Why VFD pumps produce flow more than rated flow at low speed 2

[Oldox]

I newly joined here and hope to get help. We have three identical centrifugal pumps with VFD parallelly installed in a system. The pumps are designed and tested at 560 M^3/hr @ 100% speed (1750 rpm); But these pumps could get 676 - 699 m^3/r at only 58.9-86.7% speed. What did it happen? Does the system curve too flat ( I don't think the throttle valves are installed on discharge piping). What should I do for that? Thanks a lot!

 

[Artisi]

So, it's an old plant - has the overcapacity always been evident / problem?
Has anything within the circuit been replaced, removed, or otherwise altered.
Is there any flow regulating system / equipment in the circuit?

My signature at the end of the message holds true for everything in life and pump systems .

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.)

 

[georgeverghese]

Agree, if the pump Q-h curve is flat as you say, then there would be significant difference in flow between these pumps if there is even a small variation in imposed differential head. You may need to manually chock back the pump that is running EOC for the time being - ideally, there should be an FIC loop on pump discharge on each of the pumps (FIC modulating VFD). If there is a master PIC on common pump discharge, you could run these FICs' in cascade mode with setpoint reset from this PIC. Include hi limit in FC setpoint for each pump to stop the running pump from going EOC (at full speed) when any one pump trips.

 

[Oldox]

Thanks to everyone again! We don't have any control valve installed on pumps discharge piping which is different from the pump with constant speed system. If I try to avoid creating (calculating)the system curve which seems a little complicated and I have never done it before, and also if Ops could accept reduced flows, I am thinking to manually chock back pressure, as georgeverghese mentioned above, at BEP during pumps running at 100%. Then redseal those chock valves. Pumps design required system (curve or resistance) will established, pumps could continue run with VFD, but along with the system curve. Do you think it is a simple and straigt way to solve this problem?

 

[LittleInch]

A system curve is quite easy providing you can vary and accurately measure pressure and flow at the common discharge point. Two points migh get you somewhere, but three, four or more over your normal operating range wold be better. Then you need to overlay with your pump curves to see how flat or steep they are and where they intersect

Only then can you see where your issue lies.

We can't see your system or know where the pressure gauges / transmitters and flow figures are being measured, hence it is impossible to say much more.

I don't actually know what your problem is - is it the vibration / running at end of curve or what? Throttling and running at 100% just wastes energy, but yes it could work to get you back onto a better place and depending on how far out you are, might actually use less power due to better efficiency.

Do the work, take the readings, create the system curve and then look at the pump curves. Then the solution will become apparent.

BTW how many pumps are running at once? Does each pump have its own flow meter? A decent system sketch / PFD would help A LOT.

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

 

[Oldox]

Thanks LittleInch, it looks you encourage me to creat the system curve and I will try . It might be more apparent to over reveiw whole system. To trigger this issue is that we have had mechanical seal and bearing failures and we figure out vibraiton is one of the factors, and then we reviewed the operation data and found this concern. We run two of three pumps most of time and they have their idividual flow meters and suction/discharge local gauges as well.

 

[LittleInch]

Do you have the pump curves for the different speed ranges, or even just one speed

Can you post them?

It sounds like you have a mismatch between pump and system and the only long term solution is to fix that issue.

Read and understand the posts by artisi and look up parallel pumping on this forum - there is lots of information there.

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

 

[Oldox]

We only have the pump curve with constant speed as attached. I will make the curves at different speeds. Thanks,

 

[LittleInch]

Interesting curve - pretty flat.

Any idea why the design is 560 @ 143m , but performance test is 556 @ 150m?. Probably "within spec" but clearly you're starting with a pump delivering more than the design.

Given your pump curve stops at 619 if you're operating at nearly 700 you might run into some issues, but it is strange that the efficiency curve doesn't flatten off.

If you provide that serial no to the pump vendor you might be able to get some better information on these units.

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

 

[Oldox]

Yes, I have sent email to pump OEM ysterday. Thanks!

 

[georgeverghese]

Previous response was based on a pump with a fairly responsive Q-h curve, which isnt the case now
Send us a process description with process controls brief on what is happening downstream and we'll see if we can save these pumps - we'll have to pull some other tricks out of the bag for this one.

 

[Artisi]

Time to employ a qualified pump engineer to review the operation and sort out all the conflicting information - such as earlier info about 80 psi inlet pressure etc.

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.)

 

[Oldox]

Thanks! I will review all the system when I create the system curve, and also will discuss with ops about the downstream real requirement, such as flow and pressure.

 

[Artisi]

Initially all that is required is an accurate flow and an accurate total head at the current operating condition and an understanding of what the initial operation philosophy and what is is currenrly, and why are the units speed controlled - was this to balance various flow conditions by a combination of 1,2,3 pumps in operation?

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.)

 

[Oldox]

Well, I willl discuss with ops. The system which is located at a steam plantis that the pumps pump oil emulsion from a tower where the emulsion comes from oil well to a process plant which is located around 20 km away. The system seems a bit complicated. Thanks!