ungrounded wye-delta 2

[nerb31]

the subject is 3-100kVA, ungrounded wye-delta transformer bank, 2400/4160Y-240/480V. During a spot check-up on this bank, we were surprised with the line-ground voltages 302V/320V/198V that were measured between the secondary terminals and the tank equipment grounding (as ground point). thinking it was due to grounded load or bad connections, the transformers were isolated, and terminals were cleaned. megger, winding resistance and TTR all turns good for these transformers. but after the cleaning, the same voltage magnitudes were measured when the transformer was commissioned. by the way, line-to-line voltages are balanced at around 480V. 4160V side is connected to upstream bus with LRG.

Has anyone experienced the same case or someone perhaps know what's going on with our transformers?

Appreciate your inputs.

 

[electricpete]

The low side winding of a given phase physically sits between the high side of that phase, and the core. So as first cut, the low side winding voltage tries to float somewhere between the adjacent high voltage winding and ground, also meeting the constraint that the three low side phase to phase voltages are relatively balanced.

CLH probably has more insulation and smaller capacitance than CLG. However if we assume it is not too much smaller, than it is easy to imagine that the high side has more influence on the low winding voltage than the adjacent ground (because voltage difference between low side and high side is bigger than voltage difference between low side and ground). i.e. VLH/XLH > VLG/XLG so more capacitive current flows between low and high than low and ground, so high has bigger influence on voltage.

Still with everything balanced on the high side, meeting the constraint that the 3 low side phase to phase are balanced, then I think it would put the lv windings balanced around ground. BUT what if there is a small imbalance on the supply side.... I think probably that can result in relatively bigger imbalance to ground on the low side delta winding. What do you guys think?

In any case, I don't think it's a concern because you have ruled out transformer anomaly with your testing.

=====================================
(2B)+(2B)' ?

 

[electricpete]

Attached is a quick attempt to explore what I suggested above... whether a small unbalance on high side voltages can result in relatively large unbalance on low side voltages to ground.

I assumed all CHL and CLG were 1uF (probably higher than actual, but I think all that matters is their ratio). To get rid of startup transients, I added a 1E5 ohm resistor in parallel with each cap (time constant 0.1 sec) and examined the solution starting after 10 time constants (1 second).

I assumed
VBAL~VAH
VCBL~VBH
VACL~VCH

I used 2400 peak phase-to-ground hi-side and 480 peak phase-to-phase low-side as nominal values (should've multipled by sqrt2.... sorry... doesn't change the conclusions).

I used 1 ohm internal impedance for each low side voltage source so it wouldn't hiccup at the thought of unbalanced delta loop which did not close.

The ltspice model is shown on slides 1 and 2.

Initially I increased VAH and VBAL magnitude by 1% above nominal. Result shown in slide 4 is that phase to ground voltages deviate 1.8% high and 0.7% from nominal.

Then as shown in slides 5 and 6, I changed VAH and VBAL phase angle by 2 degrees from their nominal position. That resulted in spread of voltage 262 to 292.

What it proves, I'm not sure. It did not show as much influence of high side unbalance on low side as I would have guessed, even though we chose CHL = CLG instead of more realistic CHL<CLG which would have given even less effect from high side to low. I am less inclined to believe high-side unbalance is causing your problem after seeing these results then I was before. (fwiw...I'm not sure how good a model this really is).


=====================================
(2B)+(2B)' ?

 

[waross]

Except that the high voltage is wye connected and the inner part of the winding may be connected to the wye point and there may not be a large deviation from ground on that part of the winding.

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[electricpete]

My model showed each winding as a single lumped capacitance at terminal voltage, which is a simplification from reality. I agree the further details of winding configuration can be important. If the neutral side of HV winding sits physically adjacent to LV winding, that will greatly reduce the effective CHL as you say. Likewise, if the neutral side of LV winding sits physically adjacent to the core, that will greatly reduce effective CLG. For all I know, these two effects tend to cancel since we are only interested in the ratio of CHL and CLG. I’m sure there are people on this forum that could give a better guess at the ratio CHL/CLG than me….

=====================================
(2B)+(2B)' ?

 

[electricpete]

For core form transformers, it looks like layer coil has the ability to put either line or neutral end of winding physically one id or od of coil. In contrast, disk type coil will have a continuum of voltages along od and continuum along id. See figure 1.12 here

http://books.google.com/books?id=Lz...tion for core-form power transformers&f=false

As far as capacitance calculations, there is a lot to it. Entire Chapter 7 here devoted to it. It looks like good reading for anyone who has time... not me at the moment.

http://books.google.com/books?id=Lz...that the capacitance of coil sections&f=false

=====================================
(2B)+(2B)' ?

 

[electricpete]

Likewise, if the neutral side of LV winding sits physically adjacent to the core, that will greatly reduce effective CLG.

Scratch that. There's no neutral on a delta winding ;-)

=====================================
(2B)+(2B)' ?

 

[waross]

This needs to be checked with three voltmeters simultaneously to eliminate the effects of meeter loading.
If the voltages are not congruent with the delta, look at the waveforms with a scope.

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[davidbeach]

Even three meters may not solve the problem. Refer to thread238-173479 and see what saturating VTs can do.

 

[electricpete]

It doesn't seem logical to me that meter loading would cause an apparent unbalance when sequentially measuring one phase to ground on an otherwise balanced system. The meter would create the same change in each measurement.

=====================================
(2B)+(2B)' ?

 

[electricpete]

.... however I'm in full agreement with getting more data if you want to try to find an explanation for your measurements.



=====================================
(2B)+(2B)' ?

 

[electricpete]

I also liked PHovnanian's suggestion to add some trial impedances to ground to help quantify things.

=====================================
(2B)+(2B)' ?

 

[waross]

I'm not concerned that the voltages to ground do not balance exactly. That's life in the real world. There is an another issue when the voltages to ground are not consistent with the line to line voltages.

Put the load on the transformer and see if the voltages to ground are consistent with the line to line voltages.

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[stevenal]

Pete,

Start with a simple idealized system, perfectly ungrounded, admittance to ground equal to zero. Attach a voltmeter with a non zero admittance from one leg to ground. The current through the meter is zero, since there is no current path (Note that this changes when three meters are connected simultaneously). Zero current through a finite impedance measures zero voltage. Do this sequentially and you will find each voltage to ground is zero, while the line to line measurement still measures nominal. The impedance of the meter pulls the ground up to match the potential of each corner when measured sequentially. Same effect on each phase as you said, but the effect suggests unbalance.

The real world is slightly more complicated. We have unbalanced capacitances and inductances involved and DVMs with high impedances. If you draw an equilateral triangle with sides of 480 and draw circles of radius 302/320/198 around each corner, you will see the circles do not intersect at a point, each being too short to reach both of the others at a single point. Again the same effect is seen on each phase, pulling the ground potential away from where it would reside if the meter were not present. This was what I expected.

As Schrödinger pointed out, the very act of observing affects what is observed.

In school, if we failed to account for meter effects, grades were effected. It might be insignificant, but you better include the impedance in your diagram and explain how it's insignificant. In this case, it makes a difference.

 

[electricpete]

Hi Steve – Everything you said sounds fairly straightforward to me, and at the same time, no conflict that I can see with anything I said.

We have two symptoms that we have discussed.
1 – the circles don’t meet / the measured delta voltages don’t close
2 – the line-to-ground voltage measurements are unbalanced.

I believe that I addressed #1 above in detail on 10 Apr 11 0:58 by proving that the circles will meet if the line-to-line voltage is 2.9% low, which certainly seems credible to me. (call it scenario 1A). In that same post, I acknowledged that meter loading could be an alternate explanation for symptom #1 (call it scenario 1B)…. (“For all we know the voltage could be exactly 480 and the meter loading is influencing the measurement slightly”) )

My more recent comments had nothing to do with symptom 1… I was addressing symptom 2: “It doesn't seem logical to me that meter loading would cause an apparent unbalance when sequentially measuring one phase to ground on an otherwise balanced system. The meter would create the same change in each measurement. ”

This was based on my perception that Bill was addressing problem #2… apparently I misunderstood Bill’s meaning and was projecting my own opinions onto his comments. Personally I think we have two very plausible explanations for #1 (scenario 1A and scenario 1B above), but I remain a little bit curious about 2. (why would we have unbalanced capacitances inside the transformer?.... and what are the magnitudes of associated impedances).



=====================================
(2B)+(2B)' ?

 

[waross]

Not unbalanced capacitances but unbalanced impedances.


Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[electricpete]

Which impedances do you have in mind?

I think possibly the high side magnetizing impedances could be slightly unbalanced since the center phase has slightly lower reluctance. That effect would be easy to check by comparing the hi-side currents to check for unbalance.

Any other impedance you had in mind?

=====================================
(2B)+(2B)' ?

 

[waross]

Resistance leakage. This may be less than capacitive leakage but may be at a closer phase angle to the meter impedance.

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[stevenal]

Lower reluctance on center phase? On a three phase transformer maybe? The OP was speaking of a bank of single phase (3-100 kVA) transformers. I'd say the unbalance comes from variations in the three units.

 

[electricpete]

My mistake, I was thinking 3 phase transformer. I was just trying to figure out what other impedance Bill was referring to. I had pretty much ruled out resistive myself since it was stated megger reading was ok, and I don’t think resistive leakage current wouldn’t be anywhere near capacitive leakage current unless something was extremely out of whack. Look at typical doble power factor test… insulation power factor tends to be around power factor of 0.01 or 0.02 capacitive, suggesting it is dominated by capacitance.

At any rate, I'm just throwing in my thoughts fwiw. I personally don’t know whether it is worth pursuing further or not… have never seen any other measurements on lv side of completely unloaded delta transformer to know if this is normal.

=====================================
(2B)+(2B)' ?

 

[electricpete]

Steven's comments have unspoken question for original poster (if not already asked): are these 3 identical transformers, or not?

=====================================
(2B)+(2B)' ?