Parallel Generator Solid Ground Circ Current

[KBA1]

I have read the thread "Multiple Resistor Grounding" and have gleaned good facts. Big difference here is solid grounding. Perhaps after discussion, solid grounding will be ultimately the thing to go...

The power generation is islanded, consisting of three identical 2/3 pitch generators each rated at 200 kW, 600 V. We have a 3 phase, 3 wire system (no 347 V loads). There will be times when two of the generators will co-generate, sharing the load equally between them. Additionally, VArs will be equally shared using a reactive differential compensation scheme.

The grounding of the generators' star points will be solid; i.e. no intentional impedance. In the past, a current transformer in the star point of each generator has led to erroneous indications of "ground current", due to the voltage on the two online generators being different by a small amount, causing a circulating current to flow. Short of installing impedances to minimize this current flow, does anyone have any suggestions to minimize/stop this flow or to relay appropriately so as to not get false indications? Suggestions have been varied, like tying neutrals together and then going to earth through ground fault ct (one ct source feeding three relays), and also using 3rd harmonics present in circulating current to discriminate from “true” ground faults.

Client has used solid grounding and is reluctant to change unless compelling reasons are provided. Generator manufacturer is providing bracing suitable for solid grounding.

Any ideas/suggestions?

 

[gordonl]

Instead of tying the neutrals together the CT's could be summed then fed into one GF relay versus two. Circulating currents would cancel out because they would be in opposite directions, but for a real ground fault they would go in the same direction and add up.

Perhaps a more important question here is generator heating due to circulating currents in the neutrals. With solidly grounded gens the circulating currents will be higher, thus higher temp rise.

 

[peterb]

Ground only one generator neutral at a time through a suitable disconnect switch, as stated in the reference thread. This will stop the circulating current. Consider the use of directional ground fault relays on each generator to provide sensitive ground protection - this in addition to the higher set non-directional relay scheme for distribution system ground faults.

Usually, the generator system is resistance grounded to minimize generator stator core damage due to internal ground faults. Iron burning will require a more extensive rebuild, as laminations will have to be replaced. I don't think that the generator manufacturer can protect against this phenomenon.

 

[dpc]

Solid grounding is common for small generators such as this.

The other option is some form of high-resistance grounding, which is sometimes used at 600 V and below.

The neutral currents you have seen may be due to third harmonics. Newer generator relays such as by GE and SEL will filter these out, either using an actual hardware filter or in the firmware for the relay.

Your settings for ground fault protection must be high enough to allow for some neutral current.

A common grounding connection for the three generators may help reduce the circulating currents. But for generators this small you need to make sure that the neutral is not already grounded inside the machine.

 

[KBA1]

Gordonl - Given that each of the generator breakers will already have a dedicated generator relay (incorporating a 50/51G), would routing the "summed" ct circuit through all three generator relays be reasonable? The burdens on todays microprocessor based relays will not be a problem, and since two generators will be online at the time of a fault, we will need to get both generators off-line if the fault is on the bus.
Does anyone know whether the waveform (ie harmonic content) will be different for circulating current versus (genuine) fault current. I know the magnitudes would obviously be different, but as far as discriminating on the basis of 3rd harmonic content, is that an option?

 

[peterb]

Discrimination on the basis of 3rd harmonic is certainly and option for the ground fault protection, but that doesn't stop the circulating current from flowing.
I repeat that the only way to do that is to ground only one neutral at a time - connecting all the neutrals to a common ground only means that the circulating current flows through the interconnecting bus rather than through the ground connections.

 

[dpc]

For peterb:

No argument from me that your solution will stop zero sequence currents from circulating. I was just trying to point out that many similar systems are operating successfully with each machine grounded and that there are ways to deal with circulating currents.

Since these are low voltage machines, low-resistance grounding as typically applied to medium-voltage systems cannot be safely used. High-resistance grounding could be used and this would also basically eliminate the zero-sequence currents.

dpc

For a 200 kW machine, use of directional ground relaying may not be justifiable economically.

 

[gordonl]

I would think a cleaner solution would be one relay tripping all three generators on a ground fault.

As to the circulating currents, please correct me if I'm wrong, but circulating currents due to a small difference in generator windings would be fundamental and not third, so harmonic restraint would not work.

 

[dpc]

to gordonl:

Regarding the circulating currents, the original post indicated that the generators were identical. I believe that the third harmonics will flow regardless of the design of the generators, as long as there is a current path. This occurs even when there is only one generator connected to a wye-connected tranformer.

If it is acceptable to trip all three generators for any ground fault on the system, it might be preferable to use high-resistance grounding to minimize damage due to ground faults.

 

[Electrifier]

Dears,
Considering circulating currents, amount of circulating currents should be negligible if droop and voltage level of AVRs is adjusted in same way. Amount of circulating currents should not be more than 10% of running current at any generator. Ground fault relay should must be adjusted more than this value.
Best way to protect against accidental tripping, which is recommended by various generator manufactuurers, is to ground only one generator of parralleled gennerators. neutral links of others will remain open. This will not cause the shutdown of system and will trip only one generator. After remedial tests you may normalise it.

 

[ongchazira]

Electrifier,
I feel adjusting the droop/voltage levels of the AVR will only affect the VARs sharing of the machines, but circulating current will still remain. I think that the neutral current here is mainly due to the slight unbalance in the voltages due to the winding of the machines.The possibility of third harmonic exists if there is a wye connected transformer in the system with its neutral grounded. The easiest way of avoiding this current is to ground only one neutral as suggested by many of us. This will avoid overheating of the generators too. As far as the protection is concerned, I think that all the generators should be tripped if the fault is on the bus to which all the machines are feeding. For downstream faults, proper discrimination is necessary.

If there is a wye connected transformer then only one ground of either transformer or generators should be grounded.

 

[hinrg]

KBA,
Have you reviewed the use of cross current compensation CT's. The output of these devices will feed directly into the AVRs. This will help in regulating the droop and voltage outputs of the machines, thus reducing the circulating currents.

 

[KBA1]

Thanks to all that have contributed to this thread. It has given me insight to several areas that I had not considered before, and as such, I guess I can say the thread has been a success to me personally.

I will proceed with solid grounding of the generator neutrals. The winding pitch, being 2/3 will significantly decrease any third harmonic circulating current and the AVR's will use reactive droop current compensation (not cross-current or reactive differential compensation).

We will be conscious of the magnitude of potential circulating currents, which will be measured during testing and settings of ground fault elements adjusted accordingly. As all generators will be ground fault contributors to any common bus faults, each relay will trip off it's respective generator; by contrast, if the fault is in the distribution system, downstream relaying/breakers will clear any faults. As many of you know, any attempts at "coordinating" successive downstream instantaneous relays in a single voltage system doesn't tend to work, so disabling of the 50G will likely be the norm.