Why Impeller Size Same in Multistage Pump?

Here is a good explanation of single and two stage pumps.

I was going to sketch and go into details, but this link will suffice

heres the direct link, if the above link is not working:

www.lacountyfirefighters.org/items/HALE_2stage_Singstg.pdf

Because that's the way it works. All impellers must be the same size.

What do you think would happen if the impellers were different sizes. What does impeller diameter determine?

There are many reasons I would give only 2:

- transverse dimensions: for a given flow it is a slimmer design if several rotors are in series if a high head is required.

- economical: a higher number of identical components decreases cost per component.

remark: I tried the 2 links but items have been removed!

There is no reason why the impeller of one or more stages in a multi stage pump cannot be skimmed. No impeller is smart enough to know what the properties of his neighbour is, it can only respond to the pressure difference between its inlet and outlet.

The flow through the stages are the same but the pressure contribution of the stages can differ.

Its simular to pumps in series is it not.

Yes it does but one can use a submersible in the water and use a centrifugal booster on the banks in series. No need to be the same.

Different diameter impellers in a multistage pump is not very practical but it does work.

different size impellars, I never thought of that, but then again, if your buying knew, one never would think of that.

Cavitiation may be a reason.

There is definitely a reason to sizing the impellers the same. The first, as someone alluded to earlier, is economics. All impellers being the same size makes duplication and economy of repetition a factor. However, there are other problems that can occur when the impellers are not the same. You see, when a centrifugal impeller is operating, it has a certain "characteristic curve" by which it is governed. This will tell what the delivery (flow) of the impeller is going to be and at what head (pressure). When an impeller is operating, it will deliver the flow and head as defined by that curve. The inlet pressure is of great importance in the design of the impeller and systrem. Whatever pressure is being delivered TO ihe impeller will be added to the pressure the pump will deliver. That is why the multi-stage pump is useful. In order to increase the pressure of the pump, the identical impellers will be operated in series, each adding its pressure to the pump delivery. If one impeller will deliver 20 GPM @ 25' head, the pump will delliver 20 GPM @ 75' of head if there are three impellers in series. The flow will remain the same from impeller to impeller, but the head will increase. And therein lies the problem. If one impeller is smaller in diameter than the one following it, it will follow its characteristic curve which will dictate that it delivers the same flow, but at a lower head than the next. This can cause an unbalanced condition to exist which can cause bearing and seal damage. Further, the predicted flow and pressure will change to an un-predictable number. Another consideration is the physical size of the impeller. If a larger impeller is installed, it will require that the volute is enlarged to accommodate this impeller. Again, this is a econemic consideration. I have a question for you: Why do you want to have a differnt sized impeller?

It is not that I want it that way but it will and does work. If one of the impellers in your example is skimmed it may only contribute 20 m to the head and the total head will decrease to 70m. (it may even find a new duty point) if that is acceptable it will and does work without ill effect to bearings etc.

If all impellers are connected to the same shaft and impellers are of different sizes mechanical imbalance(deflection?) will take place in the shaft and outer diameter of housing won't be uniform.Compare with an IC engine with 4 pistons firing at different times connected to a common crankshaft fitted with a flywheel.

I believe that Hendrik is quite correct. There is no reason, in theory, why you couldn't have different sized impellers.

I disagree with The Commoner 's statement: "This can cause an unbalanced condition to exist which can cause bearing and seal damage." As long as the duty point for each stage is at a suitable point on the characteristic curve for that particular stage, then there should be no problems. The shaft and bearings are designed to handle the required radial and axial loads (in many cases dual-volutes at 180° apart are used to balance radial loads anyway). It doesn't matter if one stage adds 20m to the total head and the next stage 15m or 25m.

The only two reasons why same-size impellers are used are that it is never necessary to do otherwise, and to do otherwise is more complicated and costly and therefore makes no sense.

I think the main reason is the flow. Generally multistage centrifugal pumps are in use for pumping incompressable fluides like water, the volume of the fluide remains same for all stages. So, to accomodate same flow through all stages same size of impeller to be used.

No, not really. For example, if you look at ISO end-suction single-stage centrifs, the 65-40-160, 65-40-200, 65-40-250 and 65-40-315 can all comfortably do 60 USgpm (at 1450rpm), but at different differential pressures. You could put the four of them in a system in series and you could get an overall flowrate of 60 USgpm, but with the pressure building at each stage(eg. +10m +18m +25m +40m for a TDH of 93m). Hydraulically this would be the same as a multistage pump with different size impellers.

The question was as far as my limited English allows me to understand why multi-stage pumps have identical stages. It was neither asked if other solutions are possible nor how to do it with different rotors.

I do not understand why answers do not consider the original question ? Are we so frustrated they we need to show how much we know aside the expected answer ?

I notice that the GREAT majority of answers does not respond to the question but derive.

This is good and bad, it is good because the OP will learn more but it is bad since he will not know if he has the right answer to his question.

I am sure my comment will not be well acepted because it is against the main stream but I had to make it.

I agree with you. I'll never be convinced that different diameter impellers are OK in a multi-stage pump.

As I said earlier, I still maintain that cavitation is a real possibility if this is done.

This forum hardly ever answers a question directly. Mostly oneupmanship is the root cause here. I can only find one or two posts that are answered by a single, correct answer, and not "improved upon" by fifteen other comments that may, or may not be relevant.

Usually we beat a subject completely to death, then let if lie in the dirt, unresolved.

I could be entirely wrong, but I will never be convinced that different sized impellers won't work until I see test data proving otherwise, but there is no empirical test data available, and it is unlikely anyone will ever have the requirement (or time and money to waste) to develop a multistage pump with different impellers, so we may never know for sure.

Again, I could be entirely wrong, but I don't see why cavitation should be an issue. The first stage would be most at risk (subsequent stages being pressurized by their preceding stages), and if suction conditions are poor enough to cause excessive cavitation, then it doesn't matter if the impeller sizes are the same or different, a normal multistage will also suffer cavitation. Different impeller sizes will not change suction conditions at the pump inlet.

You are right that "This forum hardly ever answers a question directly.", but wouldn't it be rather dull sometimes if they did?

I don't think you'll ever see test data., just due to the fact, it never really came up......... until now.

I want to mention that I never said it would not work but as everybody knows (I presume at least) engineering is related to economics.

You never saw a multistage pump with different rotors for several reasons. Let us look at possible reasons.

Why make a multistage pump when it is possible to obtain the full pressure in one stage provided that dimensions are such that the pressure is obtained at the usual RPM ?

Rotor dimensions would be bigger since for a given rpm pressure is related to D². But at same time the stresses in the rotor progress in same proportion which means that the material to cast or to cut the rotor has to have a higher yield stress and be thus more expensive. Dynamic disbalances which have to be corrected are easier to correct when masses are near to axis. Since flow has to be the same axial width of rotor will be less making under circumstances its casting difficult (small flow +high pressure).

The number of types would be higher and manufacturing much more expensive.

So there are -at least on my humble engineering opinion - several economic reasons to avoid when not needed the single stage pump for higher pressures and limited transverse dimensions.

Now the second question is where are multi-stage pumps used ?

and the third: Why do you believe are they the choice ?

When you found the answer you will understand why you do not see multi-stage pumps with different rotors although theoretically it is feasible.

As for cavitation it is always a problem and in multi-stage pumps it is not limited at entry of the stack but can appear between every stage since there are flow redirecting channels where in case of dirt deposit or other reasons locally velocities can be so high that in use for fluids with high vapor pressure (as water) cavitation can occur.

I maintain that correct answer have not to be dull and it can be all information presented in such a way that it is for the OP a challenge to go further and deeper in his research of information for a higher personal qualification. I further think as I wrote -against the main stream- that most of participants want to show how much they know and are in fact not interested in the question itself. Many times the "answers" were totally wrong and would lead to failures for the OP if he would follow them.

I agree that is also a chat site but our main concern has to be the poor guy who does not know what to do because very often he is obliged to solve a problem he has not the slightest knowledge about.

When we did it we can have also time to chat and make jokes more or less private.

In replying to my reply to Lyn, you are in danger of confusing things.

The OP posted two questions. The first is a "Why", and the answer is extremely simple. The second is a "What if…", and the answer is simply that nobody really knows. I ignored the second question, because it is all conjecture, and I answered the first question in my first post:

1. It is not necessary to have different impellers. There is nothing to gain.
2. It is more costly and complicated to have different size impellers.

It's as simple as that.

I stated this in my first post, and would have left it at this, but it is necessary to make some form of comment when people mention reasons for same-size impellers that are either debateable or downright wrong. There are two things that I have felt the need to comment on:

1. The assumption, implication or statement that the reason for same-size impellers is that different size impellers won't work, and thus the reason for same-size impellers is technical. This is a debateable point. Some people agree with it, others don't, but it is totally unproven either way. More importantly, it is not the reason why same-size impellers are used. It has probably never even been a consideration by pump designers because it was already so obvious (from reasons 1 & 2 above) that the impellers should be the same size.
2. Somebody stated that the reason for same-size impellers was down to flowrate, to get the same flowrate in each stage (and this somehow earned a GA). Firstly, this is technically wrong, and I had hoped my example of single-stage pumps in series would be enough to demonstrate that this is clearly wrong, and that different size impellers in a system can get the same flowrate. Secondly, this is also clearly not a reason for using same-size impellers, and again has probably never been a consideration by pump designers because it was already so obvious (from reasons 1 & 2 above) that the impellers should be the same size.

And now that the dead horse has been well and truly flogged, I will go off to ponder other more important things (like why did my toast fall peanut-butter-side-up yesterday? Surely this defies the laws of physics and of the universe? Does it have anything to do with global warming, I wonder?)

I can give you an explanation about the toast. As unbelievable as it seems a lot of tests have been done (not by me but I saw the filmed tests) and it came out clearly that it is related to the falling height. It is due to the combination between falling speed and rotation. For the normal table height and a normal slice of bread it is under repeatable initial conditions (this is a must) the repeatability is almost 100%.

Guilty.

If you read previous posts you will see that the last paragraph of my first post (#13) directly answers the OPs question.

This post (#15) is replying directly to the previous post (#14) and is simply disagreeing with its assertion that identical impellers are required to ensure the same flowrate in each stage, and that this is why multistage pumps have identical impellers.

As I see it, the whole post is effectively answering the OPs question, and the mention of ISO centrifs in series is an example to back up my assertion, not a proposed alternative solution. It should have been clear from my previous post that there is no alternative solution because there is no problem.

I hope this is clear.

Will there be any mechanical problems in shaft,seal etc?.

No. Why should there be? In this example, if you look at each pump individually, it is operating at an ideal point on its performance curve, about 60%-65% of Best Efficiency Point, which is probably the best point for a combination of good efficiency, low NPSHr, low radial shaft loading, stable curve, etc. Suction conditions are good. Why should there be any mechanical problems? (By the way, for reference, I am using the Goulds GIS curves here )

If one of the impellers is bigger or smaller than the rest its weight,inertia and centrifugal /centripetal force will differ from others thereby creating shaft deflection/wobbling which can damage bearing/seal.

The radial and axial forces on the shaft are not the result of the "weight,inertia and centrifugal /centripetal force" of the impeller (impellers are usually balanced), they are the result of the forces imparted on the pumped fluid.

The axial forces are mostly the result of pump suction, are usually minimal, and usually easily handled by using suitable bearings. This will not be affected by different size impellers.

The radial forces are mostly the result higher pressure at the discharge of the volute (unlike a concentric-cased pump, a volute pump sees a gradual increase in pressure from the cutwater around to the discharge). This is usually handled by ensuring the shaft and bearings are suitably sized. If the pressures are such that a suitably sized shaft would be impractical, then you use a dual (or triple) volute to balance the pressures and reduce radial forces. It doesn't matter if impellers are different sizes, if the shaft and bearings are correctly sized (and impellers balanced) there will be no shaft deflection/wobbling.

The biggest problems a pump is usually on the intake. with a 80/20 rule.

The biggest concern would be have a balance fluid (i.e., a start pipe length and head pressure) on the intake

That is true. But suction conditions have to be good whether the pump is a normal multistage pump, or our mythical asymetrical multistage pump. Different impeller diameters will not make good suction conditions bad.

That Idon't disagree with, I should have stated I was just making a point about the inlet side. and not the impellar config.