Speed of Synchronous Generator

The generator doesn't hunt.

The generator prime mover's governor will hunt if not properly tuned. And we're supposed to be talking about properly acting governors. Not laboratory exercises.

When your plant is connected to the grid, if you start a large motor the grid will supply that power until such time as your operators increase the power produced by one or more prime movers to return the export to 2-3 MW. In fact, if one of the turbines trips, won't the refinery continue to run on grid power? (Even if there is some load-shedding scheme in place/required?)

When you are connected to a sufficiently large grid, there is no difference between turbine frequency and grid frequency. Your little 25 MW unit isn't likely to have much of an effect on a 2000 MW grid, or even a 1000W grid, or even a 600 MW grid. The frequency of all the machines connected to the grid is the same--because they are synchronous machines. Synchronized together.

Even on a grid composed of two 25 MW machines, the speed of both machines is identical and is directly proportional to the frequency. And if the power produced by the two machines isn't equal to the power being consumed by the load, then the frequency isn't going to be at rated.

Full stop.

The question remains: How much does the speed of the turbines at your site change when they are loaded while connected in parallel with a other machines on a grid of relatively stable frequency? Even if the frequency isn't 50.0 Hz, how much does it change when you increase the fuel to one of the turbines causing it to accept more of the load on the grid? And for how long does the speed change?
 
>Part - 1 : the Theory

Vector diagrams are welcome.

The question remains: How much does the speed of the turbines at your site change when they are loaded while connected in parallel with a other machines on a grid of relatively stable frequency? Even if the frequency isn't 50.0 Hz, how much does it change when you increase the fuel to one of the turbines causing it to accept more of the load on the grid? And for how long does the speed change?
 
P

Process Value

> pick-ups of the Mark II and Mark V units from your site to see if the two units are running at different speeds while connected to the same grid. And for the Mechanical Dept. to provide the nameplate data from the load gear box nameplates. <

> As ProcessValue has pointed out, no place in the world has a perfect grid frequency, but as long as the frequency isn't changing by more than +/- 0.25 Hz in a very short period (seconds or less) we could review the results. <

> The point is that units are not normally loaded or unloaded, except possibly at ProcessValue's site, by "throwing on" or "throwing off" blocks of electrical load as he wants to do with his test. They are normally loaded and unloaded using the RAISE SPD/LOAD and LOWER SPD/LOAD buttons/switches/targets (or the Preselect Load Control enable functions, which essentially drives the turbine speed reference up and down just like the RAISE and LOWER functions). <

> What we want to know is how much the actual running speed of the turbines at ProcessValue's site change by and for how long when they are loaded and unloaded using the RAISE- and LOWER SPD/LOAD functions, which is how units are normally loaded and unloaded around the world. Not by throwing on or throwing off blocks of load, which is not a typical loading or unloading method. <

Part - 2 : The experiment

what i have said above is theory , which to the best of my knowledge is true. It is very hard to get experimental data on load angle calculations , ( there

is a load angle calcualtor of the generator on the AVR but i do not know how to get the real time data from the avr , when i asked the vendors they said such provisions is only at their test site and it is not possible on the given system :( . ) but with certain operational constraints i have done the following

experiments over a period of two days.

1. Auto loading ; machine parallel to the grid with preselect control

2. Auto Unloading ; machine parallel to the grid with preselect control

3. Manual Loading ; machine parallel to the grid with droop mode

4. Manual Unloading ;machine parallel to the grid with droop mode

5. Opening of the grid breaker under export condition

6. Opening of the grid breaker under import condition

7. Synchronizing with the grid

8. Heavy motor starting in section

I let the trend recorder run for one whole day and i got some interesting trends on how the machine behaves to grid frequency occilations and sudden and rapid frequency change in the grid. The machine was kept in preselect condition during the test. The archive file below give the trend snapshots and the trend data in csv.

http://www.2shared.com/file/G7yF_WMp/grid_frq_change.html

I trended the following signals

1. DWATT - machine MW output
2. DF - Machine frequency calcualted from the machine speed sensor TNH
3. TNH - Machine speed
4. TNR - droop reference
5. L70L and L70R - the machine speed raise / lower command which is responsible for increasing or decreasing the TNR
6. FSR - the fuel stroke rate in percentage
7. FQLM - the fuel flow in kg/s
8. SFL1 - the bus frequency , this is calculated by mark vi VTUR board from the Bus PT connection. BUS PT is in the CPP sec C bus
9. sdiff2 - the difference in the frequency between the machine and the bus. it is the value of ( df_vtur - sfl1)

The experiment setup - The present site i am on has the following configuration. The test was done on GT-3 as present in the drawing. The Gt is a frame 5

machine with 22 MW capacity at site condition. The machine is a dual fuel capability machine now running in naptha.

Pic of the test setup - http://www.2shared.com/document/YLmm69rE/test_setup.html

The complete trend snapshot and trend data in csv format is uploaded here in the archive file GT_test.rar

http://www.2shared.com/file/np0jvybf/GT_tests.html

Limitations of the test setup - The major limitation of the test setup is the lack of a frequency source which is in the grid. I had to take the bus frequency which is similar to the grid frequency under parallel condition. This is what is used for the auto sync.

1. Auto loading of the machine - The machine was in preselect mode at 14 MW and the setpoint was raised to 18. it is seen that during parallel operation to the grid the MW is hunting by 0.2-0.3 MW continuously. it is also seen that the machine frequency is hunting with the bus frequency during the whole operation. ie the machine was constantly accelerating and de accelerating over the setpoint , not by much but by but by a maximum of 0.0015 to negative of 0.002 , it also oscillates the whole way during the ramp up to the 18 MW. This is similar to the condition in "section A " of the explation . Machine parallel to the grid with increase in the fuel/steam input to the machine. here what we do is increase the fuel to the machine.

2. Auto Unloading of the machine - Here the machine in preselect mode , from 18 MW the set point was given as 14. similar to the auto-unloading of the machine but the occilations seem to be on the higher side here , the hunting was from a maximum of 0.002 to a negative of 0.0025. MW was around 0.3-0.4 MW all during the ramp down period. This is similar to the condition in "section A" of the explanation . Machine parallel to the grid with increase in the fuel/steam input to the machine. here what we do is decrease fuel to the machine.

3. Manual Loading of the machine - similar to the auto loading , except that a manual raise command was given three times. check out the results in the folder.

4. Manual unloading of the machine - similar to the auto unloading , except that a manual lower command was given , check out the results and graphs in the folder.

5. Opening of the grid breaker with export - This is very similar to a load throw off in a independent machine. an export was 2.5 MW was maintained when the grid transformer breaker was opened. this led to the sudden load throw off , sudden increase in the speed. This is similar to the "section B , B1" of the explanation. here there is a sudden reduction in the load. this is where you can really see the machine hunting. the speed during the initial throw off jumped by around 0.01 Hz . the machine frequency which was around 49.65 raised to 49.95 a increase of 0.3Hz for a load throw off of 2.5 MW , indicating a droop of around 3.34 %. This speed hunting is mainly because the machine cannot match the speed in which the load was thrown off. it can be clearly seen in the graph that even though the reduction in load is near instantaneous the fsr and thus the fuel reduction takes some time more , this is the reason for the speed spike and the subsequent speed hunting.

6. Opening of the grid breaker with import - This is very similar to the sudden loading of the independent machine. An import of 2.5 MW was maintained when the grid transformer breaker was opened. This led to the sudden loading of the machine and thus a sudden reduction in the speed. A very similar result to the above just that the speed reduction took place in accordance with the droop. one interesting aspect is that the hunting died off quite quickly in this case.

7. Synchronizing with the grid - This is where the actual Load and speed hunting in the machine can be seen clearly. This is a transient chara of the machine .

8. Sudden frequency increase in the grid - This is similar to the "section A ; A1" explanation. This trend is available in the grir_frq_change.rar archive. here you can see that the grid frequency raised from 50.1 to 50.22 suddenly , now the machine power came down from 19.5 to 18.6 before recovering as the machine was in preselect.

so that is all from the experiment side , i forgot to do manual raise lower and raise in the independent mode of operation . i will upload the data as soon as i do it. besides this all the cases in the theory Section A A1;A2 and Section B B1 is covered and it can be seen that the theory matched the results.
 
P

Process Value

" The generator doesn't hunt.The generator prime mover's governor will hunt if not properly tuned. And we're supposed to be talking about properly acting governors. Not laboratory exercises."

Not entirely true , governors are tuned for and at steady operation and to an extent small signal disturbance. For transient condition the generator will hunt before setting to the final steady state condition. an example would be , synchronizing , large motor start in the section , sudden frequency change in the grid , transient fault conditions.

I have given experimental data on all the above , in all the condition , synchronizing , large motor starting , and sudden frequency change the MW output of the machine hunted before it settled down on a steady state value. How well it recovers from the hunting depends on how well the governor is tuned. Governor cannot totally prevent a load hunting it can dampen it and stabilize it as quickly as possible. and in this regard it is not only the governor of the generator which plays the part, the AVR is also hugely responsible for maintaining a stable operating generator. Modern AVR's are equipped with

a. Load angle limiters
b. Over excitaion limiters
c. Inductive current limiter
d. Capacitive current limiter
e, v/f limiter
f. Power system stabilisers

to name a few. They help in reducing the extent of the hunting and the reduce the time needed to settle down to a steady state. Generator hunting are inevitable due to the nature of the gird and the nature of the loads.

" When your plant is connected to the grid, if you start a large motor the grid will supply that power until such time as your operators increase the power produced by one or more prime movers to return the export to 2-3 MW. In fact, if one of the turbines trips, won't the refinery continue to run on grid power? (Even if there is some load-shedding scheme in place/required?) "

True , this is exactly what happens in the steady state condition. I have given the motor starting trends , if you will see in that the machine is the one which actually takes up the load first (transient) then it reduces. yes in case of a machine trips the grid will supply power.


"Even on a grid composed of two 25 MW machines, the speed of both machines is identical and is directly proportional to the frequency. And if the power produced by the two machines isn't equal to the power being consumed by the load, then the frequency isn't going to be at rated."

Yes in a grid composed of two 25 MW machines the frequency is identical. But i do not get your point about power produced by the machines not equal to the power consumed , under steady state condition the power produced by the machines will always be equal to the load power + losses. even in a small micro island it is possible to run the machine at rated frequency.
 
The generator (alternator) is a "stupid" device. It just converts torque into amps, and in the process can handle some reactive current. (I almost said power! But I did use an exclamation point! Oops, I did it again.)

The alternator is driven by the prime mover; the reactive component is driven by the exciter regulator. The prime mover governor can hunt, and the exciter regulator can hunt. But the generator doesn't hunt. Only the devices that drive the loads connected to the generator can hunt, making it appear that the generator is hunting but it's really the drivers that are hunting.

Any instability is due to transient conditions being reacted to by the prime mover governor and/or the exciter regulator.

But we digress. Because, again, the focus is being distracted from the original question.

And, again, we're not talking about transient conditions. The question was about steady-state loading and unloading and it's effects on the speed of the turbine (and frequency of the alternator) when connected (synchronized) to a grid in parallel with other alternators and their prime movers.

So, from the data it seems very clear: There is no appreciable change in speed or frequency when the unit is loaded or unloaded while connected to a grid in parallel with other prime movers and alternators. Not including load throw-off or load throw-on, which is really subtracting or adding from the load on the grid, of which the alternator for which data is being gathered is only one portion.

Again, when we're talking about "loading a unit" or "unloading a unit" we're talking about using the RAISE SPD/LOAD or LOWER SPD/LOAD targets to increase or decrease the amount of fuel being admitted to the turbine, which will increase the amount of torque being applied to the alternator rotor. But, because the alternator is <b>SYNCHRONIZED</b> to a larger grid composed of multiple prime movers and alternators all supplying a load composed of motors and lights and computers, the increased torque can't cause the speed to increase appreciably (unless the prime mover is very large with respect to the other prime movers on the grid, or the grid is "soft") and so the increased torque results in increased amps flowing through the alternator's stator, which is referred to as "increasing the load." The MW meter of the alternator in question will increase in the positive direction, hence, the "load" increases.

In other words, in non-laboratory conditions (i.e., real world conditions), when connected to a grid in parallel (synchronized) with other alternators and their prime movers, the speed/frequency does not change when the load changes when loading and unloading the units using the governor RAISE and LOWER functions. The vector data and the "test" results prove that.
 
That explanation was quite helpful.
but i have one question .

as you said that "The total amount of electrical generation must exactly match the amount of electrical load in order for the grid frequency to remain at rated. If the total generation exceeds the load then the frequency (and the speed) of all the generators will increase above the desired grid frequency. If the total generation is less than the load then the frequency (and the speed) of all the generators will decrease below the desired grid frequency."

my question is how do we keep the total load to such an exact value as the load is the sum of all the appliances like lights, motors, computers, traction loads, how is this load matching done as there are peak periods?

so what type of system is used for load control so as to keep the system frequency constant?
 
It depends on how precisely you want to control frequency. If the grid loads (and the turbines driving the generators) can tolerate frequency wandering between, say, +/-0.5 Hz of nominal (50 or 60 Hz) then a central dispatcher can manually dispatch generation up/down to keep frequency within tolerable limits.

Much more commonly, though, the utility has a scada system which feeds MW and frequency data to an Automatic Generation Control application (AGC). AGC runs typically every 2, 4 or maybe 6 seconds and sends generator MW controls out to raise/lower generation at selected units to return frequency to nominal if it wanders too far away.

If the utility is connected to neighboring utilities via AC transmission lines, the areas meter the power exchanged between them and compare that with the amount they actually want to exchange. That difference is added to the frequency error and, again, AGC sends controls to restore the so-called Area Control Error (ACE) to some small value.

"Small value" is a matter of perspective: a utility serving 10000 MW of load might control its ACE to within +/- 20 or +/- 60 MW of zero. It depends on the nature of the load they serve, where they are electrically in the grid, etc.

Note the amount of frequency deviation that AGC responds to can be very small, easily on the order of 0.002 Hz. There is a relationship between generation-load balance and frequency: a 100000 MW system suddenly deprived of 1500 MW of generation might see its frequency drop by 0.1 Hz (it would go further if the governors didn't respond to the frequency decline.) So 0.002 Hz on such a system translates to a generation change of 30 MW - the size of many of the turbines I read about in this forum. (this is why wind generation is one of the worst things to ever happen to the power industry - and the rates you eventually pay: the intermittent and undependable nature of the generation drags frequency around, which increases regulating demands on the controllable units. It also requires much more regulation capacity to handle the enormously increased range required by the amount of variability. So utilities must build more generation, not all of which is supplying load.)

You coincidentally used terminology that is quite common in the utility industry - load control. Another term for AGC is Load-Frequency Control, or LFC. In some places, it's just "Load Control".
 
N

Namatimangan08

With due respect, CSA has small problem to understand the argument along this line.

Straightly speaking two generators in parallel most likely are never have the same frequency. As long as their Synchronism Torque Angle (STA) will not deviate by greater than 180 degree, theoretically their governors can keep them under synchronism.

Loss of synchronism occurs when STA is greater than 180 degree. In practice the deviation is kept well below 70 degree.

Whatever you want to call it: load, prime mover or generator. All of them are prompt to hunting.

It is true as pointed by somebody in this forum that a power system doesn't work if there is no frequency deviation. All prime movers (that includes generators) in the system do not recognize what is load. They can only recognize speed. Sound strange right? But it true. So their responses are based on frequency deviation.

FYI- In 2005 our grid system with running capacity of 15,000MW was hunting by the order of +/- 200MW for about 20 minutes. What I'm trying to say is when it comes to load swing (load hunting) system size doesn't matter. What matter is the size of transient load whether its magnitude is big enough to trigger the swing.
 
> With due respect, CSA has small problem to understand the argument along this line.

Okay, Namatimangan08, so exactly how much does the speed differ between synchronous generators connected in parallel to each other and supplying a common load?

Please be precise. 1 RPM; 10 RPM; a fraction of an RPM.

I will concede that there are differences in acceleration but they are transient differences.

Please enlighten me, and all of us, with your precise measurements of speed differences. I'm always willing to learn.

Thanks!
 
N

Namatimangan08

Your welcome mate. We share and we learn.

Remember my "controversial" that I have made probably last year of a year before. "scientifically there no such thing of infinite grid". I also mentioned that if you don't believe with this argument you will find one major problem to explain precisely about how a power system works without violating the conservation of energy.

Infinite bus is NOT a scientific principle. On its own, it does not exist. How we can prove it?

Go to a generator. Look at system frequency. Wait it steady state operation is achieved, ie. its frequency remains constant say at 60Hz says for 5 minutes (If you can find one). Find out the total power production from all prime generators attached to the grid. Says 15,000MW.

Next by the conservation of energy, if frequency hasn't change then generating power is equal power consumed.

For every 1hour the grid supplies 15,000MWh. The demand consumes 15,000MWh to. Therefore frequency remains at 60Hz.

From there on let us raise the generator in front of us by 1MW. So the total supply has become 15,001MW. The demand will not change since that 1MW is hardly change its terminal voltage. For a specific purpose assuming the grid system has no speed droop, no AGC and no AVR. What happens to the system after 1hr, 10hrs, 100hrs, 1,000,000hrs?

You are going to from the conservation of energy that supply side will increase as time goes. The demand remains the same. If we allow time to approach infinity we create vast unbalance between both. That indicates the system cannot exist since it cannot comply with energy balance. Stable system must comply with energy balance.

I know about inertia. That is my important point. Inertia is a part of the "infinite grid" too but less understood by many. From the above example we can't construct a working model for grid if we don't add inertia to the system.

Finally, we put speed droop, AVR, operators' interventions, AGC and control to the system. We introduce the physical parameter that is called inertia. We put systematic method to control. Then only the working principle a grid system can be explained flawless.

In summary, infinite grid terminology inclusive all these elements. It is actually the destination rather than a scientific concept that many people are inclined to believe.
 
N

Namatimangan08

If you agree that transient can exist then you have to agree with permanent frequency difference. I thought I have posted my true experienced dealing with an isolated system with 4 Diesel Gen sets that almost never able to get their frequency equal longer than 5 seconds. It was quite long explanation, Unfortunately, I think it went missing or probably I didn't press submit button.

To provide direct answer to your question, I would say most medium to big grids have little problem to deal with frequency deviation between two parallel generator up to 30rpm for a 3000rpm system. Roughly about 120 RPM, Some of hydro units might be put to partial shutdown generator by protection scheme. At 240 RPM normally the generator will be removed from system whether under electrical over speed or out of step/out of phase/slip pole/loss of synchronism protection. At 300RPM mechanical over speed protection... Just to be sure in case electrical over speed fails to deliver the required task.
 
Dear Process value, we are waiting for your valuable posting Part3 (grid operation document)
(Ref;6th Jan post)

Thanks & Regards,
Ashish
 
Namatimangan08,

You wrote:

> To provide direct answer to your question, I would say most medium to big
> grids have little problem to deal with frequency deviation between two parallel
> generator up to 30rpm for a 3000rpm system.

How long can "...most medium to big grids..." deal with these 1% frequency deviations between two parallel generators?

I'm interested in hearing from other readers and contributors about their experiences between generators operating in synchronism with a grid and other generators when the grid frequency is normal. How much deviation from rated speed/frequency do you observe when running at steady state conditions (stable power output)?

I do not disagree that load angle/torque angle--whatever someone wants to call it--will differ between generators being operated in parallel with each other. And, I believe that when loading and unloading generators being operated in parallel with each other that the momentary (on the order of milliseconds) acceleration changes.

But, I do not believe that synchronous generators being operated in parallel with other synchronous generators can operate a differing frequencies or speeds (since the two are directly proportional). That's the whole concept of synchronism. Why else would it be necessary to synchronize an incoming generator with other generators? Why not just close the breaker if the frequency of a generator doesn't need to match the frequency of the other generators?

Why does a generator being operated in parallel with other generators go into reverse power when the torque input to the generator is reduced below the amount required to keep the generator rotor spinning at synchronous speed/frequency?

Please, Namatinamgan08, enlighten us. Provide the mathematical formulae to support your position. Answer all of the above questions with actionable data, not personal thoughts and impressions.

Where is the proof of your contention?
 
N

Namatimangan08

How long? If no out of step protection, it will take as long as it takes until it goes to over speed protections or internal damage has occurred.

The first over speed protection is the unit goes to reverse power. Why? In an attempt for the governor for that unit (Via speed droop) to ensure it will not go faster than the average frequency. The governor will try to dis-accelerate the rotor by reducing the torque. But if the unit is out of synchronism already no amount of reducing torque can bring it into synchronism again. Its shaft will continue to accelerate since there is no opposing torque from the load to slow its frequency. Finally it generator turns to reverse power.

If feel doubt about this explanation, let you try to explain how a generator under parallel operation can go over speed but the grid remains intact.
 
N

Namatimangan08

Ok. Let us understand system dynamics from Newton's laws: Swing Equations for parallel generators are as follows:<pre>
J (d2A/dt) + c(dA/dt) +kA = Tm-Te

Where

J = Moment of inertia (Prime mover + generator + others) kgm2

c= Damping constant

k= Stiffness constant

A= rotor angle from stationary point of reference

d/dt = time derivative

Tm= output torque

Te= opposing torque. </pre>
If you have 20 units in parallel you have to write 20 similar equations to describe dynamics for each units. When they are in parallel they are governed by those equations. You don't need anything more that those equations to describe how a power system works dynamically whether under steady state or transient. Surprisingly, as I mentioned before there is no place for the what so called "locked into sychronism" in the equations. It was not needed.

There is such thing that is called locked into synchronism. But it meaning is not the same as many of us wanted to believe. This is similar to infinite grid concept. Let me explain further.

Steady state means Tm-Te=0. So the LHS is zero. The shaft for the prime mover will not accelerate, not going to move faster and also not going to change its displacement angle relative to rotating magnetic field.

If Tm> Te then the LHS is not zero. Energy is added to the rotating grid. This additional energy will be used up for (1) increase d2A/dt (2) increase dA/dt and (3) changing the A. Its angle may be displaced by certain radians relative to the rotating magnetic field.

The magnitude of d2A/dt is determined by the the value of J. The bigger J is the smaller its d2A/dt

The magnitude of the dA/dt is regulated by the droop via its droop set point- direct intervention of the input to prime mover. The higher its percentage set point the slower the droop to provide frequency damping. kA in controlled via manual intervention- Calibration of the unit output so that the load it dispatch is "housed" at predetermined frequency. Says 100MW at 60Hz. Note that 100MW at 60 Hz and 100MW at 59.9Hz as far as grid operation is concern is not similar. This is especially true for the units that are put under AGC -frequency control.

To conclude, "locked into synchronism" means all the above control requirements have been successfully tuned to deal with steady state (Tm-Te=0)or transient load changes(Tm-Te) is not zero. Successfully tuned means that in the event of calculated loss of generation (or demand load rejection) the controllers and the J constants for all generators are able to keep synchronism torque angle for all generators to stay within +/- 180 degree apart. Equal frequency is not a constraint that is required by the dynamic equations. Otherwise you don't need to have 20 swing equations for 20 units.

BTW: I have seen many times that two parallel generators did not have the same electrical frequency. Seeing is believing. The latest one was 2 months ago. I will tell you the story about it if the plant owner manages to solve this problem later.
 
N

Namatimangan08

> Namatimangan08 wrote:
>> To provide direct answer to your question, I would say most medium to big
>> grids have little problem to deal with frequency deviation between two parallel
>> generator up to 30rpm for a 3000rpm system.

CSA wrote:
> How long can "...most medium to big grids..." deal with these 1% frequency
> deviations between two parallel generators?

> I'm interested in hearing from other readers and contributors about their experiences between generators operating
> in synchronism with a grid and other generators when the grid frequency is normal. How much deviation from
> rated speed/frequency do you observe when running at steady state conditions (stable power output)?

---- snip ----

There are two kinds of torques that we are talking about. One part is opposing torque that is produced by the demand. The other part is traction torque that is produced by the dynamic of prime mover and generator.

Which one you are referring too? All the while I was referring to the dynamics of prime mover and generator.
 
Namatimangan08,

Realizing that you probably work in a certain Asian sub-continental region of the world famous for lack of frequency control and poor grid frequency regulation you may indeed see more unusual circumstances than many readers here. Actually, that might be changing as this site seems to be attracting more inquiring minds from that part of the world looking for explanations to unusual operating conditions.

I will maintain that a properly regulated--or "tuned" to use your terminology--will result in all synchronous generators operating in synchronism (in other words: at the same frequency/synchronous speed). You are now saying that a generator might be at 59.9 Hz versus another unit at 60.0 Hz (or 0.167%), when you previously said a machine could operate at 10% difference. It's very confusing.

Just a few months ago I had the opportunity to visit a grid control headquarters for a very large grid which is well-operated and famously stable, and I inquired about generators operating at different frequencies on the grid. The looks I received were wilting, meaning I wilted under the stares.

I think I'll stick with my "problematic" and simplistic understanding of the basic fundamentals of AC power systems, and leave the transient, freakish off-frequency operating characteristics for others.
 
P

Process Value

Grid and Grid

Wow, i return to control.com to see a old thread popped up :) and with a raging discussion going on. back to the good old days i guess. Here is my take on the situation.

Namatimangan08 Arguments

Argument 1 - "If feel doubt about this explanation, let you try to explain how a generator under parallel operation can go over speed but the grid remains intact. "

from what i get from the statement is that a generator connected to a grid can go to "overspeed" conditions even when grid frequency is within operating limits. This is plain wrong.

When a generator is connected and it has a governor ( or even no governor as in lab conditions) it will never go to overspeed conditions under small signal disturbances. Pole slipping does not mean that the the generator is overspeeding , it means that the load angle has gone beyond its stability limit (theoretically 90 deg , practically it will be around 80 due to the presence of resistance in the system). once connected to a grid you can do two things with the generator, increase its output by controlling the prime mover or reduce its output by controlling the prime mover, there will be transient speed differences when the system settled down to its new load angle but otherwise it will remain the same with some minimal hunting.

Case 1 - Increase turbine power when connected to the grid

you go on increasing power the generator output will increase till it reaches max output. in this condition only the load angle of the generator increases not the speed. suppose that you are mad enough you have manual fuel/steam control you increase the turbine output again , the generator will load , its load angle will increase and after some time when it exceeds the max stable load angle it will go to pole slipping condition. this case is possible only in laboratory conditions. when generator is designed , its Xd value is chosen in such a way that its load angle does not exceed 50-55 deg at the max output at the rated "terminal voltage" and at "minimal excitation" point of the AVR corresponding to full load. I do not believe that any generator connected to the grid has ever tripped on overspeed (tripped on pole slipping yes ,definitely possible but a rare occurrence) when the grid remains at stable frequency. Overspeed can happen because of load throw off not the other way around.

Case 2 - Decrease turbine power when connected to the grid.

you go on decreasing turbine output , the generator output will decrease and eventually will trip in reverse power. "the fundamental concept in AC power system is that the bus giving power must be leading to the bus receiving power. Thus when the generator is supplying power to the grid it is leading the reference grid bus. when the power goes down the load angle decreases and decreases and eventually will lag the grid bus reference thus receiving power from the grid ,and thus the reverse power trip.

"The whole phasor diagram in AC machine analysis is based on the fact that the two phasors representing the grid and the generator in this case have ZERO RELATIVE SPEED ie they are rotating in the same speed. " deviations in frequency is what is responsible for the load angle change. once a new load angle is reached the speed of both the grid and the generator remains the same.

Argument 2 - " To conclude, "locked into synchronism" means all the above control requirements have been successfully tuned to deal with steady state (Tm-Te=0)or transient load changes(Tm-Te) is not zero. Successfully tuned means that in the event of calculated loss of generation (or demand load rejection) the controllers and the J constants for all generators are able to keep synchronism torque angle for all generators to stay within +/- 180 degree apart. Equal frequency is not a constraint that is required by the dynamic equations. Otherwise you don't need to have 20 swing equations for 20 units. "

I agree to the above , but this is about the swing equations you have given is for transient stability analysis not small signal analysis. AGC and PSS uses the swing equations to various degrees to control but though it is certain that the frequency hunting takes place , i am quite sure that there will not be a " constant frequency " difference between the grid and the generator. ie if the gird is operating at 50Hz , the machine may swing at 50.01 and 49.99 but will not be at a constant frequency deviation say 50.1 for the entire operation; that is plain not possible.

Argument 3 - BTW: I have seen many times that two parallel generators did not have the same electrical frequency. Seeing is believing. The latest one was 2 months ago. I will tell you the story about it if the plant owner manages to solve this problem later.

Are the two generators operating independently off the grid or are they connected to the grid? If they are operating independently then there is a possibility of frequency oscillation. This usually happens due to a badly tuned iso-load sharing scheme. If the two machines are put in droop they will not hunt ( equal droop or not). here is a chart to explain what can happen in islanded operation.<pre>
Governor mode Machine 1 Machine 2 Remarks

droop droop stable
droop load control stable
iso droop load will hunt depending on the iso setpoint and current frq iso load control Disaster :p potentially unstable behaviour
iso iso needs load sharing scheme , otherwise will lead to hunting
load control load control disaster again </pre>
This is the reason if you will see almost all of the micro grids operate in either full droop , droop-loadcontrol or in iso load sharing scheme.

CSA's quiries

Query 1 - How long can "...most medium to big grids..." deal with these 1% frequency deviations between two parallel generators?

well i take it that you mean "two areas" rather than two generators. In every grid there will be inter area oscillation. These are very low frequency load hunting due to the relative change in the speed between two areas which are synchronised and connected by tie lines. the oscillations occur due to frequency deviations of about 0.02 Hz or 0.01 hz not 0.5 hz. wide area oscillations as far as i have seen has been for about 0.2 Hz during a severe sustained fault in one area.

Statement 2 - Realizing that you probably work in a certain Asian sub-continental region of the world famous for lack of frequency control and poor grid frequency regulation you may indeed see more unusual circumstances than many readers here. Actually, that might be changing as this site seems to be attracting more inquiring minds from that part of the world looking for explanations to unusual operating conditions.

ha ha ha well, we thrive in adversity don't we ;). Next time you visit a grid control station of a ISO(independent system operator) or a Load dispatch center you can ask about interarea oscillations , it will be present in the most stable of grids too. As far as frequency regulation is considered, it will be about +- 0.2-0.4 Hz for stable grids (UK and Canada comes to mind) and +- 1.2 hz for er well power deficit ones (best example would be India, sigh)
 
N

Namatimangan08

> Realizing that you probably work in a certain Asian sub-continental region of the world famous for lack of frequency
> control and poor grid frequency regulation you may indeed see more unusual circumstances than many readers
> here. Actually, that might be changing as this site seems to be attracting more inquiring minds from that part of the
> world looking for explanations to unusual operating conditions.

You are right. In our country frequency regulation is not excellently done. But that is not my point. We are talking about fundamental rather than specific method used in any country.

> I will maintain that a properly regulated--or "tuned" to use your terminology--will result in all
> synchronous generators operating in synchronism (in other words: at the same frequency/synchronous speed). You are
> now saying that a generator might be at 59.9 Hz versus another unit at 60.0 Hz (or 0.167%), when you previously said a
> machine could operate at 10% difference. It's very confusing.

My main argument is, the main operational objective is still keeping the frequencies for all parallel generators to be equal. That doens't mean it is okay to operate the generators even at 0.5RPM difference. This does not change the fact that their frequencies can deviate since the governing equation for each unit in the system is not the same.

Meanwhile the droops are tuned in such a manner that such deviation will never been allowed to go unchecked. Some of the times one or two units can go to reverse power, partial shutdown generator or even over speed protection operated. Those protections are standard protections for any any grid system in the world. That includes in your country.

> Just a few months ago I had the opportunity to visit a grid control headquarters for a very large grid which
> is well-operated and famously stable, and I inquired about generators operating at different frequencies on
> the grid. The looks I received were wilting, meaning I wilted under the stares.

You can stay in front of the grid monitoring panels for 20 years. No way you can see such deviation if don't want to see it. If you want to see it you have to install a system. We (my company) have installed one. We call the system Wide Area Monitoring System (WAMS). One of the parameters that we monitor is torque angle between inter inter connectors. Scanning time is as short as 500ms. I thought I have frequency data at one snapshot for various locations that showed they were not equal.

> I think I'll stick with my "problematic" and simplistic understanding of the basic fundamentals
> of AC power systems, and leave the transient, freakish off-frequency operating characteristics for others.

Let show us mathematically how two parallel generators can be "locked" into synchrorism the way you think that it works.

General rule that we know is something that we can put together can be teared apart. Even mechanical coupling between two generators can be cut into two pieces if enough torsion is applied. Size doesn't matter.

What is so special about the grid system under synchronism that only put together by the magnetic forces? What makes them cannot be separated at all?
 
Top