PD vers Centrifugal pumps 4

[sheiko]

Dear pros,

I used to believe that either Positive-Displacement or Centrifugal type pumps (and compressors) create flow and not pressure. That is, they conform their discharge pressure to the process' total back-pressure at their discharge flange.

Now I have come across a serious article from IMO pumps that claims the above statement is true only for PD pumps, and that centrifugal pumps create PRESSURE (and not flow).
Here is the article (see § "system response"):

http://www.warrenpumps.com/resources/pdpumps.pdf

It is mentionned that:

"In a PD pump, pressure is created by the system's response to flow.
If there was no connection at the discharge flange, the flow would exit the pump at atmospheric pressure.

Centrifugal pumps create pressure by first imparting velocity to the fluid with the impeller, then converting the velocity to pressure with the volute.
If there was no discharge flange connection, the flow would exit the pump with that developed pressure."

So my questions are:
- What is your opinion? do centrifugal pumps create FLOW or PRESSURE or BOTH? and could you please explain how?
- Do you confirm that the same reasoning is applicable to compressors?

Many thanks in advance.

"We don't believe things because they are true, things are true because we believe them."

 

[hydtools]

quote "Centrifugal pumps create pressure by first imparting velocity to the fluid with the impeller, then converting the velocity to pressure with the volute."
Sound like they are saying centrifugal pumps create flow. 'Imparting velocity' means to me, create flow. Just like positive displacement pumps, no flow, no pressure. If flow but no system resistance, no pressure. Centrifugal pump does impart dynamic pressure due to spinning fluid in the pump.

Ted

 

[Pumpsonly]

Fluid flows from 1 point to another because of pressure difference between the 2 points.
The difference between PD pumps and centrifugal is in the different way they impart energy to the fluid.

From what I read,the main point of the article is trying to sell the PD pumps advantages over centrifugal on certain application.

 

[jmw]

You need to look at the construction of pumps to arrive at an understanding of how they work and what this means.

In a centrifugal pump there is a flow path through the pump and past the impeller that will allow flow whether the pump is running or not and which allows the impeller to rotate whether there is a flow path or not. Typical examples to look at are pond pumps, washing machine pumps etc.

In a Positive displacement pump the impeller(s) are such that there is no flow path through the pump except if the impellers are turning but, if there is no flow path and the pump attempts to run then there will be breakages.
An example is the oil pump in your car.

A simple way to look at it is to consider:
A centrifugal pump will deliver a flow that is determined by the amount of pressure it can generate and the pressure drop in the system.
Pressure drop can change with viscosity. So for a higher viscosity fluid the flow rate will be lower than for an identical system with a lower viscosity fluid.
Alternatively, for the same fluid and viscosity, if you increase the pressure drop in the system by modulating a valve, for example, the flow rate will change.

So if you have a valve in the discharge and progressively close that valve, the pump will continue to operate but the flow rate will reduce.
This is because the impeller can simply rotate in the fluid. (If you don't get flow, you'll get heat.)

A PD pump will deliver a flow rate that is governed by its speed of rotation (upto the capability of the pump motor).
If the viscosity increases or decreases the flow rate will remain the same (discounting slip flow) but the pressure will vary with the viscosity.
And if you put a valve in the discharge of the pump and progressively close the valve the flow rate will remain the same but the pressure will increase.
If you attempt to shut the valve fully then something will break.

PD pumps usually therefore have a pressure relief bypass valve. This is to protect the pump, pipework and fitting.
It is often internal and at a certain pressure opens to divert the flow from the pump outlet back to the inlet.

Flow rate can be varied by diverting come of the flow back to the pump inlet or by varying the pump speed. One way to do this in the design is through gears. There are now a range of VFCs (Variable Frequency Controllers) that can be used to control pump motor speed.

To see some animations of pumps, look at:

http://www.liquid-dynamics.com/animations/pumps.htm

or for a nice Roots style go here:

http://www.mekanizmalar.com/roots_three_lobe_helical.html

What you see in this animation is a blower or pump with two interlocking rotors. These rotors have a twist in them so that they are self meshing and do not require external synchronisation gears.

In all the centrifugal pump animations you will be able to trace a free flow path through the pump. In the PD pump animations you will see that the rotating elements provide a moving seal. If the rotors don't turn then they totally obstruct the flow path.




JMW

http://www.ViscoAnalyser.com

 

[DubMac]

For every revolution of the driver, PD pumps take a finite volume of fluid from suction conditions, isolate it briefly, and then transfer that volume to discharge without regard to discharge pressure.

It doesn't matter what the discharge pressure is; the pump is going to transfer that volume regardless, i.e. "positive displacement". If discharge valve is closed, then either the motor will shut down or something will go Kablooey.

That is why PD pumps are used as metering pumps; you know exactly how much fluid is moved for each motor RPM. For a given speed, they are CONSTANT VOLUME PUMPS.

Centrifugal pumps obey a different set of rules. Their flow is completely dependent upon, and responds to the discharge pressure encountered.

The most important rule you can remember about centrifugal pumps that will always hold true is:
A centrifugal pump will always operate at the intersection of its performance curve and the system curve to which it is applied.

ALWAYS.

 

[Artisi]

The most important rule you can remember about centrifugal pumps that will always hold true is: A centrifugal pump will always operate at the intersection of its performance curve and the system curve to which it is applied. ALWAYS.

Except in those cases where some people don't believe or accept that fact [haipull3].

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.)

 

[Artisi]

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.)

 

[katmar]

I am guilty of often advising people to consider a centrifugal (or PD for that matter) pump as a creator of flow rather than pressure. I do this because it is a way of shocking people into opening their minds.

You can say that a pump with no piping has no discharge pressure, so it must be the piping that causes the pressure. But equally you can say that a pipe with no pump has no pressure so it must be the pump. It's a bit like Shroedinger's cat that is both alive and dead at the same time. Both ways are equally valid, but in some cases the one view will be easier to understand and more useful than the other view. Like Artisi said, in the end it is the pump curve that has the final say. It is for those who don't believe in pump curves that these shock tactics help.

Katmar Software - Engineering & Risk Analysis Software

http://katmarsoftware.com

"An undefined problem has an infinite number of solutions"

 

[sheiko]

Thanks for your answers!

Reading your answers, this the way I now see it:

- all pumps create flow
- pressure measured at dicharge flange connection is the system total backpressure
- depending on the pump's type (performance curve), a variation of this system's backpressure may affect (case of centrifugal pumps) or not (case of PD pumps) the pump's flow

In other words, centrifugal pumps don't create pressue.

Do you finally agree with this?

"We don't believe things because they are true, things are true because we believe them."

 

[jmw]

In other words, centrifugal pumps don't create pressue.

Do you finally agree with this?

No.
They create pressure.
They just don't create an irresistible flow.

Both pump types put energy into the system.
It is the mechanism of the pump that determines how that energy is converted and used.


JMW

http://www.ViscoAnalyser.com

 

[DubMac]

This philosophy class has gone on a bit long; sorry I'm due back on the planet Earth.

 

[sheiko]

DubMac, thanks but please feel free to not attend the class.

Dear Katmar, what if there was no pipework downstream a centrifugal pump?
Would the discharge pressure be atmospheric, or equal to the developped pressure in the pump (as mentionned in the first article I referred to)?

"We don't believe things because they are true, things are true because we believe them."

 

[katmar]

The head developed by a centrifugal pump, and as is displayed on a typical pump curve, is the "total dynamic head - TDH". At the point where the pump curve and the system curve intersect the pump is delivering the TDH applicable to the flow rate at that point. The pressure on the system curve is made up of three components - the static head, the friction head and the velocity head.

In the case where there is no pipework connected to the discharge flange there is no friction head. The static head is the small difference between the pump centre line and the discharge flange. So virtually all the TDH reports as velocity head. If you could put a pressure gauge on a tapping through the discharge flange it would read virtually zero. The centre tapping on a pitot tube in the same position would of course read the velocity head.

To summarise this long winded story into a proper reply to your question - yes, the discharge pressure would be atmospheric but the developed head would still be there as velocity head (kinetic energy).

Katmar Software - Engineering & Risk Analysis Software

http://katmarsoftware.com

"An undefined problem has an infinite number of solutions"

 

[GaTechTheron]

Thank you Katmar for pointing this out:

1. Centrifugal pumps impart energy into a system and build HEAD (not pressure) according to head vs flow curve.

2. Positive displacement pumps move a specified volume and impart energy into a system as required to move the specified volume.

In both cases, the pump will operate where the pump curve intersects the system curve. Pictures speak a 1000 words. Please see the attached, crude, drawing.

 

[hydtools]

Look at the system curves in GaTechTheron's attachment. No flow, no pressure. You need a pump to create flow to even talk about pressure. Pumps create flow. Pumps have pressure capacity, the property to deliver flow against the system pressure demand.

Ted

 

[SNORGY]

Perhaps a pump creates nothing. It is simply a conduit for energy transfer. That energy then takes the form of potential energy, kinetic energy, or some combination of the two.

I like taking the "farmer's approach" to this...

I recently installed a sump pump to de-pond my dog run. When I hooked it up to pipe and turned the pump on, water came out the end of the pipe. From this, I concluded that the pump appeared to be creating flow. But, I had forgotten to tighten up one of the rubber joint clamps, so the pipe broke apart. When I tried to put the pipe back together and tighten it up without turning the pump off, I found that this required quite a bit of force on my part. From this, I concluded that the pump appeared to be creating pressure.

When I unplugged the pump and cut off its energy supply, it created nothing.

I thereby deduce that at least my sump pump creates both flow and pressure, provided it gets energy from somewhere.

Regards,

SNORGY.

 

[sheiko]

Thank you all.

Below are my conclusions:

* Both pump types put energy into the system in order to realize the transfer required. Prerequisite for both type is power supplied to liquid by pump drive to overcome friction, elevation, velocity and pressure heads.

* Positive-displacement pump create primarily flow and discharge pressure is created by system's response to flow (ie: discharge pressure = discharge line total backpressure at pump outlet flange).

* Centrifugal pumps primarily develop head, not flow (It is indeed true that when you deadhead a centrifugal pump, it continues to rotate, and the pressure gauge displays the equivalent head at shutoff condition).
Head results in flow, as the consequence of pressure difference between pump discharge and destination point.

* In both cases, the pump will operate where the pump curve intersects the system curve.


"We don't believe things because they are true, things are true because we believe them."

 

[Artisi]

Does the academic discussion really matter or make the pump/s perform any differently than what they have designed to do? - not really - so long as the pump does what it has been installed to do without any problems.

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.)

 

[sheiko]

Artisi, I appreciate your sense of humor but this is not an academic discussion. This question arose in a professionnal context.
Or maybe it is a way to express your disagreement with the conclusions?
Regards.

"We don't believe things because they are true, things are true because we believe them."

 

[SNORGY]

sheiko...

So, within mechanical limitations of the pump and assuming 100% volumetric efficiency, PD pumps produce flow, with pressure as a consequence, since there is nowhere else for the fluid to go except into the system.

Centrifugal pumps produce head, and the potential energy of the head produced converts into the sum of energy losses (potential plus kinetic) downstream of the pump, and flow happense when the static head loss (potential energy) component downstream of the pump is less than that which the pump can produce at shut-off.

Makes sense to me.

Regards,

SNORGY.