Understanding ground fault detection scheme on a Cogen arrangement 2

[rockman7892]

I was at a facility the other day that was being set for startup when a discussion arose on how ground fault detection would work. I wasn't familiar with the project and was only a visitor so I stood in the background just listening. Of course after listening and hearing the discussion I have a few questions of my own.

I have attached a brief sketch of a One-Line for the facility in question. As you can see they will have several generators at 4.16kV (only show 2 on sketch) that will be operating in parallel with the utility to serve other loads at the facility and possibly feed back to the utility. The generators are impedance grounded at 4.16kV and feed through a step-up transformer to a 13.2kV bus to feed the other loads in the facility and the utility. The LV or generator side of the transformer is Delta and the HV or 13.2 KV side of the transformer is a reactance grounded wye.

The question first arose when it was noticed that there was no CT on the reactor or neutral bushing of the Generator Transformer primary. The question then was without a neutral CT how will ground fault be detected for faults on downstream feeder circuits on the 13.2kV bus. I did not show all of the CT's in the sketch but both 13.2kV and 4.16kV buses have a differential bus scheme protecting the bus. Saying this most people in the conversation seemed to think that becasue there was this bus differential protection faults at the downstream feeders (as shown on sketch) would be detected by the bus differential relay. I should also mention that there is a transformer differential relay scheme used on the Generator transformer.

I do not believe this to be correct for I dont think a ground fault at the downstream feeders will be detected by either the bus or transformer differential schemes because they are outside the differential zones so the differential protection current in will equal current out..

My thought was that you should be able to detect ground fault with the zero sequence CT and settings used on the feeder breakers. Would this be an adequate way of detecting ground fault?

Now I should mention that the neutral reactance on the generator transformer limits GF current to 200A.

Also if both the Generators and the utility are operating in parallel and since the utility is solidly grounded for a fault on the downstream feeders wont some of the fault current come from the utility while some came from the generator transforemr?

I'm interested to hear others input and hopefully learn from them.

Thanks for the help.

 

[rbulsara]

That is what good consultants are for.

Rafiq Bulsara

http://www.srengineersct.com

 

[rockman7892]

Rbulsara

Of course I understand this and as I mentioned I have nothing to do with this project was just simply an onlooker who overheard a discussion that sparked my curiosity.

I just wanted to take some time to digest the information being discussed and see if I could learn from others weather or not my opinion offered in my post was correct or not. I'm simply just interested in learning.

 

[rasevskii]

I should hope that they work out a primary test program and carry it out to see if the protection system actually works properly. Alas, in the USA (if that is the case) they usually consider this to be too expensive, and only do secondary tests and loop checks only.

Each of the generators (with separate excitation) is available as a current source, in principle.

rasevskii

 

[vandal06]

Rockman,

You have the gist of it. If the ground faults are downstream of the feeder relays the differential schemes employed on the 13.2 kV bus and on the GSU will not detect those faults.

They also have the issue of two varying levels of maximum ground fault current available on at the 13.2 kV level. As you correctly noted, the GSU will provide only 200A per the reactance grounded neutral. However, with the utility in service as well the available ground fault current will have two sources, 200A from the GSU and a value I would expect to much much higher from the utility because of the solidly grounded connection.

At least this facility has insulated the generation from seeing ground faults on the 13.2 kV system. I've had to coordinate systems where generation was at the same voltage level as the distribution, causing another level of complexity to the coordination.

 

[rbulsara]

In addition to what vandal06 said, you do not necessarily need a neutral CT on the grounding reactor for ground fault current sensing, as 51N function of the relay can be used as ground fault protection, which will calculate it by vector summing the three phase currents. OR you can use the neutral of the phase CTs as input to 51G function. This is true for any 3 wire loads, even on a solidly grounded system as well (say your feeder CBs).

Some authorities may want a separate residual relay for 51G. Voltage based 59G can also be an option for impedance grounded systems.

Rafiq Bulsara

http://www.srengineersct.com

 

[rockman7892]

Vandal06 and rbulsara Thank you very much for your responses.

Vandal

The GSU term is new to me but from looking it up it appears to stand for Generator Stepup Unit which I assume is referring to the generator stepup transforemr. Is this correct?

I'd like to briefly hear more about the difference in level of complexity of trying to coordinate a generator operating at the same voltage level as the distribution that you mentioned as opposed to coordinating one operating at a lower volage as in this case. Does it have something to do with determining what ground fault levels to trip at on the two different sources when at the same voltage level?

Rbulsara

I am familiar with both the 51N and 51G using the three phase CT's. When you mention using a seperate residual 51G relay are you referring to having a zero sequence CT surrounding all 3 phase conductors? If not what is the correct term for this type of GF detection method.

I have seen cases where the utility has had a CT on the transformers neutral bushing or on a neutral Resistor or Reactor to detect ground fault. What is the differences (pros/cons) between this method and the one you mentioned using a residual arrangement on the phase conductors? Is one more sensitive or accurate then the other, or is this just a form or redundancy?

It would also appear to me that any ground fault on the 4.16kV system and any feeders of the 4.16kV bus would have its fault source as the generators only and there would be no contribution from the GSU transformer since it is on the delta side of the transformer. I assume then all GF current would be limited to whatever the impedence grounding of the wye connected generators are. Am I correct?

 

[rockman7892]

rasevskii

You bring up a very good point about a primary testing program. I do not know much about this type of testing for a complete system but I am guessing it involves injecting primary current onto the system to verify that devices trip porperly. I have seen of course secondary injection testing and primary testing on single devices only, but have never actually seen or heard much on primary system testing.

This type of testing seems like a very good idea and I'm curious to hear from others how often this type of testing actually takes place here in the US, and besides cost what some if any of the downsides are to this type of testing.

Thanks

 

[waross]

You are correct in that the delta does not contribute to the ground fault directly and very little with an impedance grounded system.
However a ground fault will cause a voltage drop on one phase of the generator. With a solid ground and a solid ground fault this voltage drop will be 100% and the voltage will be zero. This results in a phase shift in the two phases.
If the wye side has the neutral solidly connected to the system neutral, the delta will try to pass enough current to restore the phase relationships and balance the phase voltages. So the the delta will contribute indirectly to a ground fault on the secondary side. If the GSU has the wye point impedance grounded, the delta is free to follow the phase angle and voltage swings and the contribution to a ground fault on the generator side of the GSU will be small. The ground fault current will be limited by the impedances to (N x 200 Amps) and depending on the relative impedances of the GSU and the generator windings, the GSU will supply a part of the (N x 200 Amps). This will show up on the primary as a current in phase with the fault on one line, a leading current on one line and a lagging current on the third line. All currents will be equal. All three currents will be 50% of the total.

This may also be a partial answer to your question concerning coordination. The transformer between different voltage levels may result in the protection coordination being less challenging.


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

 

[rasevskii]

for rockman7892:

To work out a Primary test program, the three-line diagram, showing all devices, CTs, PTs and ratios, Relays with ANSI function numbers, Settings, Disconnectors, Breakers, etc. has to be available. Nameplate data on transformers and generators essential.

Three phase and single phase deliberate faults are bolted in at certain points, and a generator with excitation source variable from zero upwards is used as the current (voltage) source. For ground fault protection tests, grounds are bolted on at certain points (after the previous tests are done) and the generator output is raised from zero upwards for these.

For the transformer differential protections (where usually mistakes or polarity reversals occur), primary testing is essential.

Obviously this is expensive, and is only possible at the original commissioning, before anything is actually energized at service voltage.

Usually such tests are carried out by the manufacturers reps, and witnessed and protocolled for the client. In the European
influenced world, it is absolutely standard practice.

Above are usually combined with tripping system tests.

Otherwise, how can one trust anything afterwards?

rasevskii

 

[rbulsara]

I had posted a reply to rockman's follow up questions of 19 Feb 11, 10:23..seems to have either disappeared or probably did not get posted..

A short version again:

1. I meant the single CT around the grounded conductor. Normally called residual/ Ground/Neutral CTs. 'Core Balance" CT that surrounds all conductor is seldom if ever ( I have not seen one) use at MV, as the phase CTS are invariably present.

2. To me there is no "technical" advantag of one over the other. They measure/calculate the same quantity. Some may see a separate CT as more convincing.

3. Bill answered that.

Rafiq Bulsara

http://www.srengineersct.com

 

[vandal06]

Rockman,

1) Yes, GSU = Generator Step-Up Unit, which refers to the transformer.

2) Bill touched on this. The delta-connected primary of the GSU will block zero sequence current contributions from the generators for ground faults on the 13.2 kV system. As such, the unit (generator) ground fault overcurrent protection (if there is any) can be set very sensitive and fast as it does not need to coordinate with any other ground overcurrent protection. If there was no GSU, or if the GSU had a connection that allowed zero sequence from the 4.16 kV system, then the unit 51G protection would need to coordinate with everything else. Compromises in coordination would probably be needed between the unit 51G and the feeder 51G. You'd typically just use an overvoltage relay on the generator neutral in that case.

 

[PowerfulStuff]

Hi,
With respect to the question of a seperate ground fault CT, or summing the phase CT's I had always thought that the main reason for putting in the seperate CT is that is is more accurate and allows for a lower protection setting i.e.-

- Under system fault conditions transients may occur that saturate the protection CT's unevenly tripping ground fault protection where there isn't one

- The protection CT's (unless specially matched) will have slight differences in output so that the sum is always greater than zero - the GF pickup must be higher to avoid inadvertent tripping

- Finally you get to pick your ground fault CT to suit the design (e.g. 200A current) rather than relying on a different ratio phase protection CT.

Regards.

 

[rbulsara]

Pwrfulstuff:

Yes, it is true for very sensitive ground fault sensing (high resistance grounding). Modern digital relays can sense very minute currents accurately with ordinary CTs though.




Rafiq Bulsara

http://www.srengineersct.com

 

[rockman7892]

Thanks for all the replies guys.

Vandal06

Without the GSU in place to block the zero sequence current, I dont really understand why there would be issues with coordination or comprimises as you mentioned if the generator was generating at the same voltage as the 13.2kV bus and feeders. Wouldn't it just be a matter of setting up the feeder breakers 51G settings below the generators 51G setting so that the feeder breakers would trip first for a fault on the feeders? Or am I missing something?

Waross

I'm assuming that the small indirect fault contribution from the Delta of the GSU is for a fault on the Delta or generator side of the GSU.

Can you explain how a fault on the generator side cause the phase shift that you mention and how the transformer with the wye primary grounded the delta tries to pass current to restore the phases and how with the wye primary floating the delta is free to follow the voltaeg shift and voltage swings. Maybe the root of my not understanding is due to the fact that I have never completely understood why you shouldn't ground the neutral of the wye priamry. If this will add too much confusion to this post then I can start another thread specific to this transformer question.

 

[waross]

A ground fault on the wye side of the wye:delta GSU becomes a single phase load on the delta side and so is not seen by ground fault devices on the delta side.

Phase shift:
The vector diagram of a healthy three phase circuit is an equilateral triangle ABC with included angles of 60 degrees and a center neutral point. If B phase is shorted to ground/neutral, the vector diagram becomes ANC with angles of 30 degrees, 120 degrees, and 30 degrees.

To explain the transformer action, first visualize three single phase transformers connected line to neutral.
Now look at the A phase and B phase transformers as an open delta. (Note to our friends across the water. I believe this is called a broken delta in IEC land.) This forms a virtual transformer for C phase. In practice an open delta may have poor voltage regulation across the open side but that is due mostly to voltage drop on the neutral conductor caused by the current that an open delta draws from the neutral conductor.
Assuming a solid supply, a vector analysis will show that the regulation and impedance of the virtual transformer are the same as a real transformer on C phase. The virtual transformer will carry the same single phase load as would a real transformer on C phase.
Spend some time to get comfortable with the concept of a virtual transformer formed by an open delta. This will help to explain the next step.
Now for reference draw a vector sketch of an open delta. Then, a little to one side draw the real transformer for C phase.
The voltage and phase angle will be equal to the virtual transformer.
Now assume that the transformers have 480 Volt secondaries. Both the real transformer and the virtual transformer are developing 480 Volts at the same phase angle and may be safely connected in true delta.
Now back up a step and consider that for some reason the primary voltage is down 10% on C phase. Our virtual transformer is developing 480 Volts and our real transformer is now developing 432 Volts. Assume that the transformers are rated at 2.5% Z. The total impedance (Z) is 3 x 2.5% or 7.5%. 10% of rated voltage will result in a current of 133.3% of rated current to circulate in the delta. This is a circulating current as it is contained in the delta and may easily destroy a transformer even with no load connected.
If one phase is faulted to ground/neutral, the circulating current will be 58% of the available fault current rating of the transformers.
The primary currents on the wye side will be in proportion to the current in the delta but power will be flowing into the delta from the high phases and flowing back to the fault from the lower voltage phase.
The four wire wye:delta is best used as a ground reference. It is very problematic as a distribution connection.
There are a number of differences between transmission service and distribution service that make the connection more acceptable for transmission service.

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

 

[vandal06]

Rockman,

You are correct, you would just try to set the generator 51G slower than the feeder 51G. That's the ideal case. But what if the feeders are set very slowly and resetting them is not part of the scope? What if the generator neutral grounding limits it's fault current contribution to 200A, but the feeders are set for 300A because of the contribution they see from the utility? It's not uncommon for feeder 51G protection to be set slow and sensitive to clear for resistive feeder faults if they are protecting overhead distribution circuits. Now you have to compromise generator ground protection in order to coordinate with the rest of the system. You'll more than likely need to set neutral voltage elements in order to completely protect for stator ground faults that would otherwise be missed because of this.

That's why it's ideal to have a GSU blocking the generation's exposure to system ground faults, you can set the ground protection as fast and sensitive as the design limits allow in that case.

These are just the experiences I've had with trying to coordinate generator protection that is connected to the same voltage level as the distribution system. But your mileage may vary.