Pumping Problem - Any Ideas?

[ChocLab]

Hello All,

I have a pumping system that has been giving a problem the past little while and we cannot seem to find the reason as to the why.

The System:
3 pumps (A/B/C)
Each pump rated for 6100 usgpm
Pump type - double split suction centrifugal
Primary Mover - Electric Motor - 3 phase 4160 VAC

Process Conditions:
Fluid - Water (80 F)
Common inlet header pressure - 80 psig
Common discharge header pressue - 450 psig
Flow - currently 12,000 usgpm, 14,000 before the problem started.
Pump A and C inlet pressure - 75 psig
Pump B inlet pressure - 35 psig
Pump A and C Motors - 200 amps
Pump B Motor - 150 amps

Additional Information:
1. There has been no obstruction found between B pump suction and the common inlet header.
2. Pump suction strainers have been cleaned.
3. Past problem with these pumps is that the seals failed between suction and discharge resulting in internal recycling. A temporary seal replacement was done and the pumps worked fine for 7 months until this happened 2 months ago.
4. Minimum suction pressure required is 50 psig.
5. The 150 amps on B pump indicates that the pump is doing less flow than A and C (probably due to the reduced inlet pressure to this pump).
6. A D pump is currently being installed but I do not think it is involved problem as the system had this problem before the D pump installation was begun.
7. I do not think the suction/discharge pump seals have failed as the symptom to this was high pump current (250 amps) and low flow rates.
8. Picture located at:

http://i203.photobucket.com/albums/aa219/ptstacey/39s.jpg

This one has me stumped. Any ideas are greatly appreciated.

Regards,

KL

 

[dcasto]

dodoyrog,
You cannot lower the energy suppy to a pump, the pump creates the demand for energy and will get it from the motor. If the motor can not supply the energy it will fail (or breakers trip). On an AC motor, if the voltage drops, the amps go up, the breaker trips.

On a DC motor the the voltage drops and the PUMP slows down and then you could see the reverse. So unless this is a DC motor....

 

[dodoyrog]

It is a very reasonable to assume that choclab water pumpset has 3 identical pumps (as shown in his sketch) with identical DC motors. An AC motor, purpose of which is to deliver a variable flowrate as required by on-off users, is only installed in systems foreseen with high demand fluctuations to prevent pressure surges and consequent piping failures, for example, in firewater system with 2 DC motors and 1 AC motor. The cost of AC-driven pump is very high compared to a DC-driven pump, so I am sure that all choclab pumps have DC motors.

Also, choclab should be aware that a 3-pump system is designed for 2 pumps duty/1 pump standby, and never 3 pumps duty as it is operating now. Meaning, that the common suction header of pump A/B/C was originally sized only basing the flowrates for 2 pumps. But with 3 pumps now in operation there is a higher-than-design flowrate (PLUS) as the suction line gets older, its internal roughness increases and its C factor changes (EQUALS) higher flow resistance, lesser available flowrates, and suction starvation to least-efficient pump/s. To determine if the suction line is adequate for 3 pumps, choclab should figure out the current water velocity in the suction line and compare it with the acceptable pump suction velocity range as given in many references. If choclab calculation is significantly higher that the reference, you have hydraulic overloading in your suction line and that is the cause of your problem.





doodscasay

 

[BigInch]

Hummm. 3 Phase DC?

"Mover - Electric Motor - 3 phase 4160 VAC"

http://virtualpipeline.spaces.msn.com

 

[kemot]

Hi Choclab

I did a simulation of these pumps and found them OK.
Acc. to me,B pump suction pressure gauge is faulty and B pump is OK.
Such big pumps got very low curve slope and you can reduce flow and you did not notice any discharge pressure drop.

B pump got some restrictions on discharge pipe and that/s why the flow through it is smaller , it can be discharge valve problem or the discharge line is fouled.
Another advice is that,
I suppose that impeller diameter inside is rated one rand there is possibility to replace it with max. diameter impeller to increase capacity all the pumps instead of to add a new D pump. Check the pumps data sheet and the impeller diameter or ask about it your maintenance gay.
You can ask your Vendor too.

Regards
kemot

 

[BigInch]

kemot, I'm curious as to how you managed to do a simulation, not knowing any pipe diameters and lengths at the suction and discharge header, no downstream system curve details, no valve info, or any pump details other than flow, inlet and outlet pressure and amps?

http://virtualpipeline.spaces.msn.com

 

[rmw]

I'd like the OP to post an actual piping configuration in some form. Often the real piping is much different than a simple sketch like we have seen.

I think there might be some nuggets of information there for those of us who are following this thread and scratching our heads.

I'm thinking there might be venturi effects.

rmw

 

[BigInch]

If the pump curves are flat, it may just be some divergence from the ideal condition where the system happens to reach a stable operating point. Difficult to simulate those effects without precise info (even with it).

http://virtualpipeline.spaces.msn.com

 

[kemot]

Hi
I did a simplified simulation. It can be done in Excel .
Suction line DP can be expressed as k*V^2, whereas k-coefficient and V-as a flow.
If we know that suction lines DP for A,C pumps are negligible so the "k" is equal to almost 0. The same estimation we can set for discharge lines and header.
If we have a substantial DP for B pump so this "k" should not be neglected.
A pump equation we can set A -B*V^2 where A is equal to a diff. head with flow =0 and B is slope coefficient.
Now you can made one test.
Having three pumps running , it is possible to stop pump B for a couple of minutes and recording suction pressure.
I stopped sea water intake pump(1200m3/h capacity) several times while 2 other pumps were running on a previous contract so is to be a safe experiment.

If then B suction pressure increases to a value about 80 (comparable with A&C suction pressure) psig that would mean that we have substantial obstruction inside B suction line
in spite of a suction strainer is reported to have been cleaned earlier and suction valve is OK.
If we got the same B suction pressure(35-40psig), pressure gauge is faulty.
To coming back to a simulation problem, you can write simple equations for suction and discharge lines as described above.

The pump curve one point we already have (6100 uspgm with 450 psig)
The second point we can estimate i.e. 500psig with 0 uspgm flow . Having these two points we can calculate A&B coefficients of pump curve.
That's seems to be enough to solve these equations and understand the problem.
If we find there is no obstruction inside suction line found then we can start investigation of discharge line.
From my experience I met several times that discharge check valve happened to be obstructed, generally check valves should be checked periodically.

B pump 150 Amps current means this pump is underloaded, and might be about at min. flow limit and check it with the Vendor please.
I suppose to clarify my point.
Regards

 

[BigInch]

But I think its not a problem with the pump curves.

http://virtualpipeline.spaces.msn.com

 

[kemot]

BigIinch, such calculations and drawing pump curves will show you that there is pump internal recycle or anything else is going on wrong if your operating points don't match the curve.
Acc. to rated flow 6000 uspgm, estimated suction line nominal diameter should be about 30" with 2ft/sec velocity. In such cases , butterfly valves is the most likely to be used. I had a case that internal rubber lining of such valve has gone and pieces of worn rubber increased noticeably DP in suction line.
Another possibility to solve the problem is to request to have gamma scaning done on the suction and discharge lines of B pump by a suitable service.
Such service is now easy available and they will find where is a obstruction causing this problem.
Regards

 

[BigInch]

I don't really consider checking an assumed pump operating point a simulation, but my point was that with all your curve kv^2 assumptions, you're still no better off than what you already know, pump B is flowing less (lower amps) and the lower suction pressure indicates suction line problem. The lower psig says that and lower amp use verifies it. Anything further, such as 30" diameter, 2 ft/sec, worn rubber, partially closed valve, broken valve, venturi flow or unbalanced header flow from poor piping configuration, etc. is 100% speculation, none of which can be deduced or confirmed from your limited simulation. Might be a crushed pipe, hardhat or dead animal too.

http://virtualpipeline.spaces.msn.com

 

[kemot]

BigInch, sorry but you are just tempting me to answer you that I would feel much better if I could know the pump curves.
As you know every pump undergoes a water performance test at Vendor site to prove its capacity and efficiency.
So this is the similar situation, if I knew performance curves of these pumps I could exclude them from further investigation and concentrate to identify a problem on both suction and discharge pipeworks.
There are many causes for pumps to decline their performance like already mentioned internal circulation, also impeller could be fouled or damaged by debris, etc..
If I have problems with pumps , compressors and so on I try to involve mechanical engineers immediately to assist me to solve it due to their duties.
As far I can see, most of us are chemical engineers and we are not enough qualified to sort out all mechanical problems.
Regards

 

[BigInch]

Kemot, All I was ever trying to say is, if its not a pump problem, we should get the piping details.

http://virtualpipeline.spaces.msn.com

 

[StressGuy]

Hmm, too bad it looks like this thread dried up. It definitely sounded like an interesting problem.

I too would like to see a picture of the piping configuration. The OP never stated that the problem started after the pumps had been successfully running for some time. He also never stated that the system was not designed for three pumps normally running. It is a bad assumption at this point that this is supposed to be a two pump + spare that is now working in three pump service.

Edward L. Klein
Pipe Stress Engineer
Houston, Texas

"All the world is a Spring"

All opinions expressed here are my own and not my company's.