Hi,

i am a shift engineer of a FR9E gas turbine-simple cycle. I've been asked to write an article to get promoted and i was thinking to write about how to eliminate flame scanners and replace the controlling to be with just exhaust thermocouples.

i need help to set the guidelines for my article. can we set some rung to come with the result of (flame established) by using the temperature change in exhaust thermocouples during firing time?

i was thinking of delta temp. versus time but i need values to be true and practical.

what about flame out (or fail to ignite)? is it possible to be controlled without flame scanners? how? combustion monitoring is already done by logic and no need to manipulate any rung (i guess) so, what i actually need is a complete picture of flame approval/ flame out without using flame scanners.

 

That sounds like a bad idea to me. Typical UV flame scanners will produce an FFRT (flame fail response time) of maybe half a second to 2 or 3 seconds, and can be self-checking besides. I don't think you'll be able to get that kind of performance measuring exhaust temp, and your system will be less safe.

 

Agreed; the concept here is a bad idea for several reasons--though I will add that as we speak work is being done to eliminate the use of UV flame detectors on combustion turbines through the combined use of axial compressor discharge pressure monitoring and exhaust temperature monitoring. The concepts presented here are much too simplistic and do not take into account load rejections (sudden loss of load) or start-up/shutdown conditions, as well as some time-delay on loss of flame that, as noted, is inherent in most flame detection systems.

Also, with some Dry Low NOx (DLN) combustion systems, there is no UV flame present in some areas of the combustor(s) yet there is still combustion (my definition of combustion being a temperature increase not necessarily in conjunction with UV flame).

Again, while there is research underway to accomplish this task, I don't believe the original poster has taken all conditions and modes of operation into account in this concept. There needs to be sufficient level of redundancy and error-checking to prevent nuisance trips. It may even be necessary to monitor load (power output) in the equation to prevent nuisance tripping. But, as presented, this concept is flawed with current technology and methods.

I do, however, like the concept of asking a candidate for promotion to document his/her understanding of some system or systems in detail, possibly proposing changes in an effort to improve reliability or availability. I have long maintained that operators and technicians both should be "tested" periodically--not to judge their current level of understanding, but to continually improve their understanding in the hopes that when an unusual situation presents itself they will be prepared to take appropriate action or to properly analyze the situation to decide on appropriate action. So, this part of the original post is a very intriguing and interesting idea....

 

CSA:

The OEM has been implementing software flame detection logic and completely removing flame detectors on W501F machines since at least 2004 - both DLN & diffusion flame units. We implemented a similar 3rd party solution in 2007 on 3 501F's. I see no reason the same concepts could not be used on GE combustion turbines, but I do not know of anyone who has done so.

Removal of flame dectectors has increased the reliability of the units, the reliability of the GE machines would benefit from such an improvement.

>Agreed; the concept here is a bad idea for several
>reasons--though I will add that as we speak work is being
>done to eliminate the use of UV flame detectors on
>combustion turbines through the combined use of axial
>compressor discharge pressure monitoring and exhaust
>temperature monitoring. The concepts presented here are much
>too simplistic and do not take into account load rejections
>(sudden loss of load) or start-up/shutdown conditions, as
>well as some time-delay on loss of flame that, as noted, is
>inherent in most flame detection systems.

 

mhwest,

The "other" OEM has also been experimenting with eliminating flame detectors on their turbines. :^) And, I mentioned as much, as you quoted--work is being done on this.

I just said the concept as stated by the original poster was not a good idea; I should have probably said it was incomplete. I did say it did not take into account all modes of operation and all possible causes of sudden loss of CPD and sudden changes in exhaust temperature.

As for your comment about increasing reliability, I would like you to explain how the Circle-Bar-W control system benefited from the modification to the flame detection scheme. Under what condition(s) was(were) the original flame detection system (employing UV flame detectors, or some similar method of flame detection) causing the system to be unreliable? During starting? During loaded operation, and if so, at what point or under what conditions during loaded operation? Was the problem caused by dirty flame detector lenses? By humidity condensing on the lenses?

The reason for the questions is that I've not run across problems with flame detectors affecting reliability--other than humidity condensation (and there's a fix for that), and dirty lenses (and the source of the "dirt" has to be understood (dust; soot; etc.)) to address the problem. The use of redundant flame detectors on GE-design machines pretty much eliminates false trips and indications, and proper attention to Diagnostic Alarms and maintenance will usually prevent multiple flame detector failures (which is something I've never seen--except when a failed flame detector was snot replaced in a dual redundant flame detection system in a timely manner). And, mostly, the problems can be attributed to poor wiring practices when replacing flame detectors, using wiring with the wrong temperature rating, and not replacing failed or failing wiring (usually when the insulation was damaged from excessive heat over time).

I don't have any personal experience with Circle-Bar-W machine control systems, but I have repeatedly heard of the reliability problems of the control systems and field devices (air-operated fuel control valves, in particular). So, that's the reason for the question--to try to determine if the 'problems with flame detectors' were actually flame detector problems or control system problems or mounting location problems or operational problems (say, bleed valves operating at rated speed under some conditions; or, poor fuel/air ratios under some operating conditions). My experience with GE-design heavy duty gas turbines is that the flame detectors are not usually the problem--it's either environmental (humidity, dirt/soot), or poor maintenance practices, but not flame detectors or even the control system the flame detectors are connected to.

For decades, there have been wiring drawing inaccuracies which caused the polarity of the UV flame detector power supply at the turbine compartment to be questionable. That is, some units would have the positive leg of the supply on the white wire and the negative leg of the supply on the black wire, and other units would be reversed. This makes connecting the flame detectors difficult, because they are polarity sensitive--with black and white leads (and I can't recall which was positive and which was negative....). So, people would not mark the wires when disconnecting them during a maintenance outage and would make an ass-umption when reconnecting them after the maintenance outage and cause the flame detector(s) to fail. Many people are deathly afraid to measure 335 VDC with their DVM (Digital Voltmeter) and just make the connections without verifying the polarity of the supply. (Thank you, Health & Safety Managers.) I know of one site that routinely failed multiple flame detectors after maintenance outages, until they were made aware of the fact that they were polarity sensitive and needed to be corrected properly (yes--it's absolutely the truth!). They were insisting for at least two years after the machine was installed that failed flame detectors were a warranty issue and demanding the OEM to replace them at the OEM's cost. Until it was discovered the failures were occurring when site personnel were re-connecting them after maintenance outages....

So, please provide some clarity and justification for the reliability improvement by describing how the previous method/detectors were causing low reliability in the Circle-Bar-W machines.

 

Hi team

Ge 9e unit with dln working on gas .

After major renovation, a phenomenon began , after transfer to premix it is low fire detection in the secondary area and at the same time also in the primary area then low detection.
About half an hour later at a base load fire detection in a secondary area gradually rises to normal values about 2000.
one thing was done to compare the situation before maintenance is the splitter position change for more open to primary because the other problem during shutdown that was not ignited primary zone .

is it could be some problem with flame detectors ?

we also change the limit low for indication cut flame from 4 to 12 because the jumping mode turbine from premix to lean lean also on base load automatically because of primary detection.

Thank you

 

leonid_r,

I am conservative when it comes to changing most Control Constants—PARTICULARLY when it comes to DLN settings. In a lot of places in the world there are no real penalties for high emissions and while these sites buy DLN combustor-equipped turbines (mostly for the optics of being able to say, “We care about the environment—we operate low emissions gas turbines!”), the actual emissions values are NOT monitored, NOT recorded and NOT reviewed. Changing the split DOES CHANGE emissions—sometimes not by too much, but other times by a lot. And while the turbine doesn’t care about emissions when the fuel flow to the primary combustion zone is increased the hot gas path temperatures are increased which IS NOT good for the primary combustion zone and Venturi section of the combustion liner. And could lead to unintended premature primary zone reignition events AND liner damage, even failure..

FURTHER, if there weren’t any issues with flame detection or intentional primary zone reignition PRIOR to the maintenance outage, presuming combustion hardware (liners and transition pieces, and nozzles, etc.) have the same characteristics (dilution hole sizes; cooling slot area; fuel nozzle flow characteristics, AND reassembly was done correctly—then there should be NO NEED to modify Control Constants AFTER the outage. None.

Now, if the IGV LVDTs and/or the gas fuel valve LVDTs were not calibrated correctly or the wrong parts were used during refurbishment of any or all of them then that could explain part of the problem. But not all of it.

If the igniters won’t reignite the fuel (which is ALREADY rich when in Premix Steady State!) then there’s either too much air flowing past the igniters during the reigniting attempt or the igniters and/or cables and/or igniter exciter aren’t working properly. If I recall correctly the igniters are energized for at least 30-, if not 60-, seconds during the reignition period—that should be MORE THAN ENOUGH time IF the igniters are properly inserted and working properly and the combustion hardware is the same as what it replaced and the IGV LVDTs and gas fuel valve LVDTs are calibrated correctly and the reassembly was done correctly.

EVERYBODY wants to blame the friggin’ control system. NOBODY believes the mechanical components or workmanship could possibly be the problem. It’s a universal constant—every problem is the result of the control system. A natural—but completely illogical—conclusion when the control system seems complicated and no one understands it or how the unit is supposed to work. “It should START and run exactly the same every time—even after a maintenance outage—and even if different components were used or not reassembled properly!!! It’s a computer!!! It’s super intelligent!!!” But out of the other side of the mouth comes, “It’s that friggin’ control system’s fault—it’s SOOOO COMPLICATED AND HAS SOOO MANY WIRES AND LEDs—it has to be the control system’s fault!!!”

DLN-I combustion IS NOT simple. It’s a very delicate balancing act—shifting fuel between combustion zones and maintaining fuel-air ratios (WITHOUT measuring the fuel-air ratio!!!).

And changing Control Constants after a maintenance outage because it’s not running like it did before the outage, is, .., well, generally not the proper solution. Of course, the mechanics are gone, and it’s likely the outage ran longer than scheduled and there is lots of pressure to do whatever it takes to get and keep the machine running WITHOUT CALLING THE MECHANICS BACK!!! But I can tell you, something is not right with the unit. It’s something we aren’t privy to (haven’t been told), or it’s a reassembly problem (like igniter insertion depth) or LVDT calibration or something like that. Sure; Control Constants can be jiggered but it’s a dangerous way to keep the unit running—and maybe not for long, and not safely.

The LAST thing to look at is: Has the way the unit is being operated changed at all? Does the unit have IBH (Inlet Bleed Heat)? Was it being used before the outage--is it being used after the outage? Is the unit transitioning combustion modes at the same approximate loads? Do the combustion reference temperatures (TTRF or TTRF1) seem the same after the outage as before for various loads and ambient temperatures? Is there any inlet air cooling in use now that maybe wasn't before the outage (evaporative cooling, for example)--or vice versa? I have seen flame detector intensity problems caused by very high humidity, and also because the cooling water to the flame sensors was not properly adjusted (this happens OFTEN after maintenance outages, when some well-meaning but unknowledgeable person (or persons!) fully opens the cooling water valves to the flame sensors thinking they should be fully open.... NOT USUALLY!!!). Something is different--that's certain. (And changing Control Constants trying to cover it up is NOT the right way to fix it!)

I’m done with this issue. On any thread—especially UNRELATED threads.

Best of luck!

I would love to know what a paid consultant (or GE) would recommend. But we’ll never know.... (I have known people who paid handsomely for good advice--and ignored it. And people who paid handsomely and followed it to the letter--with no positive results. When people keep pushing the same possible causes--it's usually because they want to believe their suggested cause(s) are THE causes and don't want to listen to any other possibilities. But without more, better data, and without an answer to the question about igniter insertion depth checking and results, we are wetting our trousers in a dark suit. It gives one a warm feeling, but no one notices.... And, it's NOT the right way to relieve one's self, either.)

 

Thank you

I will let know about next steps and the results

 

leonid_r,

We eagerly await your report.

 

Thank you very much for your consultation.
The problem was with a liner n10 , it was not connected correct.
after replacing this liner the problem solved

 

Thank you for the update.