GE 9E DLN1 secondary flame trouble

F

Thread Starter

FEFO

Hello everybody

we have four GE 9E gas turbines dual fuel running on gas.
DLN1 with 3 independent control valves system.
no IBH.

at the start up of the unit we can reach lean-lean positive without any problems. to transfer to premix steady state mode we have to stay in the lean-lean positive at load of 75 MW and wait for the secondary flame reading to rise. they start from "ZERO" value and takes up to one complete hour (some times 40 minutes or 30 minutes) to reach more than 40% on three of them so we can transfer to SS-premix mode.

once the unit converted to SS-premix, it can be lowered to any mode and convert easily to SS-premix.

i think these units are not designed to be like that does any one here have seen anything like this? and found away to solve this problem?

regards
 
Dear FEFO,

on a DLN 1 unit the secondary flame detector values should go to a normal "flame ok" value as soon as the unit goes into lean-lean mode.

Have you removed all of the flame detectors and inspected them for cleanliness? This type of slow rise is not normal. I have heard of low flame detector output values due to condensation on the detector windows which could be a problem in your case.

But to answer your question. This is not normal.
 
we cleaned the detector many times with little difference.

> Have you removed all of the flame detectors and inspected them for cleanliness?
 
Small detail, but, does this happen on only one of the four units, or all of them?

Also, when the units transitions to Lean-Lean Positive, what happens to the load when the secondary flame intensity is very low? Does it decrease, or stay the same? Because when fuel is split between the primary and secondary combustion zones when the transition to Lean-Lean occurs if there is little or no secondary flame it would certainly seem that load should decrease? Do the operators compensate by increasing the load during this period immediately after the transition from Primary to Lean-Lean?

While not all of the fuel that's flowing through the Secondary fuel nozzles is burned in a diffusion flame, some of it is and if there's no diffusion flame it would certainly seem that without fourteen secondary diffusion flames the load would decrease, by a few MW at least.

Also, what happens to the exhaust temperature spreads when this low secondary flame intensity at Primary-to-Lean-Lean transition is occuring and the secondary flame intensity is low? Do the spreads increase, and then decrease as the secondary flame intensity increases?

How long since the last maintenance outage (Combustion Inspection; Hot Gas Path Inspection; or Major Inspection) for this machine?

Are there coalescing filters on the gas fuel supply, and if so, are liquids being captured, and if so, what is the nature of the liquid (clear; viscous; dark)? Is there a lot of entrained dirt (sand; grit; rocks; etc.) in the filters or strainers?

Are witches' hat strainers installed on the gas fuel lines? Have they been checked/cleaned recently?

But, the most important questions are: How many of the four machines are experiencing this problem, and how long since the last maintenance outage(s) for the machine(s) experiencing this problem?

Lastly, what other troubleshooting steps have you taken--and what were the results?

Have you asked the OEM (Original Equipment Manufacturer) about this problem? If so, what have they said?
 
This sounds like condensation on the secondaries as suggested above. Our load profile is straight to base load. We would flameout during the transfer.

There is a hardware option for it "anti-condensateur detecteur flamme". It is a hexagonal nut which sits between the flame detector and the mounting. The problem is during certain climates and daily cycles condensation forms on the lenses whilst on cool down. The nut has three tiny holes in it which continually purge air to the enclosure. As a work around we were able to minimise the problem by isolating cooling water when the unit had shutdown and cooled appropriatly. Raising the cooling water temp also helped. It only affects the secondaries on the 9e as you have seen. Our dln 6b's on the same site do not have the problem.

Speak to you CPM about it, GE are aware of the problem, but will still ask for mindless data and trends first.
 
We have the same problem in our site with four DLN 1 gas Turbines "9E" the problem exists on the four of them, I have searched for the cause of the problem until i got it explained in the flame detector's data sheet:

PROBLEM: Periodic low reading on secondaries of DLN1 turbines.

CAUSE: Condensation on the lens, that can occur under high humidity situations.

SOLUTION: Information on a modification package to eliminate this problem is available. Contact your local GE Energy Services representative for assistance or for additional information.

That was the last row in the troubleshooting table in the data sheet. We kept talking to the GE representative about the problem and we sent them the data sheet file, but they asked us every time to try cleaning the lenses which didn't make any difference and the problem still exists after about one year.
 
Dear FEFO, as CSA has asked:

Does this problem happen on all 4 machines?

When were these machines put into service and when did the problem begin?

Does it happen all the time, on each start? Or is it somewhat dependent on the whether conditions?

What type of control system is installed?

What brand and model of flame detectors are you using?

Just so I am sure you understand the normal starting and running of the I am going to run through a quick explanation.

When the unit starts fuel should be admitted into the primary nozzles only and burn in the primary zone. All 4 primary flame detectors should indicate flame as soon as the unit lights off. Primary mode will continue through runup to FSNL and loading to approximately 20% load depending on ambient conditions. At some point the unit will transition to lean-lean mode, this transition is based on some TTRF number. Lean lean mode has fuel admitted to the primary zone as well as the secondary zone. Basically the total fuel demand is split between the primary and secondary valves and delivered to the appropriate nozzles. IF the unit began to transition from primary mode to lean lean mode AND the fuel being taken from the primary and delivered to the secondary was not being being burned(no real flame) then the unit should lose load very quickly, and TTRF temperature should go down, forcing the unit back into primary mode.

In your case it does not sound like this is happening.

I am assuming in your case that when the unit transitions into lean-lean mode that the megawatt output is increasing as it should and flame is actually present in the secondary zone, but the flame detectors are not reading it correctly due to some issue. I assume that you continue to raise unit load till you are near the TTRF temperature when the unit would be starting a lean-lean to premix transfer and you hold there until secondary flame detector values increase to the value needed for the control system to be satisfied that there is flame. Once that occurs then you can raise load further, where the unit will transfer from lean-lean to premix steady state.

DOES THIS SOUND CORRECT?????

If this is the case I still feel like you may have an issue with condensation building on the glass of the flame detectors. I am going to assume that you are using the Reuter Stokes flame detectors that are water cooled, can you verify I am correct in my assumption?
There have been changes to the secondary fuel nozzles where the secondary flame detectors mount to decrease the chance for condensation to build. The change involves shortening the pipe that the flame detector mounts, basically bringing the detector closer to the nozzle to raise its temperature. Some sites I have heard of also remove the cooling coils to raise the temperature of the detectors. The bleed air option that 309eguy speaks of has been another attempt to mitigate the problem, but I have heard it has had mixed results.

This still sounds like a condensation issue especially since you say you have cleaned the detectors many times. Can I assume that when you have cleaned them they are not very dirty? And you have looked at the glass of those detectors and it is not excessively cloudy? I assume you have probably tried moving the flame detectors from the primary zone to the secondary zone and vice versa? And that they all work properly in the primary zone?

Fefo, if you can give more information as to what other things you have done or tried, and answer some of the questions above we can all try to help you more.
 
unit one takes 40 min.
unit two takes around an hour.
unit three takes almost 15 min.
unit four takes around an hour.

if unit 3 is set to base load after short time it is usually takes a trip for loss of secondary flame (after being more than 40%). sometimes readings drops after reaching baseload so we give it around 30 min to be confident.

these 4 m/c where upgraded from standard to DLN1 with first CI
and the last unit was synched to grid a week ago.

this problem is on every start. the strange is that we clean the detector right before startup and this did nothing. I conclude that the condensation might occur during the primary fuel only period!

> somewhat dependent on the whether conditions?
weather was not studied before, but our units are around hundred meter from the sea coast !

> What type of control system is installed?
speedtronic MARK_VIE

> What brand and model of flame detectors are you using?
GE REUTER STOKES FS-9001 water cooled

> the unit should lose load very quickly, and TTRF temperature
> should go down, forcing the unit back into primary mode.

> In your case it does not sound like this is happening.

no we load up to 75 MW which is the limit of the Lean-lean positive
without any flame detected on the secondary and the load is stable and there are no alarms.

> {{{{I am assuming in your case that when the unit transitions into lean-lean mode that the megawatt output is increasing
> as it should and flame is actually present in the secondary zone, but the flame detectors are not reading it correctly
> due to some issue. I assume that you continue to raise unit load till you are near the TTRF temperature when the unit
> would be starting a lean-lean to premix transfer and you hold there until secondary flame detector values increase to
> the value needed for the control system to be satisfied that there is flame. Once that occurs then you can raise load
> further, where the unit will transfer from lean-lean to premix steady state. }}}}}

this is the right scenario we do every start up.

we take the flame detector and just point it out to the opened door and it reads correct reading (or by using a lighter).

thanks very much for your contributions
 
Adel_s and FEFO,

<b>We need to know what the ambient conditions are at the site(S).</b> Is it humid? Is it hot and humid?

<b>Is the inlet air cooled or chilled by some method?</b> Evaporative coolers? Chillers? Foggers? Wet compression?

<b>Without more information</b> (as requested) I'm more and more convinced that there is moisture condensation on the flame detector lenses. It's the only thing that makes sense. I've been on sites where the flame detectors have been removed (once it's cool enough to get into the turbine compartment) and there is no moisture--but that's because the unit has had time to cool (coastdown and whatever time the Danger Rangers required before entering the turbine compartment), the air flow through the turbine is GREATLY reduced (the IGVs are closed; the unit is on ratchet or slowroll), and the flame detectors have time to warm (from the ambient in the turbine compartment and from the radiant heat of the combustors after shutdown)--all of which serve to cause any moisture which has condensed on the lenses to evaporate. So, by the time the flame detectors are removed the lenses are DRY. And people just refuse to believe that moisture could have condensed while the turbine is running.

But it happens--frequently. In the past (before GE Belfort started building most of the Frame 9Es) the Cooling Water supply to the flame detectors had a small throttling valve in the tubing--the purpose of which was to throttle the flow of cooling water to the flame detectors <b>to prevent moisture condensation on the flame detector lens.</b>

As MIKEVI says, older machines used to have a rectangular "boss" on the pipe of the combustion can cover to which the secondary flame detector was bolted. The secondary flame detector was screwed onto a short length of pipe which was welded to another rectangular "boss" and there was an o-ring used to seal the faces when bolted together.

This arrangement added at least six, sometimes as much as eight, inches to the distance the secondary flame detector was from the combustion can cover--and from the diffusion flame in the secondary combustion zone! This "boss" arrangement also frequently resulted in an angular misalignment which made detection of secondary flame even more difficult.

In addition to this, the added length cause the secondary flame detector to be more "insulated" from the heat of the combustor (just because it was so far away from the combustor). And, too much cooling water flow could make it so cold (we know--it doesn't seem possible!) that ambient moisture would condense on the lens and block or obscure flame detection. However, after some time, when the cooling water system warmed up, and the surrounding combustors and components in the turbine compartment warmed up and eventually the flame detectors also warmed up. <b>But in the meantime, they would be affected by moisture condensation on the lens.</b>

One of the things which was done to both improve secondary flame detection and intensity of detection was to eliminate the rectangular "bosses" completely, thread the end of the secondary flame detector tubes attached to the combustion can cover, and screw the secondary flame detectors directly onto the pipe. This worked very well for preventing angular misalignment, improving detection (intensity)--and for "heating" the flame detectors and improving their ability to resist moisture condensation. It didn't always eliminate moisture condensation, but it certainly helped.

And, once the site personnel were made aware of the presence of the cooling water throttling valves in the cooling water supply lines to the flame detectors and they were properly throttled (usually about 1/3 to 1/2 open was all that was required) moist of the secondary flame detection moisture condensation issues were almost completely eliminated.

Now, there are some sites which have extremely high humidity, or which use inlet air cooling, which still have problems with moisture. Perhaps the "nut" device helps with that--but if there's too much cooling water flowing to the flame detector bodies and there's a lot of ambient humidity or added humidity because of inlet cooling then the site has to be mindful of the possibility and take appropriate action as necessary.

This is, as MIKEVI says, <b>NOT NORMAL</b>, but <b>based on the information provided</b> the most likely cause of the problem (which eventually clears up because the moisture eventually evaporates as the turbine compartment, cooling water, and secondary flame detectors warm up over time) is moisture condensation. <b>Even if it can't be seen or detected when the secondary flame detectors are removed after the turbine is shut down and the Danger Rangers deem it safe to enter the turbine compartment to remove the secondary flame detectors.</b> Again, the secondary flame detectors have had time to warm up; the air flow through the turbine is GREATLY reduced; and the moisture can--and almost always does--evaporate by the time they can be removed and inspected.

So, you've been given several things which can be investigated and implemented <b>again--based on the information provided</b> to prevent moisture condensation on the secondary flame detector lenses. It's fairly common, especially at sites which have high ambient humidity, and which have very effective cooling water systems (which is another thing which can be investigated!), or which use some type of inlet air cooling. If there are no throttling valves in the cooling water supply lines to the flame detectors, perhaps it's necessary to install some (and not quarter-turn butterfly valves--they are not flow control valves!). The "nut" device may also be helpful.

But, without more information (that has been requested) the most likely cause of the problem at both sites is: moisture condensation on the secondary flame detector lenses. It's also explains why you're not complaining about loss of load when switching to Lean-Lean mode. And, I believe the Speedtronic doesn't even "care" if secondary flame is present during Lean-Lean operation <b>when loading the unit.</b> But, it certainly does care when switching to Premix.

And, it also explains why the problem doesn't occur when the unit is shutting down and switching from Premix to Lean-Lean mode--because the secondary flame detectors are warm and moisture can't condense as easily because the unit has been running for some time.

So, all the evidence <b>provided</b> points to moisture condensation--even if you can't see it when you remove he secondary flame detectors! And, we've explained why that happens, too. And, the fact that Adel_S reports that the same issue exists on four units at his site is even more indication that it's moisture on the lenses. It's not just happening on one unit--it's happening on all the units, which, presumably ingest the same ambient air and operate under the same basic conditions.

Please write back to provide the information requested--and to let us know how you fare in resolving this problem.
 
Dear FEFO,

thank you very much for providing the information. I am never 100% sure of anything, but in this case I am almost positive your are dealing with a situation of condensation accumulating on the "eyes" of the secondary flame detectors due to high humidity and cooling for the flame detectors. This will definitely occur during primary operation.

Now what to do about it is the question. If you are still under some sort of warranty or LTSA from the OEM or another service provider then you will most likely need to work with them towards a resolution.

If not then I would start by looking at some of the suggestions from CSA. If your flame detectors are already using the short mounts then there is nothing to change there. I would suggest starting some testing with one flame detector, either installing a valve to throttle cooling water or shut it off for a single secondary detector and seeing if this eliminates the issue. I have seen sites that have removed cooling coils for all secondary detectors and have not experienced any failures due to heat. But understand you do this with some risk.

Other sites I have been at are using some other vendors that have the detector mounted remotely. This way there is no cooling needed at the fuel nozzle.

Please let us know how you proceed, but I feel you are on the right track, now you need to work to find the best cure for your site.
 
FEFO/Adel_S,

For further evidence that this problem is most likely related to moisture condensation on the secondary flame detector lenses if there were no diffusion flame in the secondary combustion zone when the unit switched to Lean-Lean mode from Primary mode the load would drop. Load is directly proportional to fuel flow and fuel combustion. If the load is stable and the fuel flow is stable when the switch has occurred, then there IS diffusion flame--even if the flame detectors can't "see" it.

Typically, approximately 50% of the fuel flow is shifted to the secondary fuel nozzles/secondary combustion zone when switching from Primary to Lean-Lean modes. This can be observed--and trended--by monitoring the gas fuel control valve position(s). If the control valve positions change when switching from Primary to Lean-Lean, and are then steady/stable AND if load is stable then there must be diffusion flame even if the secondary flame detectors aren't seeing it. If there were no diffusion flame then the load would decrease because of the lack of combustion.

Further, if the fuel valves are stable after the switch from Primary to Lean-Lean and there were no diffusion flame only in the combustors with secondary flame detectors the exhaust temperature spreads would increase--probably greatly. And this isn't being reported.

And, the fact that over time the secondary flame intensity "improves" without any exhaust temperature spreads or load decrease is just an indication that the heat in the turbine compartment, possibly combined with the warming of the cooling water in the system, is causing the moisture on the secondary flame detector lenses to slowly dry.

Now--the next question will be: Why does each turbine take different times for the moisture to dry? The most likely answer is that for some reason the cooling water flows to the secondary flame detectors on the four units is not the same. We don't know if the cooling water is in a self-contained system for each unit, or comes from a site cooling water system. There may be other reasons, like different air flows in the turbine compartments of the four units. I've been to sites with multiple turbines that had widely varying problems just due to the locations of the turbines. Sometimes there is a heavily-traveled road close to the inlet of one of the turbines, or the mist from a cooling water tower blows into one or two of the turbines' inlets and doesn't blow into the other units. Sometimes, one unit is located closest to a body of water (ocean; sea; lake) while the other units are further away.

Compressor cleanliness (which affects air flow through the axial compressor) can also have an effect.

But, again, <b>based on the information provided and the design of the combustors and the nature of DLN-I systems and historical operating experiences</b>--and the fact that another site is reporting similar problems on multiple units--the most likely cause is moisture condensation on the secondary flame detector lenses.

It does sound very unlikely--because of the heat which is generated in the combustors and the turbine compartments. But, it is a known condition which has been experienced, and continues to be experienced, with DLN-I combustors.

It's important to understand that while the OEMs/packagers tout DLN-I as a "mature technology", it's still very much a work in progress. It's a proven, viable technology--but like most technologies it has it's quirks and idiosyncrasies. The OEMs/packagers are continually working on improvements and enhancements--and this problem (moisture condensation on secondary flame detector lenses--it's one of the issues which continues to plague the technology.

Please, do not discount this issue as the likely cause even if you can't find evidence of moisture when the secondary flame detectors can be removed.

Here's a test you can perform. After the turbine has been shut down and has cooled to near ambient temperatures (monitoring the wheelspace temperatures), turn on the cooling water system for half an hour or longer. Then CRANK the turbine, or select FIRE mode and establish flame in the primary combustion zone for 5-15 minutes. Then, shut the unit down, and remove one or more secondary flame detectors as quickly as can be safely done. (It shouldn't be too hot if the unit is FIREd for a short period of time.) If you find moisture on the lenses--you've confirmed that it does indeed occur. If the ambient is humid and the unit is just CRANKed there might be enough air flow through the unit with cooling water flowing to condense moisture on the secondary flame detector lenses. I wouldn't be surprised if you ran the cooling water system for half an hour or so with the unit on COOLDOWN and at or near ambient temperature and then removed one or more of the secondary flame detectors that you would find moisture on the lenses. (In other words, it might be present on the secondary flame detectors lenses PRIOR to starting/firing the turbine.)

Please write back to let us know what you find.
 
They took air from Compressor discharge PA3-1 and injected this air in the tube that carry the flame detector through a T-connection (part7) to move the static air there to avoid condensation

Regards

> It's not very clear what this means, can you post a bit more of an explanation?
 
Hello CSA,

The ambient conditions of the site is 27 degrees Celsius and the humidity is around 70% most of the time.

The inlet air is cooled by Evaporative cooler.

Bob Johnston,

We found a documentation "GE 9E maintenance manual" for a turbine installed in Qatar which shows a modification of the flame monitoring system titled "Secondary Flame Detector Warm up System." It takes some air from the compressor discharge and injects it just in front of the flame detector's lens into the combustion chamber.
 
Top