CT for metering revenue 1

[odlanor]

- CT data
ratio: 8000 / 5 A - class: 0.3% burden: - 50VA

- Calculations:
Irated: 6350 A - total load of CT circuit: 10.13 VA

standard:
With 10% of 8000 there is no guarantee of accuracy of 0.3% to 50VA burden.

Questions:
1)For a circuit load of 10VA error CT will be greater?
How much?

2) Increasing the load circuit by adding resistance, will be improvement?

 

[scottf]

Less burden means better accuracy performance in general.

Also, by the standard for 0.3 class:

100%Inom -> rating factor = guaranteed 0.3% performance

10%Inom -> 100%Inom = guaranteed 0.6% performance

If you have the factory test results, you should have accuracy performance at 0 and full burden at 10%, 100%, and RF*Inom. Since accuracy moves linearly/proportionally with burden, you can plot the actual accuracy performance with 10VA burden at 10%Inom.

I would bet you'll find it will perform within 0.3% performance limits at that point and with that burden.

 

[LSpark]

Adding burden resistors to circutis is bad news, it increases the risk of an open circuit.

If you've got the test certificate you should have the ratio and phase errors at a minimum of two different points (usually 25% and 100% rated burden on the ones I see in the UK). As Scottf said the change in accuracy with burden is pretty much linear so you can calculate the error at the normal working point.

The tariff meters I use can have CT and VT errors programmed into them so it's no so important to get the CT operating at the lowest error point. I use the normal operating point (rated volts and full load current for a generator) to determine the CT and VT errors.

 

[odlanor]

scottf (Electrical)
for metering CT I guess by IEC standards is:
10%Inom < Ict < 100%Inom -> rating factor = guaranteed 0.3% performance.
Ict < 10%Inom = nao guaranteed 0.3% performance.

LSpark
Can you send information of tariff meters with CT and VT errors programmed ?

 

[scottf]

Odlanor-

First, the 0.3 accuracy class is not an IEC class, it's an IEEE/CSA (and a few others) accuracy class. So you need to check to which standard the unit was manufactured to, it does make a difference in performance guarantees.

With that said, I seriously doubt the unit you have which is marked as 0.3-50VA is guaranteed/tested to have 0.3% performance at 10% rated current. The ranges I listed above are per IEEEC57.13.

Attached are the standard IEC metering accuracy classes. You'll notice that as the current is below 100% Inom, the guaranteed accuracy performance drops off. For IEC, the accuracy listed must be met from 25%-100% rated burden. My company happens to always design to meet accuracy at 0VA wherever possible.

Lspark-

Most meters now contain instrument transformer correction tools, however, very few utilities that know actually use them. Just curious, do you calculate/measure the actual connected burden in order to figure out what correction to use? Also, I've been privy to many situation where correction is just made at the expected nominal operating current and end up making the install LESS accurate across the entire operating range based on incorrect correction factors. For example, many folks using correction factors use the data from the test reports at the full burden rating. In reality, most metering installs have much less than the rated burden connected to the CT. Since CT error is almost always negative and becomes more negative as the burden increases, correcting at the full rated burden can incorrectly cause higher Kwh billed than is actually consumed.


Some meters do offer CT correct factors at a number of different current points and then do linear extrapolation between the various points.

 

[LSpark]

Odlanor - all the high accuracy meters approved for use in the UK can have single point CT and VT errors programmed in. I use CEWE prometer R meters but there are other meters out there. I know Landis and Gyr have one approved, the other manufactures names evade me - I haven't had my morning coffee yet! One of my contractors raves (in a good way) about a Canadian manufactured meter.

Scottf - I do calculate and/or measure the CT and VT burdens, these days they're normally not much above 0VA unless it's a long run between the CTs and meter panel. Most of the circuits I look after sit between 50 and 75% meter rating when they're on so the corrections do improve the accuracy. The newer sites seem to have CTs with fairly stable errors at a particular burden, the %current has a much smaller effect on error. I look after generation sites mainly.

Most of my fellow metering engineers look slightly blank when I mention CT and VT error correction and totally blank when I mention power transformer loss correction. They prefer to use complex models to produce P and P*P factors rather than simple error correction in the meter.

 

[RRaghunath]

Adding resistance (additional burden) in the metering CT circuit in order to bring the connected burden to >25% rated for measurement accuracy reasons is done some times.
This also helps in improving the instrument safety by lowering the effective Fs.
For a new installation, it is better to specify CT with low burden rating as the present day digital meters hardly impose any burden.

 

[odlanor]

raghun,
do you have any reference about that?

 

[odlanor]

We have a hydraulic power plant with 2*150MVA-13,8kV generator sending energy in 500kV at a substation 120km distance.
energy sending => 68,871796 MWh
energy receiving => 69,172963 MWh
How can you receive more than you sent?

 

[RRaghunath]

Are we seeing Ferranti effect here! Receiving end voltage is supposed to be higher (than the sending end one) when the EHV lines are lightly loaded. The line generates VARs but how can it show up as active energy!!
The discrepancy in the records is most likely due to voltage drop in the VT secondary wiring (and hence lower voltage available at the meter terminals) at power plant end.

 

[odlanor]

raghun,
how to minimize this problem?

 

[odlanor]

ragun,
...The discrepancy in the records is most likely due to voltage drop in the VT secondary wiring (and hence lower voltage available at the meter terminals) at power plant end...

Do you mean the powerPlant-meter is registering less than it should register?

 

[jghrist]

energy sending => 68,871796 MWh
energy receiving => 69,172963 MWh

energy sending + 0.3% error => 69,078,411 MWh
energy receiving - 0.3% error => 68,965,444 MWh

 

[LSpark]

Voltage drop in the secondary wiring should only be a problem if you've got an appreciable amount of burden connected. Modern meters with a seperate power supply are very low burden, so it is much less of a problem on modern installations (in the UK anyway).

I've still got some older meters and voltage imbalance relays that are powered from the VT so I include the voltage drop in the VT error that I program into the meter. Some circuits have additional equipment connected so this can increase the burden too.

 

[odlanor]

jghrist,
what about loss of stepup transformer 13,8/500kV and loss of 500kV LT?

 

[jghrist]

what about loss of stepup transformer 13,8/500kV and loss of 500kV LT?

I haven't tried to account for any losses, I just wanted to point out that because of metering and CT errors, you can't determine losses with any degree of accuracy by measuring load in vs load out, and might get negative numbers. Is it possible that one of the energy meters includes loss compensation for the transformer? This would not be uncommon. If so, there might be an error in the loss compensation settings.

 

[scottf]

The chance for difference in MWh readings is much likely the result of CT errors/difference than VT secondary voltage drop (which is almost never an issue for modern applications as mentioned above).

What are the accuracy ratings and ratios of the CTs on either side. It's quite common to have CTs sized incorrectly (ratio too high) and the higher in voltage rating, the worse the sizing seems to get.

The story I normally tell to illustrate this is I once sold 30 pcs of 765 kV CTs to a US utility. The metering core was rated 5000:5A, 0.3B1.8, RF1.0. I pleaded with the utility to at least consider 2500:5A RF2.0, but they wouldn't budget. The meters were located about 800 feet away from the CTs, so they were using every bit of the burden rating using #10AWG. This was a major interchange billing point and millions of $'s were being billed across these CT every month. They paid >$100,000 for each CT and were basically probably getting actual CT errors of roughly 0.5% or so at the nominal current level. 0.5% on $10MM per month is $50k per month lost in just the CTs being sized wrong.

 

[odlanor]

CT at 13,8kV side:
ratio: 8000 / 5 A - class: 0.3% burden: 50VA
Calculations:
Iusual: 5600 A - total load of CT circuit: 10.13 VA

CT at 500kV side (remote):
ratio: 2200 / 5 A - class: 0.3% burden: - 50VA
Calculations:
Iusual: 156 A - total load of CT circuit: 10.13 VA

scottf ,raghun
How do you suggest to minimize this problem now?

 

[scottf]

odlanor-

Are you sure the connected burden on each CT is exactly the same? Hard to believe the meter is located exactly the same distance from the CT for each installation.

Is your "problem" just that the recorded MWh differ at each end by about 1%?

If that's your problem, then you need to look at your transformer and line losses first to make sure those readings aren't indeed accurate.

After that, it seems you might want to look at using a lower ratio CT on the 500 kV side and/or buy a CT with a higher accuracy/wider range, something like a 0.15 or 0.l5S class (IEEE) or 0.2S class (IEC).

 

[LSpark]

You need to consider the maximum operational current (not fault current) that flows through the circuit not the usual current when sizing a metering CT. So whilst from the given info it looks like the remote end CT is badly sized that may not be the case if additional circuits sometimes flow through that meter point.

In the UK the metering CTs and VTs should be located at the commercial boundary for all new installations (i'd have thought this was the same the world over) but this isn't possible for some of the older power stations that were build prior to privatisation. In this case we apply a correction for power transformer loss (load and no-load seperately) in the meter. The load loss usually has to be adjusted for the meter full load current rather than the transformer current and is put in as a percent rather than kW.

In your case I'd want to look at the programming of the meter to see if this has been programmed and if so what the values are. I'd also want to see copies of the CT and VT test certificates and any compensation figures for those in teh meters.

Scottf - on the last new build I had to get the metering CT ratios changed (multi ratio CTs fortunately) as they'd used the same ratio that they'd used for the protection.