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

 

[jghrist]

It seems to me that whether you use line-to-neutral or line-to-ground depends on the purpose of the measurement. If you need to measure power or voltage to neutral, I don't think the line-to-ground connection will measure the correct value. According to the SEL-547 manual:

The connection is made directly—there are no interposing voltage
transformers. These voltages are used in the following protection and control
elements:
- Undervoltage tripping (device 27)
- Overvoltage tripping (device 59)
- Over- and underfrequency tripping (device 81)
- Reverse-phase-sequence (i.e., negative-sequence) alarming
(device 47)
- Directional power tripping (device 32)
- Synchronism check close supervision (device 25)

The manual does say connect to the neutral, not the ground, although the relay is meant for direct connection, not through VTs. The voltage inputs are rated 600 volts L-L. Why not follow the instruction manual and direct connect the 480 volt inputs? The OP indicated that they were advised against using direct connection on a HRG system. Was this advise by Schweitzer? If not, what is Schweitzer's advice?

 

[davidbeach]

It can get it bit tricky. As I stated earlier, I won't/can't address the specifics of an SEL relay, but I can discuss the general issues. If you want to discuss a certain, different, Intertie Protection System I would be glad to go into specifics of the relay.

In an HRG system, the neutral does not leave the location of the grounding resistor and is not available for connection to VTs or for direct connection to the relay. Ground should be available everywhere in the system.

I'm not familiar with "power to neutral"; as far as I know, power is power and it can be calculated using phase-to-phase voltages in a manner such that voltage to ground or voltage to neutral is irrelevant. In a three wire system, which all HRG systems are, you can measure power, real, reactive, and apparent, using the two wattmeter method and have no reference to ground or to neutral. This all that would be available if the VTs were connected open delta.

If the relay is attempting to monitor for ground faults on an ungrounded or HRG system (remember the neutral does not existing in an ungrounded system and is not available in an HRG system) the relay needs to know the individual phase-to-ground voltages. This can only be achieved if there is a ground connection on the primary of the VT.

If the relay is attempting to provide a complete protection package in compliance with IEEE Std 1547, there is no way that it could depend on any measurements from the neutral grounding resistor in an HRG system. There are no provisions in IEEE Std 1547.1 (the testing standard for compliance with 1547) for testing a system that would require inputs from a CT on a neutral grounding resistor or voltage across the neutral grounding resistor.

The current into the voltage terminals of a relay is somewhere close to zero, on the order of a few milliamperes. A circuit of sufficient size to warrant relay protection is very likely to have more current to ground through the system capacitances than through the VT. As you move into the higher voltages, the current into the primary of the VT is reduced by the ratio of the transformers, so that very small value becomes even smaller, and the high voltage increase the leakage through the system capacitance. The VT primary current is essentially noise lost in the leakage currents. The current necessary to feed the losses of the VT are likely to be greater than the current being transformed to the secondary.

jghrist mentions a line-line voltage rating of the relay in question that would allow for direct connection, but for connection to an HRG system it would also need the same line-ground voltage rating. I would not direct connect any relay to an ungrounded system and would apply surge arrestors of the VT secondary if applied to an ungrounded system.

If the relay is given each phase-to-ground voltage, it can calculate the corresponding phase-to-neutral voltage. If the relay is given each phase-to-neutral voltage, it knows nothing of phase-to-ground voltages and can not calculate them.

This is pretty much all settled state of the art, and it goes back many years.

 

[TestBeforeTouch]

I agree with your analysis. The ground current is very small, probably insignificant compared to charging current, I think the code needs to be revised to clarify this situation.

I am not familiar with IEEE 1547 or the term complete protection package. Does IEEE 1547 require voltage detectors? The NEC only requires ground detection without saying how it is to be done. One manufacturer,Startco, uses a CT on their SE-330 to measure ground current through the NGR.

In addition it is common practice to use a single phase transformer in the neutral of the transformer with a resistor on the secondary of the single phase transformer and then measure the secondary current or voltage to detect a ground. Does this violate IEEE 1547?

 

[stevenal]

Gotta disagree with DavidBeach on the power. Blondel says count the conductors and subtract one. Since ground can be part of this circuit, I count four. It is true that under un-faulted conditions two element metering will suffice, but once that ground resistor starts conducting the metering will be incorrect. Might not be a huge error, but it is still an error. Unlike LRG systems, HRG systems can operate continuously with a ground fault, so the energy error can add up over time.

 

[davidbeach]

OK, stevenal, you're right. I was only thinking of unfaulted conditions. You are right, a standing fault could mess things up. To me, as a relay guy, metering takes second place to having the protection work correctly.

On the other hand, I believe that with all three voltages to ground and all relevant currents (including neutral on systems with phase-to-neutral loads) the relay can do sufficient mathematics to accurately (within the limitations of the instrument transformers used) calculate power values.

I think that, to me, the important point is that there is valuable protection information available if the VTs are connected in wye and the wye point of the primary is grounded. Information that would not be available to the relay if the wye point is connected to neutral or if the VTs are connected in delta.

 

[jwmccoype]

Guys,

Thanks for the great discussion. I'm going to print all of this out and sit down with our team to figure out how to best apply it to our situation.

John

 

[stevenal]

Three element metering using line to ground voltage has the same limitation as two element metering. Consider a grounded phase; the phase to ground measurement is zero volts on that phase, so no power is measured on that leg. The three element metering is essentially two, and the power dissipated in the grounding resistor is not measured. You could capture it with a fourth element picking up neutral to ground voltage and neutral current. Or if the IED in question allowed you to input the resistor size, you could simply measure the neutral current. If you are satisfied with just measuring the power going to the load, the phase to ground voltage measurement will suffice. If you want the total generator load, though, you'll either need to use three element metering in front of the resistor using line to neutral voltage, or include another element or half as suggested above.

 

[waross]

I think that the correct answer may be to "Bite the Bullet" and instal six VTs. Three VTs for power monitoring, connected phase to neutral, and three VTs connected phase to ground for protective relaying. If the instrumentation will not accept six voltage inputs, then decide whether you want power monitoring, or protective relaying, and connect accordingly.
One connection may not be suitable for both jobs.

Power will be correctly monitored with the VTs connected to the Neutral. The phase angles between the voltage and the current will be the same as the phase angles of the loads on the three phases, so PF, KW, KVAR, and KVA will be accurate. Any ground fault will be a single phase load added to the load on the grounded phase and will be properly metered.

If we assume a load current of 1200 amps and ground current limited to 5 amps, then when the VTs are connected line to Ground And there is a Ground fault:
The load currents and the load current phase angles will stay the same. (plus 5 amps on one phase)

While the voltage to the loads remains unchanged, the voltage reported by the VTs will drop to zero on one phase.
The voltage on the other phases will rise to line voltage.
The angular displacement of the non zero VTs will change from 120 degrees to about 60 degrees.

I think it is clear.
For power monitoring, connect the VTs to neutral.
For fault monitoring, connect the VTs to ground.
If you try to serve both functions (power and protection) with one set of VTs, one function will be unsatisfactory.
respectfully

 

[davidbeach]

waross, excellent points. If we check the original post, the question is about a relay, therefore protection. In an interconnection application it seems to me (a relay guy) that getting the protection right is far more important than getting the metering right. Yes, modern relays have metering functions, whether the relays I deal with or the relays of those other guys, but none the less they are still relays, and relays are for protection. Therefore, it is always more important to get the protection correct than to get the metering correct. No relay (using protection grade instrument transformers) will ever provide revenue grade metering, so if you really care about the metering, use the right instrument transformers and use a dedicated meter. If you want protection and would like to get some metering information at the same time, use a relay and get metering information from the relay, but remember it is a relay and connect it like a relay. To my understanding, this thread was never really about metering, but was started with a protection question. So, if I can get metering that is close enough under non-fault conditions that would seem to be good enough for me, but I absolutely want the protection to work correctly.

 

[waross]

Hello davidbeach;
Thank you for your comments and your explanations.
I was trying to resolve the controversy over the correct connection method, and to put in perspective, the comments of the protection camp, and the comments of the metering camp.
Hopefully, jwmccoype will be able to sort the conflicting advice into the "metering solution" file and into the "protection solution" file.
It is not a question of "Which connection is correct?" as it is "What is the intended application?"
I think you are probably correct that this is a protection problem.
Respectfully

 

[scottf]

Davidbeach-

I agree with your last comments in principle, although, it should be noted that almost all VTs have a metering accuracy of class 0.3 in the US. They can normally be used for protection AND metering.

 

[stevenal]

I don't agree that phase to ground voltage is preferred for protection. From the User Guide: "All voltage pickup settings are in percent of Vnom (Vnom = setting NOMV/sqrt(3) = nominal input voltage, line-to-neutral)."

If you substitute line to ground voltage for the above, undervoltage tripping will occur for line to ground faults. This eliminates the ride-through advantage of having a HRG system.

I need to admit, after browsing the manual, that this particular device makes a really poor meter. Seems it just looks at Ia and Van and multiplies the calculated values by three. One element metering, so much for Blondel.

 

[DTR2011]

I think the protection issue was WAY overlooked here. I have installed / commissioned many of these types of Small/Medium 480 Gensets. The issue, I believe is providing Ground Fault protection. Since the system is HRG, measuring the currents in a typical residual connection would be difficult. Slapping a 50A CT on the HRG and feeding that into a current input would suffice. This would be the 50/51N Function. Grounded Wye/Open Delta with a resistor across the open part, feeding a 59G is another method for GF detection.

As someone mentioned earlier, the 547 is intertie protection. Basically IEEE 1547 requires the following protective functions

32-Reverse Power
27-Under Voltage
59-Over Voltage
81 O/U Frequency.

The protective functions are provide to protect the Utility in the event of down stream faults.

Connect the relay as SEL recommends. If there is not adequate protection for the generator, consider additional relaying.