GE Frame-9E Standard Combustion Machine

We have two GE-9E machines. Recently we are changing fuel from HSD to Natural gas. Both the machines have the standard combustion with water injection facility for NOx control.

When we run the machines in temperature control mode OFF condition i.e. in Simple cycle mode the maximum load range we have got 115 MW~119MW depends on ambient temperature. (in our case site conditions are: Atm Temp=32 degC, RH=80%, Atm pressure=1013 mbarg.). But when we run the machines in temperature control mode ON i.e. combined cycle mode the maximum load range, we have got 106~111 MW without water injection and with water injection it changed to 113~115MW. As per our Combined cycle designed capacity both the GTs load have to be 115MW to 119MW with water injection. So now we are facing capacity shortage issue from off taker. We asked on site GE T/A but he did nothing. He sent data to GE Engineering team for review but still we could not find a way to compensate the capacity.

In both control mode at base load IGV opening is full and the exhaust temperature is different with a little difference in fuel flow rate. (in SC mode it is 520~528 degC and in CC mode it is 540~547degC). We all know the control of machines based on the programming in software. (it follows load order from the control block). So we also
requested GE for a review.

Anybody has any idea/suggestions please share.

Regards,
Rokan
 
My immediate thought..duct pressure losses.

>Anybody has any idea/suggestions please share.
>
>was this an existing problem...or just discovered baseD on new off take requirements?
 
> was this an existing problem...or just discovered baseD on new off take requirements?

Good question.

From the actual nameplate of the GTs (on the inlet plenum wall on the right side of Turbine compartment) what are the ratings for the two fuels?

How long since the last off-line compressor water wash for each unit?

How long since the last HGPI (Hot Gas Path Inspection)?

Have the IGV LVDT calibrations been <b>verified</b> using a machinist's protractor?

Water injection will only increase power output at Base Load.

Exhaust duct back-pressure shouldn't necessarily increase from one fuel to the other, and if it did I would expect it to be very slightly higher on liquid fuel--for two reasons. First, because the mass flow-rate through the turbine is usually very slightly higher on liquid fuel at Base Load than on gas fuel at Base Load. And, second, because amount of water being injected while running at Base Load on liquid fuel is usually higher than than when running at Base Load on gas fuel. The extra water flow also increases the mass flow-rate through the gas turbine, which usually increases the Base Load rating on liquid fuel slightly.

Any rating given to the off-taker should be based on GT nameplate rating. In my experience, all other things being equal the Base Load rating of a dual-fuel conventional combustor-equipped GE-design heavy duty gas turbine is usually slightly higher on liquid fuel than it is on gas fuel.

Yes, it's possible to tweak the Control Constants to get some extra power--but that should only be done with the knowledge that the firing temperature will also increase which should be factored into the maintenance outage interval planning which means the intervals would decrease which means the cost of operation would increase. I'm sure GE would accept a contract to do a performance test of the units and make recommendations for increasing power output, and/or making Control Constant adjustments. But they're not likely just to do it based on a (polite) request.

Answer the above questions and that may help understand what the conditions are that might be negativity impacting power output.
 
Rokan,

Do you have performance curves for both gas and HSD fuels? If so, do they indicate any difference in output between the 2 fuels?
Where you are giving a difference in output between CC mode and SC mode, are you running CC mode with the exhaust gas passing through the HRSG and SC mode with the exhaust gas bypassed to atmosphere? If so, the additional exhaust back pressure with the HRSG in service will reduce the output of the gas turbine.

At base load condition, there should be no difference in output between CC and SC modes as long as the exhaust path is the same. The fuel temperature control point is the same for both CC mode and SC modes. The IGV temperature control curves are different, but once the IGV's reach the full open position, there is no difference between CC and SC modes.

What is the configuration of the plant? Each gas turbine with its own HRSG providing steam to a single steam turbine? Or, maybe, both gas turbines exhausting into a single larger HRSG? Or, 2 single-shaft units (gas turbine, steam turbine and generator all on one shaft, and one HRSG for each)?
 
Dear All(Mr. 2c,Mr. CSA,Mr. Otised)

Sorry fellows for the late reply of the messages.

Q1:was this an existing problem...or just discovered baseD on new off take requirements?

Ans: off taker demand increased by 10% after fuel C/O from HSD to Natural Gas.Though the problem was persisting since commissioning but in HSD demand was less.So nobody concentrated on the problem.

Q2: From the actual nameplate of the GTs (on the inlet plenum wall on the right side of Turbine compartment) what are the ratings for the two fuels?

Ans: Compressor inlet conditions: 15degC,60%RH,1013 mbar
In HSD: 115300 KW,Heat Rate(LHV): 10960 KJ/KWh,Exhaust flow: 1508.6X10^3 kg/hr,Exh Temp: 520 degC.
In Natural Gas : 117900 KW,Heat Rate(LHV): 10870 KJ/KWh,Exhaust flow: 1504.6X10^3 kg/hr,Exh Temp: 519 degC.

Q3:How long since the last off-line compressor water wash for each unit?
Ans: Just before start up with new fuel(natural gas).

Q4: How long since the last HGPI (Hot Gas Path Inspection)?
Ans:HGPI of GT2 was carried out last year(November,2018) and For GT1 only boroscopic inspection carried out.

Q5: Have the IGV LVDT calibrations been verified using a machinist's protractor?
Ans: not really this time.Very beginning of commissioning of the new machines it was carried out.

Q6: Do you have performance curves for both gas and HSD fuels?
Ans: Yes, the output curves having very little difference.
Rated output: 115300KW on HSD & 117900 KW on NG.

Q7:Where you are giving a difference in output between CC mode and SC mode, are you running CC mode with the exhaust gas passing through the HRSG and SC mode with the exhaust gas bypassed to atmosphere? If so, the additional exhaust back pressure with the HRSG in service will reduce the output of the gas turbine.
Ans: CC mode-passing though HRSG.
SC mode- passing though bypass stack.

Q8:What is the configuration of the plant? Each gas turbine with its own HRSG providing steam to a single steam turbine? Or, maybe, both gas turbines exhausting into a single larger HRSG? Or, 2 single-shaft units (gas turbine, steam turbine and generator all on one shaft, and one HRSG for each)?
Ans: Plant configuration: 2GT-2HSRG-1ST

Regards,
Rokan
 
rokan_123,

You will need to provide (at Base Load steady state in CC mode):

CPD
CPDABS
AFPAP
CPR
TTXM
TTRX
TTKI_[0] (Control Constant)
TTKI_[1] (Control Constant)
TTKI_[2] (Control Constant)
TTKI_[3] (Control Constant)
TTKC_[0] (Control Constant)
TTKC_[1] (Control Constant)
TTKC_[2] (Control Constant)
TTKC_[3] (Control Constant)
TTKS_[0] (Control Constant)
TTKS_[1] (Control Constant)
TTKS_[2] (Control Constant)
TTKS_[3] (Control Constant)

Steady State operation is considered to be when no wheelspace temperature changes by more than approximately 5 deg F in 15 minutes--and it usually takes about 4 hours for a unit to stabilize at Base Load when started from a cold or warm condition (it may take a little less time if started from a hot condition, but not much less).

You need to provide ALL Process- and Diagnostic Alarms active when the unit is running at Base Load (especially if the Process Alarm "BACK_UP FSR EXHAUST TEMP CONTROL ACTIVE" (or something similar) is active). List all alarms--even if you don't think they are relevant. (It may be necessary to use ToolboxST to check the Diagnostic Alarms on all I/O Packs with yellow or red icons next to them.)

If the unit has exhaust duct back-pressure transmitter(s), can you provide the Base Load data on both SC and CC?

With the above information, we can make a couple of simple calculations to see what the CPD (CPR) should be when the unit is at Base Load for the conditions you are operating at, and, by extension the exhaust temperature.

Really, though, without being able to see the application code and Control Constant actually running in the machine it is going to be very difficult to say what is happening. otised is correct--the backpressure of an HRSG can cause the power output of the unit to decrease.

Finally, if the problem with output on Natural Gas was known at the time of commissioning, then GE should be working diligently to try to rectify the problem. Was there a Performance Test done during commissioning and before commercial acceptance of the unit by the owner? If so, and the Nat Gas power output was low, then GE should technically be responsible for rectifying the problem. Yes?

If a Performance Test was run during commissioning, can you compare the exhaust duct back pressure at that time to the exhaust duct back pressure now?
 
rokan_123,

I have been looking at the data you posted and I don't understand what happened to the load over the time period of the data in the file. Was it on Pre-Selected Load Control and the Pre-Selected Load Control Reference (setpoint) was changed by the operator?

So far, my examination of the data doesn't show any unusual conditions. The grid frequency isn't all that stable, and when that happens the axial compressor speed changes which changes the air flow through the axial compressor which affects the gas turbine exhaust temperature which causes the IGVs to move. A "significant" decrease (more than approximately 0.05%) in frequency (speed) will result in Droop Speed Control trying to increase load--which will put more fuel in the machine which will also cause the gas turbine exhaust temperature to increase. Pre-Selected Load Control, if it is active, will respond to the load increase to cause it to decrease (when Pre-Selected Load Control is active it is an outer loop to Droop Speed Control, and when there are frequency/speed issues that can affect load and IGV and exhaust temperature).

But, not knowing what was happening and causing the load changes during the period of the data file is a little bothersome. The upshot is, from looking at the data it just doesn't seem that anything is really wrong. As was said before--DLN combustion, since it relies on IGV control to maintain flame stability when operating at Part Load can make things "swing" more than one would think would be necessary. Throw in a little grid frequency oscillation, and that can only add to the oscillations. And, use Pre-Selected Load Control to grid frequency excursions and, well, you have a recipe for even more oscillations of the IGVs.

I was sincerely hoping the data you would get would be for Part Load operation <i><b>with Pre-Selected Load Control OFF.</i></b> You can use Pre-Selected Load control to unload from Base Load (or to load up to the desired load from a start) and then once at the desired load, just click once on RAISE- or LOWER SPEED/LOAD to cancel Pre-Selected Load Control. (Contrary to wildly popular and completely unfounded belief, the unit will not drift away from the current load. It will be hard--very, Very, VERY hard for the operators and their supervisors to just let Droop Speed Control control the load, but rest assured. Before Pre-Selected Load Control was invented EVERY generator prime mover in the world operated on Droop Speed Control at Part Load, and there were no major catastrophes nor widespread blackouts. People didn't lose their jobs. All was good.) Just let pure Droop Speed Control take care of what it's supposed to take care. (Be aware that if the grid frequency deviates Droop Speed Control will try to compensate--that's one of it's very important features! AND, it's a feature the grid operators WANT to work properly! So, small deviations in load may be experienced if the grid frequency isn't stable. That's normal and to be expected. Pre-Selected Load Control, if active when the grid frequency is not stable, actually makes grid frequency more unstable!!! YES--you read that right!!! It's NOT the proper way to be continuously operating a GE-design heavy duty gas turbine (at ANY load)! Just take a deep breath, and I assure you nothing catastrophic is going to happen, and no one is going to lose their job, and the unit is going to run almost exactly like it did with Pre-Selected Load Control active. The only difference will be if the grid frequency deviates, the load will also deviate (just as it would if Pre-Selected Load Control was active!) but it will actually be more stable and not oscillate as much as if Pre-Selected Load Control was active.)

Operation without Pre-Selected Load Control active will allow you to collect better data--without the influence of Pre-Selected Load Control possibly causing more exaggerated deviations. I didn't see 4 MW swings when operating at 65 MW--which is what I believe you described earlier. BUT, I don't know what was happening during the time the data was being collected, either, nor how it compares to when you see 4 MW swings while operating with Pre-Selected Load Control active at 65 MW.

Anyway, so far--nothing unusual. Two turbines, built one right after the other, and installed and commissioned at the same time, will never be identical in terms of fuel flow-rates, heat rate, and operating conditions. There are simply too many variables (internal clearances; cleanliness; IGV operation; exhaust duct back pressure; inlet filter cleanliness; ambient conditions (temperature; humidity; barometric pressure); and so on. It' just like two cars, made by the same manufacturer on the same day on the same assembly line, of the same model, with the same engine and transmission and tires and air filter and muffler--they will not operate identically. (There's a better chance of that if the cars have computer-controlled fuel injection--which monitors oxygen content and adjusts fuel flow for optimal conditions--something the Mark* DOES NOT do!!!)

Data. That's all we need. (And a little more time on my part; this has been the busiest maintenance outage season in many years. This is what happens when the economy is about to tank.... People spend money on maintenance they haven't spent in years, and then the economy usually begins to falter very shortly afterwards. It's actually very cyclical, and predictable.)
 
Dear CSA,

The following data so far I have collected from MarkVIe.<pre>
GT(SC)
AFPAP(mmHG) 756.35
ATID(°C) 20
CPD(barg) 10.9
CPR(ratio) 12
TTRX(°C) 529
TTXM(°C) 529
AFPEP(mmH20) -5.25
WQ(kg/Sec) 5.56
DWATT(MW) 124

GT(CC)
AFPAP(mmHG) 754.9
ATID(°C) 29
CPD(barg) 9.7
CPR(ratio) 11
TTRX(°C) 547
TTXM(°C) 544
AFPEP(mmH20) 353.17
WQ(kg/Sec) 4.48
DWATT(MW) 105</pre>
>TTKI_[0] (Control Constant):1100
>TTKI_[1] (Control Constant):0
>TTKI_[2] (Control Constant):1100
>TTKI_[3] (Control Constant):0
>TTKC_[0] (Control Constant):7.39098600715913
>TTKC_[1] (Control Constant):0
>TTKC_[2] (Control Constant):7.67451298701298
>TTKC_[3] (Control Constant):0
>TTKS_[0] (Control Constant):24.584
>TTKS_[1] (Control Constant):0
>TTKS_[2] (Control Constant):24.64
>TTKS_[3] (Control Constant):0

Regarding Back up FSR, there was no such alarm generated.Only the alarm was" FSR Temperature Reference Active".

The major difference we have observed in Exhaust Pressure during simple cycle(SC) and combined Cycle(CC) mode.

Regards,
Rokan
 
rokan_123

NOW we're getting somewhere. Alarms DO mean something!!! FSR Temperature Reference <b>IS</b> the Back-up Exhaust Temperature Control!!! (Also known as "Secondary" Exhaust Temperature Control. Primary Exhaust Temperature Control is CPR-biased exhaust temperature control.)

The turbine should ALWAYS be operating on the Primary Exhaust Temperature Control reference (TTRXP). Secondary, or Back-up, Exhaust Temperature Control is intended to be used if the CPD transmitters aren't working properly.

The concept is that the Primary- (TTRXP) and Secondary (Back-up) Exhaust Temperature Control (TTRXS) references (which are ALWAYS being calculated!) are supposed to be parallel to each other and NEVER intersect (at least not in any operating region of the unit). BUT, when the alarm "FSR Temperature Reference Active" is annunciated and active that means the Secondary (Back-up) Exhaust Temperature Control reference is LESS than the Primary (CPR-biased) Exhaust Temperature Control reference--and that's NOT how it's supposed to work. The two exhaust temperature control references (Primary and Secondary (Back-up)) feed into a MIN SEL block which chooses the lower of the two. And, unless the CPD transmitters are not working properly the Primary Exhaust Temperature Control Reference (TTRXP, the CPR-biased reference) should always be the lower of the two--<b>IF</b> the Secondary (Back-up) Exhaust Temperature Control Reference Control Constants were calculated properly.

You need to get GE or whoever supplied the turbine-generator package to re-calculate the Exhaust Temperature Control Constants, both the Primary- and Secondary (Back-up) Exhaust Temperature Control Constants. Again, the resulting temperature reference "curves" should parallel each other very closely--but they should NOT intersect each other in any expected operating region of the unit when properly calculated.

The "FSR Temperature Reference Active" (the Back-up Exhaust Temperature Control Reference) is only supposed to come into play when the CPD transmitters are not working properly. And, when the alarm is active the site personnel should be working to understand why the alarm is active and resolving the condition to clear the alarm, and return the unit to operation to the Primary Exhaust Temperature Control Reference (CPR-biased).

NOW, the problem may be that the 96AP transmitters may be not calibrated or working correctly--they are used to calculate CPR. And, it may also be that both the 96APs and the 96CDs are not working properly (not likely, but still possible) and that's causing a problem with CPR. But, most likely the problem is that the Secondary (Back-up) Exhaust Temperature Control Constants were not calculated properly.

You can verify this by looking at TTRXP and TTRXS when the unit is running. TTRXP should ALWAYS be less than TTRXS, by a few degrees F. If TTRXS is ever less than TTRXP then the alarm you described will be active--and that means that TTRX is <b>LESS</b> than it should be--and that means that the power output of the unit at Base Load is less than it should be!

Alarms DO mean something. They really, Really, REALLY, <b>REALLY</b> do!!! Just because the unit is still operating (and hasn't tripped) doesn't mean it is running correctly. And, when the alarm you described is active, it's not running at its full potential. It's at its full potential when TTRXP is the lesser of the two (TTRXP and TTRXS).

That's why I always try to remember to ask, "What Process- and Diagnostic Alarms are active?" when people write to ask for help with a (perceived) problem. In this case, the Mark* is doing what it's programmed to do--but something is either amiss with the inputs (easily verifiable by verifying proper operation and calibration (scaling) of the 96AP and 96CD transmitter inputs), or the factory calculation of the Control Constants. And people have become so immune to alarms they just don't believe any alarm (except ones that trip the unit--and even they can't identify which alarm actually tripped the unit!!!) because the commissioning personnel DID NOT make sure the unit would start and go to Base Load with ZERO alarms (Process and/or Diagnostic), and shut down from Base Load to Cooldown with ZERO alarms. "Nuisance" alarms <b>ARE NOT</b> normal! But because the Mark* systems are so poorly configured from the factory and because commissioning personnel are not supervised and evaluated on their ability to reduce "nuisance" alarm during normal operation (start-up, loading, unloading, shutdown, cooldown) too many units are left with LOTS of so-called "nuisance" alarms--which just shouldn't happen. It doesn't have to happen. It should be part of the commissioning Quality Control for commissioning personnel to be evaluated on the state of alarms when commissioning is "complete." But, it's not. It's never been. And the factory programs somethings to be alarms that ARE NOT really alarms--just "events" they think the operators should be aware of.

NOW, sometimes when a performance test is done during commissioning, GE adjusts some Control Constant values to achieve the desired operation and optimum performance. BUT, sometimes when they change one set of Control Constant values (usually the Primary Exhaust Temperature Control Reference Control Constants) the <i>forget</i> to recalculate the Secondary (Back-up) Exhaust Temperature Reference Control Constants.

Finally, it could be that the GCV LVDT calibration is not correct--not likely, but that could be the problem. I'm trying to list all possible problems, and state the likelihood of each one. In my experience, when the alarm you described is active it's because of improper calculation of Exhaust Temperature Control Constants. And, really only the GE factory engineers have the ability to plot CPR and FSR side-by-side to determine if they are parallel and will not intersect during any expected operating conditions (really hot days; really cold days).

I'm going to guess that when the unit is in Simple Cycle mode, the "FSR Temperature Reference Active Alarm" IS NOT active (that is--TTRXP is less than TTRXS) (IS NOT being annunciated). And, when the unit is in Combined Cycle mode, the "FSR Temperature Reference Active Alarm IS active (it IS being annunciated). Those are SWAGS, but probably pretty accurate.

Alarms DO mean something. Operators, their Supervisors, Technicians and their Supervisors, and Plant Managers <b>should NOT</b> live with any alarm they don't understand. Full stop. Period.
 
Top