from the Performance Department department...
Gas turbine power generation frame 6 load output
 Posted by Mr Oluwafemi Ogundaisi on 26 April, 2012 - 11:32 am
In between, ambient temperature condition and stable grid frequency, which of these contributes heavily to a gas turbine frame 6 load [MW} output most? How does variance in both condition affect the total load [MW} output?

 Posted by CSA on 29 April, 2012 - 5:06 am
I'm not quite sure I understand the question, or, rather the intent of the question. But, here goes.

The speed of a GE-design heavy duty gas turbine driving a synchronous generator that is being operated on a large or infinite grid in parallel with other prime movers and generators is fixed by the frequency of the grid. The speed of the generator, and of the turbine driving the generator, and of the axial compressor of the turbine is directly proportional to the frequency of the grid. And changes in grid frequency will cause the turbine--and axial compressor--speed to change.

Gas turbines are basically "mass flow" machines--the more mass (air and fuel) that flows through the machine the more power it can produce when operating at Base Load (i.e., CPD- or CPR-biased exhaust temperature control). So. anything that affects the mass flow through the machine will cause the power output to vary at Base Load when the machine is producing as much power as possible by monitoring CPD (Compressor Discharge Pressure) and TTXM (actual, average Turbine Exhaust Temperature) while the IGVs are fully open. Fuel flow into the machine while operating at Base Load is relatively stable.

Ambient temperature changes cause the air density to change. When the unit is operating at Base Load at a stable speed (frequency) with the IGVs fully open, the axial compressor is running at a stable speed, if the air density changes then the mass flow of air through the machine will change. Higher ambient air temperature will cause the air density to decrease, and since the machine's IGVs are fully open already and the compressor is spinning at a constant speed (because the frequency is stable) the mass flow of air through the machine decreases, and the power produced by the machine decreases. The opposite occurs when the ambient gets colder. A plot of ambient temperature and load when the unit is operating at Base Load over 24 hours will show that changes in ambient will affect power output.

When the air temperature increases, the CPD will decrease. Conversely, when the ambient air temperature decreases the CPD will increase. This can be observed by plotting CPD and ambient air temperature over the course of a day while the machine is running at Base Load and a stable frequency (speed).

When the unit is operating at Base Load and suddenly the frequency of the grid changes the speed of the axial compressor will change. If the frequency decreases, the speed of the axial compressor will decrease and the decrease in speed will cause the mass flow of air through the machine to decrease which will decrease the power being produced.

However, when the unit is being operated at Part Load (i.e., not at Base Load) it is being operated on Droop Speed Control, not exhaust temperature control, and the IGVs are being modulated as load changes (newer machines have modulated IGVs which vary position depending on how they are being controlled). On Droop Speed Control fuel is being controlled in proportion to the error between a turbine speed reference, TNR, and the actual turbine speed, TNH. (The turbine speed reference, TNR, is a value typically between 100% and approximately 104% (for machines with 4% Droop), plus or minus a half-percent or so depending on many factors such as ambient temperature and axial compressor cleanliness and inlet filter cleanliness and exhaust back-pressure, etc.)

If the unit is operating at Part Load at a stable load (meaning the turbine speed reference is stable) without Pre-Select Load Control enabled and active (more on that below) and the grid frequency changes, then the actual speed of the turbine and axial compressor will change in direct proportion to the change in frequency. But, Droop Speed Control senses the change in actual turbine speed with respect to the stable turbine speed reference and increases the fuel to increase the power being produced by the turbine. If the frequency increases while the turbine speed reference is stable and the unit is operating at Part Load without Pre-Select Load Control enabled and active then Droop Speed Control will see a smaller error between the turbine speed reference and the actual turbine speed and decrease the fuel to decrease the power being produced by the turbine (and generator, since the generator output is directly proportional to the power being transferred from the turbine to the generator).

As for changes in ambient temperature affecting unit power output at Part Load, while not negligible they are minimal and only really affect the heat rate of the machine (which isn't that great at part load to begin with).

The exception to the machine's response to grid frequency changes at Part Load occurs when Pre-Select Load Control is enabled and active. In this case, the turbine operating setpoint is a load setpoint, and even if the frequency changes and Droop Speed Control tries to change the load of the machine in response to the speed error caused by the frequency change. Pre-Select Load Control will try to maintain the load setpoint and over-ride Droop Speed Control.

So, the answer to your question depends on how the unit is being operated: at Base Load or at Part Load. Basically, speed and ambient temperature only have an affect on power output when the unit is being operated at Base Load on exhaust temperature control. When being operated at Part Load (loads below Base Load), ambient temperature changes really only affect machine heat rate, and frequency changes are responded to with load changes because of Droop Speed Control (unless Pre-Select Load Control is enabled and active).

[To complicate matters, some sites set a Pre-Select Load Control setpoint that is above the rating of the machine and let exhaust temperature control limit the power output of the unit. In this case, even though Pre-Select Load Control is enabled and active, it is being over-ridden by exhaust temperature control (Base Load). This is not a good practice, but it does happen. If, while attempting to reach a Pre-Select Load Control setpoint the machine reaches exhaust temperature control (Base Load), exhaust temperature control becomes the over-riding control even though Pre-Select Load Control is enabled and active. Exhaust temperature control is always the "limit" of the machines power output for the operating conditions at that given time and ambient- and machine conditions.]

I hope this helps. Again, I don't know if I really understood the question, or the intent of the question.

 Posted by Oluwafemi Ogundaisi on 12 June, 2012 - 5:05 am
This response perfectly helps. I now have the clearer picture of it.

Thanks very much.

 Posted by Vinayak on 11 August, 2012 - 1:07 am
Dear CSA, Thanks for nice explanation.

I have one more question to further understand effect of frequency.

Few days back, there was severe grid disturbance. We have 2 Frame VI GTs running in parallel (floating) with grid. Both GTs were in preselect mode at 29 MW each and frequency was around 50 Hz.
Then all of sudden frequency increased by nearly 1 Hz rapidly, this caused both GT to drop load to 12 MW.

I want to understand why GT load drops by more than 50%?

0 out of 1 members thought this post was helpful...
 Posted by Namatimangan08 on 12 August, 2012 - 2:37 am
1Hz = -17MW

1.7Hz = -29MW

For a 50Hz system, it is 3.4% frequency change from no load to base load.

So you fuel control is calibrated linearly to provide the desired response of:

No load condition (synchronized but zero output) : RPM =3000RPM

Full load condition (29MW) RPM : 3102RPM

Your droop % set point = 102*100/300 = 3.4%

If your droop set point was set at 3.4% then 17MW load rejection for every 1Hz frequency increases is the expected behaviour.

If you want to make the machine to response smaller for every 0.1Hz frequency change, then you can change the droop set point says to 6.8%. Then for the same 1 Hz transient frequency raise your GT will reject 8.5MW instead of 17MW. Note that you will alter loading characteristic too by changing droop set point.

 Posted by CSA on 12 August, 2012 - 8:51 pm
Vinayak,
What is the resolution of the data capture/historian method you are looking at? Because if the resolution is not on the order of milliseconds, it's very difficult to say for certain exactly what all of the parameters were at any given instant in time.

You also have to remember that there is a good deal of inertia in a GE-design Frame 6 heavy duty gas turbine, so when the unit is being "jerked around" by the kind of sever grid disturbances experienced a few days back, the sampling rate of the data capture mechanism may not accurately indicate the actual extremes of the frequency disturbance. Inertia works both ways--preventing a very fast unloading and also adversely affecting a very fast loading.

Also, we don't know what fuel you were running on, but there are lags in the ability of the fuel to increase or decrease very quickly when the grid frequency is unstable.

So, unless you have data from a very high sample rate it's probably going to be very difficult to draw any conclusions about how the machine really responded. And if the frequency disturbances were erratic and of short duration, they are also not going to lend themselves to detailed analysis.

I would not suggest changing any droop settings based on the data or the recent grid problems.

Hope this helps!

0 out of 1 members thought this post was helpful...
 Posted by Namatimangan08 on 13 August, 2012 - 5:02 am
I don't suggest the poster to change his droop set point too. What I am trying to propose the plant owner should know how the speed droop can help him to make the plant work better.

If 3.4% droop set point is correct then why should anybody think about changing it? But then how do we know that 3.4% is correct in the first place?

I'm talking about the best way to manage the speed droop. In this case the plant owner should know what is the correct set point by knowing the fundamental rather then accepting what has been set right by the contractor. Setting the speed droop is not so much about the machine requirement. It is required to provide fast response to mainly transient stability. As far as the machine is concern it likes to have very big percentage set point says by 18%.

0 out of 1 members thought this post was helpful...
 Posted by Namatimangan08 on 14 August, 2012 - 12:53 pm
GT + generator inertia, I don't expect to be greater than 6MJ/MVA. The highest is nuclear powered ST @ 12MJ/MVA. Hydro turbine even smaller @ 2-3MJ/MVA despite the fact that its generator is huge for the same MVA. GT inertia is not that big. That is why we need the speed droop in the first place.

Let me show you what does it means.

Assuming you have 21 X 100MW (120MVA)capacity GTs serving the demand of 2000MW. All the GTs are loaded @ the total of 2000MW (95.3MW each). The remaining 100MW is spinning reserve requirement to comply with the N-1 concept, where N is the total number of on units on bars.

All of sudden one of the GTs trips, leaving 20 X 95.MW to serve 2000MW system. Assuming grid collapse frequency is 48Hz, then how much time do we have for the remaining GTs to load additional 100MW in order to arrest system frequency decay?

Total inertia = 20*120*6= 14400MJ or MWs

Generation shortfall = 95.3MW

Total duration before rotating parts to cease rotating= 14400/95.3 =151.1seconds. (from 3000 RPM to zero, assuming linear relationship)

Free Fall Frequency Rate constant = 151/50=3s/Hz

Time taken to reach 48Hz= 6 seconds!

It is a matter of fact this is very close to what will happen in reality if there is no governor speed droop to intercept frequency decay. "Locked into synchronism concept" does not able to save the system. Not even AGC can save the day. That explains why each prime mover for a power system must have a speed droop.

If we set the droop for generator at 10% (5Hz=100MW)percentage set point, that means the prime mover will load 2% (2MW)of its capacity for every 0.1Hz frequency decay. Therefore for every 0.1Hz frequency decay the total load provided by prime movers via the speed droop commands for all 20 units is 40MW. Thus at frequency 49.75, 95.3MW generation shortfall will be recovered, assuming all the GTs can provide loading rate of that much in less than 6seconds. This is what is called fast/slow immediate response of prime movers. We are talking about transient response with duration, more often less than 10 seconds. Stable grid system shall pass this acid test...

If you ask me what happened to India Grid lately, my simple answer but accurate to a certain level of trouble shooting is the grid operators failed to preserve the above dynamic. I'm very sure I'm right about it. What made them failed? That is another story.

 Posted by CSA on 14 August, 2012 - 6:53 pm
Vinayak asked why his load decreased by so much when the frequency increased. The "inertia" and load of the compressor must also be considered; we don't have any idea about the magnitude or period of the frequency disturbance(s), but reportedly it was severe in any case and not likely to follow any simple analysis.

And for whatever reason. Using blackouts to control load never seems to work very well.

We've never hear back from Vinayak, either.

 Posted by Vinayak on 18 August, 2012 - 2:05 am
Sorry for the delay in reply.

Dear Namatimangan,
Thanks for very detailed explanation, but we are running in parallel with grid, so frequency variation is more dependent on external factors than internal load variations.

Dear CSA, we are using Natural Gas fuel on our GTs. On that particular day frequency increased from 50 to 51.2 Hz. Primarily my understanding was that due to increase in frquency, compressor and turbine rpm will increase, so part of the fuel will be consumed in increasing rpm of machine considering its inertia and higher compressor work.

Second view is droop control is causing the MW drop.
But our machines were running on pre-select load and refering to your quote in your first reply:

Quote "The exception to the machine's response to grid frequency changes at Part Load occurs when Pre-Select Load Control is enabled and active. In this case, the turbine operating setpoint is a load setpoint, and even if the frequency changes and Droop Speed Control tries to change the load of the machine in response to the speed error caused by the frequency change. Pre-Select Load Control will try to maintain the load setpoint and over-ride Droop Speed Control" unquote.

I could not understand this huge MW drop.

If at all we assume that Fuel valve opening remains same, by thumb rules, due to increase in speed by 1.02 times, compressor power required should increase by 6%. This will account for nearly 3 to 4 MW. So is the balance load drop is only due to inertia?

Or is there some other reason for getting MW drop, looking at the drop it looks like Machine is asked to reduce load due to some reason. Please explain.

 Posted by CSA on 20 August, 2012 - 7:09 pm
Vinayak,

Frankly, I don't really know how to respond. By your own admission, the event was severe. We don't know a lot about how long these severe disturbances occurred, what the frequency of the excursions were, what the peaks of the excursions were. There's a lot we don't know.

Pre-selected Load Control "sits on top" of Droop Speed Control when it's enabled and active. It looks at the actual load versus the Pre-selected Load Control reference and adjusts the turbine speed reference (TNR) to try to make the actual load equal to the reference.

Droop Speed Control looks at the difference between TNR and TNH and this error is used to control the amount of fuel being admitted to the turbine. If TNH remains relatively constant and the actual load deviates from the Pre-selected Load Control reference, then Pre-selected Load Control adjusts TNR to change the reference to make the actual load equal to the reference. This works fine--as long as there are no frequency disturbances.

In the case where the frequency deviates from desired, TNH changes which changes the error between TNR and TNH and this causes the fuel to be changed which changes the load being produced. BUT, when Pre-selected Load Control is enabled and active and the actual load deviates from the Pre-selected Load Control reference, Pre-selected Load Control will change TNR to try to make error between TNR and TNH change to make the load equal to the reference.

This leads to "fighting" between the Droop Speed Control and Pre-selected Load Control, with one trying to decrease fuel while the other is trying to increase the fuel. Eventually Pre-selected Load Control "wins"--but that's NOT what is best for the grid and frequency control. Yes, your machine continues to produce load at the Pre-selected Load Control reference, but to help the grid frequency return to desired it should reduce load as any other machine running on plain Droop Speed Control is doing.

Now, when the grid frequency increases, the frequency of all machines connected to the grid increases--because they are synchronized to each other. The grid frequency increases because there is too much generation for the available load, so part of that extra energy goes into raising the frequency of all the machines.

Think about riding a bicycle on a flat road at a constant rate of speed against a constant headwind. All of a sudden that headwind decreases significantly. What will happen to the speed of the bicycle? What will have to be done to maintain the speed of the bicycle to be the same as before the headwind decreased? Now, suppose the headwinds are very erratic, and you are trying to maintain a constant speed. What will the speed of the bicycle be during these fluctuating headwinds?

There were severe frequency disturbances, per your own admission. Notably, you did not respond to questions about the rate of the archived data collection you are using to analyze the situation. As Namatimangan08 is so quick to point out, when there are frequency disturbances there is some "fluidity" or "sponginess" (for lack of a better term) in the system that allows some machines to instantaneously be slightly out of synchronization with other machines. This "sponginess" is caused by impedance, the nature of the load, mechanical inertias of the generators and prime movers driving them, and the distances between generators and switch yards and impedances.

There are lots of factors, and we could never know all of them for the part of the world where you live under the conditions that occurred when the severe frequency disturbances occurred. These frequency differences only last for milliseconds, unless conditions are perfect, and then the differences can persist as the system tries to come back to some kind of equilibrium.

If this load drop you are worried about occurred during an individual frequency deviation and was recorded then this could explain the larger than expected load drop.

When load is significantly reduced on a prime mover driving a generator, the first thing that happens is an increase in speed--even if only for a very brief period of time. And there is a pretty good-sized inertia for most heavy duty gas turbines--much larger than for a similar-sized steam turbine.

I would suggest that if the frequency increase was fast enough that the first reaction of the Speedtronic would be to reduce fuel--and very quickly--and that coupled with the speed increase caused by the grid could have caused a larger than expected (by droop regulation calculations) load drop. Couple that with a "fluid" or "spongy" system and the plot of load (and frequency) could be very unusual.

Something you haven't told us is: How long did this 50% load decrease persist?" 1 second? 10 seconds? 20 seconds? 60 seconds? Longer?

We don't know for how long the grid frequency remained high. 1 second? 10 seconds? 20 seconds? 60 seconds? Longer?

We just don't know enough about the exact details of the grid frequency disturbance, or how long the 50% load drop existed. Any analysis without knowing more is conjecture at this point (guessing, frankly, considering the grid disturbance was, again, severe in nature).

As Namatimangan08 is so quick to point out, individual generator frequencies (and load) can deviate from the grid frequency--but this only happens for milliseconds. If this was recorded by your data archival system, it could appear skewed from other data.

Your use of this site is subject to the terms and conditions set forth under Legal Notices and the Privacy Policy. Please read those terms and conditions carefully. Subject to the rights expressly reserved to others under Legal Notices, the content of this site and the compilation thereof is © 1999-2013 Nerds in Control, LLC. All rights reserved.

Users of this site are benefiting from open source technologies, including PHP, MySQL and Apache. Be happy.

Fortune
Electrical Engineers do it with less resistance.
Help keep our servers running...