Preselect Load with Remote Control

S

Thread Starter

Sean

My machine is a GE frame 9 single shaft, with Mark V for GT/ST and Foxboro for DCS.

I've read in many posts that preselect load does not behave good in grid frequency response, which the TNR decreases when grid frequency drop. That is exactly what happened in my plant.

During normal operations, we set the Mark V control to Remote mode, and the load control is controlled by foxboro. Operators put the load setpoint in Foxboro to control machine load. In this case, the preselect mode is disabled or still alive?

My TNR dropped to the foxboro load setpoint during grid frequency disturbance, so I can pretty much assume pre-select is somehow active. The question is, how to disable it in my case so it can support the grid during frequency response?

Need some good advice to enlighten me

Thanks
 
The easiest way to disable the preselected load control is for the operator to give a momentary load RAISE or LOWER command. It will put the load control in manual mode, where it will stay until a subsequent command to go to preselected load control or BASE load control.

If you need to do this automatically, you would need a means of detecting the frequency disturbance and then automatically create the single pulse RAISE or LOWER command. I would recommend checking with the control supplier (GE, I presume) before doing this, however.
 
You need to program your DCS to stop sending raises or lowers when the frequency goes outside some "window", or tell your operators to take the unit out of Remote mode when the frequency goes outside a "window."

The signal(s) from the DCS are raising or lowering TNR, not the frequency excursions.
 
Thanks for prompt suggestions.

The concerns on my side are:
1) My DCS doesn't have the Raise/Lower buttons, sounds funny but it's true.

2) To get operator to intervene when freq runs out of the 'window' seems ok, but it would be good to make it fool proof. Human intervention is not always reliable.

3) To re-program the DCS/Mark V would be the last resort if others don't work, for the obvious reason. It would cost me a 'Fat Boy'.

Seems like the solution is obvious, to 'freeze' the TNR! I will try to play with the control deadband see whether it works.

By the way, an out of topic question. How can I confirm that my droop setting is 4% (in mark V 9FA)? I notice a lot of time TNR went up to 106 during baseload, does it means that my droop is 6%?

Many thanks
Have a great weekend!
 
Sean,

I am going to chime in a bit, but most likely not answer all your questions, but at least try to help.
I notice a lot of time TNR went up to 106 during baseload, does it means that my droop is 6%?

No the max amount of the TNR is usually around 107%, check your control constant of TNKR3, that should be setpoint for the high speed stop, in normal mode, there is a higher setpoint for overspeed testing.

Since I don't have a copy of your application code running in your MKV its difficult to say what type of droop logic your site is running. From a newer FA file(MKVI not MKV) I have, that is using the BigBlock SpdLdCsDrp, which is Speed Load Constant Settable Droop, there is a control constant DWKDG which is set to .0242. This constant correlates to a droop factor of 4%. This number is arrived at by dividing the rating of your unit, in the case of the machine I was looking at its rated at 165.29, divide this number by 4(desired droop factor) and get .0242. If you look for this constant in your code it "should" point you to the area to find your droop, which is usually 4%.

This means then that once your TNR reaches 104% then you should be at base load, assuming conditions are near ISO. Anything above 104% will typically just wind up the megawatt controller.

I plan to contemplate the other part of your question regarding frequency response of your unit. This is a complicated subject, for me at least, for which CSA has had to correct me on several times. I want to formulate some sort of input to you and him as well, but need to get all my ducks in a row first
 
MIKEVI,

Congratulations on MVP status--well-deserved!

I believe the original poster has a single-shaft STAG unit and the droop setpoint on those units are usually 5%, sometimes 6% (hopefully otised will chime in and confirm this). So, I wouldn't be too surprised to learn such a unit (if my SWAG is correct) has 6% droop setpoint. Also, if the conditions of operation include a warm/hot and dry ambient AND the unit uses evaporative or inlet cooling it could be that the unit produces more than rated power for the ambient conditions. So, there could be many reasons (including lower than spec gas fuel supply pressure and even heated fuel) which could cause extra fuel to be burned resulting in an "unusual" appearing droop characteristic. (By the way, some machines are sometimes required by the grid regulator to have "unusual" droop characteristics.)

The HSS is there to prevent an operator from exceeding 107% TNH in an unloaded condition. <i>Under normal operating conditions,</i> TNR should NEVER go above 107% when the generator breaker is closed. I guess it could happen if frequency was MUCH higher than normal, but that would be an unusual circumstance, eh?

Typically, when a DCS is controlling a load setpoint it does so in one of two ways. The first is by issuing "discrete" RAISE and LOWER commands. That can be done using relay outputs, or over MODBUS/GSM.

The second method would be an "analog" setpoint either via a 4-20 mA hard wired connection or over a MODBUS/GSM link.

In either case, the Speedtronic is going to use the signal(s) to modify TNR to try to make load equal to the DCS setpoint. So, when these frequency excursions occur AND the unit is in Remote mode something has to be done to either 1) take the unit out of Remote mode and allow the Speedtronic to be in Droop Speed Control (<b>presuming the unit is NOT at Base Load!</b>), OR, 2) the DCS is going to have be modified to stop trying to control load (when the frequency is outside of some "window").

There is ALWAYS the possibility that our Belfort Boys have implemented some unusual method of Remote load control (they would NEVER do that--<b>NOT!!</b> there's a double negative there...). So NEVER count that possibility out when dealing with a unit "engineered" by Belfort, because they will <b>ALWAYS</b> do something unusual--just because they can.

Hope this helps!!!
 
Many Thanks for the valuable feedback.

I noticed my machine achieves baseload around TNR=104.8% (so droop is 4.8%?), and it goes all the way up to 106 at baseload. I am not a control guy, so I presume the region from 104.8-106 is some kind of control tracking to allow bump-less transfer from Temp control to speed droop control. Anyway, to honor my power purchase agreement, I need to fix that droop to 4%. My machine doesn't have DWKDG, only FSKRN1 and FSKRN2 (FSR droop correction factor), mind to shed some light how to do it?

I will check how DCS is controlling the Mark V load. Anyway, I think the second suggestion from CSA is quite viable. I will try to 'create' a window in DCS. If there is any frequency excursion beyond the window, the TNR will be fixed as it is. That is the only practical way I get now.

Many thanks again
 
Sean,

<b>From the information provided</b> it appears that someone has mistakenly increased the offset between temp control FSR and speed control FSR (many people don't understand that parameter and think TNR has to go to HSS when the unit is on temp control).

I don't have a CSP to look at to tell you the name of the offset constant but I do suggest you pay for a knowledgeable person to come to site to review the configuration and settings and fully understand what you need to have done so that it can be completed once and for all.

Best of luck with your issue!
 
I don't know whether 6% droop in single shaft combined cycle is "standard" or not. I was not closely involved in that type of setting once I got into combined cycle control - it is done by the gas turbine controls engineering organization. I know the 4% setting was most common on simple cycle applications when I worked on gas turbine controls.

I am not sure how to interpret the statement that the machine achieves base load at TNR = 104.8% and it goes all the way up to 106 at base load." By definition, base load is achieved when the unit goes on fuel temperature control, and the transition from speed control to temperature control should be bumpless since a minimum value selector is used to control FSR.
 
otised,

Thanks for your input. As is the case when managerial types are posting for help with technical questions (trying to a kind paying for technical help) we don't get complete information.

It seems, after after, all the unit is not a single-shaft STAG machine and only uses "straight" or "classical" droop speed control and has a mangled offset constant--<b>from the information provided.</b>
 
Yes, 4% is the most common droop, so far never saw any 6% droop. I will pay attention on the FSR offset.

Will update if I manage to solve the issue.
 
I want to provide some clarification.

First, when a gas turbine is operating on exhaust temp control <b>it will NOT respond to grid frequency disturbances in the same way as if it were at Part Load in Droop Speed Control.</b>

Second, when Base Load is selected and active using the BASE LOAD button on the operator interface FSRN is increased and maintained at approximately 0.3% above FSRT to prevent interaction ("toggling") between Droop Speed and Temp Control. That's the offset I am referring to.

Many sites use Pre-selected Load Control to get to Base Load, by inputting a setpoint well above rated load. The unit will not go above "Base Load" and will indicate "Temp Control"-- BUT it will also cause FSRN and TNR to be well above the offset. This is a very bad, but unfortunately all-too-common, practice.

Sean, how are the operators at your plant "selecting" "Base Load"? The offset Control Constant may be fine; the problem may be the way the operators are operating. That's why I'm suggesting you get someone to come to site who can observe all the things we cannot and get all the information it would be very time-consuming to have you provide to make recommendations to help with understanding and resolving issues at your site.

Lastly, just because your goes on "Base Load" at 104.8% does <b>NOT</b> mean the unit has 4.8% droop. There may be a myriad of reasons, and it would be nearly impossible to detail them all and get all the details from you in this forum. Again, that's another reason to get someone knowledgeable to site to help with understanding and resolving the issues at your site.

And as a reminder to anyone reading this reply: <b>A gas turbine on exhaust temperature control is not going to be capable of properly responding to frequency disturbances--without special sequencing.</b> No prime mover will when it's being commanded to produce as much power as it can and ignore speed (frequency). Which is what happens to a GE-design heavy duty gas turbine when it's operating on exhaust temperature control (Base Load).

And grid regulators don't always understand this, either, and so therefore must be educated many times. As well as plant managers, plant supervisors, operators, technicians and almost always operators.

It will be most interesting to hear how the operators select Base Load. Even more if they are told they need to start doing it differently than they've been doing it....
 
Dear CSA,

Most of the time, our machine is on AGC control, where the grid dispatcher would ‘take over’ the load control. Their control method is to set the desired load (doesn’t matter it is baseload or part load to them), and if it crosses temp control, then we reaches baseload. You said it is very bad, but unfortunately it’s out of our control.

I will definitely relay the issue to either GE or Foxboro. I do not want to make any control change without OEM. Play safe makes perseverance...

Btw, talking about someone knowledgeable to come to my site to check, I wonder where are u staying? :)
 
Sean,

As suspected; there is so much more to the way your plant is operated than first thought. And we STILL don't know if the unit is a 9E or 9F, single-shaft STAG, or.?.?.?

I'm not the person for this job. Try contacting one of the advertisers here, MD&A Turbines, or GTC.

If the load setpoint is coming via AGC and they can't directly select Base Load and they are "winding-up" TNR, when they might need to lower load quickly for some reason they won't be able to. It's going to take some time working with your utility for everyone to understand what is happening, what's possible, how it might be possible to "make things work," and in the end everyone might just get pissed off and the whole thing get ugly. Nope; I'm not the person for this job, where you want a quick fix to a very complicated situation that's likely going to take weeks or even months to sort--with the potential for an unsatisfactory outcome for all.

Best of luck, Sean! You're going to need it.
 
I think you really need to get GE involved in this. It sounds like you may need secondary frequency response, which has been done at some combined cycle sites in Spain. However, this can involve essentially derating the unit to leave room for loading above (the derated) base load limit. Back when I worked at GE, one scenario for this was to operate derated only when the grid operator required secondary frequency response capability.

In order to implement something like that, changes would likely be required to both the Mark V control and to the Foxboro DCS. It is also likely that the grid operator would need to be involved. Good luck.
 
My machine is 9FA single shaft STAG.

Otised, you have hit the bull's eye (with my very minimum information provided). Primary response was acceptable, droop control is quickest to response, so initially the frequency response was purely function of (TNR-TNH). Then problem shows when TNR starts to self-correct itself to the pre-select load (by remote control), then the secondary response was taken over by load control. I wonder what had Spanish done on secondary response control...

Still, the root cause is TNR, which he should stay still. I am looking the possibility to fix the TNR before going to Mark V or Foxboro DCS.

The truth is all I want, but the luck is all I need.
 
M

MarkThe Second

Sean

Maybe the grid operator knows that your plant is more efficient than others (or for some other reason) wants your plant to operate at max output to lower requirements from other plants. They may also not care that your plant is not responding to the grid frequency and has other less efficient plants on the system on speed/droop control doing this. If this is the case then maybe all is OK?
 
That is totally not ok. For a poor secondary frequency response, not only you are not helping the grid stability, but you are pulling down the grid when frequency drops. I would not tolerate that if I am grid operator.
 
> Still, the root cause is TNR, which he should stay still. I am looking the
> possibility to fix the TNR before going to Mark V or Foxboro DCS.

TNR is the Speedtronic Mark V Turbine Speed Reference and "originates" <b>IN</b> the Mark V.

TNR is increased and decreased to increase and decrease load--when grid frequency (TNH) is stable. When anyone or any entity is trying to control load they are doing so by increasing and decreasing TNR.

Since TNR originates in the Mark V I don't understand how it can be "fixed" before going to the Mark V. When the DCS and/or AGC is raising or lowering load it can only do so by increasing or decreasing TNR <i>in the Mark V.</i>

AGC <b>is</b> secondary frequency response by most definitions. Primary frequency response is Droop Speed Control by most definitions--except for GE's ill-named PFR option.

If AGC is unaware the unit has reached exhaust temp control and keeps trying to increase load--by increasing TNR--then TNR isn't the problem.

Try this: Keep the unit in AUTO (local) mode during a start and when the load is close to exhaust temp control have the operator select BASE LOAD from the HMI and trend FSRN & FSRT & TNR for an hour after the unit reaches Base Load and remains in AUTO (local) mode with BASE LOAD active. You will likely see that FSRN tracks FSRT with a constant, or fairly constant, differential (likely about 0.2% or 0.3%). And you will also see that as FSRN increases and decreases TNR increases and decreases because FSRN is driven by (TNR -TNH). (It is presumed grid frequency is stable during this and the next test.)

Then switch control from AUTO (local) mode to REMOTE (AGC) mode, while continuing to track FSRN& FSRT & TNR. If AGC tries to increase load above Base Load then TNR and FSRN will increase--for as long as AGC keeps trying to increase load. (Again, it is presumed grid frequency is stable during this period.) This is regardless of whether the AGC signal is discrete or analogue.

Again, the truth is that TNR originates in the Speedtronic
and is driven by the operator or the DCS or AGC. It doesn't come from anywhere else and can't--and more importantly shouldn't-- be "fixed."

The "problem" still isn't defined. And the solution isn't to "fix" TNR.

The problem is most likely that TNR has been driven by "remote" control to be so much higher than "Base Load TNR" (in other terms, it has been "wound up" higher than the TNR corresponding to Base Load) and then "remote" control tries to quickly lower load. Because TNR has to be decreased below "Base Load TNR" before load will actually start to decrease it appears that secondary frequency response isn't working. It's because TNR has been wound up so much and has to be un-wound before load starts to decrease.

Do the data-gathering above. You can use VIEW2 at a very low rate, or VIEW1. The complication might be the STAG nature of the plant or if the unit has power augmentation steam injection.

Please write back with the results of the data-gathering.
 
Top