Harmonics in Grid - How to measure 1

[RRaghunath]

Measurement of harmonics is different from measuring the fundamental frequency currents and voltages.
Does it require any specially ordered CTs and PTs!

 

[scottf]

It depends how high of frequencies you want to measure accurately.

For CTs, they're generally pretty good up to 25th to 32nd harmonic.

For VTs, oil-filled VTs are normally decent up to about 1 kHz.

 

[RRaghunath]

Thanks, scottf.
We are supposed to be measuring up to 50th harmonic for computing the THD in 132kV grid.

 

[marks1080]

Is there any particular reason you are going that high? Typically you wouldn't look past 29th.

 

[MAGTiger]

Interesting project. I wouldn't expect much distortion at 132kv. So unless you are on an unusually distorted system, then you are going to have a resolution problem. The harmonics should be much smaller than the total rms but the CT & pt need to be sized form the total rms.

Neil

 

[RRaghunath]

Huge compressors with motor ratings of the order of 27MW are expected in the system. These will be connected to 132kV bus bars through transformers and 12-pulse VFDs. Since, the point of common coupling is 132kV we should be able to measure the harmonics to provide base values for designing the harmonic filter asssociated with the VFD.

 

[Skogsgurra]

This is definitely something you should talk to a specialist company about. A few thoughts, though:

If the measurements are to be used for THD calculation, I do not think that there is any practical reason to go beyond 41st harmonic. The contriubtion from higher harmonics will most likely be neglible or 'absorbed' in the total error budget.

Also, There is at least one transformer between the VFDs and the PCC. That transformer usually has a worse frequency response than the VTs and CTs that you use on the 132 kV side. So, the small contributions from 41+ harmonics will already be reduced when they reach the PCC.

And, finally, filters for 12 pulse loads are seldom designed for components higher than 35/37, and that is also quite unusual. Harmonics 23/25 is, in my experience, the highest one would need to filter. But, we do not know all circumstances. Are there good reasons to take the measurements and mitigation methods any higher?

Gunnar Englund

http://www.gke.org

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

 

[RRaghunath]

As scottf mentioned, the electromagnetic CT/VT also will be good enough only up to a frequency of 1kHz or 20th harmonic.
I wonder what methods are adopted in Industry when you need to measure harmonics beyond 20th order.
Could be that CT / VT are first calibrated at higher frequencies and the measured errors are accounted for, while actual meaurements are done at site!!

 

[Skogsgurra]

raghun,

Capacitive VTs are used when better frequency response is needed.

Gunnar Englund

http://www.gke.org

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

 

[Skogsgurra]

Fibre optic transducers based on magnetic field rotating polarisation plane (Faraday effect) is used in HVDC and will give you a very good CT with very good frequency response.

Gunnar Englund

http://www.gke.org

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

 

[scottf]

skogsgurra-

If you're speaking of capacitor voltage transformers (i.e. CVTs, CCVTs) they are the worst option for harmonic measurements, as they are specifically tuned for accuracy at the fundamental frequency.

There are devices that can be added to CCVTs to facilitate the measurement of harmonics, but they are relatively rare and somewhat new on the market.

Resistive or capacitive dividers are the most common method, but they are somewhat unique for transmission level voltages and can be rather expensive.

Optical CTs and VTs due offer very good frequency response. They are normally expensive and often have reliability issues.

Best way, in my option, is to use a conventional CT that had been tested for frequency response and a CCVT fitted with an added-on auxiliary device for harmonic measurements.

 

[Skogsgurra]

I meant capacitive voltage dividers. I call them transformers. Probably wrong. But that is what ASEA (ABB) used to call them many years ago.

Gunnar Englund

http://www.gke.org

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

 

[RRaghunath]

scottf,
Thanks for summarising. But, you mentioned that the CVTs with added-in auxiliary devices are "relatively rare and some what new". Thta is my understanding too. So, any reason (could be some experiences etc.) why you feel that is the best way to go.
How do you feel about using Electromagnetic VTs!
These are OK in our system as coupling capacitors are not required (Bus VTs) and when it comes to cost, EM VTs are not costlier than CVTs for 132kV voltage level.

 

[scottf]

As said previously, inductive-type VTs normally have a very poor frequency response above 1 kHz. The performance varies substantially by voltage class and manufacturer.

As for the aux-device for CCVTs, they are available now, just haven't been much in the past, so it can be assumed a CCVT installed in the field is not suitable to PQ measurements.

This is a link to such a device. When AREVA purchased the high-voltage designs from Ritz, they chose not to keep this offering. I originally wrote this brochure when at Ritz Waynesboro, which was a few yrs ago, so I'm not sure the current status of the offering. I would suggest contacting

http://www.pqs-consulting.com/

for questions on this technology. I know we did install some at US utilities in 2006/2007 and they worked well.

http://www.pqs-consulting.com/ritzbrochure.pdf

 

[RRaghunath]

Thanks, scottf once again. PQS Consulting is new to me and glad to know that the PQ sensor is their patent.
With PQ sensor, the accuracy of measurement for harmonics upto 20kHz is possible which is way beyond the codal requirement of 50th order harmonic.
Still, the question about current harmonic measurement remains. This will have to be done using the conventional electromagnetic CTs only and I guess the deterioration in accuracy at higher frequencies may not be as bad as that with EMVTs.

 

[ScottyUK]

scottf,

Is anyone manufacturing Rogowski coils for HV applications? In principle they are ideal in terms of bandwidth, I'm just not sure which manufacturers are in the market at transmission and primary distribution voltage. We've used them at generation voltages with great success.


----------------------------------

If we learn from our mistakes I'm getting a great education!

 

[scottf]

raghun-

The frequency response for CTs varies quite a bit, but in general, they're much better than for VTs, mainly because the insulation capacitance inter-winding capacitances really don't come into play like in a VT.

When I was with AREVA HV IT's, we did some testing on the Extended-Range Current Transformer design and it was very good up to 32nd harmonic. Don't remember how it performed above that. You could contact AREVA in Waynesboro, GA USA for more info (706-554-8800 and ask for the engineering manager).

ScottyUK-

There is always a lot of talk about Rogowski coils in HV and MV, but very little, if any, action. I think they're used more in Europe than in the US. In principle, the design of an HV CT or MV CT is the same with a normal core/coil or a Rogowski coil. The problem really lies with the design of the RC, as it's much more difficult to calculate the performance, as compared to a traditional core/coil and almost has to be manufactured using a trial and error process.

In my opinion, for HV applications, if you can't get the harmonic performance in a conventional CT that you need, then optical CTs probably are a better solution over Rogowski coils or linear couplers.

 

[RRaghunath]

Thanks a lot for throwing light on this difficult subject, scottf.

 

[ScottyUK]

Thanks scottf,

I guess that's why we never seem to hear much about them in HV. In the 11 - 25kV range we usually have so much distance around things like generator main terminals that air insulated Rogowski coils work just fine but I can imagine, though probably not fully understand, the difficulties in installing one in a conventional HV CT design.


----------------------------------

If we learn from our mistakes I'm getting a great education!