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]

What controls the discharge head on the pumps and what is the discharge head at 100% and what is the discharge head at the reduced speed?

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.)

 

[TD2K]

Building on what Artisi said, determine the actual head that the pumps need to develop based on field measurements and compare that to the design point of the pumps.

At this point with no data I realize it's speculation but I wouldn't be surprised if you find out that the "design point" the pumps were designed to operate at (x m3/hr @ y m head) is quite a bit different than the conditions they are actually operating under.

 

[MikeHalloran]

What was the suction head at test conditions?
What is the suction head at your installation?


Mike Halloran
Pembroke Pines, FL, USA

 

[BigInch]

At low flow the system piping (without additional controls) provides far less resistance, roughly proportional to (q/Q)^1/2 thereby allowing speed and discharge flow to increase, otherwise known as "run away" flow.

Get those throttle valves installed and working and you should return to normal.

 

[georgeverghese]

The pump at 100% speed produces 575m3/hr at a differential head = H1
At 60-90% of full speed, it produces 680-700m3/hr at a differential head = H2

This can happen only if H2 << H1. If this should be corrected, then find out why imposed backpressure / discharge pressure is so low at low speed.

 

[Artisi]

Just to summarise, the flow will be a result of the head imposed by the system (head,friction etc). The design and factory testing of xx@yy is rarely ever met as an installed condition, but that is another discussion.
If the site requirements are being met in terms of flow, the head and speed are irrelevant. If site head is lower than design, flow will be higher. It is not unknown for a pump spec.to be wrong or to have a wrongly specified / selected pump.
The site conditions need to be analysed and the pump selection to be re-evaluated so an engineering decision can be made to address the problem.

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.)

 

[bimr]

It is apparent that someone over estimated the system discharge head.

The pumps specification sheet is accurate. The pumps are operating at a different point on the pump curve because the system head curve has less headloss than expected.

 

[LittleInch]

Oldox you don't mention head or pressure in your initial post which is key. However the data you give is very odd. If an increase in speed from 58% to 86% makes only 30 m3 difference in 600odd then this is a very steep system curve and would not seem to suit a vfd.

What is the system this pump is feeding?

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

 

[Artisi]

Oldox hasn't mentioned anything following the initial post, so either the problem is solved / gone away or wasn't really a problem.

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]

I appreciate so much here for your kind response and help! Sorry I couldn't provide some pump ops data last weekend as they are at my office. Yes, I used to witness once locally that the local suction pressure gauge showed 80 psi and discharge pressure gauge showed 250 psi. DCS recorded flow was 690 m^3/r and speed was 87% speed. The pump head is calculated as 128 ft which pumping fluid is oil emulsion. But I haven't made converted curve yet to look if it is still on the pump curve. I have some following methods I am thinking, one is just to make a converting curve to look it running point; or operate the pump at 100% speed to look how much flow it will get; or last one to make a real system curve then konw the BEP. Do I need to do all of them or just one is enough to verify the system curve is too flat? Thanks again!

 

[BigInch]

With one speed, flow & head point you can guess. With 2 a better guess. With 4 points, 2 at different speeds and flows, should narrow it down to a pretty good idea of what's going on there.

 

[Oldox]

I will try. But the system curve is most suspicous which I worry about the pump could not be run at 100% speed as too much flow than rated. Thanks BigInch!

 

[MikeHalloran]

Um, 170 psi differential head comes to 392 feet of water.
What's the s.g. of that emulsion?


Mike Halloran
Pembroke Pines, FL, USA

 

[BigInch]

You wouldn't necessarily have to run any of those points at 100%. Try 85% and 60% speeds for example.

 

[Oldox]

Sorry, the head should be 128 meter (422 ft). S.G is 0.93

 

[georgeverghese]

In some pumping systems where the pump could go down to end of curve as the imposed backpressure drops (due to some failure or change in operating conditions), is to ensure the motor HP is good for end of curve operation also - in this case, that would be EOC at 100% speed.

 

[Artisi]

It is apparent that the system design and the actual operating conditions do not agree with each other - WHY? that is what needs to be established.
The very first step is to review design conditions compared to site conditions for variance.
Check the pump selection against the design conditions. What was specified to the pump supplier as the pump duty requirements, is this the same as the design requirements?
Did the pump supplier supply the pumps to the specified design duty.

At this stage anything else is just spinning wheels - you need to establish what was wanted against what is installed. Once that is known, any necessary changes can be reviewed to establish what should be done to correct the problem, that's if any problem actually exists or changes need to be made.
Is the pump running over capacity at a lower speed a problem?

Note that your OP asked why flow was higher and speed lower - well the answer lies in the above - once that is answered you can decide what to do next.

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 Artisi and georgeverghese! The pumps have had some vibration but I don't exactly know what were the running condition during that happened but I will talk to the vibration technicians. The document of original design seems not easily to be got it as it is a old plant but I will try. But the reality might have changed referring original requirment. By the way, the problem for this issue has being asked by Ops, do you think the good explanation is to creat the real system curve as I asked before although it might not be easy?

 

[LittleInch]

I just re-read the OP and it says you have three VFD drives in parallel?

If you have common inlet header and outlet header then you might just be getting one pump running at a different place than the supposedly identical parallel pump even if running at the same speed. If you have vibration that often indicates running at end of curve - too much flow and not enough back pressure from your system

Artisi has it correct (as usual) - first establish data from the pump side and the system side. Obtain as many flow / pressure points as you can and understand how the system control works and then compare.

If you can draw out a schematic of your system that would help us and you understand what is possibly happening here

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