Transformer %IZ Change 1

[jwilson3]

We have two transformers at a hospital that were recently tested by the manufacturer. Both are 5 years old and are 12.47kv-480v delta/wye. the 5000kva unit's Percent Impedance test revealed an average of 9.5 ohms and a calculated X/R of 1.6. Nameplate is 6.07% and typical X/R is closer to 9. The 3000kva unit also tested high, but not to the same extent. The manufacturer recommended replacing both units.

What could have caused this change? Neither has been subjected to a significant "thru-fault".

 

[Shortstub]

Who carried out the tests? The manufacturer? If so what is their reasoning? Sounds like a Bill Gates Wannabe!

 

[DanDel]

Percent Impedance is a factory test and is stamped on the nameplate. This value should never change unless the transformer is rewound.
You have stated "Percent Impedance test revealed an average of 9.5 ohms " where this value should be in percent.
I am also curious as to who did the test and where the results came from.

 

[jwilson3]

The manufacturer did the test and I can't get a real good rationale for the replacement, or why the impedance changed. They say coils could have been physically deformed, as by a heavy secondary short circuit, which has not occurred.

The 9.5 is not ohms as stated: it is the % impedance derived from their test. They tested each coil separately and I'm not sure what device they used.
The secondary coil was shorted and they applied a voltage to the primary side. The results on the center phase of the 5000kva unit is as follows:
pri. volts: 27.41
pri. amps: 3.0
sec. amps: 135.4
watts: 40.0
calc'd Zbase: 93.3 This is on a single phase basis as that's the way it was tested.

Pri impedance is calc'd by 27.41/3.0 = 9.14 ohms
% impedance calc'd by 9.14 X 100/93.3 = 9.793%

Thanks
Jim Wilson

 

[DanDel]

I would not consider changing out the transformers until I saw the results of a set of typical maintenance tests(megger, TTR, Doble PF).
The test procedures for the %Z are interesting. I believe it is usually tested by shorting the secondary and applying a low but increasing voltage on the primary until the secondary winding carries rated current. The percentage of the primary voltage used compared to the full rating is equal to the percent impedance.
I'm very curious about the test results since they applied just 0.2198% of the primary voltage. (By the way, are you sure the tap setting is on 12470/480V?)
I would expect that excitation and core loss values at low level of input voltage might account for a higher impedance, but I'm not sure from where the test procedure came. Do you have any references for the test procedures?

 

[BJC]

I think you need to see the rational for the test proceedure. If the transformer is 5 meg there should be a record of the original factrory test. I would ask to see that as well as the "test" they just did.
Did the same people provide the switchgear downstream from the transfromer? At 6.07% you have close to 100KA of fault current available. Prehaps they sold you some 65 KA gear and want to get that transformer out of there.
5,000 KVA is big for a 480 volt transfromer, what kind of switchgear do you have connected to it?

 

[jghrist]

The most likely thing to cause a change in impedance is a heavy thru fault which could distort the windings. There is something wrong with the X/R ratio. The original X/R of 9 and Z=6.07% would give a resistance (at 12.47 kV) of 0.208 ohms and a reactance of 1.876 ohms. The new X/R of 1.6 and Z=9.5% would give a resistance of 1.566 ohms and a reactance of 2.505 ohms. Resistance increased 750%. Maybe you have a bad connection internally that developed a high resistance. This high resistance would cause heating could be confirmed by a dissolved gas analysis (DGA).

What was the reason for the manufacturer's test? Impedance testing is not a normal maintenance test. DGA is a more common maintenance test that can be done without taking the transformer out of service.

 

[peterb]

Were the tests done on more than one tap? I suspect that the tapchanger may be part of the problem, as the resistance value appears to have changed radically from the original.
From the measurements given, the value of resistance is derived as follows -
VA = 27.41*3.0 = 82.23
PF = (W/VA) = 40/82.23 = 0.486
R = Z*PF = 9.14*0.486 = 4.44 Ohms

From nameplate test value, Z = (Zpu)*(kV^2)/MVA = 0.0607*12.47^2/(5.0/3) = 5.66 Ohms
Assume X/R = 9 :
tan(phi) = 9
cos(phi) = 0.110
R = Z*cos(phi) = 5.66*0.110 = 0.623 Ohms

Could contact resistance in the tapchanger be the culprit, at least partly? Also some questions need to be answered re the test method - where were the voltage drop measurements taken? How long were the test cables?

 

[jghrist]

jwilson3 said:

The secondary coil was shorted and they applied a voltage to the primary side. The results on the center phase of the 5000kva unit is as follows:
pri. volts: 27.41
pri. amps: 3.0
sec. amps: 135.4
watts: 40.0
calc'd Zbase: 93.3 This is on a single phase basis as that's the way it was tested.

Pri impedance is calc'd by 27.41/3.0 = 9.14 ohms
% impedance calc'd by 9.14 X 100/93.3 = 9.793%

The impedance test should circulate rated current in the primary (231.5A). Zbase using 1ø quantities is (kV ø-g)²/(1ø MVA) = 7.2²/(5/3) = 31.1 ohm, same as using 3ø quantities.

When testing one phase at a time, measuring power of P12, P23, P31 and volts of E12, E23, E31,

Impedance watts = 1.5·(P12+P23+P31)/3
Impedance volts = 0.866*(E12+E23+E31)/3

See ANSI/IEEE C57.12.90-1987, IEEE Standard Test Code for Liquid-Immersed Distribution, Power, and Regulating Transformers, Section 9.2.5.

 

[jwilson3]

Thanks for all the great comments. I'll try to answer the questions you've raised.

To DanBel: I agree with you about not rushing to replace the units. The TTR and PF tests showed not turns ratio change and the PF indicated good insulation. (Note: These are dry type xformers). You're right about how the impedance is tested in the factory and the field tests are done at a fraction of rated values because they use smaller test units I guess. The taps are in fact on the 12470v-480v setting. Unfortunately I did not get any details of the test procedures.
To BJC: The main secondary breaker and transformer are of the same manuf. Current limiting fuses are in series with the breaker to cover the 100KA. You are right about a 5MVA transformer. If I'd designed the system I'd have split the load up and used two transformers.
To jghrist: I don't know the reason for the % Impedance field test. I agree with you. This is a dry xformer so a DGA test can't be done. How about a partial discharge test?
Referencing your calc's, rated primary current in the coil is 133 amp which is line current/1.732 or 231/1.732.
I came up with the same Zbase as you (31) until review of GE Bull GET 3550F. Your equation is correct, but the phase voltage of 12.47kv must be used; (12.47)^2/5/3 = 93.3
To peterb: test on only one tap; the 12470 tap.

 

[jbartos]

Suggestion: Considering the price of transformer replacements and transformer test results, it would be better to have the transformers retested by an experienced firm with some good transformer testing record that would imply a testing procedure in compliance with an industry standards, e.g. ANSI/IEEE C57.12.90 or applicable one to the transformer type (it still needs to be posted).

 

[BJC]

I think they were trying to do a power factor test. Hence the recommendation to replace the transformer.
I would need to see a test diagram to determine weather or not "%Z" could be determined the way they went about it.
Were the primary coils tied together in a delta configuration or were they disconnected for the test? Was the secondary connected L1 to X0 or how.
Does the transformer have CTs on it and were they in place during the test?
I would like to see the test set up. I am always eager to learn something.

 

[jwilson3]

To BJC: They did a power factor test separately and the insulation tested OK. The impedance test was done on each phase separately. They shorted the secondary L1 to X0 and applied voltage to the primary. They didn't raise the voltage until they got rated secondary current as is done at the factory. They used lower levels for a field test, which is OK, but probably not as accurate. See my 19 May post that gives the readings for one phase. I think the methodology is OK, I just question the results.
No CT's.

 

[busbar]

Interesting situation. Have DC winding-resistance and excitation/iron-loss tests been conducted and compared to original to confirm or contradict the tester's recommendation?

 

[jwilson3]

To busbar: No DC resistance test, although another person has suggested this. No specific iron loss test, but excitation test was done. Nothing out of the ordinary was noted.

 

[Shortstub]

What is the manufacturer's "technical" reason for recommending replacement?

Have you made inqueries regarding the manufacturer's other customers and like installations?

 

[jbartos]

Question: What is the transformer cooling system?

 

[GusD]

Hi jwilson3

I don't know who has to pay for the transformers replacement?Probably you.If it was under warranty,I doubt it,if the vendor would be so quick in condemning the transformers.(from the info you provided)
They are 5 years old,they gave you no problems until the tests and it appears from your information the Transformer (Z) has changed from I don't know what to what.The 6.9% seems pretty normal.
I can't possibly understand what did the Vendor saw or measure, that calls for the changing of 1 Txr never mind 2 transformers.
Unless you have more information on the tests,than the info you provided on your post,I can't see the rational for the replacement.

GusD

 

[yenrameshn]

Hi Wilson, (to your answer to jghrist)
The base impedance per phase of a balanced 3phase system is
Zb=(KVb/sqrt3)*1000/sqrt3*Ib
=KVb^2/MVAb
where KVb is line to line base KV and MVAb is a 3phase MVA.
Zb is always a per phase value (if the numerator is per phase KV then the denominator is per phase MVA, if it is phase to phase volt in KV then the denominator is 3phase MVA)>In both the cases the result is perphase base impedance.
Hence
Zb=12.47^2/5=31.1 ohmsperphase always.

Test results ase per your report:

The secondary coil was shorted and they applied a voltage to the primary side. The results on the center phase of the 5000kva unit is as follows:
pri. volts: 27.41
pri. amps: 3.0
sec. amps: 135.4
watts: 40.0
calc'd Zbase: 93.3 This is on a single phase basis as that's the way it was tested.

Pri impedance is calc'd by 27.41/3.0 = 9.14 ohms
% impedance calc'd by 9.14 X 100/93.3 = 9.793%

My correction is Zbase should be 31.1 only. Otherwise other part of the calculation looks good.
One more question. While conducting the test was the transformer secondary earthed through some impedance? If that is the case centre phase shorted with neutral point and connected to the ground through an impedance, sec. current will be limited by the impedance value isnt it?
The turns ratio being constant the magnitude of the current through the primary should also be less.


Ramesh.

 

[jghrist]

yenrameshn,
I also thought the Zbase calc was wrong, but they tested each winding separately. The rated voltage of the primary delta winding is 12.47 kV and the rating is 5/3 MVA. Their calculation is correct on this basis. Works out to be the same, I think, as the ANSI/IEEE C57.12.90-1987 equation which includes a 0.866 factor when using single-phase winding tests.