Frame 5 gas turbine not detecting fire


Thread Starter


Dear all,

We have one Frame 5 gas turbine controlling by Mark5control system but after major overhauling we started the machine but machine not detecting flame after the spark plugs activation. Exhaust temperature is increasing a little bit but machine not detecting flame. We checked our srv gcv and flame detectors all working fine, but machine is tripping in firing mode,with loss of flame.

Can anybody help?


You really haven't given us very much information to act on.

1) How much does the exhaust temperature increase during firing?

2) What are the values of TTXSP1, TTXSP2 and TTXSP3 just before the firing time (T2F) expires?

3) You say there was a major overhaul. Were the gas valves (SRV; GCV) worked on during the outage?

4) Were the gas valve LVDT feedbacks calibrated after the outage? If so, do you have any before and after scaling data?

5) What are the values of the two Control Constants FPKGNG and FPKGNO?

6) What speed does firing occur at on your Frame 5? (What speed is the unit cranking at when firing begins?)

7) What is the P2 pressure during firing (approximately)? (The signal names is usually FPG2.)

8) Are the IGVs modulated, or are they the old "bang-bang" (open/closed) IGVs? (Do the IGVs have LVDTs?)

9) If the IGVs have LVDTs, were the LVDT feedbacks re-calibrated after the outage? If so, was the calibration verified using any measuring device (machinist's protractor) other than the pointer on the outside of the axial compressor casing?)

10) It's presumed the gas fuel supply piping was purged prior to the outage; are you certain there is gas fuel all the way up to the SRV? In other words, are you certain the air or whatever gas was used to purge the gas fuel supply piping has all been displaced and a high concentration of gas fuel is available up to the SRV?

11) Describe what, if anything, was done to the gas valves during the maintenance outage.

12) How many flame detectors are on your Frame 5 and which combustors are they installed in?

13) You did not mention checking the spark plugs for proper operation. It's a simple matter to remove the spark plugs, ground them (some sites use a car batter jumper cable to clamp to the spark plug and then to an unpainted surface), then force the spark plug power supply (usually signal name L2TVX). If you checked the spark plugs, what were the results?

14) What position is the GCV going to during firing?

15) What is the value of Control Constant FSKSU_FI during firing?

16) You say the machine is tripping during firing with loss of flame. That suggests flame is actually being detected by at least one flame detector, but that flame is lost in all flame detectors before the firing time expires. What are the values of all flame detectors during firing?

17) EXACTLY, what is the alarm that is being annunciated when the unit "trips"? Is it 'Loss of Flame' or 'Failure to Ignite'?

18) What is the value of the Control Constant K2F? (If the unit can also start and run on liquid fuel, there may be a K2F_GAS Control Constant; if so, please provide the value of that Control Constant also.)

19) Exactly when is the unit "tripping" during the firing time?

How many operator interfaces does the unit have? Are you experienced with using VIEW2 to collect data? Can you transfer data to a floppy disk or USB flash drive and then post it in a reply to this thread?

You want help; we need information. The more information you can provide, the better help we can provide.
Dear CSA,

Thanks for the prompt reply, during majour overhauling we serviced both SRV and GCV by out side parties. all LVDTS are calibrated at site that working fine. IGV also have LVDTS that also calibrated working fine. We started the machine on firing mode, during firing time IGV was open at 42 degree. also we checked both spark plugs and four flame detector in outside, all working fine. our fuel ggas pressure is coming around 18 bar up to SRV, during firing time P2 pressure going to 1.2-1.5 bar and after GCV it is around .16 bar. we started lot of time. one time two flame detectors catch fire, but firing in not unifrom. we have 10 chambers some of the exhaust temp thermocouple showing 100 degree and some showing around 60 degree during firing. GCV position during firing is around 20%. we are getting alram line breaker trip alarm during trip time, but one machine run continueously when two flame detectors detect fire without trip. but that time also exhaust temp was not uniform. There is no liquid fuel available, and during firing time only machine tripping. There is no alarm coming like "loss of flame or failure to ignite".

We are getting the alarm "L63fpg2h" p2 pressure excessive flow before tripping. Our L63fpg2h value is 1.7bar. we could not notify the flame detector values were checking logic changing or not. after GCV fuel directly going to primary nozzles and secondary nozzles through another on-off valve VGN. purge air also connected with outlet of this on-off valve and air flow is coming through 2purge valves.

You were specifically asked what the P2 pressure reference Control Constants are, what the firing speed is and what the actual P2 pressure is--in an effort to understand what the parameters for control are.

It also seems the GE-design Frame 5 heavy duty gas turbine has a gas fuel system that is designed for possibly two gas fuels, one of them being a low-BTU gas fuel. That probably means the P2 pressure reference Control Constants have names that are atypical which makes it harder to easily look up.

Without being able to see the application code running in the Mark* it's going to be very difficult to say too much more, because we don't have the information requested.

<b>Based on the information provided</b> it seems the SRV is unable to control P2 pressure during firing. The SRV was worked on during the outage. Therefore the SRV is suspect.

In addition, the GCV seems unable to provide sufficient fuel gas flow to establish and maintain flame during firing. The GCV was worked on during the outage. Therefore the GCV is also suspect.

That's about all that can be said, based on the information provided.

There is also another possibility: The hydraulic system (which is common to both gas valves) is unstable during firing, which is contributing to the gas valves being unable to do their job.

Again, based on the information provided it doesn't seem possible to say much more. I was prepared (against my better judgement) to say that the firing FSR might need to be increased slightly. But based on the new information my disinclination to make that recommendation was sound.

Something is causing the gas valves not to do their job properly. Use the troubleshooting capability of the Mark* to trend the various valve positions during firing, the P2 pressure, the speed, the flame detection logics, the the P2 pressure reference, the hydraulic system pressure, if available on the Mark*, and other signals and values to get actual data instead of relying on visual recall after a trip. This is the way to troubleshoot and resolve issues--with actionable data obtained using the troubleshooting capability of the Mark*.

Please write back to let us know what you find as you resolve the problem.
Dear CSA,

Thanks for the reply.

One time machine fired with all flame detection and gone to full speed after that start up, again same problem came. We have 4 flame detectors. You asked FPKGNG value is 0.1278bar and FPKGNO value is -1.87bar. Firing time P2 pressure is going to 1.5 bar, and we checked both SRV & GCV are working fine.

Perhaps someone else will choose to help, but you have not provided the data requested and the information keeps changing. It's clear you have some thoughts about what the problem(s) might be, and are looking for validation of you theory(s).

You won't tell what the spreads are, you said once that the unit was tripping on high P2 pressure, and now you say it made rated speed once but 'problems still persist' without ANY other details.

You will learn, eventually, (or maybe you won't; a few never do): Process Alarms really do mean something. And that Diagnostic Alarms do, too. It's extremely unfortunate that many sites believe a barrage of alarms during starting and operation is normal and the only thing that matters is to keep STARTing until the unit finally synch's and loads--because, eventually, most do. So lots of alarms doesn't really mean anything, and since troubleshooting ALL those alarms is just too hard, well, they can be ignored.

Maybe this will help: The standard is that every condition thst results in a turbine trip has a Process Alarm to indicate what caused the trip. So, it IS possible to determine--just by looking at the Alarms--what caused a turbine to trip. And, if there ISN'T an alarm then the programming or configuration needs to be corrected--but that's a RARE occurrence (that there is not an alarm to indicate why a turbine tripped that needs to be fixed). Believe it or not.

I can tell you that a GE-design heavy duty gas turbine with a Mark* turbine control system CAN START and operate a unit WITHOUT nuisance alarms. It's not common, unfortunately, but it can and does happen. But, it takes a desire and effort--and even then there are still the occasional nuisance or odd alarm, but they are the exception rather than the rule.

You have never said if these problems existed before the outage, and you are unwilling to recognize the problems (if they didn't exist before the outage) could be caused by work done during the outage. That's your choice.

The overwhelming majority of alarms and trips <b>ARE NOT</b> caused by the Mark*. Rather they are the result of inattention to alarms, failure to learn and understand the P&IDs, and lack of maintenance of auxiliaries.

And, contrary to popular belief--the Mechanical Department can and does make mistakes. You keep saying the gas valves were tested, but you won't say how or what the results were. It's obvious you believe the problems aren't the result of the work done on the gas valves, and the problems are not related to the hydraulic system. So, what else is there?

The Mark* is only as good as it's inputs and outputs. And, if one chooses not to troubleshoot and resolve alarms, then problem resolution is going to take a VERY LONG TIME. And, in the end, the problems will eventually be blamed on the Mark*, regardless of whether they are truly attributable to the Mark* or not.

It's very likely that at least part of the problems could be LVDT calibration--but that's an OPERATOR or TECHNICIAN error, not a Mark* error. And LVDT calibration has been covered MANY times before on; we're not going to cover it again in this thread.

Finally, IF the problems started AFTER the outage, and everything worked just fine before the outage. If no work was done on the Mark*, THEN HOW CAN THE PROBLEMS BE CAUSED BY THE MARK*?
Thanks CSA for your reply and sorry for the delay......long time I was away from the site, but I understood machine got fired and went to full load without doing anything, after so many start attempts. Still there is problem, once it get tripped it is not easy to start, after a lot of start attempts only it will go to FSNL. SRV, GCV, IGV all are calibrated, all spark plugs and flame detectors are checked. Problem is after spark command flame is not enough to catch the flame detectors, but exhaust temperature showing 100+ degrees. Machine not accelerating P2 pressure showing approx. 1.2 bar GCV also opening. But some times machine is starting and going to FSNL smoothly, so we are confusing. When I get the data which you mentioned I will forward to you once again thanks for your valuable support.
I am going to add three things for others reading this message. The original poster has, as many people often do, offered that the SRV and GCV and IGVs are all "calibrated."

First, the SRV is a pressure control loop--it will be sent to whatever position it needs to be in in order for the actual P2 pressure to be equal to the P2 pressure reference. Yes, there are LVDTs on the SRV which need to be calibrated--<b>BUT</b> the accuracy of the SRV LVDT calibration is NOT critical. In fact, the SRV LVDT calibration can be grossly out and the SRV is still going to work properly. Let's say the SRV LVDT calibration is 0.25 inches high--meaning that the Mark V thinks the SRV is 0.25 inches more open than it actually is. Doesn't matter--because the Mark V is positioning the SRV to make the actual P2 pressure equal to the P2 pressure reference (calculated in the CSP). So, even if the SRV is only at 0.625 inches of stroke but it thinks it's at 0.875 inches (because of the inaccurate calibration), if the actual P2 pressure is equal to the calculated P2 pressure reference the valve is going to stay at that position (even if the calibrated LVDT feedback is incorrect).

And, second: "Calibrating the SRV" (or the GCV or the IGVs or the LFBV) <B><I>DOES ABSOLUTELY >>NOTHING<< EXCEPT SCALING (CALIBRATING) THE LVDT FEEDBACK.</b></i> People--a LOT of them--mistakenly believe that AutoCalibrate does something to improve the stability of the device being "calibrated" by doing something magical to the electro-hydraulic servo valve or the hydraulic actuator or the signal to the servo. AutoCalibrate <b>ONLY</b> calibrates (scales) LVDT feedback. Nothing else. Full stop. Period.

Third, and finally, unless one physically compares the scaled LVDT feedback to the actual position of the device when calibrating the LVDT feedback there is NO CERTAINTY that the LVDTs were properly calibrated. If one just clicks on "AutoCal" and lets the Mark* do it's magic, the Mark has ZERO way of knowing what the actual position of the device is during the calibration, and does NOT verify the actual position against the calibrated (scaled) feedback.

So, one may believe the device is "calibrated"--BUT, ONLY the LVDT feedback scaling (calibration) is affected when an AutoCalibrate is run, and unless the scaled feedback is verified by checking the actual position against the scaled (calibrated) feedback--one can have a false sense of security.