Substation Voltage Regulator - question on source and load amps 2

[Manindemand]

Below is an email I sent to engineering after changing out a bad MW Transducer in one of our stations. I got this data off our server (TOIS) that night via VPN. The XFMR is 30 MVA (161kv/13.8kv) and Regulator is 3333 KVA 10% Raise and Lower. Were all kind of scratching our heads thinking... is this right? This is critical because in August the load was showing 28 MVA in SCADA. We're trying to justify installing a 2nd XFMR.

Question: When the regulator is on 16 lower it lowers the voltage but does it raise the amps on the load side of the regulator and lower the amps on the source side?


"I took these values out of TOIS. It looks like the readings are correct now.

01-OCT-2003 19:15:00 (13.8 Bus kv) 7.8 (Feeder Amps) 157.99 + 118.92 + 149.2 + 167.76 + 159.95 + 147.25 + 168.25 + 164.83 + 184.37 = (Total Amps) 1418

(T1 kv) 8.59 (MVA) -12.09 (MVA1) 10.16 (MW) -11.16 (MVAR) -4.63

The regulator is probably on about 13 lower. Keep in mind that the total current flowing through the metering CT's (in the XFMR) is also the total combined current flowing out of the feeders. However, the regulator has dropped the kv from 8.59 down to 7.8. The MW and MVAR xdcrs are getting their voltage from the 8.59 kv source. To calculate MVA using feeder amps (to compare to SCADA), you have to use 8.59 kv instead of 7.8 kv.

1418 x 8590 = (MVA) 12.18 That's pretty close to the reading in TOIS at 19:15 this evening. Keep in mind that the T1 kv reading may come from the highest of the three T1 kvs.

The point I'm making is, it might look like the MVA reading is too high if you use 7.8 kv value - (1418 x 7800 = 11.06 MVA). On Aug. 17 the load was high enough that the T1 and Bus kvs were close so this wasn't as much of a factor."

 

[jbartos]

Suggestion to the previous posting: If there were some modifications made as stated below:
""The reason the regulator is running at 13 lower is because the XFMR is on the lowest tap- "E Tap" -which makes the secondary voltage too high. It ran near neutral until they added another section of transmission line. Now the high line is running a lot higher.""
Is not there a chance that the SCADA is sensing voltage or current or both from incorrect locations or instruments?
12.09/11.06=1.093, which may be just a matter of improper connections. Have instruments been modified or wired during the transmission line addition?

 

[Manindemand]

alehman - yes, three 3 phase feeders.

You gentlemen are saying "... you cannot sum scalar current values. You must sum the feeders vectorially (e.g. with a summing current transformer)."

This is the first time I've heard this. Thank you. Can you give any links so I can study this?

What's more confusing is that other technicians and managers use the same calculation that I've been using. I can connect to our server and check 10 different stations comparing calc MVA with SCADA MVA and they will always be very close.

But like I said, this station is unusual due to the XFMR on lowest tap.

Thanks again.

 

[Manindemand]

jbartos - excuse my thick skull, good idea! I'll take a Poly Meter with me and check phase angles to verify all my potentials are connected to the right terminals on the Watt and VAR transducers. Like I said, I verified current circuits but not potentials.

 

[jghrist]

Both of your formulas are mathematically identical and correct as long as you use the right current, which should be the magnitude of the vectorial sum of each phase current, and all phase voltages are equal. If all loads have the same power factor, the vectorial sum will be the same as the arithmetic sum. I don't know of any links to show that a vectorial sum is needed, but the transformer measurement is of the sum of the feeder currents, and the currents are vectors with different angles so they have to be summed vectorially.

To see the possible difference, assume that the first three feeder currents given in your first post are phase A for three separate feeders, that phase A voltage angle is zero, and the the power factors are 90%, 70%, and 60% for the 157.99A, 118.92A, and 149.2A currents respectively.

Current angles are acos(pf): -25.84°, -45.57°, -53.13°
Real currents are I·pf: 142.19A, 83.24A, 89.52A
Imaginary currents are I·sin(angle): -68.87A, -84.92, -119.36
Sum of real currents = 314.95A
Sum of imaginary currents = 273.15A
Magnitude of vectorial sum = sqrt(314.95²+273.15²) = 416.9A
Arithmetic sum of current magnitudes = 426.11A


Not a whole lot of difference even for this wide range of power factors. Actual power factors are undoubtedly much closer, which is why your simplified formulas work pretty closely.

 

[stevenal]

"Is it possible that we really do have 12.09 MVA on the source side and 11.06 MVA on the load side?"

I don't think so. Apparent power is a conserved quantity, so you would have a 1.03 MVA loss in your regulator. Sounds rather extreme.

If you're checking individual CTs and PTs, don't forget you have a single phase set in your regulator. The CT will reflect the phasor sum of the feeder quantities on that phase. I'm assuming regulator uses line to ground quantities, sometimes line to line quantities are used. Check the nameplate. Also, instrument transformer ratios used may seem a little strange. 1000/0.2 CT ratio for example.

I agree with jghrist's post following mine, arithmetic addition just makes the discrepancy more confusing. If power factors are similar on the feeders, a simple arithmentic sum may provide a fair approximation. Kinda sorry I muddied the waters with that post.

 

[Manindemand]

The results are in and the case is closed. All phase angles were correct. Below is the email of my findings that I sent to engineering. Comments are welcome and thanks again for the help.

"I think I found the cause for the 1.03 MVA difference between SCADA and calculated. It turns out there was a combination of small problems. I didn’t notice them during my original check. Pardon all the details, I wanted a good record of this to put in my notes for future reference.

The MW and MVAR xdcrs are accurate but the “zero” trim potentiometer on the Analog Control board (in the RTU) was off and made each reading approx .25 higher than actual (.25 mw high & .25 mvar high). I adjusted it.

The Bus kv transducer was giving a reading that was 45 volts low ( DOC read 7802 but it was really 7847). The xdcr was slightly off and I adjusted it.

While troubleshooting, I also noticed the RTU cards and card cage were covered with dust. The card headers and edge connectors also had corrosion buildup. I gave all parts of the RTU a thorough cleaning. This type of problem commonly causes inaccurate analog readings but I’m not sure how much effect it had in this case.

I finished all the repairs about 16:45 on 10-9-03. Since then, I’ve checked MVA readings a couple of times and they match the calculated value: FEEDER AMPS x BUS KV = SCADA MVA.

I believe I’ve also solved the question of how much MVA drop we have across the regulator. I read the three voltage inputs and the three amp inputs to the MW xdcr and calculated 9.924 MVA. I read all the feeder amps (and again compared to DOC - ok) and bus kv and calculated 9.542 MVA. This was with the regulator on 11 lower. Below are the values:

Watt xdcr - Amps = 1.66 + 1.64 + 1.58 = 4.88 x 240 = 1171
Volts = (121.2 + 121.2 + 120.8) / 3 =121 x 70 = 8475
1171 x 8475 = 9.924 MVA

Feeder amps x Bus kv - 1216 amps x 7847 volts = 9.542 MVA.

So, .382 MVA drop is more realistic (at 11 lower and with these load conditions). It does show that you can have higher MVA on the source side of the regulator. I’ll have to wait until load is back up again to find out what the difference is during heavy load. I suspect stations with regulators running near neutral will have very little difference.


BOTTOM LINE - we need to go by the test results from the bad MW xdcr. Those results show we were would have read 75 % of the true MW value (when true MWs were between 23 and 30). The peak MWs reading in TOIS on 8-17-03 was 20.36 but the actual MWs were 27.15. With MVARs at 9.39 we would have had 28.73 MVA. I can’t say how much effect the “zero” trim pot and the dirt/corrosion in the RTU had but it does appear we’re getting close to max load on the transformer. I feel confident that all the analog readings are now accurate.

SIDE NOTE - my original estimate that the actual MVA was 25.3 on 8-17-03 must have been off. The total feeder amps in SCADA were 4.24 % too high. I assumed 6 % after bench testing two different model amp transducers. I dialed in a set 5 % error at 1 amp then plotted the error through 5 amps. I assumed the xdcrs at ********** would have given similar readings at approx 27 MVA. It looks like I was off a hair - my bad.

Unless I can be of some help, I'm considering this "case closed"."

 

[stevenal]

Thanks for the wrap up. Too many threads simply die away without any of us knowing the final result.