pump damage?

[Thealanator]

We have a centrifugal pump that is started with the discharge valve wide open. Valve is closed after 10 seconds. Of course, the pumping-of-gravel sound is present while valve is open.

The pump is started once, and sometimes 2 or 3 times a day. The pump runs far back on its curve (about 15 GPM over minimum flow). I have also noted that npshr is 2-3 times what it is at BEP.

We are on our third pump in about 3 years (seal leakage).

Does this seem normal? I'm suspicious of both off-BEP operations. Which do you think is worse?

 

[bimr]

No, neither situation is a normal pump operating technique.

It is normal to throttle a pump when the pump is starting up against low head. Otherwise, the pump will run out.

It is not normal to throttle the pump back so that you are operating near the minimum pump flow. Operating like this will cause the pump to overheat.

The worst situation of the two that you have described is probably the one where the pump is operating under the conditions for the longest time.

 

[1gibson]

Why? I think whoever wrote the startup procedure got that backwards... Start at closed valve, open it after 10 seconds.

 

[SNORGY]

I am not sure what is meant by the statement:

"I have also noted that npshr is 2-3 times what it is at BEP." Do you really mean npsha?

What is the nature of the fluid being pumped?

In any event, if you have had seal failures already and you hear the grevelly sound, you might want to look for signs of cavitation on the impellers and casing. Suggest a pump overhaul.

Premature seal failure can be due to overheating (mentioned by bimr) and / or by shaft deflection caused by unbalanced radial forces arising from operating left or right of BEP, which also increases the heat load on the seal.

Re-examine the hydraulics for the piping system on the suction side of the pump in particular.

 

[Thealanator]

Whew! I brain-f**ted. Instead of discharge valve I should have said vent valve.

The discharge path actually presents a good amount of resistance. The vent valve has the low resistance path. It gets opened to help chill the pump rapidly (cryogen) and then gets closed 10 seconds after pump start. The ten second period is where the gravel comes in. The period used to be longer, but they got overload trips (surprise, surprise).

This pump feeds a high-pressure recip pump. The recip doesn't take enough flow to let the booster run near BEP. I expressed concern about the situation after seeing the pump curve when the pump was proposed. I was told that all would be fine and got subtle hints that my comments were "unwelcome".

I was not kidding about npshr. It dropped from 6 feet at minimum flow down to 2 feet at BEP. I think it rose back to 3-4 feet at maximum flow. I have heard about recirculation near minimum flow. Is added liquid head (increased npshr) supposed to suppress it?

I expect the pump repairs to get more frequent as additional pump sets get operating time.

 

[GHartmann]

Not sure how you determined the NPSHr. I have never seen a pump curve where the NPSHr increased as the flow decreased.

You might consider looking into the "accelleration head loss" created by the reciprocating pump. Your centrifugal might be in the flat head zone where these pulsations cause the pump to have an unstable flow. Hopefully your recips have a suction stabilizer bottle inlet the pump.

Flashing liquids, such as cryogenic fluids can also be difficult to pump and hard on the seals.

What damage do you see in the pumps other than seal failures?

Rarely will a centrifugal pump operate exactly at its BEP. In theory, as long as you are above minimum flow conditions the pump should be OK.

Perhaps you can share details of the pump head, capacity, and type of pump (ANSI, API610, multistage, etc)

Good luck

 

[scjames]

GHartmann, I work as an application engineer for a pump manufacturer. We have some inline, multi-stage pumps that increase NPHSr as the flow decreases. I attached the pump curve to this post. I believe that we have only one product line which behaves this way. It is probably not very common, but it does happen.

 

[1gibson]

Pumps with inducers, or high NSS, can have a "U" shaped NPSHr curve instead of one that rises continually with flow (what most would consider normal, or intuitive.)

For example, multistage verticals, there is often a dedicated first stage, low NPSHr impeller design. It will have higher NSS, lower NPSHr, and you should be able to see at least a small rise in NPSHr at lower flow. Fewer vanes, and usually larger eye area than a series impeller. A typical series impeller will usually not show the rise in NPSHr at lower flow.

 

[Artisi]

scjames, a rising NPSHr requirement as flow reduces is not that common but it does happen. Most mass produced pumps are probably never tested all that well at the low flow end of the curve as there is really no need to fully understand what is happening. NPSHr curves are usually extrapolated from the high end flows down to the lower flows - hence the continually falling curves (usually)
Having work for a major pump organisation who built large to very large custom pumps I have seem a number of rising NPSHr curves as flow reduces. It is all a bit academic anyway as pumps shouldn't be run that far left of BEP - unless of course you like to go looking for trouble.

It is a capital mistake to theorise before one has data. Insensibly one begins to twist facts to suit theories, instead of theories to suit facts. (Sherlock Holmes - A Scandal in Bohemia.)

 

[SNORGY]

Maybe one needs to consider the pump manufacturer's reasons *why* the NPSHr is higher for reduced (near shut off or near MCSF) flows. Is it a condition where the impeller churns and creates bubbles that it can't get rid of? Is it a condition coincident with a droop on the far left of the Q-H curve denoting instability? Vibration? What specifically are they setting as the criterion or criteria for the NPSHr, in the event it is something other than cavitation?