UV Flame Detector Working - Honeywell -LG1093AA

K

Thread Starter

kichaa

Dear sir,

I would like to know my interpretation regarding the working of Flame scanner installed in GE Gas turbine with Mark VIe&es control system is correct or not.

Flame detector:

1. Sensor has two electrodes Anode in the center and cathode itself surrounded as wall and sensor is filled with inert gas.

2. Very high DC voltage around 330 volt will be across the anode and cathode electrodes.

3. Whenever the flame occurs on the combustion chamber the radiation strikes the sensor and the inert gas inside the sensor with electrodes creates or ionizes positively charged ions and electron pairs.

4. So that positive ions attracted to cathode and electrons to anode.(Electric field built s up)

5. Thus near the anode which gain sufficient energy can ionize additional gas molecules. (the region which is ionized with ions and electrons is called avalanche region).

6. Thus by above 5 th step the gas multiplication effect will take place with respect to the flame UV ray intensity.

7. So now the current conduction will occur through the medium of IONIZED INERT GAS.

8. Thus the current pulse rate will change accordingly with respective to flame intensity.

And at last i have question, and i want to know how the reading in the system is showing in HERTZ.
 
The type of flame detector (sensor) you are describing is known as a Geiger-Muller flame detector (sensor). There is lot written about it on many places on the World Wide Web. Conduction is a function of intensity, and when conduction takes place, it's essentially a short circuit, which reduces the voltage in the circuit to near zero volts for a very brief period of time.

An oscilloscope would show very clearly how the output waveform can be interpreted as a frequency. It is a square wave, the frequency of which is a function of conduction which reduces the voltage to zero before it rises again to 335 VDC.

If you're experiencing problems with flame detection, if you would describe the problems we might be able to provide some help or assistance. UV flame detectors are very old, proven technology--the only real "problem" with them is that they require such a high voltage to operate, and some pretty sophisticated circuitry to measure the frequency (though that circuitry has been industrialized and made very robust over the years).

Newer flame sensors use 24 VDC/4-20 mA circuitry, which tends to reduce the fear level of Danger Rangers (safety managers) and untrained technicians when working with them. It also reduces the circuitry required at the monitoring end of the circuit, since intensity is measured by the amount of current flowing in the circuit. The electronics are all self-contained in the sensor.

There are other flame sensors in use which use fiber optic cables to "look" at the flame in the combustor, and conduct the UV light to electronics modules outside the turbine compartment that are then connected to the turbine control panel. These have their own advantages, and disadvantages.

But, again, if you're having problems with the legacy, Geiger-Muller, UV flame detectors please try to describe the problems and we can try to help. If you're trying to develop an interface for a different type of sensor and GE Mark V Speedtronic turbine control panels, we're not good at that kind of help on this forum. It's been done with varying degrees of success, but because of the way the Mark V was designed flame detection <b>MUST</b> be connected to the inputs designed for Geiger-Muller UV flame detectors which will provide 335 VDC and needs to measure a frequency with a zero-based axis in order to provide a flame indication for the Mark V. It's a very fundamental design and programming characteristic that requires flame detectors/sensors be connected to these inputs. The control signal database is only configured for the flame sensors to be connected to these input points, and they must be connected to these input points.

Hope this helps!
 
I started wondering, where is this post going? But it's a good question.

My own thought, but could also be wrong.

There are voltage/current threshold levels for flame intensities.

Because the flame intensity fluctuates within these levels, a measured frequency is possible.

> And at last i have question, and i want to know how the reading in the system is showing in HERTZ.
 
We have fixed the new sensor in the combustion chamber. Before fixing we have tested the sensor with UV torch at that time it was reading around 75 Hertz. We thought it's OK with the real flame intensity when combustion occurs it will give more intensity thought and fixed. But flame detector intensity more or less same. So what might be the problem?. We fixed two one was reading 45 hertz and 75 hertz before and after fixing.

And I also want to know how to configure spare terminal for flame scanner in Mark VI control system database. Sorry for asking so many questions.

And if the intensity is for ex 185 hertz then do its fluctuating that much time in a second. 0 voltage to 335 approx.
 
Is this a retrofit, or new commissioning?

I don't remember seeing that voltage level sensor on a MKVIe.

Look into the scan rate settings (ms/s) of the logic.

> And if the intensity is for ex 185 hertz then do its fluctuating that much time in a
> second. 0 voltage to 335 approx.
 
Honeywell published a textbook on its industrial flame safeguard controls. The text below is the part of section on UV flame detectors.

Flame Safeguard Controls: A Honeywell Textbook 71-97588 (1989)

pages 95, 96

Ultraviolet UV Flame Detectors

Operation

Ultraviolet flame detection systems depend on the ability of the sensing tube to respond to ultraviolet radiation and remain insensitive to radiation in the infrared and visible light ranges.

The UV sensing tube consists of 2 electrodes sealed in a gas-filled quartz envelope. (Normal glass blocks UV radiation, so the tube must be constructed of a special quartz type glass.)

http://i61.tinypic.com/21mgq2u.jpg

Fig 20. shows a drawing of the anode and cathode arrangement of the sensing tube in the MiniPeeper Ultraviolet Flame Detectors (C7072A, C7035A, and C7044A). The sensing tube in the Purple Peeper Ultraviolet Flame Detectors (C7012A, C, E and F) is constructed differently, but is operating principle is similar.

The figure shown is simply a schematic representation of tube construction. The actual power tube sensor has 4 pin construction to provide increased resistance to vibration.

When a high enough voltage is applied across the electrodes and the tube is exposed to a UV source, the cathode emits electrons, which izonize the gas in the tube. When the gas fill is ionized, the tube becomes conductive and current flows through the tube. We say that the tube "fires" - since the tube glows visibly while conducting. The appearance of the tube when it fires is a function of the gas fill used. Honeywell sensors fire with a reddish glow. Some tubes constructed by other manufacturers have a blue glow which is not as apparent.

When the tube conducts, or fires, the voltage potential between the electrodes drops sharply. When it has dropped sufficiently, it is no longer high enough to cause electrons to be emitted from the cathode. The tube restores itself to the nonionized state. During this phase, we say the tube is "quenched."

We can summarize the operation of the UV tube as follows:

"Any time that the UV tube is exposed to sufficient ultraviolet radiation (or other radiation of higher energy which is capable of penetrating the tube envelope - see page 13), and sufficient voltage potential is applied across the electrodes of the tube (from a properly designed electronic network), the tube intermittently fires (becomes conductive) and is quenched (restores itself to the ready condition)."

These tubes do fire randomly even when not intentionally exposed to a UV radiation. The firing rate under these conditions is referred to as "background count." This random firing may be observed by looking at the face of the tube when it is powered; it is normal, and will not activate the flame detector relay under usual circumstances.

<snip>

An electronic amplifier in the flame safeguard control counts the firing rate of the sensing tube. When the rate reaches a level indicating the presence of a flame, the amplifier switches power to the flame relay in the control. The flame relay pulls in and stays in as long as the firing rate is high enough.
 
Hertz are Hertz, even for flame detectors. 185 Hz is 185 conductions (shorts) per second.

You haven't told us the type of combustors on the machine: conventional or DLN? if DLN, which perversion of DLN? And which combustion zone is the suspect flame detector(s) monitoring if DLN-I?

Have you meggared the wiring between the Mark VI and the flame detectors? If so, what was the result?

Is the unit experiencing high exhaust temperature spreads--higher than normal? Because if not, then it doesn't sound like there's a real flame intensity issue. If the exhaust temperature spreads are higher than normal, and they "move" when load and IGV angle changes, then there may be flame/combustion issues.

What fuel is the machine burning when the flame intensity is low?

Have you tried swapping flame detectors--or just the wiring at the Mark VI--to see what happens? If so, what happened?

Have you measured the voltages (using a True AC RMS voltmeter of a working vs. non-working flame detector--presuming there aren't high exhaust temperature spreads? If so, what were the results?

I don't have a Mark VI Toolbox .m6b file right now to be able to tell you how to enable a spare flame detector input; sorry. But you should be able to look in Toolbox and see how an existing one is configured and then look at a spare one and see what needs to be done; you can ask questions here if you need help after you've looked at the configuration(s).

I have seen combustion liners with the wrong flame detectors openings installed. Also, I've seen liners without flame detector openings installed in the wrong combustors. Did this problem start after a maintenance outage, or a trip from high load?

Please provide the answers to all the questions; it will be most helpful if you do. When posting to a forum like this for help, it's most helpful if you tell us what is/isn't happening, what you've done to troubleshoot the problem, AND what the results of your troubleshooting were--in the original post, And, tell us as much about the machine and fuel(s) as you can; we're not there with you and we don't know and can't see what see. So, the more information you can provide in the original post the more concise the responses will be--and likely faster, too.

Hope this helps!
 
Thanks for your info. could you please send me a link to download the
Textbook of honeywell. and i cannot able to see image in the link you posted.
 
I can't post a copy or a link to it, it's not in an electronic format. I typed the two pages that were relevant to the discussion from the paper copy on my bookshelf. The book is over 25 years old and was published in that era before electronic (pdf) files were common formats. Occasionally used paper copies show up on used book sites. I am unaware of Honeywell publishing the book in any electronic format.

Although long URL have a space or two inserted into them on this forum, I've never seen that happen with a short tinypic URL. Here's the same image on a different image hosting site:

 
Dear CSA,

Sorry for late reply.

Combustor Type: DLN.

Exhaust temperature spreads are normal.

IGV operation is normal.

Fuel: Natural Gas-(Methane @ temperature around 55 Deg C)

We didn't swapped and checked the flame detectors.

Now i am clear about the operation of flame detector from you peoples (Marcopolo, David and CSA). And regarding the fault as i told before. Just in my assumption do any chance of carbon particles just settled and blocking the actual path between the sensor and flaming area.

And just for your info we have four Flame sensors.

Two is normal mean one showing around 365 hertz and another 125 hertz. but other two are problematic one is reading around 60 hertz and 35 hertz. - (The problematic sensor which i am mean is the new sensors which we fixed in last turbine shutdown time).

Now turbine is running with two sensors which is reading acceptable values.
 
> carbon particles just settled and blocking the actual path between the sensor and flaming area.

Dirt, oil or soot on the quartz (or special glass for UV transmission) lens of the UV sensor will attenuate (decrease, block) the UV passing through the lens.

From the Honeywell text book:
"Ultraviolet radiation from the flame is absorbed by dirt and dust, oil vapors, and other contaminants, and byproducts of combustion in the combustion chamber. <snip> problems due to UV absorption do not generally occur in [natural] gas [methane] burners." (my [additions]).

The problem of UV absorption refers to other fuels and the 'screening effect' that the use of other fuels might have regarding UV absorption.

Yes, check the tube lens for cleanliness.
 
kishore,

Okay; I'm guessing DLN-2, or some perversion of DLN-2, and that probably means a Frame 9, F-class. Since these are UV flame detectors I'm guessing the unit is older, since most newer turbines use the 4-20 mA Reuter-Stokes SiC Flame Trakkers (or Trackers) not the 335 VDC UV flame detectors.

As for the carbon particles, they would have to come from the inside of the flame detector; they are sealed pretty well because of the high pressure in the combustor. Not very likely--unless they weren't handled with the proper amount of care during assembly/disassembly.

There have been lots of problems with flame detectors getting too much cooling water flow, causing moisture in the ambient air (humidity or from, say, an evaporative cooler or fogger) to condense on the lens. This is a huge problem, and while it usually affects all flame detectors, sometimes it doesn't--because usually there are cooling water flow control valves for each flame detector's cooling water coil, or for groups of coils.

Moisture condensation is a tough one to prove, because usually by the time the unit gets shut down and it's cool enough for someone to enter the turbine compartment to remove the flame detector the moisture has already evaporated. Have a look at the Cooling Water System P&IDs and the actual piping to the flame detectors to determine if throttling valves were used, and if so, if they are properly adjusted. I've even seen throttling valves have to be replaced--or installed--to correct this problem. Really humid environments or evaporative coolers/foggers can be very problematic if too much cooling water flows to one or more flame detectors.

Also, air gets trapped in the cooling water piping of these flame detectors and that can cause problems, also. Be sure all of the air is bled out of the cooling water tubing/coils.

Another problem which has occurred from time to time is that the tube the flame detector is mounted on gets bent during maintenance outages, causing the angle to be wrong for the flame detector to see the diffusion flame "ball."

There have also been instances where the combustion liner has not had the hole through which the flame detector must "look" to be in the wrong location.

Problems with interconnecting wiring have also resulted in low intensity readings.

But, if there are no high spreads that means there are no combustion problems--that the flame intensity in all combustors should be relatively the same. Since there are multiple fuel nozzles the flame detectors are generally "aimed" at one of the fuel nozzles, and there may actually be some problem with that/those nozzle(s). There's not usually much fuel being burned in the diffusion flame tips of DLN-2 machines, especially considering there are, again, multiple nozzles per combustor.

So, you have several things to check out. None of them are too difficult, and I never recommend the shotgun approach to troubleshooting--trying several things at once. Because of the problem gets solved it can be very difficult to determine what the cause/resolution was. And, since many Plant Managers only care about solving the problem as quickly as possible they don't want to try un-shotgunning things to determine the actual cause/resolution, and that leads to problems down the road--lots of problems, usually. Best to take a logical, methodical approach to eliminate things one by one until the root cause is found.
 
As part of a project I would like to replace LG1093AA26 flame detectors with flame detectors model ITS flame scanner 184X0254M029 with a 4-20mA signal supplied with 24 VDC ;my system is mark VIe from GE.
Those new ITS flame scanner will be installed in compartments(auxilliary,turbine and coupling).
Please tell me if I can use these detectors in these compartments?
 
Ngassaki,

The LG1093AA26 is a flame detector--which is used to detect the presence of diffusion flame in a combustor (such as when the turbine is burning fuel in a diffusion flame). Combustors are located only in the turbine compartment of GE-design heavy duty gas turbines.

I suggest you refer to the website flamescannet.net for application information and specifications. I don't know if these ITS flame scanners require cooling water jackets or not. Also, you can contact the manufacturer/supplier for help with application information.

The Mark VIe can power and accept signals from many different types of 4-20 mA devices, so that aspect should not be a problem (connecting the flame scanner to unused 4-20 mA inputs of the Mark VIe). What will mostly likely require some skilled and knowledgeable assistance is the reconfiguring of the application code in the Mark VIe to use the 4-20 mA flame indications in place of the Honeywell flame signals. And, I'm CERTAIN ITS can assist you with that if you purchase their flame scanners.

Best of luck--let us know how this works for you!
 
CSA,

Thank you for this return.
I will get back to you on the subject.



Ngassaki,

The LG1093AA26 is a flame detector--which is used to detect the presence of diffusion flame in a combustor (such as when the turbine is burning fuel in a diffusion flame). Combustors are located only in the turbine compartment of GE-design heavy duty gas turbines.

I suggest you refer to the website flamescannet.net for application information and specifications. I don't know if these ITS flame scanners require cooling water jackets or not. Also, you can contact the manufacturer/supplier for help with application information.

The Mark VIe can power and accept signals from many different types of 4-20 mA devices, so that aspect should not be a problem (connecting the flame scanner to unused 4-20 mA inputs of the Mark VIe). What will mostly likely require some skilled and knowledgeable assistance is the reconfiguring of the application code in the Mark VIe to use the 4-20 mA flame indications in place of the Honeywell flame signals. And, I'm CERTAIN ITS can assist you with that if you purchase their flame scanners.

Best of luck--let us know how this works for you!
 
Ngassaki,

Methinks you are possibly confusing fire detectors with flame detectors.?.?.?

In GE-speak, a fire detector is a device that senses intense heat (which would be indicative of a fire) where there should be none, such as in the Accessory Compartment, the Load Compartment, and even the Turbine Compartment. A flame detector is a device that senses the presence of UV flame in the turbine combustors--where there should be a flame when the turbine is combusting (burning) hydrocarbon-based fuel. The newer GE flame detectors, made by Reuter-Stokes (which I don't think is a GE company any more since the divestiture of Baker-Hughes) are call flame "trakkers."

Some packagers of GE-design heavy duty gas turbines do provide UV flame "scanners" to detect the flames of a fire where there should be none (such as in the Accessory, Load and Turbine Compartments (outside of the combustors, that is)). Those are usually connected to a stand-alone fire or fire & gas system. They are not usually connected to the turbine control, because they (the fire or fire & gas systems) need to operate even if the turbine control system is incapable of operation.

Isn't the English language wonderful? (Most of the time.)
 
Top