I have two machines in my plant operating in parallel: frame-V 20MW + frame-III 8.5 MW. running in NG gas fuel.

Average load is around 18MW (13.5 + 4.5) MW.

Frame-V machine is in isochronous mode and frame-III machine is in droop mode.

It is observed that when a sudden load cut occurs say 1.5 MW, both GTs throw load almost equally(0.8MW) following a MW oscillations.

Later on the MW oscillation damps out in 10/15 seconds.
I am searching for the explanation for load decrease in droop mode. Any help in this regard will be highly appreciated.

 

Barindra75...

I suggest you search the Control.com archives. Some 70 threads are listed!

Regards, Phil Corso

 

It's very interesting that you completely neglected to tell us how much the frequency changes and if it goes high or low during the disturbance.

How long has this problem been occurring?

If it has started recently, what has changed between the time the two units were operating properly and now? Was there some mechanical work done on either unit, say on the fuel control valve?

Was there some maintenance work done on the turbine control systems--calibration or card replacement?

What should happen when there is a load change--sudden or not--is that the Isoch machine should immediately respond to the load change (which causes the frequency to change which affects speed) to maintain the system frequency. If the frequency change is limited fast enough, the Droop machine's load will not change by very much, if at all--only the load on the Isoch machine should change.

If the load on both machines is changing then it would seem the Isoch machine is not controlling frequency and speed very well. A machine operating in Droop speed control looks at the error between the machine's speed reference and the machine's actual speed to determine how much energy (fuel) to put into the machine. (The turbine speed reference is almost always higher than the actual speed which is how energy flow-rate is greater than that required to maintain rated speed/frequency.)

When the frequency/speed is at rated and stable, changing the speed reference changes the speed error which changes the energy flow-rate into the machine (which changes the power produced by the generator). Increasing the speed reference increases the error and increases the energy flow-rate which increases the power output of the generator. (Again, this is presuming a stable frequency.) Decreasing the speed reference decreases the error which decreases the energy flow-rate which decreases the power output of the generator. This is how Droop machines are normally loaded and unloaded.

However, if the machine speed reference is stable but the frequency of the system changes then that will change the speed error which will change the energy flow-rate into the machine which will change the power output of the generator. If the frequency increases this will decrease the speed error which will decrease the energy flow-rate into the machine which will decrease the power output of the generator. (Again, this presumes the machine speed reference is stable during the frequency disturbance.) If the frequency decreases the error between the machine speed reference and the actual speed will increase which will increase the energy flow-rate into the machine and increase the power output of the generator.

So, what seems to be happening is that the Isoch machine is not responding very quickly to the frequency increase when load is "cut"--which is causing the frequency to increase which is decreasing the speed error of the Droop machine which is decreasing the energy flow-rate into the Droop machine which is decreasing the power output of the Droop machine-driven generator. Then when the Isoch machine finally corrects the frequency the load on the Droop machine should return to what it was before the load was "cut."

All of the above presumes there is no Power Management System or Load-Sharing System in place at the site, that the two machine's governors operate independently of each other and the only "communication link" or shared signal between them is the system frequency.

It would seem, <i>based on the information provided</i> that the Isoch governor is not responding very quickly to the load "cut" and as a result the frequency is increasing until the Isoch governor reduces the frequency back to normal. At which time the load on the Droop machine returns to the same as it was before the load reduction. Typical causes can be incorrect current gain on the Isoch fuel control servo loop, a sticking servo-valve (caused by poor oil quality, worn and/or sticking fuel control valve actuator, worn fuel control valve plug/seat, incorrect Isoch regulator gain, problems with speed sensing (not likely). The Isochronous regulator is a P+I (Proportional plus Integral) loop, so this may require still more tuning.

If this sudden load "cut" is something relatively new it could just be that the Isoch governor loop needs "tuning", which isn't very likely if the governors are Speedtronic (but it might be a possibility).

It's always best to eliminate the mechanical possibilities rather than just jumping right in tweaking the control system--which is generally accepted to be the root of all evil. But, what typically happens is the mechanical stuff gets overlooked and a lot (LOT) of time is spent tweaking and tuning the control system only to find when the mechanical problem gets so bad or the faulty part is replaced during normal maintenance that the machine is unstable again, and the control system gets still more blame and requires still more tweaking and tuning which would never have been necessary if the faulty or failing mechanical part had been properly identified and replaced to begin with.

Hope this helps! <b><i>From the information provided,</b></i> the problem does NOT seem to be the Droop machine, but you didn't tell us what happens to the frequency immediately after the sudden load "cut" so it's difficult to say anything else.

Please write back to let us know how the problem resolution progresses.

 

barindra75,
Do you have no comment or feedback?

Can you not tell us if the frequency rose slightly after the load "cut" and then returned to normal along with the load on the Droop machine returning to pre-load "cut" level?

Was the explanation not clear?

 

>I suggest you search the Control.com archives. Some 70
>threads are listed!
>
>Regards, Phil Corso

dear phil...please mention the keywords. i have searched like damm hell.

 

respected CSA sir,

I beg pardon for being out of network for 3/4 days.

I am agree with the reply part

>"It would seem, <i>based on the information provided</i> that
>the Isoch governor is not responding very quickly to the
>load "cut" and as a result the frequency is increasing until
>the Isoch governor reduces the frequency back to normal. At
>which time the load on the Droop machine returns to the same
>as it was before the load reduction. Typical causes can be
>incorrect current gain on the Isoch fuel control servo loop,
>a sticking servo-valve (caused by poor oil quality, worn
>and/or sticking fuel control valve actuator, worn fuel
>control valve plug/seat, incorrect Isoch regulator gain,
>problems with speed sensing (not likely). The Isochronous
>regulator is a P+I (Proportional plus Integral) loop, so
>this may require still more tuning."

For your kind information, frequency raises up to 50.5 during a sudden approx. 4MW load cut.

Another thing,

Recently we have a Hot gas path inspection of frame-V machine. With all instrumentation skid cable replacement.

After that this problem is coming, however no job carried out in the gas control valve or hydraulic oil line.

Another thing we have noticed that, IGV temperature control mode was in OFF condition, we made it ON. This will make the response faster as at 14MW load previously IGV was 87% open and now it is 56% open.

Secondly,
Frame-III machines are double shaft machines, with control in governor as well as 2nd stage nozzle angle.
Also these 8.5 MW generators are directly connected with the generation BUS, however 20MW Fr-V machine has a generator transformer (27 MVA) in between.

Another problem we face,
When only frame-III machines runs in parallel (one machine in ISO + 2 machines are in Droop, continuous "load and frequency hunting occurs" and the system becomes unstable i.e. if load fluctuations not controlled manually, the MW & frequency oscillation will increase unlimitedly, finally will cause tripping of generator in reverse power mode.

We have tried all the permutation / combination for ISO & Droop with all three machines, the problem persists.

Hence we need to depute an operator exclusively for only manually controlling this MW & Hz oscillation.

This problem is perennial. the turbine governor is old Mark-IV type in service since 1994.

Now we are going to commission Mark-VIe replacing the Mark-IV in coming 6 month time by BHEL-GE.

Please suggest what things can be taken care of.

 

barinda75,

Your post has contradictions and new information.

>Recently we have a Hot gas path inspection of frame-V
>machine. With all instrumentation skid cable replacement.

>After that this problem is coming, ....

Later you write:

>This problem is perennial. the turbine governor is old
>Mark-IV type in service since 1994.

So, which is it? The problem is perennial, or it started after the "instrumentation skid" cable replacement? (By the way, what is an "instrumentation skid"?)

If the problem just started after the cables were replaced, well, then, one would presume there is some problem with the wiring, most likely the fuel valve servo polarity was not checked after the cabling was replaced prior to start-up.

By the way, there is nothing wrong with Mark IV Speedtronic turbine control systems--except that most have been poorly maintained and troubleshooting can be difficult due to the lack of ability to record and plot real-time operating data. (There are third-party vendors who provide HMIs for Mark IV systems that can provide this functionality. It would probably surprise you to learn that there are still a couple hundred or so Mark IV systems still in service around the world, working very well, thank you very much, and even older Mark II and Mark I and even Fuel Regulator turbine control systems (the earliest GE-design heavy duty gas turbine control systems) which--when well-maintained by knowledgeable personnel--work without any problems.)

>Another thing we have noticed that, IGV temperature control
>mode was in OFF condition, we made it ON. This will make the
>response faster as at 14MW load previously IGV was 87% open
>and now it is 56% open.

How is this so, that "response" will be faster with IGV Exhaust Temperature Control ON than if it is off? When operating at Part Load with IGV Exh Temp Control ON the Mark IV is now also having to modulate IGV position as load changes in order to maintain exhaust temperature, whereas if IGV Exh Temp Control is OFF the IGVs are at a stable opening (except when modulating normally from Minimum Operating Angle to Maximum Operating Angle)--and the gain doesn't usually change for IGV control based on whether or not IGV Exh Temp Control is ON or OFF. Changing IGV angle when load changes abruptly is going to change the air flow through the machine which is going to have a large impact on power output, and to a certain extent, stability--versus if the IGVs were at a stable angle and air flow was relatively constant (presuming frequency was relatively constant).

>Frame-III machines are double shaft machines, with control
>in governor as well as 2nd stage nozzle angle.
>Also these 8.5 MW generators are directly connected with the
>generation BUS, however 20MW Fr-V machine has a generator
>transformer (27 MVA) in between.

>When only frame-III machines runs in parallel (one machine
>in ISO + 2 machines are in Droop, continuous "load and
>frequency hunting occurs" and the system becomes unstable
>i.e. if load fluctuations not controlled manually, the MW &
>frequency oscillation will increase unlimitedly, finally
>will cause tripping of generator in reverse power mode.

>We have tried all the permutation / combination for ISO &
>Droop with all three machines, the problem persists.

So, there is more than one Frame 3 machine (which was not clear or mentioned in the original post), and the Frame 3s have no impedance separating them, but all three are separated from the Frame 5 by the Frame 5's step-up transformer (or a transformer of some kind). Well, in this kind of set-up (no impedance separating generator-sets connected to a common bus supplying a load) usually special tuning is required for the governors to operate properly and "share" load--without some kind of over-riding power management/load-sharing system. It's not surprising that an operator has to be dedicated to monitor system operation under these circumstances. Again, without tuning the governors to operate properly in this circumstance they're not likely going to respond very well to load changes without manual intervention.

The Mark VIe, by itself, isn't going to solve the underlying problem that the system just doesn't lend itself to islanded operation without some kind of power management/load-sharing system--especially when multiple generator-sets are connected to the same bus without any impedance separating them.

You need to have a Power System Study performed to understand the configuration and make recommendations about improving system response under load changes. There are likely some relatively simple things which can be done, but, we aren't likely to get the full, correct story in this forum and we can't see drawings and configurations. And, this should be performed by a paid consultancy experienced in power system studies and islanded operations.

Best of luck with your endeavour!

 

Barindra...

Use Control.Com's "Search The Site" box and type in +hunting. There are some 450+ posts!

Phil

 

Barinda... has the problem been solved?

If not, I would like to propose a tried and true method. But first, answer one question:

Have you observed if hunting occurs for specific load-ratios?

Regards,
Phil Corso