I would appreciate some advice from anyone who has delt with this problem.
I am working on a cooling water pump with a 14" diameter discharge nozzle that has a rubber expansion joint connected to it. The expansion joint does not have any control rods to absorb the pressure thrust. I can anchor the piping downstream of the joint to hold the pressure thrust. The pump will act as an anchor for the other side of the expansion joint. My question is: Do I need to worry about the pressure thrust load on the pump nozzle which exceeds the load allowed by API 610?

Hi Joe,

Should you be concerned about longitudinal (along the pipe's axial centerline) pressure thrusts that are the consequence of using (installing) an expansion joint in pump piping?

In a word, absolutely!!!

Remember that an expansion joint is 1/3 of the total expansion compensation system. The other necessary components are the anchors and the pipe guides. As you state, the pump is one of your anchors and your design will supply the other (opposite) anchor. Then you have to think about guides.

Many times elastomeric expansion joints are used at pumps to "uncouple" the piping from the pump and thereby theoretically "eliminate" pump vibrations from being transfered to the piping. This should never be done capriciously. Since the EJ disrupts the path of the longitudinal forces (due to pressure), these forces must be accomodated in some other way before they impinge upon ANY strain sensitive piece of rotating equipment. This might be done (in a relatively cool piping system) by installing a line-stop somewhere between the pump and the EJ - just be careful not to force too much expansion/contraction force into the pump (you can approximate the location of the line-stop by quick hand calculation and fine tune the choice of the location by several CAESAR II runs). Obviously, this will force the bulk of the expansion/contraction deflection into the EJ. Then you have to be careful to assure that the designed capacity of the EJ to absorb expansion/contraction is not exceeded (consult the EJ's OEM - also ask them about the annular "thrust area" or the total axial thrust acting at the EJ).

Once you satisfy all those considerations, consult the EJ's OEM regarding the location of the required guides - those closest to the EJ and those further away. Remember that the pipe is acting like a column with a rather large slenderness ratio and it wants to "buckle" under axial (longitudinal) loading. The EJ is a structural discontinuity which acts to facilitate this buckling. The guides prevent this buckling by redirecting the buckling forces into a different load path - so remember to check to see that the available adjacent structure is strong enough to accomodate the loadings on the guides. The same cautions apply to loading imposed upon the structure at which the "anchor" is located.

A very good reference for this work is:
Guidelines for the Design and installation of Pump Piping Systems.
V. A. Carucci and J. A. Payne
Welding Research Council Bulletin 449

Good luck with your project.

Best regards, John.

[This message has been edited by John Breen (edited March 06, 2001).]

[This message has been edited by John Breen (edited March 06, 2001).]

I agree with John. Of course you should be concerned with pressure thrust loads when expansion joints are used.

However, I don't think John directly addressed your query -- "Do I need to worry about the pressure thrust load on the pump nozzle which exceeds the load allowed by API 610?" I see only minor pressure thrust loads on the nozzle.

The way I handle pressure thrust loads is to get inside the pipe (figuratively speaking) and look up and down the line. The surfaces I see are the surfaces loaded with pressure and therefore, the location of the pressure loads. In the joint, looking away from the pump I see the inside surface of the elbow. The pressure thrust here is taken directly by your anchor. Turning around and looking the other way, I see the pump internals. These pressure loads go right into the pump shaft, casing and then supports. So far, no pressure load on the pump nozzle. That is, except for the exposed area of the expansion joint convolutions. The difference between the large ID of the bellows and the small ID of the bellows will provide an area upon which the pessure will develop a compressive load on the nozzle. It's a little more complicated than that if you want to consider all the stiffnesses involved -- the joint stiffness, support stiffness and pipe stiffness but I don't think the numbers will change much.

------------------

I agree wholeheartedly with both Dave and John and I can add some of my own first hand experience.

I was asked to investigate a pump systems setup at a power plant... (No Names on any of these facts). The installation kept "knocking" the pumps out of alignment. The pumps were large bore 36NPS suction slurry pumps. The suction side was installed with untied elastomeric expansion joints on the flange of the pump(s) itself.

The engineer of record had ignored the pressure head on the suction (about 40' as I recollect) and had only provided guides on the suction piping.

When I looked over the components I noticed that despite the massive pump housing, variable speed motors etc. the actual components holding the pump to its baseplate were rather gossamer like.

So I had a pump suction and discharge piping isolated and drained. Then I had dial indicators installed and zeroed. Finally I opened suction and then discharge valves recording dial indicator readings as pressure came up. Lo and behold the pump "walked" over 0.030".

I contacted the engineer of record and advised them they had some problems to reconcile! They denied there was a problem... the owner retained my services and I restrained for pressure thrust unloading the delicate pump mounts... problem solved.

Hope this helps,

------------------
Best Regards,

John C. Luf

Hello Dave,

I have the same situation, only that in my case is a rubber expansion joint. The effective area of the below is less than the internal area of the pipe. So my "ring area" (Aring=Aeff - A int)is negative, would that make any sense??

I've never seen a bellows penetrate into the flow stream, but it doesn't make any difference. You should be conservative and enter the pipe ID for effective diameter.

Hi Dan,
What about if we have tied Expansion bellow, still do we need to input the Effective dia.. as i presume tie rods in the epanison bellow will take the pressure thrust.
In the CAESAR II application manual, in one of the sample problem (attached), Effective dia. is in the input. However for tied bellow ths should not be entered. Please clarify.
Best regards,
Shaikh.

I'd recommend always entering the effective diameter except for a couple of very special circumstances. It's the conservative a safe thing to do. When using a tied joint, the rigid element and C'nodes will restrain the pressure thrust.