Problem with delta to wye VT's on HRG system 5

[jwmccoype]

I'm involved with the replacement of the diesel genset on a 480V high resistance ground system.

As part of the project, a Schweitzer SEL-547 distributed generator interconnection relay is replacing several individual relays protecting the utility interconnection breaker. The OLD relays were all connected using two VT's in an open delta configuration.

The SEL-547 needs a four wire wye voltage input, and is designed for a direct connection to a 480V system without VT's, which initially sounded quite convenient. However, we were advised against such a connection in an HRG system. Something about the neutral begin unusable.

The recommended solution was to connect three VT's in a delta-to-wye configuration, thus providing a four wire wye output to connect to the relay. In order to supply 480 volts L-L on the secondary, a VT ratio of 1.73:1 was needed. The only VT's readily available at that ratio were actually labeled for 208:120 V service.

This scheme was wired up and the primary 2A fuses were closed with the secondary fuses open. The VT's heated up quickly and either two or three of the primary fuses blew.

Is the problem simply from using the VT's above their design voltage, or is there something more complex at work here?

Thanks,
John

 

[cuky2000]

Check if ferroresonance is a possible cause.

 

[waross]

Do I understand that you applied 480 volts to a 208 volt rated transformer winding?
respectfully

 

[davidbeach]

You need to start over and do it right with three VTs in wye from the lines. That initial open delta configuration blocks the zero-sequence voltage and putting another set of VTs in wye downstream of the delta VTs won't get it back. If the relay needs a 4 wire VT connection, you need to have a true 4 wire VT connection rather than something that looks like a 4 wire VT connection but is missing critical information.

 

[jwmccoype]

Hello davidbeach,

The original open delta VT's are gone. Three new ones were installed, but their primaries are connected in "full" delta with their secondaries connected in wye. All of this was done to derive a "usable" neutral. With that clarification, please expound a bit on what we should do differently.

Thanks.

 

[davidbeach]

You should have connected the primaries in wye with the wye point grounded. VTs should never be connected delta-wye, and in general I can see no reason to ever connect VTs in any configuration with a delta primary. All you do with a delta primary is throw away all of the zero-sequence voltage information; information that can never be recovered.

Unless you are using the secondaries of the VTs in a broken delta (not open delta) to directly measure 3V0, the secondaries should also be connected in wye. If connected to a grounded system, the wye point of the secondary should be grounded. If connected to an ungrounded system, the grounding of the VT secondary needs to be evaluated and it may be necessary to ground one of the phases (generally B) rather than the wye point. This is to avoid having a ground fault on the secondary be detected as a ground fault on the primary. On solidly or low impedance grounded systems the VT secondary ground fault can't draw enough primary current to be an issue.

In the days of electromechanical and solid state relays, the 2-VT open delta configuration probably had a reasonable economic explanation as the V0 (or 3V0) information was generally not used and there was no way of recording it for later analysis. Now with numeric relays, the oscillography produced during faults can be vastly more informative if the VTs are connected in wye, plus there are things the relay can do with the voltages that used to be much more specialized. Fault analysis is much easier if you know the phase-to-ground voltages in addition to the phase-to-phase voltages.

 

[jwmccoype]

Thanks for all of the good info, davidbeach.

Like I said in the beginning, the relay is designed to be connected directly without VT's, for systems up to 480 V. Do you think I have any chance of making this work with the three phase voltages connected directly to the relay (through fuses, of course) and the relay's neutral input grounded? This was my plan before I was warned by this other fellow about possible problems in an HRG system.

Thanks

 

[davidbeach]

I really can't go into the specifics of how to connect or use an SEL relay as they are the competition.

I doubt that a directly connected relay would impact the system grounding. If the voltage inputs on the relay (any relay) have a phase-to-ground voltage rating equal to the phase-to-phase voltage of the system it should be possible to connect them to an HRG system. If the phase-to-ground rating of the relay is less than that, you need the VTs. On an ungrounded system (capacitively grounded system) you should always have VTs between the system and the relay.

 

[rbulsara]

The reason transformers burnt up because, as waross, suspected, you applied 480V to 208V rated transformer windings.

You still need three (3) 480V:480V voltage transfomers with proper VA rating, to follow davidbeach's advice.

 

[apowerengr]

I'd suggest you contact SEL - they have great technical support. The connection of wye connected VT's to a high-impedance grounded system should concern you as the neutral to ground voltage can get very high. Obviously, the use of 208 volt VT's on the 480 volt system speaks for itself...................

 

[TestBeforeTouch]

I recently hooked up 3 PT's in a wye-wye configuration on a high resistance grounded system and it did not work well because the voltages on the primary were not perfectly balanced (1% unbalance). The problem is with the neutral floating on the primary side of the PT's the voltages on the secondary become distorted. My power quality analyzer said I had 18% THD.
The NEC in 250.36 does not permit the primary side of the PT's to be grounded on a high resistance grounded system as this would constitute a line to neutral load. Therefore on a high resistance grounded system you are probably better off hooking the PT's in open delta or delta-delta.
I also tried hooking up the PT's in a delta-wye configuration. The problem with this hookup is that the KWH meter and the motor protection relay's are not designed to take into account the 30deg phase shift on the voltage. This is a problem in that the power factor and therefore the power are calculated wrong.

 

[davidbeach]

The NEC deals with instrument transformers (VTs and CTs) separately from any type of load and considers them something distinct from any thing else connected to the system. In this case the VTs need a full 480V rating. I'm not sure why you (TestBeforeTouch) would have considered your VT installation to have been a failure; if the VT secondary accurately reflected the primary, where was the problem? Wye connected VTs are regularly used on systems of all grounding types. 250.36 would not apply any way as the VTs are not connected to neutral, the wye point of the VTs is grounded, always, regardless of the system grounding. The relay needs to know the voltage to ground, voltage to neutral would be of little use in a HRG system.

 

[jwmccoype]

TestBeforeTouch, thanks for your post. Good to know someone else has been in a situation similar to ours. We don't feel quite as crazy for trying the delta-wye connection now. What was your final solution?

As for applying 480V to a 208V VT, another member of the team convinced me that a voltage ratio was a voltage ratio, and as long as it was a 600V class device, we'd be OK. In hindsight, we would never think of doing that with a power transformer, so we shouldn't do it with a VT, either.

davidbeach, from you posts it sounds as if you recommend we should always ground the primary wye point, and either ground the secondary wye point or float the secondary wye point with a phase grounded instead. What factors would you evaluate to determine which secondary node to ground?

If it helps, the grounding resistors on the the generator and the utility transformer are sized for 5 amps.

 

[TestBeforeTouch]

The problem I had with an ungrounded wye on the primary is that the station KWH meter gave false information saying I had 18% THD when, in fact, I did not. The THD was due only to the ungrounded wye connection on the primary. The secondary line to ground voltages did not reflect the primary accurately because with slightly unbalanced voltages on the primary it caused the neutral point to have a sine wave on it, therefore distorting the secondary voltages.

There is a good chance I am missing the section, but I cannot find where the NEC allows you to hook the primary of the instrument transformers directly to ground.

 

[davidbeach]

I did a search on the word "instrument" in the NEC and there is no prohibition on connecting the VT wye point to ground. Otherwise there would be no means of measuring phase-to-ground voltage. Leaving the wye point floating will cause all manner of trouble, might as well not bother with the VTs if you are going to leave the wye point floating. The NEC is woefully lacking when it comes to use of instrument transformers and relays, tending to concentrate too much on fuses and circuit breakers with internal trip units.

Evaluate a ground fault on the secondary of the VTs. If the current on the primary of the VT is high enough to cause a zero-sequence current relay to pick up, you should ground a phase of the secondary (usually B) rather than the wye point. The relay (or meter) will still see accurate voltages, but "ground" reference used by the relay (or meter) won't actually be ground.

 

[TestBeforeTouch]

I believe the NEC does not allow you to connect the PT primary wye point to ground. The grounding electrode conductor is only allowed to be connected at one point per 250.36 (F).

If the PT primary is connected to ground then current would normally flow over the grounding equipment (if the voltages are unbalanced) and would cause a voltage to appear on the neutral resistor.

It may be that the only way to correctly connect PT's on a resistance grounded system per the NEC is to use a delta connection. This does not give you phase to ground voltage but phase to ground voltage loads can't be used anyway. An ungrounded wye would be legal but as I have found out does not work correctly.

I believe that if you have a solidly grounded system a wye connection would be allowed, but only if you connected the star point to the neutral conductor and not the ground, otherwise you violate 250.30 (A).

 

[davidbeach]

On what grounds do you base your belief that the NEC won't allow a proper connection of a VT? The VT is not a load and there is essentially no current flow. I've seen many installations with wye connected VTs on systems required to comply with the NEC and in all cases the wye point of the VTs is connected to ground. A grounded-wye - grounded-wye transformer does not provide a system grounding point, NEC 250.30(A) does not apply to the VTs. By your logic, it would be impossible to have any means of detecting ground faults on ungrounded systems and no means of monitoring the ground shift associated with ground faults on HRG systems.

The NEC 250.21 requires ground detectors. It is impossible to meet the ground detector requirement without connecting something between phase and ground. Remember this is a code requirement to connect something between phase and ground. In 480V systems, this is sometimes done using light bulbs (3 - 240V incandescent lamps in series between each phase and ground). In many 480V systems, and almost universally at higher voltages, the only way to do this is using VTs connected phase to ground. Again that is to meet a specific requirement of the NEC.

So, once again, what basis do you have for you belief that connecting the wye-point of a set of VTs to ground is a violation of the NEC, particularly when the NEC actually requires this connection?

 

[TestBeforeTouch]

I am just trying to have an exchage of ideas in a non confrontational way.

It is not necessary to connect something phase to ground to detect a ground fault. A very common connection is to measure the current through the neutral grounding resistor with a CT. A second method is to measure the voltage across the neutral grounding resistor (neutral to ground not phase to ground). A third method is to use a zero sequence CT.

I am aware that this is a common practice (common practices though are not always correct by the code, e.g. the issue of a obtaining a quiet electronic ground ).

But looking at it strickly from what the code book says I can't find a section that says it is allowed. I can find, however, where one could intepret that it is not allowed as I stated in the references above.

I do agree with you that the code is lacking in this area.

 

[waross]

Hi folks:
I am wondering what conditions the relay is intended to react to? Am I wrong in thinking that under conditions of ground current flowing, the phase to neutral phase angles and magnitudes will be different from the phase to ground phase angles and magnitudes.
If the meter is intended to monitor power functions such as KW, KVA, KVAR, and power direction, does it not have to have the wye point connected to the neutral to maintain the correct phase angles with the current transformers?
Likewise, if the wye point is grounded, the relay will be able to infer ground faults, but it will not be able to accurately monitor power functions if there is any unbalanced ground current leakage or faults.
In regards to potential transformers connected to ground on HRG systems:
I would assume that the currents drawn by the PTs will be very little.
Can you take the relay burden and divide by the applied voltage to estimate the PT primary current? The fundamental currents will cancel at the wye point and only harmonics will flow in the grounding resistor. Aren't we looking at a small percentage of a small current flowing in the grounding resistor as a result of grounding the wye points. The voltage resulting from the PT current flowing in the grounding resistors would be so low as to never pose a hazard.
davidbeach, I am asking here, and I respect your expertise in this field. If my comments are in error, please edit or red flag this post out rather than let me cause confusion.
thanks
respectfully

 

[scottf]

I concur with davidbeach. It is wuite common to connect VTs line-to-ground in a HRG system.

As to Waross's question....yes, you can approximate the current flow in the VT winding by taking the applied burden and secondary voltage to derive a secondary winding current then reflect that the the primary side. Of course, you also have to include the excitation current of the VT, which is also very low. With modern electronic relays and meters, the current in the 480V VT's primary winding is probably going to be in the order of 2-3 mA.