Detecting Utility Side Faults with Generator Backup Protection in Co-Generation Facility 1

[rockman7892]

This thread goes along with another thread I currently have ongoing but shifts topics a bit so figured I'd start a new thread to discuss.

The issue here is weather or not the generator backup protection 51V/50P and plant distribution feeders 50/51 elements will be able to detect faults at the utility. This application again has a 2MW generator at 4.16kV connected via a feeder with 50/51 to the main plant bus which then connects to the utility service through a main breaker with 50/51. The plant service is coming in through a 22.8kV-4.16kV transformer which I am being told is 10MVA and currently trying to obtain more information as to the impedance and winding configuration. Others have indicated that that there is not a need for a 67 element at the plant's main breaker to detect faults at the utility because these faults will be detected by the generator backup protection as well as the plants feeder and main 50/51 elements.

My concern however is for faults that may occur on the utility side of the service transformers and weather or not these protection elements will see these faults. I just don't see this backup protection being able to detect and disconnect the generator for faults that may occur on the utility system. The 51V element has a pickup of 600A while the feeder breaker and main breakers have 50/51 pickups of 400A and 2000A respectively (assuming that these 50/51 elements will see current in the reverse direction for faults at the utility and don't just see fault current in the forward direction)

When running a quick simulation with an assumed transformer impedance (using 10MVA rating) I come up with a reverse current flow through the transformer of 400A for faults on the utility side of the transformer supplied from the plant generator (ignoring plant motor contribution). Looking at this, it appears that the feeder breaker and main breaker 50/51 elements would not pickup for this fault flow however I'm not sure if the 51V element would detect this fault? Would the voltage collapse at the utility be enough to pickup the 51V for faults on the utility side of the transformer?

Also I'm still waiting to confirm the utility transformer winding configuration but if its a Delta-Wye configuration then the plant feeder or generator backup protection will not be able to see ground faults that occur on the Utility Delta side of the transformer. In this case would a 67N definitely be needed to detect these faults? If the transformer is a wye-Wye configuration then can the generator 50/51N and 501G as well as the plants feeder 50/51N see these external faults?

Maybe I'm incorrect here in assuming that the generator needs to trip for faults occurring on the utility side of the transformer but if it does not, then I see issues occurring when the utility re-closer tries to reclose and the generator is still online.

I appreciate any thoughts.

 

[davidbeach]

If you're looking to separate from the utility for utility faults, you may find that you have better luck with undervoltage and/or negative sequence voltage than using current based detection.

 

[FreddyNurk]

There appears to have been quite a bit of consideration around what to do with non-rotating machine generation in terms of both network protection and islanding events, at least here in Australia. A lot of the reasoning isn't that different for rotating machines, apart from the obvious difficulty in modelling the equivalent fault behaviour of inverter based equipment.

I'd expect that the plant would want to be islanded for a number of reasons, including the issue of reclose. Other reasons would also be for potential to supply part of the network for a utility trip, which ties in with delta utility side and requirements for NVD / 59N, often mandated by the utility. Your requirements may vary.

To be honest, I can't really think of a reason not to separate from the utility for utility faults, although I'm happy to be convinced otherwise. Whether that means tripping onsite generation or islanding though, is a separate consideration. davidbeach is correct that current based detection is probably not going to work very well for detecting utility issues.

 

[Kiribanda]

rockman7892,
As far as I know, normally the Utility recommends to have an effectively grounded network when your generator is syncd
to their system. That means the step-up transformer should be DELTA/ WYE GROUNDED where the Utility side connection
is WYE GROUNDED. Then for a Utility side ground fault a normal ground fault protection scheme can be used. But if the
transformer is having DELTA on the Utility side, then a ground fault will go undetected which is a safety issue.

 

[ScottyUK]

The British ENA produce guidance note G59 which is available at

http://www.amps.org.uk/static/asset...170914updatev1280115trackchangev4.pdf?attach.

I'm not intimately familiar with the document because I (used to) work with big central plant connected to the transmission system rather than smaller embedded machines, but it is fairly readable as standards go. Section 10 covers the protection requirements which operators of embedded machines on the British distribution system are required to comply with.

 

[rockman7892]

I have been doing some reading up on the 51V element and it looks like that in order for it to trip it must cross the generators decrement curve at some point. If there is a fault at the utility side of the utility transformer remote from the generator then the generator terminal voltage may not drop enough and the 51V will use a higher curve than it would for a reduced terminal voltage. If this 51V curve happens to fall above the generator decrement curve in order to coordinate with devices downstream of the generator then the 51V element will never trip for generator faults, is this correct. The 51V curve must cross the decrement curve at some point in order to trip?

Assuming a case where the 51V does cross the decrement curve and there is an external fault at the utility what fault current would you use to evaluate the trip time of the 51V element? Would you simply assume full fault contribution from the generator, and see where the 51V element crossed the decrement curve in order to determine trip time? Or would you used some reduced generator contribution value due to the remote location of the fault?

It appears that if a 51V element were to be used for external faults then it may be better suited to be located at the main breaker.

The more I think about it, and the more others mention it above perhaps the generator is better off being disconnected based on an undervoltage condition for faults external to the plant instead of the 51V element. If there is an external fault at the utility then it is likely that the re-closer will open based on its time delay. Once the re-closer opens the plant will sense a loss of voltage and thus may trip on an undervolage condition before the re-closer tries to close back in at 400ms. Is sensing an undervoltage condition quick enough to remove generation before the re-closer tries to close again?

So my thought/question is why in this case we need to specifically sense ground faults at the utility if the upstream re-closer will open for these faults? Perhaps for cases where the re-closer does not trip and sense these faults? (in which case the generator would not trip on under voltage like I mentioned above). If the undervoltage relay did not trip because of the re-closer opening, then I guess it would need to trip based on the reduced voltage level at the main bus due to the fault?

I will also find out tomorrow the winding configuration of the utility transformer to determine what type of ground fault sensing is needed.

 

[davidbeach]

Probably best to have that discussion with the serving utility. We can give you plenty of perfectly good answers, for other locations, but if your serving utility doesn't go for that solution it won't matter how well it might be accepted elsewhere.

 

[rockman7892]

Davidbeach

I actually have a call setup with the utility tomorrow to discuss such items but I was hoping to learn as much as I could from some of you experts on this forumn in order to see what was typically done elsewhere. As I get answers from people in this forum they tend to lead me down different paths of research (white papers, texts, etc...) on particular subjects in order to gain more knowledge on this particular application. I figure the more knowledge I can gain before speaking to the utility will be beneficial. I appreciate all of the answers I have received thus far.

Any thoughts on my 51V question as its role in this application is fairy new to me.

 

[davidbeach]

I've not seen 51V used for interconnection protection. What works here is some combination of voltage based islanding detection and directional overcurrent based system fault detection. Each system is looked at individually, generation size vs. load size, system strength, interconnection level (primary vs. secondary), transfer trip or not, other things, all play into the final answer.

 

[rockman7892]

From looking further at the 51V relay function I can see now that it must be coordinated with other system protective devices which results in the relay having a fairly long time delay which will not clear fast enough to isolate the generator for external faults on the utility system before reclosing happens. From what I see the 51V element usually trips somewhere above a second or so which is way too long for isolating generator from utility.

I see that that the 51V pickup should be below the generator steady state output current (synchronous reactance) with a 0 voltage restraint (25% pickup) but is there a sweet spot where the curve should intersect the decrement curve? From what I found it should just intersect the decrement curve defined by constant excitation (IF=1pu), not the decrement curve defined by field forcing (IF=3pu). I see that it usually crosses this decrement curve somewhere above 1s.

Obviously the 51V pickup setting is based off of a close in fault at the generator terminals so only considers the generator impedance. If you wanted to evaluate the 51V for picking up at a remote location would you have to factor in the system impedance between gen and fault in order to determine a new generator armature current during a fault? I'm thinking you could do this in order to determine if the 51V relay would even pickup for a remote fault since the new total impedance of gen and system may increase the fault current to above the relay pickup. Are you able to shift the decrement curve to the left to account for remote faults in order to compare against 51V curve?