Controlling Speed of Turbine Generator when in parallel with Grid


Thread Starter


I am doing an investigation to an incident within my workplace and unfortunately, some of the questions being raised fall outside my area of expertise.

We have a 132kV feed from the grid which is transformed to 11kv as soon as it comes on site. We have a normal load in the vicinity of 30MW of which approx 20MW is supplied by an onsite Steam Turbine Generator in parallel with the grid.

We recently had a situation where the Steam Turbine did an emergency shut down on a 110% overspeed trip.

The on site DCS, indicates that the main 11kV incomer tripped approx 2 seconds after the overspeed condition occured. This goes against my (somewhat limited) understanding of the principles of parallel operation in that I thought that if the governor failed on the Turbine Generator and tried to overspeed, that it would just take more real load, rather than actually overspeed.

Is it possible that our TG could overspeed whilst still in parallel with the grid, if all indications from the grid show that there was no over frequency condition on the grid.

Any help would be greatly appreciated.

PS. I hope I've posted this in the correct section.
It seems that a SUDDEN loss of load occured. It is obviously that you are working in a "islanded" section. I suspect that something tripped, leaving the lead generator to accept the sudden LOSS of load. I do find it interesting that reverse power relays were not the initial trip. Then again "islanded" operation. Speed droop settings. Check the settings to ascertain if the control system responded as set. If the generator responded as programmed. Check your switchyard.
Many thanks for your feedback. It is my personal belief that the main incomer tripped due to other reasons and the sudden load decrease (and a now verified sluggish governor valve) allowed the Turbine to overspeed.

My argument is that the incomer actually tripped first even though the DCS printout contradict this.

I really just want to be able to say, hand on heart, that there is NO way the Turbine Generator can 110% overspeed whilst in parallel with the grid.

Can I say that?

Appreciate any and all advice.

Speed drop settings are generally set at about 5%. Overspeeds are set at twice that ie 10%.

Assume a two pole generator running at 3600 rpm to provide 60 Hz. To reach the overspeed, the generator would of been running at 3960 rpm and providing 66 Hz.

Surely, someone monitoring the grid would of noticed this runaway generator trying to push every other generator off the grid. The speed droop settings should of closed or nearly closed off the throttle valves long before the generator reached 110% speed.

DCS systems are limited by the I/O that is configured. the speeed at which the I/O can be scanned and the speed at which a HMI/SCADA can scan what is being reported. Suppose you are scanning 100 points. Suddenly a change of state in a I/O point occurs at point 49 but the DCS is currently scanning point 51. Point 49 will not be reported unit the remaining points after point 51 and the points in front of point 49 are scanned.

When a trip occurs, generally a several I/O points report a change of state suddenly. The DCS has to handle those alarms as well. The difference between an engineering marvel and a engineering disaster in less than a millisecond. You cannot always trust the timestamping from the DCS.

I am not saying that the throttle valve could not of stuck in some position that caused a runaway turbine. If the throttle valves, speed droop, etc. check out fine after the trip, that basically tells me something happened in the switchyard that the turbine could not respond to within its control parameters.
Another test might help in troubleshooting. If the DCS has the ability to open the 11KV incoming feed. I would like to see the timestamps from the open command to the proof of open feedback. It may take nearly 2 seconds for feedback to be recieved from the 11KV feed.
Fantastic feedback! Many thanks indeed.

I probably should have gone into a bit more detail with the initial problem but was trying to categorically determine whether a single Steam Turbine Generator in parallel with the grid could physically overspeed? It has always been my understanding that the generator would somehow be electrically held at the same speed.

We have ascertained that there was a problem wherby the main steam control valve, and the oil pressure that controlled it were both not up to standard.

In the case of the servo control valve that drives the Live Steam Control Valve (LSCV), it peformed poorly in bench test after the original failure. In fact it took 10 seconds to close rather than it's specified 0.5 sec.

My hypothesis is that:
1. The 11kV incomer tripped due to a trigger from a Vector Surge Relay which detected some anomoly between us and the grid (no way of determining this as we did not capture first alarm around the switchyard monitoring equipment);
2. The electrical plant load shedded as specified (DCS evidence to suggest this occurred properly);
3. The large and sudden decrease in load, reduced the electrical load;
4. The servo valve started closing the LSCV (have good evidence from the Steam Turbine Control System) - way too slowly as mentioned above.
5. Before the turbine had a chance to slow down, it oversped and subsequently tripped Emergency stop valve and started to wind down;
6. The Generator Circuit Breaker tripped on under frequency as it wound down (40.4 hz, which even on the Australian 50 Hz system I think is too low);

I am very keen to know if it's possible for the turbine Generator to overspeed whilst in parallel with the grid. It is my understanding that it is not possible. This is the basis of my argument that the 11kV tripped indication on the DCS is false and results from latency in the system.

Great resource this... many thanks. Pete.
It's impossible for the turbine generator to overspeed whilst in parallel with the grid.
In your case the breaker tripped first and then generator oversped. The puzzle may lie with the feedback signal of breaker tripping.

Yes, it is possible.

I disagree with MAhsan.

Your situation proved the possibility, but not for very long. In a millisecond world of frequency.

MAhsan; I apoligize!!!

No!!!!! A generator cannot remain synchronized with the grid at 10% over-frequency.

I previously posted that a test should be conducted to ascertain the time between the command to open the switching mechanism and the feedback to the actual opening.

Many HIGH VOLTAGE switchers; although they change states quickly; require some time to relay the 125 VDC feedback.

Thanks for you help Gents.

Unfortunately, the DCS does not send a trip command to the 11kV incomer breaker so I'm unable to determine time delay between command and feedback indicating breaker opened.

CTTECH, I got a little confused in your last reply. Are you saying that it can happen, but only for a second, AND it can't happen for very long. There was a 1.42 second delay between the steam turbine tripping on overspeed and the 11kV incomer breaker tripping, according to the DCS.

Thanks again,

Phi Corso, PE

Some additional food for thought:

1) No, it’s not impossible for a generator to overspeed while connected to the grid! It could have suddenly become an induction-generator!

2 Suppose upon loss of local-load the generator’s rotor-excitation was also lost. Then, the machine would certainly operate as induction generator.

3) As a result of losing 1/3 of its load, it would overspeed, but the grid frequency would have remained what it was before the incident. No droop-control, mechanical or electronic, can react fast enough to cut-back inlet steam!

4) Do you have any suspicion that the mechanical overspeed-trip malfunctioned?

5) On the other hand, if the grid tripped first, then loss of 2/3 of its load would have caused the machine to overspeed with an increase in generator frequency. Since your sequential-events-recorders appear to be useless, then proof of overspeed could be seen in any old-fashioned, paper-type circular or strip-chart recorder!

6) What initiated the DCS load-shedding signal?

7) Did you lose lighting in the control room? Instrument power?

BTW, to all of you involved in operations and forced to accept scanning-type event recorders, here is a trick I used in the middle ‘70s… train a cam-corder on the control board. However, use one whose tape-speed can be increased!

Regards, Phil Corso, PE ([email protected])
Mr. Corso,

I am curious about the induction generator. I am only familiar with the induction motor. I have witnessed, due to DC field failure, the excitation field loss of a generator.

I have witnessed the combat between the driver end and the driven end. The driven end became a motor and the driver end refused to relinquish.

All of the above was caused by the malfunction of the reverse power relays. Could you enlighten me?

[email protected]
In the dynamic world of synchronization. I DO NOT believe a generator can remain synchronized during a overspeed. SOMETHING has to release the generator from the grid. Although I yield to those that have witnessed otherwise.

Phil Corso, PE

Yes, CCTech, I’ll certainly take a crack at it! But first, I would like to address several misconceptions about a TG-set operating in parallel with the grid, that is, a grid much larger in capacity than the TG-set in question. By “much larger” I mean at least one order-of -magnitude, where the generator has no influence on the grid. The grid, then, is referred to as an “infinite-bus!”

Misconception # 1: TG-set speed is able to change while connected to the grid!
While connected to the grid TG-set frequency is synchronized to, therefore locked to, therefore exactly equal to, grid frequency. In other words, the grid can’t be 59.623 Hz, and the generator 59.624! (BTW, increasing the number of significant figures, doesn’t add credence to posted answers! If instrument readings are different, that makes the meters suspect, not speed!)

Misconception # 2: changing the turbine’s speed changes the generator’s frequency!
Changing power input to the prime mover can’t change system frequency. Technically speaking, not even one-millionth of an Hz! However, there is an associated change in the physical-position of the TG-set “rotor” relative to the voltage sine-wave of the system! That displacement is called the “power-angle!”

Misconception # 3: changing field excitation changes terminal voltage!
Changing exciter voltage can’t change system voltage. Technically speaking, not even one millionth of a volt! However, there is an associated change in the generator’s output current. Some describe it as a change in the generator’s kVAr output! (This subject has been covered extensively on this forum, although thus far, there has been only one dissenting “voice” regarding my interpretation!)

Misconception # 4: a generator’s speed can’t be different than that of the grid!
The scenario I posited in an earlier post, that is, a generator, yes, an AC synchronous-generator, can operate as an induction-generator… when excitation is lost, but while mechanical power is still available. That means, while its shaft speed (in frequency terms) is different than that of the grid, the generated voltage is still synchronized to the grid. Of course, while its field-excitation is nil, excitation is provided by the grid. Hence operation is exactly like that of an induction motor operating above synchronous speed. The key element is that turbine input must be sufficient to accelerate the generator so that its frequency, like that of an induction motor, must be greater than synchronous (grid) speed! Will SOMETHING happen? You betcha! Instrumentation will certainly reflect the change over from synchronous to induction operation. That’s why I suggested a cam-corder be trained on the instrumentation panel.

For Pete Visy: although you stated the problem “fell outside your area of expertise”, what theory did your local experts present?

CCTech: Were you with kcpl during the HS boiler “problem?”

Phil Corso, PE ([email protected])
How long, and how often, did this mythical synchronous generator operate as an induction machine?

What was the output of the machine at the time it was being operated as an induction machine?

What was the power factor of the machine at that time?

What was the VAr loading of the machine at that time?
Most of us who have operated power plants have observed system voltage changes when changing excitation. There are many purchased power agreements which specify that independent power producers provide, and I'm quoting, "...voltage support..." during specific periods and conditions. This is done by increasing excitation and "pushing" kVAr's out onto the grid.

These system voltage changes can be dramatic depending on distance from nearest power plant and/or substation, impedance (reactance) between power plants and/or substations, and size (capabability) of generator(s).

Theory is great; reality is, well, reality.
One possibility is that the mechanical drive power somehow increased to the point where the machine angle exceeded 90 deg, and it began pole-slipping. This would certainly cause major electrical fluctuations, and also somewhat extreme mechanical effects!

Thanks, Phil, and everyone else who has assisted.

Unfortunately my local "experts" aren't exactly as expert as one would like. As a matter of fact, I think I'm fast becoming the local expert on these matter given the amount of research that I've conducted to date!!!

Since this subject started, I have found that the control/lubrication oil had built up a waxy, varnish like substance on bearing surfaces, pipes, steam control valves, etc., etc. To the extent that the pistons on the governor, which should be removable by hand, had to be hammered out!

This would be the reason why the Live Steam Control Valve took 10 secs (vice 0.5 secs) to close during subsequent bench test.

It is my belief that this sluggishness in response of the governor caused some transient to arise that was sufficient to energise the Vector Surge Relay and trip the main incoming breaker.

The resultant loss of power from the grid caused us to load shed which removed load on the TG from about 20MW to about 5MW. The TG started to speed up as a result and due to the hopelessly slow performance of the governor control valves, reached 110% overspeed and tripped the emergency stop valve.

The TG started to slow down, the Generator Cct Breaker tripped on low frequency and we enjoyed some hours of darkness.

The grid in our case could be considered "infinite grid" and the fact that I have firm data from the TG control system showing me how it performed during this event means I am discounting the report from the DCS telling me the incomer tripped nearly 2 seconds after the TG oversped.

I have no reason to believe we lost Generator excitation.

Our DCS does not record sequence of events, and normally runs at about 85% of load. It is my belief that in this case due to the 100s (1000s?) of alarms that were present the DCS became overloaded, causing erroneous event times in the DCS. I have a chart of the DCS loading trends (albeit only 10sec resolution) that shows it loading out to 100%.

Thanks again,
Mr. Corso,

I understand the infinite bus concept, yet the misconceptions are confusing me. I am new to this forum and wish to share my humble experience and also educate myself and perhaps others. My motto: The only way one can know everything is for one to accept that one does not.

Misconception #1: Speed cannot change while connnected to the grid.

I had believed that the prime mover must to be able to respond to speed/load changes within it's speed droop parameters. While this is dynamic in this "infinite bus" and the speed changes are subtle, are we not controlling the power input to our prime mover based on speed control. As load/power demand is applied to the generator, the control system detects the prime mover is slowing slightly and additional steam/fuel is applied to the prime mover to compensate for the load imbalance. Frequency at the generator set is upset slightly while the generator set raises or lowers it's load/power output. The grid and all the other generator sets on the grid raise or lower their power output proportionately to bring the frequency upset back into balance. It is transparent to most. If the generators in a particular area are unable to supply the load for the grid, the imbalance could become a "brown-out" on the way to a "black out".

Misconception #2: changing the turbine's speed changes the generator's frequency!

Or should of it been grid frequency.

On an electrical island, it should do just that. A 2 pole generator operating at 3600 rpm produces 60 Hz (isoch). Change the rpm and change output frequency; ie 1800 rpm = 30 Hz. Now we abandon our electrical island and join the grid (synchro), we should still affect our partners on the grid when the generator set is changing load. Back to misconception #1.

Misconception #3: changing field excitation changes terminal voltage!

We often push terminal voltage up (VAR) to compensate for the "voltage control" on the grid. The location of the generator set on the grid determines just how much voltage we need to compensate for voltage imbalance on the grid.

Misconception #4: a generator's speed can't be different than that of the grid.!

When the system is in balance, I would think generator speed should match frequency; ie "locked to" the grid. Refer to misconception #1 for imbalance of the grid. If load is suddenly removed from the generator, I would hope that the reverse power protection systems in both the generator controls and the generator protection relays would detect power being applied to the generator from the grid and open the main generator breaker. Also, misconception #4 seems to argue with misconception #1; i.e. "locked in".


Phil Corso, PE

CTTech, your 17-Jun (19:08) comment regarding voltage excursion, and how IPs are ordered by system operations to provide "voltage support" and "pushing" kVArs, is obviously a case of a network in deep doo..., ahem, trouble! It is not the simple case of a generator connected to an infinite bus. Certainly, not the situation Pete described!

The scenario you described, that is, the existence of a signifiant impedance between the generator terminals and the grid is not atypical, but in fact, a normal situation, i.e., generator-to-grid connection via a step-up transformer! In this case your observations related to electrical parameter excursion, is certainly correct. Furthermore, I agree with your stance about distance, impedance, etc.

Now, regarding your statement, "Theory is great; reality is, well, reality." Once again, I bow to your position, that is, operators are often the most knowledgeable when it comes to electrical crisis management. But, if you recall, the originator asked a relatively simple question. Paraphrasing, he asked, "can a TG operate at a speed greater than syschronous speed?" I provided such a scenario... asynchronous operation!

Admittedly, simultaneous loss of excitation and local load is unusual, but not unheard of! In fact in the '60s and '70s, a great deal of work was done to overcome asynchronous operation, therby avoiding the inevitable (at least in my experience) "black-out!"

Regards, Phil Corso, PE ([email protected])