Centrifugal pump design and manufacturing has been around for over 200 years. If there were an ideal impeller vane number, it seems likely someone would have discovered it by now. However, just because an ideal vane number has not yet been established is not sufficient reason to consider it doesn’t exist.
Single stage pump and first stage, high suction performance impellers tend to have fewer vanes – often 3, 4 or 5. Multistage pump series stages, pipeline pumps and large, high efficiency pumps tend to have 7 or more impeller vanes. Vane number might be dictated by special application requirements. Such is the case with solids handling impellers which often have just 2 or 3 vanes and sometimes only 1.
I recall an anecdote I heard decades ago from a senior hydraulic engineer that only 3-vane impellers are found across the entire range of specific speeds. The problem, especially for high speed, high head pumps, is a 3-vane impeller produces significantly less head than, say, a 7-vane impeller. Head rise to shutoff and pressure pulsations are also issues.
But thinking about this quasi-factual anecdote, I’ve occasionally considered that an ideal vane number does indeed exist. If there is a magic vane number, what is it? If there could be only one impeller vane number for all pumps, what would it be? Pump engineers think these kinds of thoughts.
Dear Randal,
I find interesting your note. I have a question. If a pump is designed with a 6 vanes impeller, and for availability reasons during maintenance the pump is assembled with a 5 vane impeller, what is to expect in performance (flow, head, energy consumption, torque…)?
Thank you in advance for your comments.
Lizardo
Lizardo,
The question can be answered generally, but may not necessarily be correct for your specific pump. If only the vane number is reduced from 6 to 5, whether or not the vane design is changed, a reduction in pump head vs. flow can be expected, as well as reduction in torque and power. I might guess a drop of 10% – 15% head at a given rate of flow, but I wouldn’t wager any money on it. NPSHR is likely to improve, but this would be of no consequence if the existing 6-vane impeller already has adequate NPSH margin. The overall effect of this vane number change will also depend on the system characteristics. I hope this helps.
Randal
why no. of vane are kept in odd no of centrifugal pump in most of the case?
Avijit,
This is an excellent question!
A rotating impeller produces a circumferentially varying pressure field, between vanes, at its exit. The periodic pressure fluctuations interact with the stator, usually a volute or a diffuser. This rotor/stator interaction (RSI) can combine and amplify pressure pulsations and pump or piping vibrations.
When an even vane number impeller is paired with a double volute, vane pass pressure pulsations can reinforce each other. Volute type pumps, except for solids handling designs and small volute sizes, are often a double volute configuration. While even vane number impellers should generally be avoided in double volute pumps, proven designs with these do exist. Attention given to the impeller-to-volute clearance Gap ‘B’, shroud radial clearance Gap ‘A’, and the shroud sidewall gaps can help mitigate pressure pulsations and RSI troubles. There are many applications where the pump energy levels are relatively low, for instance, not greater than 200 meters head (650 feet) or 225 kW (300 Hp) per stage as per API 610, and the incidence of RSI and pressure pulsation problems is rare.
In high head-per-stage machines, even or odd vane number impellers can be paired with appropriate diffuser vane numbers with due attention given to the periodicity of the RSI as described texts such as Karassik, et al, ‘Pump Handbook Fourth Edition’ and Gülich, ‘Centrifugal Pumps Second Edition’.
In lower head machines, especially those less than 30 meters (100 feet), RSI problems due to a vane number combination are all but non-existent.
Pump hydraulic designers are aware of vane pass pressure pulsation and RSI issues. Most newer designs for double volute pumps, which comprise a substantial share of total pump production, will not have even vane number impellers. Additionally I suspect this tends to create a bias against designing with even impeller vane number impellers except when they are essential for performance purposes.
Thank you for posting this one, Randal
how to increase centrifugal pump efficiency more than 90%? is it possible or not?
Jugal, Yes, it is possible to achieve greater than 90% pump efficiency. In fact, both theoretical evaluations and actual experimental results indicate pump efficiencies well into the mid-ninety percent range are attainable. There are many different pump designs and a wide range of sizes to meet requirements for an endless variety of applications. Pumps tend to be optimized for several competing design constraints and while efficiency is important, it is not always the primary design parameter of interest. In fact, for industrial applications, reliability and safety are often top priority. The highest theoretical efficiency is actually seldom achieved given the range of design criteria and practical factors that determine the outcome in a given pump product. Best holiday wishes, Randal
sir, can we increase the pump’s discharge head by increasing vane numbers?
Kalpesh,
It depends on the design you are starting with. If the design has a low vane number for the given pump specific speed, then increasing vane number is often effective. Some higher specific speed impeller designs do not respond favorably to an increase in vane number. Often for impeller vane number changes, the vane design itself has to be modified to ensure adequate inlet and outlet areas. In other words, one may not be able to simply change the vane number without also modifying the vane geometry. Sometimes the vane inlet and discharge tips can be adjusted and sometimes an entirely new vane design is necessary.
Besides increasing head, an increase in vane number can possibly introduce instability in the head rise at reduced (throttled) flow conditions. This is potentially a problem when there are pumps operating in parallel.
Good luck with this,
Randal
Dear Randal,
All of these informations are really interesting.
I’m fully agree with you about the pump efficiency is not the first priority of petrochemical industry, and I think will be not in near futur. Just due to the capacity reduction of the companies located in Europe…
But for me the rated ratio flow/head is moving for a lower flow rate with similar head.
If increase the vane numbers is occur instability (vibration), does the multistage pump is still the best solution?
Or a vane geometry can easily allow more than 140mCL for 315mm diameter single stage?
Thanks for supporting,
Richard.
Dear Richard,
The trend for increasing impeller vane number is a reduction in vibration levels. Except, if the new vane pass frequency is near a resonant condition, then vibration levels can be much higher.
For process pump applications it is almost always more desirable to select a single stage pump over a multi-stage pump unless there is a substantial penalty in motor sizing.
Here’s what I understand of your inquiry. Let us say your motor speed is assumed to be approximately 2970 rpm. Then 140 meters head from a single 315 mm diameter impeller is possible, but detailed hydraulic design evaluation of both the impeller and the stationary casing components would be required.
I hope this helps.
Randal
Dear Randal,
How to choose good vane design for sand dredge pump? (Is this a relevant question?)
I have seen some impellers for sand dredge pumps with similar vane design as for water. But some others have simplified designs (4-vane but shorter ones, similiar to vane design for 6-vane).
What is the purpose of this simplified design? How does it affect the dredge pump performance?
Do you have any experience on this?
Thanks,
Richard
Richard,
A dredge pump is designed to pass solids. The internal passageways are designed to pass a specified minimum object size, and this is rated as ‘minimum sphere’ size. The minimum sphere size affects the impeller and the casing passageway sizes. Particularly in the small-to-medium scale pumps, the choice of impeller vane number and the vane geometry are often the result of design compromises aimed at achieving the specified minimum sphere size. The ‘simplified’ appearance you speak of is probably because the vanes have to have an excessively large passageway opening that can limit the degree of vane curvature and overlap, and limit the vane number as well.
The hydraulic design compromises for a solids handling pump typically results in a steeper H-Q rise to shutoff which can be a desirable performance attribute for many solids handling applications, but with reduced efficiency compared to a non-solids handling pump.
These are excellent questions.
Thank you,
Randal
Dear Randal
What effect does higher vane pumps has to the lower van pumps with respect to production in sand dredging.
Dear Chidi Cub,
The requirements for sand production are satisfied by selecting the appropriate dredge pump and its operating speed. The pump manufacturer chooses impeller vane number based on geometry that will satisfy solids handling requirements. Please refer to my comment reply posted Sept 19, 2014.
Regards,
Randal
Dear All,
Why the numbers of vanes in impellers shall be different to the numbers of vanes in diffusers?
Thanks,
Sofrani
Sofrani,
Good question. To economize on the radial dimensions or diameter of the pump casing, and in the interest of optimal performance, the desired hydraulic geometry of the diffuser typically demands a larger number of passageways compared to that of the mating impeller. To avoid excessive pressure pulsations and acoustic response from the interactions of impeller and diffuser flow fields, particularly with higher energy pumps, matching vane number combinations or their even multiples are definitely to be avoided. Certain even versus odd vane number combinations can also be undesirable depending upon the resulting beat frequencies.
Best regards,
Randal
The flow rate demands of our pump have recently changed to a higher capacity. We have also had vibration issues (6x vane frequency). The suggestion on the table is to change the impeller to 7 vanes to help mitigate these issues. Do you feel that this is the right approach?
Thanks
Dear Reese,
Whether changing impeller number is the best solution for either of these two technical issues cannot be answered here in this blog. Changing from 6 to 7 impeller vanes may very well be the answer, but there is some risk that the vibration problem does not resolve.
On the vibration side, there are both excitation and response frequencies. The various harmonics, side bands and possibly other attributes of the frequency spectrum must be considered for both the rotating and stationary components. Expert vibration analysis coupled with rotordynamic and structural frequency analyses may be required.
Concerning the higher capacity, the new system head versus flow conditions will need to be looked at to determine the appropriate pump hydraulic modifications. These could include design changes or alterations to both the impeller and the volute or diffuser.
Of course, if it determined that a different impeller vane number will solve the vibration problem, then the hydraulic modifications would incorporate that.
I hope this helps.
Best,
Randal
Hi,
On a 6-inch centrifugal mixed sand and gravel pump with a large diameter impeller and a small diameter impeller what is the difference in discharge volume? I also would like to know what will be the difference between a 1-vane impeller, a 2-vane impeller and a 3-vane impeller on the same type of pump in terms of discharge capacity.
Thank you for your assistance.
Dear Danny,
The impeller diameter and pump speed determine the pump total head. So for a given speed, the larger impeller diameter would be expected to generate more head than a smaller one, but both yielding approximately the same volume of flow. The differences in pumping volume based on different vane numbers is generally that the higher vane numbers will yield more flow and some additional total head. This holds true up to a point where increasing the number of vanes may not yield additional flow, but may, in fact, cause a reduction in flow due to vane blockage effects.
I hope this answers your questions.
Best regards,
Randal
In multi stage pump, impellers r always in odd numbers??? If it correct Pl. Give the reason of it.
Dear Pritesh,
Because your comment pertains to a post about impeller vane number, it is not clear whether your question is about the number of impellers or the number of vanes. But I will at least give a short answer to both possibilities.
Concerning the number of stages for a multi-stage pump, there is no set number. The number of stages can be either odd or even. It primarily depends upon meeting the required head. If it is a custom design, the manufacturer will usually select a design specific speed that provides an optimal balance of the number of stages, head per stage and efficiency. For very high head, mostly types BB3 and BB4 applications, the manufacturer will deliberately choose a specific speed to limit the maximum number of stages.
Concerning the number of impeller vanes in multi-stage pumps, type BB3 pumps will normally have impellers with odd vane numbers in order to avoid two vanes passing the double volute lips simultaneously in order to avoid reinforcement of pressure pulsation magnitude.
I trust this answers your question.
Best regards,
Randal
Mr. Ferman,
Is there a set rule regarding impeller size (diameter) vs suction and discharge diameter for solids handling recessed 4 vane impeller style gravel pumps. I have built a 6×6 gravel pump utilizing 10″ diameter flat-plate 4 vane recessed impeller and the pump performed with great suction power. I have built the same style pump in 8×8 w/20″ impeller and it too created tremendous suction. I am interested in building a 16×16 gravel pump and am trying to figure out my impeller diameter based upon the other 2 successful pumps. My rpm will not exceed 1800. Any advise would be appreceated.
Erich
Erich,
Glad you asked! Not because I can directly answer your question about the 16×16 gravel pump. But you may be saving yourself some grief. In the world of pump design, an assumption that one can simply scale up a working design may yield unanticipated and undesirable results.
What I would do, as a minimum, is develop the suction velocity triangles for the two successful designs, and then compare those with the proposed new design. To do this, you need to know a design flow, for each case. I would also consider velocity magnitudes – those are always a consideration and with solids handling equipment: you do not want to find that you are too frequently replacing/repairing components due to an excessive rate of erosion.
Developed total head versus flow is relatable to the impeller discharge velocity triangles and I would compare those as well.
I hope this helps,
Randal
Erich,
Another comment if I may. Your inquiry starts with “Is there are set rule….” In the practice of pump hydraulic design, I seldom find a “set rule.” The design space is multi-dimensional and it is often the case that a desired performance characteristic can be achieved with a variety of different hydraulic geometries. Of course, the more that constraints are placed on the problem and when design optimization is done, the range of possible geometry variations is reduced. On the other hand, in the industrial world, it is often the case that a “workable” design is all that is needed and that opens the door to a wider range of possible solutions.
Randal
Mr Ferman,
What is the effect of the impeller internal volume on flow rate? If I decrease the impeller volume by a half and keep the same rotational speed will I decrease the overall pump flow rate by a half?
Thanks,
Bob
Bob,
It depends upon how it is done. If the objective is to cut the flow in half, then I suggest ‘side-cutting’ to trim down the B2 dimension (impeller outlet), starting from the eye shroud side, to half the width. The impeller eye has to be adjusted so that the velocity triangles are ‘correct.’
This is relatively easy to do if you have a semi-open impeller. If it is an enclosed impeller, you might be better off creating a new one or just finding another pump.
Another option is to design a new impeller and have it cast or 3D printed in either plastic or metal, and finish machined.
Let me know if you need help with this.
Best,
Randal
Randal,
Thank you for responding so quickly. The pump in question has an enclosed “double-sided impeller” so a new impeller with the same external dimensions that can be “dropped in” is probably the way to go. This is a special pump for an Aerospace application that is being adapted for another system.
Although I am an engineer this is outside of my area so I have been doing research into the design of turbo pumps that use centrifugal pumps. Although I have found a lot of material (Including the Affinity laws) the information on the impeller seems to be limited to the diameter and the rotational speed with some information on the number of blades. I have not been able to find anything on the internal volume of the impeller which should have an impact on the flow volume. Do you have any references? Also, I may be able to share some more information if I can contact you directly. Is it possible to email you? My email should be in the information I supplied with this reply. Thanks Again. Bob
Hi Randal,
Your replys which I have read so far had been highly accurate and helpful, thank you! Here is a non standard question which perhaps you could help with. I have a spinning tube or it could also be a cone shape that I want water drawn up in 2 different ways. One to achieve the highest pressure possible and the second for the highest volume. I played with some axial flow pump impeller designs but as there is no central shaft and the drive needs to come from the spinning tube. Figuring out an effective way to do this is tricky as there is no diffuser or I have not figured out a way to integrate one yet. I could perhaps bolt the assembly to the floor and have the shaft stationary with the diffuser connected on the other side of the impeller but there are probably better ways to do this!
If you have some creative ideas it would be nice to touch base.
Cheers
Che
PUMP:13 DURCO GR2 MK3 ANSI3 FML SGL AFLA.
How do I know how may vanes the impeller/pump if I do not have the pump in front of me? The pump information is above
Dear Jimmy,
Unless you can find documentation indicating the impeller vane number, that information is essentially impossible to know, short of direct observation. The most reliable observation would be to look at the impeller exit perimeter because sometimes an impeller will have partial length vanes that might not be visible from the suction inlet side.
Good luck,
Randal
I suspect that theoretical calculations would indicated that an infinite number of vanes of zero thickness would be the optimum for a fluid with zero viscosity. As viscosity rises and both the frictional losses due to vane surface friction and the tolerable stress on the vane material are take into account, the desired number of vanes will go down. The optimum number of vanes for a real fluid will be influenced by the viscosity of the fluid which drags on both sides of each vane, causing both head loss and stress on the vane. So, the optimum number of vanes is going to depend on the viscosity of the fluid you need to move. Some really smart (not I) engineer could calculate the optimum number of vanes, but the results will come to a small set of practical choices; more than 2 and probably less than 16. Pump manufacturers are not going to give the freedom to choose any number of vanes you like (unless CNC machining to customize impellers takes over) and, the better pump manufacturers have done their own tests to determine best impeller design for fluids of different viscosities. They should offer guidance with that choice.
Dear Steve,
The optimal number of impeller vanes will depend upon the pump type and application. For instance, the need for low NPSHA suction performance often dictates fewer vanes. Or the need for a rising head characteristic over a wide range of conditions, including parallel pumping operation, may dictate vane number choices. In the realm of engineered-to-order high pressure multistage pumps, it is not uncommon to find a vane numbers for the first stage impeller plus a couple of different impeller vane numbers for the series stage impellers. These choices can be dictated by both suction performance and the specified operating point along the head versus flow curve.
Today, a user could hypothetically design his own impeller and a service, repair, parts replicator shop would be happy to make it and install it in the pump of his choice. I see some degree of that happening when a pump user hires a consultant or engages a pump manufacturer to modify or re-rate a pump.
The significant development I now see is an individual’s or a user’s ability to 3D print a prototype part. This democratization of manufacturing is having profound effects in the making of molds, dies and parts for prototype testing and parts for actual industrial machinery service.
For the most part, the discussion between manufacturer and user regarding impeller vane number mainly occurs in the context of vibration or acoustic excitation frequencies and harmonics and is normally only a concern with the higher head per stage pumps.
Thank you for the comment,
Randal
How much does the height at the inlet and outlet sides of the vane affect the performance when designing the impeller?
If there is a good way to select this, I would like to know.
Yu,
The inlet geometry is normally based on velocity triangle calculations.
The exit can be based upon:
1) modeling from pumps of known performance and similar specific speed,
2) textbook or handbook design charts and formulas,
3) velocity triangles with impeller slip factor from textbook or handbook formulas, or
4) some combination of these.
Design experience helps.
I trust this is useful.
Yours,
Randal
Hi there! First of all, thanks a lot for all the knowledge put in here.
I have a question about fixed displacement unbalanced vacuum vane pumps with oil bath. Specifically I would like to know how the rotor width and diameter relation can change pump efficiency (mantaining a constant volumetric capacity). Is there any rule of thumb for this?
Thank you!
Leonardo,
The article does not address positive displacement vacuum pumps, but your question may have some relationship to centrifugal pump hydraulic design. With these, the more efficient hydraulic passageway shapes are closer to being either square or round, as opposed to having an elongated or ‘flattened’ aspect ratio. Also, having less impeller rotating disc area relative to its normal inlet and outlet areas – this is related to specific speed – tend to be more efficient. Centrifugal pump hydraulic designs which minimize the wetted surface area will tend to be more efficient. I should note that there is often a broad range of design solutions for which the differences in efficiency are small.
Not sure that there is a ‘rule of thumb’ on positive displacement rotary vane vacuum pump designs. It would be a valuable exercise – not a trivial one though – to create an analytical spreadsheet performance model to check out possible design geometries. That said, tribology and mechanical wear are likely to be your largest design and practical concerns with the rotary vane vacuum type of pump.
Interesting question – thank you!
Randal