Sudden Phase Shifts in Supply 3

[ScumPunk]

Hi,All
We've recently had a potential customer at our lab performing some tests on various drives. They are mainly interested in performance/recovery from mains disturbances. This started out as the usual mains dips/brownouts, etc. but as the day progressed, one of the engineers began asking us to shift ALL 3 PHASES by up to 90 degrees instantaneously. I asked how this could happen in the real world, and got a long-winded reply about lightning strikes. I get the impression he was "playing" with our programmable mains source, and having asked a few of our engineers, who are much more knowledgable than myself, have come to the conclusion that this situation is not possible. Can any of you correct this impression?
Thanks,
Mort

 

[Skogsgurra]

I also thought that such a disturbance wouldn't be possible. But I have seen a similar thing. See attached recordings. Still do not understand what might have happened. There were only two VTs, so only two phases recorded. Phase shift is clearly visible, though.

Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[waross]

In the event of a serious fault downstream of the monitoring point, the phase angle of the current will be mostly dependent on the impedance of the transformer and supply conductors. This current will be mostly reactive. If the normal load was running at a leading power factor you could see a phase shift of the current of 90 degrees or more. Expect smoke and fire.

Another possibility is a system running on standby power and a closed transition is made to a much larger source 90 degrees out of synchronization.
In addition to smoke and fire anticipate possible sheared drive keys, broke couplings, bent motor and generator shafts and broken motor and generator shafts. Transformer type equipment may suffer severe transients.
This may become more likely in the future. There is a move to extremely fast open transition type Automatic Transfer Switches for standby sets. The extremely fast open transition avoids issues with protection and utility back feeds and has many of the advantages of closed transition.
These ATSs use a synchro check circuit and inhibit out of sync transfers.
However in the case of either an operator initiated transfer or mis-operation of the sync circuit your equipment may see an almost instantaneous phase shift of the supply of 90 degrees or more.
The wisdom of testing with a 90 degree phase shift would depend on the nature of the equipment under test and the intended market.


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

 

[ScumPunk]

Hi, Skogsgurra,
That is very interesting. Looks like around 100degrees shift? I'm sure the guys here will find it very interesting and puzzling when I show them. The other day I refused to spend time creating this transient. I actually said "show me a plot of it happening for real and I'll reproduce it". Don't I look a fool now...
Anybody out there know what could cause this?
Thanks,
Mort

 

[ScumPunk]

Hi, waross,
The potential client was looking at various VFD's. Some have conventional (diode) font ends, some have full active front ends, and our newer drives have a kind of hybrid AFE. This disturbance only caused trips with the full AFE drive.
The application is a pressurised production area making enzyme based drugs, in UK. They claim that they have problems with brownouts/blackouts which, with the drives they have at the moment, sometimes cause them to lose a lot of product (=money). The reason is that the VFD's there at present take too long to reboot/start on fly to keep the pressure up. I believe we have demonstrated that a modern drive improves this situation, but then we started getting asked about all sorts of other problems..
One of these guys is doing his thesis on VFD's and he said this was the first time he had used a drive, and he was quite obviously impressed with our mains source. As I say, I got the feeling we were actually doing his Phd work and not the original request.
Thanks,
Mort

 

[waross]

Gunar, am I correct that the third trace indicates a reversal of phase rotation? Is it possible that it is a metering glitch and nothing actually happened on the lines? I am thinking that the scope or recorder may have suddenly inverted the traces, or something of that idea.

ScumPunk my first thought is "Can you support this Phd candidate and benefit from his research?"

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

 

[Skogsgurra]

No, Bill. Nothing like that. This is the 70 kV distribution tied to the Scandinavian power grid and no local generator that gets hiccups.

The reason we monitored the 70 kV is that this site had a problem with fuses on CSI type VFDs for pumps. Monitoring was on for about three months during summer. Mostly to see what happened during a thunderstorm. The three events shown in my picture are the only ones we got for those three months. There was no thunder for the whole period, which is uncommon and verifies Murphy's Law in a backwards way.

The recorder is a Yokogawa memory recorder and it was set to trig and output a strip whenever the voltages got outside a +/-5 percent mask. If you zoom in on the anomaly, you can see that the millimeter grid pattern is unaffected so the possibility that the recorder feed has gone wrong can be ruled out.

Also, the waveforms do not change in microseconds but milliseconds, so a measurement glitch seems to be out of the question. That it should occur on both channels is not probable, not even possible if you study how the recording system is built. And, it is clear that something "physical" has happened. Something that "falls back" again after about eight milliseconds. Perhaps a huge load being (short?) connected and tripping after a very short period?

Have been thinking of a compensating capacitor being switched in. But that usually starts heavy ringing. And that can not be seen here.

More tips?





Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[davidbeach]

Gunnar,

One possibility is that a single phase capacitor bank opened at that time, but I'm not sure that can explain the full amount of the shift. The lack of ringing might be explained if the capacitor were far enough away as the impedance at the ringing frequency is very high and the ringing is damped out with distance.

I doubt it was load because what was the leading phase becomes the lagging phase and any load large enough to make a noticeable phase shift - and not nearly that much - should also affect the magnitude of the voltage, at least temporarily.

 

[Marke]

Looks to me very much like a very fast network switching to another source with the phases rolled round.

Best regards,

Mark Empson

http://www.lmphotonics.com

 

[Skogsgurra]

I will have to ask the guys at the utility about this. Wonder if they have any records from 2002? Or they may be able to tell what could have happen. If it is possible at all. I mean, switching 70 kV like that.

Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[davidbeach]

Switching 70kV is no big deal. We feed many of our distribution substations at 115kV, others at 57kV, and switching operations can make all sorts of changes. We wouldn't, though, switch only one phase of a capacitor bank.

 

[Skogsgurra]

Hey Scum,

You somewhere in Shropshire?

Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[slavag]

Hi.
Intresting graph. Need more information about this 70kV line.
Fast switching to alternative source, possible. Gunnar you need ask your utilities, what type of ATS them used.
I would like add graph of switching (6.6kV bus)to other source.
Regards.
Slava

 

[Skogsgurra]

Hi Slava,

I tried to reach guys at the utility. But they didn't work Saturday nights and Sunday mornings! Where is this World going? ;-)

If I interpret your recording correctly, it is a case where you got such a high load that the frequency starts to drop (island?). When it has dropped to just under 25 Hz, that branch is disconnected and another supply is switched in.

This is probably a very good wavform to program into Scum's (very impressive) mains disturbance simulator. It will make that thesis more interesting. That's for sure.

Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[Marke]

It could also be a classic case of a supply disconnect with a high motor loading. The motors act as generators and as they slow down, the frequency and voltage drops.

Best regards,

Mark Empson

http://www.lmphotonics.com

 

[Skogsgurra]

I am not so sure, Mark. A disconnect usually looks less "wild".

I did some "document excavation" and found this. It is from year 2000, so it took a while to find it.

There is a clean disconnect of the supplying grid, but there are some motors running at the farmers in the neighbourhood. The motors continue to feed the disconnected part of the grid with decaying rotor magnetization. It is over in a few cycles.

The sudden drop at disconnect time is typical. It reflects the impedance in the motors that are now running as generators.

Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[waross]

Hello Gunnar. I guess you realize that we were posting at the same time and my first post was not a comment on your first post. Were your voltages measured phase to phase or phase to ground? The first trace looks like an event that took the voltage down to about 50% for 70-75 millisecods. It may be explained by inadvertently switching in and out a bolted fault.
Or, not to forget the human factor, someone with very fast fingers could have switched the recorder to a different scale and back. If that was a disturbance on a 70kV feeder it is almost certain to have also been recorded in the station. I hope you are successful contacting your fiends at the utility on Monday.
The third trace seems almost impossible. It brings to mind Matthew 19:30 "But many that are first shall be last; and the last shall be first."

(King James Bible, Matthew)

The cause and effect of such a phase reversal would almost certainly become part of the folklore of the industry. That the phases could be smoothly reversed and not create massive system disturbances would be very unusual.
I note that the traces reverse instantly and within a few milli-seconds have stabilized.

I am not familiar with your recorder, but on the old oscilloscopes that I have worked with I can easily visualize the following discussion:

"That recorder has been sitting for weeks and has not done anything. Is there any way we can check it?"

"Well, it responds to system upsets. Try quickly switching it to a higher range and back."

"Wow, it is working!"

Several weeks later.

"That recorder hasn't done anything for weeks again. Are you sure it is still working?"

"Gosh, I don't know, but after the commotion the last time we tried to check it, I'm not going to change ranges again!"

"Yes you are right. How about that other switch that switches channel A and B?"

The disturbances may be reconciled with contact bounce in a mechanical switch.

If you have access to the recorder you should be able to check this out on a 220V circuit.
I'm not challenging you Gunnar. You have more information on this than I have. But, whenever I forget the human factor it bites me in the butt.
Yours

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

 

[slavag]

Hello.
Gunnar, is not O.K., why they didn't work on Sunday. I'm worked today and they not.
Mark, you are right, my graph is part of testing motor bus transfer ( sorry I dosen't wrote it)system. This voltage is voltages produced by several 6.6kV motors. According to our tests those motors (fans) keeped voltages on the bus up to half second.
Bill, same discussion I hear today near to new recorders also.
Regards.
Slava

 

[Skogsgurra]

Sorry folks.

I did some more digging and found the report as well. The two VTs are connected between phases on the secondary of the transformer. So they show the 10 kV. But there is no breaker between transformer and VTs.

The transformer is a 63 MVA unit with 11.2 % uk and there are many more loads connected to it. DOL pumps and also a synchronous compensator. The latter I do not know the size of, but it is probably in the 5 - 10 Mvar range. Most compensators that I see are about that size.

Picture shows the part of the 10 kV system that feeds the inverter for the pump.

The recorder used was an OR 300, description can be read here

http://www.dqm.it/pdf/yokogawa/REGISTRATORI_OSCILLOGRAFICI/OR100E-300E_BU7840-20E_010.pdf

There is no way of changing range or channels when it is recording, which it was when all these recordings were made. To do that (change channels or ranges, you have to stop recording first. And then it doesn't record. The time stamp above the recordings show when the triggering took place and are proof that no human factor was involved. The recorder was in an electrical room (10 kV) with very limited access.

I may be the human factor, myself. But, in this case, I was about 250 km away, so I am not so sure about that either. I will try and contact the guys tomorrow.



Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[slavag]

Gunnar.
Please check with utilities:
1.Are this line have 3-ph autorecloser and what is a time of first autorecloser?
2.Are this autorecloser include synchrocheck?
Regards.
Slava.