Average response time in droop mode 3

[sparkyken]

Hi,

We have a power system that consists of 4 steam turbine generators and 2 gas turine alternators. In normal operations we have 4 of these generators to be running at the same time. The average power consumption (load) is about 10MW but it varies with the different times of the day.

All the generators are set to operate in droop mode and the frequency is set to be 50 Hz. If there was a load change, not a sudden load change but more or a gradual increase or decrease in load, how should the frequency of the generators respond? I am looking at the graphs, and when the load increased by about 1MW over about 10 hour period, the frequencies of the generators dropped gradually over the 10 hour period. The frequencies did not drop below 49.9Hz so i believe that this behaviour is completely satisfactory.

My only question is should the generators try to come back to operating at exactly 50Hz, say a couple of hours after the load increase was introduced, instead of waiting for the total load demand to decrease and then return back to 50Hz? Or is this behaviour normal?

Thanks for your time.

 

[waross]

This is normal behavior.
The factory setting for many smaller, preset generators is 3% drop and +3% offset. The sets run 3% fast at no load and run at the rated frequency at full load.
There are three operating modes that I would consider for a plant such as yours.
1> Set your offset so that you are at 50 Hz at normal maximum load.
2> Set your offset a little lower so that you are at 50 Hz in the mid load range.
3> Trim the frequency at regular intervals.
Offset: 3% droop corresponds to 1.5 Hz at 50 Hz base.
The governors are set at 50 Hz plus 1.5 Hz, or 51.5 Hz at no load. At full load the frequency drops to 50 Hz. although you have set the frequency at 51.5 Hz. This difference is called the offset. Your droop may be different than the small diesel sets and you are at liberty to set your offset to the value that gives the best compromise for your situation.
You may be interested in the discussion at thread238-231507.
Of particular interest may be the excellent link provided by catserveng.

http://www.woodward.com/pubs/pubpage.cfm?web_section=power

THE FINE PRINT SAYS SEARCH FOR DOCUMENT #26260

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[sparkyken]

Thanks for the quick reply Bill. It was really helpful. I'll check out the thread you have recommended as well.

 

[itsmoked]

Bill is it normal for an islanded system like sparkyken's to run all the generators in droop with none designated as isosynchronous?

If his system was connected to the grid then they would be lock to the system frequency and would not droop. Correct?

Keith Cress
kcress -

http://www.flaminsystems.com

 

[ScottyUK]

Keith,

Droop is what enables a machine to operate connected to a large system to operate in a stable manner in the absence of a more sophisticated controller. Over here the term 'frequency response' is used by the transmission system operator to describe how the machine loads as frequency drops and deloads as frequency rises. This is distinct from 'frequency correction' where the machine will actively try to restore system frequency to a target setpoint, within the limits of its capacity, by loading and deloading. This has similarities with a fairly aggressive droop controller.

On a CCGT station it is quite common to see the machine controlled by one of the following: a MW controller holding the unit to a fixed output regardless of the frequency shifts, which isn't great for the system behaviour but makes for an easy time commercially; or a thermal limiter where the engine is running as hard as it can at base load and both frequency and MW are uncontrolled. In an electricty market where under-generation and over-generation are penalised it seems the money men have won over the engineers in creating a system where generators look after their commercial interests by meeting their MWh despatch and to hell with system operability. Bloody accountants...



----------------------------------

If we learn from our mistakes I'm getting a great education!

 

[itsmoked]

Droop is what enables a machine to operate connected to a large system to operate in a stable manner

But OP wise the droop is resulting in a frequency deviation. On The Grid no frequency deviation can be allowed on a unit away from the grid frequency.. Is this just quoted droop you speak of some other kind of droop? Voltage only?

Keith Cress
kcress -

http://www.flaminsystems.com

 

[itsmoked]

OOPS! Quote should be attributed to Scotty.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[ScottyUK]

That's right, the grid frequency of a large system can't be influenced to any practical extent by a single machine: the droop characteristic causes the machine output to drop off as the grid speeds up, and to load up as it falls. In an islanded system, which is what I interpret the OP to be describing, as the load increases the speed has to fall slightly because that will open the throttle to meet the higher load. Normally droop is used to ensure that a number of sets share the load rather than one hogging the load.

The more sophisticated control schemes have droop enabled for load sharing purposes within an islanded group but also use a supervisory trim controller which will slowly restore the island frequency to nominal when averaged over a period of time.


----------------------------------

If we learn from our mistakes I'm getting a great education!

 

[itsmoked]

Thanks Scotty.
My understanding is intact.
Whew.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[waross]

Hi Keith;
Scotty pretty well covered it but i'll try to say the same thing in different words.
Case 1;
Droop in parallel with the grid. Assume that the grid frequency is stable. If you set your droop governor to 60 Hz. while connected to a 60 Hz. grid, there is no transfer of power, (but you can transfer vars). Assume 3% droop. The normal speed setting is 61.8 Hz. When islanded, at full load the frequency drops to 60 Hz. Now if we set the same governor to 60.9 Hz, at 50% load the frequency will be 60 Hz.
When the frequency of the set is controlled by the grid, adjusting the speed setpoint from 60 Hz to 61.8 Hz will give stable control of the loading.
Case 2;
Islanded plant with parallel machines in droop. Now the whole plant operates in droop mode. Droop on the individual machines allows different size machines to share the load in the ratio of their sizes.
My little system in the tropics used to run diesels in droop. Every 15 minutes the operators were required to record;
Temperature
Oil pressure
Current for each phase
Voltage for each phase
kW loading for each set
Total kW loading
Frequency
If the frequency was drifting they would tweak the frequency.
As the load increased they would add sets.
As the load decreased they would drop sets.
This meant that the frequency excursions were generally much less than the maximum 1.8 Hz that 3% droop would allow with a 100% load change.
Now a load control panel interacts directly with the electronic governors and or throttles on the individual sets, balances the loading and keeps the frequency stable.
Where we could do an excellent job of control with mechanics and hydraulics, the computer driven load control panels can do a perfect job. But it takes a lot of the challenge out of the job and you can't fix it with an adjustable wrench and a multi-meter.

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[itsmoked]

Ok, So when setting droop in a gridded situation the speed really is locked, so the droop setting is messing with the throttle which is only going to alter torque, and hence kW. In this case you can change any particular sets 'offset' to control its contribution to the grid.

I was seeing droop as only a load/speed thing when it is really more of a load/throttle relationship. With a non-rigid network droop will translate into a speed,(freq), result and with a rigid network it will result only in a load/kW only result.

Thanks for the explanations!

Keith Cress
kcress -

http://www.flaminsystems.com

 

[rovineye]

I run up to 4 generators in parallel on most of my ships (islanded). They are all set up with 3% speed droop. The slower responding power management system will then adjust speed back to base 60 hz.

I do also use voltage droop, again at 3%, based on VAR, 1 1/2 % each side of base 6,600 volts. Power managenet will then adjust back to base voltage.

This helps get close to proportional load sharing without reverse power/VAR tripping.

I do dynamic reactive load testing but have never tried while in parallel with power managent off to see how stable I would be with droop alone. If I even have a ship where nothing fails I could play with that....if if if.

 

[waross]

You have it Keith.
When you are locked onto the grid you could just pull the throttle to get the desired loading and then lock it there. That is the throttle not the governor. The problem is that if you are running at full load and full throttle and lose your load, you may overspeed your set to the point of destruction. When you use the droop to control the throttle position the governor takes over and limits the speed if you lose the load.

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[sparkyken]

Thanks for the useful explanations and Keith for adding a bit of spice to the discussion. In my case, the power system is an islanded system with 6 parallel generators, but we normally have 4 or 5 of these generators running at any one time. This power is used for operating a gas processing plant.

While we are discussing about droop, can I also ask you guys something? At the moment all the generators are set to operate in droop mode. We are considering of changing the operation to all isochronous except for the largest gas turbine generator to run in droop. The problem is we have different sized generators, inertia constants and control systems all operatiing together. I have read somewhere that to run in isochronous mode, it is recommended to have similar sized generators and control systems.

Is the statement above true? What might be the possible problems we might encounter? In your opinion, would you recommend the move from droop to isochronous for a small islanded power system with different generators?

Thanks again for your input.

 

[wayne440]

If I understand your original post, you are able to reliably stay between 50.0 and 49.9Hz with all of your machines in droop mode. To me that is a compelling reason to leave it as is.

 

[waross]

When you put one machine in isochronous, you can consider it to act like the grid. It will set the frequency and hold it stable.
But if the load goes too high or too low, the system will fall back on droop operation.
Too high and too low are relative to the number of sets running and the settings.
I will assume that the sets are 1 MW each, for ease of example.
A possible operating sequence may be;
1 Start one machine in isochronous. This machine will now control the frequency.
2 The load is up to 800kW. Start another machine and place it in droop. Set it to supply 500 kW. (50%)
The isochronous machine will now control the frequency through load changes from 500 kW to 1500 kW. If the load changes out side these limits the system will function in droop from 0 kW to 500Kw and from 1500 kW to 2000 kW.
You may add more machines for increased load but you will always have a "Window" of isochronous operation determined by the size of the set in isochronous mode. Outside the window and you are in droop mode.
If your load swings are slow enough that you can add or drop sets or change governor settings to stay in the isochronous "window" you may go isochronous.
Remember though that this is old technology. A new plant may have a load control panel. My inclination would be to stay in droop mode. If you are having serious problems with droop mode, you may consider implementing a modern load control. panel.

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[itsmoked]

Thanks sparkyken!

Let clarify some more:

I don't think you would want more than one generator in isosynchronous mode at one time. They would struggle to force each other to get to their individual 'correct' frequencies as they would never actually be identical.

Toss in three or four in iso and

?

Or is this not correct?



Keith Cress
kcress -

http://www.flaminsystems.com

 

[waross]

You are correct, Keith.

Bill
--------------------
"Why not the best?"
Jimmy Carter