I am doing the discharge piping analysis for sea water vertical pump (vertical suction line from sea). The operating temperature is 38 ºC, CS Pipe, discharge line size is 18”. The routing is quite rigid as attached in dwg, discharge of 3 pumps go straight (Approx 6 m) into the header, 2 tapping from discharge line for psv and test line as shown.
On analyzing in Caesar II, I am getting tremendous loads on nozzle as one should get with this type of routing.
The pump piping I am analyzing is almost if not exactly similar to existing pump piping and these existing pumps are running successfully without any failure.
In order to qualify the pump nozzle loads I need to change the routing drastically, even may need to use spring / expansion tied bellow as the allowable loads of pump are very less.Please suggest me if this is right approach to qualify the nozzle load of pump which is operating at 38 ºC or is there any other approach to solve this problem?

Are you calling 21deg C (70 deg F) your ambient temperature? If so, there is a thermal growth of about 1mm between the pump discharge and the header (6m). Do the guides you have on the 28" header, test line, and psv discharge lines have any gaps? Friction will cause loads for sure, but I don't know what you mean by "tremendous loads." Supports in caesarII have a stiffness of 1x10^12 lbf/in (convert to metric...), so perhaps these supports are not truly rigid? Most supports give in some way... Remember that just because a pump is running does not mean that it is not overloaded. The pump recommeded loads are to ensure it runs well for a very long time, but overloaded pumps can still run its just they wear out faster. That just means that the pumps in use right now are running, yes, but that doesn't mean they are loaded properly. Maybe this helps, good luck.

Ambient temperature is 21 ºC, so there is very little expansion at the header (about 1 mm), all guides are with gap, and no guide support is acting, but friction is playing major role .The loads coming on the nozzle are about 10 times allowable values (Radial loads is about 4000 kg against 400 kg allowable loads for 14” nozzle) . Loads will reduce if I reduce transient support stiffness. But the same default stiffness used for all other calculation in project as agreed with client.

As you pointed out the existing pump should be overloaded, but what should I suggest to client? It is difficult for me to convince myself to change pipe routing or use spring or expansion bellow in order to reduce loads on pump nozzle for pump operating at 38 deg C temperature

stressedengineer,

You may want to compare comments in the similar post by dnyanesh, Jan 21,

http://www.coade.com/ubbthreads/ubbthreads.php?ubb=showflat&Number=32716&page=3#Post32716

discussing the top suction pump loads due to water weight. The deletion of water in the vertical pipes could reduce the sustained loads on the pumps, as suggested by ohliger.

Good Morning.
I have heard many times that in the model of the vertical pipe, the water can be removed, for the check of loads on that nozzle.
I guess this is possible because that weight has a minor effect on the nozzle. Can somebody tell me if my guess is right?
And can somebody give me a reference to codes or articles where this topic is discussed? Thank you.

The weight of liquid column in vertical line after top nozzle will act on pump sacing and not on the nozzle. Hence the fluid density for this vertical portion can be deleted for nozzle load calculation purpose.
But The pump I am analyzing is not top suction or discharge. Its end discharge and there is no vertical piping associated with the nozzle as you can see in the drawing attached.

stressedengineer,

The pump discharge is horizontal in the sketch, and I got confused by the 'Vertical Pump Discharge piping' title of the post, along with the other post of top suction pump. Maybe it is time for new eyeglasses!

If the piping loads on pump are excessive, I would suggest 3 tests for the source of the loads. First, test for the pump model by deleting the elements for the pumps, and placing anchor restraint at the discharge piping connection. Second, if you suspect friction in supports is causing high loads back to the pumps, make friction zero at the supports. Third, place anchor restraints on the discharge header/ recycle header to stop external piping loads from pushing on the pumps. This could narrow down possible sources of loads in your piping model. If the excessive loads are mostly in the vertical direction, then it would indicate need for more support of the large valve weights - but your lateral or horizontal loads seem to be the problem from your description. Are the excessive loads occurring in the sustained case and / or in the operating case?

I would also question your process engineer. 38°C is 100°F. Where are Earth are you working that you get 100°F seawater?

The excessive loads are in Operating case. As pump anchor point is below nozzle centerline, hence the loads in Y direction (vertical) are also more than allowable, as first support from nozzle is not acting. In order to reduce vertical loads, I have put spring support at first support location that is between pump nozzle and valve. The loads due to friction are also high even with sliding support. I have raised the discharge header in order to give leg in Y direction in to reduce radial and lateral load on pump. Even then loads on pump are more than nozzle allowable. If I use rigid strut, then the loads comes within limit. But is it viable to use strut and spring for pump piping operating at 38 deg C in order to qualify pump nozzle load?

Another thing I want to know is which formula shall be used to calculate the psv reaction force for water service open discharge system? I haven’t considered any force taking into account the water service.

stressedengineer,

If the excessive loads are in the operating case, and piping is lifted off at the first suppport from pump, then it appears that the pump thermal expansion is causing liftoff at the first support. This could be confirmed by deleting the pump elements while restraining the node of the pump connection as simple anchor - no thermal displacement. The vertical loads would then be much less. The solution would be to locate the pump's base and the horizontal piping support elevation to be nearly equal so that there is almost no differential thermal growth verticallly between the pump discharge connection and the piping support elevation.
The piping support elevation could be modeled more accurately as located below the piping centerline - a zero weight rigid element from center line to the bottom of pipe shoe support might be nearly equal to the pump discharge elevation from its structural base support. If there is pipe reducer, then there could be some more offset from centerline to include for decreasing the differential thermal growth of pump discharge.
The ambient 21 deg C could be adjusted to be closer to the 38 deg C of the seawater to further reduce the differential thermal expansion. It might also require instructions to shim up the first support to closeup gap so it provides support in the operating condition.

Honestly, there's really no good reason to build a Caesar model for a 100°F pump system and this story is an example of exactly why. You indicate that the first Y support lifts off due to the anchor point of the pump being below the nozzle. I expect you are going to find your vertical movement at that "lifted" support to be less that .01". While Caesar will calculate such a movement and tell you all the load has shifted from the support to the pump flange, it's not really going to happen.

That movement is so much less than even the installation tolerance as to be meaningless. Further, there's flexibility at the support/pipe interface, there's flexibility ( a little) in the pump case, there's flexibility in whatever your pump is attached to - all the world is a spring!

Caesar is a very useful tool, when used correctly. It can help us locate problem areas and examine solutions to them. But, it can also tell us that the world is getting torn apart by a tornado when just looking out the window will show a clear sky.

The computer is not a substitute for thinking. The computer does not solve problems, that's the engineer's job.

Alright, that's enough time on the soapbox for me today.

Thank you Edward and Richard. I had thought in similar way as Edward explained and logic of "all the world is a spring" and that’s why I raised post in order to know how to do analysis with help of Caesar and make an analysis report for this type of arrangement.
I talked with client with the help of your reply and he also got convinced for not preparing report as it’s not practical to clear system in Caesar as it requires unnecessary flexibility or use of hangers. Also, the expansion coefficient is as less as 0.2mm/m at 100 ºF and the sea water will not attain even this temperature (100 ºF) but exact temperature can not be predicated.

I have one doubt though.

Practically thinking, Can the axial force on the pump nozzle get reduced due to cushioning elements such as gaskets?

Adding to Mr. Ed's reply, at the most run a SUS case only and check/show the loads are OK if someone demands a computer analysis. Simply state on the title page that the thermal effects are insignificant given the small delta T.

One practical company simply stated in the stress procedures that Formal/Thermal stress analysis is not required for pump systems at temperatures <= 150 F. QED.