Speed of Synchronous Generator


Thread Starter

Muhammad Salman

I have confusion as how increase in electrical load lead to decrease in generator rpm? And why if we increase the generator rpm (by injecting more fuel in gas turbines) increase the power?
The only way that an in increase in electrical load can decrease speed is if the load is greater than the generation. So, if the load is 20MW and the frequency is stable, if the load increases to 21 MW (actually any value above 20MW) and fuel is not increased then the speed (and the frequency will decrease).

Think about riding a bicycle along a flat road at a stable speed which is about three-fourths of your maximum speed and you need to maintain that steady speed for the duration of your ride. And you are approaching a hill. A hill that is long and fairly steep.

As you start up the hill, you increase the torque you are applying to the pedals to maintain your speed. But, the hill is long and you reach the limit of your ability to produce torque and maintain the speed and the bike begins to slow down.

A generator is a device for converting torque into amperes. Power is expressed by multiplying volts times amps (basically) and since generators generally run at a relatively constant voltage (usually no more than plus or minus 5 percent of rated terminal voltage), the power produced by a generator increases and decreases as the amperes that are flowing in the generator stator windings.

And, to increase the amperes it is necessary to increase the torque input into the generator from the prime mover (be it a reciprocating engine, a combustion turbine, a steam turbine, a wind turbine, a hydro turbine--any device that can produce torque). More fuel equals more torque; less fuel equals less torque. Just exactly like on the bicycle, in the example above and in the real world.

When a generator is paralleled with other generators on an alternating current grid, the load is the sum of all the lights and motors and computers and other electrical loads. 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.

And one of the most important aspects of any AC grid is its frequency (in most parts of the world).

The frequency of an AC machine is expressed by the formula:

F = (P * N) / 120

where F = Frequency, in hertz
P = the number of poles of the generator rotor
N = the speed of the generator rotor (in RPM)

When synchronous generators (alternators) are connected in parallel with each other on an AC grid, they are all operating at a speed that is directly proportional to the frequency of the AC grid. No generator can go faster or slower than the speed which is proportional to the frequency.

There are great magnetic forces at work in the synchronous generator to keep the rotating magnetic field of the generator in locked synchronism with the magnetic fields induced in the stator winding by the current flowing in the stator windings. These forces keep every generator's rotor locked in synchronous speed with the frequency of the grid with which they are connected.

The production of electricity is all about producing torque in one place, to be used in another. The torque is converted to amperes, transmitted over wires, and then reconverted into torque. So, generators are just supplying torque from where they are located to many different locations--anywhere wires can be run.

So, I would like to know how and when you experience a decrease in speed when the load increases. Are you speaking of the entire grid? Or just a small isolated generating "island" and it's load?
i think what you have provided is wonderful and very useful. but a question in my mind is that, like in synchronous motors, is it not possible that the speed remains constant and angle b/w the excitation voltage and terminal voltage change so that the net torque required is met by the the current due to change in angle. normally what happens in motor the reverse of that can be assumed to be true in generators. but in synchronous machines the case is coming out to be different. motor is constant speed, whereas in generator i think nothing is constant.

also a second doubt. if we look at a vector formula then :-
E = V + IZ

now in generator, when load inc IZ inc , also speed decreases so E decrease. due to armature reaction V also decrease(generally). how can all the 3 parameters change simultaneously ?
<b>If</b> there is no automatic governor on the prime mover driving the synchronous generator providing some kind of speed/load control for the prime mover and the generator rotor, <b>and if</b> there is no automatic control of the generator terminal voltage, <b>and if</b> the generator and prime mover are being operated independently of other generators and their prime movers, <b>then if</b> the load (real or reactive) changes speed can change and generator terminal voltage can change and armature reaction has an effect on speed and load and voltage.

In the real world where prime movers used to drive synchronous generators have automatic governors to control speed or load, and synchronous generator excitation systems have automatic regulators to control generator terminal voltage (or VAr setpoint or power factor setpoint), when loads change the governors respond appropriately to maintain speed, load, generator terminal voltage.

It would be disastrous to operate a prime mover and generator in any kind of application (supplying a small isolated load independent of other generators and prime movers, or synchronized to a grid with other prime movers and generators) without a governor that was capable of responding appropriately to load changes and without an exciter regulator capable of responding to VAr changes.

All of this counter emf stuff is wonderful in that it helps us to understand what's physically happening inside the generator and possibly how generators are designed and constructed, but it doesn't help us to understand how generators and their prime movers are operated in the real world, and how to operate them reliably and properly to supply power to loads at stable speed (frequency) and load and reactances.

It's very important to understand all of the components of equipment that produces electrical power. There must be a source of torque and there must be a source of excitation. The torque is provided by the prime mover (turbine, reciprocating engine, etc.), and the excitation is provided by the exciter regulator, sometimes called the AVR (Automatic Voltage Regulator). When there is a source of torque (provided by the prime mover) and voltage (provided by the exciter regulator and the rotating magnetic field of the generator), the generator can convert the torque into amps. The torque must generally be provided at a specific frequency, which is directly proportional to the speed the prime mover is spinning the generator rotor. The speed of the prime mover can be controlled by the prime mover governor.

Further, when a synchronous generator and its prime mover is operated in parallel with other synchronous generators and their prime movers, the speed of all of the generator rotors (and hence their prime movers if directly coupled to the generator rotors) is fixed by the frequency of the grid. If the grid frequency goes up, the speed of all the generator rotors goes up at the same time. Conversely, if the grid frequency goes down, the speed of all the generator rotors goes down at the same time. It is the job of the grid/system operators to control the amount of generation so that it exactly matches the load on the system so that the frequency remains relatively constant; failure to do so can result in unstable grid frequencies.

And if the grid operators don't account for the reactive "loads" of the grid, then the grid voltage can also be unstable or too high or too low.

Everything is related, but it's really important to consider all the conditions of operation, and many times textbooks and references don't properly state all the conditions when they're explaining these very important principles.

So, yes, in the absence of any kind of "response" to changes in load then speed and voltage and counter emf can all change. But, that just doesn't normally happen in the real world of electrical power generating equipment.
While agreeing with all that was said above, I am not sure whether Salman's situation was actually referring to situations where you have an isolated and very finite grid. Isolated grids have an added limitation in that keeping exactly 50Hz (or 60Hz for all that matters) is somewhat difficult, or puts too much demand on controlling/governing systems. In such environments it is normal to accept some deviation from the nominal frequency.

Given this it is observed that the actual load is dependent on the frequency at that time. One must keep in mind that certain loads, especially motor driven loads, are speed/frequency dependent, and so if the frequency falls, their power demand falls also. Water pumps are a case in point.

In such situations one observes that the system manages to stabilise at a different frequency with minimal control system intervention, eg. 49.95hz for 50Hz systems. Obviously to bring the frequency back to 50Hz, one needs to increase the power input to the generator prime mover, thus ending up increasing the speed/frequency of the system.

And this I believe this is what Salman was commenting upon when he said that if they increase generator rpm they increase the power of the generator.
Thanks a lot. A large part of my doubt is clear now. if you know any good reference book do tell me. i want the effect of load on an ISOLATED alternator, without keeping anything constant, i.e. just an isolator supplying a variable load. And thanks again for the information.

Frankly, I didn't really know how to respond to Muhammad Salman's post, but made an effort anyway. His summary asked about decreasing speed with increasing speed, and his post talked about decreasing speed with increasing load, and increasing load by increasing speed.

Would you agree that if a synchronous generator is being operated in parallel with other generators that an increase in load (on the system!) would generally have little or no effect on a properly operated system (grid)? A change of 0.001 Hz is virtually imperceptible on the operator interface of most prime movers.

I really wonder if some operators actually watch prime mover speed when they are "changing load". I believe the assumption is that as fuel is increased during starting and acceleration, that when load is "increased" after the unit is synchronized to the grid with other generators that speed will change with "load" changes. In other words, that increasing the fuel flow when the unit is synchronized will have the same affect as increasing fuel flow when it's not: that speed will change.

All they are really doing, when they are connected to a grid with other generators, is changing the amount of load their generator is providing. They aren't changing the load of the grid. And, the speed of their unit usually doesn't change by a perceptible amount, or an amount that is proportional to change in fuel flow that would be experienced if the unit was not synchronized with other generators.

The point of reference of most of the questions regarding speed and load and counter emf of synchronous generators is not clear at all.

As Muhammad Salman has not responded, we don't know if his question was answered or not.

Your reasoning is perfectly right, as long as the size of the generators is small with respect to the total grid capacity. Continental systems are huge, so a 1250MW generator will only supply a very small percentage of the total demand. Changing its output will have a minimal effect on the system frequency.

However in small island systems, e.g. where a 30MW generator supplies 10% of the total system demand, a change in generator output will have a big effect on the system frequency. In such situations, one observes that to maintain system frequency, the prime mover governors are constantly acting, due to the constant corrective action of the droop control. Droop settings of 4% to 5% are used to create some system stability, obviously at the expense of not maintaining the system frequency exactly 50Hz (or 60Hz as the case may be). The load manager then corrects the generators' setpoints to maintain the frequency close to nominal, and this is done at a much slower rate.

Hope this clarifies my point.

I have done some shipboard generator control work and troubleshooting, and the "islands" don't get much smaller than that. (Though some naval vessels are bigger than some countries I've worked in!) And, in all cases, even when dockside and loading and unloading with large shipboard cranes, the emphasis was always on frequency control, keeping the frequency as close to the setpoint as possible. This takes some good operational skills, and some good governor tuning, as well as good communications between the deck and engineering crews. The Chief Engineers would get really upset if the frequency wasn't controlled stably, and the Chief Mates would get upset if the cranes became sluggish and unresponsive. It's a learning experience for all, and some learnt the lessons better than others.

Your point about smaller grids is well taken, and I do presume that people consider the relative size of generators compared to the grid when taking into account the effects of adding or removing generation. Most of the smaller grids I've worked on used 20-, 25- and 40 MW gas turbines for main generation, with some smaller steam turbines (less than 15 MW, if that large). A few locales had multiple 70- or 110 MW gas turbines, but those were larger peninsular grids with more load and generation than smaller regional grids.

Thanks for your help with these posts; your clarity is most welcome!

I wish I could understand all of these esoteric questions about emf and counter emf. I recall when I was in university how confusing the texts were about this versus our laboratory exercises where we had to learn about frequency control, isochronous and droop speed control, and paralleling/synchronizing. I just consider all of that emf and counter emf stuff as transmission method "background" stuff; it's nice to know but it's not ever measured in the day-to-day operation of power generation equipment.

At any rate, we shall continue to endeavor to persevere, shall we not?
Dear CSA,

Your explanation is good. However there is a grey area wherein you talk about the additional torque being converted to Amps. It is the only possible result when the speed of the synchronous generator is held constant and if you try to increase the torque. Can you try to explain the transfer of torque to amps.

Electrical machines (motors and generators) are devices for converting electrical power into mechanical power (motors) and converting mechanical power into electrical power (generators).

There is virtually no difference between a motor and a generator. In fact, in some applications the electrical machine is used as both a motor and a generator. The difference is whether or not torque is being produced by the electrical machine (in this case it's a motor), or whether the machine is converting torque into electrical power (in this case it's a generator).

That's what generators do: They convert torque into amps. Motors convert amps into torque. Wires connect motors to generators. So, in effect, the torque being produced by the prime mover (turbine; reciprocating engine; etc.) is supplying the torque that the motors connected to it are supplying to their loads--through the wires that connect the motors to the generators.

That's why we produce electricity: To easily transmit torque from one place to many places using wires.

There are mathematical formulae that could be used to explain this, but you can find those on many Web sites if you need them. (I note you didn't question how motors produce torque from amps. It's exactly the same phenomenon, just in reverse.)

Process Value

well once again i have to respectfully disagree/ slightly change what csa has said in his post regarding converting of torque into amps in a generator.

The total current in a generator is a vector addition of active current and reactive current , torque is only responsible for the active current not the reactive current. a more correct explanation would be that the mechanical torque applied to the shaft is converted to active electrical power.

in a steady state condition the mechanical torque applied by the turbine is equal to the backward electromagnetic torque produced by the generator. this is the reason that the machine is running at a constant speed. thus ignoring the losses the mechanical power is equal to the electrical power produced. if there is a change in the mechanical or the electrical load in the machine acceleration or deacceleration of the machine takes place. for example if the machine mechanical input is increased, the machine accelerates till a new steady sate is reached.

The originator of the thread asked why the machine speed decreases in case of a electrical load increase. This can be explained by the governor droop chara.

The droop chara of the machine is given as 4-5%. droop is defined as the percentage negative change in speed when the machine is loaded from no load to full load. a very simplified explanation is given below

under steady state condition Power mech = Power electrical power
if there is a small increase in the electrical power then
deacclerating power = electrical power - mechanical power

this difference causes the machine to deaccelerate and thus reduce the speed. if there is no governor and the mechanical power remains the same the speed does on decreasing. however when a governor is present , it senses a reduction in speed and increases the fuel input thus increasing the mechanical power , thus reduction in speed is limited to the droop percentage.

Synchronous machine chara.

Though the above machine is a simplification of what happens during a machine loading and unloading. The machine actual parameters can be visualised by the operating diagram of the machine. I am uploading a small picture here. it represents the machine operating region. here

E - Generator field voltage
V - Generator terminal Voltage
xd - Machine impedance
I - current output
delta - machine load angle
phi - power factor of the machine


The region OPQR represents the stable operating region of the generator. P represents the maximum power output from the machine. The curve QR the maximum

excitation of the generator. lines parallel to the X axis are constant power lines while circle (arcs) drawn from the center O represent constant excitation.

The Tip of the point E is the operating point of the generator. you can see in the diagram that the operating point E is at the intersection of the constant power line B and constant excitation circle B'. The equation is given by

E (angle delta) = V (angle zero) + I(angle phi)*Xd

Varying the excitation without changing the power

In the initial operating diagram we see that the operating point lies in the constant power line B. if no power input is increased the generator power output follows this straight line. now if the excitation is increases the magnitude of the E increases and shits to a new excitation arc C'. thus the new operating

point is E' which is a intersection of the constant power line B and the new excitation arc C'.The effect of such a operation is that the power factor of the machine reduces as shown in the figure.

Varying the power without changing the excitation

In the third diagram , we actually increase the power input to the turbine which increases the power output from the generator. now the power line shifts from B to C. but as the excitation is not varied it remains in the same curve B'. thus the new operating point is a intersection of the new power line C and the old excitation line B'. the effect of such a operation is that the machine load angle increases. The deaccleration power mentioned in the post is the one which increases the load angle thus facilitating the transfer of more active electrical power.

what i have presented is a idealised case , but in real operation both the power and excitation is varied thus the final operating point is a combination of the two.
Attempts are made to try to keep the explanation as simple and real-world as possible, and some assumptions have been made.

And we're talking about units with properly acting governors, in any mode (droop or isochronous), and we should be referring to stable grid frequency (unless otherwise stated).

In general when someone asks about loading a unit, they are not usually asking about changing the reactive component of the load. If they are asking about the reactive component of the load of an alternator they will usually use the term VAr or power factor, or, they will talk about "reactive power" (which as we all know is a no-no on this site).

In a real world situation where there is no VAr or Power Factor control in operation on the generator (alternator) exciter and when an operator "loads" a unit the excitation generally remains constant because it is manually operated. Until the loading is complete, at which time a duly vigilant operator will check the VAr meter or the power factor meter and then adjust the excitation to maintain the desired setpoint.

Without vectors and formulae and neglecting the reactive current component, which the originator did not ask about, the explanation given was valid. We have to be very careful here on control.com lest the 'Exclamation Pointer' chastise us for incorrectly referring to reactive "power" in our discussions, which makes this forum different from all the other discussions of alternator operation and loads in the world. (Aren't we lucky?)

It would be very interesting to know how much the speed (RPM) of the prime movers and generators (more correctly called alternators) at the site where ProcessValue works decreases when loaded (or increases when unloaded), either "electrical power" (amps) or reactive current. As used in this context, loading and unloading refers to increasing, or decreasing, respectively, the amount of torque being produced by the prime mover and transmitted to the alternator. And this question would be presuming a stable grid frequency which is at or very near rated frequency.

The answer should also include the time period the speed changes are observed to occur. Is it 0.00095 seconds? Or is is 9.995 seconds? Or, does the speed/frequency drop when some load is applied and remain there until an operator takes some action?

And, how would a prime mover and alternator behave if a unit were synchronized to the grid without a governor and then electrical load were "applied" to the unit, what would the speed do? (Without a governor there would be no droop- or isochronous speed control.)

In the real world, when alternators and their prime movers are operated in parallel with other units on a grid with a stable frequency (an "infinite" grid as some would say), when an alternator is "loaded" there is an imperceptible change in speed. Granted, there is an acceleration/deceleration and change in load angle (which is invisible to the eye, naked or not) but for all intents and purposes the speed of the unit, and the frequency of the grid, doesn't change as machines are loaded and unloaded unless there is an imbalance between generation and load.

In any case, we are not talking about percentages of speed, like 1% or 0.25% or 2.34%. We're talking about hundredths and tenths of RPM for split seconds and fractions of a Hz for split seconds.

So, ProcessValue, how much does the speed of the units at your site change when they are loaded and unloaded and for how long does this speed change exist? The question is regarding increasing (or decreasing) the amount of torque being provided by the prime mover to the alternator. You can state if reactive current is constant or not, but the question should be viewed in the context of stable grid frequency when the units are operated in parallel with the grid. And, you should also tell us by how much the speed/frequency changes and for how long if the units are being operated independent of the grid ("island mode").

Anxiously awaiting your response--with or without vectors or formulae.

Inquiring minds want to know what kinds of speed changes we're talking about and for how long these speed changes last on real world units with properly acting governors, in either droop- or isochronous speed control.

Process Value

CSA makes my life difficult by assigning my difficult tasks, lol ; just kidding. well this one is going to take me a little bit more time. I am taking 3 days off and will be going to the site on 3'rd. but this is what i am goint to do.

the question piqued my interest on how to simulate the loading and unloading of the machine that too by a significant load and sudden load. the normal raise / lower will not work here. the reference change too small to be recorded. to really feel / record the change the speed a minimum of 2 MW ( at present at a site with frame 5 machines) sudden loading and unloading should be there. i cannot go and ask for a sudden feeder charge or turn on a motor, but i can do this.

" Test procedure " ;)
The machines at the site feed to a section load of 14 - 16 MW and have a grid transformer back up.

a. parallel the machine to the grid. keep the machine under 2-4 MW export, ie keep the machine more then 2-4 MW more than section load so that the excess is exported through the grid transformer. Now open the grid transformer thus simulating the grid islanding, the machine will be suddenly unloaded by 2-4 MW.

probably what is going to happen is that there will be a speed oscillation and will settle according to the droop chara.

b. the reverse of the operation, ir keep the section under 2-4 MW import and open the grid transformer. this will simulate the sudden loading of the machine.

in both the above cases i will keep the VAR export/import as zero, thus in effect simulating a loading and unloading of a unity PF load :).

i will also run a trend recorder for DWATT anf DF with 40ms resolution, hopefully i will get the desired results with this.

CSA, if you have anything to add to the above test procedure, or want a remakeover of the procedure please tell me. i will consult with the operations team and try to do it. there is a TG under shut down at site and they are keeping all the units under import and continuous import from grid. hopefully they would have started the machine by Monday and i will be able to do the above " experiment ". but as i said i need a little time, but i will post my answers by jan 1st week.

I am also working on a simulation with ETAP, the models are coming on nicely i will post the results here. perhaps then we can compare the simulation and real world results.

And i nearly forgot, CSA, Wishing you a Happy and Prosperous NEW YEAR.
This test that ProcessValue has devised <b>does not</b> meet the requirements of the request posed to him.

It seems what needs clarification is the definition of "loading" and "unloading" and "electrical load" (or "load").

I will grant that <b>one</b> definition of "load", or "electrical load", is the amount of motors and lights and computers and other devices drawing power from a grid.

But, operators don't have control over the electrical load on the grid, only the "load" the prime mover(s) and alternator(s) they are operating can "assume" from the total load being supplied by all of the alternators and prime movers connected together on the grid.

So, <b>another</b> definition of "load" is the amount of the electrical load that is being produced by the prime mover and alternator, which is a part of the total electrical load of the grid with which it is being operated in parallel with other prime movers and alternators.

Operators don't have control of the numbers of motors and lights and computers, or the load on the motors. They can only control the amount of power their prime mover and alternator is contributing to support the total electrical load. Operators refer to the amount of the power their unit is providing to the system as "load". This is the "load" of their unit, which is only a portion (sometimes a very minute portion) of the total electrical "load" on a system to which is connected.

In other words, when an operator raises, or increases, the "load" on a prime mover and the alternator it is driving he/she <b>is not</b> changing the electrical load on the system--only the amount of electrical load being provided by the alternator and prime mover under his control. That's what the operator knows as "load"--the amount of power being produced by the unit under his/her control.

My definition of "loading" and "unloading" a prime mover and alternator being operated in parallel with other alternators is to increase the load being supplied to the electrical system by increasing the energy input to the prime mover, not by adding or subtracting electrical load from the electrical system to which the alternators and prime movers are connected.

Now, when the "load" of an individual unit (prime mover and alternator) on an electrical system increases while connected to a normal grid in parallel with multiple alternators and prime movers, <b>if the electrical load on the grid does not change</b> then what would tend to happen is that the grid frequency will tend to increase. (In this scenario, the total electrical load on the system (the numbers of motors and lights and computers) would need to increase by the same magnitude at the same rate in order for the grid frequency to remain absolutely constant. Or, ... read on....)

But, the diligent and proper operators of a grid will--either manually or automatically--unload another alternator and prime mover (or multiple alternators and prime movers) in order to keep the total generation equal to the total electrical load thereby keeping the grid frequency relatively stable. If they have a machine operating in Isochronous mode (or several operating in Isochronous Load Sharing, or some other kind of automatic frequency control scheme) the grid frequency will remain relatively stable.

In actuality, what happens on many grids is that the droop action of many of the prime mover governors will sense the change in speed and will therefore reduce their power output in order to maintain speed (presuming the units are not using GE's version of Pre-Selected Load Control!). In other words, when the differential between the turbine speed reference and the actual turbine speed changes the governor will counter that action to maintain the same differential.

The originator has never responded (not with the same "name" as originally used anyway) to clarify his open question. So, some assumptions were made, and hence the response that was given.

A subsequent poster (at least someone using a different name who did not identify himself as the original poster) asked about alternators operating independently of a grid and without any governor/control, which is not the original question.

And, still, I have yet to encounter a real prime mover and alternator being used to produce electricity that didn't have some kind of governor: electric, electronic, hydraulic, mechanical or some combination thereof. A prime mover driving an alternator supplying a variable electrical load without a governor is a laboratory experiment, not a revenue-producing machine.

So, if the turbines and alternators at ProcessValues's plants are loaded and unloaded by manipulating the amount of motors and lights and computers connected to them, then the test would be a valid one. These are the normal kinds of conditions that operators encounter--not the manipulation of electrical loads (numbers of motors and lights and computers) to change the load of the alternator. Unit operators don't increase and decrease the "load" of the unit (prime mover and alternator) by changing the energy being input to the prime mover--<b>NOT</b> by changing the number of motors and lights and computers connected to the system to which the unit is connected.

But, if the units at ProcessValue's plant are loaded and unloaded by increasing or decreasing the amount of fuel being admitted to the turbines, then test he has devised is not a valid test per the question.

A valid test would be to run a trend, while the unit is operating on a stable grid in parallel with other units at a stable frequency and to increase or decrease the fuel being admitted to the turbine and observe the change in speed while monitoring the "load" being produced by the alternator. We want to know how much the speed (frequency) changes when fuel is changed and the electrical load being produced by the alternator being driven by the turbine changes while being operated in parallel with other prime movers and alternators on a system with a stable frequency.

No more and no less. Just as during normal unit operations in every part of the world where multiple prime movers and alternators supply an electrical load consisting of motors and lights and computers.

Rahul P Sharma

<b>stable grid</b>

Process Value's site seems to be in India...... and I am not sure which state in India can claim to have a 'stable grid'..... :).... So I guess, the primary condition itself will not be met for the test.....

Sorry for diverting the thread with an insignificant input.....
That's one of your strong points, Rahul P. Sharma. We're still waiting for the frequency readings from the speed 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.

Process Value

CSA , the present site i am working on and also incidentally which happens to be my "parent site" is a refinery, and almost all the sites i have worked on are refineries (that is where they send me usually). here the electrical load manipulation works in many ways. you can view my site as a bottom up co-generation plant. the GT predominantly supply electrical power to the section load in the refinery but the GT's are also kept in parallel to the grid as there is a PPA with the state power operator. thus a min power export of 2-3 Mw is maintained. Inside the refinery the operators are the people who give clearance to the HT or large motor starting. Thus in effect the plant operators have certain degree of control to the electrical load in the refinery. and also they need to maintain the export so they also have control over the prime mover to increase or decrease the load when under parallel to the grid. so in effect both the operations specified by you are done.so i have went ahead and done a battery of tests. There are certain restrictions in the operations and certain limitations in the test setup (i have explained them in part 2 of the explanation), but i have done as much as i can , which included cajoling , pleading , begging :p plant supervisors . lol .

One of the main reason why i said devised the test about sudden loading and unloading is because , it is only then we will able able to get an adequate hunting in the generator which will hopefully give the disparity between the turbine frequency and the grid frequency. for a normal loading and unloading which is done at a constant and controlled rate it will not be possible to "see" or get this experimental data. the machines in refineries are small, they will have no effect on the grid whatsoever , imagine trying a auto loading of 2-3 Mw on a 22 MW machine in a 48000 MW grid. and CSA in your earlier post you have mentioned on the frequency change in a independent machine also. the frequency change in the independent machine can be done in two ways . one is by fuel change and other is by load change. my experiment was specifically aimed at that. any ways i am presenting the data in three parts go through them and tell me if you are satisfied ( yes , yes i know you are a hard man to impress :p ) , and also if any more clarifications are needed in the explanations given.

I am spacing my answer into three parts

Part - 1 : The theory - this section will explain with vector diagrams the response of the machine in parallel and independent condition.

Part - 2 : The experiment - i did a few experiments to compliment the theory part , this is to give a real world example of the theoretical conditions explained. and also as asked by CSA the experimental data for his quiries. This section contains trends , and trend data in csv files for readers here to take and analyze.

Part - 3 : The Grid - There seems to be a lot of talk about how the grid operates. I am explaining my view on it. I am also uploading a document on the operation guidelines and philosophy of the southern gird (India). people who are interested on grid operation can read it as it is an excellent material on grid operation in india and also in general.

Process Value

Part - 1 : the Theory

The machine acts differently for different operating conditions. Yes i am a fan of vector diagrams and i am using them again :)

section A - Machine in parallel to the grid

I am assuming that the machine is put in droop mode of operation. in the vector diagram there are two main vectors. The vector V which is the terminal voltage while the vector E is the generated voltage in the alternator. The vector V in effect represents the grid and the vector E the machine. Both these vectors rotate ; V to the grid frequency and E to the machine frequency , and the normal convention is to that they rotate in the anticlockwise direction. when a machine is synchronised to the grid both the vectors E and V rotate in the same frequency , thus the relative speed between the two is zero. This is reason why we can draw a stable vector diagram. The angle between the two vectors , the load angle determines the power output of the machine. when connected to the grid the machine and the grid frequency will remain the same under steady state condition. Thus the hallmark of a grid under steady state condition in stable frequency.

Now two separate conditions can happen to the machine connected to the grid.

A.1 change in the grid frequency - when the grid frequency changes , the vector V accelerates twords the vector E and then finds a steady state position. Thus this causes a reduction in the load angle and thus the power supplied by the machine to the grid reduced. the reverse happens for a reduction in the grid frequency. Here the change in power output with the change in frequency follows the droop reference of the machine.

A.2 Change input to the machine - when fuel/steam input to the machine is increased the machine accelerates ie in the vector diagram the vector V remains the same but the vector E accelerates away from the vector V. thus the load angle in the machine increases and thus the power delivered to the grid increases. the reverse happens in case of a fuel/steam reduction to the turbine/gen set.

I am uploading a diagram here which explains the above scenarios with VECTOR diagrams. I have also induced the corresponding droop Chara diagram also.


Section B - Machine independent , and supplying to a section load

The explanation of a independent machine is a little difficult as there is no fixed reference point of observation. Machine and bus load angles are always expressed from a reference bus. i am taking a intermediate generation bus as my reference. The machine is put in droop mode of operation. The vector V represents the section load terminal voltage , while the vector E the generation voltage. The vector V is controlled by the load while the vector E is controlled by the generator. both vectors are moving in the anticlockwise direction with the the speed equal to the speed of the generator. The relative difference between the speeds is zero for a steady state condition. The angle between them , the load angle determines the power transfer.

Now two separate conditions can happen to the machine when it is independent

B.1 - Change in the section load - when there is increase in the section load the vector V decelerates and reaches a new steady state position. this deceleration will cause the increase in the load angle which will cause the increased power transfer. the machine speed also drops in accordance with the droop chara/ref. the reverse happens for a decrease in the section load.

B.2 - Change in the input to the turbine - when fuel/steam to the machine is increases when it is operating independently with a constant section load ( in real life cases this is not the same as most of the loads in a system are motors and an increase in the frequency will increase the power output from the motors) both the vectors E and V accelerate together so as to keep the load angle a constant. This increase in speed is due to the shift in the droop ref chara. the reverse happens for a reduction in the fuel/steam input. However the increase in the fuel is minimal as assuming a constant section load the increase in fuel is just to compensate the rotational losses for a higher speed.

I am uploading a diagram here which explains the above scenarios with VECTOR diagrams. I have also included the corresponding droop chara diagram also.