VAR swings following gate swing on 100MW hydro unit 1

[low1]

Hey everyone, I'm hoping someone here will be able to give me some input.

I'm trying to troubleshoot some issues we have been having with one of our units. The primary issue is that the governor system, Vevey Type 200, will intermittantly slam the gates closed while running online. It appears as though there is either an input that is failing intermittantly, causing the governor to drive the gates closed, or one of the control paths in the governor itself is failing, causing the gates to slam closed. I've experienced a similar problem here, when a different unit was being placed on Joint Load control, the load setting signal was being grounded out, telling the governor to immediately go to 0% load. We are monitoring a number of points with an event recorder to determine the actual point of failure. However, here's where I need some help.

Looking at the data captured on our UCMS system, we see that the gates swing closed, followed immediately by a significant increase in field voltage, field current, and VARs. The actual event looks like this. The gates drop from approx 80% to approx 40%. Then the gates swing back up to about 60%, then down to full close, back up to about 15%, back to full close, up to about 30%, full close, then start climbing back to normal gate position. The speed at which these swings occur seem to be limited only by the actual gate timing itself. When this happens, every time the gates swing closed, the unit's VARs swing up. When the gates swing open, VARs swing down. The VAR swing is approximately 40-50 MVARS, peak to peak.

I understand rapid gate movement will cause the REAL power (watts) to surge like this, but I can't understand if these VAR swings are a natural consequence of gate swing. If it is, then the problem should be resolved when we fix the governor issue. I'm just not sure if that's the case, or if we are looking at bigger problems that a intermittant fault on the governor system.

Could someone please tell me if VARS should swing with rapid gate movement?

Thank you very much in advance.

 

[ScottyUK]

How long does the excursion last? Are you talking about a few cycles, or for seconds? As a machine of that size suddenly accelerates or decelerates the load angle will change fairly rapidly and the AVR will certainly react as it settles to the new steady-state condition.


----------------------------------

If we learn from our mistakes I'm getting a great education!

 

[rasevskii]

low1:

Firstly, you should not even run the unit if the governor is acting up in that manner. The large swings in the wicket gate opening can and will cause pressure surges in the penstock, possibly resulting in a disasterous situation in the waterways system.

To see what can happen to large hydro units that have gone unstable just Google "Sayano-Shuskenskaya" the large hydro plant that was wrecked and flooded by a water surge in one of the 600MW units.

You must call th OEM (Vevey of Switzerland) to get a service engineer out to your site ASAP. I have worked with them on a project and they are absolutely first class.

Do not worry about what is happening to the VARS, The safety of the plant is foremost.

Likely the problem is in the gate position feedback sensor, or the MW sensor that is giving false or unstable information to the governor. It can be a simple loose connection, or a loose linkage only for example.

If your are not an expert on governors. or there is no such expert around, do not touch it.

rasevskii

 

[dpc]

Excellent advice from rasevskii.

 

[low1]

Thanks Scotty, rasevskii.

Scotty, the swings occur at the rate of maximum gate timing. I'd have to check the charts again, but the entire event lasts a number of seconds. My thoughts are that there shouldn't be any acceleration or deceleration, as the unit breaker is closed and the generator is tied to the system. There are a number of LO MWATT alarms, and Reverse power alarms, as would be expected, but my understanding is that the machine will spin at a constant speed as long as the breaker is tied. I do suppose that there would be enough of a load angle change to have the AVR to react to it, it just seems to me that it is reacting much stronger than it should.

rasevskii, I used to be an Electrical Technician at this plant, and at the time had a heavy involvement with the control systems on the exciters and governors. I have been away from the plant for a number of years now, but still get called back from time to time when site staff has problems they can't solve on their own. We do have a dedicated technical services department, but they told site that they would not be assisting, as they have no experience with the equipment. That leaves me as the "expert" by default. I am likely the person with the most experience and familiarity with the system who still lives in the area. I have taken a number of courses on generator control systems, but I'm not sure if I would call myself an expert. More like "the best we have" at the moment.

I do know that gate timing is critically important to avoid catastrophic penstock failure. "Slamming" closed is perhaps not the correct wording. Closing at the maximum rate allowable is perhaps a better way. Our river plants here have essentially no penstock, head is about 26 meters, through the intake gates and into the scrollcase. I'm definitely not downplaying the safety aspects, but have been assured that the gates are operating within specified timings. These timings are dictated by limits in the actuator head, which doesn't appear to be the source of the problem at the moment. I'm not sure if Vevey has been contacted by site yet or not, but I will definitely suggest that they do.

Regarding the problem specifically, I'm looking primarily at the power setter/MW feedback circuit as I know there a number of logic components that have been prone to fail in the past, as well as possible issues with the speed signal generator. These systems are 25-30 years old and a lot of the components are past the end of their useful life. I have been going through the prints to determine what will drive the output to full close, and there are a number of possible component failure scenarios. The joys of intermittent failures, and troubleshooting by second-hand information.

My question with the VAR swings is in regards to other issues we have had in the plant in the past, where simultaneous ground faults have caused DC control voltages to collapse, causing erratic operation in a number of equipment at the same time. I'm just trying to find the likelihood of our governor issue causing this var swing problem, or if the problem is perhaps in the unit control further up.

Thank you both very much for your input.

 

[rasevskii]

low1:

I fully understand and agree with your situation. Since it is a low-head plant the resulting pressure surges will be less dangerous but even on a short penstock with gate closing times acting at the design speed (hydraulically limited) it is not an absolutely safe situation.

The unit should really not be operated.

If there are that many faults in the control equipment (DC failures)and lack of any qualified staff at the plant, and there is no (or no longer) any central group that is responsible for governors, excitation, and relays, (a protection and control group) then that is a sad tale indeed. You yourself can only be asked to do so much, not work miracles in an old plant.

To be safe, it would seem advisable that the OEM be called in to make an assessment on the governor.

If you are in the USA, it might be an idea to get in one of the several companies that specialize in hydro upgrading to quote a complete replacement of the P&C for all the units.

rasevskii

 

[waross]

The VAR swings may be a result of the load swings. It depends on your excitation control. As the gates close the MW load drops and the real current drops. If your excitation doesn't drop of as well you will be over-excited and VAR flow will increase.
Once the governor is fixed the VAR issue may disappear.

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

 

[low1]

Thanks everyone for the replies.

waross, that makes sense.

rasevskii, thanks again. I guess I should mention that this gate swing issue is very intermittent, it's happened 2 or 3 times, all in the past year. The other unit that I had mentioned that experienced a similar problem had a failed NAND gate IC, and that was 4 or 5 years ago. The other issue I spoke of regarding multiple ground faults was the result of some poor planning, lack of insight, and Mother Nature. We had an existing ground fault on the system and three failing battery bank chargers, so the entire DC system (3 separate islands) was tied together. Then, one night during a rain storm, moist air entered the powerhouse and caused a relay to fault. This shorted out the entire DC system and we lost 9 out of 10 units.

The lack of qualified staff is a known issue, and something we struggle with daily. We are in a remote location with very high turnover. I actually left the plant on a temporary job with the purpose of attempting to capture the knowledge of technicians before they left the area, documenting work procedures, tips and tricks, etc. I've since been more focused on helping establish a new training center in the area to help develop our technicians. Our central group of technical support is also in pretty rough shape, mostly EITs right out of school with zero experience, being told by their superiors to focus on project work at other plants. I had the privilege of working very closely with the last good engineer we had up here who worked specifically on governors and exciters.

At the end of the day, though, I'm an electrical technician by trade, in an engineering technician position focused on training. I simply can't make the call to shut down the unit. Believe me, I've tried. The color of my hard hat prevents me from making those decisions. I can only make recommendations. The unit is coming out of service on the 19th for an opportunity outage, so I am trying to put together a plan of attack for site staff.

Thanks again everyone. I'm feeling more confident that the var swing is indeed related to the gate swing.

 

[rasevskii]

quote: "The lack of qualified staff..."

This speaks worlds about the state of the power industry. The retirement/downsizing of the old bods who knew the old equipment, and the plants like the back of their hands, have been replaced by newbies who often know little and are not really interested. They have not yet seen a real disaster happen in a plant.

The nature of hydro is of course that qualified staff cannot be expected to stick around at the isolated sites, where absolutely nothing else exists. The better orgs have central P&C departments that have travelling staff with a load of test equipment that go around and take care of governors and excitation systems. The companies in Brazil, for example, are set up very well in this respect. The local operators are not allowed to work on this equipment.

Are the Vevey governors original or are they a retrofit of and older system? Vevey has been part of the VA-Tech group of Austria for many years.

rasevskii

 

[wolf39]

low1:

Bill is exactly right with his comments dated 12Dec11 21:47. If the excitation current remains constant (in case the AVR got stuck) during MW load changes such VAR variations are inevitable. However, don't forget the field time constant. For large hydro generators such a time constant can be as high as 10 seconds. If the MW load change is happening in a matter of a few seconds only, even a functioning AVR would be unable to follow fast enough and the VAR issue therefore may not disappear completely.

Wolf

http://www.hydropower-consult.com

 

[rasevskii]

I have identified the plant, but won't mention the name here. It has 10 Kaplan Turbines and Generators originally built by English Electric from about 1959 to 1963. That means the Vevey governors were retrofitted at some later time and are likely digital units.

English Electric used their own mechanical governors at that time, still to be found in many hydro plants elsewhere.

The fault with the battery system, causing 9 units to trip was due to human error during maintainance. The DC was switched off by mistake. (one unit was not running?)

All the more reason to get the OEM in and check out everything.

rasevskii

 

[wolf39]

rasevskii:

English Electric built generators mainly at their Stafford plant (formerly Siemens. Bros.) until 1968. Then EE merged with GEC (General Electric Company). To my knowledge English Electric never built hydro generators in the 100 MW class, not even close to it.

Send me an e-mail for discussion.

Wolf

http://www.hydropower-consult.com

 

[low1]

I have identified the plant, but won't mention the name here...

rasevskii

I think you're close, but not exactly right. 10 units with fixed blade propellers, generators built by Canadian General Electric between 1971 and 1979. I'm postive that the governors, as well as the Westinghouse exciters, are original equipment. The Type 200 is an analog electric governor. It's old, and not without it's problems, but from a technician's standpoint, it's maintainable. A resistor is a resistor, diode is a diode, and op amp is an op amp. The prints, manuals and card descriptions we have are actually very well written, its just that not too many people up here have taken the time to read them.

The DC problem was both human and equipment error. The chargers should have been replaced years before, all banks should never have been left tied together, and the original ground fault should have been given more priority. When the unit started up and the relay faulted to ground on the opposite leg, the DC voltage collapsed, and the reference voltages for the governors, exciters, joint load control panel and joint var control panel all went haywire. There was also human error by the operator on shift at the time, although he saved the last unit from tripping off by removing it from joint load and putting it on full manual control. It was a fun night, to say the least.

Thanks again everyone for your input. I finished going through the prints today and have a plan of attack to discuss with site staff tomorrow. I am also pushing for out Technical Services department for more involvement.

Thanks again everyone.

 

[rasevskii]

I Stand Corrected...

The other plant that I thought it was also had a DC fault that tripped off 9 or 10 units...

rasevskii

 

[rasevskii]

Very odd that they would take "reference voltages" from the station battery. In other plants the station battery (usually) can actually be switched off for a particular unit, and it will continue to operate under load. All the protection will be lost, however.

Normally governors were at that time supplied from a PMG, and the excitation system was self-supplied. In more modern units, loss of auxiliary supply including the station battery will result in transfer to manual.

Regarding joint control, often it is included in the equipment, but never used or actually disabled to avoid anything that will disturb all the units together.

rasevskii

 

[low1]

Reference voltages aren't from the batteries directly, but from dc-dc converters. They are regulated outputs based on a "slightly" variable input. The input varied much more than tolerances allowed.

Our Vevey governor, as well as the type 500 (I believe, Mipreg digital) on the next plant up the river, have speed signal generators comprised of 16 magnets bolted to a steel collar around the turbine shaft itself. These magnets pass by 2 sets of sensors, diametrically opposed to compensate for shaft wobble, which provide the speed signal to the governor.

Our excitation units are all static in the local area, and can field flash from either station service or station battery, normally the AC station service. The power supplies are self-contained, however the Joint Var panel again uses dc-dc converters for voltage supply.

All three powerhouses in the area run predominantly on Joint Load and Joint Var control, and usually from our System Control Center about 900km to the south of us, unless there is a compelling reason not to. One of the plants even has a Joint Efficiency control, but it is not used.

To be honest, it was never 100% verified exactly what went wrong, the best assumption was the the station DC voltage collapsed and everything went haywire. Understandably, they didn't want to repeat the event to find out for sure.