NOx injection in Gas Turbine

A

Thread Starter

Awaneesh

We have recently commissioned steam injection for reducing Nox emissions in our GE Frame V gas turbines. But we are facing the problem of 'Flame detector trouble' in one of comb chambers (#3) due to loss of feedback of flame detectors.

Also, we have checked/replaced the flame detectors due to their possible failures and also cleaned the, for probability of carbon deposition inside. There is no visible variation in exhaust thermocouples near that Combustion Chamber.

Kindly suggest what could be the probable cause?

Thanx in advance
 
Awaneesh,

Hmmm.....

Are the flame detectors cooled with water (coils placed around the flame detector bodies)? If so, it sounds like the cooling water flow for at least that one flame detector is too high, causing moisture to condense on the lens. By the time you shut down to check/clean the lens, it's evaporated.

But, one would think if one water-cooled flame detector is having a problem, then all of them would be--unless the throttling valves are not all set properly or some of them have air trapped in the coils. There are usually throttling valves which must be field-adjusted--just for this reason, to prevent excessive cooling causing moisture condensation on the lens. Sometimes ambient humidity is so high that excessive cooling can cause condensation on the flame detector lens--though that's not likely in this case; just mentioning it because it happens on occasion.

If the flame detector is not water cooled, then things like this can happen if the steam piping was not properly blown down ("flushed") to remove dirt and scale prior to admitting steam for the first time. The spray nozzles can be plugged with small particles of dirt causing an uneven spray pattern, but one would think the exhaust temperature spreads would be uneven.

You say the exhaust temperatures "...for that can..." are not unusual. Combustion gas flow through the turbine is not usually directly straight through the machine, so, for example, Exh. T/C 1 is not reading the flow from Combustor 1. There is what's called "swirl angle" and it changes with IGV angle, and load (mass flow through the unit). GE publishes swirl angle charts for most machines, but not all. Frame 5s being the workhorses of the GE fleet, I would imagine there are swirl charts for your machine--if you ask GE, or whomever supplied the steam injection system.

What is the maximum spread when running NOx Steam Injection? Is the "cold spot" or "hot spot" approximately 90 degrees from Combustor 3?, against rotation?

It's really hard to imagine an even distribution of steam in all the cans and yet one can with plugged NOx Steam Injection spray nozzle(s) causing the steam flow to be so high in the area of the flame detector sight tube that it blocks the flame.

NOx steam is supposed to have a minimum of 50 deg F of superheat, so it doesn't "flash" when injected into the combustor. That's why steam is less efficient at reducing NOx (more mass flow of steam is required for a particular emissions reduction that for steam for the same emissions reduction)--because it's not as effective at quenching the flame temperature. So, it shouldn't be flashing.

Is the #3 combustor near where the steam enters the steam injection manifold? Could it actually be that flame is being lost in that can because a slug of water makes it through the steam injection piping into the manifold, and the #3 combustor is closest to where the steam line connects to the manifold around the compressor casing? I have seen this happen--but, there is usually a very high spread when this happens.

Does the loss of flame detector intensity occur immediately after NOx steam injection is started? Or some time after; if so, how much time? Does flame intensity decrease over time, or does it just go to zero?

This is very odd--that's there no exhaust temperature spread, or a small spread, when this happens. Either there's too much steam going into one combustor (which should cause a cold spot exhaust temperature spread) and blocking the flame detector, or there's an uneven flow of steam into one combustor because one or more of the spray nozzles on the fuel nozzle is plugged resulting in too much steam being injected in the area of the flame detector sight tube opening, which would mean that there is probably less steam flowing into that combustor as compared to the others--which should result in a hot spot (exhaust temperature spread).

Or, perhaps one or more of the NOx steam Injection spray nozzles in the area of the flame detector sight tube opening has too large of an opening resulting in excessive steam flow in that combustor blocking the flame detector sight tube. But it would still seem likely there would be an elevated spread, in this case a cold spot.

You could try swapping the fuel nozzle of the #3 combustor with one in a combustor that doesn't have a flame detector, and see what happens?

Are you certain the combustion liner is properly positioned in the can? That the flame detector sigh tube is not being partially blocked by the can not being properly held in place?

Is the flame detector sight tube for the #3 combustor bent such that it's not looking at the proper location in the combustor and when steam is flowing the flame moves a little farther downstream and is being "missed" by the flame detector?

Something just doesn't add up; but there are some things to try and investigate.

Please write back to let us know how you fare!
 
Top