Genset and AVR. Recordings from a load dump? Anyone? 2

[Skogsgurra]

When there's a bad connection between a 1 MW generator and the load (a deformed connection coming lose from vibrations), the fully loaded generator will produce a transient high voltage. Not only from the du/dt, but also from the high excitation. The inductive kick will be of very short duration and the overvoltage will be dependent on the dynamics of the AVR.

What are the typical values? Amplitudes, durations etcetera? Any good recordings showing typical transients and waveforms?

Gunnar Englund

http://www.gke.org

--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.

 

[Skogsgurra]

I cannot believe that there are no recordings from an event like the one described in the heading. Or any experience from such an event.

It is not practical to go and do a full scale experiment. I am prepared to do it, of course, but the owner of the genset does not like the idea.

So, I ask again. Please?

Gunnar Englund

http://www.gke.org

--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.

 

[waross]

No recordings but an anecdote Gunnar. Our little plant in the Caribbean is just over 1 MW. We generate at 480 V and step up to 13200 kV with a delta:wye transformer bank. Single phase switching is common on the 13.2 kV side. Also a full load loss will leave the station transformer connected to the generator. We have not had any known issues with voltage spikes from load dump.
The delta winding facing the generator divides a single phase load on A phase in the ratio of 50% on A phase and 50% of the load on B and C phases combined.
In years past we had serious consumer issues with low voltage burning out refrigerators and freezers when we picked up the load one phase at a time. I got rid of some problematic wye:delta transformer banks and we have not had that issue for years now.
I suspect that with the loss of one phase, the load on the remaining phases will pull the flux down very rapidly.

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

 

[Skogsgurra]

Thanks Bill

The background is this: A large manufacturer of diesel MG sets (possibly the one you have, if it isn't a CAT) asked about a case where one VT had exploded. The VT had been run before with no issues. When the plant was inspected after the incident, they found that one of the connections to the generator was loose and there were arcing marks.

The thinking is that vibrations in the MG set made this loose connection intermittent so that there were repeated inductive kickbacks and also increased voltage due to full excitation when load dropped to zero.

Any thoughts? No recording worthy of the name available.

Gunnar Englund

http://www.gke.org

--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.

 

[waross]

Some random thoughts;
With an averaging AVR the effect on the output voltage may be moderated.
With a bad connection I would first expect the voltage to rise to the breakdown level and then arc over. I would expect the arc to be sustained until close to the next zero crossing and so negligible inductive kick.
When you add vibration all bets are off. The circuit may go open at any point on the sine wave, and more important, specifically any point on the current wave form.
If the vibration opens the circuit at the point of maximum current, the self induction will increase the voltage to try to keep the current flowing, but any load on the transformer secondary will tend to limit the voltage "kick".
Rather than a simulation with a generator, how about back to back transformers with a "loose" connection in one of the leads between the transformers. I leave it to your ingenuity to invent some way of vibrating the loose connection.
Was the PT loaded or are we just concerned with magnetizing current?
Yours Bill

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

 

[catserveng]

Gunnar,

I've been going thru old event files, can't find one from an event like you describe. I should have a bunch becuase we went thru a phase where one of our tail end suppliers did a cost reduction that resulted in using cheaper lugs on the generator cables, and caused a number of generator leads coming off in service.

Most resulted in differential trips or ground fault events. I can look at a couple of older ones, just can't print or move them to anythig useful, in the two cases I can see that I know came from that type event, the voltage rise during the fault was less than 10%, closer to 5%.

All of the units that we had that suffered this problem, and I can remember over 20 of them, all occured on units between 1 and 4.5 MW in size, all driven by recip engines.

The only one I can remember that we lost a single phase that resulted in a high voltage was on a rental set that was driving a material plant for a road contruction project, it started with a blown out contact on the breaker, and the AVR went full field, the last recorded voltage was over 600 VAC, and substantial downstream equipment damage occured. That being said, we relied on the package control panels electrical protection, which in this case were terribly inadequate. It was one time I had lost the battle with our management about putting on a "real" electrical protective relay for a set providing power to a large load, something about the cost was too high.

I have seen a number of generators where we lost a single phase as you described, either a generator lead breaking at a lug, mostly at the generator lead side but a few occurances on the customer connection side, mostly due to poor workmanship by the installing contractor. In only a couple of rare cases like the one described above did I see any indication of resultant damage due to high voltage. I can't think of a case of resultant high voltage event that had a properly applied and setup electrical protective relay, especially if differential protection was included.

I keep looking to see if I can find any event files, unfortunately I left most of that info with my former employer when I left.

Mike L.

 

[waross]

AVR went full field, the last recorded voltage was over 600 VAC, and substantial downstream equipment damage occured.

What was the set voltage Mike? Thanks.

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

 

[catserveng]

On that unit it was nominal 480, and actually set at about 495 to compensate for some line loss problems.

I also should have noted that the bulk of our generator lug problems were on MV machines, mostly 4.16 and 12.47 kV machines.

Mike L.