Generator breakdown

[Denan]

Hi all,

A generator was damaged few days ago, and I received some pictures showing few curves at the moment of the failure. I am not sure what the cause of the failure is/could be, so I wonder if someone out there might have seen similar before. There are few more curves I have but I think these are best visible (I got a scanned copy)

The generator voltage is 10,5kV, and it was running parallel with grid.

Stator windings are damaged, rotor seems to be ok.

We are planning to have an inspection to try to fing the cause but I don't know when that might be.

 

[Denan]

Hi,

I wonder if the other breaker (grid breaker, utility breaker) might have some faults, because in log of prot. unit, we could see that this breaker did change status many times lately (before the fault) which is not normal there. Also, several times, there was an alarm we call Grid breaker fault, also before the fault. It might have been that the breaker did not change status after opening while engine was running, causing high voltage on the generator (AVR is in PF mode when parallel with grid).
Also, several times before the fault log shows High voltage on the generator..

Unforunatelly, the generator is already on the way to some workshop for dismantling. But the breakers are still there, I think that they are ABB make.

But I have got few pictures from site:

http://files.engineering.com/getfil...-4d59-baf1-7563a65d2fe9&file=all_pictures.pdf

 

[Denan]

Here are logs from both protection units (generator and grid breaker)

http://files.engineering.com/getfil...-23cf-4fb3-a7a8-f55b76aa39c1&file=Gen_log.pdf

http://files.engineering.com/getfil...9a3c-492e-ae95-bc80ac49a678&file=Grid_log.pdf

 

[rasevskii]

It would appear that there was an earth fault from the damaged coil to the core laminations at the coil end, resulting in some splatter onto the surroundings. The stator however, does not look particularly dirty, only a normal dust accumulation.

The protection should have tripped on earth fault (apparenly did) resulting in minimal damage at the point of fault in the winding.

It would seem that the control system has some/many issues. The breaker (grid breaker) may have been giving false indication due a faulty aux switch on it.

The PF control has to be only enabled when the Grid Tie is closed, otherwise the voltage may/will go too high if the unit is actually in PF control due the false indication from the grid tie breaker which was actually open (the unit in island operation.) However, there should be an overvoltage and under/over frequency protection to trip the unit in this case.

What is not clear in the recording is why there is still stator current even though the excitation had tripped. Where are the CTs located for the recorder input? If the gen brkr had actually opened, there would be no possible stator current. It may be that the unit stayed on line without excitation. The further oscillations in current seem to indicate this.

There may be a lot more damage as well as blown diodes in the rotating rectifier.

What is the OEM for the protection and what relay type? (unless that is better left out of the discussion...)

just some ideas...

rasevskii

 

[rasevskii]

BTW, the many strange indications in the SCADA report may be due to simple loose connections giving a momentary false input. If the unit has been running OK for 6 years, vibration, etc may have caused bad conns to develop.

Particularly in the AVR sensing circuits, loose PT secondary connections or bad contacts on the HV side in drawer type PTs can drive the excitation system wild, and it will happen when nobody is around to see it...

rasevskii

 

[potteryshard]

I don't know what model ABB breakers your are using, but I've had one fail to open on one phase. (A week later, after the plant fire was put out, while several of us were staring at the breaker, it finally popped open.)

 

[Bramb]

Hi Denan,
unfortunately I'm not able to get any suggestion by the SCADA logs...

As nothing was found in protection's alarm log, I'll start to consider the possibility of a fault which has occurred after generator CB was open while the machine was de-excited (but with residual flux) and starting to decelerate....

Ciao
Erminio

The difference between overload and short circuit lies in the nature of the fault, not in the value of the current.

 

[Denan]

Thanks guys for you coments!

From the log it looks like both Gen. and Grid CB had multiple operations, apparently without anything initiating it. So, it may be that the generator got the damage on 03.sep. at 18:58, according to the Gen log. Breaker closed at 13:40 and did not open until 18:58. Protection unit reacts at 13:40:55 with low freq. and low volt, as well as grid CB fault, but the breaker did not open.

Note also what happened 02.09. at 14:53 and again 15:07 - Phase sequence error. For us it is not clear why this happened.

On 02.09 there was no CB OPEN until late in the afternoon but there were lines without voltage present, in which case it would normally be without generator running. Can it be that the breaker closed when engine was not running?

Rasevskii,
We are also thinking of damaged diodes, since that excitation rose high while voltage was not too high.

CTs are mounted inside the generator, near the neutral bar.

potteryhard,
I don't know what type of breaker is used there, assume it is ABB but not sure.

 

[rasevskii]

Denen:

CTs mounted at the neutral bar (neutral side of the stator winding) may well account for the current that existed in the stator after the Gen CB opened (if it did), that was current flowing from the earth fault in the stator to earth over the neutral. (possible scenario)

It really sounds like the control equipment is faulty in several ways and the protections also. If we can believe the logs.

Was the equipment ever properly commissioned and tested? Are there test reports and signed docs from a consultant or the power company? It is very unlikely that the power company would allow any connection to there system at the HV level if the equipment had never been properly tested.

This will be the case when the unit is recommissioned as well.

Is there a PLC involved that is suspect?

rasevskii

 

[Denan]

Equipment was tested during commissioning. If it was tested after that again, I am not sure. Have no idea about those reports etc. you mention.

I have already asked my colleagues at local office to ask the company that installed these breakers to check them all for possible faults. We will probably have to do the same with our part after the generator is brought back.

I don't think there is any issues with a PLC, there is one, but it does not do anything directly with interface with the breakers.


By the way, I forgot to mention there is apparently no earth fault relay in the system (at least on our side). On similar plants, we have broken delta detection, but I can not see it on the SLD from this plant. We should maybe add it, right? Generator neutral point is not connected to the ground.

 

[rasevskii]

Denan:

I assume that you are the OEM for the genset, and the switchgear and remaining plant is from "others". Always a sticky situation after a fault has occurred. The question is mainly about the relay protection for the unit, did you supply it or "others"? In any case the protection has to be retested, including all tripping and alarm functions, and it is in your interest to get a signed protocol from these tests.

On the earthing, the generator neutral is unearthed? (is this true), in that case a bad idea. The 10.5 Kv system must be resistance earthed somewhere, at the (unshown) grid tie transformer neutral?

If the gen neutral is left unconnected, then the only earth point is over the wye connected PTs, with broken delta additional secondaries conn to the earth fault relay to give some degree of protection before the unit is connected to the grid. (from your explanation, I think this is what you have?)

Look up some of the threads on this Forum regarding neutral point earthing on industrial systems.

rasevskii

 

[Denan]

Hi,

yes, we supplied the genset, and also control system, including protection. Not the switchgear etc..

About generatro beeing unearthed, that is common in the Netherlands, we have several plants like this one, and none of the generators are earthed in any way (only PE of course). But as you say, usually thafo neutral point is earthed, but in this case I have no idea where the trafo is located, how far from the genset.

We don't have any earth fault relay in our system, I realise that we probably should have it, but that is how it is now. I see that broken delta PT is available (shown on the SLD). I will try to find out if others have it in their scope.
On the later plants we had earth fault relay from broken delta PT (either we or the others).

 

[rasevskii]

The broken delta PT connection connected to an E/F relay will detect an earth fault in the generator, but also an earth fault anywhere on the 10.5 KV system, in other words it is not selective, any may result in spurious trips of the generator due an E/F somewhere else.

A possible solution would be to disable this protection after the unit is online to avoid this situation. In that case the generator would be protected against E/F only by the protection elsewhere in the system.

A protection study of the whole system should really be done to co-ordinate all the protections. It is likely that the 10.5 KV system is resistance earthed at the neutral of the supply transformer that has to exist to connect with the utility system.

It is quite possible to get by with uncoordinated or not-thorughly-studied systems as long as nothing ever goes wrong or no fault ever occurs...

Since you are not telling us what kind of industrial situation this is, we can only make some educated guesses from experience elsewhere with cogen and industrial generators connected to large grids. In other words, the more large, extensive, and complicated the 10.5 KV system is, the more difficult the situation with neutral earthing and earth fault protection becomes.

rasevskii

 

[Denan]

It is a CHP plant, used for greenhouse (is it called like that?)..

I have asked to find out more about E/F elsewhere in the system, no replay yet.

 

[Bramb]

Hi Denan,
the public utility, which the plant is connected to, should have requested to install some protection functions on the grid breaker which are needed both to protect the plant from grid faults and to protect the grid from plant internal faults.

Both actions are done by means of opening the grid breaker:
- in the first case (trip for grid faults) the plant will keep on running in islanded operation
- in the second case (trip for internal fault), the trip of the grid breaker isn't sufficient to clear the fault: trip of the generator is necessary.

In order to work correctly in the first case, the grid breaker protections, requested by the public utility, should be faster than internal protection.

On the other hand, the grid breaker protections aren't sufficient to clear internal faults, unless they are equipped with two stages: the faster one will be used to trip the grid breaker and the slower one to trip the generator.

This solution, in my opinion, can be used (or maybe is already used) in your plant to achieve selective behavior in case of earth fault as the generator isn't equipped with stator earth fault protection.

I hope this can help.
Ciao
Erminio

The difference between overload and short circuit lies in the nature of the fault, not in the value of the current.