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

 

[dpc]

We used to do basically the same thing with drive isolation transformers for multiple pump installations. Order with delta-wye transformers with opposite 30 degree phase shift and or use combination of delta-delta and delta-wye. It works well.

Once we stopped isolation transformers on every drive we got away from this, I guess.

 

[gepman]

If you want to fix harmonics in an existing system you can us an active harmonic filter similar to an Accusine.

http://www.squared.com/us/products/...F65D385256C7D006D6717/$file/activepfcpage.htm

The active filters work but designing a solution into the system from the beginning like you suggest is probably better if you can convince the client to pay for it up front. If you need to fix it afterwards this is less expensive.

 

[sreid]

A question related to this thread. Are drive manufacturers using switching patterns to minimize harmonic distortion?

 

[Skogsgurra]

Harmonic distorsion is a matter between mains and the input section of the inverter. It is usually a passive rectifier, so switching pattern does not have any influence.

But, there are inverters with an active front end (AFE). They use power factor correcting techniques (PFC, not power factor compensation) to draw a perfect sine shaped current. There, of course, the switching pattern of the input system is optimized for lowest possible harmonics.

Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[ozmosis]

Mark
Your proposal is an economic method of harmonic mitigation but being able to guarantee equal loading with pumps is not so easy, and if the loading is off on one leg, then problems within the transformer windings will ensue.
Active filters should not always be seen as a problem solver after the event. If you know the problem harmonics before installing the drives, then designing an active filter solution isn't always as expensive as you think as you simply design the AF for the harmonic load you are looking at unlike an AFE that has to be designed for the FLC of the drive connected.
Hey guess what, we've just launched a brand new range of AF....
I'll be in NZ late Feb if you want to catch up Mark!

 

[Marke]

The current situation is a rural area with 22KV overhead lines. Each pump has it's own transformer 22KV/400V. There are around 2000 pumps in this area of which probably half are controlled by VFDs. The harmonic distortion on the supply is greater than 7% and there are more pumps being installed all the time. The pumps all run almost continuously from October through March so balancing would not be an issue.
All pumps are three phase.
Active filters are a problem because we would need to add one to each installation, or use a large transformer to step down to 400V and install one large unit to correct the main supply.
My thought was to install all new pumps with a lead or a lag phase shift to help to a) cancel the effects of harmonics on new installations and to reduce the effects of existi9ng installations.
The existing transformers used are all delta to star. An advantage of this configuration is that it cancels the triplen harmonics. We could use star to star transformers on new installations, giving the effect of a 12 pulse input, but we would get much higher 3rd harmonic on the 22KV lines from this configuration.
We could also use active front ends for all new drives, but this would have no effect on the existing harmonics. and would drive up the price.

Best regards,

Mark Empson

http://www.lmphotonics.com

 

[Skogsgurra]

Is that THDV or THDI? If volts, then it is on the high side, but still not too bad. If it is amperes, then I would just laugh at it.

But, knowing you, I suspect it is THDV. The question is then if it is on the 400 V side or if it is also on the 22 kV side. If on the 22 kV side, then one should definitely not laugh at it.

Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[Marke]

Hi Gunnar

It is THDV and on the 22KV side. There are some concerned faces.
Because the load is increasing, it is preferable that we specify the additional load to reduce rather than increase the THDV.

Have a good day,

Mark Empson

http://www.lmphotonics.com

 

[ozmosis]

wow, 7% THVD on the primary, that is a problem. What is it like on the secondary? I'm surprised the network operator allows you to connect at all. If this was the UK using G5/4, they would enforce that certainly for the future but also retrospectively for all the existing issues.
I'll send you some info offline that might be of use.

 

[Skogsgurra]

Just one more question. You did measure via VT? And it was connected line-line?

I once had a panic job where someone had measured THD using a VT connected from line to ground. That produces lots of "imaginary harmonics" on a non-grounded 22 kV line.



Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[Valvecrazy]

OK, since this is a pump application, I have to throw in my two cents worth. You say that of the 2,000 pumps, about ½ are running on VFD controls. I can assure you that the other ½ of the pumps that are not running on VFD controls, are not causing harmonic problems. And since the pumps run continuously, there is no problem with high currents from multiple starts. Therefore, it stands to reason that if you do not use Drives to control the pumps, you also will not have the problems with harmonics.

I have seen this many times. The electronics or drive experts are left to try and solve problems that should not have been present to start with. By knowing how to size and spec the right pumps, and by using mechanical means of varying the flow if necessary, there are NO energy savings by using VFDs. Since there are no energy savings on most pump applications by using a VFD, the addition of a VFD just complicates the issue by adding problems like harmonics, bearing currents, resonance frequencies, and voltage spikes to the motor. VFD controls also make the supply of water less dependable than if the pumps were controlled by simple mechanical means.

The more VFD’s added to a system, the worse the problems with harmonics. There are many other good uses for Drives such as machine tools, conveyor belt systems, and positive displacement pumps. If you do not use Drives on centrifugal pumps, that could be just as easily and as efficiently controlled by ATL or Soft Start only, then in this case you would have 2,000 less VFD’s adding increasing amounts of harmonics to the system, and 2,000 less problems. It would stand to reason that if you do not cause a problem by using a VFD, then you do not have to find a complicated and expensive way to solve the problems later.

I do not understand why there are so many intelligent people, like the ones on this thread, who understand all the complications of electronics in use today, and so few people who understand pumps well enough to know that VFD’s do not save energy, and are usually more trouble than they are worth.

I know there are a lot of you who would like a VFD on every machine for your own job security. However, the first indication of real knowledge about Drives, should be when and when they are not the best control for the particular application.

Many end users are having the carrot dangled in front of them claiming that the VFD can save 20% to 50% in energy. Many articles and papers are being written on how this hospital or that high rise building, has saved tremendous energy by switching pumps to VFD controls. If you read these articles carefully, you will see that the energy savings actually came from downsizing the pumps, eliminating dump valves, and cutting waste, rather than the implementation of a Drive. Anytime you vary the speed of a pump, you are using more energy per gallon than if the pump were running at it’s designed BEP. Knowing this, how can anybody say varying the speed of a centrifugal pump with a VFD can save energy?

The use of filters, extra grounding, zig zag transformers, six phase rectifiers, and countless other band aids, would not be needed if you did not install 2,000 harmonic generators (or VFD’s) to a system that could have more easily and effectively be controlled by other means.

 

[Marke]

Hello Valvecrazy

By knowing how to size and spec the right pumps, and by using mechanical means of varying the flow if necessary, there are NO energy savings by using VFDs.

In a constant flow situation, I would agree with you, but in a variable flow situation, I can not agree with you and there is sufficient evidence to show that using variable speed does save considerable energy on pumps operating under variable flow situations. For example, a farmer has three irrigators of different sizes fed from one pump. The flow requirement varies depending on which irrigators he has on. One solution is to use correctly sized pump per irrigator, but that means three wells, three pumps etc and the single solution with a VFD is more attractive.
If you reduce flow by throttling the pump, using a valve, then the pump operates at a much lower efficiency and thus wastes energy, so using valves or similar to reduce the flow is not efficient.
If you have a better option, I would be very interested to hear it.

Gunnar, the harmonic voltages have been measured on the output of a three phase transformer.

Ozmosis, the THDV level kind of crept up on the network operator, they knew it was not good, but did not realize how bad it had got.

Best regards,

Mark Empson

http://www.lmphotonics.com

 

[Valvecrazy]

As someone said in an earlier thread, the well depth plus the pressure required by the sprinkler sets the Minimum Possible Head or TDH for the pump. This also limits the minimum speed that the pump can spin that will still produce the head required. In the case of the 100 HP and 1200 GPM pump in the attached curve, minium speed is only 3290 RPM. The power required still falls off further as the flow required decreases, the same as it does when the pump is running at a constant of 3550 RPM. Notice at 100 GPM flow the variable speed curve shows this pump dropped from 100 HP to 38 HP. However, the power required by the full speed pump dropped from a100 HP to 42 HP simply by restricting the flow to 100 GPM. This shows only 4% difference at low flow between VFD and NO VFD. Add back in the parasitic losses of the Drive and reduced efficiency of the motor compared to running the motor on Across The Line, and there is no difference in power consumption of VFD or NO VFD.

Moving the sweet spot of the curve and maintaining maximum efficiency is just VFD propaganda, when the electric meter is still spinning at the same rate regardless. It is a common misconception that a VFD can slow a properly sized pump down enough to save energy.

Using your example of the variable flow situation, 1200 GPM, 800 GPM, or 400 GPM, the two curves follow each other so closely that all things considered, restricting with a valve reduces energy consumption as much as varying the speed.

The reason the problem crept up on the operator is because the number of VFD's continues to increase. Changing to VFD's on pump applications does not decrease energy consumption, it increases the THDV level and causes other problems.

 

[Skogsgurra]

This is very interesting!

I do not have the time to do my own comparison, but you seem to have a valid case here. Anyone done the same comparison with same or different results?

Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[Marke]

Hello Valvecrazy

That curve is what we refer to as a flat curve pump and they are difficult to get good results with VFDs because the curve of flow against pressure is essentially flat across quite a range of flow, but if we look at that pump and consider the situations that you nominated with a flow of 1200GPM and 100GPM. If we assume that the pressure is relatively constant, then there is a lot of energy that is being wasted in the pump when it is throttled back to 100GPM.
Hydraulic power is the product of flow, pressure and a constant K.
Taking the case of 1200GPM and 230ft, we have a hydraulic power of 1200 x 230 x K HP = 276000 x K HP.
If we assume that the pump efficiency is 75% at this point on the curve, then the value for K will be about 0.00028

Now look at the case of 100GPM. The head is about 270, so the hydraulic power is 100 x 270 x 0.00028 = 7.56HP
The absorbed power at this flow is around 40HP, so the losses in the pump are 32.5HP (24KW) The efficiency of the pump is 18.9%

If you were able to slow the pump down and stil have sufficient head, then the efficiency of the pump would rise towards 75% and you would save nearly 30HP (22.3KW)

In the case of the pump curves that you have suggested, then it will be difficult to operate the the pump at a head of around 240 ft and achieve an energy saving with a VFD. If however, we used this same pump at a head of say 120 ft, then we could us a VFD to achieve an energy saving at 100GPM.

Most pumps used in this region for irrigation are submersible and have a very definite slope on the flow/head curve and these definitely allow the VFD to slow the pump down to achieve appreciable energy savings.

So, yes, I agree, with the pump used in your example, a VFD would not achieve energy savings under the nominated conditions, but there is still considerable loss when the flow is reduced.
A different pump with the same operating conditions would enable the pump to achieve the required head at reduced flow and speed, and the operating efficiency of the pump is much enhanced.

Best regards,

Mark Empson

http://www.lmphotonics.com

 

[Marke]

This curve would be more typical of the type of curves that are commonly found on submersible pumps.

http://www.lmphotonics.com/images/pump1.htm

At low flow, the power lost in the pump when it is throttled, is higher than a flat curve pump because the head is much higher, but it is able to be controlled with a VFD to gain a significant improvement in efficiency.

Best regards,

Mark Empson

http://www.lmphotonics.com

 

[waross]

Mark;
I ran across a line of transformers about a year ago that may be of interest.
The transformers were called "Mitigators".
They were three phase dry types in the range of 50KVA to 100KVA.
There were two types installed.
The first type had three windings on each leg of the secondary. The first winding was about 80% and the second and third winding were each about 20%.
The connections were: A Phase, the 80% winding plus a 20% winding from B phase and a 20% winding from C phase.
The configuration was repeated for the other phases.
The second type was similar except that the main windings were about 60% and the second and third windings were about 40%.
At the time I was not able to get any information from the contractor doing the installation. I spent some time Googleing with no joy.
I have again spent over an hour googleing with no positive results. Possibly this technique may be of use to you and possibly someone will recognize the transformers and be able to give us more information.
On pumps, I am on the fence here. I suspect that the answer may be, "It depends!"
I haven't had to check pump curves for a few years but as I remember the pumps that I was concerned with, the motor loading and the flow dropped off as the head increased. However I would not assume that this applied to all pumps.
I believe that there are instances where a VFD is beneficial and instances where the advantage of a VFD is marginal.
It depends!
Respectfully

 

[edison123]

Very informative discussion. LPS for all.

* I would go green if only I were not yellow *

 

[LionelHutz]

I have a question. How much effect will using separate transformers to create the 12-pulse have? Any time I have seen that done a single 3-winding transformer was used and the harmonics canceled in the transformer. Will the harmonics cancel in the high voltage line between a Y-Y and a Y-Delta transformer?

There's a company in Canada called Mirus that makes a line filter called a Lineator. It is a 3-lead in/3-lead out filter that installs like a line reactor. That thing works amazingly well for reducing the harmonics of a drive. Unfortunately, Mirus knows they do and asks a premium for them. Regardless, these would be a very reasonable solution to use due to the ease of installation and the fact that they would cancel the harmonics at the source.