Generator load Ramping rate

B

Thread Starter

BMP

We operate 2 units of Combustion Gas Generators GE-NP MS-5001-PA firing with Gasoil to supply electrical power for our Petrochemical complex. Total power requirement is 20 MW, our 2 CTGs both running in ISOCH mode, IGV control Auto ON with 10 MW load each CTG. When 1 CTG trip, the other CTG can load up till 18.4 MW within 0.95 second but the frequency drop to 48.05 HZ causing our plant Under frequency load shedding activated (UFS relay setting in 48.5Hz).

Do you any idea to speed up the ramping rate or to avoid the bus frequency drop below 48 Hz.

Thank you for your help to solve our problem.

Best regards,

Budi MP
[email protected]
 
I doubt we have all of the information about the configuration of the plant and control systems to be able to make any recommendation. If two units are operating in Isoch mode, then it's likely there is some kind of Isoch Load Sharing scheme in place, or, Isoch control has been "de-tuned" to allow the two governors to operate in parallel in Isoch mode (which is not a typical configuration). It may also be possible that the servo regulator has been "de-tuned" to make the two units operate in parallel in Isoch control.

And, it may be possible (probably likely) that some combination of all of the above was used to allow the two units to operate in parallel in Isoch mode. Again, this is presuming that the typical Speedtronic Isoch speed control scheme is used in the turbine control systems (governors).

It may even be possible that there is some kind of external load control sending raise/lower signals, either discrete or analog, to the governors. We just don't have enough information.

Isoch control really doesn't have a "loading rate" (at least in a typically configured Speedtronic panel). If properly tuned, it should respond very quickly to change fuel flow to keep speed and frequency constant, and this is not done (typically) with loading/unloading rates. It's usually just PI control.

Something we also don't know is the maximum power output of the turbines at the site. Is 18.4 MW at or above the maximum power (Base Load) for the units at the site in the ambient conditions at the site? If so, then one unit can't take all of the load on its own.

But, there are too many things we don't know. We also don't know if the plant was designed to be operated in this manner, or if plant operation has been changed recently to try to operate in this manner and what kind of studies were done to determine if this is possible and what would be required.
 
BMP... a possible method is to add inertia to retard the deceleration-rate.

Start with a Power-System Stability analysis using the "Stored Energy Constant" H, (MW-sec/MVA) of each machine. They are calculated from the Moment-of-Inertia, rated-speed, and rated MVA. If the two machines have different capacities, find the equivalent SEC, Heq.

Finally, if the H-constant is unavailable, you can certainly use representative data.

Regards, Phil Corso
 
Phil Corso,

How would you propose to add inertia to a combustion turbine-generator?

What form would the inertia take?

How would the inertia be attached to the shaft?

How would the addition of inertia affect the power required of the starting means to break the unit away from zero speed, and to accelerate it to self-sustaining speed?

How would the addition of inertia affect the self-sustaining speed?

How much would you estimate it would cost (in USD) to add sufficient inertia to achieve the stated goal, including any modification to the starting means?

Six (6) simple questions, Mr. Corso, to help us all understand the practicality of the proposal you are making.
 
==== quote

Actually our plant total power requirement is 29 MW. 20 WM supplied by our two CTGs (for essential and critical power load), and the rest 9 MW is supplied by Grid. This system is island mode without synchronized with grid due to unstable supply from Grid (16 times per year outages). Generator rated load is 21484 kW at pf 0.85,rated voltage 11000 V, 50 Hz and rated current 1327 A. FSR speed Proposional gain settled on 5%/sec. Liquid fuel used is Blended gasoil with 10380 BTU/lbs LHV. Our tested base load CTG #1 is 21.3 MW and CTG # 2 is 20.95MW (ambient condition temperature is 30 deg C). When in the incident our CTG trip coincidentally there was lightening strike. According your experience, Is there any influential lightening strike to the generator ramping load from 10 MW to 20 MW. Thank you for your help to solve our problem.
 
CSA... using the KISS principle here are the answers:

1) What form would the inertia take? Rotating mass!

2) How would the inertia be attached to the shaft? Details, Details!

3) How would the addition of inertia affect the power required of the starting means to break the unit away from zero speed, and to accelerate it to self-sustaining speed? More umph from the driver!

4) How would the addition of inertia affect the self-sustaining speed? Make it more stable!

5) How much would you estimate it would cost (in USD) to add sufficient inertia to achieve the stated goal, including any modification to the starting means? Some, maybe lots, especially if USD!

CSA... my idea, was to determine, first feasibility, then, practicality! It is a rather simple exercise!!

Of course, if you want detail, you know how to contact me!

Regards, Phil
 
Phil Corso,

Rotating mass. So, you're proposing adding some sort of flywheel to the shaft to solve this original poster's problem?
 
No; lightning won't affect loading/unloading rates.

And, the conditions of operation are completely unclear to me now after this most recent post.

I still maintain we don't have all the information, and that multiple, typically configured Speedtronic panels are not to be operated in parallel with each other when both panels are in Isoch mode. Only one Speedtronic panel should be in Isoch mode at any time when separated from the grid (island mode).

I have seen some special, unique sequencing provided to some certain sites which allowed Isoch control on multiple units, but we don't know precisely what is running in the Speedtronic panels at your site--nor what, if any, external load control signals might be connected to and driving the Speedtronic panels. And that sequencing does cause some small excursions (depending on tuning) from nominal frequency, but it does still respond very quickly to sudden load changes of just about any magnitude.

Without being able to see the electrical schematics for the plant and tie-line systems, and without being able to view the sequencing/application code in the Speedtronic panels at your site, we can't make much more of an educated guess at what might be wrong.

Except to reiterate that typical Isoch control doesn't used loading/unloading rates; it's just PI control using turbine speed (which is directly proportional to generator frequency) and which should be able to change load (on a typically configured Frame 5) much faster than you are describing. Which leads one to believe that some modifications have been done in an effort to get two governors (Speedtronic panels) to operate in parallel which both are in Isoch mode.

And, without being able to see what's running in the Speedtronic panel we can't add much to what's already been said.
 
CSA... yes!

BMP... before I rule-out lightning I would check all lightning-related protective devices and components for physical damage!

Regards, Phil Corso
 
Thank you for your advice. I think our generator control and speedtronic doing well.

After we collect the operational data and Mark VI trending, the machine fail to hold the frequency may caused by unbalance liquid fuel supply. The fact that we found is there were unbalance pressure in the flow divider (chamber # 6 indicate 330 psi while the others is 375 psi). Also we suspect to the varnish formation in the hydraulic oil system specially for servo valve of bypass Fuel Main Gear pump.
 
BMP,

It's not likely that the pressure differential you cited caused the problem you described. The general rule is that liquid fuel pressures to each combustor should be within approximately 10% of one another, and that any combustor with a pressure approximately 10% higher or 10% lower than the average of the others is probably a contributor to a high exhaust temperature spread.

The drop in frequency is definitely related to an <i><b>overall</i></b> drop in liquid fuel flow, not just a low pressure on one combustor/nozzle.

In general, a frequency decrease means an overload condition and that can occur when the load exceeds the unit rating (at rated speed), or when the overall (total) fuel flow-rate can't increase quickly enough or is somehow restricted or prevented from increasing quickly enough in order to produce enough power to supply the load at rated frequency.

I had a question about the 9 MW of "non-critical and non-essential" load you mentioned previously. In your originally stated scenario, if this load hadn't been "shed" (disconnected) when the tie breaker opened the total load would have exceeded the capability of a single unit.

I'm not clear on the operating situations from your previous post, as previously stated. And, it's not clear if you've had problems previously on separation from the grid or if this was a one-off event.

Oil conditions and servo cleanliness can affect the travel time and response of the LFBV (Liquid Fuel Bypass Valve). So it could be possible that a "sticky" servo wasn't responding to the servo current.

Lastly, when the frequency decreases the turbine speed (which is directly proportional to turbine speed) decreases. When turbine speed decreases the air flow through the machine decreases, which means the amount of power that can be produced by the turbine also decreases.... I don't know what the proportion is (load drop per delta speed change) but if the unit is at 48.5 Hz the speed is at approximately 97% of rated and the power output cannot be 21+ MW; that power can only be achieved when the unit is at rated speed, 100%.
 
As I mentioned before that "Non critical and non essential load" mean mainly the load for Offsite (Tank farm and marine facilities) and Utility plant, these load are supplied separately by National Government Grid. No UFS applied for these load.

The critical & essential load (20 MW load) is for Aromatic plant consist of LPG plant, prefractionation unit, NHT, platforming unit and Aromatic unit. These loads are supplied by 2 CTGs (21.484 MW isolated) in island mode. In these load applied 5 steps UFS load shedding when 1 CTG trip (UFS step 1 at 49.0Hz by PMS (0.5 MW load), UFS step 2 at 48.5 Hz (2.1 MW load), UFS step 3 at 47.8 Hz(6.0 MW load), UFS step 4 at 47.6Hz(3.5 MW load) and UFS step 5 at 47.5 Hz CTG open breaker).

 
You might have mentioned it before, but it wasn't clear (at least to me).

And, UFS is not a term I'm familiar with, and is one of the points of information that I maintain we aren't fully aware of and don't know enough about.

Look, every installation of GE-design heavy duty gas turbines is not the same as every other installation of GE-design heavy duty gas turbines. A Frame 6 is a Frame 6 (and a Frame 5 is a Frame 5, etc.) in that they all suck and squeeze and burn and blow, but they all do so with different auxiliaries and in slightly, but importantly, different situations and applications. And while they all might have a Speedtronic control system, those can also be programmed and configured slightly differently depending on the application and plant design/operation philosophy.

Not every site has a "UFS" (whatever that is, some kind of power management system I would imagine) and when you use a term like that it's best to provide some kind of definition or explanation of how it works and what it does, or at least what it's supposed to do. I have seen a number of power management systems, or load sharing systems, or load shedding systems, or frequency management systems, or whatever someone wants to call them. And most of them don't work "as advertised."

Why is anyone's guess, but mostly, because they weren't designed or applied properly. (The exception to this rule is some applications of the Woodward isoch load sharing system on prime movers equipped with Woodward controls and configured and applied by people familiar with and knowledgeable about Woodward control components. And, these are usually on much smaller machines than GE-design heavy duty gas turbines.)

And, another "problem" with these systems is that the people using them are not always very familiar with electrical generating fundamentals and they make a lot of assumptions about how the equipment should work and try to say that the equipment isn't working properly because it's not working the way they think it should work. And, it's their perception that's incorrect in many cases.

Analyzing an event like this on an Internet forum is very difficult. We don't know enough about this "UFS" and you haven't provided enough information, and it would be very difficult for you to provide all of the details necessary for anyone here to make an educated presumption about what might have happened and how to prevent it from happening.

Further, you have been asked and have not answered if this problem occurred just once or if it has happened during previous separations. If it's just this past time, well, then you need to understand what was different about this past separation from previous separations. Was the "UFS" in some different mode? What was different that may have caused the problem?

Certainly dirty oil flowing through the servo-valve could cause a delay in getting the fuel flow-rate to increase fast enough.

But, without knowing all of the details of the installation at your site and its configuration it's virtually impossible to say any more than has already been said. (Except to consider the addition of rotating mass (inertia) to fix the problem....) You haven't provided, and likely cannot provide, sufficient information to be of help. As was said before, a typically configured Speedtronic system isn't designed to be operated in Isoch mode in parallel with another typically configured Speedtronic system also operating in Isoch mode. Again, I'm speaking about typically configured Speedtronic systems.

And, further, Isoch mode doesn't use loading/unloading rates like Droop speed control does. Issuing RAISES/LOWERS to a machine in Isoch mode is changing the turbine speed reference, and hence the generator frequency reference. And the PI controller in the Speedtronic will respond, if typically configured, very, very quickly to either a reference change or a speed change caused by load.

So, there is something unusual about the application and configuration at your site, and the existence of the UFS and how it's interfaced to the Speedtronic panels is not clear, and likely cannot be made clear on an Internet forum such as this. The amount of information required (drawings and manuals and schematics) just precludes providing and analyzing the amount of data which would be required to provide you with some information about how your system operates and how it can be operated.

I would suggest having a read of the plant instruction manuals and design information, usually provided with by the architect/engineer with the documentation for a plant. It will usually describe the intent and the actual implementation of such schemes for a particular plant. You may find some very useful information there. If not, try contacting the A/E or the plant constructor for more information.

Best of luck with your analysis!
 
N

Namatimangan08

> And, UFS is not a term I'm familiar with, and is one of the points of
> information that I maintain we aren't fully aware of and don't know enough about.

UFS-Under frequency load shade. I won't be surprise if you are not familiar with it. UFS is the manifestation of a failure. It happens when the grid management or a power system operator has run out out of idea how to get the system going without interrupting consumers demands.

UFS should be a part of grid system but for a very stable system we won't realize that it is there.
 
N

Namatimangan08

CSA point is right. Just be careful to ensure that not to get it bigger than the value the shaft and the start up torque motor can handle. Otherwise torsion force can be pretty too big for the shaft to handle and it is pretty too big for the start up motor to crank the shaft.

I know one of mini hydro power stations in my country has to change it shaft coupling once every 2 months! I think that station has pretty big inertial of rotating mass that its coupling can handle or the frequency for the system that it serves hunting above the expected range or the combination of both.

Bottom line is adding inertial or rotating mass to an AC system can be done. Depends on how critical the motivation to have it there. We don't have to limit the flywheel attachment to GT shaft only. We can have a motor driven flywheel within the AC system that our GT is serving. The energy consumption will not be as big as what we might think. It is equivalent to "synchronism" power. In this case equals to windage resistance plus bearing loss + vibration (Excluding start up energy).

In short, inertial energy of rotating mass can come from supply or demand or combination of both as long as rotating mass is synchronized with the system.

Adjusting ramp rate cannot guarantee to solve our poster's problem. If ramp rate is to fast but inertial energy of rotating mass is relatively too small then we will have load hunting. It follows by AVR hunting. As a result you will see a lot of unthinkable damage on our mechanical and electrical equipments.

Increasing inertial energy of rotating mass, in theory will able to solve our poster's problem or similar problem. Whether it is practically feasible or not to design it is another issue.
 
N

Namatimangan08

We start with stored energy with rotating mass

Kinetic energy =0.5Iw^2 Joule

Where

I =mk^2 (kg-m^2)

w= angular velocity (rad/s)

k= radius of gyration (m)

m = mass (kg)

It is not mass that matters most. It is radius of gyration since its value is squared.

A 1,000,000kg of mass that is rotating at 1 rad/s with radius of gyration of 0.1m will have inertial energy of 10,000J

A 10,00kg of mass that is rotating at 1rad/s with radius of gyration of 1m will have the same inertial energy of rotating mass.

When you want to design a big inertial energy of rotating mass, locating radius of gyration for the given mass is utmost important to ensure you can minimize the cost of material to design it.
 
N

Namatimangan08

In this case equals
>to windage resistance plus bearing loss
>+ vibration (Excluding start up energy)

Shall be read " In this case equals to
windage resistance + bearing loss
+ loss due to vibration + electrical loss(Excluding start up energy)

 
For those interested... the world's largest Grid-Scale Flywheel Storage system was commercially produced in Stephentown, N. Y.

Last June, some 20MW was provided by Massachusetts-based Beacon Power to assist frequency-regulation service to the NYISO (New York Independent System Operator) System

Details can be found in the August Issue of "Power Magazine!"

Regards, Phil Corso
 
N

Namatimangan08

> For those interested... the world's largest Grid-Scale Flywheel Storage
> system was commercially produced in Stephentown, N. Y.
----snip ----
> Details can be found in the August Issue of "Power Magazine!"

Very userul info. Thank you Phil.
 
Again, for those still interested... the July issue of I.E.E.E Spectrum Magazine, has an article covering "Flywheels Keep The Grid in Tune!"

There is also mention of a study, carried out several years ago by USA’s D.O.E., proving that frequency-dips could be corrected by temporarily shutting-off home electric water heaters and dryers!

Regards, Phil Corso
 
Top