Impeller Size vs Efficiency 1

[wrd2032]

We're trying to squeeze some electricity out of our oversized pumps, but they already have VFDs. Typically they run no higher than 75%. I heard you can improve efficiency by making sure the impellers are maximum size. Is that so? Any other improvements available to a pump on a VFD already?

Thanks.

 

[1gibson]

No. "Making sure impellers are maximum size" refers to the original pump selection. Depending on the pump type and industry standards, you can't have full diameter impellers because there must be provision for future head rise. And if it comes up short, you have to re-select the pump. Not something you want to do as part of commissioning.

If you want to retrofit non-metallic wear rings (Vespel, PEEK, etc)they can have tighter clearances than metallic wear surfaces. These need precise interference fits, and special 'pinning' to prevent rotation, so don't just try to slap them in. If appropriate for the service (no idea) look into coating the hydraulic passages, Belzona has a number of products.

But, your best bet is to make sure you are running at a speed that puts the pumps at or near the best efficiency point.

 

[DubMac]

No need to look for a rule of thumb on impeller diameter; look at the pump's original performance curve to see where the BEP lies, and go from there.

As 1gibson says, there are a few different techniques available to possibly squeeze a couple pts out, but they usually have a tradeoff. Tightening up your clearances absolutely will make the pump more efficient, but it can lead to quicker failure if the pump is not designed to handle them (too flexible a shaft or bearing design). This will depend on your service as well (???).

Impellers can often be "hand-dressed" with backfiling and polishing to gain a point or two, but then again this is more of a black art and may not be applicable to your pump.

If you were to let us know what pump make, model, size you have and what service you have it in, I imagine someone on this forum could give you a much clearer answer to your question.

 

[stanier]

You could save energy by putting in decreased size impellers and doing away with the VFDs. VFDs do nothing to save hydraulic power consumption and burn up energy in their use, their cooling and they have embedded energy.

More efficient to have a correct sized impeller and a control valve downstream to keep the operating point on BEP.

For research on pumps use these websites.

http://www.mcnallyinstitute.com

http://www.pumpsystemsmatter.org/

http://www.pump-zone.com

http://www.pumpfundamentals.com/

“The beautiful thing about learning is that no one can take it away from you.”
---B.B. King

http://waterhammer.hopout.com.au/

 

[DubMac]

Why were the VFD's added in the first place? Were they a necessary part of the service. i,e, was there a need to adjust the speed/flow on the run? Or were they put on because " you just gotta have a VFD because everybody else does?"

If due to the latter, then listen to Stanier, they have inherent inefficiencies. You can do a cost/savings analysis balancing the efficiency($$) savings against the cost of new impellers and make your decision on length of payout.

 

[Artisi]

You need to answer DubMac's question before anything else can be discussed that will add to the discussion.

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.)

 

[bimr]

If your pump is operating at less than full capacity, then you would want to decrease the impeller size, not increase it in order to have the pump operate at a better duty point. Unfortunately, this solution will enhance the risk of cavitations in the pump, and there is no possibility to increase the flow or head if it should become necessary.

However, a VFD mounted on the pump is by far the most efficient way to change the duty point of a pump system, and reduce the power consumption. The VFD gives much more flexible pump control and reduces the risk of cavitations. See the graph.

In order to develop recommendations on the most efficient operating scheme for your pumping system, it is necessary to prepare a system curve and pump operating curves for various conditions, as others have noted.

 

[hydromech]

VSDs are employed to adjust the pump speed to reduce energy consumption at times when the pump performance can be reduced. If the flow or head requirement reduces as part of a process cycle, or if the heat energy transfer is reduced for a period of time, the pump speed can be reduced. However, the impeller will have been designed to suit the highest load in terms of flow and head and trimming the impeller to reduce the performance will mean that pump will never achieve the design performance. Over speeding the pump will increase inlet velocity and will lead to cavitation.

If the pump is oversized because of a permanent process change, then change the pump because trimming the impeller will increase the efficiency losses and the potential savings will diminish.

VSD are not a magic bullet for improving energy efficiency. Getting it wrong will cost more in the long run.

Adrian

 

[stanier]

The original post spoke ofthe pumps not running > 75% (speed) hence a reduced size impeller is solution. The vfd could be kept and the motor run at 100% speed. ie its most efficient.

“The beautiful thing about learning is that no one can take it away from you.”
---B.B. King

http://waterhammer.hopout.com.au/

 

[TenPenny]

I don't think it makes much sense to worry about the motor efficiency without first looking at the pump efficiency.

Let's consider a 100 hp 4pole motor, efficiency will range from 95% at full load to 94% at 50% load.
Over a similar range, a centrifugal pump efficiency could range from, say 84% to 60%.
I pulled these numbers from curves chosen at random, so they are typical.
I know where'd I focus my attention, if I was concerned about wire-to-water efficiency. How much improvement in motor efficiency are you ever going to achieve?

 

[EnergyMix]

I'm going to echo Dubmac's advice from the very start -- have you looked at your pump curves and compared them to where your pump is currently running? Your VFD may have moved your pump off its BEP. You also should look if something else in the system changed. It would be a shame to spend a lot of money replacing the impellers only to find out you're still only getting 75% (of whatever) because of something you didn't account for before rushing into changing a component.

I'm a proponent of figuring out the problem before enacting a solution.

Want to know the do's and don'ts of Eng-Tips? Read FAQ731-376.
English not your native language? Looking for some help in getting your question across to others or understanding their answers? Go to forum1529.

 

[wrd2032]

Thanks for all the help.

There are several centrifugal pumps, Goulds 3196 3x4-8, 2x3-13 and similar. The motors turn at 25%-75% of 60 Hz when the pumps are in service. There aren't any records of what impellers are actually in the pumps, so looking at pump curves is somewhat guessswork. Service is water. Load is highly variable - I mean to say each pump varies from 25% to 75% load, not that each is a constant within that range.

 

[EnergyMix]

But you can plot a new pump curve and compare it to the original. Just saying that changing out the impellers can be expensive in terms of materials and downtime. If they're not the problem, then you go through all that and still have the same problem.

Wouldn't it be cheaper to do some up-front work first?

Want to know the do's and don'ts of Eng-Tips? Read FAQ731-376.
English not your native language? Looking for some help in getting your question across to others or understanding their answers? Go to forum1529.

 

[1gibson]

Yes there are, contact your distributor or the OEM and ask for a replacement impeller. State your concern (you have a VFD and might want to increase/decrease diameter after you look at things in more detail) and request that they include the impeller diameter on the quote. If it is a distributor (or helpful OEM aftermarket team that will work with their app engineers on it) you may even get them to send you some curves at different diameters.

 

[DubMac]

It is odd that the original diameter is not on the nameplate...however, per 1gibson, it is not that difficult a task to track down with the S/N.

Any chance you could throw us a bone and describe the service a little better??? It sounds like some sort of HVAC/booster application, but who knows.....

With a hugely varying load, this may well be a service well suited to VFD's....but who knows?

 

[TenPenny]

First step is to define what you need the pumps to do, ie, flow and head range. Not just % of speed at current trim, but actually figure out what the pumps are doing and what you want them to do.

Hopefully, you have some pressure gauges on the system, or you can put some on there. With gauges installed, figuring out the installed impeller diameter is easy, just run the pump at shutoff for a few seconds, and read the suction and discharge pressures.

 

[ccfowler]

wrd2032,

Am I correct in inferring that the VFD's are actually retrofits? If so, this seems to be a classic case of presuming that adding a VFD'S will inherently improve efficiencies all by itself. In the case of most pump installations, the simple addition of a VFD to an existing pump system almost always increases energy consumption due to the combination of the parasitic losses of the VFD itself plus the fact that the pump is almost always forced to spend all of its operating time farther from its BEP and therefore is forced to operate at lower efficiency. In nearly all cases, the energy losses associated with the use of a control valve will actually be smaller than the losses associated with a retrofit VFD.

The closer the system is to having a well selected pump operating near its BEP in an application where the duty is a varying flow rate in a pure circulating duty, the greater the benefits of applying a well selected VFD to this pump. Conversely, the greater the constant head requirement (a duty such as a boiler feed pump being an excellent example) the greater the likelihood that a VFD will be a near perfect waste of time, energy, and money.

Valuable advice from a professor many years ago: First, design for graceful failure. Everything we build will eventually fail, so we must strive to avoid injuries or secondary damage when that failure occurs. Only then can practicality and economics be properly considered.

 

[LittleInch]

wrd2032. In answer to your OP, you can improve efficiency, but if you're operating below the duty point of the pump, this could require smaller impellors. Max size impellors, even if possible, could create excess dead head pressure. However there are some inconsistencies and gaps in the info to allow people to comment effectively. This includes - range of flows, you say 25% to 75%, but it's not clear what this is a percentage of. I can't believe it's speed as I don't recall a VFD being able to go that low, but is it pressure, flow or load? By load do you mean power?. You say this is highly variable, which is perfectly OK, but how much time does it spend in each range.

Whilst there are a few VFD sceptics responding here, in many cases VFD drives will give you lower overall power consumtion, but efficiency may suffer. Hence if efficieny drops to 50% from 70%, but flow drops by 50%, then you use less energy, providing your VFD is relatively modern, compared to a fixed speed unit which then throttles the flow turning your fluid energy into heat. You may find that a mixture of fixed speed pumps and smaller VFDs to make up the variability would work better.

Well before you get around ordering different impellors, you need to have your system looked at from a hydraulic point of view to establish your system curve(s) and operating points. If your requirment is for variable flow and variable pressure (head) that is a lot different from a fixed or nearly fixed head, but variable flow. As said above, the pump manufacturer usually holds records dating back to the dawn of time about what was supplied, so a search for the pump serial no is also required.

Bet you didn't think a simple question would generate so many variable responses....

My motto: Learn something new every day

Also: There's usually a good reason why everyone does it that way

 

[wrd2032]

LittleInch,

Indeed I didn't think my question would generate this much response.

To give some background, we corporate engineers were visiting a plant when I asked this question. We'd audited many of the water pumping systems at another plant, and wanted to repeat the process. The systems involved are chilled water, CIP water, cooling tower water, and perhaps a couple others. Fairly constant pressure, except for variable friction losses, but great flow variability. Since these pumps were put in, we've reduced our utilities loads greatly, so throughout the company they tend to be oversized. In the original plant audited, the pumps were almost all single speed, running at 100% flowrate, with the unneeded flow recirculating back to the tank (Often 60% of the pump's flow). It was decided to throw some VFDs on, since the flowrates were always going to be variable, as they're driven by batch manufacturing processes.

When I got to the new plant, I found that the plant maintenance manager had already put VFDs on all such pumps. Whether this was part of the original design or a retrofit, I'm not sure. This pump audit was one of many things we were scheduled to do on the trip, so we didn't call ahead to get a handle on the existing situation. Once there, I wondered if there were any other adjustments that could be made to the few pumps in the system that never reached 100% design flowrate, given their variability. So I asked my original question, about impeller size.

Ever since that week, I've been back at the corporate office, a thousand miles away. I'm trying to get access to more pressure and flow data for the systems out there, but it's slow going. From the debate back and forth, I'm getting the feel that there's not dramatic electricity savings available for a pump with varying flow. Once I get access to real time flow and pressure data, I'll go back to the manufacturer to get pump curves and double check to see if a new impeller could pay for itself, or if a control valve would beat a VFD, but our ROI requirements are pretty strict, so I don't think this is going to go anywhere.

Thanks again for all your advice.

 

[1gibson]

ROI aside, make a note of your findings for next maintenance/repair cycle. If you happen to need a new impeller anyway, good to have the results on file with instruction to maybe order smaller diameter. Even if the impeller is good and you just need to replace wear rings, that involves teardown, inspect, remove old WR, install new WR, and confirm balance. Adding a trim operation in there will not significantly increase cost.

And if nothing else, having your findings on record may avoid or streamline performing the same review in the future.