Centrifugal Pumps Operating in Parallel 3

[Rha257]

Hello Everyone,

What do you guys think about adding a different pump model from a different pump OEM next to a set of fixed speed pumps running in parallel considering that the new pump will also operate in parallel with the existing pumps? Of course the new pump will be designed for the same rated point as the other pumps, however, being a different pump, the performance curve of the new pump will be somewhat different.

Do you think it could cause operational issues?

What will happen if the back pressure changes in such a way that shifts the operating point of the pumps to a point different than the design point? Will the new pump allow a different amount of flow than the other pumps but maintain the same discharge pressure?

Will it make any difference if the new pump is fixed speed or variable speed?

Thanks in advance!

 

[bimr]

It is not possible to predict what will happen if you add a third pump without knowing the system discharge head. The increased pressure from the additional pumped volume will shift the operating point on the pump performance curve to the left. The degree of the shift will depend on the pumped volume and the increase in system discharge head.

 

[LittleInch]

Parallel operation works best when the pump curve is quite steep. Any change in head then makes little difference in flow between each pump.

Parallel operation, even with so called identical pumps results in different flows. So a new pump isn't much different.

Whatever happens if the pumps are parallel they will all have the same discharge pressure.

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

 

[goutam_freelance]

You need to match the pump characteristics of old and new pumps. All should have drooping Q-H curve. The curve slopes should also be similar otherwise there will be uneven load sharing between pumps. Introducing variable speed pump may lead to instabilities in the sense that if the speed becomes lower than the required, pump flow of new pump will sharply reduce Or even may become zero.

Engineers, think what we have done to the environment !

https://www.linkedin.com/in/goutam-das-59743b30/

 

[mech-engr-experience]

You should be sure to have a certified pump curve from each OEM manufacturer and know your system curves for what you are trying to obtain. I would look in the Cameron Hydraulic Data Handbook at the end of chapter 1 for what typical curves look like. You need to know the terms the pump industry uses to get the answer you require to have a successful installation.

If your first pump in parallel, it would benefit you to partner with the people selling you the pump or pumps and have them assist.

The are number of variable involved and mistakes can be a real headache.

VFD drive for small (approximately) +/-20 RPM tweaks can always help during start-up - I have worked on a few parallel applications and you really need to do a failure mode analysis of the system to be sure you do not miss something, as nobody can truly know your application fully.

(Whoever deleted my previous reply to this post for this thread - I was NOT amused)

 

[LittleInch]

Occasionally posts don't register but unless your reply was somehow offensive I don't think it was deleted.

That's a good point about pump curves. Only if you get a specific test on your actual pump can you trust the curve. Otherwise the accuracy can be quite large compared to the theoretical curve but still within the specification.

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

 

[Artisi]

You really need to give us more data for a meaningful answer.
Pump curves.
Static head inlet and discharge side
System curve or data to calculate total head with the intitial pump operating.
Given necessary data it's a simple exercise.

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.)

 

[Comm5]

You can use Pump System Analyzer software (

https://pumpsystemengineering.com/projects/)

to evaluate the behavior of the combined pumps. You will need their head-capacity data.

 

[Marke]

In addition to the pump curves, you need to consider the hydraulic circuit.
The flatter the Q-H curve, the greater the effect that the hydraulics can have. Connections to the manifold should all be the same, a swept (curved) connection will offer a different resistance to a "T'd" (square connection). Steep curve pumps are far more tolerant of hydraulic variations.

Mark Empson
Advanced Motor Control Ltd

 

[KevinNZ]

I have seen this problem solved by using variable speed drives and matching the pump motor torque. One pump is setup as the leader. The process control loop uses this pump. The other pumps are the followers and match the lead pump torque.

 

[1503-44]

Oh yes. Ive seen that before. Software can fix anything, even a bad design.

 

[Wkcoo]

For the best analysis, use system curve (try to get if from process engineer) and the pump curves to predict what will the new operating point be.

Practically speaking, the new pump head curvd should not differ that much of the old pumps. If say the new pump can deliver higher pressure, it will force the existing pumps flow to decrease to match the discharge pressure, and may operate below min stable flow without knowing if the instruments are not too complete.

Using VFD may solve the issue, but i don't have much background in this.

 

[accapumps]

VFD will make it work; however, check on the efficiency of the pumps with VFD.

 

[Rha257]

Thanks everyone for your feedback. The intention behind the query is only to know, conceptually, if two different pumps with relatively different performance curves can run in parallel with no major issues.

As I understand from the response, the discharge pressure of both pumps is governed by the backpressure seen on the discharge side so the discharge pressure of both pumps will be identical (as long as both pumps are running in the stable operating region). However, the flow may not be the same for both pumps due to the different performance curves. Please correct me if my understanding is not correct.

Practically speaking, can the addition of a VFD correct or improve the difference of flow in both pumps somehow or it simply does not make a difference? Please note that I am only focusing on the difference in flow rate and not thinking about the efficiency, discharge temperature, energy saving, cost, etc.

 

[Marke]

Hello Rha257

Yes, you are correct. The actual relative performance of the two pumps is entirely dependent on the pump curves and the hydraulic circuit.

If you add a VFD to one pump, then you can adjust it's speed and effectively change the impellor diameter at any point in time. This will enable you to increase or reduce the flow in this pump and theoretically, gives you the ability to "match" the flows at all points on the curve. There are several schemes for achieving this, but I would caution that while this is theoretically possible, and I have done this on many occasions, there can be problems. If the pressure/flow curve is very flat, then a very small change in speed will make a significant change in flow and can yield an open loop gain of speed against curves such that it is very difficult to create a stable result. If there is a slope of the pressure / flow curve, there is more impedance in the system and the effective open loop gain will be lower and a closed loop system will be more stable and track more reliably.

You can use two flow meters, one in series with each pump and use a "flow" follower arrangement, or you can compromise and use a torque follower scheme which will generally give a pretty reasonable result.

Mark Empson
Advanced Motor Control Ltd