Flame Failure During Startup -Frame6B


Thread Starter


We are operating GE Frame 6B GT with Mark V- TMR control, Simple Cycle. (HMI -Windows NT, CIMPROJ).

During the re-commissioning of one of our GT after the Combustion Inspection, all the four flame appeared at 19% TNH and two flames disappear at 22% of speed (can# 3 & 8) and again reappeared at 28% (approx) speed (on Gas Fuel). ( flame failure trip logic is 3 out of 4) Unit loaded, checked all parameters found normal.

During the normal shutdown ‘CHAMBER FLAMED OUT DURING SHUTDOWN’ alarm appeared at 39 % (approx) of speed.

Subsequent start-up and shutdown same thing is repeating.

Gas Inlet Pressure is 23 Bar (FGUP), we are having 3 Nos FPG2 transmitters, 2 LVDTs each for SRV and GCV, not using any barriers, SRV-GCV is a single moulded unit .

Please see the startup trend in below link

Liquid fuel operations, startup and base load and fuel changeover everything found perfect.

Following parts replaced during the CI: Dual Fuel Nozzle, Liner, Transition pieces and cross fire tubes all from OEM. We do not have much detail about the parts.

After the incident following checks were carried out:
SRV and GCV calibration checked (Stroke length, feed back etc) found OK, not changed existing IO-config values. Not used Auto-Cal feature. Not changed moog servo valves.

Current Gain: 1.93
Current Bias: 2.7
Current Gain: 1.92
Current Bias: 2.7

Following constants as found

FSKMINU2—FSRMIN FSR corner2-start up(=FSRWU)--11%

Shutdown FSRs


Shutdown speed

FPG2 pressure Transmitter(3 Nos)calibrated,found OK.

flame scanners rechecked , found OK .

Unit is in operation now (on gas).

Kindly give your advices to sort out the issue.

Thanks in advance.
This is not an unusual occurrence after a maintenance outage. I believe you neglected to tell us that once the unit accelerates to near FSNL (probably around 80% TNH, maybe higher) that the flame is re-established in the cans where it was not present.

It's probably pretty likely that prior to the outage that one can "flickered" during starting, or if not, was very close to it.

During starting, the "firing" FSR is set to provide a little extra fuel to help ensure flame is established. Shortly after flame is detected, fuel is cut back to "warm-up" value for approximately one minute, reducing the fuel to try to reduce the thermal shock of establishing flame and thereby reducing the thermal stresses on the hot gas path components.

Sometimes, especially during the cut-back to warm-up, flame can be lost in one or more combustors (cans). And during acceleration flame does re-establish itself in all the combustors.

Increased air flows into the combustor can cause this problem, from new or refurbished liners and/or fuel nozzles that are slightly out of spec. Incorrect reassembly can also cause similar problems. You didn't tell us what the exhaust temperature spreads were at FSNL after flame is established in all combustors.

For the start-up problem, it might be as simple as slightly increasing the warm-up FSR by 0.1-0.3% FSR. The Control Constant names is usually FSKSU_WU. It looks like in your case, the Control Constant name is FSKSU_GASWU, and that's kind of a significant drop in my experience, of nearly 33% from 17.6% to 11.3%.

I've always felt that the time delay in the Speedtronic between flame detection and cut-back to warm-up is a little too short. I've had good luck increasing that time delay a little (sorry; I can't recall the Control Constant name and I don't currently have access to any Speedtronic drawings/documentation). But, that's just my personal experience.

As for the shutdown alarm, it means that flame was lost in one or more of the combustors with flame detectors before it was expected to be lost (usually around 20% TNH on gas fuel, but may vary on some units). This likely a problem again with either increased air flows or problems with fuel nozzles installed during the outage.

The reason that fuel is shut off when this occurs during shutdown is to prevent high exhaust temperature spreads during shutdown from causing thermal stress on the hot gas path parts.

I can't make any recommendation about changing a shutdown Control Constant without inspecting the CSP for that particular unit. The fact that it's occurring at around 39% speed means it's not too bad; it will mean the unit will likely coast down to cooldown a little faster, too.

So, the upshot here is that the Control Constants were likely very close to marginal before the outage. Now that new components (or refurbished components) are in the machine, something has changed (increased air flows into the combustor or decreased gas fuel flow through the fuel nozzles) because of excessive tolerances on the parts. Combine these two and that can result in the problems you are describing.

Hope this helps!