SRV opens to 5%, FPG2 gets to 2 ~ 3 bar, SRV closes, FPG2 drops to 0.03 bar, then SRV opens again. The cycle continues until flameout and loss of flame trip.
MOOG ServoValves and LVDT for SRV and GCV have been changed with new ones twice.
The exhaust temperature attains 115degC before the flame out.
The combustion cans are clean with no signs of debris or FOD.

Please, what could be the cause and possible correction?

 

@Psalmywisee,

1) What alarms are present BEFORE the START attempt, and what alarms are annunciated during and after the START attempt?

2) You say the machine loses flame at about 36% TNH, but when is that during the START sequence? How long (time-wise) does it lose flame after flame is originally detected? 20 seconds? 40 seconds? 60 seconds? 90 seconds?

3) Do you have any trends/graphs of the problem? Do you have a Trip History print-out of one of the problem starts you can share?

4) When did this problem start? Before the servo-valves and LVDTs were replaced or after?

5) What kind of combustion system does the machine have? Conventional diffusion flame combustors or DLN (Dry Low NOx) combustors?

6) What is the control system being used to control and protect the machine?

7) Does the machine have an AC motor-driven Auxiliary Hydraulic pump?

8) What is the hydraulic system pressure doing during this time? Is it stable or is it erratic?

9) Does the machine have a hydraulic accumulator--and if it does has the accumulator bladder pressure been checked and adjusted if necessary (typically the accumulator bladder pressure should be at least 50% of the normal system hydraulic pressure when there is NO hydraulic system pressure (the Aux. Hyd. Pump is NOT running; the machine is not running/cranking)?

10) So, you've replaced the SRV & GCV electro-hydraulic servo-valves and the SRV & GCV LVDTs. Have you checked the calibration of the P2 pressure transducer(s) to be sure they are working correctly?

11) What is the gas fuel supply pressure doing during these firing attempts--is it stable? Is it erratic?

12) What is the GCV doing during these firing attempts--is it going to one position and remaining their stably or is it hunting open and closed?

During firing, a GE Mark* turbine control system will usually open the GCV first to a stable reference value (for example, 18.9% or 20.2%, as examples) and then the SRV is opened to control the pressure between the SRV and GCV (called the Interstage Pressure, or the P2 pressure) as a function of turbine speed. The SRV will usually go to whatever position is required to make the actual P2 pressure equal to the P2 pressure reference during firing, and if the turbine speed is increasing slowly (which it usually will be) then the P2 pressure reference and the actual P2 pressure reference will also be slowly increasing. When flame is detected in the machine the GCV will usually close just slightly (for example, by 1% or so, sometimes more, sometimes less) during the warm-up phase of the START process. This will usually cause the SRV to close slightly as the flow through the GCV is slightly reduced which will tend to slightly increase the actual P2 pressure--so the SRV will close ever so slightly to keep the actual P2 pressure equal to the P2 pressure reference.

You say the machine trips on LOSS OF FLAME--is this the actual alarm you receive when the machine trips? Or is this what you think is happening when the machine trips? Because a LOSS OF FLAME trip means that the flame in a "majority" of the combustors with flame detectors was lost BEFORE any other condition which would trip the machine was detected (Exhaust Overtemperature, or Low-Low Lube Oil Pressure, or High-High Vibration, for example). A GE-design Frame 5 heavy duty gas turbine usually has between 2 and 4 flame detectors depending on the age of the machine and the type of combustion system (conventional or DLN), AND the Mark* won't usually trip the machine on LOSS OF FLAME until AFTER the warm-up period is over (which is about 1 minute after flame was originally detected). So, it's important for us to understand exactly when during the START attempt the machine is tripping.

I know--this is a LOT of questions. And you may think some of them aren't related to the problem you are experiencing. But, we don't know anything about the machine at your location. Every GE-design Frame 5 heavy duty gas turbine IS NOT like every other GE-design Frame 5 heavy duty gas turbine--they all have the same basic turbine and axial compressor section, but they don't all have the same starting means (some have electric motor; some have diesel engine), they don't all have the same hydraulic system or even IGV system (some machines have modulated IGVs, some have IGVs that are either open or closed with no intermediate positioning); etc. Some turbines power generators; others drive centrifugal compressors or pumps. Some have a single shaft, others have two shafts. There are several versions of GE-design Frame 5 heavy duty gas turbines. Some Frame 5 machines use a SIMPLEX turbine control system; others use a TMR control system; and still others use a DUAL REDUNDANT control system.

But if you're just looking for a, "Do this and your problem will be solved!" answer--be sure the P2 pressure transducer(s) is/are calibrated and working properly. Be sure the valving for the P2 pressure transducers is correct (there is/are usually three-way valve(s) in the sensing lines of the transducers which are used for calibration--they should be in the SENSING position during STARTing and normal operation). Be sure the hydraulic accumulator--if there is one--is properly charged and the valves of the hydraulic accumulator assembly are in the proper positions (one of the valves should be open (the blocking/isolation valve; the other valve (the bleed valve) should be closed). Be sure the hydraulic pressure is stable during firing. Be sure the gas fuel supply pressure is stable during firing. If the gas fuel supply line was recently purged for a maintenance activity (either maintenance of the turbine or maintenance of the gas fuel supply system) be sure the gas fuel supply piping is properly purged of nitrogen and/or air prior to attempting STARTs. Be sure the LVDT wiring is correct after replacement, that the terminals are all properly tightened (not gorilla tight, but firmly tight) and if crimp connectors were used be sure they are all securely and firmly attached to the wires. Be sure the calibration of the SRV & GCV LVDTs is correct after the replacement, that the zero-stroke voltages are correct and the actual positions of the valves match the reference positions on the turbine control system when verifying the calibration is correct. Be sure the electro-hydraulic servo-valves servo current is correct for each coil of each servo (this should be done BEFORE the calibration procedure).

That's a good start. But, when writing to a World Wide Web forum like this for help with a problem it's best to remember--ALL GE-design Frame 5 machines ARE NOT identical; we can't see the machine/equipment; we don't know when the problem started. Provide as much information as you possibly can--ESPECIALLY alarm text messages. You will get better and more concise responses. When we don't know a lot about the situation/machine, we have to ask a lot of questions (as above) because these are not simple machines and, again, they aren't all exactly alike.

PLEASE do your best to try to answer as many of the questions above as possible. It will really help us to help you.

 

Check the fuel gas supply line for pressure fluctuations, restrictions, or obstructions.

 

@WTF? Thank you for the detailed description and questions. I must say that they are substantial to address any problem of this magnitude.

1) What alarms are present BEFORE the START attempt, and what alarms are annunciated during and after the START attempt? - Diagnostic Alarm <C><Q>, Generator Cooling Air Outlet Temp.High A, Voter Mismatch <T> L26GC2T, Voter Mismatch <S> L26GC2T, Voter Mismatch <R> L26GC2T.

2) You say the machine loses flame at about 36% TNH, but when is that during the START sequence? How long (time-wise) does it lose flame after flame is originally detected? 20 seconds? 40 seconds? 60 seconds? 90 seconds? - During the warming cycle. The status field changes from Warming to Firing and Warming to Firing until flameout. The duration is within 50 secs.

3) Do you have any trends/graphs of the problem? Do you have a Trip History print-out of one of the problem starts you can share? We did not set up a trend for the troubleshooting. We only analyzed the trip history and the loss of flame was pronounced trip signal.

4) When did this problem start? Before the servo-valves and LVDTs were replaced or after? Before the servovalves were replaced. The machined tripped when there was overvoltage on site.

5) What kind of combustion system does the machine have? Conventional diffusion flame combustors or DLN (Dry Low NOx) combustors? DLN

6) What is the control system being used to control and protect the machine? Mark V - TMR

7) Does the machine have an AC motor-driven Auxiliary Hydraulic pump? Shaft-driven hydraulic pump and dc-driven auxiliary hydraulic pump.

8) What is the hydraulic system pressure doing during this time? Is it stable or is it erratic? The pressure was 80 bar and not erratic.

9) Does the machine have a hydraulic accumulator--and if it does has the accumulator bladder pressure been checked and adjusted if necessary (typically the accumulator bladder pressure should be at least 50% of the normal system hydraulic pressure when there is NO hydraulic system pressure (the Aux. Hyd. Pump is NOT running; the machine is not running/cranking)? There is no hydraulic accumulator installed for the machine

10) So, you've replaced the SRV & GCV electro-hydraulic servo-valves and the SRV & GCV LVDTs. Have you checked the calibration of the P2 pressure transducer(s) to be sure they are working correctly? The loop test of the P2 transducers confirmed the transducers as okay.

11) What is the gas fuel supply pressure doing during these firing attempts--is it stable? Is it erratic? The supply pressure was 20 bar and stable

12) What is the GCV doing during these firing attempts--is it going to one position and remaining their stably or is it hunting open and closed? The GCV opens to 16% during firing and holds to 14 bar after firing.

Thank you for the submission to address the problem.

 

@Psalmywisee,

THANK YOUR for taking the time to answer all of the questions! We very rarely get that kind of response to questions, so it's much appreciated.

Hmmmm...

Soooo, the machine loses flame after the warm-up period expires and during warm-up more than one of the primary flame detectors are flickering.

The "problem" started AFTER the machine tripped from load. Can you describe what happened when you tried to restart the machine after the trip? Did it start doing the exact same thing as you are describing--hunting SRV position and flickering primary flame detectors and LOSS OF FLAME TRIP? How many times did you try to START the machine before deciding to replace the servo-valves?

What load and combustion mode was the machine running at when the overvoltage trip occurred? Did the Mark* V lose power after the trip?

The main purpose of the SRV is to act as the gas fuel stop valve--BUT it has an almost equally important function of controlling the pressure between the SRV and GCV (called P2 pressure) as a function of turbine speed. During firing the machine speed (TNH) is probably around 20% or so which means the P2 pressure reference (FPRG) is low--and that's a good thing because you say the gas fuel supply pressure is 20 bar (and stable), and the SRV reduces the gas fuel supply pressure to something around 2-3 bar so that the GCV can open to a controlling position (able to control the gas fuel flow rate which is very low during firing, and if one gas valve was trying to reduce the pressure from 20 bar to about 3-4 bar (the pressure downstream of the GCV during firing) the valve would be almost closed and couldn't control the flow/pressure very well. So, the SRV helps the GCV to limit the pressure and flow during firing and acceleration.

The SRV is primarily a pressure-control regulator--but it does have LVDTs which serve to try to add stability to the pressure control loop. The SRV will be driven to whatever position is required to maintain the actual P2 pressure (FPG2) equal to the P2 pressure reference (FPRG)--at a rate of 128 times per second the actual P2 pressure is compared to the P2 pressure reference and if necessary the Mark* V moves the SRV to make the actual P2 pressure reference equal to the P2 pressure reference.

This means that if the gas fuel supply pressure (upstream of the SRV) is stable and the GCV position (downstream of the SRV) is stable that something seems to be causing the Mark* V to believe the actual P2 pressure is not equal to the P2 pressure reference and that if it hunts/oscillates that the actual P2 pressure is not stable.

This could be caused by a problem with the SRV internals (plug, seat, stem, seals, closing spring) or the SRV actuator (the single-acting piston that opens the SRV against its closing spring pressure may have a worn spot which allows high-pressure oil to leak by or the actuator piston o-rings or seal can be worn and need replacement). Based on the Diagnostic Alarms it does not appear the Mark* V is experiencing problems with SRV regulator in any of the three control processors or problems with the P2 pressure reference.

It could also be that after the servo-valves (and LVDTs) were replaced that there is some problem with the wiring of the servo-valves and/or LVDTs, though if it was a serious problem it would most likely be accompanied by Diagnostic Alarms indicating the problem. One thing that happens often is that the polarity of the servo current being applied to each individual servo coil is not verified (and corrected, if necessary) after the servo-valve is replaced. Also, there are regulator gains which can be "tweaked" (adjusted) over time, and when installing a new (or rebuilt) servo-valve the regulator gains should be reset to the value specific in the Control Specification provided with the machine. Finally, there is this thing called Servo-valve Null Bias Current which seems to get misadjusted and can cause problems with control and which ALSO should be reset to the value in the Control Specification before "calibrating" the SRV.

About that--"calibrating" the SRV. It's really a bad thing to say or call the act of using AutoCalibrate on a device (including the SRV) "calibrating" the device. Because one isn't calibrating the valve (or the IGVs), or the servo--AutoCalibrate ONLY calibrates the LVDT feedback. Nothing else. Full stop. Period. People often (mistakenly) believe that when a servo-valve is replaced that it's necessary to perform an AutoCalibrate of the device the servo-valve is controlling the flow of hydraulic oil to--and that's INCORRECT. One MUST verify the polarity of the servo currents to each servo coil independently when replacing a servo-valve (whether the servo-valve is new or refurbished) but changing the servo-valve DOES NOT affect the range of travel of the device or the LVDT feedback. So it's NOT required to AutoCalibrate after replacing a servo-valve (unless something was done to affect the LVDTs or the range of travel of the device (like a refurbishment of the SRV or GCV, or the IGV actuator was replaced, for example).

Now, at some point the SRV LVDTs were replaced--so THAT definitely requires recalibrating the LVDT feedback (actually, it's scaling the LVDT feedback that AutoCalibrate does which can technically be called "calibrating" or "re-calibrating" the LVDT feedback).

Without a lot more data (obtained by using VIEW2) I can't really say what's happening. If one of the servo coils has the wrong polarity connected to it that can cause the Mark* V servo currents to be imbalanced and that can cause a problem with the stability of the device in some cases. It would be necessary to know what the three servo currents to the SRV coils are doing when the SRV is manually moved to several positions (10% stroke (position), 25% stroke, 50% stroke, 75% stroke) in order to rule out incorrect servo current polarity. It would be necessary to know what the SRV regulator gain values in the Control Spec AND in the Mark* V are. It would be necessary to know that the SRV Null Bias Current values in the Control Spec AND the Mark* V are.

You say, " The loop test of the P2 transducers confirmed the transducers as okay." Was pressure applied to each of the P2 pressure transducers and the feedback observed on the Prevote Data Display? Because just checking the "loop" (electrical circuit) doesn't prove or verify the accuracy of the P2 pressure transducer calibrations.

And, if pressure was applied to the P2 pressure transducers using the three-way valves on the transducer sensing lines, are you absolutely 100% certain that the three-way valves were all returned to the proper positions after the pressure testing was finished? Even on partially closed valve can cause problems.

The calibration of the SRV LVDTs is not critical--but the calibration of the GCV LVDTS IS critical. Again, the SRV regulator is a pressure control loop and the Mark* V is going to move the SRV to whatever position is required to make the actual P2 pressure equal to the P2 pressure reference (which means if the LVDTs tell the Mark* V the valve is 15% stroke when it's actually at 9% stroke when the actual P2 pressure is equal to the P2 pressure reference it doesn't matter that the LVDT calibration is incorrect). BUT, if the Mark* V thinks the GCV is at 14% but it's actually at 12% because the LVDT calibration isn't accurate that means that too little fuel will be flowing into the machine during firing--which can cause flickering flame detectors during firing and warm-up and lead to LOSS OF FLAME trips. Again, you said this problem started before replacing the SRV servo-valve and LVDTs, so this isn't likely the original, root cause of the problem. BUT it could be.

We also don't know what kind of SRV and GCV valves the machine has. It's possible it is still using the combined SRV/GCV assembly which contains BOTH the SRV and GCV in one single casting. If so, there should be a roll pin which holds the LVDT bars of the two valves in a stable position (because the LVDT bar is clamped to the valve stem and can move if not pinned to the valve stem), and there have been repeated cases of the roll pins shearing off and allowing the LVDT bar to slide down--especially when the machine has been tripped many times (because both the SRV and GCV will be slammed shut by their closing springs when the machine is tripped).

Finally, when calibrating the LVDTs of a combined SRV/GCV assembly it is necessary to use feeler gauges between the actuator rods and the valve stems to make sure the valve stems are touching the bottoms of the valve plugs. There is typically a 0.030-0.050-inch gap between the top of the valve stem and the bottom of the valve plug which must be "accounted for" during LVDT calibration. (When the feeler gauges are removed, it's COMMON for the indicated LVDT position to be slightly negative--because there is an intentional gap between the valve stem and the valve plug--so do not be alarmed to see a NEGATIVE 1-3% position on these two valves after a proper LVDT calibration is completed and the feeler gauges are removed.

None of the above really explains why this started after a machine trip and continues after the servo-valves and LVDTs were replaced. The only thing I can think of here problems with fuel nozzles being plugged with debris, or serious cracks in the combustion liners around the fuel nozzles that were made worse when the machine tripped. If the machine experiences a lot of trips (an excessive number of trips) that could explain serious damage to the liners and/or nozzles. It's a serious thermal shock to the machine when the fuel is nearly instantly shut off and the flame is extinguished, especially when the machine was running at a high load. If the machine regularly runs, even for short periods of time, in Extended Lean-Lean combustion mode that's NOT good for the combustion liners (if I recall correctly, GE says 1 hour of Extended Lean-Lean operation is equal to 10 hours of Premix operating--Extended Lean-Lean is pretty hard on the combustors.

Based on the information provided, I don't believe the problem is with the Mark* V--possibly with the inputs to the Mark* V, but not with the Mark* V. Good high-speed data might make me change my mind.... And, I've been wrong many times before and will probably be wrong many times in the future. It's more likely something to do with manual three-way valve positions in the Ps pressure transducer sensing lines, or P2 pressure transducer calibrations, or SRV actuator or -valve problems. Poor LVDT calibrations would make the problem worse (presuming the original problem and the current problems are caused by the same thing(s)). But, again, based on the information provided it's not likely the problem is being caused by the Mark* V--again presuming all of the inputs and outputs have been verified to be good. (Sorry; the Mark* V gets unfairly blamed for a lot of things and is often seen as the "root of all evil" when it comes to GE-design heavy duty gas turbine problems--but, very often it's not the root cause, it's bad inputs or output or poor maintenance or just poor understanding and lack of documentation. ( suspect this is another case of the same--based on the information provided.)

Please write back to let us know how you resolved the problem!

 

@Psalmywisee,

By the way, you can add photos and files to posts and responses here on Control.com. (If you send photos of HMI displays, please be sure they are clear and readable.)

 

Be sure to check the fuel valve actuators, under flowing conditions