Generator Reverse Power Trip on synchronizing 9

[edison123]

Trying to commission a 11 KV, 2.5 MW steam turbine driven generator. The generator recently rewound and tested OK. Phase sequence also checked ok.

On synchronizing, the generator trips on reverse power every time. The turbine engineer says the turbine, the governor and the steam valve are ok.

I noted that on closing the breaker, the power factor goes haywire and the Woodwards Multifunction Relay reads about 100 amps in all 3 phases but the power is reading (-) 358 KW and the excitation is reduced to zero. (I will post two photos of the control board meters - before and after synch)in the following posts.

The AVR is Basler. I think the AVR is malfunctioning. Any other possible reasons for the RPR to act every time ?

Muthu

http://www.edison.co.in

 

[edison123]

BTW, thanks everyone for your valued tips/suggestions. My conclusion is that the AVR malfunctioned and killed the rotor.

Muthu

http://www.edison.co.in

 

[catserveng]

Muthu,

If the AVR fails and then results in a failed rotor, I think the protection scheme is not adequate, at least in my opinion. For a loss of excitation to cause a rotor failure means that after the AVR failed the breaker remained closed and the unit now absorbed VARs rather than generates them, do it long enough and the rotor can fail.

It can also be that the rotor failed, and the diodes and AVR are subsequent damage. Do you know at this time failure mode of the rotor?

DECS 15 has a UEL function, depending on how it's integrated into the overall control and protection system will dictate how well it works for acting as a 40 device.

It's not that I'm a huge fan of the DECS 15, and a DECS 200 will give your client much more capability, it's just that I've replaced a fair number of these regulators when they were blamed as being the problem, only to get the system back on line and find the real root cause.

A few years back CAT went thru a huge number of diode failures, and all kinds of things got blamed. In the course of investigating a number of failures, I found that at least on a lot of units I was dealing with, that a 40Q loss of field protection only worked as expected on a total loss of field. We ate a fair number of rotors, and a few stators due to partial field failures and coupled with poor capablity of the installed PM's, we had AVR's sitting at an available full field point, a large leading power factor, still on-line, only to be taken off line by thermal protections or an attentive operator. I had a hard time putting a unit into srvice after that without reverse VAR protection as well, usually at a level equal to about .8 leading power factor. It wasn't the fix for the primary root causes of the problems, but did limit the number of failures caused by subsequent damage.

I like your idea, do it right completeky.

 

[catserveng]

And a star to you for sharing the whole story, nice to hear the follow-up.

 

[dcset]

"BTW, thanks everyone for your valued tips/suggestions. My conclusion is that the AVR malfunctioned and killed the rotor"
No, i don´t think so, i´ll share with you a document about a bad synchonization... and the process, my experience tells me that in a loss of excitation, you have to be some time to kill the rotor, cause the problem is in one way, you can have a out of step, and finally high current incoming or you keep running like a asynchronous generator, and your damp winding will be burned, cause their are no dimensioned for such flow of current, they act like a squirrel case, and for this you need time, and as you said you are out cause of reverse power, but take care maybe you are using relay from GE, and they have one relay,now i don´t remember the name, that measures on the same relay inverse power and loss of excitation, but in any way,
I think about a bad syncronization, but it you do it, and diodes are broken you need to feel a high current coming into windings, the same if you have a out of step,
And about loss of excitation, decs 15 have a UEL (under excitation limit) that senses the reactive component on the current sensed and depending on the adjunted amount of that reactive it give an alarm if its properly configurated, and conected to alarm, cat, talk about decs200, yes this is better cause you can integrate the curves of the generator on the subexcitation side, but sometimes is not so needed, and i use to say that avr is no a protection relay, it´s just an avr, and for some applications decs15 can work good enough, the question is that basler stopped manufacturing such regulator for that is difficult for spare avr.
i hope this could be helpful, and you can cheek one report i do on a bad synchronization.

 

[dcset]

and i just thinking, if generator was rewound, the avr is the same that was working with the generator, and it have to work perfectly,
Maybe the generator come with a problem from the repair shop? maybe on the rotor?
About the diodes, if you rewound generator, maybe you have dameges diodes? open diodes, or just damaged, that is enough for getting voltage, but not enough for manage excitation enough for load....and also, the rotor was damaged when you have to rewound.
Could we think about this???

 

[rasevskii]

Yes catserveng and dcset are absolutely right. The AVR is not a protection system. It may have various limiters and signalling functions included, but in this case these functions had no effect, nothing was tripped.

Whoever set up the generator protections on that project is responsible for the resulting damage. We have still no information on the protections that were provided, and if they were actually tested. Secondary testing only, is not sufficient. Primary testing on open circuit and short circuit is essential using the generator itself as the source. This is absolutely standard test procedure on power station units. I have done this together with the clients on many sites.

In addition, protocols of each test result have to be taken
and signed. Only then can the unit go in service. This is in the interest of all parties on the project.

regards, rasevskii

 

[edison123]

The generator was tested for OCC & SCC and phase sequence at the client's site with a separate DC source for the exciter field. All electrical (field current vs generator voltage/generator current) and mechanical parameters (generator vibrations less than 1 mm/sec peak at 11 KV) were normal indicating nothing was wrong with the rotor when it left my shop. After these tests, they used the AVR for excitation control and built the voltage to 11 KV several times without any issue of excess exciter field current or generator vibrations confirming again that the rotor has no issues.

Only when they tried to parallel to the grid, they had issues of first reverse power trip, then overcurrent trip (when they increased the timing of RPR) etc. and in all cases the PF was going haywire along with negative watts (whatever that is). And after their last paralleling attempt (seventh one), they blew the diodes and the rotor along with it. As seen from the one of the photos I posted above, the exciter current and voltage went to zero on synchronizing before the machine tripped on overcurrent.

So I base my conclusion on these facts. Though I admit AVR is not a protective relay, in a brushless system it must provide some sort of protection for loss of excitation.

Anyway, I recommended to the client that the whole system including the turbine control needs an independent review by a system expert.

Muthu

http://www.edison.co.in

 

[edison123]

dcset - From your report, they synched the machine without checking the phase sequence ?

catserveng - "it's just that I've replaced a fair number of these regulators when they were blamed as being the problem, only to get the system back on line and find the real root cause." - What was the real root cause ?



Muthu

http://www.edison.co.in

 

[rasevskii]

In fact it seems to be that everything was normal when doing the SCC and OCC on the generator itself, using a separate DC source for the Exciter field. This is the normal method, not using the AVR in manual mode.

So it could be that the phase rotation was actually wrong.
In this case on paralleling there will be severe damage immediately. If in fact that was the case they are fortunate that nobody was hurt. But the Instantaneous OC on the stator should have tripped the CB in milliseconds limiting the damage. But I rather doubt that incorrect phase rotation was the case, something would have been wrecked beyond repair.

It appears that the AVR did malfunction, see the above posts regarding cross-current compensation for example.

Was the synchronizing properly tested? A 180 or 120 degree out of phase closing will result in severe damage, the diodes would at the very least not survive.

Let us hope that after the rebuild, better qualified people are available.

regards, rasevskii

 

[edison123]

I personally checked the phase sequence using only the generator PT for both the sources. (Open bus breaker, feed the generator supply to the PT and check the rotation. Open generator breaker, feed the bus supply to the PT and check for the same rotation without disturbing the phase sequence meter connections). I even photographed and videographed the whole phase sequence testing for future rcords.

And then the relay test engineer double checked it with in his own method by using both the bus PT and the generator PT.

Synchronization was done properly, first automatically and then manually.

Muthu

http://www.edison.co.in

 

[rasevskii]

So, then we know that the synchronizing and phase rotation were correctly checked according to normal procedures. That leaves only an AVR problem that sent the excitation to zero, and protection settings that failed to trip the unit promptly, allowing the unit to run as an induction generator for long enough to destroy the diodes and cook the rotor windings.

What did the turbine governor do at this time? Did it open up the inlet valve and attempt to pick up load? Perhaps it closed the inlet valve seeing a slight overspeed due the positive slip.

The root cause is probably the AVR cross-current compensation being backwards or disabled, and the damage is due to too high settings or non-operation of the stator overcurrent protection. It appears that no primary testing of the protections was ever done first. On the next attempts, reduce the OC protection setting to less than 0.2 PU stator current and a time lag of a few seconds only for the first sync. Can be raised later when the unit goes into normal service. Have the client see to it that the tripping circuits actually operate.

regards, rasevskii

 

[catserveng]

Muthu,

I few of the things I've seen in situations similar to yours, a running unit taken down for repair then put back into service,

Faulty diodes, tested ok with DMM but under load failed, have seen this in CAT, KATO, EM and AVK tailends. I'm pretty hard about changing diodes and surge suppressors on repaired tail ends these days, the couple hundred bucks (US) for the parts sure makes it good insurance in my mind. But boy do a LOT of people not do it.

Over the years I have found three instances of connection problems with rotor pole jumpers after repair, one was a bad solder joint, one a bad wire crimp, and a loose bolted connection. All were very hard to find, resulted in AVR and diode failures until they were found. One of these was found doing a pole drop test (initial pole drop test performed at install was good), the other two putting the AVR in manual and watching the field, then doing a physical inspection when data indicated a problem with the rotor.

Numerous times have had the droop CT hooked up backwards, both primary and secondary on reinstall of tail end. Even in a couple cases it was not reconnected at all.

Numerous times bad field and PM connections, usually due to cage type terminal blocks and them being damaged by technicians with improper tools and/or procedures.

Once a small adjustable wrench left in the excitor stator housing (caused intermittant problems for two months), finally made a big enough problem to make a real mess and cause us to look under the cover again.

A few occurances of problems with the pole piece windings, both on new and repaired machines, usually due to nicking wires during wrapping, and problems not showing up for a while. There was a batch of KATO tailends sold to CAT for a project that all generators either failed or showed high vibration after 2-6 months, they said root cause was something to do with a tensioner on a wrapping machine nicking the wire. We swapped all the tail ends and no more problems that I am aware of at that site.

And in a few cases, just a bad repair process by the repair shop, usually everything looked ok initially, but problems cropped up soon after being in service, sometimes right at first close.

I miss field voltmeters in panels, on older sets they were common, and when the unit started and came up to speed you could see the relationship between the field output and the generator voltage. It was also nice to see what the field did right after sync, especially since DMM's and digital controls don't update very fast, and being able to see an analog meter gave you a great idea of what was happening. When I do startups or troubleshooting now I drag out the Simpson 260, not very high tech, but is helpful to me. I have talked a few customers into going back to field meters in their switchgear, mostly small island utilities where lightening strikes and tree branches are common, using the data from the field meter helps indicate if a tail end problem is brewing by a change in field output for a given load.

I can only think of once, in what I felt was a properly protected system, that the AVR was the root cause, it was an early CAT DVR and it went no field to full field after getting to about 1/2 load, it did it so fast that all the operator saw was a jiggle in the amp meter. But the unit had multiple diode failures, then a rotor failure, then I got called in to look at it, when we got it back on line with a repaired rotor and new diodes it was fine, came up on load, operator noted that the "jiggle" was back, had a meter hooked up to the field and took a look, saw the on-off. We replaced the AVR, jiggle went away, and I never went back. It was happening so fast that the 40 protection didn't catch it, the overall system appeared to buffer the transient so it was hard to notice. Even had the tech who had made the previous repairs with me, he said he had looked at the field, he had an older DMM, so we compared my analog meter to his digital, found we saw the analog meter deflect and the DMM only showed a minor change.

Well, hope it helps, and good luck. Remember if it was easy someone would have already fixed it!

Mike L.

 

[waross]

Remember if it was easy someone would have already fixed it!

Amen to that. But fixing the tough ones keeps life challenging.

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

 

[edison123]

Thanks Mike. Your experience borders mine. I still value my AVO analog multimeter.

Attached is the photo of the generator rotor. It's a 4 pole, cylindrical rotor with many turns of round copper enameled wire. No direct access to any of the pole coil connections. Not a great product, imo.

Muthu
www.edison.co.in

 

[dcset]

i agree with catserver, and if you test all the points, something is wrong on generator repair, i mean that when they dismounted rotor for rewinding maybe they hit it or something get injured, and the same as diodes, if there´s a problem on stator as short circuit, is posible diodes get injured and when you measure with multimeter, you don´t fix it, i agree also with catserver to change all diodes.
Loss of excitation problem, i see and the effect in a Stamford generator 2 MW, maybe 6 years ago, was that rotor was ok, but damp windings are destroyed, this is cause the generator was working like an asynchronous one, but for that your need a long time, and about diodes blowing up in syncronism, maybe they are defective, or maybe they blow up on overcurrent, on that overcurrent you get a high voltage on the rotor, and then when diode is broken high current...
And yes, you need to see with a oscilloscope or something like this the current and voltage on the field, cause dec15 when it´s off, you will see zero current and voltage...
I was near some units with no excitation, but syncroniced to the net and nothing occurs, as you are low power, you get a very bad power factor and keep working..
for three or five minutes.
and about my report, yes, they synchronize, without checking the real phase rotation on generator