Hello there?

Recently we are facing Combustion trouble- alarm in one of our GTs (frame 6B) with DLN system. It has Mark6e control system. During de-loading of that machine, while it reaches around 21 MW from Base load it generates Combustion Trouble-alarm just before of DLN transfer. Surprisingly CT-alarm does not appear all the time of de-loading. For example if we de-load 10 times from Base to 20 MW load it generates CT-alarm for 6 or 7 times.

From Mark6e trend it was observed that GCV starts hunting with GT load hunting (21 MW to 23 MW) and CAN 2 & 8 (FD-6 and FD-7) becomes flame out (lower threshold limit 16%) while P2 pressure and GSV remain steady.

Several times GCV and GSV servo replaced and calibration done accordingly to troubleshoot the combustion problem. It should be noted that servo current and LVDT FB found very normal during calibration each time and also during runtime of GT all gas valves reference command and Feedback remain almost same (very insignificant error observed). For your better understanding I would like to share two trends (one with CT-alarm another without CT alarm) in excel csv format of the same GT. Please follow the link below:

https://www.dropbox.com/s/xap0mwegrzm0m38/With CT Alarm.csv?dl=0

https://www.dropbox.com/s/7f2t0gtep26shbg/Without CT Alarm.csv?dl=0

Please don't be confused about P2 pressure in the trend. For some other reason SRV is adjusted to maintain P2 pressure around 19 bar earlier it was set to 15 bar.

 

There is a lot of data here to study but my first thought is to look at the DLN modes that the machine is in between the 2 sets of data.
How are you shutting the machine down, manually or Automatically? What I mean is are you reducing the load manually and then opening the breaker and shutting down? Or are you doing a Normal Shutdown (Press the STOP button and let the machine unload and shutdown)?

What I see is that you are reducing the load very slowly when the the machine is transferring from Premix to Lean-Lean. take a look at your "With Alarm" data and look at the DLN Mode. I think you are dithering and not getting the transfer complete quickly enough. Try to reduce load a bit quicker if you are doing it manually.

It's a first thought. I'll wait for your answer.

 

SMS,

If flame is lost in one or more combustors then power output of the unit will be lower than when flame was present in those combustors. As such, the Mark VIe will try to increase fuel to maintain load--especially if the operators are using Pre-Selected Load Control. So, that's probably the reason for the load swings.

How long has it been since the unit was through a maintenance outage (CI or HGPI or Major)?

How long has it been since the unit was tuned--DLN tuning?

Replacing a servo-valve DOES NOT require "calibration" of the device.

"Calibration" ONLY affects LVDT feedback--not stability of the device (unless one or both of the LVDTs has problems--which calibration will not solve either). It's entirely incorrect to say, "We calibrated the GCV and GSV," as if it means something was done to affect the gain or the stability of the GCV or GSV (or SRV or IGVs)--because AutoCalibrate only affects LVDT feedback scaling. When using AutoCalibrate one is only calibrating LVDT feedback--not the device or the servo-valve output/operation. And unless something has been done to change the physical stroke of the device or the stroke of the LVDTs then there is almost NEVER a reason to have to calibrate/re-calibrate the LVDT feedback from a device.

LVDTs are extremely reliable and do not drift or change calibration as many other types of field devices and instruments do. They are used on aircraft and spacecraft because they are so reliable and drift-free--the SAME LVDTs as used on GE-design heavy duty gas turbines. AND, LVDTs are just like pressure switches or temperature switches or limit switches--when testing them to see if they are working correctly, one first compares the present feedback to actual position (the as-found condition), and if and only if the feedback is not equal to actual position is the calibration changed--and that's noted as the as-left condition. Just like with a pressure switch or a temperature switch or a limit switch--the as-found condition is noted, and if and only if an adjustment is required to make the actual condition equal to the required condition is an adjustment made, and then the as-left condition is noted on the Calibration Data Sheet for the device. LVDTs are no different.

And LVDTs almost NEVER drift or require re-calibration.

And changing a servo-valve does <b>NOTHING</b> to change the physical stroke of a device.

And calibration <b>IS NOT</b> required when a servo-valve is replaced. A polarity check of the servo currents being applied to each servo-valve coil IS required after a servo is replaced--but a "calibration" of LVDT feedback is not require.

The problem you are facing is most likely related to a DLN tuning issue--and if you have been re-calibrating the GSV LVDT feedback and it hasn't been done correctly that could definitely be affecting the fuel splits which can be affecting the stability of the flame in the combustors at low Pre-Mix Steady State load operating conditions.

I haven't been able to download and review the files you have provided, but it would seem the unit may be using Inlet Bleed Heat to extend Pre-Mix Steady State load range. If there is something amiss with the IBH control valve and/or the IGV operation that could also be affecting flame stability at low Pre-Mix Steady State load operating conditions.

But, almost certainly--as your testing of the LVDT and device operation has revealed--the problem </b>IS NOT</b> caused by the servo-valves of either the GCV or the GSV, or even the LVDT calibration of the GCV or GSV. If the unit is losing flame in one or more combustors during low-load Pre-Mix Steady State operation the load of the unit will also decrease, which will likely send the GCV into an oscillating condition as it tries to maintain load--especially if Pre-Selected Load Control is being used to operate the unit. (Actually, most DLN combustor-equipped units use Constant-Settable Droop Speed Control, which also uses a PI regulator for load stability which can also cause GCV instability if flame is intermittent or lost during loaded operation.)

Please do not fall prey to the typical, false myths regarding "calibration" and servo-valves. You have needlessly spent a lot of time and money replacing servo-valves which were probably perfectly fine, having watched "trained and experienced" OEM and third-party field service personnel do something similar when an "instability" problem occurs (incorrectly, though with the best of intentions).

Please do write back with the answers to the questions above--but, most likely, the unit needs DLN tuning and/or is due for a maintenance outage to replace combustion hardware which will likely reveal some problems (plugging/blockage; cracking; etc.) of combustion components which may also be contributing to the problem. And, if the unit has IBH, then a problem with either the IBH control valve or the IGVs may also be causing a problem with flame stability during low-load Pre-Mix Steady State operation.

Hope this helps!!!