System Frequency Drops on Gas Turbine with Mark IV control system

If one GT (GE frame 9 with Mark IV control system) is running on droop mode/Speed control then if system frequency goes down for tripping other machines how it will response? Whether it will share load? As theoretically it should share load but i do not find in Mark IV logic/control sharing load. (What i find from control system only it will raise load but do not know how much rate and again it comes back to previous position as in this case TNR remain fixed.) Could anyone explain from Mark IV control view with specific logic/control name and whether my understanding is correct?

Since a droop machine does not try to control frequency, when the frequency of the grid/network to which the droop machine is connected decreases the frequency of the gas turbine-generator will also "decrease"; read on.

This may be (part of) the problem you are experiencing; in reviewing all of your posts you have been complaining that if one unit trips then another unit will trip on high exhaust temperature. Gas turbines have kind of a nasty little secret: When the rotor speed of a unit decreases with a decrease in frequency, the air/mass flow through the unit decreases. If the unit is operating at Base Load and the fuel isn't reduced very quickly the exhaust temperature will increase--possibly enough to cause the unit to trip on exhaust overtemperature.

IF A GAS TURBINE-GENERATOR IS OPERATING AT BASE LOAD AS THE FREQUENCY OF GRID/NETWORK TO WHICH IS CONNECTED IS DECREASING the CPD-biased exhaust temperature control reference will be reduced--and, hence, the fuel will be reduced. Which means the power output of the unit will be reduced--which is EXACTLY THE OPPOSITE of what one wants to occur on a grid/network when frequency decreases!!!

What is happening in this situation is that the unit (or units) being used for Isochronous control has reached it's rated power output and can't increase its fuel/power output in order to maintain the grid/network frequency. This is the theoretical explanation: that one "large" unit is being used to control frequency and that is has the ability to "absorb" the loss of any of the droop units connected in parallel with it.

The unit being used to control frequency (the Isoch unit) should NOT be operated at or near its rated power ouotput, and if its load approaches its rated power output then additional droop unit(s) should be connected to the grid or loaded to reduce the load on the Isoch unit.

The difference in the power being supplied by the Isoch unit and its rated power output is sometimes called "spinning reserve" (not to be confused with GE's auto-synch setpoint they refer to as Spinning Reserve...isn't this fun?). Spinning reserve on a grid/network means that additianl generating capacity (a unit (or units) which is (or are) NOT being operated at their rated power output) is connected to the grid/network and is available to provide additional load if required to support the grid/network frequency--such as if a droop unit trips. If there isn't any spinning reserve on a grid/network and a droop unit trips, the grid/network frequency is going to decrease. If the spinning reserve is not available from the Isoch unit, then droop unit(s) operating at part load must manually be loaded to maintain the frequency.

In general, the formula used in Mk IV for speed control is: FSRN = (( TNR - TNH) * FSKRN2)) + FSKRN1 (refer to Sh. 27D of the Mk IV SpeedTronic elementary). FSRN is Speed Control FSR, in percent; TNR is the Turbine Speed Reference (i.e., the droop speed control setpoiont) in percent; TNH is the actual turbine speed, in percent; FSKRN1 is the assumed value of FSR at Full Speed-No Load (i.e., 100% speed with the generator breaker open), in percent; FSKRN2 is the gain, or, regulation, epxressed in %FSR/%Speed.

The above formula is essentially y=mx+b, where the variable, "x", is (TNR - TNH); the gain, "m", is FSKRN2, and the offset, "b", is FSKRN1. If you look at the formula, it says that if TNH is constant (remember, grid/network frequency is being maintained by the Isoch unit(s)) that as TNR is varied the amount of fuel will vary. Increasing TNR will increase the fuel; decreasing TNR will decrease the fuel.

This is Droop Speed Control at its simplest--TNH is being maintained at a relatively constant value by some external source.

Now consider what happens when a unit is operating at Part Load (i.e., NOT at Base Load and NOT on Preselected Load Control), and the grid/network frequency decreases. In this case, TNR remains constant and TNH decreases, which results in an increase in fuel to try to help support frequency. BUT--the droop unit is NOT in control of the grid/network frequency, and the increase in fuel MAY OR MAY NOT BE ENOUGHT to assist with increasing the grid/network frequency back to the desired setpoint. And, if the grid/network frequency does increase then TNH will increase and the fuel will be reduced as the error between TNR and TNH decreases.... It's an interesting set of circumstances which isn't easy to explain or understand. It generally takes a lot of reflection and review to grasp the concept.

GE heavy-duty gas turbine speed control philosophy is "documented" (sort of and very briefly) in the GE Service Manual, Vol. I, in SCPR-1 of the 'Control and Protection Systems' tab. Also, Section 2 of the Control Specification-System Settings document usually has some brief, but sometimes insightful, information on speed control.

The reason a unit whose governor is in droop control "shares" load with other units is simple: If the frequency/speed of the unit is being controlled by some external source (i.e., TNH is constant because the electrical grid/network frequency is constant), the amount of fuel being admitted to the unit is strictly a function of the error between the speed reference (TNR) and the actual speed (TNH). Increasing TNR will increase the error--and increase the fuel; decreasing TNR will decrease the error--and decrease the fuel.

Some electrical grid/networks have very unusual characteristics, and can have very unique problems. The "design" of the grid/network (transformers, power correction capacitors, location of generators in proximity to loads, length of transmission lines, etc.) ALONG WITH the quality of the governors of the generators connected to the grid/network and of the regulating body (system "operator"/overseer) all contribute to the quality/stability of the grid/network.

>From all your posts on this tripping problem--and there have been several--you are still not providing alarms and information that has been requested. Rather, another post is opened and the same question is posed in a slightly different manner--WITHOUT PERTINENT INFORMATION AND FEEDBACK.

For example, has this problem existed for some time, or is it a relatively recent phenomenon? Has something recently been changed/modified/rebuilt in the fuel delivery system? Has something been recently modified in the electrical grid/network to which the units are connected?

There are simply too many parameters and variables to properly troubleshoot your problem in this forum (not to mention that requested information has not been provided). There are many firms--including GE Energy--which can (for a fee) come in, gather information, and properly analyze your system and suggest enhancements/solutions.

Now, having made the above explanations you are IMPLORED TO RESIST THE TEMPTATION.

I have not got a clue that why system frequency goes down if GTs are on droop mode. does it mean that you drop the speed of another GT?

in our Mark Iv control FSRN calculated from (TNR-TNH)X FSKNG (FSR speed ref prof gain). Our system is Mark IV not Mark IV+ system.You said droop machine share the load only if TNR changes. What about TNH changes? If TNH (system frequency) goes down the droop machine share the load or not? Could you without ambiguously tell me that?

another question what is the difference part load and preselected load control? aren't they same?

I said before machine trip with Over Exhaust thermocouple high & trip alarm. and nothing done after trip and again started and it is running well.

I open the post as i want to clarify my knowledge with others like you (Really a great help).

my mail is [email protected]. You can reply on it also if you wish.

With the standard droop setting of 4% or thereabouts, a drop in frequency of 4% will change the fuel flow to the turbine from 0% to 100% - so this will also contribute to high back end temperatures.

One paragraph in the first reponse (which was mistakenly posted before it was proofread and complete) attempted to describe what will happen when the unit is operating at Part Load (not on Preselected Load) when the system frequency drops. Yes; if TNR remains constant during the frequency drop the error between TNR and TNH will increase and fuel will be increased--and load will be increased. Droop control is proportional control control only (for those that prefer that terminology).

Droop governor mode allows units to "share" load with other units on the electrical grid/network because droop governor mode is straight proportional control. "Sharing" load means a unit operating in Droop control will only accept a portion of the total load (up to its rated power output) and will "behave" and not cause any instability on the electrical grid/network.

WITHOUT SOME SORT OF ISOCHRONOUS LOAD SHARING SCHEME, multiple units operating in Isochronous control and connected to the same electrical grid/network will generally "fight" each other to try to control the frequency. Their operation is usually erratic and unstable. Droop control is what allows multitple units to "share" load on an electricl grid/network without causing any instabaility or erratic operation.

Will a unit or units operating in Droop mode "share the load" when the system frequency decreases? Depends on what your definition of "sharing" is. If three 120MW-rated units each with 4% Droop setpoints are all operating at 60 MW, and suddenly the system frequency drops by 1%, then each unit will pick up approximately 30MW--AS LONG AS THE SYSTEM FREQUENCY STAYS AT 1% BELOW SETPOINT because the amount of load will be proportional to the error between TNH and TNR. AS SOON AS TNH STARTS TO INCREASE, the amount of load each unit picked up will start to decrease. Are these units "sharing the load"? Or, are they each just accepting an equal amount of load because they have the same rated power output and the same Droop setpoint?

Some GE System Description documents in the 'Control and Protection Systems' tab of the Service Manual describe what might happen when multiple units operating in Droop governor mode suddenly experience a drop in the system frequency as "sharing the load" if they each have the same droop setpoint. Technically, each unit in such a scenario would pick up a similar amount of load by each taking on an amount of load in proportion to the percentage that the system frequency decreased. That could be described as "sharing the load"--sharing the load that was shed when one or more units tripped off the line and the unit with the Isochronous governor could not make up for. But that's NOT the same as "sharing load" with other generators on an electrical grid/network under normal circumstances--which is what Droop control is primarily for.

Preselected Load is a FORM of Droop control. When Preselected Load Control is enabled, the SpeedTronic will Adjust TNR as necessary to maintain the Preselected Load Control Setpoint. In other words, Preselected Load Control could be considered an "outer" loop to Droop Control. In preselected Load Control, the amount of load is being controlled by increasing or decreasing TNR to make the load equal to the Preselected Load setpoint.

When the unit is NOT on Preselected Load Control and is NOT operating at Base Load, the "load" is a function of the difference between TNR and TNH. To increase load, one increases TNR; to decrease load, one decreases TNR--but the AMOUNT of load is NOT being controlled.

In this author's opinion, if a unit were operating on Preselected Load Control and the system frequency dropped, the unit would pick up some load initially because the difference between TNR and TNH increased, but that load would be dropped because it would exceed the Preselected Load setpoint.

So, if the system frequency drops very suddenly the error between TNR and TNH will increase very suddenly and the SpeedTronic will increase the fuel very quickly, perhaps quickly enough to trip the unit on Exhaust Overtemperature, especially if the unit was being operated near rated load at the time of the frequency excursion. There is usually some kind of ramp rate on FSR increase/decrease just to try to prevent trips due to sudden fuel increases. Check your SpeedTronic elementary to see if ramp rate control exists.

Ambiguity is not the intent; this is not an easy topic to explain. My apologies for the verbosity. This author believes in the adage, "If you give a hungry man a fish, you feed him for a day. If you teach a hungry man to fish, you feed him for life."

Lastly, it would be very helpful to others reading these posts (and to the authors of the responses) if you would provide feedback if your issue is resolved or not--for all your posts. If you're not shy about posting questions, don't be shy about posting feedback--even if you're not satisfied with the response(s). Perahaps others could clarify the response or add to it or present it with a different perspective--but certainly, everyone could benefit from knowing whether or not your question was answered or your problem resolved with the responses posted on this site.

Having been involved with a small GT-powered generator on an industrial site in New Zealand, I have had plenty of opportunities to observe the behaviour on frequency changes - the 2 islands are connected by a DC link and the
connected load in each is relatively small in comparison with the largest generator. Frequency would dip from 50 Hz to 49.5 or so about once a week, to 49 about once a month, and occasionally to 48 - emergency load shedding
cut in at 47.5. The GT was run in standby mode at about 50% capacity, and when the grid frequency dived would try and single-handedly pump the system back up to 50 Hz. I have seen the nominally 2.5 MW generator put out over 3.5 MW on the panel power indicator - but for only a couple of seconds. Time delays and other dynamic effects, both within the set and in the external system, make this whole behaviour on load changes/frequency changes very complex.

Cheers,

Bruce

There is one thing which has been alluded to but never really epxlained in the context of SpeedTronic and GE heay-duty gas turbine control philosophy which might be of help.

The amount of Droop, someitmes called the "percent of regulation," reflects the value of the turbine speed reference at rated power output of the unit. For example, if a unit has a 4% Droop setpoint the turbine speed reference at rated power output will be approximately 104% (assuming the system frequency is at rated and stable). For a heavy-duty gas turbine, rated power output is related to a number of different factors: compressor cleanliness, inlet filter cleanliness, hot gas path part condition, ambient temperature, etc. For a heavy-dut gas turbine, rated power output will be considered to be CPD-biased exhaust temperature control, Base Load.

When the unit is at Full Speed-No Load (FSNL, synchronous speed, 100% speed, etc.), TNR is 100% (very slightly higher for synchronizing purposes). At 101% TNR, the unit should be at approximately 25% of rated power output. At 102% TNR, the unit should be at approximately 50% of rated power output. at 103% TNR, the unit should be at approximately 75% of rated power output. And, at Base Load, TNR should be approximately 104% (on an isobaric day, with a new and clean compressor, with reasonably new hot gas path parts, with reasonable clean inlet filters, etc.).

So, since TNR is related to turbine speed, and hence electrical grid/network frequency, if TNR remains constant while not on CPD-biased exhaust temperature control and the system frequency decreases then the amount of fuel will increase--until the unit reaches CPD-biased exhaust temperature control, or, until the error between TNR and TNH reaches approximately 4%.

Yes; it IS a very complex subject which is affected by a large number of factors. The design of the electrical grid/network, the number of generators, the number of transformers, the types of loads (resistive, inductive, capacitive), the types of transmission media--there are just so many variables.

To this author's mind, the real problem is: Why is the system frequency dropping? How much load is being suddenly beinng put back on the system when one or more units trip? How much "spinning reserve" is generally kept on the system for just such occurrences? Is there a large unit being operated in Isochronous mode somewhere on the system that can't pick up the load to maintain system frequency?

This was a difficult topic in university; it's still a difficult topic even with decades of job experience. In this author's experience, there are only a few power plant operators (and technicians) who really understand what they're doing from this perspective. Most just twist a switch and watch load go up and down, and don't have any understanding of what droop and isochronous control mean and how they work. It can be made even more problematic when documentation is non-existant or uses terms which can be misleading in certain contexts.

Fortunately, most electrical systems are fairly stable things, and most operators and technicians don't ever seem to experience the problems the originator of this post is experiencing so they don't really need to have a good understanding of the dynamics and the variables. It would be really helpful to understand if this problem has been ocurring for some time at this site, or if it has just started recently.

well, The gas turbine load control can be done from mark iv local control or remote control. if mark iv in local control other machine do not know about this machine unavailability. but it is not the correct way to do load sharing. Basically the load sharing control done by seperate control system(PMS). Assume we have 4 machine with 250 MW each. First we have to set the base load, minmum load, maximum load(only emergency) in PMS. now PMS give remote setpoint commant to each system. it is closed loop control. if any one machine trip then the PMS will issue maximum load demand to each system. from there mark iv system control system adjust the load to it is remote setpoint.

I have another question then, what if I have two turbogenerator units, one on Isoc and the other on droop, and the isoc TG tripped for any reason, the droop TG will definitely pickup the load, but what will make it see it and pick it, is the freq, or the voltage... or what??

Regards

markviguy, huh, ???

Had to do a double-take there...! Almost looked like markvguy was asking questions! (He is; he wants to know when the Droop questions are going to end....)

The original poster of this thread had Mk IV units....

Okay, let's say two generators are connected to a small "grid", one's prime mover is operating in Isochronous speed control and the other is operating in Droop speed control with a Droop value of 5%. Let's also say the Isoch unit has a rated output of 50 MW and the Droop unit has a rated output of 50 MW. Further, let's say the total load on this small grid at this instant in time is approximately 40 MW, and the governors of the two units are adjusted such that the Isoch unit is supplying 10 MW and the Droop unit is supplying 30 MW and the grid frequency is excactly at 100% of rated (50- or 60 Hz--it doesn't matter). This would meand that the Droop machine's Droop Speed Control reference (setpoint) would be 103% (since 30 MW is 60% of 50 MW, 60% of a 5% Droop setting would be 3%).

In your scenario, something causes the Isoch unit's generator breaker to open--meaning that there is no control of grid frequency. The electrical load on this small grid hasn't changed--it's still 40 MW of motors, lights, transformers, etc. The immediate effect on the small electrical grid would be that the frequency would drop.

The immediate effect on the Droop machine is that the actual speed of the generator and the prime mover drops [because the grid frequency drops: N = (120 * F) / P ], so the error between the Droop Speed Control Reference and the actual speed increases by approximately 1% (since 10 MW is equal to 1% of Droop)--which causes the Droop machine's governor to increase the amount of energy being supplied to the prime mover to power the electrical load of the grid (40 MW).

The grid operators in this scenario are "lucky", in that the total load on the grid did not exceed the Droop unit's rated output! The lights stayed on, but the grid frequency dropped.

Now, the grid operators CAN use the Droop machine's governor to increase the energy to raise the frequency back to rated (probably wouldn't take much)--but if there is any appreciable change in load, the grid frequency will change! If the Droop unit is capable of Isoch operation, they could also select Isoch and let the unit control the frequency until the problem with the other unit could be resolved and re-connected to the grid; that way, they could go back to sleep and not worry about grid frequency changes as load changes. (Isn't automation wonderful?)

markvguy

I am new to this forum and want to thank you for the education you have given me regarding droop and Isync control. You seem to have a very great understanding of generation control and the MarkV fuel controller. I am currently a novice working with the Mark VI and look foward to being 10% as knowledgable as yourself. Thanks for the education!

Keith

Please read the step load as 20-25% instead 5%, in my previous post. it was a typo error.

The way I understand, the system works in this fashion:
Suppose 3 machines are running in synchronisation and let's say each is having a load of 20 MW, i.e. total 60 MW.
Now if one of them trips, 20 MW will be dumped on the other two. This is because although 1 generating unit will be lost, electrical loads will remain the same, i.e. 60 MW. Now additional current equivalent to 20 MW will be supplied by the other two units. The additional load (power) on the generator is compensated by a proportionate drop in frequency of the generator in keeping with the the droop characteristic of the governor. Now if the speed of the turbine drops it has to be adjusted manually in case of machines in droop mode. But if one of the machine is in droop mode and the other in isoch., the isoch will try to regain its original set frequency (let's say 50 Hz) by increasing the FSR/opening of its gas control valve. Most of the load will be taken by the isoch machine. And the other machine will be forced to adjust its frequency (by adjusting its load) as they are in synchronisation.

Now here another factor is there, whether a machine of 30 MW running with 20MW laod will be capable of handling 10 MW or more at one go or step. Here the step-load concept comes in and mostly 5% of the rated load is considered safe for step loading in a machine and additional load is fit to be disposed by means of automatic loadshedding.

If someone would like to put additional inputs, it would be welcome.

Regards,
jcd