I desire to calculate the %age droop setting of turbine at our plant but I am having some problems.

Normally 4% droop is claimed by GE manuals. There are two gas turbines at our plant, one operating on Iso and other on part load droop mode. The turbines have design output of 24.5MW at base load and 26.45MW at peak load. Design conditions are 15 deg C ambient temperature and 14.7 PSI atmospheric pressure. It is running on natural gas and has HRSG on exhaust. Since a long time has passed after its manufacturing now we normally get 18MW at base load and 19.2MW at peak load.

Now the question. When we talk about %age droop setting and say that "it is amount of frequency (or speed) drop when the load moves from 0% to 100%", is this 100% load the base load or peak load?

Secondly, when I saw the reference speed of my turbine, it was 101.74% when load was 13.6MW (ambient conditions are 40 deg C temp and 14.3psi atmospheric pressure). Even for design conditions, at peak load and 4% droop, the reference speed should be 102% for 13.2MW (am I right?). How come this reference speed is there? I wanted to know the factors which have decreased (or changed) its droop setting. Shouldn't the droop value increase with the passage of time? Isn't the reference speed setting calculated with help of base load value?

We have one turbine operating with MarkV and other with MarkII control system. To get the exact droop setting, I have seen the control constants in CSP. Somebody on this forum told that it is DWKDG constant that says the droop setting. We have its value 0.05%/MW. Other than this there is also a constant FSKRN2 (called as 'droop speed correction factor') whose value is 7.969%/%. The droop setting calculated on base of DWKDG constant comes out to be very low (1.2%) so there must be some other constant governing the droop setting. Can anybody tell the control constant that shows the droop setting or that leads to calculation of droop setting?

Thanks and Regards.

 

What the ..!..?..?..!..??

Did the name of this site get changed from www.control.com to www.droopcontrol.com when I wasn't looking? (Nope; I just tried www.droopcontrol.com and neither my web browser nor google nor dogpile.com could find any website by that name or alias.)

I'm going to apologize right now for the length of this response, but it's going to serve as the last response to droop-related questions that I'm going to make. After reading many of the older threads on this topic on control.com (in which I learned a thing or two!) and responding to a few myself in the past year or so, I believe I have finally come to understand what most of the confusion is about. I'm going to try to summarize several years of questions and answers and put this line of questioning to rest for a while. So, this is going to be a rather lengthy response, but I believe it will answer your question, SWAZ, if you are patient, and those of others who might be reading this thread.

Thank you for the good information you have provided; it should serve as a model for other people wanting help with their questions. A couple of important pieces of information that you didn't include
were: the FSRs at the various conditions you cited, and it would have
been helpful to know what speed the unit was operating at in the conditions you cite: rated or higher or lower than rated? Also, it would have been helpful to know if the unit were equipped with IGV Exhaust Temperature Control (sometimes called "Combined Cycle Mode") and if it was enabled during the conditions you cited.

I want to answer one of your questions about your assumptions regarding droop speed control first. The droop setting of a GE-design heavy duty gas turbine equipped with a Speedtronic turbine control system refers to the power output of the turbine at nameplate-rated Base Load *unless otherwise noted*, for a unit which is in a new and clean condition being operated at the conditions listed on the nameplate for Base Load, not Peak Load. Peak Load capability is something over and above Base and not every machine is purchased with Peak Load capability, nor can every machine be configured to provide Peak Load capability (without severely reducing hot gas path parts life). Most units are only operated at Peak Load for brief periods of time, and this is not the optimal operating condition for GE-design heavy duty gas turbines; it is only intended for limited periods of operation ("emergencies").

Many units are rated at 59 deg F (15 deg C), 60% Relative Humidity, at the local elevation above sea level, as yours seems to be. Others are rated at 90 deg F (32 deg C), 60% Relative Humidity, at the local elevation above sea level. The conditions you describe the units are being operated at are significantly above the nameplate rating, which would mean the power ouptut of the unit would be less than nameplate. New and clean condition means the axial compressor is clean, the axial compressor inlet filters are clean, and the hot gas path parts are relatively new and in good condition, with proper clearances and seals. Achieving rated output also requires that the IGV LVDTs be properly calibrated and that the CPD transducer(s) are properly calibrated. (These two things are usually checked/adjusted prior to any contractual performance testing, and exhaust duct back-pressure is also usually monitored as are axial inlet pressures.)

When unit performance tests are done, they can rarely be done at the exact nameplate-rated conditions, so there are curves for correcting the performance back to nameplate-rated conditions (these are usually supplied with the units in the unit instruction manuals; please, nobody write to ask for a copy; look in your instruction manuals or ask your packager or the OEM).

The characteristics of the fuel(s) being burned over time will also have a small effect on the power output of the machine versus the droop characteristic. I would also say that for some machines the gas valve condition would even have some small effect on the power output of the machine versus the droop characteristic.

Units being operated in combined cycle mode (with HRSGs) at part load usually have some form of IGV Exhaust Temperature Control enabled to maximize exhaust temperature, which reduces turbine efficiency slightly in order to maximize overall cycle efficiency. If IGV Exhasut Temperature Control is enabled at part load operation, it can adversely affect the apparent droop setting (not the droop setting, just the appearance of the droop setting versus the power output; read on).

So, there are lots of factors affecting turbine power output at any given instant in time. Many people choose not to buy their hot gas path parts from GE or the packager, and while they are functionally equivalent, they may or may not be built to the same tolerances or perform equally as well (and some may actually be superior). Some owners buy refurbished parts, also, which usually won't perform as well as new parts but are functional equivalents. Some owners buy air filter replacements from different vendors, which have different air flow characteristics which can also affect power output. Some owners also make modifications to their HRSGs which can increase the back pressure on the turbine exhaust. Some owners add HRSGs without having the control curves for the gas turbine re-calculated to reflect the added exhaust duct back pressure. When a nameplate rating is calculated for a gas turbine, it is for a new unit, which includes the filters being provided, and the exhaust components being provided. If these change over the years, then the power output will also change at Base Load, regardless of the ambient conditions.

Droop speed control is amazingly simple, and yet very powerful. It is how generators can be operated in parallel with each other in a stable fashion and supply loads that are much larger than any single generator, *and* when grid frequency is unstable it is how prime movers can increase or decrease their output to assist with maintaining the load on the grid, which in turn actually helps support the frequency of the grid. The first sentence of this paragraph, the part about generators (and their prime movers!) being operated in parallel with each other in a stable fashion is the "sharing" part of one of the common definitions of droop speed control. Without droop speed control, or something equivalent, generators and their prime movers could not be operated in parallel with each other.

I agree with a previous contributor in an earlier thread about the definition of droop that you cite: That definition is one of the *worst* definitions ever. Technically, it's correct, but in the real world we never operate machines like that. Not ever; never. It's just an absolutely insane definition, and not a practical definition by any stretch of the imagination, and it contributes to this "mystique" about Droop Speed Control that has so many people baffled.

The other common definiton used to describe droop speed control talks about "sharing the load" which is one of the two main features of droop speed control; the other is how the governor (the control system) of the prime mover responds to changes in frequency. The common "sharing the load" definition is also technically correct but equally obtuse and leads to even more confusion. The part of this definition about increasing or decreasing output to assist with maintaining the load on the grid and how that is done is related to the definition you cited. Actually, the people who perpetrated these two definitions on us each tried to use a single definition to cover two distinct, but related, features or aspects of droop speed control and in the process have created a lot of confusion and misunderstanding.

For the overwhelming majority of the time when units are operating in parallel with other units, droop speed control doesn't really care about the actual turbine speed; it's straight proportional control and presumes that some other machine or over-riding "governor" is controlling grid frequency which directly affects turbine speed. When the turbine speed reference is 102.6%, the turbine is still running at 100% speed; because droop speed control is straight proportional control there's nothing to drive the actual turbine speed to be equal to the turbine speed reference. In other words, the actual turbine is being allowed to droop, or be less than, the turbine speed reference.

I can hear the screams already; you're saying, "Droop speed control depends on the turbine speed reference and the actual turbine speed!" And you would be almost 100% correct, except for the fact it's not the value of either the turbine speed reference or the actual turbine speed at any given time, it's the *DIFFERENTIAL* between the turbine speed reference and the actual turbine speed that drives droop speed control. *AND* that each of the features or aspects of droop speed control makes an assumption: that either the actual turbine speed, or the turbine speed reference, will be constant for either of two conditions.

On a normal grid which is operating at a stable frequency, the speeds of the generators and the prime movers driving them will be equally as stable as the frequency. So, when the turbine speed reference increases the actual turbine speed doesn't increase, and droop speed control relies on that fact. As the reference increases and the actual speed remains the same, in other words as the differential increases, the governor admits more energy to the prime mover, which the generator converts to amperes. This is the first feature or aspect of droop speed control that I was referring to, increasing and decreasing the energy in relation to the differential between the turbine speed reference and the actual turbine speed. This feature or aspect of droop speed control is predicated on the assumption that the actual turbine speed (which is a function of the grid frequency) is stable and isn't changing.

So, this aspect of droop speed control uses the differential between the turbine speed reference and the actual turbine speed to control fuel flow (energy) to the turbine. When an operator twists the Speed/Load switch on the Generator Control Panel in the Raise direction or clicks on Raise Speed/Load on the operator interface screen, the turbine speed reference is increasing, and that increases the differential with the actual speed (which should be constant on a stable grid) and the turbine control increases the fuel flow (energy input) to the turbine and that energy becomes torque which the generator converts into amps. When an operator twists the Speed/Load switch on the Generator Control Panel in the Lower direction or clicks on Lower Speed/Load on the operator interface screen, the turbine speed reference is decreasing, and that decreases the differential with the actual speed (which should be constant) and the turbine control decreases the fuel flow (energy input) to the turbine, which results in less torque being produced which results in fewer amps being generated. So, this feature or aspect of droop speed control uses the differential between the turbine speed reference and the actual turbine speed (which is assumed to be constant, and is for most grids for the majority of the time) to increase or decrease the fuel. The turbine speed isn't changing, but the turbine speed reference is and the differential is being used to control fuel flow.

Another way that the differential between the turbine speed reference and the actual turbine speed (which is directly proprotional to generator frequency) can change is if the grid frequency changes. If the grid frequency decreases, the differential will increase; if the grid frequency increases the differential will decrease. And droop speed control will change the fuel flow in proportion to the change in the differential, just like when the operator changes the turbine speed reference.

So, for the second feature or aspect of droop speed control, another assumption is made with regard to the speed differential: that the turbine speed reference is not changing when the grid frequency is changing. When a GE-design heavy duty gas turbine with a Speedtronic turbine control is being operated at Part Load *WITHOUT PRE-SELECTED LOAD CONTROL ENABLED OR REMOTE LOAD CONTROL ENABLED* (that's a *VERY IMPORTANT* distinction), the turbine speed reference is constant (unless the operator is changing it). (The way that Pre-Selected Load Control or Remote Load Control is usually implemented in a Speedtronic turbine control system is that the turbine speed reference is raised or lowered in response to the differential between the turbine power ouput and the load control setpoint; that's why the distinction is made. This can actually have a negative effect on the unit response to grid frequency disturbances and is why it's important to make the distinction.)

The assumption for this second feature or aspect of droop speed control is that the turbine speed reference is not changing when the actual turbine speed is changing. This is the part of droop speed control that regulatory bodies are concerned with because it has to do with grid stability and response to grid frequency distrubances. And it's an equally important feature or aspect of droop speed control.

So, the droop setting of a GE-design heavy duty gas turbine with a Speedtronic turbine control system is really about the magnitude of the change in turbine power output for a given change in the differential between the turbine speed reference and the actual turbine speed. And that differential can change for either of two
reasons: operator-initiated load changes or grid frequency changes.
It's *NOT* about the instantaneous value of turbine speed reference versus power output or the actual power output versus the nameplate rating or anything else. It's about the change in power output for a given change in the differential between the turbine speed reference and the actual turbine speed, which should not change appreciably over time even though the turbine output might vary for any number of reasons as outlined above.

For a droop setting of 4%, the power output of the unit should change by approximately 25% of nameplate rated Base Load for each 1% change in the differential between the turbine speed reference and the actual turbine speed. And this should not change by very much over time, even if the machine is old, dirty, being operated in an ambient which is greatly different from the nameplate rating, has poorly calibrated IGV LVDTs and poorly calibrated CPD transducer(s). Again, it is this aspect of droop speed control that the regulatory agencies are most concerned with. The amount of load change for a given speed differential change should remain relatively constant, not exactly 100% constant, but relatively constant. We're dealing with real machines in the real world, not simulators in ideal conditions.

In your case, the unit output at Base Load is greatly reduced from nameplate for a number of reasons. Age and condition of hot gas path parts (including internal alignments and clearances), the ambient conditions the unit is being operated at, and we have no idea about the axial compressor condition (clearances, cleanliness), the inlet filter differential and flow-rate, the exhaust duct back pressure, and the calibration of the IGV LVDTs and the CPD transducers. But, this is typical for most machines of the age you describe. And it should *NOT* have any appreciable affect on the amount of change of power output for a given change in the differential between the turbine speed reference and the actual turbine speed. If the unit has a droop setting of 4%, the unit output should change by approximately 25% for each 1% change in the speed differential, regardless of how that differential is changed (either by the operators or by a change in grid frequency).

One thing which is maddening to many people when it comes to the Mark V is that the "set" of Control Constants which are present in the control system are "max case" which means that many of them, sometimes as much as 50% of them, are not actually used in a particular CSP. So, just because DWKDG exists in a Control Constants display does *not* mean it's used in the CSP for that unit. You would need to check the CSP for your unit to see if Constant Settable Droop was programmed in the CSP and if DWKDG was used and what, if any, other parameters might be affecting droop control. I've seen some pretty amazing (and just plain wrong!) things done by Field Engineers/ T.A.s just to get the units commissioned and finish the installation. Also, I've seen some Customers demand that changes be made to the turbine control system without really understanding exactly what the overall effect of those changes would be. So without being able to see your CSP it's pretty difficult to comment about how Droop Speed Control is implemented at your plant. Just because a Control Constant is listed in the Control Constants Display of your unit does not mean that Control Constant is actually used in the CSP of your unit. Or even that it's not somehow "manipulated" or modifed in the CSP of a particular unit.

There are basically two types of droop speed control: "straight" droop speed control, which doesn't take into account turbine power output at any instant in time, and Constant Settable Droop, which adjusts the power ouput based on the load transducer feedback as a function of turbine speed. Constant Settable Droop would typically use DWKDG, but straight droop speed control would not (typically). I should have corrected the respondent who said that DWKDG was *the* "droop setting" but didn't. Not every machine uses Constant Settable Droop (which is a horrible name for the function it performs!) and only by examining the programming of the Speedtronic can we determine which version is in use for that particular control panel.

DWKDG is expressed in terms of %/MW, where the % term of the engineering units refers to the differential between the turbine speed reference and the actual turbine speed. For a value of 0.05 %/MW, and presuming a Droop Characteristice of 4%, I would say the unit output "should" be approximately 80 MW (4 % / 0.05 %/MW = 80 MW). In other words, every MW of output would be worth 0.05% of turbine speed reference. Or, for each one percent change in turbine speed reference, the ouptut should increase by 20 MW, or 25% of rated.

I would expect that if Constant Settable Droop were implemented in the standard fashion on your unit, with a Base Load Rating of 24.5 MW, and presuming a Droop Setpoint of 4%, DWKDG should be approximately (4 / 24.5 = 0.1632) 0.16 %/MW. In other words, every MW would be equivalent to a change of 0.16 % of turbine speed reference. Or, every 0.16% change in turbine speed reference would result in a change of 1 MW. On a new and clean turbine being operated (or corrected to nameplate rated Base Load) the unit should reach Base Load when the turbine speed reference is approximately 104%. So, I'm going to go out on a limb and say that Constant Settable Droop isn't used, or isn't implemented in the standard fashion, or is modified by some other function or variable on your unit.

I would further say that at 13.2 MW, which is just slightly above 50% of nameplate rated load, and for the ambient conditions which you cited which are much higher than nameplate (at least the ambient temperature), that a turbine speed reference of 101.74% would be just about right for a Droop setting of approximately 4%. I would estimate that you would reach Base Load at a value much less than a turbine speed reference of 104%, since you say your normal base load these days is about 18 MW instead of 24.5 MW. But, the amount of change in power output for each 1% change in the differential between the turbine speed reference and the actual turbine speed should change by approximately 25% if the unit has a droop setting of 4%. Approximately because we're talking real world conditions here, based on calculations.

Constant Settable Droop also has some integral action with respect to the load (since it uses a "load-corrected version of actual speed in the calculation), so generally, there isn't too much drift on the droop setting. That's another reason I believe that Constant Settable Droop isn't in use, or is modified in some fashion, in the Mark V at your site.

Another possible reason for the differential you cited might be the Control Constants for "straight" droop speed control were never properly adjusted, or the Gas Control Valve LVDT calibration isn't spot on, or the fuel characteristics might have changed since the unit was installed, and it might just some combination of any or all of these conditions. From several previous threads on this subject straight droop speed control is usally accomplished with the equation: FSRN = ((TNR - TNH) * FSKRN2) + FSKRN1, where FSRN equals the Droop Speed Control Value of FSR for any value of *the differential between the turbine speed reference, TNR, and the actual turbine speed, TNH* (I think we've seen this somewhere before in this thread!); FSKRN2 is the "gain"; and FSKRN1 is the offset. If FSKRN1 is not equal to Full Speed-No Load FSR, or the Gas Control Valve LVDT calibration isn't spot on or if the fuel characteristics have changed since the initial installation (or some combination or any or all three factors), then the nominal droop setting might not be exactly 4%.

Without being able to see the CSP for the unit at your site, it's impossible to say exactly how the unit is programmed or whether Constant Settable Droop or "straight" Droop Speed Control is being used. Although Constant Settable Droop is used almost exclusively on most new units in recent years, it isn't always, and it has been known to be modified (rightly or wrongly) for a variety of reasons.

So, while it might seem that the unit doesn't have exactly 4% droop because it reaches Base Load at 103.1% turbine speed reference or it reaches 50% load at slightly less than 102% or turbine speed reference, the important point is the *change in power output for a given change in speed differential*. If the unit power output was 75% of rated at 101.74%, then I would be concerned that something is amiss with the droop setting, and in this case the unit would probably not respond to grid frequency disturbances as desired. Or, if the power ouput was 23.7% of nameplate rated at a turbine speed reference of 105.1% (which I've personally seen) then I would say the unit definitely does not (and in that case did not) have a droop setting of remotely close to 4% and would not respond to grid frequency disturbances as a unit with a 4% droop setting would.

For a unit with a droop setting of 4%, the change in power output of the unit should be equal to *approximately* 25% of nameplate-rated Base Load for each 1% change in the differential between the turbine speed reference and the actual turbine speed. If the unit doesn't reach Base Load at exactly 104% turbine speed reference, that doesn't mean the droop setting isn't 4%. As we have seen above, the ability of the unit to produce rated output is affected by a lot of factors, but it doesn't affect the "nominal" droop setting of the unit. The "nominal" droop setting (and by nominal, I mean the planned or designed value: the value which is set or programmed into the turbine control system) is a function of the amount of change of power output for a given change in the speed differential.

I think what a lot of the recent queries about droop control (and this one, too) are related to what might be termed the "instantaneous" droop setting, based on the actual or current Base Load output versus turbine speed reference, not the nameplate rated Base Load, or the instantaneous value of power output versus the instantaneout value of turbine speed reference (without taking into accout the actual turbine speed, because we are presuming the units are being operated at rated speed and I'm becoming more and more aware that in many parts of the world the grid frequency is consistently below rated and that would have an effect on the "instantaneous" droop setting). And, all of these values should be regarded as approximations, since they are calculated values. Actual machine conditions and characteristics will affect the real numbers. Even generator conditons and PT (Potential Transformer) conditions can have an effect of the real numbers. Again, we're talking about real turbines operating in real world conditions, not simulators or ideal turbines in ideal conditions.

Lastly, EVERYBODY *PLEASE*, *PLEASE*, *PLEASE*, *PLEASE* **DO NOT** START CHANGING ANY CONTROL CONSTANTS TO TRY TO MAKE YOUR "INSTANTANEOUS" DROOP VALUES EQUAL TO A DROOP SETTING OF 4%!!!! Don't change it for any reason!!! *YOU WILL ADVERSELY AFFECT LOTS OF THINGS, INCLUDING LOADING AND UNLOADING RATES!!!!* IF YOU DO CHANGE IT WITHOUT UNDERSTANDING ALL THE RAMIFICATIONS, YOU WILL BE CHANGING THE "NOMINAL" DROOP SETPOINT WHICH MEANS THE UNIT WILL REACT DIFFERENTLY TO CHANGES IN THE DIFFERENTIAL BETWEEN THE TURBINE SPEED REFERENCE AND THE ACTUAL TURBINE SPEED, WHETHER THE DIFFERENTIAL IS A RESULT OF THE OPERATOR INCREASING OR DECREASING LOAD OR IF THERE IS A GRID FREQUENCY DISTURBANCE. (In other words, "Kids, don't try this at home!" There, the warning has been issued; absolution is complete. Don't complain here if things go awry if Control Constants are changed.)

Again: WHAT MATTERS IS: HOW MUCH DOES POWER OUTPUT CHANGE FOR A GIVEN
CHANGE IN THE SPEED DIFFERENTIAL. The instantaneous values of turbine speed reference, or actual speed, or actual load versus turbine speed reference have nothing to do with the "nominal" droop setpoint and how the control system reacts to changes in the speed differential.

That's it. Period. Droop speed setting is about the amount of change in power ouput for a given change in the differential between the turbine speed reference and the actual turbine speed.

Nothing less, and nothing more.

And, with this, my contributions to droop-related threads end.

Even if they change the name of this site to www.droopcontrol.com.

Again, I apologize for the length of this response (and to control.com's moderator for posting it!).

 

An excellent and comprehensive explanation...

Bruce.

 

Thank you CSA for such a comprehensive and detailed reply. It really helped me a lot, and for everyone on the forum, in fact.

Yes, its right that we have 'straight droop control' implemented in our system. I have not found DWKDG in the CSP. The logic for droop speed control in implemented with the same formula that you told i-e FSRN = ((TNR - TNH) * FSKRN2) + FSKRN1. When Isochronous mode is selected, another factor is added in the above thing which is (TNRI-TNH)*FSKRN3, which ultimately makes the Isoch control.

The FSR values for our system are 18.2% for FSNL (FSKRN1 in above formula) and 36% for Base load. As far as 'change in power output for a given change in speed differential' is observed, the %age droop factor still seems to be deviated. Although we maintain a high standard of maintenance at our gas turbines but still I think this deviation may be due to GCV LVDT calibration.

So, the crux of the story is that it is the slope of the 'power v/s speed differential' curve that matters, not the instantaneous value.

Thanks again.

 

Can you explain the meaning of symbols in detail those are used in the formula?

 

It was really excellent explanation about droop control. I am having below queries.
We are having Siemens make extraction cum condensing turbines of capacity 105 MW x 2 nos and 93 MW x 1 no. its CPP plant for oil refinery.
We are always connected with grid and exporting around 20 MW power to grid and rest power will be utilized in refinery itself.
We have not tested these turbines for islanding operations and once when we were islanded from grid whole plant was blackout as these turbines could not control the frequency.
Can you help us on this
Suresh Kalaskar
[email protected]
8238015436

 

sukalaskar,

Were the turbines/plant DESIGNED (intended) to operate independently of the grid (island operation)?

These days when there are multiple generator sets (generators and their prime movers) to be operated in island mode there is also usually some kind of external 'power management system' which will adjust load(s) on the various units (or most of them anyway) and attempt to control frequency, usually without using Isochronous Speed Control governor mode on one (or more) of the generator sets. At least that's what these PMSs (Power Management Systems) are supposed to do--but many of them are not properly configured or adjusted or tuned, and there are usually so many variables (load shedding, etc.) that it's difficult to coordinate everything all the time under all circumstances and scenarios.

In the past, this was historically done by switching one generator set's governor to Isochronous Speed Control mode and leaving the others in Droop Speed Control--but sometimes this isn't properly configured or tuned and plant needs usually change over time but no one thinks or remembers to consider island operation. And, it really, Really, REALLY requires knowledgeable, trained and experienced operators (which are in very short supply these days--what with owners and Corporations wanting to automate everything based on what the automation salespeople say the automation can do (can we say vaporware?)).

You haven't provided enough information about the configuration of the plant and it's design for us to be able to help you. You should really get someone knowledgeable in large, islanded operations to come and work with power plant and refinery personnel to determine what's required, and work from there. It could well be that the plant was designed for islanded operation, but conditions when the scenario you briefly alluded to were unusual and weren't considered in the design. We just don't know enough to be able to be of much help.

Sorry!