Third Pump - Significant decrease in flows 8

[catplank]

Hello All,

I’m a little stumped the following… any able to offer any insight?

Normal Operating conditions 2x pumps in parallel. Third pump kicks in on level control. When third pump kicks in its flow/current in the middle pump (pump 2) is significantly lower.

Pump 1 & 3 = 40L/s @120A
Pump 2 = 30L/s @100A

If pump 2 is operating under duty conditions it is inline with the other pump (maybe 1-2L/s lower)

I've attached a section view of the pump station, my thoughts is that it needs an inspections to wear ect but I've been told that its always happenened (even other pump i.e. 1 or 3 were doing the same).

 

[TugboatEng]

The standing wave occurs at specific frequencies so changing the velocity highe or lower should break it down. If increasing the velocity improves the pump performance then that would certainly point towards some type of resonance.

 

[1503-44]

From the linked tip of the month...

"... you have to attack the forcing frequency. If it’s a rotating machine you can change the speed, isolate the machine from the pipework or fix the vibration issue such as unbalance or misalignment. If the problem is flow induced or VIV you will have to do the analysis to positively identify the culprit and make changes in flow rate, piping arrangement or even production procedures."

Normally you want to be operating at a frequency well outside resonance, 0.7 to 1.3 x primary modes, which may require substantial change to rpm.

 

[electricpete]

This standing wave proposal is a little far out for me as a source of op's problem, at least in the way described in the link.

The linked tip of the month talks about causes of piping vibration, not causes of degraded pump performance (pump performance is not mentioned).

Part 1 talks about what you might call a mechanical standing wave... or more commonly called mechanical piping resonance. It results in vibration. It does not affect performance. (the comments about isolating source of vibration like misalignment / unbalance apply to this type of resonance).

Part 3 talks about what you might call a fluid pressure standing wave... or more commonly called acoustic resonance. The reason this phenomenon is highlighted again is in the context of troubleshooting piping vibration (not performance). Acoustic resonance would result in pressure oscillations which are minimum at a “node” of the pattern but NO change to the average pressure at that location. I’ve never heard this implicated as a cause of pump performance problem (has anyone else?). Maybe that's not enough to rule it out, but to get a fluid pressure standing wave, you need a source of periodic fluid pressure variation… typically it is a positive displacement pump involved somewhere. Centrifugal pump blade pass frequency pulsations seen on suction side are very unlikely.

fwiw personally (I'm not a pump guy) it sounds like looking at suction NPSH is worthwhile and also the various types of turbulence that can disrupt suction based on piping configuration. But then let me back up on one thing. For op you said this is not a new system but this is brand a new problem? Did you run 3 pumps before and saw equal load sharing? If so then I guess you look closely at what has changed since then and also at degradation mechanisms or foreign material.


=====================================
(2B)+(2B)' ?

 

[LittleInch]

It's see definetly a strange one.

If the pumps are 20 years old the no 2 pump may have been used more? Or just has more wear?

Only if you can run a pump curve test for each pump will you see if the pumps are in fact not equal. Maybe at higher flows you didn't notice the difference so much as the curve is steeper.

Another issue is whether at the higher flow rates are you getting vortex issues at the inlet which is somehow affecting the no 2 pump more? What's the liquid height above the inlet nozzle in the tank?

A bit of testing will either eliminate some of these or show up the issue.


Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[1503-44]

I think it may have more to do with vortex shedding off the first or second header outlet. Not a standing wave.

 

[Artisi]

All becoming too academic, run the pumps and get results for the various operating configurations - you might start making some sense.

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.)

 

[catplank]

Just added a couple more of photos.

I am hoping to get testing done in the next week or so. Frustratingly the pressure guages we have on site are more discharge related & scaled 0 - 2500kPa, not useful on the suction side..

 

[MJCronin]

New culprit ... that cheap-ass butterfly valve on the suction piping to the number two pump.

Is it fully open ??? Is it defective an cannot be turned fully open ? Has the butterfly disconnected from the stem ?

Is it a new valve or did you get it from the bone-yard or take it from another project ?

After you compare actual suction pressures among the three pumps, it would be lots of fun to take this valve out and examine it.

..... or better yet, swap the suspect number two butterfly with the butterfly of either two properly operating pumps.... see if the problem moves with the butterfly ...

My opinion only


MJC

MJCronin
Sr. Process Engineer

 

[Artisi]

For interest, is pump 2 the same model as 1 & 3, different bearing frame and appears to have a different volute, although not really clear

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.)

 

[LittleInch]

And what exactly are those large blue valves in the discharge line?

Look suspiciously like pressure or flow regulating valves to me ...

If pump 2 is new, even if supposed to be the same model, it rarely is to any level of accuracy.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[LittleInch]

Same issue applies with those gate valves.

Placing them horizontal is never a great idea and it doesn't look like you can easily judge how open they are.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[catplank]

Hi all, to confirm a couple of comments.

The butterfly valves on the suction side are all operational - doesn’t seem there is any obstructions.

The gate valves on the discharge side are all fully opened.

The valves that look like flow regulating valves are NRVs.

I was able to complete some minor testing today. Probably muddies the water a little bit (I didn’t have the equipment to record pressure which is a pain) and I was only able to run each pump on its own (there is access issues ect which can be difficult)

Running pump 1 on its own - 55Ls
Running pump 2 on its own - 63Ls
Running pump 3 on its own - 53Ls

I feel I need to confirm pump data with the supplier… I’ll wait to complete the full testing when I have appropriate pressure gauges.

 

[1503-44]

That would help!
+24% difference

 

[Artisi]

Next step is pump curve for #2 and a head pressure for 3 operating.

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.)

 

[LittleInch]

That flow data is suspect to me. 63 is well off the end of the curve.



Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[Artisi]

LittleInch, good point, 55 is also off the curve for 1 & 3 and 2 is an uknown - - so we are still guessing.

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.)

 

[1503-44]

OFF THE CHART!

Reboot!!

 

[electricpete]

If normal operation is 2-pump or 3-pump, and op moved to 1-pump operation just for data collection (but did not have ability to adjust system demand), then it wouldn't be surprising that the single-pump operating point is to the far right on the curve. So the data isn't necessarily suspect to me, but I understand that the operating point so far to the right would make it impossible to compare individual pump performance to the published pump curve, even if we had a head measurement.

At any rate, for me those measurements elevate the scenario that the performance curve of pump 2 might be different (maybe compressed down vertically and stretched out to the right horizontally), when combined with Artisi's observation about pump appearance.

=====================================
(2B)+(2B)' ?

 

[danschwind]

He can still get the individual pump performance by throttling the discharge.

It does seem that pump 2 has a different curve than the other ones - probably a less steep curve that whenever you get to a 3 pump operation it goes to a very horizontal part of it and flow drops a lot.

Daniel
Rio de Janeiro - Brazil

 

[MedicineEng]

If pumps are the same, swap pumps' positions and measure flows again.
That would give you a clear indication if the problem is pump or position related.