pressure thrust will not be absorbed by the nozzle but the restraints and the equipment anchor. When I entered the Effective ID, there is a remarkable amount of load at restraints and nozzle.

To answer your questions above:

1) The piping will only compress the bellows if the forces on each flange of the bellows can overcome the pressure thrust. And then you have those same forces acting on the restraint and pump.

2) If you have an expansion joint with directional anchors on either side of it, then I believe it will act like a tied-bellows. The pipe is "free" to laterally displace, but the piping is locked-in axially.

Have you looked at pressure-balanced bellows? They make an inline type, and a type that basically has a bellows at the nearest elbow. Perhaps they can help.

I hope this helps. Bellows design is confusing and talking with a manufacturer's engineering team will help.

Hello Devink,

Pressure balanced bellows is one of my option but we have to change the routing and also for me it is very complex and difficult to model in Caesar with this kind of bellows. Based on the answers above and the previous replies, I think Thrust force should be absorbed by the area perpendicular to thrust direction (maybe an elbow or a pump casing) not on the nozzle as per force balance. The force acting on the nozzle is the force required to deflect the bellows (spring rate) plus the force developed by the difference between the bellow outside diameter and bellow inside diameter in a pressurized bellow.

Maybe its better to inform the pump manufacturer about the existence of pressure thrust, am I right? Because we dont want the nozzle to carry the Thrust force, then how about the pump casing? What I do know is that the pump already considered the thrust force during the design of pump but I dont know how they calculated the thrust force.

Regards,

Clyde

Dear Clyde

Please note that two nuts set are there in both end of tie road. So tie rod will neither allow to compress the bellow nor allow to expand the bellow ( except minor rod thermal expansion effect).

Tied bellow only allow lateral movement of pipe not any axial movement.
So when you use tie rod complete pressure thrust will be absorb by the rod.

using a untied bellow for high pressure & large bore pipe diameter is very dangerous.

Pressure balance bellow is highly expansive, should not be used unless there is real demand for it.

Best Regards

Habib

Hi Clyd,
Is it a rubber bellow or typical Carbon/Stainless stell espansion joint

How many pumps in the system? If more pums there, it is better to consider them togather by putting them in one model and setting different operating cases. Caesar is able to analyze all these for you.
Thanks

@ Habib,

Yes, I got your point. Tied bellows are meaningless when used for axial movement, they are best used for lateral movements. For me even control/limit rod is not advisable, its better to use untied bellow. Anyway, I have a 8.5 kg/sq.cm system pressure in a 20" line (imagine the thrust force), based on your experience, is it advisable to use EJ? What do you think?;)

@ Learner2011,

It's a metallic expansion joint. Any thoughts?

@ Carter,

The system has 3 pumps, and ofcourse it is combined in one system, is there something to do with EJ? interesting.


Thank you all guys. Have a nice day.

Clyde

Clyde,

The results could be various depanding on the layout and locations of the EJ, anchor, shut off valve, and also the operating cases of 1 pump run, 2 pump run and 3 pump run. If the model is properly modelled, Carsar will give you the resuts for each case. I think the pressure trust load is high and it is not secondary load. Peng's stress analysis book has a detail discussion about that. But your case is more complicated since there are 3 pumps.

Carter

********I’m working on the piping system connected to a 20” pump nozzle and the designer put an expansion joint after the pump nozzle and a line stopper after the expansion joint.*********

Dear Clyde

As per your initial question.
Yes it is quite usual practice of some good companies for cooling water line.
I guess you also handling cooling water system.

1)Do not forget to use Tied bellow.
2)Just after bellow put rest+guide+limit stop
3)You can not model this system in Caesar & justify
4)It is assumed that tied bellow with gasket flexibility will take care of small amount of thermal expansion between bellow & limit stop
5)This is not accepted for line with considerable temperature


Regards

Habib

Habib, that's very dangerous, you're going to put limit stop after a tied bellow? give me a reason not to disagree with you. It's better not to put limit stop, just guide and resting type..

Clyde

Dear Clyde

I suggest that kind of system only for cooling water service not for any system with considerable temperature.

For cooling water system you can find many existing system so rigid but performing very well.

People popularly call cooling water system will crate problem only you try to analyze the system in caesar.

Anyway I don't insist to follow my recommendation, I share my experience only.

Regards

Habib

Although it's typical to run the system at Design Pressure I would suggest running a load case at operating pressure to get realistic pressure force loads on the piping system. I would agree that the thrust force need not be considered on the pump nozzle. The casing will take most of that load.

Dear all,

I'm glad to be part of this forum.

I read your comments but I'm still doubting If I really modelled properly my system.

It's an straight line (aerial branch) that comes from the pump conection to a buried header. This branch enter to the header with a 45º elbow. The length is around 13-15 meters.

This line have some guide supports & flange support which are guided. So, there are no fix support (axial stops).

The Diameter of the line is 2000 mm, before the pump there is 3 meters with a reducer 2000/1600 & pipe of DN1600, so the nozzle is DN1600. The max design temperarure is 60 degrees.

I used an expansion joint with tie-rods:

In order to calculate the nozzle loads in the pumps & support, I created 2 models, exactly the same, but with the following differences:

1) Nozzle loads model (NZ): If we need to check the nozzle loads (only for that), the nuts of the tie-rods in the expansion joint must be fixed for expansion (EXP = 0), but It could have a gap for compression (COMP = 10 mm). Applying this settings, the way to check the nozzle loads is deleting the Effective ID, otherwise, the nozzle pump will receive the Pressure x Section because for Caesar is like a “cap” or “blind flange”.

2) Support Loads Model (LS): In this case, the Effective ID is introduced and the expansion & compression of the expansion joint is 0. Applying this setting we obtain the worst loads that can receive the supports, movements, stress, etc…

Are those assuptions correct?

A coleage suggest me to introduce an axial stop in the line, but If I do that, the axial loads increase up to 80 tons...

What's your opinion about that?

Many thanks in advance.

It's recommended you start a new thread, rather than dredging up a very old one.

With that said, if you click on the "switch to full reply screen" and upload some screenshots, it'll make it easier to respond.

However, generally when you have enormous loads as such, it's because you're telling CAESAR it has to compress piping axially an equal amount to that which it grows, which is understandably not going to work.

Calculating and plugging in stiffness values for such supports will reduce them somewhat, but they will still be typically high, I find.

Many thanks Michael.

I will create a new thread.

Regards,