We are tuning to governors on our large Island system and we have a couple units with Mk V controllers. One set is on two 66MW steam turbines (Mk V triple redundant, EX2000, EHC, WinNT HMI). Other one is on a frame 5 40MW gas turbine running liquid fuel (diesel) ONLY and an iDos.

I'm familiar with the system and organization of constants, etc. I've viewed the constant txt file. However, I am having trouble finding which constants control the characteristics of the governor in various modes like speed, isoc, base. We have a GE TA coming to help tune, but I'd like to get these values ahead of time to analyze.

I would like to know the following and any other applicable constants used: droop %, ramp rate, time constants, gain, mode change limits, etc. The steam units are very slow to increase load when running in droop. However, they reject load pretty fast (I can view trends using historian).

Secondly, on the gas turbine unit, the testing company wants to record fuel valve position. On the steam units he just places a temporary LVDT on the actuators--obviously this doesn't work on the liquid fuel valve. He doesn't have a LVDT signal conditioner so he can't "piggy back" on the position signal from the fuel valve.

Is it possible in MarkV Idos to create an analog output (voltage or current) of the fuel valve position? There are plenty of spare I/O's on the cards.

Thanks in advance.

 

Droop on GE steam turbine governors has always been a mystery to me. Mostly because most of the steam turbines I have worked on have been at cogeneration facilities, where they are primarily "followers" instead of "prime" generators. So, I can't help with that.

The subject of GT droop has been covered ad nauseum on control.com (which became speedcontrol.com for a while!). That's not an easy topic, either, but it has been covered very recently (actually, it never seems to die).

Well, since there is no feedback from the liquid fuel bypass valve on most Frame 5s, it would be pretty difficult to use a spare analog output and drive it with valve position. There is only a flow-rate reference, expressed as a percentage of maximum expected flow-rate, called FSR (Fuel Stroke Reference). For liquid fuel, FSR is actually converted to a liquid fuel flow-rate reference, and then speed feedback from the liquid fuel flow divider is used as the feedback. So, the liquid fuel control valve (the liquid fuel bypass valve) is positioned to make the actual flow divider feedback equal to the flow-rate reference, which is a function of FSR.

Your GE TA should be able to help you answer all of your questions.

The loading rates for the GT are usually contained in an array call TNKRn, where 'n' is a number from 0 through 7. BE VERY CAREFUL because the longname descriptions of those Control Constants can be wrong--and therefore, very misleading. One needs to look at the CSP for the particular unit in question and then determine exactly what each Constant does based on how the logic is configured/written.

 

Thanks for the info. On this thread: http://control.com/thread/1294347789#1294426997 it is shown how to calculate the droops. Funny how GE does things way more complicated than other governor makers.

I will look for the loading rate in the CSP and see what constant it calls. I've found the ramp rates allowed at the HMI (20% offline, 15% online) when the operator makes a speed reference change.

I'm sure there is a reason, but the other governors on our system are simply PI or PID loops. I'm just looking for the GE equivalent of gain and time constants. And of course the documentation is vague and varies from site to site.....

Hopefully we get a good TA........you'll probably find better, but you won't pay more (try $450 an hour right now + 10k mob/demob +400/day per diem). Wish some people here would list contact info so I could hire 'em.

 

The quote is:

"You may find better; but you'll never pay more."

In its purest form, droop speed control is proportional control, pure and simple. There is no integral component to pure, "straight" droop speed control, so if you're looking for one, stop. (That's regardless of the governor manufacturer, though I've seen some programmers try to use an integral component with some very, very poor results.)

You need to be very careful about what you read on control.com. That formula in the thread you cited is a very old GE formula that they have basically stopped using decades ago. (Some of the packagers of GE turbines have not stopped using it so it still makes it way into some newer packages. It works, but it's not the method you may find in your GT control panel unless it's a very early Mark V or it was provided by one of the packagers of GE-design heavy duty gas turbines and they chose to use that particular formula.)

The formula in that thread is for pure, straight proportional droop speed control. FSKRN2 is the gain.

You should also look in the Control Specification for the Mark Vs at your site (they were usually provided with every Mark V) and you might find some useful information in Sect. 3 about the droop method used in the Mark Vs at your site. (There should be one for the STs and one for the GT.)

Not too long after the Mark V was released, GE went to a form of droop speed control called Constant Settable Droop Speed Control for gas turbines which does use an integral component as it converts the turbine speed reference into a load reference and then uses an integral component to achieve the load reference (sounds more confusing than it is, but then, we are talking about GE, aren't we?). Constant Settable Droop has also been covered on control.com, and it's usually also briefly described in the Control Specification.

There are much better (older) threads on control.com about droop speed control as implemented in GE-design heavy duty gas turbines.

Unfortunately, there are so many of them.

I'm curious to know what "tuning" is thought to be necessary....

By the way, you've got quite a machine in that 40MW GE-design Frame 5 heavy duty gas turbine.

 

steam turbine droop setting -
i do not have much experience in steam turbine controls involving speedtronic. but i remember that the droop, ramp up rate and valve position is all done using
single auto extraction control - xsaxcoo
double auto extraction control - xdaxcoo

i do not have any steam turbine application code right now and i have worked with cogeneration turbines with double extraction. if you have a similar turbine your constants will be in present in that particular block. this is as much as i can help you about. tuning that monster of the block is not easy.

well , you seem to have found out the thread about droop setting calculation. so you are welcome (always hoping that you will say a mental note of thanks to me for posting it ) lol.

and you seem to me mistaken about the ramp rates, 15% ?? i hope you are referring to load ramp rates. it is configured in tnrv1 block in the l83jd# array. this will give you the auto load rate , the manual load rate etc.

as far as steam turbines are concerned , the load rate is reduced as you have a boiler master pressure to take care of. load it suddenly to a large extent and the boiler firing is not keeping up with it you will face a drop in the master pressure. for this reason , the load rates of steam turbines are lower than GT. as far as unloading rates are concerned they are the same. there are no different auto rates for loading and unloading. but in case of a load throw-off ,or a sudden unloading then for a steam turbine it is much more critical than a gas turbine. sudden load throw off in a steam turbine will result in overspeed condition if steam input is not fast valved. GE terminology for this is called early valve actuation it is a firm ware present in VTUR in mark vi , you will have similar one in Mark V. i do not have info on it but i will look it up.

and finally GE TA's get paid 450$ an hour !!!!!!!!!. no wonder refineries train and keep guys like me so that they do not have to call a GE TA. i wonder if GE would hire me , seems like a good proposition hmmm .... perhaps they will also pay me extra so that i do not come here and post in control.com all their secret tuning techniques at least regarding GT. (dry humour again , seems i cannot stay away from them)

 

"That formula in the thread you cited is a very old GE formula that they have basically stopped using decades ago. (Some of the packagers of GE turbines have not stopped using it so it still makes it way into some newer packages. It works, but it's not the method you may find in your GT control panel unless it's a very early Mark V or it was provided by one of the packagers of GE-design heavy duty gas turbines and they chose to use that particular formula.)"

what ?? that FSRN block was in the 4 machines i commissioned within the last two years. and all of them are in mark vi units !! and also in other units as well. seems that GE is shipping only old stuff to india al teast the turbine control library , or is it a problem with BHEL i do not know. CSA can you tell me the block name for the droop which has the Constant Settable Droop Speed Control. i searched my GE turbine control library to no avail. there is no other FSRN block.

as an after thought seems GE will not be vying up to hire me , i know only the old stuff sigh . but hey i got a new commissioning coming on and hopefully they will have the new control library with the constant settable droop speed control , i am a eternal optimist

 

ProcessValue,

There is nothing wrong with that formula, <b>as was said in the reply</b>. It's just a very old formula, and it is still used in some packages (even with Mark VIe) by some packagers, and I would imagine that even GE still uses it on some applications as well, particularly if the Customer asks for it or requires it. (I'm speaking primarily of new unit applications in this reply.)

Some applications require a different method for calculating droop; it seems that BHEL doesn't believe the applications at your refineries do.

GE, when they package and provide a turbine, seem to have standardized on using Constant Settable Droop for most all of the generator drive applications they provide. (Many, if not most, of the turbines GE packages these days have DLN combustion systems, and Constant Settable Droop helps to smooth combustion mode transfers in some cases. So, it seems that may be one reason for why they have "standardizec" on it.)

Some packagers, such as BHEL, may have decided that "If it ain't broke, don't fix it." In other words, if it's working, why change it? Their engineers and service people understand how the formula works, and so they stick with it.

There is an old saying to the effect of, if you haven't learned something new today then you must be dead. Two years' experience doesn't qualify for knowing everything there is to know about GE-design heavy duty gas turbines and control systems. So, start living--and learning. And stop thinking you've learned all there is to learn.

 

>"You may find better; but you'll never
>pay more."
>
>In its purest form, droop speed control
>is proportional control, pure and
>simple.

Let's move back a few steps. As much as I appreciate discussion, I'm not here to learn about the theory of speed control and subsequently power generation control. I know what droop is--yes it's just the P and has no reset in it's pure form. I know how droop is calculated, why it's important etc. I also fully understand it.

What tuning? Well, most likely, NO tuning has to be done on the gas turbine. But we want to get data so a model of the ENTIRE power system (grid) can be modeled. We are on an island that sees about 250MW peak load. We have four 40MW units, two 66MW, and two 25MW units that are base load. The 40MW frame 5 is a peaker/backup.

The base load units are all being "tuned" because essentially, NONE of them respond to system disturbances. We will have under frequency load shedding if there is a 10MW drop in generation. A 40MW trip causes multiple stages of UFLS. This is completely ridiculous and it's obvious (through trending, etc) that no units respond and push on the gas. I've been here less than a year and we are fixing it--utilizing an outside expert in small grids--by not only tuning the units to respond, but also get data so a computer model can be created for system stability simulations.

That's the background.

Now, my original question is still not fully answered. The problem is with the 66MW steam units. I want to know what settings control the droop and response rate. Right now, when frequency drops, the unit(s) do not open the control valves as fast (response) and as much (droop) as I think they should. If they did, we would not see the system freq drop like it does.

No unit is running ISOC on our system.

The frame 5 doesn't run much, so it can't help out on day to day events, but my guess is that it doesn't need any "tuning".

We completed one plant that has two 40MW slow speed diesels. The ramp rate was original as commissioned...2.3MW per minute.

So again, what constants, settings, or logic section dictate the droop and response rate on MK V. I know ya'll can help me as the experience here is truly amazing--just please don't get lost in the nit-pick details.

The control valves do not open fast enough on the steam unit. The rate they open is CONSTANT and always a constant slope when viewed on the historian. This rate is too slow.

 

>The quote is:
>By the way, you've got quite a machine
>in that 40MW GE-design Frame 5 heavy
>duty gas turbine.

You'd be right if I didn't mistype--it's a frame 6. Sorry about that.

We also have two frame 5 and two lm2500, but they rarely run.

 

>ProcessValue,
>
>There is nothing wrong with that
>formula, <b>as was said in the
>reply</b>. It's just a very old formula,
>and it is still used in some packages
>(even with Mark VIe) by some packagers,
>and I would imagine that even GE still
>uses it on some applications as well,
>particularly if the Customer asks for it
>or requires it. (I'm speaking primarily
>of new unit applications in this
>reply.)

Yes ,CSA i know you did say that there was nothing wrong with the formula provided , i was just curious to know how the droop is calculated with the new constant settable method , that is why i asked for the block no.

i have not worked in DLN machines , perhaps it is the reason why they do not have this particular block. DLN machines are not prevalent in india (only present in 4 sites at present), and i have heard that BHEL has only "one" person who is trained in DLN tuning.

and yes , i know learning is a continuous process, gas turbines form only one part of my responsibilities and i am learning a new thing every day

 

Kurush ,

we do not know the electrical system of your network. we do not what is the normal operating configuration of the system.

It is not known if your island , is normally connected to the gird or not? the load rate you mentioned 2.3 Mw/min is only possible when you are connected to the grid. in a typical island, ie you are not connected to the grid the system operates in droop and takes up load instantaneously abit with a frequency drop corresponding to the droop. there is no time delay in taking up the load in a independent system.

so, if you can post your electrical SLD and then a normal operating configuration then we will be able to help you more. better still give an example indecent. more detail you give us , better we will be to help you.

 

>I'm just looking for the GE equivalent of gain
>and time constants.

That sounded like you were looking for P & I values, to me anyway.

I might have given the wrong Control Specification section for Speed Control. I believe Sect. 03 is for Temperature Control, and Sect. 02 is for Speed Control.

Again, I'm not really too savvy on how GE controls droop with their steam turbines; sorry.

I also apologize if you get caught in the middle of the deteriorating banter between ProcessValue and I. No collateral damage intended.

 

Spent a little more time digging. I got the csp printout and found the XTNCB02 "Speed control block". Some of the constants it uses follow. I see nothing explicit for droop. My GUESS (which is non educated) would be that KTNE_G Speed error gain constant is key.<pre>
KTNED_HOLD -- Speed error differential hold constant % <0.0 %>
SEQ_TRB1 97

KTNE_DIFF -- Speed error differential constant % <0.4 %>
SEQ_TRB1 97

KTNE_G -- Speed error gain constant %/% <20.0 %/%>
SEQ_TRB1 97

KTNE_I -- Isochronous startup time constant sec <10.0 sec>
SEQ_TRB1 97

KTNE_IG -- Speed error gain for intercept valves %/% <20.0 %/%>
SEQ_TRB2 72

KTNE_IT -- Isochronous speed error deadband % <0.05 %>
SEQ_TRB1 108

KTNE_ITH -- Isochronous speed high error deadband % <0.50 %>
SEQ_TRB1 108

KTNE_LAG -- Speed error lag constant sec <1.0 sec>
SEQ_TRB1 97

KTNHDEV -- Speed deviation from setpoint limit % <3.0 %>
SEQ_TRB1 97</pre>

 

>That sounded like you were looking for
>P & I values, to me anyway.


Well, I at first I didn't know what I was looking for LOL. You can see in an above post I found the constants that drive the "Speed Control Block". I think the "Speed error gain constant" is the droop I'm looking for. 20 % per % would mean valve opens 20% for each 1% drop in speed...that would be 5% droop.

One problem with the moderated posts is the lag time before things show up. Also, I don't really understand the non-linear nature of how posts show up, but please look above for my other post with the constants.

Thanks again and side banter is ok.

Proccess...no need to get too detailed and talk about electrical system. We literally are AN ISLAND in the middle of the ocean.

 

To give an update and closure to this thread:

I found the "droop". It's given by GE in %/% and the value is 20.

There was something limiting the opening rate of the valves on the steam turbine. Digging through the logic I found there is a inlet pressure scheme that had a clamp/ramp block that was limiting how fast the turbine would pick up load.

The variables were undocumented and also were not recorded in the historian. I added the points to the historian and came back after a system disturbance and the data confirmed my suspicion.

The speed control block was outputing the correct reference value to open the valves during system freq deviations. However, the output of the speed loop went into another block that selected the min input--and the min input during low frequency events was always the inlet pressure limiter.

So, all the information was in front of me...but in normal GE fashion the documentation was pathetic and I had to spend some time digging.

 

Thanks for the feedback, though it would have been nice to have the signal name since it's likely different for the steam turbine than for the gas turbine(s). I don't have any steam turbine Mark V software and since it can be different depending on the application anyway, it would have likely been difficult to give any specific assistance (without being able to examine the software running in the panel at your site).

The answers can always be found in the sequencing and logic. And trusting longname descriptions has caused many to be severely disappointed and worse. So, it's always best to do one's own research and verify signal names and operation because what's running in the machine is THE gospel, not comments or longnames or what's written in some description in some manual. Everything except what's running in the Speedtronic is "intent", and should only be trusted as such. The only thing to trust is what's running in the panel, and while it make take some digging and head-scratching, I think you have to admit: It's all there! (Except the SAX and DAX blocks in steam turbine controls!) By following the signals through the sequencing and blocks, with a rudimentary knowledge of ladder logic symbols one can usually decipher what's in the sequencing/application code.

Anyway, thanks again. Best of luck in you island operation scheme (adventure)!

 

Kurush , thanks for the feedback . it would be great if you can provide the details, the tag no and the block in which the droop is configured. and yes write back to what happened with the island scheme modeling

 

The block that is causing the pain is a XPLAG00 Prop+Lag Control. The variable in it that is limiting the opening of the valves is the "Lag" variable (time constant). It is 10second. I did testing dropping it to 5, 2, .5 and 0. At 0 it finally let the valves go where the speed control loop wanted them to go.

This logic is a mess. There is no defined SPEED REFERENCE. There is TPWR, which is basically the speed ref. However there are tons of other pressure limited etc references.

The XPLAG00 block is in the inlet pressure limiter section. It makes no sense because when we do a step swing on the unit of about 10% output, the steam drum pressure hardly moves.

When the time constant was at 0, it performed nicely. And to note: closing the valves is NOT affected by this block because later in the sequence there is a select min block that select the lowest ref signal and that is eventually turned into valve demand. Going down, the speed control loop always puts out the smallest value--so the valves are allowed to close fast.

Now everybody says "that limit must be there for a reason, we can't just change it". Sigh...it was there to limit the drop in boiler steam pressure--but somebody messed up. The boiler is taking the swings fine and nothing is being harmed (I'm a boiler guy in former life).

I'll be writing a paper eventually on about this tuning and running a true island grid. I can drive the system freq with the few clicks of a button on a 40MW unit...not many people can do that LOL

 

A lot of smaller, so-called "industrial" steam turbines were not designed nor were they applied in frequency control applications. The idea is that when they are connected to a grid in parallel with other machines that the frequency is governed by the grid. So, instead of a turbine speed reference, like is used on GE-design heavy duty gas turbines, they use a <b>T</b>urbine <b>P</b>o<b>W</b>er <b>R</b>eference (TPWR) and use a modified droop control based on rated power output to allow them to stably contribute to the system load.

A lot of smaller steam turbines used in cogeneration plants are controlled the same way, because they are "followers" not frequency control machines. As you're no doubt aware (but others reading this thread might not be) many of those units are just put in "power" mode with the inlet control valves wide open and they just convert any available steam flow from the HRSG to as much power as they can make. Some of these even have extractions for feedwater heating and process applications.

It could be that when those turbines were purchased they were not purchased with the intent of being used in frequency control mode. Or, that the intended "AGC" (Automatic Governor Control) or load-sharing scheme was never implemented and/or it wasn't successfully implemented.

It could even be that the implementers of the AGC/load sharing scheme didn't program the scheme properly, not understanding how the steam turbines were programmed to be applied (I know; that's inconceivable, but stranger things have happened).

And finally, it could be that the whole plan for frequency control on this island has changed for whatever reason.

They should just be happy you've arrived on the scene to save the day, what with that <b>messy</b> logic being as bad as it is.... lol

We look forward to reading your paper and learning about a successful island frequency control.