Distributed Generation Anti-Islanding 3

[redfurry]

I'm currently involved in evaluating a distributed generation proposal for a utility. It includes 600kW of 1-ph synchronous machines connected to a long 1-ph line at 16kV (p-n). This line is protected by a 1-ph recloser.

The gen owner is proposing the normal set of protection (overcurrent, under/over voltage, under/over frequency).
For anti-islanding, as required by IEEE-1547, nominally the installation should provide transfer trip to the local recloser as the size of the generation is around the same size as the load supplied from the recloser.

The customer is instead proposing to set the generation to run at unity power factor and to use reverse VAr sensing to trip the generation if any VArs are output from the generation to support local loads (normally the VArs would be provided by the distribution system).

Has anyone seen reverse VAr sensing used for anti-islanding?

It seems like a simple approach that avoids the need for transfer trip capabilities. Can you see any drawbacks to this approach?

 

[rasevskii]

It would seem, at first glance, that the recloser cannot be used onto a line on which synchronous machines are connected. If there were to be a trip and reclose (such as a lightning strike) the generators would be reconnected out of sync, with damage resulting. There would have to be an undervoltage trip and lockout on each generator to prevent them being reconnected out of phase. Perhaps then an autosync after a suitable timelag when line voltage has returned.

Others invited to comment here...

regards, rasevskii

 

[waross]

You will have to set the VAR detection high enough to alow for VAR transients such as motor starting or transformer energisation. This setting may be too high to detect reverse VARs at light load conditions.
Furthermore, the utility has the final say on the power factor that you operate at. This can change. Should the utility demand that you export VARs as a condition of renewing a contract, you will be dead in the water.
Run the machines in droop mode and use over frequency trips as most others do.

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

 

[redfurry]

Thanks for the input.

rasevskii
- the machine protection is supposed to operate before the recloser closes back in (i.e. machine is dropped if the recloser operates)

waross
- in this case the generator is obliged to operate between .9 lagging and .95 leading (the utility doesn't actually have a say in this according to the regulator) so planning on operating at unity PF is an option.
- I was under the impression that over/under frequency protection was inadequate for anti-islanding protection (in cases where the islanded load may be similar to generator output)
- I like your point about VAr transients. I'm guessing it wouldn't be easy to quantify how large these would be.

 

[davidbeach]

If the generator is operating in Watt/VAr control or in Watt/PF control, you won't get an export of VArs anyhow. If it should be supplying VArs and isn't then you will get a voltage collapse. If the generator is operating in a voltage control mode, then you would get the VAr transient/export when you separate from the utility but you wouldn't be running at unity power factor except by happenstance.

 

[dpc]

I think there are lot of possible scenarios where this won't work very well. Certainly not as reliable as transfer trip.

It would be necessary to install voltage sensing at the recloser to block reclosing if the voltage on the load side of the recloser was more than a few percent of nominal.

Undervoltage tripping of the generator will probably be more reliable than "reverse" VAR flow detection (it isn't really reverse) for line faults.

David Castor

http://www.cvoes.com

 

[jghrist]

Even if reverse var tripping worked (and I agree with davidbeach that it wouldn't), it wouldn't be fast enough to prevent reclosing while still energized. A combination of frequency and voltage relaying should handle the islanding protection, but this won't be reliably fast enough for reclosing either. The utility needs to have voltage sensing at the recloser, as David said, to prevent reclosing.

 

[slavag]

Possible use another anti-islanding protection functions, or LOM-loss of main:
ROCOF-df/dt
voltage jump
U1< PPS undervoltage.
Of course option of voltage sensing at the recolsere is very good option.
Best Regards.
Slava

 

[rasevskii]

How does each machine sense that the recloser has tripped? There has to be a transfer trip signal from the recloser to each machine to insure that that machine is actually off. Then the recloser can reclose, if no voltage exists as said by others. An exotic communication system is required...?

I suppose that these are some sort of biogas units in a rural setting in view of the single phase line. An interesting challenge indeed. The timing on the reclose is usually less than 2 seconds, not time enough for the frequency or VAR protections oo each unit to clear, as said above.

Why is the supplier afraid of supplying VARS? They are almost free, actually. Just have an AVR working on droop as suggested.

Tell us more...

regards, rasevskii

 

[redfurry]

Thanks everyone. I appreciate all the input / thoughts.

rasevskii
- good guess! - it's rural biogas

- The one reason to run at unity power factor is that it prevents excessive voltage rise at the generator end of the feeder. If they export VArs it tends to increase the voltage to the point where there is overvoltage under light load conditions or the potential for flicker issues as the generation starts/stops.

davidbeach
- good point. I'm not sure what control mode they have in mind for the generators. I'll find out.

slavag
- those are certainly more conventional options for anti-islanding.

The generator claims they can set the undervoltage/over-frequency/under-frequency to trip in 0.4 - 0.5 seconds so the claim is that this will take care of recloser timing issues. I agree with the comments that this is questionable / potentially unreliable.

I think they threw in the 'reverse-VAr' sensing as anti-islanding so they can try to avoid the cost of a tranfer trip arrangement.

I think I'll go back to the engineer for the generator and have a chat.

 

[dpc]

Without transfer trip, and maybe even with transfer trip, I think voltage sensing at the recloser is necessary.

Or eliminate reclosing.

David Castor

http://www.cvoes.com

 

[Marmite]

Another option could be a fault thrower on the load side of the recloser. Recloser trips, fault thrower operates and puts an earth fault on the line which is then detected by the generator earth fault protection. Crude but cheap. Not sure if you use fault throwers in the US but its widespread in the UK in rural areas where the cost of an intertripping channel is prohibitive. Normally they close and stay in, but in your application it would need to reset in the dead time of the recloser.
Attached is a useful document on anti islanding and a discussion on the advantages and disadvantages of each type of protection.
Regards
Marmite

 

[redfurry]

Marmite - thanks. That's an excellent paper on anti-islanding.

 

[rasevskii]

Another point that has to be looked at, is how the protection systems at the generators actually operate. I suppose that these are small engine-driven gensets which in turn are controlled by an integrated controller from Woodward or Basler, for example. Programmable with many inputs/outputs. Is the auxiliary supply from battery or AC, and what happens to the controls when the AC line goes dead...

Perhaps there should be an independent overcurrent/undervoltage/etc electromechanical protection to clear the generator off the line if it gets really ugly, electrically speaking.

Our friend Catserveng hopefully can add some input here...

Has the genset supplier done this before somewhere and what were the results..

rasevskii

 

[catserveng]

When you say "600 kW of machines", what kind of a mix are we talking here? 2-300kW machines, 6-100's, 10-60's, 20-30's, or what? Since you stated they are all single phase I'd expect several smaller machines, sounds like a bunch of small dairies on a rural line.

Does the recloser have a sync check function?

Do the machines have to be synchronus? Do they need to be able to island or run isolated? Is there a future requirement for the units to provide voltage support? Can they be induction machines instead?

What kind of governing? Will they have real power control or small units leaned against the grid in droop?

What kind of voltage regulation? Does the regulator have a built in PF/VAR control mode, will it operate in voltage droop, or take an interface from a PF/VAR controller?

I've done a fair number of this type project, never found a one size fits all solution. Unit size, what kind of installed controls and protections, site load requirements, line voltage stability, fuel quality, maintenance practises, utility requirements all had impacts and each site ended up with unique problems. Sorry for all the questions, but I usually ended up on projects like this after they were installed and things didn't work and I was looking at a piece of damaged equipment (and a gang of upset people).

Great discussion, and you're getting lot's of great comments, but without more details on what you're putting on the ground and what you actually expect it to do, it's hard to provide a precise solution.

Marmite, great paper! Wish I'd had that a few years ago.

 

[jghrist]

I don't think the expense of transfer trip is justified. The recloser can tell if the generators are still on line by sensing the voltage. They could be set to reclose as soon as the voltage goes away.