In our captive power plant we run Four Frame-V gas turbines of BHEL-GE make. All GTGs run on Naphtha firing.

It was observed that, Fuel flow (FQL1) to two of the GTGs which are running at same load (~ 17.0 MW) was 48% and 62%.

Both machines are having same type of flow divider, magnetic pickups and the magnetic pickup speed gaps. Both machines' fuel servo valves are of same type.

Does different FQL1 means that the GTGs are running with different fuel flow rates? One machine is more economical than other?

 

One bit of information you didn't give is what is the HP Fuel Pressure before the Flow Divider? If the pressures are different, the actual fuel flow could be the same. Do you have any other fuel flow meter installed anywhere on the system, even on the LP side?

 

Fuel pressure before the flow divider differs by approx. 2.0 Kg/Cm2. We have fuel flow meters on the LP side which are reading same amount of fuel flow (hardly 0.3 m3/hr difference).

If the fuel flows read by the LP side flow meters differs greatly, then I would have tried to explore the reasons for higher fuel consumption by one of the GTGs. In this case, LP side flow meter of both GTGs are reading almost same value, and now I'm concerned about the FSBV assembly. Should I suspect the flow meters or FSBV assembly?

> One bit of information you didn't give is what is the HP Fuel Pressure before the Flow Divider? If the pressures are
> different, the actual fuel flow could be the same. Do you have any other fuel flow meter installed anywhere on the
> system, even on the LP side?

 

Is the LP Fuel Pressure the same on both machines, both before and after the LP filters? If the LP pressure before the filters is different, make it the same on both machines. If it is the same before the filters but different after the filters, check the filters. Remember a change in LP pressure will make a change in HP pressure, that may be the difference. I don't really think that you have a problem, just a difference between the the machines but your actual fuel flow is the same.

 

2 kg/cm^2 equals 28.45 psig, which is a large difference in liquid fuel supply pressure. Both machines should have the same liquid fuel supply pressure. Please adjust the the liquid fuel supply pressures to be approximately equal and you should find the flow divider feedback will be more nearly equal--<b>presuming the two machines were installed at the same time and have the same high pressure liquid fuel pumps, flow dividers and liquid fuel bypass valves AND the same turbine control systems with the same liquid fuel flow divider feedback scaling.</b>

 

Greetings Mr.CSA.

The two machines are installed and commissioned at different times, one is in 1999-2000 and the other one in 2008-09. However, both machines are of same type (Frame-V), uses same high pressure liquid fuel pumps, same flow dividers, same liquid fuel bypass valves (moog), same turbine control system (Speedtronic Mark-V control) and same liquid fuel flow feedback scaling.

After reading the logics in Mark-V control, I understood that the Droop correction factor-FSKRN2â plays vital role in calculating the FSRN or in other words FSR. This value is different for both of the machines.

Here I would like to name the machine running with lower FQL1 as machine #1 and the other as machine #2.

Machine #1's droop correction factor is 11.92% and FSNL FSR constant (FSKRN1) is 15.0%.
At 16.7 MW load TNH: 100.1% and TNR is 102.9%.

Machine #2's droop correction factor is 13.0 % and FSNL FSR constant (FSKRN1) is 14.7%
At 16.7 MW load TNH: 100.1% and TNR is 103.7%.

As per the logic display, the formula to calculate FSR is ((TNR-TNH)*FSKRN2)+FSKRN1.

Now, I would like to find out what is the droop percentage of machine #1 and machine #2.

I would also like to know the factors to be considered while fixing the droop correction factor.

 

Greetings,

When you post here at control.com for information such as this (to understand difference between to machines) it's most helpful--and expedient--if you will include as much information like you have below in your original post so that we can provide the most concise and appropriate response, instead of having to ask a lot of questions or make a lot of assumptions.

Droop speed control has been covered numerous times on control.com. There is a 'Search' field cleverly hidden at the far right of the Menu bar at the top of every control.com webpage. (We highly recommend using the Search 'Help' the first few times; while the search engine is very fast the format of search terms is not intuitive for most, though it's easily learned and quite powerful.) When searching it's most helpful to review several of the search results if you don't immediately find exactly what you're looking for as many times you can learn new technical words and terms you can use in subsequent searches to improve the results. So, try several searches, using different words and terms, in conjunction with the available formats to improve your search results.

At it's simplest, Droop speed control is about how much the load will change for a given change in the error between the actual prime mover speed and the prime mover speed reference. For GE-design heavy duty gas turbines with Speedtronic turbine control systems the actual speed is TNH and the turbine speed reference is TNR. For a machine in new and clean condition with clean inlet air filters and fuel that is very close to expected (heat content) and with ambient temperature, -pressure and -humidity at or very near nameplate conditions the TNR will be approximately 104.0% when at Base Load (or Peak Load for machines with Peak capability) when the grid frequency is at rated and is stable.

<b>In general,</b> FSR (or FSRN, as you correctly indicated) is usually designed to be in the range of between approximately 65% and 75% when the machine is in a new and clean condition, the inlet filters are clean and the ambient conditions are at or very near rated with fuel having the expected heat content.

FSKRN1 is the "typical" expected FSNL (Full Speed-No Load) for a machine in new and clean condition, with clean inlet filters and when the ambient is at or very near rated and fuel heat content is as expected. Let's say the value for a typical machine is 20.0% (which is pretty typical for most machines). So, if we subtract 20.0% from the midpoint of the average range of 'Base Load FSR' (which we said was between 65% and 75%), or 70%, the result is 50%. And, if we divide 50% by 4 (for 4% droop) then FSKRN2 is calculated to be 12.50% FSR/% speed error.

Now let's plug all the number into the equation and see what happens. When the grid frequency is stable and at rated (let's say 50.0 Hz), the turbine speed will be 100.0%. When TNR is 104.0%, the error between TNH and TNR will be 4.0%, multiplied by 12.50% will be 50%, and when we add FSKRN1 (FSNL FSR) to that the sum is 70% FSR (FSRN)--which is the midpoint of the normal range of expected FSR (FSNR) for rated gas turbine power output when the machine is in a new and clean condition, the inlet filters are clean, and the ambient conditions are at or very near rated and the fuel heat content is as expected.

It's just that easy. FSRN = ((TNR-TNH)*FSKRN2)+FSKRN1.

Now, as for why FSKRN2 is different for the two machines--that's anyone's guess (because, unless we know exactly what transpired during commissioning that's all it will ever be: a guess).

<b>NOTE:</b> I have taken great pains to emphasize that the approximate values of FSKRN1 and FSKRN2 are for turbines in new and clean condition (so, a clean axial compressor, with clean inlet guide vanes), clean inlet filters (low d-p), with ambient conditions (temperature, pressure and humidity) at or very near turbine nameplate conditions, and for fuel which contains the expected heat content at the time the turbine was manufactured and installed and commissioned (unless someone has had the fuel nozzles reworked because of a large change in the fuel since original installation/commissioning). Most turbines are not in a new and clean condition, having been operated for many years, some with normal maintenance and many with less than normal maintenance, most with not-so-clean axial compressors and inlet guide vanes, with dirty inlet air filters, and with internal hot gas pat parts not made by the OEM and so not with the same tolerances or life expectancy as OEM components. And, fuel does change over time (usually only slightly, but sometimes by more than enough to affect performance).

All of this is important because many people want to know, "How much Droop does my machine have?" And, they expect that after reading the above and ignoring all of the caveats (which most people do) that if they load them machine to Base Load that TNR will be 104.000000% and so their machine has 4% Droop--and when TNR only goes to 103.29% at Base Load they start screaming, "The OEM has screwed us! Our machines does not have 4% Droop! It's not efficient! It doesn't meet guarantee!" And, all while the machine is operating at an ambient temperature of 40 deg C (when the nameplate is 20 deg C), the inlet filter drop is 7.2 mm H2O (when clean filters have a 1.25 mm H2O d-p), and it's been six years since the last Major Inpsection, and two years since the last off-line compressor water wash, and the fuel was purchased from a least-cost supplier and has low heat content. And, they still expect the machine to produce rated power, or at least to have "4% Droop" and if it doesn't, well they're been cheated.

So, how to know how much Droop your machine has? Very difficult to say, or even to calculate or "simulate" while running, because how often is the ambient temperature and pressure and humidity exactly equal to the turbine nameplate value? At the same time the axial compressor was just washed (and properly washed and rinsed!), with new (or newer) inlet air filters, and with fuel which has approximately the expected heat content, and when the IGV LVDTs have been properly calibrated? And, the internal turbine nozzles and buckets and fuel nozzles are in near new condition with OEM tolerances? Almost never.

Here's about the best and quickest way to check how much Droop your machine has. Let the internal turbine temperatures stabilize while operating for a few hours at, say, a TNR of 101.0%. And record the load being produced by the machine after the internal gas turbine temperatures have stabilized; let's say the load was 6.2 MW. Then load the machine to a TNR of 102.0%, and after an hour or so with TNR still at 102.0%, record the load, and let's say it was 12.4 MW. (We're presuming the grid frequency--and therefore the turbine speed--is at rated, or 100%.) So, for a change of 1.0% in the speed error (from 101.0% to 102.0%), the load changed by 6.2 MW. If the gas turbine nameplate said that rated power output (Base Load on this machine) was 24,980 KW, or 24.98 MW--25 MW, then a change of 6.2 MW would represent a change of approximately 25% (25 / 6.2 = 4.03%)--for a Droop of approximately 4%.

If the load at 101.0% TNR had been 5.0 MW, and the load at 102.0% TNR had been 10.0 MW, the change in load for the change in speed error would have been 20%, or a Droop of 5% (25 / 5 = 5.0).

This is how to tell how much droop your machine has--approximately. Now, for machines in poor condition, and with "bad fuel" and with poorly calibrated fuel control valve LVDTs (for fuel control valves with LVDTs), this doesn't work very well--but then that's not the fault of the procedure, is it? (No.) And, this does also presume the IGV LVDTs have been calibrated properly, also (for machines with modulated IGVs and LVDTs).

So, make of this what you will. But, I <b><i>absolutely and sincerely</b></i> recommend AGAINST changing any Droop Control Constant unless you know exactly what effect it will have <b>AND</b> you are absolutely certain of the condition of the machine, the IGV LVDT calibration, the fuel characteristics, and what effects any change will have on unit operation. This information is provided without any warranty, express or implied, and with the admonition, "Kids--don't try this at home." The machine loading and unloading rates are directly affected by the values of FSKRN1 and FSKRN2. FSNL TNH is directly affected by the value of FSKRN1. There are lots of other possible knock-on effects of changing FSKRN1 or FSKRN2--none of which can be exactly known or predicted without being able to examine the CSP of the Mark Vs at your site (which may or may not be exactly the same!).

You make changes--you take the responsibility. And, the liability. And accept the consequences.

You now have enough information to be able to answer your own questions. Except, that without being able to see the CSPs in the Mark Vs, we can't say exactly what the effects of changing values will be--except that the loading/unloading rates, and possibly the FSNL speed--will change if you muck around with FSKRN1 and FSKRN2.

Hope this helps!

 

Thank you very much Mr.CSA. The detailed explanation had greatly helped me in understanding FSR related constants and droop setting.

I now want few more suggestions for adopting ISOCH mode of speed control, if you have time to spare.

Right from the inception we are running our GTGs in DROOP control only. As per OEM the machine is configured to run on both DROOP and ISOCHRONOUS controls. In the CSP file also we have seen all the logics related to ISOCH mode are available.

The following is our plant operating philosophy

1.We run four numbers of Frame-V gas turbines in "Island" mode and in DROOP control.

2.All four GTGs feed power to individual bus bars, at a given instance of time a GTG feeds a prescribed set of process units only.

3.We don't have load shedding scheme.

4.We often run GTGs at a load close to its base load (With marginal gap of 1.5 to 2.0 MW).

With those conditions, what is the best recommended mode of speed control?

If, its good to run in ISOCH mode of speed control then, it would be better for me to know the precautions to be taken or factors to be considered before proceeding for ISOCH mode (we proceed only after consulting our maintenance team and in presence of autorised service provider).

Thanks.

 

You're very welcome; glad to help when I can.

This is going to be very brief.

There's no way we can tell you how your plant should be operated. We aren't privy to plant design details, or plant configuration. I can tell you that if each of the four turbine-generators independently supply individual buses independent of each other and any grid while operating in Droop speed control mode then there is most likely an external control system sending commands (either via discrete inputs or a 4-20 mA signal) to control load based on frequency variations.

OR, the load is fairly stable and predictable and the operators can manually change load and control frequency (you haven't said how stable the frequencies are ...). Which isn't very likely, but possible.

But, changing the way a power plant operates is like trying to change the path of a star or make clouds appear or disappear. And, without a lot more information which you likely can't provide in a timely manner, we just can't say how (or why) your plant should be (or is) operated.

Good luck with this.

As for loads being close to rated, well, as long as they can't exceed rated, that's not really a problem. It's when they do exceed rated--then bad things start happening when gas turbines are the primary power source.

 

I'm thankful to you for sharing valuable information in this thread. I would like to end this thread with one single post.

In out frame-V machine Mark-V, I had observed one soft button "Ext load". It is placed just below Preselect, Base, Peak load buttons.

We are aware of Preselect, Base, Peak load but not "Ext Load"

In the alarms list I have observed an alarm related to this "Loss of External load setpoint signal" and in CSP file some logics are also defined on this. L83EX (External load control command) is one of the logic.

I believe that we are not using any of those in our operations but I'm little curious in knowing about a bit about this. Would you share some information on this.

 

When the power consumed by a process unit changes, the respective bus & GTG which are feeding to that process unit gets loaded/ unloaded and results in frequency drop/ raise. That frequency drop we are compensating by manually giving Raise/ Lower commands from Mark-V HMI. This is how we are operating.

We like to avoid the manual interventions to maintain the frequency why because we do have synchornous motors which are susceptible to high frequency variations. We thought changing the speed control to ISOCH would help in this aspect. We are getting in touch with OEM on this but before that I was just trying to get some inputs here.

Thanks for helping.

 

kranthikiran326,

In my personal opinion it should be possible to switch each of the Speedtronic control panels to Isoch mode--and then the operators would not have to make continual manual adjustments to load. At the very least, it is definitely worth a try for a few hours on each machine individually to see how well it works at your site for the load swings being experienced at your site.

Again, this is ONLY if the machines are truly isolated from each other, each supplying it's own individual loads.

I think you'll be pleasantly surprised--and we'd really like to know if you are. Or if you aren't. So let us know how it goes!

 

kranthikiran326,

External Load Control is a method for using an analog signal (usually) from a external source as the load setpoint for the turbine. Usually, a 4-20 mA signal, proportional to the rating of the unit (for example a turbine rated at 25 MW would have an External Load Setpoint input signal from the external controller scaled for 0-25 MW @ 4-20 mA). When enabled, the value of the 4-20 mA signal would be the reference for the load control function of the Speedtronic, and TNR would be adjusted as required to make the actual load equal to the load setpoint from the external source.

I don't think this would work very well at your site, based on the information provided. You would need to know "in advance" what the load would be (increasing or decreasing) in order to change turbine load to control frequency if that's what you were trying to do with this method.

Hope this helps!

 

CSA:
Thanks your excellent comment on droop.
Neo

 

Very useful information to give feedback by Mr. CSA.

Take care
G.Rajesh