Harmonics and VFDs 22

[Marke]

One of the areas that I operate in, has a high density of VFDs on pumps on relatively weak supplies. The result is that the high levels of harmonics on the VFD inputs has accumulate in the supply and is causing a high THD of the supply voltage. While we do have supply regulations covering harmonics, in this instance, the harmonics are higher than they should be.

There is an option of using zig zag transformers and six phase rectifiers as a means of reducing the harmonics drawn by drives however in this case, the drives are already installed.

There is a transformer for each drive and sizes range up to 200KW.

One thought that I had, was that for future installations, and there are new installations going in all the time, that the new supply transformers be designed with a zig or a zag winding to give a phase shift, and install equal loading on the leading and lagging phase shifts. This should act like a twelve pulse input on one drive, only it will be across two drives.

Any thoughts on this??

Best regards,

Mark Empson

http://www.lmphotonics.com

 

[gepman]

This question has been asked before in this forum. See

http://www.eng-tips.com/viewthread.cfm?qid=28370

You can also google "understanding pump curves".

 

[nightfox1925]

Thank you very much gepman!!!

GO PLACIDLY, AMIDST THE NOISE AND HASTE-Desiderata

 

[waross]

I had some pump adventures years ago. Two pumps on alternating service pumping sewage uphill. Changing regulations resulted in the discharge being moved much further up the hill. This put the pumps so far up their curves that one pump could no longer keep up with the flow. Tried both pumps in parallel ad the slight increase in dynamic head lowered the flow through each pump enough that the water boiled inside the pump housings from excess shear.
Higher pressure pumps meant bigger pumps and motors and a couple of hundred feet of buried conduit to upgrade to handle larger feeders.
The solution was to purchase a foot mounted pump and belt drive it over speed for the main pump. The other two pumps were run in series as the back-up system. This still meant replacing and adding feeders and special permission from the AHJ to install the conductors in the existing conduits.
If VFDs had been available then, I could have just installed two on the existing motors, run them up a litle over speed, and gone home. Torque, and Hp at over speed,- not a problem that far up the curves. Efficiency,- so what. But better than what we had. Compare optimum run times with kludge fixes.
respectfully

 

[waross]

For those of you who are following this with interest:
For a dramatic illustration of the effect of speed on pump curves, many performance booklets give curves for the same pump at 1800 RPM and 3600 RPM. Take a quick look and you will see that speed has a dramatic effect on pump performance and why many pumps run at 3600 RPM.
Respectfully

 

[waross]

An observation for comment by others. Current measurements are common when setting up pumps, whether VFD or valve controlled.
Because of power factor issues, current is an unreliable indication of real load particularly at light loading. (Remember we are talking about savings or losses of real power in the range of 5% or less.)
What is not so common is accurate flow measurements.
If reduced flow results in 10% or 20% longer running times, an apparent 2% gain in efficiency may be misleading.
Rather than comparing relative efficiencies and energy consumption of motors with various drive methods, it may be more accurate to try to compare the total energy required move at given mass of liquid from point A to point B.
As an example, consider a hypothetical deep well where the draw-down is great and is proportional to the pumping rate.
The greatest efficiency may be approached by pumping continuously at a rate that results in the least draw-down.
It will take more energy to lift 1000 gallons 250 feet than it will to lift the same mass 200 feet.
In regards to pump curves we must not confuse efficiency with effectiveness. As the head on the pumps in question is increased, the effectiveness (and the flow) drop off faster than the head rises. The result, less energy required, but also much less flow or delivery.
The wholistic approach is to consider the total energy required to do the job, not the rate at which the energy is consumed.
Respectfully

 

[waross]

Mark, on your original post, am I correct in assuming that these are individual customers?
Will the power company consider delta/delta connections on new installations?
Respectfully

 

[Valvecrazy]

I have studied the Energy Savings Calculator in the Standard Performance Contract at the following link

http://www.aesc-inc.com/download/spc/)

given by gepman. If the utilities are giving cash incentives based on this "calculator", they are paying out a lot of money for nothing.

This so called "calculator" has no provision for even looking at the pump curve. It has no provisions for figuring the minimum possible pump speed. It recommends that you plug in 50% as the minimum pump speed and 96% as the efficiency, which are both bogus numbers. Anytime you plug in the "recommended" numbers, it shows a significant savings in energy. When I plug in the real numbers, it shows that there are NO SAVINGS with a VFD.

Since head drops off by the square of the speed, the Minimum Possible Speed is the most important part of the calculation. Everybody seems to think that you can just keep slowing a pump down to decrease the flow rate, even though this is not possible when a static head or constant pressure is required. From the example curves given in this thread, 10% to 15% is the most you can slow a pump down and still produce the TDH required. Even the curve from the oversized pump given by Marke, can only be slowed down by 28%. None of these real numbers show any energy savings when plugged into the so called Energy Savings Calculator.

This "Calculator" also recommends that you use 96% as the efficiency. This is a fictitious number concocted by VFD manufactures. The true Wire to Water efficiency should be figured and is usually more like 80%.

If they really have a fleet of 40 pump testers, shouldn't there be at least one of them who knows how to read a pump curve and figure minimum speed using the Affinity Law?

If you would like to see how pumps really work, I have several systems on my pump test stand in Texas that can easily be switched between VFD and Valve control. They easily prove everything I am saying is correct. If you have a pump test stand set up that proves otherwise, then you have either way oversized the pump for the application, or no one even looked at the Brake Horse Power curve when they were choosing the pump.

As for the sewage pumps that could have had there speed increased to solve the problem, horse power increases by the cube of the speed. A slight increase in pump speed nearly always means that a larger horse power motor is required. If these motors were not already maxed out, then a larger diameter impeller could also have been used to solve the problem without needing to vary the speed.

In all my years, I have yet to see a situation where a VFD could save enough energy to even pay for itself, when compared to a correctly chosen pump being throttled by a valve. The propaganda that says VFD's save energy is so widespread, that not only do most people blindly believe it but, they also get very angry when I try to show them the error of their ways. Even the "Energy Savings Calculator" shows no savings with a VFD when you plug in the correct numbers instead of the "recommended" numbers.

 

[waross]

Hello Valvecrazy;
Thank you for considering my post. Changing impellers was not an option and that far up the curve, so little work was being done that the cube law was not a problem.
The drive gurus will point out that VFDs on old existing motors is not a good idea. Rewinding to VFD quality would still have been an economical option.
Respectfully

 

[waross]

gepman and Valvecrazy;
I understand that both of you have access to pump test stands.
Can either or both of you set up a test to measure the energy consumption (KWHr) to raise a given volume to a given head?
Filling an elevated tank or wasting the discharge out an elevated pipe if the volume can be accurately measured would be the most acceptable test.
Then, apply in turn, a VFD and valves to reduce the flow 20%, 30% or 50% and measure the total KWHr required to pump the same quantity against the same head.
This test should also reflect any savings from reduced dynamic head, although dynamic head issues are much dependent on site conditions.
Respectfully

 

[gepman]

valvecrazy
I agree with your assessment of the SPC calculator and that is why I do not use it. I have pointed out to the utilities the problems with this portion of the software. I only mentioned it since you stated you work with these utilities. I do not know when it will be corrected since it has to go through at least two large utilities (PG&E and SCE) and the CPUC (California Public Utilities Commission). However the ABB savings calculator which I recommended in my previous post does not have any of these problems. It is the best calculator that I have seen which does not have the actual pump curve built into the calculation procedure. The calculator is available at

http://www.abb.us/product/ap/seitp322/24b03100d005c31ac1256e040043f4c1.aspx.

Since this calculator has inputs for both static head (which it uses to calculate minimum pump speed) and pump shutoff head (which it uses to estimate the shape of the pump curve between the nominal operating point and the no flow condition) it takes into account all of the issues that you mention. It also takes into account the efficiency of the VFD (if you use a VFD at 100% flow for 100% of the time it will show a negative savings). Another good feature is that it calculates savings based upon several different existing control strategies including pump on/off control and control valves.

Again I will state that I do not disagree that pumping systems with a high percentage of static head/total head are not good candidates for VFDs and a lot (if not most) of well pumping applications have this high static head characteristic. Play around with ABB calculator to see what the break even point is. Even when you break even on energy you then have to consider the capital and maintenance cost of the VFD. I also disagree with your previous statement that most pumping applications have a high percentage of static head. Most process applications, which have a far greater number of installations that well pumps, are not high static head. I also totally disagree with you that throttling the flow does not consume energy across the valve. A pump system operates at the intersection of the system curve and the pump curve. When you control via a valve you are adjusting the system curve and when you control via a VFD you are adjusting the pump curve. If you have a centrifugal pump curve which is either flat or rising towards the shutoff point (which is almost all centrifugal pump curves) then when you reduce flow with a valve you are always pumping at a higher pressure than when you reduce flow with a VFD. It is also most likely that the efficiency at this reduced flow point will be better with the VFD than with the valve. Your comments show that you do not have a good understanding of hydraulics. Remember that VFD savings come when you VARY the flow not keep it constant. If the flow is constant you should try to design your system without any throttling. However it is common for the required flow to be known however in sizing the pump a 10 to 20 percent safety factor is added to the head. This is done because it is difficult to calculate accurately the exact head loss in a system. For pumps that operate continuously it can pay to install a VFD to reduce the pump curve to where a valve is not throttling this 10-20% head loss. I encourage you to try this out.

You are also wrong on your efficiency numbers for VFDs. I have instrumentation which is capable of measuring power consumption on both the line and load side of VFDs and I have never seen efficiencies below 90% even at extremely low loading of the VFD and much higher is typical.

I would like you to produce some articles or research that show that throttling flows with valves is more energy efficient than controlling flow with a VFD.
I will try to setup a test at the following test facility

http://www.sce.com/RebatesandSavings/EnergyCenters/AGTAC/exhibitsanddisplays/.

See the bottom of the page, "Pumping Technology Station". I don't think that anything will change your mind because you have too much invested into believing your sermon on valves. It is nice to think that you know something that the rest of the world does not understand.

Also I would not belittle the pump testers. Each of these professionals accurately test more pumps in a year than you do in a lifetime. I have been trying to upload a 6.5MB file on their pump testing which might help you understand wells, ag pumps, hydraulics, efficiencies, etc. however my guess is that this file is larger than the limit for this site. There is no website link to this paper.

Regarding your comment that you have yet to see a system where a VFD could pay for itself, take a look at

http://www.bpa.gov/Energy/N/Utilities_Sharing_EE/Workshop2007/pdf/DonCasada.pdf.

This is the kind of work that I do for SCE and PG&E. I don't know if you understand the concept of varying load or capacity. These systems were not oversized, they just have to deal with varying flow or loads.

I am not going to waste my time responding to you anymore however I will try to set up the test I mentioned above or maybe another test out in the field.

 

[gepman]

valvecrazy
Post a pump curve (either flat or rising upward at shutoff) and indicate the static head of the system, the nominal operating point (flow and head with valve wide open), and the throttled operating point. I will post hand calculations and ABB calculator calculations.

Remember VFDs are applied where there is substantial flow variation so the throttled point must be significantly less than the nominal point. Also the static head can't be the system head but if you give me that value then I will also calculate it at 80%, 40%, and 0% static head/total head values.

 

[Valvecrazy]

I don't think that anything will change your mind because you have too much invested into believing your sermon on "VFD's". It is nice to think that you know something that the rest of the world does not understand.

I guess you don't know how many pumps I have tested in my lifetime. It will take me a little time but, I will try to video a test and get it on line.

If you want to do the calculations with the ABB system, use the curve I posted earlier. Remember that this system needs 231' of head at low flow as well as high flow, as most pumping applications I am familiar with do. Check it at 100, 600, and 1200 GPM for equal amounts of time. Although, I can see these numbers on the curve. The only thing not included on the curve is the loss from the drive itself. If you say a drive is 96% efficient, then that means you are losing 4% just for using the drive, and that should be added back to the numbers from the variable speed curve.

I don't know if you understand the concept of varying load or capacity. I do know there have been too many VFD salesmen and engineers involved in making these Energy Saving Calculators. Feathering their own nest, so to speak.

 

[Marke]

Hi Bill

am I correct in assuming that these are individual customers?
Will the power company consider delta/delta connections on new installations?

Yes, all individual customers and yes, I am exploring options with the Power Co, and phase shifting transformers are an option. We would probably have to go for star/star rather than delta/delta because we operate an MEN system out here, plus delta driving VFDs is not a good move.

Best regards,

Mark Empson

http://www.lmphotonics.com

 

[Marke]

Hello gepman

I have been trying to upload a 6.5MB file

Contact me (via my website) and I can arrange to host that file for you so that all can see.

Best regards,

Mark Empson

http://www.lmphotonics.com

 

[gepman]

valvecrazy

As promised here is my reply. I don’t expect to convince you since I can only offer engineering principles, calculations, manufacturers literature, articles, research publications, and similar material while you offer your unsupported statements that valves save more energy than VFDs.

I have taken the curve that you posted for the Goulds 12WAHC pump (by the way I could not find that particular pump on the web site) but they are a very good manufacturer. Now you stated that the system needs 230’ of head at low and high flows. Please see item #6 in my large post above to see why constant head requirements are basically impossible in a real hydraulic system. To summarize the flow causes a head loss through the piping system which is proportional to the square of the flow. Your requirements would necessitate no piping system including the valve which you advocate. What I have done is take a minimum piping system with the valve (that you described previously) which has a head loss of 7 psig (16.2’), 4.8’ of head loss through the pump discharge pipe in the well, and another 9’ of miscellaneous head loss through foot valves (check valves), elbows, and other fittings in the system. This gives a total head loss of 30’ at 1200 gpm. I have used this system data in the ABB calculator. I have used annual operating hours of 4000 which is probably high for many farmers irrigation system but it wouldn’t be high for a processing plant or a large farm with one well (unlikely).

The first calculation (page 1 of attachment) shows that at 1200 gpm, just by pulling out your valve, the energy savings would be 56,911 kWh per year. (When a value is entered in the "head over open throttling valve" it will calculate the savings eliminating this valve). At 600 gpm (50% flow) (page 2 of attachment) the savings is only 61,253 kWh per year. If you had lower static head the savings would have been much higher. At 240 gpm (20% flow) (page 3 of attachment)the savings would be 97,373 kWh per year. This is due to the slight rise in the curve near the zero flow point which allowed the pump to be slowed down more than a perfectly flat curve. If you would have selected a pump with a more sharply rising curve (see page 4 of the attachment with the same efficiency at the original 1200 gpm @ 230’ pump, the Goulds 9TLC) the savings on all of the above would have been even more, 187,833 kWh at 240 gpm. (page 5 of the attachment, all that changed is the value for the maximum head).

I am only going to calculate the energy savings by hand for one example to compare it to the ABB calculator. (See page 6 of the attachment). Let us use the original pump curve at 600 gpm. With a valve the pump would have to pump 600 gpm, approximately 260’ head, and 69% efficiency. With a VFD the flow would be 600 gpm, the head would be the 200’ static head plus .25*13.8’ (600/1200 squared times the friction loss without the valve) which is 200+3.45=203.45’. The point on the corresponding 3450 rpm curve has an efficiency of approximately 73%. I would estimate this point as 684 gpm @ 260 gpm. Using 88% speed this would give flow at 602 gpm (.88 * 684) and head at 201’ (.88^2 * 260). Using the horsepower equation (Q*H)/(3960*eff) the bhp of a throttled pump is (600*260)/(3960*.69)=57.09. Converting this to kWh for 4000 hours of pumping including motor efficiency of 95.4% is 4000*57.09*.746/.954= 178,570 kWh. For the VFD pump this is (600*200)/(3960*.73)=41.5 bhp. Converting this to kWh we have 4000*41.5*.746/.935=132,444 kWh. Note that this includes a 98% VFD efficiency (.98*.954). Therefore we have a savings of 46,126 kWh which compares reasonably well to the ABB calculator value of 61,253. The difference in the calculator value is probably due to estimating the efficiency both for the throttled pump curve and the VFD pump curve. The estimating algorithms in the software would probably work more accurately for systems which had a lower percentage of static head. The percentage of static head in this system is 200/230=87% which is very high. The savings on the original system with 50% static head (115’) would be much greater, 114,609 kWh (see page 7 of the attachment).

I believe that I have validated the ABB pump save calculator (at least as a preliminary tool). You can always calculate using the pump curve if you have it. I have yet to run into a farmer that has his pump curve. Most don’t even know what pump is installed especially if it is a submersible or a line shaft pump.

Here is what you should learn from this:

1) Just adding the valve costs you energy.
2) You should like VFDs because using a VFD is like trimming the impeller, it just shifts the pump curve slightly downward as you reduce speed. It helps to keep the BEP at the operating point while a valve shifts the operating point away from the BEP.
3) You can save even more money using a VFD if you select a pump curve with a rising slope towards the shutoff point with NO loss of efficiency at your nominal operating point. Of course if you are going to use a valve this will just make things worse.
4) Even in high static head systems a VFD can save you energy although it is more difficult to save energy in these systems.
5) A reasonable savings calculator is available which takes into account all the issues that you have raised regarding static head, VFD efficiency, etc.

If you care to see more propaganda regarding VFDs saving energy in pumping systems see the study at

http://www1.eere.energy.gov/industry/bestpractices/pdfs/variable_speed_pumping.pdf

I will post another paper regarding wells, pumps, etc. if I can get it onto Marke’s web site which he has graciously allowed me to use.

I don’t know if you took a look at the link that I gave regarding SCE’s demonstration pumping facility. If you are really willing to travel to it I will try to setup a test. If not I don’t want to waste my time. This will most likely be my last post in response to you because I have better things to do than try to prove the obvious.

 

[gepman]

The pages in my attachment above are not in the correct order since I uploaded the file before I saved the changes in page order. The first page (a pump curve) should actually be shifted to be page 6 in order to match my post above.

Sorry.

 

[nightfox1925]

I was informed that putting a throttling valve on the suction side of a centrifugal pump oftenly produces vapor which can harm the pump as well.

GO PLACIDLY, AMIDST THE NOISE AND HASTE-Desiderata

 

[gepman]

See the following link for a very good brochure on pump testing and wells. There is also some information on hydraulics and pumps in general. It was written so that a farmer could understand it but it is technically accurate also. Thanks very much to Marke for hosting the site containing this brochure.

http://www.lmphotonics.com/gepman/Edison Pumping Guide 12-27.pdf

Valvecrazy

As you can see you need to accurately measure the water level in the well, flow, pressure, and power in order to calculate the efficiency of a pump (and the effect of using a VFD versus a valve).

 

[unclebob]

For those who are interested in Mitigators Transformers,
here's what I found in my old books.

Using the Symetrical Components, one can show that multiplying the secondaries is a way to eliminate harmonics.

I don't have any references about these type of transformers.

I suspect the cost to build one of these may be... ah... out of range.

 

[waross]

Thanks unclebob.
Diagram "B" in your posts is the vector representation of the "mitigators" that I was trying to describe. I saw two versions.
In the first version the zig-zag was about 10%. In the second version the zig-zag was about 50%.
Respectfully