Analog Electronics Design Project -

A particular blast furnace uses natural gas powered burners for heating. A problem exists in that, if the flame was to extinguish, the gas will continue to fill the chamber and be a major explosive hazard. Consequently an electronic flame-out detector is required, that will sound a warning soon after flame-out.

The problem is made more complex, in that the detector should take into account for momentary flutters in the flame. This could be due to either gas pressure changes or from the flame being blown away from the detector.

A flame has a number of properties that can be used for detecting the presence of flame. The first is the heat from the flame. However, the furnace could already be at a certain temperature such that the cooling from flame-out will not be detected soon enough.

Another property of the flame, is that it has the capability of supporting a current. That is, the flame has a low resistance and thus can be utilised as a closed switch, which opens on flame out.

Any form of help will be greatly appreciated.Thnx

 

Find them. Search for possible detector to detect these properties.

Construct experiment to find out how much current flows across two electrodes when placed inside a burning flame. How much voltage is needed to pass sufficient current? Then design and build a suitable detector/detecting circuit.

__________________
L.Chung

 

Electrical conductivity detectors are standard fare in laboratories, namely in gas-liquid chromatographs. There are numerous manufacturers; although, I am not sure what would be available in Australia. But in my experience, the circuits for the detectors were available in the service manuals and are simple. Typically, the gas jet is one electrode and a small cylinder above or in the flame is the other electrode. They are quite sensitive to impurities in the flame, hence their use as detectors in chromatographs. However, since you are interested in basically and on/off application, that should not be a major problem. Instability of the flame might be a problem, as it would be with anything that did not have a large thermal mass; however, various damping techniques and positioning of the electrode(s) should be able to address that problem.

A different alternative would be to look at the spectrum of the flame. For example, the emission spectrum (or absorption spectrum) should be quite different than the background thermal emissions. For leads on that method, I suggest looking into atomic absorption or atomic emission spectroscopy. Both of those methods would be more complex than simple electrical conductivity.

And last, one might monitor CO2, CO, or H20 production, or unburned fuel gas by IR absorption or other specific methods, including simple mass spec. Again, the flame would produce those gases, but the background heat might not, unless of course, one were monitoring near the molten metal. John

 

A few thoughts:

A. Take a look at equipment that is commercially available to see what they do.

B. The sound or noise of a burner is quite distinct - it would seem that a system that could "listen" for the roar - or the not roar.

C. The color of light from a gas flame might be different than the surrounding hot surfaces in the moment following flameout. A photocell sensitive to the color of the gas flame might be useful.

__________________
stevez

 

ron787:
good topic, thanks. This reminds me that I need to buy that $14.00 digital camera selling at Radio Shack.
I read in a magazine that if you remove the filter in the front of the light detector and insert a dark film in between such as a negative you can turn that camera into an infra red camera, therefore detecting the absence of heat , if you place a peice of wire or something to absorb the heat and kind of heat charge it like a capacitor that will help detecting heated surfaces.
Am just dreamming here, i know is a wild untested idea ... for $14 bucks ...
rs1

 

The proper spec. thermocouple and a thermal-switch wouldn't work for your needs?

__________________
Don't make me reach through this monitor to slap you a good one!

 

You could use a conductivity sensor like jpanhalt is talking about and add a low-pass filer to eliminate the effect of the fluctuation of the flame.

__________________
I also post at the following sites:
http://www.stop-microsoft.org http://www.heated-debates.com
Screen name: Aloone_Jonez
And http://www.silicontronics.com, same screen name as here.

 

Isn't this just a school homework question anyway?, and not an actual requirement?.

__________________
PIC programmer software, and PIC Tutorials at:
http://www.winpicprog.co.uk

 

It is for the jet plane he flies to school each day. But he keeps running out of fuel and needs the warning.

__________________
Uncle $crooge

 

There are actual flame sensors - very short wave UV sensors that ignore pretty everything except combustion (i.e. sensitive enough to see a match a couple meters away or something).

http://www.electro-tech-online.com/e...t=flame+sensor

 

Honeywell makes industrial flame detectors for boilers, furnaces, etc. They're sensitive to the ultraviolet energy from the flame and commonly called "purple peepers".
JB

__________________
Nothing is impossible for the man who doesn't have to do it himself - Weiler's Law

 

I have no experience in the Hamamatsu UVTRON but noting the rerquirement of the high voltage supply I have a feeling of going back to the vacuum tube time !

If selective detection of UV at wavelengths 185 to 260 nm can do the job, how about using a UV photodiode such as JEC 0.1 from Electro Optical Components Inc, or SG01S-C from Scitec, and many others.

I have just assembled recently a UV meter and the circuit has been posted here:

http://www.electro-tech-online.com/e...n-meter-2.html

The photodiode OS100 I have used covers all UVA, UVB and UVC and is therefore for flame detection purpose I would think that it needs to be replaced by any of the above-mentioned.

With regard to the circuit, the output from the Op Amp TL081 can be followed by a voltage comparator which can trigger an alarm circuit if so required for announcing the lost of flame.

It would be an interesting project !

__________________
I am NOT a llama !

 

I think the main advantage of the UV flame sensors is that they're insensitive to everything except "far" UV. It claims to not see any of the wavelengths in sunlight.

Building a HV source is a bit of a headache, but there are a handful of boost converter chips which are capable of 50+V. Adding a diode multiplier to one of them would make for a very compact power supply.

[err typo]

 

The earth's atmosphere ozone layer is full of holes now so maybe it can detect far UV from "the ball of fire" when one of those holes is over it.

__________________
Uncle $crooge

 

I think the gas which blocks out the far UV is supposed to be oxygen, so we'd probably have some *REALLY* sincere issues if one of these things works outdoors.