parallel pumping of dissimilar pump with VFD 1

[moideen]

What happens when Variable Frequency Drives (VFDs) are used to match the head and flow of two different pumps for parallel pumping? This is not a made-up question; it is a real problem. Thank you ahead of time for taking this into account.

The problem with the world is that intelligent people are full of doubts, while stupid ones are full of confidence.
-Charles Bukowski-

 

[pierreick]

Hi,
A sketch and pump curves will help. Tell us about the background.
Pierre

 

[georgeverghese]

Knee jerk suggestion would be to run each pump on flow control. And split range the master level controller output to reset each FIC to suit level demand at the source vessel (ie use a cascade control scheme with level controller as master). And if necessary, use downstream user constraint control to override the flow controller output.


Describe this system and its constraint parameters, and add the current control sketch also.

 

[1503-44]

You cannot match the pressure AND flow of two different pumps. At least if the difference between them are their pump curves. How could they ever match? OK, maybe they could "match", but only if the two pump curves intersected with the system curve. Getting 3 curves to intersect at the same point could be difficult.

You can try controlling flow, or pressure of each pump,
Or setting flow on one pump and controlling the other to match that flow and accept whatever was pressure discharged, or by controlling to an equal pressure and accept whatever flow was produced.

Since we know you can't match two different pump curves, what defines a "match".

Here's what happens if I try to control pressure in each pump. Say matching means each pump will have the same flow.

1) The same flows will produce different pressures.
2) The different pressures will equalise at the junction point, that pressure being somewhere lower than the lowest discharge pressure of either pump.
3 ) The common equal pressure will energizer the system curve and a system flow rate will be established.
4) The flow rate will be divided into each of the two pump discharge pipes.
5) Each pump will see a new flow rate and adjust to produce its head accordingly.
6) If pump discharge heads do not change a lot, flows will not change a lot and a stable pressure and flow are established in the pump and piping system.
7) if the flows in each pump change a lot, then the system is still unstable, so start over at step 1 and repeat the process over and over until stability is reached. If pump curves are wildly different, or are not suited to the system, stability may be impossible.

B) Control pumps on flow

A process similar to the above occurs. Substitute pressure for flow and flow for pressure in my sequence above to understand how flow will be established.

So, rather than define match, let's just find out what flows and pressures we can do with two pumps and a pipe system. Following those steps will define the answers.

--Einstein gave the same test to students every year. When asked why he would do something like that, "Because the answers had changed."

 

[LittleInch]

You need to look closely at the two curves and the system curve to figure out what is going to happen.

Sometimes you need to start one pump first which has the lowest head curve at 100% and then start the second and slowly increase speed watching pressure and flow. You really need flow measurement on both pumps individually otherwise you are guessing somewhat as to what is happening.

But you can also watch power in tot he pumps.

Not easy without the pump curves and the system curve.

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

 

[1503-44]

Or pressure. Or all three.
And, If not suited to the system curve, maybe you get nothing, or too much.

--Einstein gave the same test to students every year. When asked why he would do something like that, "Because the answers had changed."

 

[LittleInch]

"What happens"?

usually you have some sort of complex control system to try and achieve what you want, but only the pump curves and system curves can tell you if it's actually possible.

You pump curves then also need to be changed to show the curve at different speeds.

Basically it's a mess.

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

 

[MJCronin]

In my opinion, this is what happens when the Plant Manager/MBA gets the bright idea to reuse an old, rusty "beat-to cr*ap" pump that is very different from the reliable, hard working newer pump in service ..

His eyes turn into dollar signs ... He does not care about control details ....

He runs to his bosses and screams "Look at how much money I have saved !!!!"

.... and, when all is said and done, and large amounts of money has been spent on engineering, control systems and testing....

He ends up with a pumped system where he has ONE Olympic Track Star and ONE Senior Citizen pushing on the same yoke ...

His reworked Frankenstein system is far more expensive than a similar system with two identical pumps ...

I have seen this done over and over

Guess which one will fail soonest ???

MJCronin
Sr. Process Engineer

 

[LittleInch]

Oh MJC,

You make my day sometimes with your utter hatred of MBA's.

Hope you're having a good day.

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

 

[MJCronin]

MBAs purposefully made my life miserable for many years, LI ....

I was insulted and degraded on an almost daily basis .... while I was the only one moving projects forward.

I was only trying to earn a living for my family .....

Permit me a modicum of solace, here ... in my declining years ...

Do you disagree with the point that I was trying to make regarding the reuse of pumps in paralell service ?

MJCronin
Sr. Process Engineer

 

[LittleInch]

MJC,

No but we don't know if this is what is happening.

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

 

[moideen]

Thank you for the responses.
I will explain the situation briefly. Currently, the condenser pump flow rate is set at 10,500 gpm with a head of 100 feet. The plant operates with three chillers, each equipped with three pumps, totaling a flow rate of 31,500 gpm. There are plans to add three more chillers, bringing the total thermal capacity to 21,000 tons. To accommodate this expansion, three additional condenser pumps with matching specifications are required.
After careful evaluation, Grundfoss pumps were selected to ensure compatibility with the existing pump, resulting in a combined flow rate of 63,000 gpm from six pumps. At this juncture, a variation was introduced by the project consultant by increasing the flow rate from 10,500 gpm to 12,000 gpm. This adjustment requires the selection of three new pumps capable of handling the higher flow rate, and the existing pumps will be upgraded by adjusting the impeller.
It's important to note that the head remains consistent at 100 feet, while the impeller size differs between the existing pumps (382 mm) and the new ones (415 mm). This change aims to provide a fail-safe mechanism in case of pump failure, ensuring uninterrupted operation. With the new configuration, the collective flow rate will be 60,000 gpm from five pumps, slightly below the previous 63,000 gpm.

However, a client has expressed questions about these changes for several reasons: client's point of view is the following,
1. The consultant's insistence on a spare pump for the condenser raises concerns, especially with six chillers already in operation.
2. Merging two hydraulic systems into a common circuit can lead to complications due to differing shutoff heads and flow rates.
3. The proposed changes could unnecessarily complicate the control system, leading to potential operational challenges.
4. Instead of the proposed approach, a more practical solution could involve stocking spare parts or an identical pump.
5. There is a discrepancy between the required flow rate (63,000 gpm) and the proposed rate (60,000 gpm with five pumps at 12,000 gpm each).
6. Increasing the flow rate to 12,000 gpm with a 600 mm inlet pipe could lead to cavitation concerns due to high velocities.
7. These concerns highlight the need for thorough analysis before implementing the proposed changes.
Your insights on this matter would be greatly appreciated.


The problem with the world is that intelligent people are full of doubts, while stupid ones are full of confidence.
-Charles Bukowski-

 

[moideen]

pump curve-1

The problem with the world is that intelligent people are full of doubts, while stupid ones are full of confidence.
-Charles Bukowski-

 

[1503-44]

No File extension. What is it?
Can you not post an image. Click the camera icon.

6 x 10 or 5 x 12 = 60
Doesn't sound like much of a change so far, but I don't know what's going on with your piping.
600mm ??

--Einstein gave the same test to students every year. When asked why he would do something like that, "Because the answers had changed."

 

[ThatShouldHaveWorked]

They run all three pumps on each chiller at the same time? 9 pumps are currently running all the time?

Your current design will end up running 5? Seems like each triplex skid would have a spare pump on it after the upgrade.

We run triplex pump skids for just that reason. Close valves to pump A and open valves to pump B. Turn the switch and we are in business. Pull the bad pump and rebuild it with the spare parts we keep on the shelf.

The other thing I wanted to mention is the pumps I have on inverters are happier if I start them at 60 HZ for a minute before I ramp them down.

 

[1503-44]

See. Now I thought they had 3, adding 2 and a spare.
But is all that going through a 600mm?

--Einstein gave the same test to students every year. When asked why he would do something like that, "Because the answers had changed."

 

[ThatShouldHaveWorked]

When you say condenser are you actually referring to the chiller's water-cooled condenser? I.E. You are pumping to a cooling tower.

Otherwise you would be talking about what I would call cooled process water. I.E. going to radiators with blowers to cool a room. I'm sure there is fancy sounding term for it.

 

[georgeverghese]

a) None of the files uploaded are visible.
b) "The plant operates with three chillers, each equipped with three pumps, totaling a flow rate of 31,500 gpm."
This means you've got a total of 9 existing pumps. Think you meant to say 3 chillers with 3 pumps total
c) Post a typical sketch of chiller, condensor and condensor pump.
d) Does each chiller plant take source from its own dedicated source ?
e) What is current control scheme - pls post sketch