Centrifugal pump question

On Sunday, May 28, 2017 at 9:54:50 AM UTC-4, SbzWr?? ? ?????? ? ?? ?????? ??AJFeU wr ote:

umps threw me. Since water isn't compressible, I don't see how the multi-st age pumps work. For gas, no problem, but I don't get it for liquids.

e/volume relationship, though, isn't in agreement with Boyle's law. Gases a pproximate it. It's easy to imagine a multi-stage non-positive-displacement compressor that keeps building pressure in a material that obeys Boyle's l aw. It's much harder to imagine it with liquids.

n't hold pressure that way.

e kinetic aspects of a turbo pump (velocity) and the potential aspects (pre ssure). A turbine pump that's pumping a liquid must be producing potential energy from kinetic energy.

ay that an ordinary turbine pump can hold the pressure generated by a previ ous stage, unless the entire thing is kinetic, which we're then measuring a s potential energy (pressure).

They are very lossy machines. The part that's of interest here is compresso r efficiency, which ranges from 0.70 to 0.85 in the best turbo machinery, i ncluding stationary and aircraft gas turbines. At the high end, 85% compres sor efficiency, they're losing 15% to gas friction.

But again, you're talking about a gas turbine. Gas has a close ratio betwee n pressure and volume (Boyle's law). Compressing gas with a machine, whethe r it's positive or non-positive displacement, like a turbo compressor, is n ot a problem. An example of positive displacement types is a vane-type supe rcharger. An example of non-positive-displacement types is a turbocharger. Or the compressor stage of a gas turbine engine, such as an aircraft jet en gine.

Liquids don't behave according to Boyle's law.

How do you know that a turbo fan engine can compress water?

Obviously, it *does* work. The question is "how."

snipped-for-privacy@gmail.com wrote on 5/28/2017 10:27 AM:

pumps threw me. Since water isn't compressible, I don't see how the mult i-stage pumps work. For gas, no problem, but I don't get it for liquids.

ure/volume relationship, though, isn't in agreement with Boyle's law. Gas es approximate it. It's easy to imagine a multi-stage non-positive-displa cement compressor that keeps building pressure in a material that obeys B oyle's law. It's much harder to imagine it with liquids.

ldn't hold pressure that way.

the kinetic aspects of a turbo pump (velocity) and the potential aspects (pressure). A turbine pump that's pumping a liquid must be producing pote ntial energy from kinetic energy.

way that an ordinary turbine pump can hold the pressure generated by a p revious stage, unless the entire thing is kinetic, which we're then measu ring as potential energy (pressure).

essor efficiency, which ranges from 0.70 to 0.85 in the best turbo machin ery, including stationary and aircraft gas turbines. At the high end, 85% compressor efficiency, they're losing 15% to gas friction.

tween pressure and volume (Boyle's law). Compressing gas with a machine, whether it's positive or non-positive displacement, like a turbo compress or, is not a problem. An example of positive displacement types is a vane

-type supercharger. An example of non-positive-displacement types is a tu rbocharger. Or the compressor stage of a gas turbine engine, such as an a ircraft jet engine.

If you daisy-chain two centrifugal pumps together, I am sure you will get a lot higher pressure output than using just one (too optimistic to expect 2 times the pressure).

threw me. Since water isn't compressible, I don't see how the multi-stage p umps work. For gas, no problem, but I don't get it for liquids.

This is where I have trouble. Assuming these are regular centrifugal turbin es, the outlet of the first stage is fed into the axis of the second stage. The pressure from the first-stage outlet is retained at the second-stage i nlet, but from there it feeds into the whirling blades of the second stage, the outlet volume of which is LARGER than the inlet volume between any two blades.

Pressure, thus, is converted to velocity. Unless the machine *compounds* th e velocity at each stage, I don't see how it works. And, in order to compou nd velocity by a factor of, say, three, either the shaft driving the stage either has to be turning at (square root of 3) times that of the first stag e, or the the second stage has to have a completely different scroll design .

But you can carry that only so far. Go to three stages, or four, and the sh aft rotational speeds become outrageous, or the scroll design does.

Obviously, I'm missing something here, but I haven't yet seen what it is.

Ed Huntress

This is where I have trouble. Assuming these are regular centrifugal turbines, the outlet of the first stage is fed into the axis of the second stage. The pressure from the first-stage outlet is retained at the second-stage inlet, but from there it feeds into the whirling blades of the second stage, the outlet volume of which is LARGER than the inlet volume between any two blades.

Pressure, thus, is converted to velocity. Unless the machine

*compounds* the velocity at each stage, I don't see how it works. And, in order to compound velocity by a factor of, say, three, either the shaft driving the stage either has to be turning at (square root of 3) times that of the first stage, or the the second stage has to have a completely different scroll design.

But you can carry that only so far. Go to three stages, or four, and the shaft rotational speeds become outrageous, or the scroll design does.

Obviously, I'm missing something here, but I haven't yet seen what it is.

Ed Huntress

===========================

The pressure of a 10 meter head of water is close to one atmosphere. Section 3.2 describes pumps in series. Fig 3.2.5 and 3.2.6 show how the pressures add when the pumps are of equal or mismatched sizes.

-jsw

A centrifuge pump creates pressure by spinning fluid away from the center (hence "centrifuge").

You are confusing yourself by of processing too much information at one time. To make it easy for you, let's consider the output valve is closed (the system is creating pressure but not expelling anything).

The first pump supplies the pressurized fluid to the second pump, the second pump spins the pressurized fluid away from the center to add more pressure to the housing wall. The pressure gauge in the second stage should register more pressure than the first stage.

Does this help you understand better now, Ed?

ps threw me. Since water isn't compressible, I don't see how the multi-stag e pumps work. For gas, no problem, but I don't get it for liquids.

rbines, the outlet of the first stage is fed into the axis of the second st age. The pressure from the first-stage outlet is retained at the second-sta ge inlet, but from there it feeds into the whirling blades of the second st age, the outlet volume of which is LARGER than the inlet volume between any two blades.

• the velocity at each stage, I don't see how it works. And, in order to co mpound velocity by a factor of, say, three, either the shaft driving the st age either has to be turning at (square root of 3) times that of the first stage, or the the second stage has to have a completely different scroll de sign.

e shaft rotational speeds become outrageous, or the scroll design does.

That's well said, but how is the fluid pressurized once it's fed into the s econd stage?

Take a look at the pump impellers in these photos:

(or, Tiny URL):

Remember that the inlet is at the center of the impeller. The liquid then m akes a 90-deg. turn and enters the involutes. As the liquid travels from th e center to the periphery, the volume *increases*. Poof! There goes your pr essure out the window.

But I think I have it figured out now. The problem starts with the concept of "pressure." Except when you're dealing with gases, that's always problem atic. The physical measures that are involved here actually are mass, veloc ity, and force. Forget about pressure for a moment. Think energy instead.

If the previous stage can supply enough liquid to fill the subsequent invol ute more than it would be filled without that previous stage, then that sub sequent stage increases the energy of the water by increasing its velocity. "Pressure" is irrelevant. The energy going in is the product of mass and f orce (forget the actual formula for now). The energy coming out is the same thing, but along the way, an increase in velocity has increased the force. Restrict that mass and force at the pump exit, and you get pressure.

So that's what I think is happening.

ines, the outlet of the first stage is fed into the axis of the second stag e. The pressure from the first-stage outlet is retained at the second-stage inlet, but from there it feeds into the whirling blades of the second stag e, the outlet volume of which is LARGER than the inlet volume between any t wo blades.

You are over thinking the situation. A multistage pump has a bunch of iden tical sections all turning at the same speed. Each stage increases the pre ssure. So you might have a 6 stage pump with each stage increasing the pre ssure by 10 psi. Which makes for a fairly efficient pump which will supply water at 60 psi. Google it.

In Eric's case it is a little different. He wants to circulate water. Th e system is pressurized to 80 psi and the pump has 80 psi on the input. Th e output is at 90 psi into a heating loop. and the friction of the water f low reduces the pressure so that at one end of the loop you have 90 psi an d at the other end you have 80 psi. The pumps are not in series.

Dan

That's well said, but how is the fluid pressurized once it's fed into the second stage?

Take a look at the pump impellers in these photos:

(or, Tiny URL):

Remember that the inlet is at the center of the impeller. The liquid then makes a 90-deg. turn and enters the involutes. As the liquid travels from the center to the periphery, the volume *increases*. Poof! There goes your pressure out the window.

But I think I have it figured out now. The problem starts with the concept of "pressure." Except when you're dealing with gases, that's always problematic. The physical measures that are involved here actually are mass, velocity, and force. Forget about pressure for a moment. Think energy instead.

If the previous stage can supply enough liquid to fill the subsequent involute more than it would be filled without that previous stage, then that subsequent stage increases the energy of the water by increasing its velocity. "Pressure" is irrelevant. The energy going in is the product of mass and force (forget the actual formula for now). The energy coming out is the same thing, but along the way, an increase in velocity has increased the force. Restrict that mass and force at the pump exit, and you get pressure.

So that's what I think is happening.

Right. I have no problem with steam injectors or with centrifugal pumps. The problem I was having is with compounded centrifugal fluid pumps.

After brushing up on impulse and momentum, and remembering that "pressure" is a frequently misleading term, I'm pretty sure I see what's happening.

snipped-for-privacy@gmail.com wrote on 5/28/2017 2:50 PM:

umps threw me. Since water isn't compressible, I don't see how the multi- stage pumps work. For gas, no problem, but I don't get it for liquids.

turbines, the outlet of the first stage is fed into the axis of the secon d stage. The pressure from the first-stage outlet is retained at the seco nd-stage inlet, but from there it feeds into the whirling blades of the s econd stage, the outlet volume of which is LARGER than the inlet volume b etween any two blades.

ds* the velocity at each stage, I don't see how it works. And, in order t o compound velocity by a factor of, say, three, either the shaft driving the stage either has to be turning at (square root of 3) times that of th e first stage, or the the second stage has to have a completely different scroll design.

the shaft rotational speeds become outrageous, or the scroll design does.

he second stage?

974

en makes a 90-deg. turn and enters the involutes. As the liquid travels f rom the center to the periphery, the volume *increases*. Poof! There goes your pressure out the window.

ept of "pressure." Except when you're dealing with gases, that's always p roblematic. The physical measures that are involved here actually are mas s, velocity, and force. Forget about pressure for a moment. Think energy instead.

nvolute more than it would be filled without that previous stage, then th at subsequent stage increases the energy of the water by increasing its v elocity. "Pressure" is irrelevant. The energy going in is the product of mass and force (forget the actual formula for now). The energy coming out is the same thing, but along the way, an increase in velocity has increa sed the force. Restrict that mass and force at the pump exit, and you get pressure.

No, that's no what's happening.

You are confusing yourself by making things complicated. You don't need force/energy equations to figure this out.

A centrifugal pump is a "centrifuge". It is spinning the fluid (gas/liquid) outward. The fluid is continuously being thrown against the

peripheral wall of the housing which results in fluid pressure against the retaining wall. It is just like you use your hands to push against the retaining wall to create pressure on the retaining wall, except that

the fluid is being spun around continuously to create the radially outward pressure all around in 360 degrees against the retaining wall that is preventing it from flying out.

Of course, the center of the centrifuge will be a partial vacuum. That is where the input valve is usually placed so it can suck in fluid if the pump is properly primed.

On Sunday, May 28, 2017 at 6:14:05 PM UTC-4, izlsb?? ? ?????? ? ?? ?????? ??TSwdF wr ote:

e:

umps threw me. Since water isn't compressible, I don't see how the multi-st age pumps work. For gas, no problem, but I don't get it for liquids.

turbines, the outlet of the first stage is fed into the axis of the second stage. The pressure from the first-stage outlet is retained at the second-s tage inlet, but from there it feeds into the whirling blades of the second stage, the outlet volume of which is LARGER than the inlet volume between a ny two blades.

ds* the velocity at each stage, I don't see how it works. And, in order to compound velocity by a factor of, say, three, either the shaft driving the stage either has to be turning at (square root of 3) times that of the firs t stage, or the the second stage has to have a completely different scroll design.

the shaft rotational speeds become outrageous, or the scroll design does.

he second stage?

74

en makes a 90-deg. turn and enters the involutes. As the liquid travels fro m the center to the periphery, the volume *increases*. Poof! There goes you r pressure out the window.

ept of "pressure." Except when you're dealing with gases, that's always pro blematic. The physical measures that are involved here actually are mass, v elocity, and force. Forget about pressure for a moment. Think energy instea d.

nvolute more than it would be filled without that previous stage, then that subsequent stage increases the energy of the water by increasing its veloc ity. "Pressure" is irrelevant. The energy going in is the product of mass a nd force (forget the actual formula for now). The energy coming out is the same thing, but along the way, an increase in velocity has increased the fo rce. Restrict that mass and force at the pump exit, and you get pressure.

That doesn't explain how it adds to the pressure at the inlet.

No, it isn't. In fact, in this case, it's 80 psi.

I suggest you watch this centrifugal pump video to clear up your brain fog:

snipped-for-privacy@gmail.com wrote on 5/28/2017 7:04 PM:

ote:

pumps threw me. Since water isn't compressible, I don't see how the mult i-stage pumps work. For gas, no problem, but I don't get it for liquids.

l turbines, the outlet of the first stage is fed into the axis of the sec ond stage. The pressure from the first-stage outlet is retained at the se cond-stage inlet, but from there it feeds into the whirling blades of the second stage, the outlet volume of which is LARGER than the inlet volume between any two blades.

unds* the velocity at each stage, I don't see how it works. And, in order to compound velocity by a factor of, say, three, either the shaft drivin g the stage either has to be turning at (square root of 3) times that of the first stage, or the the second stage has to have a completely differe nt scroll design.

d the shaft rotational speeds become outrageous, or the scroll design doe s.

the second stage?

=974

then makes a 90-deg. turn and enters the involutes. As the liquid travels from the center to the periphery, the volume *increases*. Poof! There go es your pressure out the window.

ncept of "pressure." Except when you're dealing with gases, that's always problematic. The physical measures that are involved here actually are m ass, velocity, and force. Forget about pressure for a moment. Think energ y instead.

involute more than it would be filled without that previous stage, then that subsequent stage increases the energy of the water by increasing its velocity. "Pressure" is irrelevant. The energy going in is the product o f mass and force (forget the actual formula for now). The energy coming o ut is the same thing, but along the way, an increase in velocity has incr eased the force. Restrict that mass and force at the pump exit, and you g et pressure.

A centrifugal pump creates a partial vacuum at the inlet (usually at the

center for a centrifugal pump)

The partial vacuum is filled in by the incoming fluid because the outlet

is shut off. Once the tap at the outlet is turned on, the partial vacuum

will suck in additional fluid to fill the void.

In both cases (either the tap at the outlet is turned on or turned off),

a lot more than 80 psi will be at the casing radially outward farthest from the center.

Just watch this video, then you will understand:

On Sunday, May 28, 2017 at 7:05:42 PM UTC-4, ZScPb?? ? ?????? ? ?? ?????? ??tLolr wr ote:

pumps threw me. Since water isn't compressible, I don't see how the multi- stage pumps work. For gas, no problem, but I don't get it for liquids.

ure/volume relationship, though, isn't in agreement with Boyle's law. Gases approximate it. It's easy to imagine a multi-stage non-positive-displaceme nt compressor that keeps building pressure in a material that obeys Boyle's law. It's much harder to imagine it with liquids.

ldn't hold pressure that way.

the kinetic aspects of a turbo pump (velocity) and the potential aspects (p ressure). A turbine pump that's pumping a liquid must be producing potentia l energy from kinetic energy.

way that an ordinary turbine pump can hold the pressure generated by a pre vious stage, unless the entire thing is kinetic, which we're then measuring as potential energy (pressure).

essor efficiency, which ranges from 0.70 to 0.85 in the best turbo machiner y, including stationary and aircraft gas turbines. At the high end, 85% com pressor efficiency, they're losing 15% to gas friction.

tween pressure and volume (Boyle's law). Compressing gas with a machine, wh ether it's positive or non-positive displacement, like a turbo compressor, is not a problem. An example of positive displacement types is a vane-type supercharger. An example of non-positive-displacement types is a turbocharg er. Or the compressor stage of a gas turbine engine, such as an aircraft je t engine.

I know how a centrifugal pump works. That video doesn't address the issue i n question: What happens when the input pressure is higher than the example in your video? And how does it work?

Notice that you did not address the issue of the involute volume increasing as the liquid flows from the center to the periphery, and the effect that has on pressure.

On Sunday, May 28, 2017 at 7:22:40 PM UTC-4, Pdoih?? ? ?????? ? ?? ?????? ??XSEKB wr ote:

ote:

pumps threw me. Since water isn't compressible, I don't see how the multi- stage pumps work. For gas, no problem, but I don't get it for liquids.

l turbines, the outlet of the first stage is fed into the axis of the secon d stage. The pressure from the first-stage outlet is retained at the second

-stage inlet, but from there it feeds into the whirling blades of the secon d stage, the outlet volume of which is LARGER than the inlet volume between any two blades.

unds* the velocity at each stage, I don't see how it works. And, in order t o compound velocity by a factor of, say, three, either the shaft driving th e stage either has to be turning at (square root of 3) times that of the fi rst stage, or the the second stage has to have a completely different scrol l design.

d the shaft rotational speeds become outrageous, or the scroll design does.

the second stage?

=974

then makes a 90-deg. turn and enters the involutes. As the liquid travels f rom the center to the periphery, the volume *increases*. Poof! There goes y our pressure out the window.

ncept of "pressure." Except when you're dealing with gases, that's always p roblematic. The physical measures that are involved here actually are mass, velocity, and force. Forget about pressure for a moment. Think energy inst ead.

involute more than it would be filled without that previous stage, then th at subsequent stage increases the energy of the water by increasing its vel ocity. "Pressure" is irrelevant. The energy going in is the product of mass and force (forget the actual formula for now). The energy coming out is th e same thing, but along the way, an increase in velocity has increased the force. Restrict that mass and force at the pump exit, and you get pressure.

That's the same video you posted the link to in another message. It still d oesn't deal with compound or pressure-inlet conditions.

There is no "partial vacuum" if the inlet is at 80 psi of positive pressure .

entical sections all turning at the same speed. Each stage increases the p ressure. So you might have a 6 stage pump with each stage increasing the p ressure by 10 psi. Which makes for a fairly efficient pump which will supp ly water at 60 psi. Google it.

From Wik

Multistage centrifugal pumps Multistage centrifugal pump[5]

A centrifugal pump containing two or more impellers is called a multistage centrifugal pump. The impellers may be mounted on the same shaft or on diff erent shafts. At each stage, the fluid is directed to the center before mak ing its way to the discharge on the outer diameter.

For higher pressures at the outlet, impellers can be connected in series. F or higher flow output, impellers can be connected parallel.

A common application of the multistage centrifugal pump is the boiler feedw ater pump. For example, a 350 MW unit would require two feedpumps in parall el. Each feedpump is a multistage centrifugal pump producing 150 l/s at 21 MPa.

All energy transferred to the fluid is derived from the mechanical energy d riving the impeller. This can be measured at isentropic compression, result ing in a slight temperature increase (in addition to the pressure increase) .

Dan

On May 28, 2017, snipped-for-privacy@gmail.com wrote (in article):

I think that the missing piece is Bernoulli?s Equation:

. .

For water (and air at low velocity compared to the speed of sound), read the stuff about incompressible flow.

Joe Gwinn

identical sections all turning at the same speed. Each stage increases the pressure. So you might have a 6 stage pump with each stage increasing the pressure by 10 psi. Which makes for a fairly efficient pump which will su pply water at 60 psi. Google it.

e centrifugal pump. The impellers may be mounted on the same shaft or on di fferent shafts. At each stage, the fluid is directed to the center before m aking its way to the discharge on the outer diameter.

For higher flow output, impellers can be connected parallel.

dwater pump. For example, a 350 MW unit would require two feedpumps in para llel. Each feedpump is a multistage centrifugal pump producing 150 l/s at 2

1 MPa.

driving the impeller. This can be measured at isentropic compression, resu lting in a slight temperature increase (in addition to the pressure increas e).

Thanks, Dan. I read that -- and maybe 100 more pages over the past few days . None of them really explain it. To say that the energy is derived from th e impeller is axiomatic. It doesn't explain what's going on inside the seco nd stage.

Rather than try to go through it in detail, I'll post something if I find a good explanation.

Thanks, Joe. That is a good way to deal with the conversion and conservatio n of energy. If I can find out what the dynamics are inside of that second stage, it may help.

Without going into details, here's the basic dilemma. Note that the volume of the involutes increases as you progress from the center to the periphery . Illustrations usually show that volume filled at the center, but only par tly filled at the periphery. I don't know if the illustrations are correct or not. If they are, then there is no pressure involved inside of the invol ute -- only velocity and mass.

If they *are* correct, then the velocity must *decrease* as you progress fr om center to periphery, to conserve energy with the larger mass involved. T hat's the static view. It's possible that a dynamic view allows for both an increase in volume and an increase in velocity, due to the energy added by the rotation of the wheel.

I don't think that's what happens. I think it's a case of velocity increasi ng. If that's the case, the energy is imparted by the second stage by the v elocity imparted by radial acceleration -- which is what we're often told i s the way a centrifugal turbomachine works.

Now, if that's true, then what is the effect of feeding the second stage wi th water at high pressure? What happens with that pressure inside of the in volute? It can't be conserved because, if the involute isn't filled, it's u nconstrained and it simply fills up the empty volume near the periphery. En ergy is conserved because the mass*velocity is conserved: greater mass, les s velocity.

Is that what happens? I've found no explanation or illustration of it so fa r.