Gas Turbine Generator suddenly drops load and switches mode 2

[hayaku]

Hello there,

We have a 10 MW GTG operating in parallel with several more identical units. At the time of a thermocouple alarm (which was just acknowledged and reset) the unit dropped a few MW of load and switched operating modes from ISOCH to DROOP. Any ideas if these two things are indeed related?
The TC's measure the turbine inlet temperature which i believe, affects the maximum power output of the turbine.
Anyone dealt with generators behaving in this way and what were the likley causes of sudden load drops and/or switching modes?

Thank ye kindly

 

[Alex68]

Hello Hayaku,
you should ask direclty to the GT manufacturer.
Anyway... Usually a GT governor includes the temperature control that maintans the exhaust gas temperature below a fixed limit and works in parallel with the load and frequency control.
Maybe the temperature control of your governor changes the load control from isochronous to droop mode.

 

[hayaku]

Thanks Alex68, I have been going through the GT control system manuals to see what parameters affect load control and what are all the possible output actions.. so far haven't seen any case for GT to switch modes like that..


Hmmm.. i suppose in order for the machine to switch to droop..the input power was not sufficient to keep the set frequency at the current load conditions. (5.2 MW)
Its probably likely the machine switched modes and when the speed regulation was still too tight even in droop mode, it began to shed load?

 

[mc5w]

As the inlet temperature rises the power output of the gas turbine decreases. As humidity rises power output also drops because there is slightly less oxygen in the air and the moisture affects the adiabatic compression and expansion. For a summer peak utility this might actually used to advantage by operating the generator at say 70% or 60% power factor to provide more reactive power.

Something is causing this gas turbine to think that the inlet air temperature is higher than what it actually is.

I would look for a loose, oxidized, or improper terminal connection. For oxidatio problems your best weapons are silicon carbide abrasive paper for cleaning wires and so forth and there are also some high temperature rated antioxidant compounds. There is a reason why there are 2 anitoxidant compounds on the market that are formulated for use only with copper wire and brass conduit threads.

I also encountered and instance where a 20 HP permanent magnet variable speed motor would not start unless rotated 1/4 turn. Changing the 3 phase shaft position encoder did not solve the problem. Further investigation showed that in the press control box a technician inserted the encoder wires too far into a terminal strip with the result that the terminal setscrew was grinding through the insulation to contact the conductor rather than contacting the stripped portion of the encoder wires. This was essentially a workmanship problem because everybody uses $8 per hour apprentices to do the wiring.

I also had a problem with a servo feeder for a punch press that started double length feedeing, then lost track of which direction was forward, then shut down. Problem was a broken encoder wire inside of a military style multipin connector was the problem.

Another instance was that somebody forgot to tape the shield wires inside of a military style connector with the result that the encoder power got shorted out. Somehow erased the personality PROM that customized the drive for the particular motor.

Thermocouples also need a cold junction and a means to measure the temperature of the cold junction or to make the cold junction run at constant temperature.

I know that the loose wire, broken wire, and oxidized wire theory seems a bit too simple, but industrial machines get this a lot due vibration and bad workmanship. Bolt forging machines have to rewired once every 2 or 3 years and I one time renamed a machine ( fake French Accent ) Ze Lemon. The hard part is finding that bad wire.

You could also have an analog to digital converter that is shorting the +12 volt or +15 volt analog power into the 5 volt digitsal logic. I have seen this happen with a digital to analog converter that was sending 0.7 amp pulses into the microprocessor every time the microprocessor updated the D/A converter. Never did figure out how the D/A converter or the microprocessor did not catch on fire or blow a hole.

 

[itsmoked]

I saw the Ze Lemon.. It was a leather splitting machine. They actually use a 4 inch wide razor blade to split hides. The machines had 35 ice cube relays in it that ran logic. Once a week one relay's contacts would fail..

 

[ScottyUK]

This sounds like a fairly serious problem. Isoch mode is associated with islanded operation where the turbine governor maintains constant shaft speed (hence Isochronous). Droop mode is intended to allow load-sharing operation between multiple sets or to allow a generator to operate in parallel with a large system or grid.

I would be suprised if a machine operating on its thermal limit - i.e. base loaded - was also operating as an islanded machine. If the unit has IGV control it is possible that the IGVs are closed and the unit is operating on its thermal limit but below base load. This is about the only way I could imagine being on the thermal limit while in Isoch mode, and this would only be applicable for a CCGT application where the HRSG requires high exhaust temperature even though the unit may be at part-load.

On the face of it, having a unit transfer from a control mode associated with parallel operation into a mode associated with islanded operation as a result of thermocouple failure sounds very odd. You need to provide more information regarding the configuration of this machine and give some information about the system it is connected to before I could help more.





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

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

 

[mc5w]

Droop mode also has to be connected to the Area Frequency Control computer which tweaks the commanded speed of each generator to make the whole system isochronous. In the case of a cogeneration outfit that is a small part of the grid the setpoint of the droop mode control can just as easily made to produce maximum power, 75% power, whatever is economical to sell to the utility.

What a standalone isochronous governor does is that it has a small amount if integral gain with low bandwidth that brings the speed back up to 60 Hertz or what ever after a load change. An isochronous control still has a droop characteristic that has a fast bandwidth to follow load changes rapidly. Servo controllers for industrial machines have a similar concept that helps that servo system close in on an exact position in spite of friction that causes the motor to falter at low speed and torque.

I also had a redundant air valve for a pneumatic punch press clutch that would eat 3 amp ice cube relays. Only happened with just 1 of 4 contacts. The contacts of 2 relays are in series that in turn drive the 2 solenoid coils of the redundant safety valve that fed compressed air to the clutch. Evidentally, 1 solenoid was not sealing and when this 1 contact opened first it had to interrupt excessive current. The ice cube relays were the output of an electronic controller that was dual channel for safety reasons.

 

[ScottyUK]

Hi mc5w,

Isolated single machines do not have a droop mode. There is no requirement for this characteristic: droop is a means of providing load sharing between units. An isochronous governor on a single unit should not have any droop; it is designed to maintain constant speed as its name suggests. Multiple paralleled machines islanded from the grid use a speed reference modified by the droop controller to allow them to load-share.

Grid paralleled generators do not operate in speed control. They can't: they would cycle load as they tried to track the small fluctuations in grid frequency. Normally they are used in either temperature control (a gas turbine unit), MW control (any type), or IPC control (a steam turbine). I'm not at all sure of the terminology used by the hydro operators. Speed control of large turbines is reserved for non-synchronised operation - acceleration and synchronising, or islanded operation.



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

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

 

[hayaku]

Hi ScottyUK, mc5w,

If I understand correctly, you are saying that droop mode is used for load sharing. We have 5 10MW units that all run in Isoch as standard operating procedure and they all share the facility load.
Other plants where i have worked we had OMIB system so our little dinky generator would run in droop mode and in parallel to the main grid and hence our machine would follow the main grid system frequency.
Here there is no connection to the main grid, just 5 GTG's running parallel. I would have thought that perhaps one machine would be in Isoch to 'set' the frequency and the others in droop to follow.. but in fact they all put to operate in Isoch mode.
Any ideas how this is advisable? Perhaps since all the units are the same size it makes no difference?


concerning the thermocouples, seems we did not think we had turbine inlet temperature contol even enabled on the machines, but judging from the behaviour of the machine when 2 of the 17 TC's failed high we suspect we do!

The units are rated for 9.17MW at 26 deg C.. but its often much hotter than that at our location, so in fact we use 8MW per machine as a safe average. At the time the machine was doing 5.2 MW and then it switched modes and dropped to 3.0MW.

 

[Alex68]

HAYAKU
In a case like yours, in which 5 small GTG feed an isolated grid, it is better to manage all the machines in isoch mode.
If you have only one machine in iso mode, then in case of fault only that will work to restore the system frequency, while the others will try to produce the same power. That's not good depending on the system transient, the machine size, the load loss, etc...
In your case all the machine cooperate to regulate the frequency and maintaining the power production and consumption in balance.
On the other hand you can risk oscillations among machines. But you must activate the PSS to avoid active power oscillations and the negative compound of the AVRs to avoid the reactive power oscillations.
Some transient stability studies will allow to set the parameters.

About your particular problem of regulation mode change: What is the model of the GTs?
What are the governors?
If I am able to retrive the manuals of these components, maybe I will help you better.

 

[ScottyUK]

Alex68 has provided a good answer. An isolated group like this should operate in speed control, but they must have droop enabled otherwise the 'fastest' machine will hog the load. The small differences between the governors will cause this to happen even if all units have same target frequency unless droop is enabled.

One thought about the apparent change in mode is that if the t/c failed 'high', the high temperature reading will be propagated through the control program and the turbine controller will think the engine is hotter than it really is. This could cause the temperature controller to override the speed controller and may have caused a shift in the load controller state. Temp control is a base-load mode which is inherently a droop-enabled mode, and the machine may have remained in this mode. You should investigate how your controller behaves in terms of rejecting failed theremocouples. Normally critical t/c's are duplex pairs and the controller selects the 'best' of the two. We had some really badly written software which did not gracefully handle t/c failure. The problems were not identical but the source of the problem sounds similar.



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

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

 

[hayaku]

Thanks ScottyUK,

I have seen that situation where one of our TG's grabs all the load, overloads and trips out while the others just sit idly by and let him run off over the edge without trying to help the poor fellow. Then they all start going down one after the other.. like sheep following eachother off a cliff.

This is why as you say they must have droop enabled.. by this you mean we must have some automatic way of switching to droop when frequency unbalance arises then? Because right now all the TG's run in Isoch. And well up to the last couple weeks we thought they only go into droop mode if we do so manually.

 

[ScottyUK]

Hayaku,

Droop should always be enabled when units run in parallel. Essentially the machine governor control has a speed reference and speed feedback which are fed into an error amplifier and PID block to produce a demand signal to the governor valve. This is a simplistic description of a single unit running is isochronous mode. The droop controller modifies the above response by reducing the speed setpoint in proportion to the MW output of the unit. Typically the droop may reduce the speed setpoint by 3% or so at full load.

If this scheme is implemented on all machines in a paralleled group, as the load on the group increases the frequency will drop slightly. If one machine tries to hog the load, its droop controller will reduce its governor signal and lower the MW output of that unit. Thus the system is a stable closed loop.

It is possible for a supervisory controller to modify the speed reference signals to the whole group in order to maintain the actual output frequency at a constant value while allowing the droop controllers to load-share between units. This is a 'nice-to-have' and doesn't really matter in terms of how the system behaves.

The cascade trip scenario you describe is typical of a system which has no droop control enabled. Running a plant in this manner would quite probably explain one unit - the 'fastest' one - being in temperature control when it reached its thermal limit. You need to study some theory on parallel operation of generators - there are many textbooks which describe droop control in detail and give worked examples. It would be well worth you spending some time understanding how it works. It is hard for me to describe here without the ability to draw a diagram or two.



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

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

 

[hayaku]

Hey ScottyUK,

Thanks for the info...makes alot of sense to me and follows from what I have done in UG and PG. I came into a new environment and they do things a bit differently to other places i have worked.. but i am pretty inexperienced so i dont like to rule any options out. I'll definitely do some more research into parallel operation of gen sets.

 

[EddyWirbelstrom]

What is the recommended generator control mode for the following powerhouse system :-

- The powerhouse common 'synchronizing bus' is connected to the Utility via an interconnector transformer.
- 4 off steam extraction turbines generators, each connect to the 'synchronizing bus' via current limiting reactors.
- Plant load is supplied from the individual generator buses.
- Three of the extraction turbines are condensing and one back-pressure.

 

[EddyWirbelstrom]

Data I omitted from the powerhouse description :-

- 4 off generators each rated at 30 MW
- 1 off Utility Interconnetor transformer rated at 30 MVA

 

[mc5w]

mnewman,

Your 4 steam turbines would need to operate in droop mode with a speed setpoint adjustment coming from a wattmeter and master controller that monitors the utility tie transformer. You could then operate the transformer at constant power flow.

This is similar to the Area Control Computer idea that utilities use to make the system as a whole isochronous with each generator running in droop mode.

Also, if you need to produce maximum reactive power to help your utility you should leave 2 of your 4 voltage regulators in droop mode with the commanded voltage bumped up a little. When you operate a generator voltage regulator in constant reactive power or constant power factor mode some other voltage regulator will have to do the voltage regulating.

 

[EddyWirbelstrom]

Thanks for your helpful posting mc5w; can you please comment on my following assumptions and questions to your posting:-

Assumption 1.
With the Utility Interconnector in-service, the 4 off generators are constrained to operate at the Utility frequency.
Required MW sharing between the generators would be achieved by Droop speed control. ie. the generators would droop to the Utility frequency at which they deliver the fixed MW power determined by the governor zero speed setting frequency and the speed droop settings.

Question 1.
Any change in plant MW load would have to be supplied from the Interconnector transformer. You mentioned the Interconnector transformer operating at 'constant power flow'. Can you please explain the transformer 'constant power flow'. Where would the change in plant MW load come from ?

Question 2.
Assuming any change in plant MW load comes from the Interconnector.
If one of the reactors tripped, the generator and its load would be islanded and the generator would have to supply the MW previously supplied by the reactor ( Interconnector ). Would this require an automatic switching of governor control mode from 'Speed Droop to Utility Frequency Control' to either isochronous or Speed Droop to a lower frequency with adjustment to desired frequency after the generator MW load change ?

 

[Alex68]

mnewman
your assumption is correct but remember to activate the frequency bias. In this way your gens work in load control but are also sensible to the frequency and cooperate with the grid to the frequency control.

Answer 1: if the load increases, then you must ask the extra-power to the grid and viceversa. The cited "constant power" doesn't includes the load changes.

Answer 2: Yes, it does! You can do that in several ways but the simplest one is sending a command signal from the lock out relay to the governor to change from load control to frequency control and probably also activating a load shedding logic, if required.

 

[EddyWirbelstrom]

Thank you mc5w and Alex68, your assistance has been of great help. It has confirmed my generator modelling for steady-state loadflows I am doing.

When all generators, reactors and the Utility Interconnector are in service, I have modelled the generators to :
- maintain constant MW, and
- maintain set voltage control within the generators VAR capability ( PV generators ).
I have modelled the Utility as a swing bus and modelled source impedance.

For the case where the reactor trips, I have modelled the islanded generator as a 'swing generator' which supplies ( or absorbs ) MW and MVAr within generator capabilty limits.

Concerning stability :
What are your comments on the requirements for stability studies for the step load changes in this application ?
For the loss of a generator, there will be no rotor power angle change because the MW step increase is supplied from the Utility.
For loss of reactor however there will be a rotor power angle change due to the step increase in generator MW load from 23.8 MW before reactor trip to 25.8 MW after reactor trip.
The generator is rated at 27.5 MW.
Is a transient study justified for loss of reactor ?




.