Continue to Site

Eng-Tips is the largest engineering community on the Internet

Intelligent Work Forums for Engineering Professionals

Register Log in
  • Congratulations The Obturator on being selected by the Eng-Tips community for having the most helpful posts in the forums last week. Way to Go!

An interesting set of Voltage Measurements 2

Status
Not open for further replies.
davidbeach

davidbeach

Electrical
Mar 13, 2003
9,493
I was at a project site on Friday and saw some interesting voltage measurements that I would be interested if anybody could explain. Now, I know these aren't mathematically possible, but I saw them, other saw them, and they had been reported the day prior, which is why I went to the site; so nobody need respond that there was an error in the readings.

The relevant portion of the system begins with a 480V delta transformer winding. The three phase from this transformer are fed into some 480V switchgear. Also connected to the switchgear, but not running (and breakers open) are three generators which can be run parallel to the utility, that's why the delta winding. Connected to the 480V bus are a set of 277V-120V grounded-wye/grounded-wye VTs (yes, they should be 480V-120V and that will be corrected but that shouldn't matter for this). The output of the VTs are connected to a CAT/ISO control module, an Eaton Power Meter, and an SEL-351 relay. The control module has only phases A and C connected plus the neutral of the VTs. The Eaton Power Meter has the three phase voltages input and has a separate ground connection. The SEL-351 relay has all three phases and neutral connected. All three devices have a VT ratio of 2.31 programmed in. The SEL relay and the Eaton Power Meter report about 281V line to ground/neutral for all three phases as could be expected. So far, so good.

Now the problem; the CAT/ISO control module reports 148V secondary/342V primary. OK, that's easy - a bad module. Well, if only it were that easy. Two different Fluke DMMs were used to make measurements (I didn't get the model numbers, both had calibration stickers indicating calibration in 2006) and this is where it gets very interesting. Measuring on the bus of the 480V gear, the Flukes read 485V to 490V across each of the three phase pairs - completely reasonable, and an indication that the meter isn't wildly off. They also read between 340V and 345V phase to ground. On the secondary of the VTs, the Fluke meters read 210V to 215V phase to phase and between 145V and 150V phase to neutral.

I had the thought that on this ungrounded system, the Fluke meter was somehow causing a neutral shift, but a couple of tests tend to disprove that. One test was having both Fluke meters used at the same time, one A to ground and one C to ground, and they both read about 348V. Another test was measuring the voltages at the terminals on the rear of the SEL relay while watching the metering display on the front of the relay. With a VT ratio of 2.31 programmed into the relay, and a metering display of 281V phase to neutral (primary), the relay thinks that it has 122V applied to the terminals. Measuring across the same terminals with the Fluke meter, the fluke read 148V and the metering display on the relay did not change.

Any ideas? The voltages listed above were all read from the devices indicated by more than one person. One oscillographic record was trigged and downloaded and it showed that there was between 119[°] and 121[°] between phases, so that isn't the cause either.
 
Replies continue below

Recommended for you

Sort by date Sort by votes
Hello davidbeach
This is an interesting problem, but I've overlooked something in your post I guess and I'm rather confused. Could you clarify:

The 480V windings are connected delta(three wire) and there are only three phase conductors brought to the switchgear?

There are no earth grounds connected to the delta either at the transformer or the switchgear?

The VT's are nameplate rated 277v/120v and are connected grd Y/grd Y ?

The VT's are simple wound units and not CVT's of any sort?

Thanks
 
Upvote 0 Downvote
David, some ideas:
I summarise that your phase to phase measurements are OK, both on primary and secondary of the VTs, but the phase to gnd measurements are too high on primary and secondary.

This might imply that your star point of the VTs is not at the same potential as gnd.

The two devices that measure phase-n correctly may deduce this from 3 phase measurements, whereas the device that measures the same as the fluke may be measuring phase-n, same as the fluke.
 
Upvote 0 Downvote
I'm going to have to sketch this lot out, but as a quick thought: what is the grounding arrangement for the VT secondary? I know you said 'grounded Y' but legacy UK practice frequently grounds the 'Y' phase of the the RYBN (red / yellow / blue / neutral) group. Some electronic instruments really dont like this connection and produce some extremely confusing results. From memory the early GE Multilin PQM II's and the Yokogawa WT series power analysers behave strangely on installations where the star point is not the grounded connection. I have been spotted chasing wild geese for a few fraught hours on a couple of occasions wondering WTF was going on!

----------------------------------
image.php
Sometimes I only open my mouth to swap feet...
 
Upvote 0 Downvote
Hello David;
My stories may not be relevant but here goes:
Story #1. I sent some students out to the shop to connect some
transformers with the secondaries in wye and verify the voltage ratios between the line to line voltage and the line to neutral voltage.
There was a significant error. We tried with different volt-meters. Still bad readings.
I rolled out a 'scope and looked at wave forms. It was a long time ago and I don't remember which connection had the distortion, but one waveform was distorted. The meters were the old Simpson AVOs. They read average voltage and used the standard form factor to indicate RMS. The distorted wave form had a different form factor.
Story #2. A pump station start-up. The designers may have been a little unsure of themselves and seemed to be ultra conservative.
The station service was grounded wye. Solidly grounded wye.
To the solidly grounded wye they had added an artificial neutral. I am sure that the heat production was beneficial in keeping the switch gear dry in a potentially damp enviroment.
To this was added ground detector lamps.
There were also phase to phase panel voltmeters and phase to neutral panel meters.
The designers tried in vain to locate 480 volt motors for the pumps. The closest that they could come was the industry standard for North America, 460 volt rated motors.
They asked the utility to lower the voltage taps on the transformer bank as low as possible.
Come the start-up. The only day that the tides were suitable during the day fell between Christmas and New Years day. I interrupted my vacation and drove through two mountain passes in a mild storm to be present as the contractors rep.
The designers had a utility crew called out to do a final voltage check before energising.
Now the problem. The phase to phase panel meters did not show the proper ratio between line to line voltage and line to neutral voltage.
They asked the utility crew to check. I checked with two meters. All the meters were d'Arsonval movement meters. All were responding to average rather than RMS.
I remembered the previous experience and gave the assembled engineers a quick tutorial on distorted wave forms and form factors.
Everyone stepped back and let me close the main breaker. As soon as there was some load on the transformer bank, the voltage ratios lined up perfectly.
As a first step, I would see if any io the meters were average reading, and I would consider adding a load (probably some light bulbs) to the vt secondaries.
The wave form was roughly the shape of a positive half-cycle sin wave, but only about 80 degrees duration. There was then about 10 degrees of zero voltage, then the negative half wave, followed by about 10 degrees of zero voltage.
It will be interesting to see if this factor is at play here.
Respectfully
 
Upvote 0 Downvote
subtech, ungrounded delta windings. There is undoubtedly some connection between the switchboard ground and the case ground of the transformer, but only three phase conductors and none of them are grounded. The name plate of VTs, wound as far as I can tell, are 277V primary, 4/sqrt(3):1 ratio, which works out to 277V:120V. They were connected grounded-wye/grounded-wye when I arrived and during the testing the primary ground was lifted, but that didn't change anything.

DaveScott, A continuity test was not made between the secondary wye point and switchboard ground, but since direct measurements on the primary of the VTs was showing the same problems as on the secondary, I doubt that was the problem. If I'd seen different results primary and secondary I'd agree. As far as I know, the SEL relay only measures voltages between the phase terminals and the N terminal; it won't provide voltage readings in open delta applications unless the grounded B phase conductor is connected to both the B terminal and the N terminal.

ScottyUK, Brand new installation, grounded wye at the wye-point, not at the the B phase. Because it is a system that can operate ungrounded, the VTs should probably have been grounded on B instead of the wye-point, but they weren't; another sign that the people who put it together don't do much with ungrounded systems
 
Upvote 0 Downvote
waross, interesting; we were apparently posting at the same time. The oscillographic record downloaded showed nice clean voltage waveforms, but it was filtered data. I'll see if it is possible to get a raw data record of the same event.

Some people have tried to brush it off by saying that the voltages will get to where they are supposed to be once there is some load, but the only way to get load is for the voltages to be close enough to where they should be to allow the generators to sync to the bus - so far it was just the transformer with no load. Actually that's not quite true, and I may have just thought of something. A bit bigger picture - on the high side, there is one 12.47kV feed that feeds two 2.5MVA transformers paralleled to the bus each through their own breaker. Also on the bus are three 1.5MW generators. The connection to the transformers is there to allow the generators to back feed the utility and those breakers are interlocked with other "load" breakers that serve transfer switches to supply power to hospital loads on loss of utility. Transformers or transfer switches, but never both at the same time. I wonder if circulating currents between the two transformers could have been causing problems some how - but I don't really know how they could.

Also, for what it's worth, my visit to the installation on Friday was my first involvement with this, I have no vested interest in any outcome other than eventually understanding what is going on.
 
Upvote 0 Downvote
Hi David.
My first guess would be that the anomaly is being developed by the wye connected CT's, not the delta connected power transformer.
I suspect the VT's
Both my experiences have been with wye connections. Not much you can do to load the power transformer, but you may be able to load the VT's with light bulbs for a test.
respectfully
 
Upvote 0 Downvote
David,
I have been working on the exact same problem at our refinery.

Our readings were normal phase to phase (480v) and around 140 L-n, 200 L-n and 400 L-N.

We thought at first we had a transformer problem. So we took the transformer offline and perfromed a megger test and turn to turn ratio tests all off which passed.

The next step was verifying the ground grid at this substation. Since it is located in our tank farm it was not directly connected to our overall ground grid. To make a long story short, we drove 3 ground rods at each of the 4 corners of the sub and this changed our voltage readings to about 250 L-n, 280 L-N, and 340 L-N.

The next step is to investiggate the individual VT's that create the ground detection system. If the Z is not close on all of them, it may throw off the system balance.
 
Upvote 0 Downvote
arcflash99, I can see your measurements as within the realm of possibility; I have tried to work out the math to figure out where the neutral might be, but you have an offset neutral. In this case where a balanced system should produce about 280V line-to-neutral primary, we were seeing 348V line-to-neutral primary on all three phases, what one would expect from a 600V system, not a 480V system. An offset neutral wouldn't concern me as long as the voltages (and phase angles) would produce the phase-to-phase voltages. The only way to make the math work on this set of readings would be if the neutral point were not on the same plane as the three corners of the phase-to-phase voltage triangle; and that doesn't make sense.
 
Upvote 0 Downvote
waross--

Two thoughts:

1. Have you tried these measurements with a low-impedance meter, like an old Simpson or something. High impedance meters such as Flukes are notorious for giving strange readings on ungrounded systems.

2. I'm really interested in voltages and phase angles in this situation.

old field guy
 
Upvote 0 Downvote
davidbeach
Given your statement about the 12.47 KV feed to this transformer, I'd go back and take a serious look at the grounding grid between the switchgear containing the
VT's in question and the incoming wye feeder. It seems to me that there are many workers who don't understand the importance of good neutral bonding in general and grounding grids as well. I've worked in many substations where I found the grounding grid cut to pieces during previous projects and never put back together.
The more this is discussed, it smells like the same open ground grid problems I've seen before. I'm sure it's completely possible I'm missing something else here, but I'd sure check out the grid integrity first if I could get to it.
By the way, there are no resistors or reactors in the upstream transfomer neutrals are there? I once saw a rather large transfomer on a utility that employed a neutral reactor that got virtually eliminated from service by one misplaced 4/0 copper jumper on the transformer tank near the X0 bushing............
 
Upvote 0 Downvote
David, I'm trying to understand more clearly:
//...One test was having both Fluke meters used at the same time, one A to ground and one C to ground, and they both read about 348V....
...They were connected grounded-wye/grounded-wye when I arrived and during the testing the primary ground was lifted, but that didn't change anything...
...A continuity test was not made between the secondary wye point and switchboard ground, but since direct measurements on the primary of the VTs was showing the same problems as on the secondary, I doubt that was the problem...//

You were measuring from VT connection to ground, and got some strange readings. When you disconnected the ground (where?), it didn't make any difference. It is not known if there was continuity between VT star point and ground.

If the primary and secondary VT star points were commoned and not grounded properly, but you measured from VT to ground, then this could be one possible problem, since you are measuring the voltage across the VT plus the additional impedance to earth.
 
Upvote 0 Downvote
Hi David:
I've been thinking about this all day. Then it hit me. As I understand the connections, your VTs have a floating primary neutral on a wye/wye connection. With a delta transformer winding, connection of the wye point to ground is immaterial because there is no system neutral connected to ground.
If the VTs were delta connected on the secondary, the delta would force proper phase angles of the primary voltages and this in turn would result in the expected voltages on the secondarys.
With a floating wye primary/wye secondary, the voltages on the secondaries are inversly dependant on the loads on the secondaries.
With no load on the transformers, the secondary voltages will be dependant on the exciting currents of the VT primaries.
The current in each primary will be dependant on the impedances of the three primary windings. Hysteresis will render these impedances non linear.
Rather than a neutral point out of the plane of the other voltages, think of it as a rotating neutral point.
As the voltage in "A" phase reaches a peak, the impedances of one or both of the other VTs become proportionatly less than the impedance of the "A" phase VT. The neutral point shifts away from "A" and the resultant voltage is higher.
There will be the same effect as each phase peaks. The neutral point becomes a rotating vector and revolves around the geometric neutral point.
Proof of concept:
Look for some resistors suitable for 480 volt service.
Connect them in wye to the delta transformer winding.
Measure the voltage from the VT wye point to the resistor wye point.
Connect the resistor wye point to the VT wye point. That is, use the resistors to generate a true neutral point. I anticipate that the VT secondary voltages will be corrected.
With time limitations, I would use resistors which draw several times the exciting current of the VTs.
If you have the luxury of a lot of time and resistors, it may be interesting to vary the resistors ratings so that the resistor current ranges from about 10% of the exciting current to several times the exciting current.
Comments are welcome
Respectfully
 
Upvote 0 Downvote
Ah, yes. I think you are right. I just checked what we do with electric arc furnaces (floating delta), and there is a set of 0.1uF capacitors, with fuse protection, which are in star connection to provide the metering phase-gnd measurement.
 
Upvote 0 Downvote
waross, an LPS for you, I think you may have solved it, thanks. I don't know if the testing you suggest will be able to take place or not. When the 277V VTs are replaced with 480V VTs, it sounds like they will also be connected in open delta (not my preference, but again it's not my project).
 
Upvote 0 Downvote
David,
I don't understand how I can have an offset neutral with a delta secondary.

If it is offset, how does one correct it?
 
Upvote 0 Downvote
arcflash99, The offset neutral is just where the leakages and stray capacitances have put it. I'm not sure you can correct it without causing your system to become grounded. It's not really a problem for the electrical system, just a bit annoying for your phase-to-ground voltages. In a delta system, the only real meaning that can be attached to phase-to-ground voltages is for the calculation of zero-sequence voltages, everything else should be based on phase-to-phase voltages.
 
Upvote 0 Downvote
The phase to ground problems is what started our investigation into the phase to ground unbalances. All motors, MOVs, and other 3 phase loads have been working just fine.

Thanks for the info
 
Upvote 0 Downvote
David,
What waross says makes a lot of sense if the PT neutral is floating. I thought I understood you to say the neutral was originally grounded and now is not. Can you access the floating PT neutral and measure the voltage to ground?

Another item that might cause rotation of the neutral would be if the PT's are wye-ground on primary and saturating (you said they are 277V).

Also, I recently read an article that described a problem with wye-ground PT's on ungrounded systems causing ferroresonance.
 
Upvote 0 Downvote
Status
Not open for further replies.

Similar threads

Cesme
  • Question
Soil electrical resistivity measurements for a wind turbine location
Replies
3
Views
5K
jghrist
jghrist
Distribution73
  • Question
Circuit Breaker Transient Recovery Voltage Parameters. Interpretation of Time Delay (Td) line.
Replies
0
Views
1K
Distribution73
Distribution73
KillBill7
  • Locked
  • Question
Overhead line structure footing resistance measurements 1
Replies
8
Views
267
wcaseyharman
wcaseyharman
NickParker
  • Locked
  • Question
Auto Transfer Switch in place of Low Voltage breaker
Replies
7
Views
266
wcaseyharman
wcaseyharman
jmelo28
  • Locked
  • Question
EN 50522 Calculation of Permissible Touch Voltage 1
Replies
1
Views
242
jmelo28
jmelo28

Part and Inventory Search

Sponsor

Back
Top Bottom