First post in this forum, so forgive me if this question has already been asked/answered. I searched through old posts, but didn't find it

I am working in a corporate engineering office, so my access to several Mark V's is limited to what I can convince my buddies at the plants to do for me (sometimes they have better things to do than fool with my stuff, so I may not be able to get data in a short amount of time). Our operations are US operations, i.e. 60Hz. My terminology is consistent with:

http://www.wecc.biz/documents/library/generator/wsccgov1.pdf

Does the Mark V have an intentional frequency (speed) deadband? If so, is it typically set at 0.036Hz (or some lesser value to give an effective overall value of 0.036Hz)? Is it applied after the Droop calculation, before the low select block? How do I check to see what the deadband is set at in our Mark V's?

Does the Mark V have an unintentional frequency deadband (0.005Hz in the article)? If so, what is a typical value for the Mark V? Is there a quick an easy test that is used to nail that number down?

Thanks,
Bill

 

MANY THANKS for the very good document; now we'll never have to answer another droop-related question again! We can just refer to this document! (Oh, if this were only true! Documents like this just seem to breed more questions, as they generate more questions than they answer sometimes!)

As a GE turbine-generator "owner" you should be able to ask these questions via a PAC (Power Anser Center) Case.

To this author's knowldege, there is no "intentional frequency deadband" and it is not known if there is an "unintentional frequency deadband."

Let us know what you learn--and thanks again for the fine treatise!

markvguy

 

As far as I know, you are right (I'm still working on an official answer), the Mark V does not have an intentional deadband.

The problem is that certain units on the grid have some form of a deadband and it's very likely that it's intentional (giving them a cost advantage over the rest of us).

The question that keeps coming up is... Why doesn't the Mark V have an intentional deadband?

Bill

 

While the paper you cited is not very clear on the subject, one would think that "unintentional" frequency deadband--called inherent frequency deadband in the WECC document--would be due to characteristics of the control system such as response timing, actuator slew rate, etc. How to measure that in the field would be a good question; it's probably something that's modeled on some computer somewhere during the design review of a control system, but is not published or commonly known or easily obtained.

>From the paper, it seems intentional frequency deadband would be something like introducing a lag time delay in order to prevent hunting or oscillation of the prime mover governor.

The paper makes an important distinction between droop speed control (or, speed droop control as it's called in the paper) and speed regulation. The latter uses a load (watt or megawatt) transducer feedback in addition to speed error, and adjusts the megawatt output to provide additional load during a low frequency disturbance. GE's 'Constant Settable Droop' can probably be more correctly termed speed regulation

The paper also makes an important point in Sect. 3.0. That point is that even with speed droop control (or, droop speed control...) or speed regulation control that governors alone cannot restore grid frequency to normal levels during a frequency disturbance. Their power output will increase--but ONLY as a result of the frequency disturbance--and to restore frequency some external means (AGC or manual operator intervention) must be used to increase fuel and, ultimately, speed/frequency.

Thanks again for providing the link to this document! It is providing much food for thought and contemplation. It is sure to generate a question or two here on control.com.

markvguy

 

It's really unclear (to this author) what is meant by a cost advantage being "given" to units with "some form of a deadband", intentional frequency deadband to be exact.

And how is it known that certain units have this intentional frequency deadband. It would seem the result of such a deadband would only be to "delay" response to frequency excursion, and it's hard to understand how that would be a cost advantage (unless one would think that this would result in some kind of fuel-savings....).

It was this author's understanding that organizations such as Western States Coordinating Council "encouraged" units to respond to frequency distrubances to try to help maintain a stable grid frequency. They even mandate testing and documentation of results on a regular basis.

Gas turbines (which have the largest concentration, by far, of units with Mk V turbine controls) would seem better suited to respond to frequency excursions, they should do so faster than steam turbines or hydro turbines--unless they are on exhaust temperature control, which a WHOLE 'NOTHER subject.

So, could you please explain why it would be advantageous--from a cost perspective--to have intentional frequency deadband.

It's sure that if GE has implemented such a scheme because their Customers have determined it to be economically advantageous, they've attached an "appropriate" price to the option. But this author's not familiar with the concept or the economic advantages.

So, your question is better asked of GE! But, in the meantime, can you 'splain what you meant to the rest of us?

markvguy

 

Well, I got my answer.

There is an intentional deadband in the Mark V. It is set by constant TNKEDB. A typical value is 0.14Hz on 60Hz machines. So far, the statement has been made that that number is not adjustable. But, we all know that money talks.

The unintentional deadband is whatever slop/backlash is in the turbine/generator. If you want that number, you'll have to extract it from your own data (which is exactly what I've been doing over the past couple of days....lotsa gremlins in that task).

Bill

 

> There is an intentional deadband in the Mark
> V. It is set by constant TNKEDB. A typical
> value is 0.14Hz on 60Hz machines.

Oops, that should be 0.014Hz.

> And how is it known that certain units have
> this intentional frequency deadband.

When there's an "event". The EXPECTED system droop (from all generators) is typically somewhere between 5% and 8%. The actual ERCOT droop (from "event" data) has recently been anywhere from 15% to 22%. In other words, some governors are blocked (intentional???), and others have such a wide deadband that they don't "kick-in" until well after the "event" has started. From the data, it's real obvious who's doing what.

Right now droop is not a "paid" service in ERCOT. However, if they want to keep things stable, that will have to change.

> unless one would think that this would result
> in some kind of fuel-savings.

What's cheaper, to hold the gas valves fairly still, or to open/close those valves following the urges of frequency? Think about what's going on with the air/fuel numbers when you keep moving the valves. Also think about the wear/tear on the turbine and compressor that is avoided if you can minimize raising/lowering the load.

> It was this author's understanding that
> organizations such as Western States
> Coordinating Council "encouraged" units to
> respond to frequency distrubances to try to
> help maintain a stable grid frequency. They
> even mandate testing and documentation of
> results on a regular basis.

"Encouraged" is not the same as monetary penalties/rewards.

> Gas turbines (which have the largest
> concentration, by far, of units with Mk V
> turbine controls) would seem better suited to
> respond to frequency excursions, they should
> do so faster than steam turbines or hydro
> turbines

If frequency was a "paid" service we would get an economic answer to that idea. Right now, all units with functional governors that aren't at some limit are "encouraged" to participate. Pretty sloppy rules if you ask me.

> It's sure that if GE has implemented such a
> scheme because their Customers have determined
> it to be economically advantageous, they've
> attached an "appropriate" price to the option.

We'll have to wait and see how things evolve. This may or may not be just a GE issue.

Bill

 

Thanks for sharing the information!

Are you referring to Mk Vs being used with heavy-duty gas turbines or aero-derivative gas turbines?

On heavy-duty gas turbines, TNKEDB is only "active" if L83SCDB is true (a logic "1"). Several heavy-duty gas turbine CSPs were reviewed and L83SCDB is enabled on some, not on others. And, the values of TNKEDB were all different than the one you listed, and appears to be Frame size-specific, but that's just a guess.

markvguy

 

> Thanks for sharing the information! <

You're welcome. Everyone wins when we don't hoard information.

> Are you referring to Mk Vs being used with
> heavy-duty gas turbines or aero-derivative gas
> turbines? <

Standard 7FAs.

> On heavy-duty gas turbines, TNKEDB is
> only "active" if L83SCDB is true (a
> logic "1"). Several heavy-duty gas turbine
> CSPs were reviewed and L83SCDB is enabled on
> some, not on others. And, the values of TNKEDB
> were all different than the one you listed,
> and appears to be Frame size-specific, but
> that's just a guess. <

Actually, the number I was given is 0.02344, so to make the calcs match, change the deadband to 0.014064Hz on a 60Hz machine:

TNKEDB = 100 * 0.014064/60 = 0.02344

What CSP document/manual are you looking at? Do you have a CSP for a typical 7FA? If so, is it available? e-mailable? Does this site have a place to post .pdf's, .gif's, and other files? If not, do you know of another site?

Bill

 

> While the paper you cited is not very clear on
> the subject, one would think
> that "unintentional" frequency deadband--
> called inherent frequency deadband in the WECC
> document--would be due to characteristics of
> the control system such as response timing,
> actuator slew rate, etc. <

There's slop in any machine. It gets bigger as the machine ages. You either fix it, or model it.

> How to measure that in the field would be a
> good question; it's probably something that's
> modeled on some computer somewhere during the
> design review of a control system, but is not
> published or commonly known or easily obtained. <

As far as I know, you have to extract it on your own.

> From the paper, it seems intentional frequency
> deadband would be something like introducing a
> lag time delay in order to prevent hunting or
> oscillation of the prime mover governor. <

I'm fairly sure that it's to keep from bumping the valves unless you have to. From a modeling point of view, it changes the nice damped response signal (like you see in a college textbook) to an ugly non-linear squirmy-thing. It also complicates what is the "right" output from the AGC that is feeding the Mark V.

> The paper makes an important distinction
> between droop speed control (or, speed droop
> control as it's called in the paper) and speed
> regulation. The latter uses a load (watt or
> megawatt) transducer feedback in addition to
> speed error, and adjusts the megawatt output
> to provide additional load during a low
> frequency disturbance. GE's 'Constant Settable
> Droop' can probably be more correctly termed
> speed regulation. <

Here's what GE says about their term "Constant Settable Droop":

"....Constant settable droop control is a
special speed control system capable of
accommodating variation in gas fuel heating
values. When operating on droop speed control,
this system maintains the droop relation even
as the heating value of the fuel changes. To do
this a watts transducer and algorithms “Turbine
Load Droop”, DWDROOP, “Load Turbine Speed
Reference”, TNRL and “Setpoint Tracking” 70TRACK
are added to the typical software program. Refer
to the Control Sequence Program and Control Specifications for details of the Constant
Settable Droop Speed Control System...."

> The paper also makes an important point in
> Sect. 3.0. That point is that even with speed
> droop control (or, droop speed control...) or
> speed regulation control that governors alone
> cannot restore grid frequency to normal levels
> during a frequency disturbance. Their power
> output will increase--but ONLY as a result of
> the frequency disturbance--and to restore
> frequency some external means (AGC or manual
> operator intervention) must be used to
> increase fuel and, ultimately, speed/frequency. <

That was the question that I asked on another thread. The governor is the primary method of reacting to changes in frequency. However, it won't correct it (no integral). That's what the ISO does, add/subtract generators/loads in an attempt to balance supply/demand without cooking transmission lines. The frequency control integral term is where the ISO forces the Area Control Errors to cross zero ever so often, or has incentives to minimize supplier's Schedule Control Errors.

> Thanks again for providing the link to this
> document! It is providing much food for
> thought and contemplation. It is sure to
> generate a question or two here on control.com. <

Thanks for the answers. It's rare to find anyone with experience in this area. I'm still learning and have a long way to go.

Bill

 

The ORIGINAL intent of Constant Settable Droop was for units which used fuels from different sources. For example, take a unit which runs on natural gas and some kind of refinery "tail gas"--some by-product of a refining process which would otherwise be burned in a flare to dispose of. This by-product gas may have a very high BTU content or heating value, which means that for the same power output the Gas Control Valve might be only at 20% stroke instead of 65& stroke. Or it might be at 90% stroke instead of 55%....

Constant Settable Droop was developed to factor into the Droop Speed Control equation the actual load produced by the unit, regardless of the fuel control valve "position."

Why was this necessary? Because gas fuel flow-rate and BTU content/heating value are NOT used in the control and protection of the unit. Rather, a relationship between Gas Control Valve position and flow-rate is used to derive the calculations for positioning the Gas Control Valve, based on flow-rate/positon relationship.

That description you cited was VERY OLD. When DLN (Dry Low NOx) control was first being developed, there were very large load swings when fuel was being "transferred" betwen combustion zones, and Constant Settable Droop was applied to help reduce the load swings.

Droop Speed Control is simple proportional control--only when Constant Settable Droop is applied is there any integral action. AND ONLY UP TO EXHAUST TEMPERATURE CONTROL. Once a heavy-duty gas turbine reaches exhaust temperature control, it behaves exactly opposite of what one would like during frequency excursions....unless some kind of additional over-riding control is used, such as in the UK and some other parts of the world (coming soon to a North American location near you, though!).

Subsequently, Constant Settable Droop was applied to all units and it really does assist with load control during frequency excursions, also--while operating at Part Load on Droop Speed Control.

markvguy