Hi,

What does the mean of Grid failure and what are the steps required to do after knowing grid failure at Coal thermal power plant? Please help me to get knowledge on this matter. Your replies would be appreciated.

 

kk1989,

Ah, grid failure. A topic close to my heart.

What you're probably referring to is when the breaker that connects the coal thermal power plant to the grid opens("trips"), resulting in an inability to export power to the grid. There can be MANY reasons why the breaker connecting the plant to the grid would be opened ("tripped"). And before that breaker is closed again, the plant personnel must be satisfied they understand why the breaker tripped (opened) and that the cause is resolved. A breaker connecting a coal thermal power plant to a grid can be tripped (opened) by plant protective relays or grid protective relays--however, plant personnel must be satisfied that any cause has been investigated, understood and resolved, whether the plant-or the grid protective relays operated to open (trip) the breaker. This is VERY important: knowing and understanding the cause of the "grid failure" (the opening of the breaker connecting the olant to the grid).

Many times the opening (tripped) of the breaker connecting the plant to the grid will result in a blackout of the power plant, but in some cases the steam turbine may continue to provide power to the plant. So, the steps to take when a "grid failure" occurs differ depending on the power plant and how it is configured. It could be as simple as re-synchronizing to the grid across the breaker that opened and separated the plant from the grid. Or, it could as complicated as starting the entire plant.

Again, this is probably the most "grid failure" but unfortunately people use the term to describe many problems--such as severe grid frequency excursions resulting in breakers openin to protect the generators and/or their prime movers (turbines; engine; etc.).

Hope this helps!

 

So, what are the primary steps to do after grid failure? is it shutting down unit? or, anything else?

 

kk1989,

Depends on the configuration of the plant. If the "grid failure" you're talking about is one where the breaker connecting the plant to the grid opens for some reason:

1) Investigate, understand, and if necessary resolve, the condition that caused the breaker to trip

2) Resynchronize to the grid when possible

If the plant lost power when the breaker opened (and lots of plants will), then there's going to be a LOT of people running around trying to get things ready to re-light the boilers and basically re-start the plant once the power from the grid is restored.

We're talking about a lot of steps that are particular to the situation (whatever the "grid failure" is or was) and what happens to the plant as a result of that grid failure. Most large power plants rely on grid power for most of their auxiliaries and pumps and motors and such. No grid power; no lights in the plant--except for those powered by batteries. Even if the plant has an emergency diesel generator, or a gas turbine-generator, it's going to take some time to get them started and producing power. In the meantime, there are going to be lots of safeties (safety valves blowing steam to atmosphere--lots of steam) lifting. A plant will (should) have a written procedure for steps to take in the event of such an emergency.

It's the rare thermal power plant that can withstand being separated from the grid by supplying its own power, running in what's called "island" mode. Usually, the plant load is so low that it's difficult to keep the boiler fired on coal, so it's necessary to switch back to oil or natural gas (whatever's used to start the plant). And that takes time--and it doesn't always go very smoothly. Again, even if there are emergency generators (diesel engine- or gas turbine driven) they take some time to start producing power).

And, no two power plants are alike. I was just working at one plant that is supposedly a "twin" of another plant on the other side of the valley and when I went to the other plant--there are some pretty significant differences. Over the years (the plants have been running for almost 60 years) the two staffs have diverged in their maintenance and upgrade and replacement activities and philosophies resulting in some very different operating procedures and equipment--even though the basic plants look identical from the street. Both of them have different SOPs (Standard Operating Procedures) based on the differences, and, even though the plants were built at the same time by the same contractor using the same design and equipment the way they are connected to the grid was different at the time they were built. So, even when you have two seemingly "identical" plants, they ain't always exactly identical.

I've provided about as much information as one can provide for such a general question. Again, everyone's definition of "grid failure" can be different--you haven't even told us what YOUR definition of "grid failure" is. So, the answer is going to be just as general as the question--and every coal thermal power plant is not the same as every other coal thermal power plant. I've been at coal-fired power plants as small as 10 MW, and as large as 2400 MW, and, believe me--they are VERY different, even though they were only 5 km apart. The way each plant would respond to the same scenario would be VERY different, similar, but different. Just as described above. (Both of those plants, by the way, would go "black" if the breaker(s) connecting them to the grid opened. The small one only had one breaker connecting the plant to the grid; the large one had several, and two of them were designed to only open under extreme circumstances--because losing AC power to the plant when it was running at 2400 MW would be not only loud, but ugly, and dangerous. That much steam with no place to go would be very loud, Very, VERY, <b>VERY</b> loud.

 

Dear CSA,

i have some questions regarding this issue.

Suppose the "Power grid failure" leads to complete shutdown of many plants leading to a case of complete blackout. Now we have to restart many plants. Some steam plants can be hot restarted. Some take like 6 hours to resynchronize. Hydro plants take few minutes.

1.What makes the the process so different for hydro and steam startup? Also i heard some steam plants can be hot restarted within 30 to 45 min after shutdown. Why is this variation from 45 mins in hot starts (for drum type boiler) to 6 hours (in once through boiler steam plants)?

Also plz tell how the approximate time to startup vary with respect to plant size, type of boiler, type of fuel used (oil firing and coal firing)

2.After synchronizing,during restoration, do all units placed in manual control or some units in agc and some units in governor droop mode with manual setpoint changing? or all units are placed under fully automated central agc control?

3.After restarting, What percentage of total output is the minimum load of a unit and with what rapidity should it reach the minimum load to avoid undesired heating or other effects?

 

pikachoo99,

Your question makes you sound like many inexperienced power plant managers and -owners. "Why does it take so long to start my generator?!"

The reason for the question is that in today's world of renewable energy, thermal power plants can make LOTS of money if they can start and get synchronized to the grid very quickly.

But, most prime movers weren't built to be started in ten minutes or less--not without seriously degrading the life of the equipment which means maintenance intervals are greatly reduced which means more costs for parts and labor and less generation (because the equipment is being repaired more frequently).

Think about the difference between a hydro turbine power plant and a steam turbine power plant. And then think about the differences between a small steam plant and a very large steam plant. Hydro turbine power plants operate at a constant temperature whether they are starting or running they are just converting kinetic energy to torque to amperes--at a constant temperature. They don't have to be concerned about the effects of thermal cycling on the various parts of the plant.

Steam plants, on the other hand, come in all different sizes and operate at widely varying temperatures and pressures and burn widely different fuels to covert chemical energy into high temperature and -pressure steam which is then admitted to a turbine that has stationary and rotating parts with very small clearances. Raising the temperature of any boiler, piping and turbine is stressful for the metals and other materials involved. Doing so very quickly can lead to some very damaging consequences for one or more parts of the plant. Every component has a thermal stress limit and the overall plant start-up time is a function of the "weakest" component.

Additionally, steam being admitted to a steam turbine has to be "dry" and that means the piping between the boiler and turbine has to be warmed up to prevent condensing the steam into water which, if introduced into the steam turbine at high pressure and velocity would cause serious damage to the turbine.

Steam turbine rotors and shells (casings) also have different masses and grow (expand) at different rates. This must also be factored into the starting time so as not to wreck the steam turbine. Smaller steam turbines take less time than larger steam turbines.

There are also different types of gas (combustion) turbines each of which has very similar concerns for thermal cycling during starting and stopping.

Does this help you to analyze each type of power plant to recognize the differences in starting times? It's all very understandable when you take the time to recognize and understand the differences.

Of course, if all one thinks about is money, then all one need concern themselves with is how much will it cost to start this plant in one-third or one-quarter of its designed starting time. It can certainly be done--but it has a cost.

Power plants are not like automobiles which can be started and run up to highway speeds very quickly and have relatively long lives if maintained properly. Of course, if you drive like most teenage males or many women and keep the pedal to the floor at all times--either the accelerator pedal or the brake pedal--and don't perform proper maintenance then the automobile isn't going to last very long--or the repair costs are going to be extremely high, and the down times are going to be frequent and long, as well as costly. Automobiles have been designed for the kinds of service they are expected to encounter; operate them differently and they aren't going to last as long or they're going to require more maintenance.

Hope this helps!

 

Thank you CSA.

Your explanation does help me.

You mean the tolerance level of different steam plants to rapid temperature and pressure raising of steam varies from plants to plants,depending on their sizes. i have read that in boiler drums, headers and in turbine steam chests there are set limits on differential temperature between inner and outer walls which must not be exceeded, and the rapidity with which these jobs can be done takes into account the stress development in these metal parts.

i am not an industry guy. i am just a student. and thus, am more interested in specific real data. I am confused between the claims made by many authors in their paper regarding restoration.

1. Many authors claim that some steam units with drum type boiler can be hot restarted within 30 mins to 45 mins. However they are awfully silent regarding the unit size. i think they are oil fired and size within 100-200 MW. Do u think my deduction practical? let me know.

This so called 'hot start' units have metal temp. 400 deg centigrade and some author claim that if they are not provided start up power within 45 mins,they may need some cooling period and increased shaft rolling up to 4 hours. thus having a cold restart, to reduce shaft vibration and eccentricity. They mean within 45 min rotor metal temp will decrease much below 400 degree and thus matching with steam temp. will be difficult. What confuses me is that some operator previously raising steam temp will now have to stop warming and cool just due to exceeding some interval. Do u find these matter of fact?

Other base drum type units or super critical once through units can only be cold restarted after 5-6 hours from trip out. I know for restarting a unit, one has to lightup boiler, keep the turbine in turning gear. then slowly increase steam temperature-pressure. then roll the turbine up to rated speed. then synchronize.

what i want to know from u specifically is what is the approx size (in MW), fuel type of these units having this start up time variation? i just want some real datas. Also can u refer handbooks/material carrying this infos.

2.In general i studied that power plant personnel consider 'hot start' if a unit after trip out can start within 12 hrs. their turbine temp don't reduce much and less problem in steam-metal temp. matching. 'cold start' is up to 72 hours. whereas in restoration terminology they call hot start as those drum boiler units starting within 45 mins and other starting even within 6 hrs as cold restart. why this difference in definition? i fail to understand. Moreover, My guide (being an industry man) is of the sound view that minimum starting time shouldn't be less than 1.5 hrs for any unit. I just don't know whom to believe

3.Do you know what does 'boxing up' of boiler means and how it helps in quick boiler firing?

And lastly i would request others to respond to point 2 and 3 of my previous post in this thread (regarding generation control after synchronizing and minimum load and time specifications), and also request CSA to respond to the same, which he may have missed.

Thank you once again for taking the time to answer my query.

 

pikachoo99,

The answers to your Points 2 & -3 would almost require a book, or at least a small pamphlet. It would require a lot of statements and definitions and explanations--all about Droop- and Isochronous Speed Control.

Let's take a grid of, say 500 GW (that's Gigawatts). And let's say that a 200 MW machine running at 200 MW tripped off the grid because of a faulty protective relay. The condition was resolved and the unit was restarted and resynchronized to the grid. It was most likely operating in Droop Speed Control mode before it tripped, and it will most likely be operating in Droop Speed Control mode once it is re-synchronized to the grid. That means that it's load can be whatever the operator or grid regulator wants it to be--as long as the prime mover can operate in the range it's desired to be operated.

When the unit tripped off-line--meaning that the total generation was reduced by 200 MW--the immediate effect on the grid, because there was no change in the total load (the number of electric motors, lights, TV, computers and computer monitors)--was for the grid frequency to decrease. Which means that EVERY generator (and its prime mover) saw it's frequency--and speed--decrease all at the same time and by the same amount. Now, when the grid regulators saw this happen their first action was to increase the output(s) of one (or more) generator(s) by a total of 200 MW. This will serve to bring the grid frequency back to rated. This can ONLY be done on units which were NOT operating at their maximum output rating. 200 MW of generating capacity was lost, resulting in a decrease in frequency (you can calculate the amount by dividing the lost generation by the total load), and so to restore the grid frequency to rated the load(s) of one (or more) generators had to be increased by an amount equal to that which was lost.

The way that individual loads are controlled on machines is done through Droop Speed Control. This is the prime mover governor method that allows multiple machines synchronized together to act as one generator to "share" load without any oscillation--in other words, to "play nicely" with each other, to stably contribute to the total generation capacity of the grid. When the total generation matches the total load at the rated frequency, everything--and everyone--is happy. And, when multiple machines and their prime movers can act together as one they are being operated in Droop Speed Control mode, which allows their individual loads to be increased or decreased smoothly and with negligible effects on grid stability and frequency.

Now, when someone synchronized their machine to a grid which is operating at rated frequency and the load (the total number of electric motors, lights, TVs, computers and computer monitors) is stable, then as they increase the share of load they are producing what will happen--the immediate effect on the grid--is for the grid frequency to start to increase. But, grid regulators/operators see this happening and they have to reduce the load(s) of one or more generators and their prime movers by an equal amount in order for the grid frequency to return to and remain at rated.

It's all a big balancing act--and all of the machines (the generators and their prime movers) are all synchronized together, operating at the same frequency, and are all operating in Droop Speed Control mode, which allows the loads of individual machines to be increased, decreased, or held steady, as required. It's all about Droop Speed Control.

Because we are talking about an AC (Alternating Current) system. And, all of the same principles are true, whether or not power plants tripped off line and are being re-synchronized to the grid, or whether plants are being started and synchronized to the grid, or shut down and taken off the grid.

Many threads on control.com talk about Droop Speed Control. And many use either a train or bicycle analogy to try to explain Droop Speed Control. AC electrical power systems are designed to operate at a specific frequency--which means a specific speed (synchronous speed). And AC electrical power systems are all about converting torque into amperes in a generator, and then converting the amperes back into torque or some other form of useful work (again, if you can call watching TV or YouTube videos useful or work). The electric motors doing the converting of amperes into torque aren't really doing the work of driving the pumps and refrigeration compressors or elevators--it's really the prime movers driving the generators that are doing the work. They are producing the torque, which is converted into a medium that can be transmitted long distances over wires, and then converted back into work at the other end of the wires.

It's just like using a train or bicycle to move freight and/or passengers from one point to another--at a constant speed. It takes torque to move the people and/or goods, and it also takes a certain amount of torque to maintain the desired speed. If you want to move a LOT of freight and/or passengers, you may need more than one locomotive engine, or if you're using bicycles you might need more than one person pedaling. The bicycle could be a tandem, with two riders, or, it could be many riders all hitched (mechanically connected) together to provide the torque to move the load. And if you want them all to move the load at a constant speed, when the load is changing, then there has to be some method of keeping them all working together and "playing nicely" together to keep the speed constant as the load changes. And, in an AC electrical system that means is Droop Speed Control.

The grid operators, could, when they see a 200 MW machine trip off line trip off 200 MW of load--people's homes and businesses and water pumping plants, etc. That would "balance" the loss of the 200 MW of load and keep the grid frequency stable. Until such time as the generating capacity could be increased by 200 MW which would allow them to restore power to those homes and businesses and water pumping plants. But, that's not practical or possible--and people would be very unhappy about that.

Droop Speed Control is the method by which multiple generators--and the prime movers driving them--can be controlled so that each generator and its prime mover can contribute to the total load and do so in stable and controllable fashion. It has other benefits and aspects as well--but that's its primary benefit.

Now if the loss of generation is very large--say multiple large generators--then it might be necessary to trip blocks of load off the grid, called "load shedding" to keep the other generators and their prime movers from becoming overloaded and the grid frequency from dropping too low. In this case, the grid regulators/operators have to balance adding generation--and load--back onto the grid so as to maintain frequency as best as possible.

Again, it's a very big balancing act--and there are so many variables that to try to explain every one would be impossible. There are maths to do all of this, but, I'm not one for using a lot of maths--because, to me, it's more important to understand what the maths will prove. You may wish I would use more maths, but that's not what I'm about here on control.com.

And, while one may be able to use maths to predict something or to prove it, it <b>DOES NOT</b> mean they can explain it to anyone so that they can understand it using only maths. I've heard, and read, a lot of people using using maths to try to explain Droop Speed Control and grid dynamics--and most of it is rubbish. Because they concentrate too much on the maths--and not enough on the causes and effects. The maths can prove and predict--but the maths don't explain. Not so most people can understand. Without also understanding a lot of maths. And I'm not about making people understand maths to understand concepts.

The unfavorable "thumbs-down" (dis)likes I received for the answers I've given in this thread--and other similar threads--are all given by an individual who would use maths for every terse explanation, without any cause and effect. Just the maths. So, he doesn't like my explanations, because I don't use maths. But, I find that most power plant operators, their supervisors, and their managers only understand enough maths to calculate their paychecks--and trying to use vectors and calculus and differential equations to explain what happens when they twist that knob or click on that target doesn't help them. So I don't use maths. Maybe, after they grasp the concepts I might, but I don't want to confuse people--I want to try to help them understand how power plants operate. And I don't need a lot of maths to do that.

AC power system dynamics are a very intricate and involved set of principles and involves a LOT of maths. Much more than I care to delve into here (or anywhere). If I thought understanding more of the maths would make be a better technician or a better teacher of power plant operators, their supervisors and their managers--I would study maths again. But, it's not going to do any of that for me, or for the people I work with or try to help here on control.com or the other World Wide Web-sites I contribute to. Concepts and causes and effects are what's needed, not a lot of maths. Maths, again, are useful predictors and proofs of causes and effects--but one needs to understand the principles in order for the maths to aid in the understanding.

I had a curmudgeonly electrical professor in university who tried to use maths to explain concepts and causes and effects--and almost none of us (his students) learned much or enjoyed his classes. We couldn't ask a question without being asked another question--and being referred to an equation or a vector diagram. It was maddening. And, it took me many years of reading and listening and operating to get the concepts and causes and effects I could have learned in a few semesters--but didn't, because maths were the only vehicle being used to teach. No explanations; just maths. And we were expected to "learn" from the maths.

Anyway, I digress. I challenge another contributor here at control.com, Mr. Phil Corso, to answer your Points 2 & -3. And, then I want to hear how you like his answers (though as a student you may like them more). Heck, send him your particulars (name; course of study; place of study) and ask for his white papers. His email address is: [email protected]. He also likes to take threads off-line, but that doesn't help the other people who read these threads and try to learn or enhance their understanding. It only helps the one person who's getting off-line help--which defeats the best thing about these World Wide Web forums: that MANY people can be helped and learn from the questions of one person.

And, once received and read those white papers, give us some feedback. We'd like to know. They have a lot of maths--you'll probably love them!

 

Immediately after a boiler-turbine combination trips, the internal temperatures will be within tolerance. While the set is shut down, the temperatures will gradually drift away from the running values. If the set can be restarted quickly, while the temperatures are within tolerance, there is no need to spend time warming up or matching differentials.

"Boxing up" is a procedure used to halt all flows through the equipment, so that the only drops in temperature are those due to losses through the insulation. With a boiler, the main thing is to shut off air flow as this will cool down the internal metalwork quickly and mean that a warm-up interval will be needed after a shorter shut-down interval.

Strangely enough, the time values don't seem to depend a lot on equipment size or rating. A larger boiler will have more metal and therefore a greater thermal capacity, but will also have higher flow rates. Thermal limits may depend quite critically on the detailed design of the machinery, especially for turbines - one with generally the same thickness throughout will be more able to handle higher temperature differences than one in which there are large changes of metal section.

 

Thanx to CSA and Bruce Durdle

To Bruce Durdle,I say this- I infer from ur post that boxing up is basically some way of closing boiler side gas doors so that the hot air can't escape. conserving this hot air would also mean conserving the metal temperature, and thus when boiler is refired after outage the problem of matching metal temperatures with increasing steam temperature is mitigated. right?

However, i have one concern here that sufficient boiler draught/draft is required for combustion or firing. i.e,entry of new air and passage of previous flue gases outside. If boiler side doors are boxed up or closed, don't u think boiler will have difficulty in refiring? or do u mean the previous air which has not been allowed to escape by boxing up is sufficient for firing the boiler?

To CSA, i say this-yes, u r right maths are not so handy in concept formation period in mind, but it does help to hone and verify one's concepts once we get some idea. so i respect both kinda guys. One who provides equations and one who provides not. lol! And i will surely contact Mr Phil Corso and let u know what he says. thank you!

Yes i know that effective load sharing is only possible via droop speed control. but we need integral resetting action in the load reference or speed changer setting to eliminate the small frequency drop (error), even after primary response of governors and this is called AGC response. Now i have learned about three kind of controls. Let us assume that there is sufficient unutilized running generation or spinning reserve of all units connected to grid for my following points:

A) Within an area, let 4 thermal units be running, three units are placed in droop control and one (the highest capacity unit) in isochronous control. so that after a step load increase, it can act to reset the frequency back to original after droop governors have given their primary responses. but as soon as the ISO unit tries to bring back the frequency to its original value (50 HZ), the droop units are given a manual resetting action in their speed changer/load reference setpoint so that they do not overload the ISO unit. So from my block diagram in mind, we have three (1/R) feedbacks to droop units with partly manual control. and one (D+1/R)*K/s) feedback along with a separate 1/R feedback for isochronous unit with no manual control, in our turbine governor model.

B) Considering the same area: In another mode of control, there is a central AGC controller which calculates just the required the integral resetting action for all four units ([D+1/Ri]*Ki/s) and send this signal via telemetry or pilot wires to speed changer motor of all four units. so that all four units after providing initial governor (1/Ri) response can automatically adjust their load reference point, and thus this scheme is free of any manual intervention.

{Obviously,this scheme is made more complex by calculation of base point and participation factors for each unit from economic allocation point of view. but i am not going into that as economic considerations are of secondary consideration while controlling generation during restoration period. we are talking about POWER GRID FAILURE}

C) third scheme is a very coarse scheme. this scheme is applied for limited time to some particular unit as per operator's decision and confidence where i heard the governor is changed from 'auto' switch to 'manual' switch. and even the basic primary torque increase is as per the whims of operator, leading to risky frequency regulation of power system. so no 1/R and no agc feedback in this case.

Now i have question in this regard. i am working on restoring an IEEE 40 bus system purely for academic purpose. There at a certain point during restoration, let us say there are 4 units all under AGC control. i.e all with governors and all having setpoint control by central AGC response from control room. Problem is one governor is having very quick frequency response than others, and lets say i pick up 30 MW at some bus. now this fast units almost takes 50% of the 30 MW i.e. 15 MW by itself and other 15 MW is taken by other three slower gen units. Due to this, the line connecting this particular generator with the load bus is overloaded exceeding line limit but generators are within their limit. don't worry. Now i know i can do load shedding, parallel line closing or generation shifting/rescheduling to offset the overload. but can i do this?->


1) Can i take the fastest unit in full manual control (The third control i mentioned above) for some time being and do not increase its torque despite frequency drop due to load pickup. i.e, operate it at constant generation level for the time being, and give the task of load and frequency regulation to the other less fast units as per their governor response and agc response to pick up the 30 MW fully. Assuming they all have sufficient spinning reserve and they can maintain frequency drop within 3% of grid set value (50 Hz)

or 2) can i increase the regulation coefficient of fastest unit so that its (1/R) is reduced so that it takes up less load out of 30 MW and thus do not overload its corresponding connecting line.

Now i request the viewers to tell me whether the above two control scheme is possible? If possible, which one is more preferable? Also i would request all to respond to these statements point wise (so that i can keep track they are responding to which point of mine). and also to point out if any points i raised in discussing the above three generation control scheme (A, B, C) is wrong or not complying with practical cases.

Thanks in advance!

 

Naturally, you need to introduce air for combustion when relighting, and this air will have a cooling effect on the setting. However, if there are no problems with lighting-off, the amount of cooling will be insignificant.

An excellent reference for boilers and their operation (somewhat dated) is "The Efficient Use of Fuel" put out by the UK Government in 1944. An update is available dating from the mid-1970's - "The Efficient Use of Energy". You might find these interesting.

 

pikachoo99,

It's not clear what the end-game is here, and it's not clear where you're getting your information--besides references and textbooks, and some hearsay. I can tell you that there are so many "schemes" out there and there are more perversions of these schemes that have just evolved from misunderstandings and lack of understanding and just plain necessity--having to keep the lights on. And, I've been to sites where a lot of research and maths were done to try to develop schemes which proved to be unworkable in the real world because there are always a LOT of intangibles which can rarely be accounted for in designs and calculations. In reality, the simplest schemes always work best, but there are some who will just not believe that more maths and more algorithms and more assumptions and suppositions will result in a "better" control scheme.

It's laudable that you are trying to learn as much as you can and that you are using as many resources as you can. Something that you're missing is that people use words and terms interchangeably and incorrectly. AGC means many things to many people, and "boxing up" isn't a literal term, not to mention we have never defined "grid failure."

The last two times I was involved in this kind of exchange it turned out there were a lot of undefined and undeclared variables that kept being introduced. It's very hard to hit a moving target, and I have too many of those on most every job I'm assigned. I contribute to control.com primarily to try to help people understand concepts and causes and effects, and also to try to provide examples of logical troubleshooting and analysis of real problems. These kinds of open-ended questions with literal theoretical references are more than I can devote any time to.

Bruce Durdle has indicated in previous posts that he has some experience with islanded systems, so his assistance will be very valuable.

My personal advice is to make sure you clearly and completely understand the basics. And then try to supplement your understanding with personal experience. And to remember that even though you are in a very technical industry there are lots of words and terms which are used to describe the same or similar processes or procedures that have little to do with what's actually taking place--except to the mind(s) of the person(s) who coined the terms. I enjoy trying to learn the origins of many of these terms--it really helps to understand them many time. Take the term "Droop" Speed Control.

Sometimes, it's just impossible to relate the terms to reality--now or in the future.

Best of luck!

 

thank u Bruce D. I will surely check that reference.

Dear CSA,

The end game is to maintain frequency in the power system and keeping all generators within their operational limits. I am talking about a single area so tie line flows are not of any concern here. and my references (as far as understanding generation control is concerned) are books like 'Power generation,operation control by Allen J Wood and wolleneberg' and' Power system stability and control' By Prabha Kundur and Olle I Elgerd. and many papers on AGC simulation. i have tried to make my understanding as plain as i can. I dunno what went wrong. My be u found some inconsistencies in my statements. Would u be kind enough to plz point out those inconsistencies? i am always grateful to learn.

 

> Problem is one governor is having very quick frequency response than others, and lets
> say i pick up 30 MW at some bus. now this fast units almost takes 50% of the 30 MW i.e.
> 15 MW by itself and other 15 MW is taken by other three slower gen units. Due to this, the
> line connecting this particular generator with the load bus is overloaded exceeding line
> limit but generators are within their limit. don't worry. Now i know i can do load shedding,
> parallel line closing or generation shifting/rescheduling to offset the overload. but can i do this?->

If a load increase is applied to a small system, the frequency will start to fall. The rate of fall depends on the amount of the load increase as a proportion of the total capacity,and the inertia of all the rotating plant (generators and motors) connected to the system. A useful value here is the inertia constant, which is the total kinetic energy of the system divided by the rated capacity of the generators. This is a time which can be interpreted as the time constant of the speed decay curve, assuming the power drain is proportional to speed (it's not). This inertia constant depends on the generator type but is typically between 4 and 10 seconds.

So your 4 generators are all going to see a ramp of speed from 50 Hz.How the individual governors react to this will depend on their internal time constants and damping, rather than the droop setting. The combined time effects will also determine the maximum drop in speed, before the system can recover under the influence of droop settings. In a large system I have worked on, containing a mixture of hydro and non-reheat steam turbine generation, the frequency fell to reach a minimum after about 2 seconds before increasing again to a steady value as determined from droop settings after 8-10 seconds.

There is no point in trying to make your isochronous machine react in a very short time - it probably can't anyway, and if the droop machines are still changing dynamically in response to the load increase you will have severe interactions between the units. I would suggest you use a time constant for the reset action on the isochronous machine that allows the other units to settle out before there is any significant action on its part. In the situation outlined above, an integral action time of around 15-20 seconds would probably be acceptable. If you try to tune the isochronous response too tightly, you will find that changes in system dynamics with connected load, total inertia, and probably machine operating point will make the whole thing a bit unstable - and if there is one thing that panel operators are very good at, it is switching controllers from auto to manual as soon as they think they see a hint of "instability" or cycling.

> 1) Can i take the fastest unit in full manual control (The third control i mentioned above)
> for some time being and do not increase its torque despite frequency drop due to load
> pickup. i.e, operate it at constant generation level for the time being, and give the task
> of load and frequency regulation to the other less fast units as per their governor response
> and agc response to pick up the 30 MW fully. Assuming they all have sufficient spinning reserve
> and they can maintain frequency drop within 3% of grid set value (50 Hz)

It shouldn't matter how fast the fastest unit responds - it is possible for a turbine to react to the initial upset and then recover without suffering too many ill-effects. I have seen a nominal 2.5 MW gas-turbine powered generator reach 3.5 MW as indicated on the panel meter for a second or so. In fact, the main problem we had with this machine was that the shear bolts on the coupling between the turbine and gearbox were routinely bent because of short-term transient load increases.

If you are interested in playing around with mathematical models, it is comparatively easy to set up a model of a small system such as you have described. You can do an energy balance on the total system with a single inertia equal to the sum of all the connected inertias, which will give you a speed transient. Then estimate the response of each of the individual prime mover - governor combinations (just use a single time constant for this initially) and use the results to give you a value for the "energy in" component of the balance. This will let you play with various combinations and identify options more realistically. You don't need a lot of high-powered maths and Excel or similar tools will be quite adequate.

Hope this gives you some more ideas!

Bruce.

 

That was of help.

BruceD,

Yes i know initial frequency drop depends upon total system inertia at present level and the load damping/frequency sensitive constant.

Just as u said, i designed a single area system with 6 generators having inertia equal to summation of total inertia. the frequency is unified and i observed the percentage of generation increase of each generator governor individually. i gave my isochronous machine sufficient time to react only after all the droop governors have stopped reacting to avoid instability and overshoot.

Regarding the last point of mine that u quoted and responded, u said, "it is possible for a turbine to react to the initial upset and then recover without suffering too many ill-effects." But sir, i just want to know whether that is possible or not practically. i.e., taking out a generator out of auto governor control for some time (with other 3 generators in governor auto control) and then again to place it back under governor control when the conditions are suitable? -whether it can be done? Since u hinted that all operators can quickly shift from manual to auto control, i infer that it is possible after all.

Also, In general i heard that a single unit is placed on iso control and others on droop. All units aren't made iso since speed settings of no two units can be similar. and thus, if done, then one unit (with higher setting) will take all the load and other generators (with less speed setting) will back out. Here can u plz look on the 2nd point i made on the 2nd post before this. where i tried to know that other than this "single iso and other units with droops with manual speed setpoint change of the unit on droop" scheme can we use a single central Agc scheme for all units where frequency deviation (with load pickup) is calculated centrally. and then suitable actuating signal depending upon individual generator unit characteristics is fed to all units (on Agc) via pilot wires to their speed changer motor, to eliminate frequency error after initial primary governor response. What i want to tell is that all units here are given the job to regulate frequency and load demand as per the characteristics, not a single iso unit. is that okay? I have always seen in books that this above scheme is always applied to "SELECTED UNITS ON AGC." what i wonder is that 'why not all units?' since the unstable phenomenon of matching of speed settings of more than two units will not pose a problem in the central agc scheme, assuming they all have sufficient spinning reserve during a particular step load pick up.

Lastly, sir, can u plz say something regarding minimum load requirements of a generator and how quickly it needs to be achieved after initial synchronization of alternator to grid? and what percentage of total load is this minimum load? if u can write this in the format: <UNIT CAPACITY--PERCENTAGE FOR MIN LOAD---MAX TIME TO ACHIEVE THIS LOAD> for some unit sizes, that will be cool 4 me to understand. As far as i know, the reasons for this minimum load achieving within a short time is to stabilize furnace flame and ensure cooling effect of steam in turbine exhaust stages. plz VERIFY that too.

Thanx in advance.

 

> But sir, i just want to know whether that is possible or not practically. i.e., taking out a generator
> out of auto governor control for some time (with other 3 generators in governor auto control) and then again to place
> it back under governor control when the conditions are suitable? -whether it can be done? Since u hinted that all
> operators can quickly shift from manual to auto control, i infer that it is possible after all.

It might be possible, but in general it is not wise to rely on operators to undertake this sort of action - the need for it will arise when the plant is in an upset state and their reaction times will vary widely, (and there is also a high chance that any such action will be too late or incorrect). It would be possible to design a feature into control logic that would stop load changes - in fact, in my first power station job, I was involved in analysing a device which was designed to put a dead band into the governor characteristic to stop the turbine responding to small frequency changes.

>What i want to tell is that all units here are
>given the job to regulate frequency and load demand as per
>the characteristics, not a single iso unit. is that okay? I
>have always seen in books that this above scheme is always
>applied to "SELECTED UNITS ON AGC." what i wonder is that
>'why not all units?' since the unstable phenomenon of
>matching of speed settings of more than two units will not
>pose a problem in the central agc scheme, assuming they all
>have sufficient spinning reserve during a particular step
>load pick up.

I'm sure you could devise a computer-based load sharing scheme where a single processor is used to calculate the governor signals for each unit. However, there is a good engineering principle - KISS (keep it simple) which makes the single Isochronous + 3 droops the most widely-adopted system. There are also reliability issues, and if a single computer is used to control 4 units there is a relatively high chance that all 4 units would be unavailable if the computer goes down - unless the existing scheme is used as a back-up. You would need to use the single controller to operate the governor set-points within the isoc + droop system.
>
>Lastly, sir, can u plz say something regarding minimum load
>requirements of a generator and how quickly it needs to be
>achieved after initial synchronization of alternator to
>grid? and what percentage of total load is this minimum
>load?

As a good rule of thumb, I have always recommended that the initial load on a generator should be between 10 and 20 % of unit rating. This is mainly to make sure that the machine doesn't trip on reverse power during the dynamic disturbance that can arise immediately after synchronising, but also helps to get the controls on other equipment such as boilers into a working range where they have a good automatic response. For instance, a boiler drum level control system may have a minimum downturn of about 6:1 and will need at least 16 % flow to give good control. So this will depend more on the nature of the prime mover and its associated controls than on unit size. This initial load should be applied on synchronising, and is the reason why the incoming machine needs to be running faster than the grid to which it is connected. On a small system such as the one you are describing, the connected machines will have to shed load to allow for the incoming one to pick it up, and this may also complicate things. In general, there are too many variables such as types of prime mover, machine characteristics, and the effect of connected equipment to make simplistic statements based on unit size alone.

Bruce.

 

Dear Bruce,

U have managed to clear most of my doubts. thank u so much. Can i plz know ur email id in case i face similar doubts in future?

Some last queries:
1) In case of single isoc + 3 droop units the speed changer setpoints/load reference setpoints of droop units are manually controlled ryt?

2) i know,the primary response of governor adjusts its main steam valve, so output is increased. i just wonder how raising of speed changer setpoint leads to a similar opening of steam valves in practical situations today. i have studied the mechanical hydraulic fluid governor model in my books where both flyball governor movement and speed changer raise/lowering lead to independent opening/closing steam valve with a linkage mechanism of rods and piston. but is that same mechanism in practice today in governing?

3) u are telling that the so called min load is 10-20% and that must be provided at moment of synchronizing in alternators to minimize reverse operation in low loads. now let us say the unit is of 500 MW (DURING RESTORATION) AND 10-20% OF THIS 50-100 MW. now let us suppose this load is not locally available. Say 30 MW is locally available and for remaining 50-70 MW, one has to charge a long distance line to search for that load in a far off bus. Now this line charging may lead to large MVAR production leading to overvoltage.---->>>>

Now wat do u think the operator should do?--

a) provide 30 MW locally and wait for a suitable time before providing 70MW by line energising on that far off bus, when system load has increased with help of other generators which makes the present system series reactive loss=line charging MVAR (I^2Xl=V^2wC),thus reducing overvoltage possibility.

b) or should the operator wait fully for other generators to pick up loads in surrounding far off buses which reduces the overvoltage possibility and then our present synchronized generator at once charge the required line and take up full 100 MW just at synchronizing.

4) i have heard ramp rates of hydro units are 50-100% of full capacity per minute. Is this anyhow dependent on unit size among hydros. like hydro units range from few KWs to 250 MW in my country. so it makes ramp rate of say few kws/minute to 125-250 MW/minute for different units. is that okay? is there no mechanical fatigue in hydros.

Also i heard steam drum type boiler units have ramp rates of 1-3% capacity per minute and supercritical once through boiler units have ramp rate of 4-7% capacity /per minute. is that right? If so can u plz comment on why is it so from constructional/operational features?

 

Hi pikachoo99,
You can contact me off-line at bmdurdle at clear dot net dot nz - but I'd rather keep things on the forum as other people can then benefit.

> 1) In case of single isoc + 3 droop units the speed changer setpoints/load reference setpoints
> of droop units are manually controlled ryt?

Usually - yes, but with modern systems it is relatively easy to automatically adjust the speed set-point.

> 2) i just wonder how raising of speed changer setpoint leads to a similar opening
> of steam valves in practical situations today.

Droop control is actually a simple proportional control system, with the corrective action being proportional to the difference between a speed signal and the "speed set-point". When synchronised to a grid, the speed can't change so adjustment of the "speed set-point" changes the error value which then feeds through to give increased power - when isolated, a change in the error will alter the speed as the power is set by the load. In a small isolated system such as you have described, load sharing means that both load and speed can change. (This will depend on how the load varies with speed, but if the load consists of induction motors driving fans, power is proportional to the cube of speed, so there can be a significant effect.)

The control scheme can be implemented in a number of different ways, but the overall operation is basically the same regardless of the type of hardware used. Modern electronic systems may have more bells and whistles, but are essentially the same in function.

> 3) u are telling that the so called min load is 10-20% and that must be provided at moment
> of synchronizing in alternators to minimize reverse operation in low loads. now let us say
> the unit is of 500 MW (DURING RESTORATION) AND 10-20% OF THIS 50-100 MW. now let us
> suppose this load is not locally available.
> ( SNIP)

This is a very complex situation, and really needs to be tested using a fairly accurate model of the system and all the possibilities. I can't give you a useful answer here.

> 4) i have heard ramp rates of hydro units are 50-100% of full capacity per minute. Is this
> anyhow dependent on unit size among hydros.

The loading rate on a hydro station is more often set by hydraulic factors. If the station is "run-of-river" with the turbines located at the bottom of a short penstock, there is only a small amount of water to accelerate or decelerate, and the loading time can be very quick. This type of station is sometimes used as "spinning reserve" where the generator is kept synchronised so the rotor does not have to be accelerated but with wicket gates closed: on a loss of other generation, the gates are opened and water is admitted as fast as possible. Such machines can load fully in a few seconds. On the other hand, if the water supply is via a long tunnel or a canal system, the rate of loading and unloading is limited by how quickly the water column can be accelerated or decelerated without getting problems from effects such as water hammer, caused by the change in pressure as the water changes speed - this could be several minutes to load fully. Governors on a hydro system will have a more complex arrangement, including what is referred to as "temporary droop" to limit the rate at which load can be picked up or dropped off. In control terms, this adds a deliberate adjustable lag to the governor response.

> Also i heard steam drum type boiler units have ramp rates of 1-3% capacity per minute and
> supercritical once through boiler units have ramp rate of 4-7% capacity /per minute. is
> that right? If so can u plz comment on why is it so from constructional/operational features?

With boilers, the main factor is the temperature gradient and other thermal effects. In a drum boiler, there may also be issues with the rate at which water can be added to the drum without causing problems. One of these is the possibility of "squashing" a drum by adding too much relatively cool feedwater - this reduces the temperature of the water and increases the density - so that adding water to the drum may actually cause the water level in drum to fall rapidly. This is not an issue with supercritical boilers, which don't have a drum.

Hope this helps,
Bruce.

 

BruceD,

One more query related to this response of yours.

Question.. In case of single isoc + 3 droop units the speed changer setpoints/load reference setpoints of droop units are manually controlled ryt?

Answer->Usually - yes, but with modern systems it is relatively easy to automatically adjust the speed set-point.

But that is precisely AGC again..isnt it? It means in addition to proportional control in the governor (1/R * change in Hz), we are devising an automatic method of producing a signal which is possibly an integral action (ki/S) to read just the speed setpoints of all units on droop governors.But this mechanism is also otherwise,done manually without AGC. (i.e with automatic proportional response but manual setpoint changing action to governors).

I just wonder, assuming what i inferred is correct, that if we operate a system where all 4 generators are under proportional control (by droop control) and their setpoints are automatically changed by 'this' auto (AGC) central response, IS THERE ANY REQUIREMENT OF A SEPARATE ISOCHRONOUS UNIT? I think not as if setpoints are adjusted after primary governor proportional response, depending on which their generations will further change a bit from their last values set by primary droop response, the job of frequency regulation is already done. leaving no requirement for an isochronous unit in a grid connected small system. i mentioned in my above posts assuming all units are within operational limits, i.e., between minimum and maximum active power limit. Am i wrong?

 

To me, the difference depends on the time scale involved. In general terms, the response of a power system to an upset has three separate phases. In the very short term - one to ten cycles - the conventional controls such as voltage regulator or governor do not respond and the behaviour is dependent only on the electrical parameters such as reactances. This is often referred to as the "transient" behaviour.

In the next phase, the control systems react to regulate the voltage or current. The extent of this time depends on the settings of these units, and for governors (as I have already said) will be determined by the prime movers and their characteristics. I refer to this as the "dynamic" response of the system.

Once all the transient and dynamic responses have settled down, the system will go into steady state mode. Loading etc will be set by adjustments to the controller set-points. These set-points can be adjusted manually by operators, according to some written guidelines, or adjusted automatically according to a specified program in a supervisory computer. As an example, one of the items of equipment I worked on was a GT which could operate at a fixed load, but more usually was in a mode where the generator power matched the incoming power - so the set produced about 50% of the plant electrical energy. In a larger installation, this approach can be used to maximise efficiency by loading the least expensive or fuel-hungry machines as much as possible, and trimming load on the most inefficient. Call this automatic load control or load scheduling if you like.

This is different from a system where the dynamic response of a number of sets is controlled by a single controller, trying to optimise the dynamic response so as to minimise frequency or voltage disturbances. This is what I would refer to as AGC. The individual machine control algorithms in this will have to be set based on machine dynamic response, which is not a concern with load scheduling - in fact, if the load scheduler reacts too quickly there is the possibility that the to systems will fight each other. The time scale for load scheduling must be greater than the time taken for the dynamic response to settle out.