Failure to Latch in on Synchronous Idle

The crystal ball is fogged today.....

1. Yesterday, what trip stopped the set?
2. What does "Today no dice" mean? Did it fail to close breaker or trip after close?
3. Is there a "check-sync" relay?
4. Can you close OK manually, using synchroscope, with very low slip rate?

Basic thought is a) Auto sync is doing its job and some fault is tripping after closure - you imply the synchroscope showed zero angle offset at close; or B) The auto-sync is connecting out of phase or unmatched volts.

You need a recorder for the beat waveform to check the ASU although if slip rate is slow & synchroscope shows 12 O'clock at close then ASU appears OK.

67model

Answers:

1. The "trip" on 9/11 apparently was momentary, and it appeared to be in the latch-in relay that only briefly went into the proper state to tie into the grid, then dropped out. Operator intervened and was successful getting the unit online for the dispatched deadline.

2. "no dice" meant that no attempt to latch up online was successful on 9/12/16, it simply failed to close the breaker. At some point, during the second or third attempt in the afternoon, XND ref on the gas turbine was not updating (that is a problem). Also a high diff. press. was found on HVT only on one phase. Substation group will be investigating that today. During circuit troubleshooting, I&E found that the panel switch was showing "made", but the relay ("latch" relay) was not changing state (stuck?). It made no difference whether latching in was attempted in auto or manual. At some point also, there was a condition on one of the field relays (safety relays) that was addressed, but this had no effect on latching in. Finally, there was a hard trip (not sure why), and the entire control cab went black, no power anywhere on the unit. This morning is troubleshooting mode for all hands.

3. Yes there is a "check sync" relay, synchroscope, etc.

4. As stated in #2 above, no difference in results between auto sync and manual sync. We seem to be pretty sure one of the problems is the relay between the panel switch and the field breaker. Not sure at this point about the hard trip, or the loss of aux. power to control cab. Everything still in the dark this morning.

The field and generator voltages appear to be aligned well within tolerance during attempts to latch up. There were no ground fault alarms, no other fault alarms, and no indication of the latch-in relay firing during manual or auto-sync. We are very suspicious of the relay.

I think having the beats waveform for the ASU is a great idea. I may drop that hint to I&E, to see how they respond. Bear in mind, I am only the chemist, and have no authority on the matter at all, just trying my best to be of service.

Stuff is still missing, and your Texas jargon is a bit different than my Joisey slang, so please bear with me as I try to decipher what we both mean!!

What "trip" was momentary? In my experience a "trip" initiates a lockout to prevent further operation of any kind until reset, if you had a relay "bounce" that's an entire 'nother story that requires investigation. Am I correct in assuming that your latch-in relay is the permissive to close the generator breaker (GCB)? If so, did the latch-in relay stay "latched-in" and/or the GCB trip-free?

While everything seems to point to the latch-in relay and the logic that drives it, however the relay is a relatively simple device and if there were no firmware and/or s/w updates applied then that chain could be ruled out after swapping the relay out.

I'd be looking at the GCB and its internal control circuitry. That's a subsystem unto itself, and I remember many instances where everything from the permissive back was working fine, but the breaker either wouldn't close or tripped-free. Either the closing spring wasn't charged enough, or it would close and immediately open (trip-free) because some of the complex mechanical linkages weren't set properly and/or the mechanically driven interlocking switches weren't making up properly. Sometimes just mechanically charging and re-racking the breaker cleared up the problem, other times we swapped it out for a known good one, in the worst case it involved replacing the guts of the cubicle.

Firstly, I apologize for not having the terminology down pat. It is supposed to be standard terminology of the industry, but I am a mere chemist working in a power generation plant.

Second, I must assume there to be intervening logic between the select latch switch and the relay, (in addition to latch mode switch for selecting automatic synchronization and manual). If all the ladder rungs in logic are set, then the latching relay should energize the field breaker, correct? If 9/11 incident was a relay bounce, it apparently took a long time to bounce. It is more like a permissive momentarily dropped out, then magically popped back into skew before trip timer could engage?

I do know that from time to time, we are starting to experience more anomalies when taking units offline at night, and this could be the outfall from that.

I suspect we will be going through the entire chain of devices with a fine toothed comb today, looking for the one thing that is out of place, out of time, and out of luck.

What it tripped on yesterday should provide a clue....sounds like something is about to take a crap....

Based on what I just reported to 67 model, I think it already did take one.

Pretty sure we have issues (again) in the HVT - to be diagnosed today.

Almost certain it is also in the latch in relay not changing state (I am not sure whether this relay is NC or NO, but I suspect it is NC (with opposite sense) so that if power is lost to this relay, the field breaker is switched open. I might be wrong on this, but how would an NO relay get "stuck" open if it is indeed a mechanical relay and not solid state?

How long did it take to trip "...right after latch-in...", what flags dropped, what meters took a bump, and what exactly did the control system annunciator panel report as the reason for the trip?

Do you have a synchrocheck relay? If so the acceptance window may be a bit too tight, and/or the fuel control didn't ramp fast enough once the breaker closed. We need the exact sequence of events to be more specific.

on 9/11, the trip was well within one minute. The reason left no trace behind. That is why I am writing about it. If we knew what was going on, we would not be poking about in the dark, now would we?

The strange thing is it also apparently latched right back in (and there were problems with that). Operators tripped, then reestablished sync idle, and latch-in in auto after that. No latch-in achieved whatsoever under any of the conditions on 9/12. Then we had a "hard" trip in auxiliary power, so indeed we are "groping about in the dark" this moaning.

Bear in mind, this has been a reasonably reliable unit since 1999, that it is GE controls, an LM6000 gas turbine, relatively few controls hardware issues up to this point, but the HMI computers are getting pretty old by now, and upgrades have been forestalled, but are pending. I think the desktops we were using in the control cab and in the control room were Windows 3.1 back in 1999, not sure, maybe XP.

Right now, the sequence of events is that the unit reaches sync idle with no issues. The synchroscope appears to be functioning normally with lights slowing phasing on and off with clockwise dial rotation (increasing phase from lag), and there is no actual firing of the latch relay, and hence there is no engine ramp up. She's ready to give us all she's got, but the tribbles are multiplying this morning.

I spoke with the group leader on site with distribution, and he mentioned the straps on the output side of HVT were hot enough at some point to melt the solder out of them!

He added this is due to loose straps. I can easily envision how this would prevent load up of the unit. If there is an excessive resistance at the HVT connection, then as soon as local generator attempts to increase load on the ramp signal, one or more phases would show a voltage out of range and cause a differential trip. Right?

Perhaps this even the supposed "relay bounce", we were considering earlier. Electrical events happen very fast, so it could seem "instantaneous" to human observer.

Hi James,

The names for electrical events in a power plant can be confusing. Mostly those protections which stop a set immediately to prevent damage are called "trips".

Unfortunately, all major circuit breakers have closing and trip (opening) mechanisms. Most circuit breakers (distribution) in a high power system spend their life closed and only open if there is a fault or maintenance requirement. Mostly closing involves storing energy in springs and takes a relatively long time, even 0.5 sec , opening just requires a retaining claw or similar to be "tripped" over (to release the springs) and happens as quick as possible, like less than 1/100 second. Almost all faults require breakers to be "tripped", so "trip" has become associated with all sudden protective shutdowns or actions - unfortunately, for generators, tripping the main breaker at the end of a run is routine, not a fault. Also, a fault which is judged to be in the load, not the generator/source may just open the breaker rather than stopping the generator - a basic example is a lightning induced flashover on a line, disconnecting and reclosing after a few cycles usually clears the fault.

Looking at your reports, it seems there was no problem with the synchronising and breaker closing to connect the generator to Grid. You report that the first trip occurred a minute after connection, which is beyond the synch/connect process. The problem was compounded it seems by failure of annunciation of the cause of trip and loss of auxiliary power. The dreaded multiple fault and no flag/alarm/log situation.

As a matter of fact, it is very difficult to electrically identify and trip on a high resistance joint at HV. The LM6000 is around 50 MW capacity and generators and transformers have very high efficiency, like >98% at this size.

Of the losses, the resistive loss is probably no more than 0.5% of transformer rated output - corresponding to a voltage drop of 0.5% of the rated voltage. This is compounded by the inductance voltage drop of a transformer being around 10% rated volts.

The relation of the total voltage drop across the series combination of the winding inductance X ohms and its resistance R ohms is given by a right angle triangle with the total voltage drop given by the longest side (this is because the peak voltage across the inductor is displaced 1/4 cycle or 90 degrees from the peak voltage across its resistance - the same current is flowing through both). Working it out for X = 10, R = 0.5, the longest side is 10.0125. If you suppose your bad joint is equal in resistance to the winding, you get X = 10, R =1, but the long side of the triangle is only 10.05 - only 4 parts in 1000 different. Note that 0.5% resistance is about 80 kW loss in one phase connection.

My guess on what tripped you is basically chemical. All that high temperature around the faulty joint, probably with arcing, decomposes the transformer oil into gases which bubble up to the top of the transformer - into a closed-top chamber with a float which operates alarm and trip switches as the liquid level is pushed down by the gas. The arrangement is known as a "gas" or Bucholz relay. It has a supplementary action where a surge of oil and gas due to a heavy fault moves a paddle which operates a "trip" shutdown switch.

So enlighten us by revealing the protection which actually shut you down.

67model

Your last line is the one that is giving me pause. You say "Indicator lights were lighting up when the phase offset rolled (slowly) through 0."

If I remember the functioning of the sync lights correctly, when the Incoming and Running voltages (Gen & Bus) are in phase, the lights should be off. The connection of the voltages to the sync system is that when they are in phase, they are opposite direction on the scope and lights so they "cancel out."

Has some recent wiring change been made that could have reversed the polarity of one of the voltage connections? Look at that first.

Pause accepted, I am having a brain functionality issue this morning, and perhaps it has been dislocated to my posterior. Yes, the lights were functioning normally off at top dead center 0. Proper null indeed, my bad, for making such an obvious boner statement, and it appears my short term memory is cross-wired.

Is there such a thing as age-induced mental dyslexia?