When a single-shaft gas turbine is operating at Base Load and grid frequency decreases the unit can't produce any more power.

And <b>in fact,</b> the unit actually produces <b>LESS</b> power because air- and mass flow through the unit both decrease as axial compressor speed decreases.

So, even if the "load reference" increases the turbine power output decreases.

It's how gas turbines work.

 

My apology CSA, I think I have missed out much of the detail.

The problem happened not during baseload, it was in part load (around 85% load). With the frequency dropped (to approx 49.6), TNR dropped as well to pre-select load. So the load did not pick up as expected. Worst more, with TNR dropped to the pre-select load, TNH was low which in turn caused lower air flow and lower load. So, my machine actually pulled the grid down.

 

There are several problems with this thread.

The frequency--and therefore the speed--of a synchronous generator and its directly-connected prime mover(s) is <b>NOT</b> controlled by the Speedtronic. It's completely controlled by the frequency of the grid.

What <b>should</b> happen when a grid frequency excursion occurs <b>and</b> a GE-design heavy duty gas turbine-generator is operating at Part Load is that when TNH changes TNR should not change--which is what happens when Pre-Selected Load Control is active <b>OR</b> when external (in your case, Remote) load control is active.

You are confused and you have confused us with your comments about Base Load being reached at 104.8% and TNR continuing to increase to 106% and not being able to respond to frequency excursions.

 

Dear CSA,

I think I have put in unnecessary info and had confused everyone. If you noticed that in my early post, I mentioned the 104.8% to 106% is a side topic, just another issue popped out from my messy mind. So, from now on, we all should forget the issue of TNR=106% at baseload. My bad.

The original issue is:
1) Machine in part load
2) freq dropped
3) Load dropped at secondary frequency response due to TNR dropped.

From the chronology, I believed the problem is that TNR is not fixed, because Remote Load Control was active.

 

> The original issue is:
> 1) Machine in part load
> 2) freq dropped
> 3) Load dropped at secondary frequency response due to TNR dropped.

> From the chronology, I believed the problem is that TNR is not fixed, because Remote Load Control was active.

Sean,

You have talked about Pre-Selected Load Control, Remote control, Remote Load Control, DCS, Foxboro, AGC and secondary frequency response as if they were all the same thing and were interchangeable.

Pre-Selected Load Control <b>usually</b> can only be active when the unit is in "local" (AUTO) mode.

Remote control can mean many things. DCS control, either via discrete or analogue signals, hardwired or over a communication link (MODBUS, GSM, etc.). Remote control might also mean AGC control, which <b>usually</b> does NOT come to the Speedtronic through a DCS, but stranger things have happened.

AGC might be used for secondary frequency control (response), and it might not be. Secondary frequency control/response is not <b>usually</b> done "in" or "by" the turbine control since it's usually done remotely.

The difference (or, error) between TNR and TNH is used to determine how much fuel will flow to the gas turbine. I find it somewhat unusual that a 9FA STAG unit uses "classical" Droop Speed Control and not Constant-Settable Droop Speed Control, but, again, stranger things have happened. The "truth" is the formula for classic Droop Speed Control is:<pres>
FSRN = [(TNR - TNH)*FSKRN2]+FSKRN1</pre>
FSKRN1 & FSKRN2 are constants, that is, they are not variables.

<b>BOTH</b> TNR & TNH are variables. Except that under most conditions TNH is stable, because grid frequency is usually stable. So, by varying TNR when grid frequency is stable the fuel flow and the load is varied.

When grid frequency (and TNH) is stable Pre-Selected Load Control can be used to keep load near to an adjustable setpoint--even though it should <b>NOT</b> be necessary to do so. "Why," you ask, "should it not be necessary to use Pre-Selected Load Control to maintain a stable load?"

Because at Part Load when <b>NO</b> Load Control is active TNR doesn't change unless the operator clicks on RAISE- or LOWER SPD/LOAD. When TNR is not changing and TNH is stable the difference between TNR and TNH is not changing (it's stable), the fuel flow will not change, the load will not change, and so there's no need to use Pre-Selected Load Control.

And I maintain that Pre-Selected Load Control is usually poorly tuned--if at all--that it usually results in continual hunting (oscillation) around the Pre-Selected Load Control setpoint which results in increased wear of fuel control valves, valve actuators, and fuel control servo-valves.

As an example, when grid frequency (and TNH) is stable on a non-STAG unit with 4% Droop, if the operator loaded the unit to 50% of rated TNR would be approximately equal to 102%. And as long as TNR stays at 102% and TNH is stable the error between them will not change (it will remain stable)--which means fuel flow and load will remain stable--so there is no need to use Pre-Selected Load Control to adjust TNR to maintain a stable load! If a poorly tuned Pre-Selected Load Control is used and it results in continual changes of TNR because of over-shooting the deadband of Pre-Selected Load Control it's ultimately damaging to the unit's auxiliaries.

Pre-Selected Load Control is, essentially, trying to maintain a constant difference, or error, between TNR and TNH. A constant error means a constant fuel flow which means a constant load.

Which is why when grid frequency (TNH) changes load will change <b>in the exact opposite direction that it should change</b> when grid frequency changes.

From your "description" above it's not clear if DCS or AGC or secondary frequency response/control was active. If the "remote" signal was simply trying to maintain a constant load regardless of grid frequency then what happened was to be expected. Period.

If the "remote"signal was trying to maintain grid frequency (secondary frequency response) then it did <b>exactly the opposite</b> of what it should have done.
Right?

So, we still don't have the details required to properly analyse the event.

 

> Pre-Selected Load Control <b>usually</b> can only be active when the unit is in "local" (AUTO) mode.

CSA,
At least up until I retired, when GE supplied the overall combined cycle plant control, we did make use of Preselected Load Control to implement (remote) dispatch load control. We generally received the dispatch load setpoint into the DCS and then used that setpoint to adjust the preselected load setpoint in the gas turbine control. This scheme was more common with Mark VI and Mark VIe turbine controls than with Mark V.

 

First of all to CSA, thank you for the MVP Kudos, I'll try not to let it go to my head.

To Sean, I am going to restate some of the comments from CSA, for your benefit and mine, and to have CSA sanity check my thoughts and understandings.

You say in your post:
"The original issue is:
1) Machine in part load
2) freq dropped
3) Load dropped at secondary frequency response due to TNR dropped.
From the chronology, I believed the problem is that TNR is not fixed, because Remote Load Control was active."

So you operate this unit in some sort of Remote load control mode, some sort of power control center sends an AGC mwatt setpoint to your DCS. The DCS then sends this Mwatt setpoint to the MKV via a 4-20mA signal or through communication. If you can clarify this it would take out some of the assumption.

Prior to the drop in frequency the mwatt setpoint to the unit is lets say 220 mwatts, frequency is normal and everyones happy. Now frequency starts to decline due to some event on the system. At this point the mwatt setpoint has not changed and so TNR (Turbine Speed Reference) is also unchanged, but TNH (Turbine Speed High Speed Shaft) speed is decreasing, being dragged down due to decreasing system frequency. This change creates a larger error between TNR and TNH which causes the turbine controller to increase fuel to raise output. It will do this until either the unit reaches temperature control, or frequency and machine speed (TNH) rises back to nominal. This is basic "primary frequency response" or droop, or whatever term I am supposed to use.

Now you say that the "secondary frequency response" is not proper, since remote load control was active. This part of your post needs further description for me to clearly understand it. As CSA says "AGC is secondary frequency response by most definitions".

Lets go back to your situation, mwatt setpoint is 220mwatts and frequency is normal. Frequency begins to decay, TNH falling, control system increases fuel to reduce error between TNH-TNR. Primary frequency response, or droop is attempting to bring the system back to normal.

In a properly operating AGC system a new mwatt setpoint will be sent through the DCS to the turbine asking for more mwatts or power since the AGC system "sees" the low system frequency. So now TNR will start to increase since the mwatt setpoint is rising, and TNH is already low, so the unit should increase output until it reaches base load depending on the duration of the event.

In your case it sounds like the setpoint from AGC was reducing at the same time that frequency was falling. This is exactly what you do not want to happen in this situation. I have seen this exact issue before and it is usually the fault of some issue with the AGC control at the power control center.

Sean if you can clarify for all of us if any of these scenarios is what is occurring at your plant I know I would feel a lot better offering any more advice.

 

MIKEVI,

Thanks for asking Sean to clarify his terms and usages and actual configuration. I think the main problem here is that Sean, by his own admission, is not a "controls guy" and so he is operating on his understanding of what he is being told and is having some trouble getting his understanding up to where it needs to be to comprehend the whole problem. He's probably getting told multiple things from multiple sources--including site personnel and grid regulator/operators. And sometimes neither of these two groups really understands what should happen when, and who's on first, and what's on second, and, well, you know the routine.

I do feel the need, however, to clarify a point. It's something that took me a long time to get clear, also, and I think it's worth mentioning here because it could confuse others who might be reading other passages and question this one.

>Frequency begins to decay, TNH
>falling, control system increases fuel
>to reduce error between TNH-TNR.

It's precisely because the error between TNR and TNH increases that the fuel increases. Increasing the fuel increases the power being produced by the unit, and that's what any kind of frequency response to a grid frequency decrease should do. Not until TNH increases as grid frequency increases will the error between TNR and TNH decrease.

And, technically it's TNR-TNH, not TNH-TNR.

During a grid frequency disturbance when Load Control (Pre-Selected or "Remote") is NOT active, the power output of the unit will remain high or low until TNH returns to rated.

It's when some kind of Load Control (Pre-Selected or "Remote") counters the increase or decrease of Droop Speed Control (Primary Frequency Response) that the problems begin.

Secondary frequency response means many things to many different agencies and entities around the world. Just today I did another World Wide Web search with my preferred search engine and found a new UK National Grid definition of secondary frequency response--one that is seemingly unique to wind turbine installations! I had formed my understanding of secondary frequency response a few years back by reading the European grid regulatory agency's documents, which, at the time, seemed to have the best description for the situation I was dealing with (which was a nation that was emulating that document's guidelines). I still believe that was the best and most comprehensive definition but, again, I have (painfully) come to the knowledge that before commenting on precisely what secondary frequency response is in any particular situation I need to know what <b>their</b> definition is--not mine--and adjust my definition to theirs for the duration of the discussion with them at their site about their problem.

So, thanks for asking Sean to help. As I've said earlier in this thread--I believe the "solution" to or any recommendations to assist with the issue(s) Sean has raised in this thread are going to require much more time and in-depth, on-site knowledge of the configuration and operation of the units at Sean's site before any comprehensive analysis and recommendations can be made. And if my past experiences are any indication, the likelihood that there will be an outcome that is satisfactory to most (certainly, not everyone!) is somewhere between a slim chance and no chance at all. It's interesting to speculate, but I feel we aren't really helping "not a controls guy" Sean by speculating much more. I think he, or his company, needs to "bite the bullet" and spend the coin to get this resolved once and for all. While I'd like to know the outcome, usually the consultants in these discussions get blamed for all of the problems (while getting handsomely paid for the privilege of being blamed for telling the Emperor he is not wearing any clothes)

 

The response is so overwhelming, I wish my GE support can be as efficient as you guys. Now I need to answer carefully so I don't mess up all queries.

The first response (I called it primary response) is exactly like what MIKEVI described, as though he was there at my plant. One correction that DCS remote control to MKV is not 4-20mA but a hardwired.

Quote MIKEVI:
<i>"So you operate this unit in some sort of Remote load control mode, some sort of power control center sends an AGC mwatt setpoint to your DCS. The DCS then sends this Mwatt setpoint to the MKV via a 4-20mA signal or through communication. If you can clarify this it would take out some of the assumption.

Prior to the drop in frequency the mwatt setpoint to the unit is lets say 220 mwatts, frequency is normal and everyones happy. Now frequency starts to decline due to some event on the system. At this point the mwatt setpoint has not changed and so TNR (Turbine Speed Reference) is also unchanged, but TNH (Turbine Speed High Speed Shaft) speed is decreasing, being dragged down due to decreasing system frequency. This change creates a larger error between TNR and TNH which causes the turbine controller to increase fuel to raise output. It will do this until either the unit reaches temperature control, or frequency and machine speed (TNH) rises back to nominal. This is basic "primary frequency response" or droop, or whatever term I am supposed to use."</i>

Then come 'Secondary Frequency Response'. To answer CSA question, my definition on secondary response is completely based on my country's Grid Code, where the load response after 30s to 30 minutes of frequency excursion is considered secondary response. Primary response is 10s to 30s. During the secondary response, at frequency of roughly 49.5Hz, the load dropped back to partload of 220MW (carry on from MIKEVI's chronology). That was not a good response because the load should not go back to setpoint of 220W while frequency still low, TNR-TNH was still a positive value and we expect fuel increase. Then we found out that load dropped because TNR changed.

If I may summarize in a layman's chronology:

1) Load stable
- TNR set at 220MW
- Actual at 220MW partload.
- On Remote Control, Pre-select load disabled

2) Primary response (First 30s)
- Frequency dropped to 49.5 hz
- TNH dropped with respect to 49.5hz
- Load increased to baseload

3) Secondary response (after 30s)
- TNH maintained with respect to 49.5hz
- TNR dropped
- Load decreased to around 220MW

I believe CSA have pointed out the problem:
Quote CSA: <i>
"From your "description" above it's not clear if DCS or AGC or secondary frequency response/control was active. If the "remote" signal was simply trying to maintain a constant load regardless of grid frequency then what happened was to be expected. Period." </i>

From what we have observed, the DCS/AGC is designed to maintain a constant load, or a "pre-select load" in disguise. In reality, the TNR dropped has proven the DCS doesn't care about frequency.

We are checking a way to incorporate frequency band in DCS, so that TNR would freeze when TNH drop to certain value, i.e. frequency excursion. Looks OK to me so far.

I hope I don't confuse you all even more...

 

Some correction-
Quote:<i>"One correction that DCS remote control to MKV is not 4-20mA but a hardwired."</i>

I actually tried to say DCS to MKV via hardwired 4-20mA but not Modbus. Don't know how I ended up that....

 

Sean,
You state:
"One correction that DCS remote control to MKV is not 4-20mA but a hardwired."

A 4-20mA signal is a hardwired analog signal. Another hardwired possibility is hardwired contact Raise and Lower command signals.

Also possible is a data link, which may use "hardwired" copper connections or fiber optic connections, but would (with Mark V) be either RS232 (or RS485) MODBUS protocol, or Ethernet (I don't remember what message protocol was used on Mark V Ethernet).

Can you be more specific as to the interface?

 

Dear Otised,

It is a hardwired contact to raise/lower command signals to MKV TNR.