Could you please detail explain, why two or more engines (gen sets) work in Isochronous mode without load sharing will make the system unstable. Is it impossible that two engines (gen sets) in Isochronous mode without load sharing run and carry the different percentage of load but the speed (frequency) is stable?

 

Erwin,

It is possible to synchronize two or more generator-sets in isochronous mode without load sharing <i><b>if the governors are detuned to allow simultaneous operation.</b></i> HOWEVER, when this is done the frequency is not stable and it's very difficult to control which generator-set has what percentage of load--except by dynamically changing governor tuning parameters.

I highly suggest you find a text or reference book to understand the type of regulator used for isochronous speed control--a proportional-plus-integral regulator. It is the function of a PI regulator to not allow any--or very little--error. And, the setpoint for isochronous speed control (true isochronous speed control) is synchronous speed--the machine speed that corresponds to desired frequency. Any deviation from that, or anything other than a very small deviation, will result in a very fast response by the prime mover's governor using a PI regulator tuned to maintain a specific speed setpoint to return the speed to the setpoint.

If two machines are synchronized together in isochronous speed control, and the load is stable, it's very possible for them to stably control the frequency without too much "fighting." But, as soon as a light is switched on or off, or a motor is started or stopped, the frequency will start to change and when that happens both governors sense the change and immediately want to counter the speed change. When two machines do that at the same time, the results can be "dark"--meaning that what usually happens is that the loads of the two machines start swinging back and forth between the machines, wildly, until the under- or over-frequency relays or the reverse power relay of one of the machines actuates and trips that machine, and if the total load is higher than one machine can sustain, well, that machine is going to trip very soon (probably on overload or under-frequency, depending on the governor and generator protections being used on the machines).

Again, this is what happens when two generator-sets operating with their governors in (true) isochronous speed control are synchronized together. And, it's because of the type of regulator that's used for isochronous speed control that causes this to happen. There are governors which do not use true isochronous speed control (and use some other parameter(s) besides or in addition to speed) that can be operated simultaneously in "isochronous" mode--but those two machines usually have to have the same type of governor (make/manufacturer) in order to have similar operating characteristics and tuning parameters.

Now, as for being able to adjust the load of individual (true) isochronous speed control machines--it's not possible. The setpoint is speed, and speed will change with load. If the operator changes the speed setpoint of one isochronous machine (let's say it's increased) the other isochronous machine will say, "I don't want the speed to increase!" and will reduce its load--and keep reducing its load until the speed changes back to its setpoint (which wasn't changed). This will cause the first machine to load and load and load and the second machine to unload and unload and unload--until the very same scenario listed above occurs.

Can the generator-sets at your site be tuned to operate simultaneously in isochronous speed control mode when synchronized together? We can't know the answer to that question--only the generator-set manufacturer or service personnel can say for sure. Is it true isochronous speed control? Who knows. Is it adequate isochronous speed control? Maybe; until a large block of load suddenly is added to the system or is isolated from the system, and then--who knows what will happen.

PSM (Power System Management) schemes can be used to operate multiple machines in droop speed control mode when isolated from a grid and still maintain system frequency fairly well. Some are better than others.

But, isochronous operation of two (or more) generator-sets without some kind of isochronous load-sharing (or without some serious de-tuning of the governors) will result in unstable frequency, and usually, eventually, black-out.

The best analogy is a tandem bicycle with two riders trying to maintain a constant speed on a flat and smooth road while carrying packages. If some packages are added to the bicycle the effect will be to cause the bicycle speed to slow if the two riders do not increase the torque they are applying to the pedals of the crank-sets to maintain the speed. And, if both riders increase their torque then it's likely that the speed will increase above the desired value, and if both riders then decrease their torque at the same time then it's likely the speed will then decrease below the setpoint.

But, if one rider applies a fairly constant pressure (torque) to the pedals of his crank-set and the other rider--by agreement--varies his pressure as packages are added or removed from the bicycle then it will be easier to maintain the speed of the bicycle at the desired setpoint.

Generators and their prime movers are exactly the same. Most tandem bicycles have the two crank-sets chained together so that they spin at the same speed as torque is being applied to them--this is synchronism (synchronous speed). Neither rider can pedal faster or slower than the other rider--but if either rider increases his pressure (torque) or decreases his pressure (torque) while the load is not changing then the speed of the bicycle will increase or decrease, respectively, and the speed of the two crank-sets will also increase.

Two riders both trying to respond to changes in load (or even changes in road condition--slope, roughness, etc.) in order to maintain speed will usually result in unstable speed. Unless the two riders somehow work together and share the load, then the speed will not be very stable (if the load is changing).

That's a very simplistic analogy--but it's applicable all the same. If you want to know about the maths and formulas, get a textbook or reference. But, (true) isochronous speed control is about maintaining a speed setpoint (which is maintaining a frequency setpoint) as load varies. Two machines can't share load when they're both trying to respond very quickly to speed changes when load changes. And, there's no way to control the load carried by each machine in such a scenario--without some kind isochronous load-sharing scheme, or PSM, or some other higher-level control system/scheme.

Hope this helps!

 

If the governors are detuned to allow simultaneous operation.

Is that not the same as adding droop utilizing the governor?

 

CuriousOne,

Effectively--except it doesn't allow for changing the load setpoint very easily.

Detuning is really not an option for allowing load sharing, but I've seen it done with reasonable results on a couple of occasions--but the system frequency wasn't very stable. Not very stable and not very constant.

 

First of all, thanks for your response. Most helpful. But there are something that I am still not clear.

1. as you mention, "If the speed setpoint of one isochronous machine increased, the other isochronous machine will reduce its load and keep reducing its load until the speed changes back to its setpoint"
Does "reduce load" or "unload" mean "reduce fuel to engine"?

2. Is it possible that more engines(gen sets) paralleling with infinite Utility (grid) work all in true isochronous mode without load sharing and the speed (frequency) still stable to the desired speed (frequency)? In my opinion, it is possible. Because the frequency of the infinite Utility (grid) can not be changed. Am I right?

3. what about one engines (gen sets) in isochronous mode and the others in droop speed mode working together and carrying the common load? why it is stable in this way? the speed (frequency) will be stable to the isochronous one or to the droop one? how the common load will be shared between them?

Erwin

 

Not very stable and not very constant.

So we have broken the isoch machine in search for a machine that can be switched to droop.

 

Erwin,

1) Yes. Load is equivalent to fuel flow-rate (for internal combustion engines; steam for steam engines; wind for wind turbines; water for hydro turbines; etc.). More energy flowing into a prime mover (turbine, engine) driving a synchronous generator means more load that can be supplied (carried) by the generator-set. Reduce the energy flow-rate into a prime mover driving a synchronous generator and the load which can be supplied by the generator-set decreases.

2) No. And no. Your opinion is not correct. Many grids around the world post the real-time frequency of the grid on their websites. It's extremely interesting to watch. Even a well regulated grid has peaks and valleys around the nominal frequency, and depending on the parameters programmed into the generator-set's governor they may or many not respond to frequency changes at the same time or in the same manner. But, the same problem still exists for multiple units synchronized together while operating in Isochronous speed control mode: There's no way to control the load on those machines without some Isochronous load-sharing scheme.

Think about a very large grid--one with several gigawatts of load (and generation supplying that load). If you synchronize even a 250 MW gas turbine-generator to that grid and try to run it in Isochronous speed control and the frequency drops below the deadband of the PI regulator of the governor, it will increase its fuel flow-rate into the gas turbine as long as the frequency remains below the deadband value--and this one little 250 MW machine isn't going to increase the frequency--and speed--of all the other generator-sets connected to the grid by an amount sufficient to increase the frequency by very much. And if it did, then it would not be possible for the operator to control the load on that machine. And, if you synchronized several of these 250 MW gas turbines all operating in Isochronous speed control mode they would all be responding to the same frequency change and it would be impossible to limit their interactions and oscillations and they would quickly become unstable.

Only one generator-set's prime mover can be operating in Isochronous speed control mode at a time (without some kind of Isochronous load sharing scheme) to control frequency. And, the load on that machine is controllable by an operator--it will be what it will be because it will increase or decrease as the load on the system increases or decreases.

3) Have a read of this response to a similar thread on control.com (this topic has been discussed many times on control.com):

http://www.control.com/thread/1369738472#1370191185

 

Dear CSA,

Plz answer this. .

Suppose 4 steam turbines each of 250 MW capacity are operating in an area whose net customer load requirement is 1000 MW. So at final steady state load is equal to generation. but presently say the load is at the 500 MW level, so still 500 MW is to be added or restored.

Now if the above four 250 MW generators, presently at 125 MW loading, each is allowed to operate in isochronous governor control mode (i.e., proportional + integral control mode). and they have same speed setpoint being equal to 50 HZ at all load level (isochronous being a constant frequency line parallel to MW axis,that is all loads from 0 to 100 percent in droop speed curve). will there be any stability problem if a 30 MW load pick up (in steps) at some portion is attempted?

---->As far as i can think,each 4 governor will sense the same frequency deviation while responding to increase generation i.e., deviation = (present frequency minus same speed setpoint).

They will not be any fighting with each other on account of bringing system frequency to each units' own speed setting as the speed setting of all is fully identical. So one generator will never take the whole load and consequently other generators are never going to back out or unload in an effort to bring back frequency to their own setting as they have essentially same setting. and they will continue to load each other peacefully until 1000 MW rated load is achieved in the system. Is this thing possible or have i made any error somewhere?

Plz let me know.

 

They're gonna fight--and it will be ugly.

If you believe every governor is exactly identical and the actuator strokes are identical and the hydraulic flow-rates are identical and the steam flows are identical and the boiler combustion controls are identical and the condenser vacuums are identical and the cooling water flows through the condensers are identical, you're not right. Steam plants have a LOT of variability. Open the control valves too quickly and the pressure AND temperature drop. And you have to increase firing rate and feed water flow-rate as steam flow increases--then back off after load levels out.

But mostly because the governor just won't "share" the load when in Isochronous mode. Not without some kind of Isochronous load-sharing scheme. Which is essentially a de-tuned Isochronous mode P+I controller with external inputs and feedbacks.

Mate, you need some real world operating experience. And if you're in charge of the grid everybody better be carrying--a torch (flashlight)!

 

LOL, u are correct. May be they will really be carrying a torch. but i am just in the process of ensuring that situation never comes, and that i clear all my misunderstandings till i really become an operator/manager.

I have not indicated that all the variables u mentioned in ur first para have to be identical for the 4 units concerned. It is obviously possible that combustion controls-firing rate-flow controls are different. BUT i just want to ensure that at least the speed sensor senses the same speed deviation at all times throughout the dynamic response of the 4 units on account of maintaining their speed setpoint same. (assuming the power system frequency deviation due to a step load increase is similarly available at all unit terminals). Isn't that achievable? let me know.

Coz if its achievable, then even the units with different ramp rate limits will try to minimize the integral frequency error by increasing generation as per their own generation rates. The ramp rate limits are different apparently due to different time constants of boiler combustion control, turbine temperature control, or feedwater flow control. Suppose the faster unit 'A' among the four has quicker response than others 'B','C' and 'D', it increase its generation faster than them in time t1. and due to total generation increase (though with different rates), the net integral frequency error also decrease, and this sensing is available to other 3 units also. but even after time t1, the response of unit 'A' is faster, so the remaining integral error is also mitigated with more generation increase from 'A' and less from 'B', 'C', 'D'. Thus error continue to decrease to zero and steady state comes. If the response time is divided to n stages, then at every stage the plant responds accordingly to the decreased integral error out of the operation of previous step for a single load increase. The scope of fighting between the four in transient state is not apparent to me. i know i am missing something but cant figure it out!

If the integral gain of all controllers is different so that their reaction time is also different to frequency deviation, do u still reckon 'fighting' will take place?

Lastly, can u let me know some material or tell yourself about 'isochronous load sharing' or' detuned isochronous mode'. i have tried viewing 1 post here regarding the same topic, but there the asker wants to know about PMS operation in an island mode with three units setpoint calculation. and there the responses quickly digressed to PMS in island mode, leaving no actual data on isochronous load sharing topic.

Thanks in advance.

 

pikachoo99,

for all intents and purposes, they will all see the same decrease in speed--they are synchronized. The differences will be in the rate of change, and that is machine-specific. But, the way they will respond will be different--even if they all are the same turbine, same governor, same boiler, same condenser--everything. They just won't respond identically, and so there will be load swings.

I don't have any material about isochronous load sharing. I suggest you look at the Woodward Governor website--they have lots of manuals and white papers (documents) about various forms of generator control and load-sharing (isochronous and otherwise).

Best of luck on your studies!

 

> BUT i just want to ensure that at least the speed sensor senses the same speed deviation at all times throughout the dynamic
> response of the 4 units on account of maintaining their speed setpoint same. (assuming the power system frequency
> deviation due to a step load increase is similarly available at all unit terminals). Isn't that achievable? let me know.

Even with the same speed signal applied to each of the controls, there will be slight differences in the actual internal signals produced because of measurement error. With an integral response element, even a very very small error will be integrated to a significant level over time.

So if you have two isochronous machines in parallel, both running at an actual speed of 50.000 Hz, one may be measuring the speed as 50.001 Hz and the other as 50.002 Hz. Although both are reading a high frequency and will tend to cut back on generation, the one with the higher error will ramp power out faster and so behaviour of the two sets will differ.

 

for question 2):

in this way the droop for all gensets must be same? it is impossible!

>1) Yes. Load is equivalent to fuel flow-rate (for internal
>combustion engines; steam for steam engines; wind for wind
>turbines; water for hydro turbines; etc.). More energy
>flowing into a prime mover (turbine, engine) driving a
>synchronous generator means more load that can be supplied
>(carried) by the generator-set. Reduce the energy flow-rate
>into a prime mover driving a synchronous generator and the
>load which can be supplied by the generator-set decreases.
>
>2) No. And no. Your opinion is not correct. Many grids
>around the world post the real-time frequency of the grid on
>their websites. It's extremely interesting to watch. Even a
>well regulated grid has peaks and valleys around the nominal
>frequency, and depending on the parameters programmed into
>the generator-set's governor they may or many not respond to
>frequency changes at the same time or in the same manner.
>But, the same problem still exists for multiple units
>synchronized together while operating in Isochronous speed
>control mode: There's no way to control the load on those
>machines without some Isochronous load-sharing scheme.
>
>Think about a very large grid--one with several gigawatts of
>load (and generation supplying that load). If you
>synchronize even a 250 MW gas turbine-generator to that grid
>and try to run it in Isochronous speed control and the
>frequency drops below the deadband of the PI regulator of
>the governor, it will increase its fuel flow-rate into the
>gas turbine as long as the frequency remains below the
>deadband value--and this one little 250 MW machine isn't
>going to increase the frequency--and speed--of all the other
>generator-sets connected to the grid by an amount sufficient
>to increase the frequency by very much. And if it did, then
>it would not be possible for the operator to control the
>load on that machine. And, if you synchronized several of
>these 250 MW gas turbines all operating in Isochronous speed
>control mode they would all be responding to the same
>frequency change and it would be impossible to limit their
>interactions and oscillations and they would quickly become
>unstable.
>
>Only one generator-set's prime mover can be operating in
>Isochronous speed control mode at a time (without some kind
>of Isochronous load sharing scheme) to control frequency.
>And, the load on that machine is controllable by an
>operator--it will be what it will be because it will
>increase or decrease as the load on the system increases or
>decreases.
>
>3) Have a read of this response to a similar thread on
>control.com (this topic has been discussed many times on
>control.com):
>
>http://www.control.com/thread/1369738472#1370191185