System Voltage Phasing Measurements

[redtrumpet]

I had mentioned in a previous post that I have a client who has several 63 kV - 13.8 kV delta-wye ANSI Group 2 transformers. Some of these transformers have A-B-C sequence on the primary, some have A-C-B. The client is attempting to parallel secondaries. Obviously, there is a problem with paralleling transformers with opposite primary sequences - a 60 degree phase shift is introduced between secondaries (30 lead on the A-C-B + 30 lag on the A-B-C = 60 degrees).

The client has made some changes to cable secondaries to at least get proper rotation when doing a cold transfer, realizing a parallel (hot) transfer is not feasible right now. Voltage measurements were taken across the line and load terminals of two open tie breakers involving three transformers.

Based on visual inspection of the primary connections and the transformer nameplates, the transformers appear to be connected H1-H2-H3 as follows:

T1 - B-C-A primary, 30 degree lag on secondary
T2 - C-B-A primary, 30 degree lead on secondary
T3 - A-B-C primary, 30 degree lag on secondary

For convenience designate the two systems a-b-c and a'-b'-c' read left-to-right at the back of the 13.8 kV switchgear breaker cell, recognizing a and a' may not necessarily correspond to the X1 transformer bushing. Measurements were as follows:

T2-T3
Va-b = 12.94 kV
Vb-c = 12.88 kV
Vc-a = 13.29 kV

Va'-b'= 13.45 kV
Vb'-c' = 13.50 kV
Vc'-a' = 13.85 kV

So far, so good - line-to-line voltages read between phases as expected.

Va-a' = 7.75 kV
Va-b' = 15.60 kV
Va-c' = 7.75 kV

Vb-a' = 7.70 kV
Vb-b' = 7.77 kV
Vb-c' = 15.54 kV

Vc-a' = 15.54 kV
Vc-b' = 7.07 kV
Vc-c' = 7.82 kV

Based on these measurements, the roughly line-to-neutral voltages correspond to a 60 degree shift, and the roughly twice line-to-neutral voltages correspond to 180 degree shifts. From this I inferred that the systems are rotating in the same direction, and that a-a', b-b' and c-c' are 60 degrees apart. This matched my expectation based on the primary connections of T2 and T3. It appears the A and C cable leads of one side of the tie were swapped to correct the secondary rotation.

However, a set of voltages measured across the tie breaker connecting T1 and T3 has me stumped.

T1-T3
Va-b = 12.95 kV
Vb-c = 12.95 kV
Vc-a = 12.87 kV

Va'-b'= 12.78 kV
Vb'-c' = 12.78 kV
Vc'-a' = 12.65 kV

Again, line-to-line voltages read between phases as expected.

Va-a' = 13.22 kV
Va-b' = 7.82 kV
Va-c' = 7.66 kV

Vb-a' = 7.10 kV
Vb-b' = 13.22 kV
Vb-c' = 7.82 kV

Vc-a' = 8.38 kV
Vc-b' = 7.11 kV
Vc-c' = 13.22 kV

Now, it appears that there is a 120 degree shift between phases based on a-a', b-b', and c-c' readings. However, the line-to-neutral voltage readings indicate a 60 degree shift - an impossible situation to resolve in my mind. Based on primary connections, these two systems should both be A-B-C rotation and exhibit either 0 or 120 degree phase shifts.

I was not present during the voltage testing. Am I missing something obvious here, or do the second set of voltage readings appear inconsistent? I want a second opinion before I approach the client about redoing the measurements.

 

[busbar]

Are the transformers delta primary, wye secondary? If I understand the situation correctly, the only fix is changing ‘internal’ secondary interconnects, which is difficult to do in a enclosed 'three-phase' transformer. The voltage in the "low" ø-ø measurements is based 52% of the full ø-ø voltage. The only way to obtain nearly zero-volt differences is through internal changes within each “3ø” transformer.

This misunderstanding had led to some very serious accidents in government facilities. It is reviewed in an electrical-safety manual used by the USDOE in a chapter titled “Phasing Two 3ø Systems Together.”

If you need this text, please reply for availability information. {Please do not ask me for copies.} It is a text that explains the situation fully in a step-by-step manner. Properly done, 15 measurements are needed to phase/parallel two transformers.

 

[redtrumpet]

busbar - yes, I would be interested in the DOE publication, if it is in print and available for order.

We have thought of gauging the feasibility of changing the internal bushing connections on at least one transformer. It may be difficult, but no harder than changing the primary connections given the present substation configurations.

However, my question related to the measured voltages. A comparison of A-B-C/A-C-B primary connected systems (measured between secondaries of two Dy1 transformers) should yield either 60 degree or 180 degree phase shifts. An A-B-C/A-B-C comparison should yield either 0 or 120 degree phase shifts. I can not physically draw a phasor diagram of two three-phase systems with phases displaced 120 degrees in time that will give me both a 60 degree and a 120 degree shift between systems as indicated by the voltage measurements between T1 and T3. Or can I? That is the question I am wrestling with.

I am using the following figures I obtained from drawing my own phasor diagrams:

Vpu = V line-to-neutral (7.97 kV nominal)

Vpu = .518 corresponds to 30 degree phase shift
Vpu = 1.0 corresponds to 60 degree phase shift
Vpu = 1.414 corresponds to 90 degree phase shift
Vpu = 1.732 corresponds to 120 degree phase shift
Vpu = 1.932 corresponds to 150 degree phase shift
Vpu = 2.0 corresponds to 180 degree phase shift

Admittedly more measurements will have to be made before we can definitively parallel. However, if I can't resolve these initial voltage measurements, I'm in pretty sorry shape for completing the rest of the job.

 

[redtrumpet]

busbar - did you post to powerlineman.com on this issue back in January? I read a thread on phasing two 2400 V delta systems and someone named busbar listed the 15 measurements to be taken. I take it that was probably you

 

[stevenal]

Redtrumpet,

I can't find fault with your analysis. The only explanation I have involves a double contingency, so this may be out there. Your meter reader is dyslexic, mistaking a 5 for a 3 on a digital display. T1 is miswired, with all polarities backward. (150 degrees shift from high to low) T1 is therefore 180 degrees out of phase with T3.
You might consider doing a polarity test on T1. Hope there's a better explanation. I don't see one, though.

 

[redtrumpet]

stevenal, I like your idea of checking the transformer polarity. I am also taking the tack of assuming the voltage readings are correct, and trying to envision the system configuration that would give the readings. For example, a transformer neutral may have opened with a simultaneous ground fault. This would give a neutral shift that may account for the voltage readings - I haven't examined it in detail yet.

 

[busbar]

Transformer phasing pitfalls are covered in §19 of ‘Electrical Safety Handbook’ by Richard Harris – American International Training Institute, Nashville TN – 615.399.0520

RT—Yes, that was my comment at JLC. War story: Some years ago at a rural research facility, a 2-mile tie line was built between two 115/12kV 20MVA banks, with flopped hi-side phases on one bank. Some doodling convinced me that there was a potential problem that couldn’t be fixed by any lo-side flopping or rolling of cables. Because a change on the hi-side was going to involve an overtime crew from the serving utility, for several months there was reluctance to deal with the problem head on. So, a mockup of the existing equipment was needed. We ended up connecting six little 100VA machine-tool transformers in various delta/wye pairs to demonstrate the problem.

 

[dpc]

redtrumpet,

This whole thread is giving me a headache just reading it, so you must be having a great time.

I'm not sure how you're measuring the voltages and phase angles, but is it possible there is a PT (or aux PT) polarity reversed somewhere?

dpc

 

[redtrumpet]

I didn't witness the testing, but am reviewing results submitted by the client. I have sent a note back to the client's electrical engineer asking if he knows what's going on.

I think the best course of action is to do polarity, phase-relation, and phase-sequence tests at each transformer as per ANSI C57.12.90. This will at least confirm that the transformers match nameplate and are indeed Dy1. The client mentioned one transformer has been rewound, I wonder if the internal leads weren't brought to the wrong bushings after the rewind.

Given a choice, is use of a ratio meter preferable to the voltmeter method described in ANSI C57.12.90?

 

[stevenal]

DPC,

I was assuming that the testing was done with phasing sticks. Only one measurement is usually taken at a time, so only magnitude is measured. Redtrumpet is coming up with possible phase angles based on the magnitudes.

 

[redtrumpet]

You have hit the nail on the head, stevenal. All I have are voltage magnitudes across and between stabs of an open tie breaker. I don't have phase angles or polarities. I am trying to infer the phase sequence and shift from the voltage magnitudes. I have done this rather simply, by using the law of cosines to figure out the magnitude of measured voltages for phase differences in steps of 30 degrees. Perhaps my method isn't valid. In one case, it predicts the configuration accurately. In the second case, I have both 60 and 120 degree shifts which seem to me to be an impossibility for two three-phase systems, each phase displaced 120 degrees in time.

For the second case, there are line-to-line voltage magnitudes across corresponding stabs. If I choose a reference stab, I would expect that measuring from the reference stab of one system to the other two stabs on the second system, I would get one reading of line-to-line voltage and <10% line-to-line voltage for the other reading. I should be able to roll the leads on one system and eventually end up with <10% L-L voltage across all three stabs. However, I have line-to-neutral voltage magnitudes across the other two stabs. Is this an error in the measurements, or an inability on my part to envision other system configurations that could give this measurement result?

 

[busbar]

A side note: If you need to document the systems, ANSI/IEEE C57.12.70-1978 §8 cites available variations. I think that, for the combinations listed, they are all symmetrical giving equal readings for all 'pairs' of readings.

 

[redtrumpet]

Agreed, busbar. The client's engineer agrees the second set of measurements look strange. We will redo the measurements across the tie breaker and probably do phase-relation and phase-sequence tests for each transformer.

 

[NormGA]

Are all these transformers in the same yard? If not, has line phasing been checked and rechecked? To me it appears that they are not in the same substation yard and are separated by some distance and the line phasing is not actually what it appears to be. In my company I have seen many older stations where the high side phasing posted on the frame, and on the prints, is not what is actually there. These were all 46/12 stations. It was sometimes very interesting trying to connect a mobile substation to some of these older stations.

In your first posting you stated that&quot;line-to-neutral voltages....&quot;. Do you actually mean &quot;line-to-line voltages...&quot;. I don't understand how line to neutral voltages, without a phase angle, can help in this situation.

 

[redtrumpet]

NormGA - in my first posting I should have said line-to-neutral voltage magnitudes, and line-to-line voltage magnitudes. With a 60 degree phase shift between two systems, you will read line-to-neutral voltage magnitudes between identical phases of the two systems, although you are obviously not measuring the line-to-neutral voltage.

One transformer is not in the same yard as the other two, but I have walked the primary lines in their entirety and have verified the primary transformer connections.

 

[redtrumpet]

Just an update - I determined that the questionable set of measurements could not exist between the two systems. I asked the client to retake the measurements. The new measurements indicated 60 and 180 degree phase shifts, not the 60 and 120 degree phase shifts indicated by the first set of measurements. These measurements match what I expect to see. Although I had expected to see either 0 or 120 degree shifts between the two systems, visual inspection of the primary connections to one transformer showed that they were reversed from what I had originally been told. Therefore, 60 and 180 degree phase shifts showed up.

We are now doing a feasibility study to correct the phasing to allow parallel operation. Ironically, although this customer has five transformers with conventional A-B-C primaries and only two transformers with reversed A-C-B primaries, because of substation primary bus arrangements it appears it will be easiest to change all transformer primaries to reversed A-C-B connections.

 

[busbar]

RT-- Hopefully your client will fund updates on signage and documentation.