Can a Thermometer Be Engineered to Detect Fires at High Temperatures?

[T.O.E Dream]

Is there a way that I could create a fire detecting thermometer that would detect a fire if the temperature went above a certain point? For example, if a thermometer went above 200 degrees C this would signal a fire? How would I create this?

 

[cronxeh]

Hmm..
1) How long would it take for ambient temperature to affect the thermometer
2) What area of the room would be under fire after the temperature reached the threshold.
3) Would smoke reach the detector faster than the ambient heat?

Given these 3 questions I think the current smoke detector system is a lot better at detecting fires

 

[Moonbear]

If I'm not mistaken, sprinkler systems work by detecting heat. When the temperature reaches some threshold when you should assume there's a real fire burning, something in it melts to turn on the water...I think. You wouldn't want to still be in a building when the room temperature is already turning into an oven, so a temperature sensor for an alarm system isn't such a good idea. Just once the fire is enough out of control to be raising the room temperature to a certain level, it's time to dump water on everything.

 

[S_Happens]

To see where this goes I'll ignore the two popular styles of smoke detectors.

Temperature can be measured in multiple ways, so there are quite a few ways to accomplish this. Right now I'm sitting within a few feet of thousands of various temperature alarms. Can you be more specific in what you want to do or what your limitations are as far as materials?

 

[T.O.E Dream]

Let me explain more, I'm trying to detect a fire on a stove. I think maybe if I use an infrared thermometer it would be better.

 

[cronxeh]

Depending on air flow and room volume, the ambient temperature in fire will be drastically different, so using a thermometer may not be the best idea. IR camera could be useful if you create exclusion zones (for stove outlets) to be ignored, and program an alarm for all other areas which will trigger fire suppression system

This looks promising (but probably expensive)The cheapest option I can think of which will involve great deal of ingenuity is using near infrared pass filters (650nm-1050nm) on cheap CMOS pinhole cameras (1.3 megapixels and up). Could cost about $200 dollars if you get DAQ + CMOS + NIR pass filter on ebay

 

[cronxeh]

Just hook up DAQ to Matlab, play with image acquisition and histograms, set up the detection zones you interested in and program outputs for same DAQ to activate.. a fire extinguisher? Play around with it.

In case you wondering what those NIR pictures may look like..

 

[T.O.E Dream]

cronxeh, I believe that the temperature of a fire (not even a very hot one) is still dramatically hotter than room temperature. As long as we know the minimum temperature that a fire might start at, we can use an thermometer. Let's just say that the minimum temperature of a fire is 100 C (throwing numbers around). This is obviously, hotter than room temperature so anything hotter than 100 C will trigger an alarm.

 

[cronxeh]

cronxeh, I believe that the temperature of a fire (not even a very hot one) is still dramatically hotter than room temperature. As long as we know the minimum temperature that a fire might start at, we can use an thermometer. Let's just say that the minimum temperature of a fire is 100 C (throwing numbers around). This is obviously, hotter than room temperature so anything hotter than 100 C will trigger an alarm.

By the time the ambient room temperature (which is affected by air flow, distance from thermometer to fire - which you really can't predict where the fire will start at so your placement is randomly placed from fire source) - the fire will be in full blaze, it could take 5 minutes for the fire to raise the thermometer temperature to 100 deg C.

Look up pictures of toys in fire room, some plastic toys are not even melted (and plastic melts at 350 deg C)

 

[FredGarvin]

There are plenty of flame detectors used in industrial fire suppression systems available.

http://en.wikipedia.org/wiki/Flame_detector

 

[xxChrisxx]

Most University halls have fire detectors using heat rather than smoke in the kitchens.

Stops the problems of the firebrigade coming to put out some burnt toast that a drunk student has tried to cook after a night out.

IIRC it worked on a thermostat (really not sure what type), that was triggered in the high 60 degrees c ambient.

 

[ideasrule]

It should be feasible to make an IR camera watch a stove, calculate the number of pixels hotter than a certain temperature, and trigger an alarm if that number is greater than a certain limit. The stove element is guaranteed to be colder than any fire that might be of concern.

 

[Phrak]

If I'm not mistaken, sprinkler systems work by detecting heat. When the temperature reaches some threshold when you should assume there's a real fire burning, something in it melts to turn on the water...I think.

You've piqued my curiosity. In some sprinkler heads there's is a low melting point solder joint between two mechanical elements. When it melts, it let's a cap pop off the end of the sprayer. http://books.google.com/books?id=64...esnum=1&ved=0CBEQ6AEwAA#v=onepage&q=&f=false" (Some Indium-Gallium-Tin Alloys are even liquid at around room temerature.)

 

[mgb_phys]

I worked on a system once that detected fire from existing CCTV cameras by looking at the changing light levels. Flames have a very characteristic flicker frequency,
The idea was to detect fires in large warehouse/factories before the smoke/heat built up enough to trigger a smoke detector or sprinkler.

ps. Sprinklers detect heat directly, they have a low melting point plug. By the time they trigger it's pretty much too late - they are just there to try and save the structure, and give time to people in other parts of the building.

 

[T.O.E Dream]

If it helps, i want to measure the heat in one certain place. If that place gets to hot how can I sound an alarm? This assuming were using an infrared thermometer.

 

[edward]

If it helps, i want to measure the heat in one certain place. If that place gets to hot how can I sound an alarm? This assuming were using an infrared thermometer.

If you just want to measure heat in one small area any thermostat will work in a low temperature range, say up to 95 degrees f.

There are a number of off the shelf thermostats available in the heating and cooling industry.

Most forced air furnaces have a thermostatic switch built into start the fan motor at about 120 degrees f. Many of them are even adjustable. On the other hand they are not highly accurate beyond + or - five degrees or so.

As far as sounding an alarm use a door bell or furnace transformer and a bell wired through the thermostat.

 

[cronxeh]

I think another way would be to use symbolic logic and a few temperature sensors. If you spread the sensors apart at a certain distance and surround the stove area, you can create a simple algorithm for when the alarm should be triggered, given that probability that over 50% of the sensors would never show a temperature increase of +2 standard deviations (you can trigger the parameters to whatever may suit your experimental setup) unless it was a fire

 

[Phrak]

If it helps, i want to measure the heat in one certain place. If that place gets to hot how can I sound an alarm? This assuming were using an infrared thermometer.

That's a lot easier. A sensor diode can be placed remotely. Photodiode sensors detect photon radiation, as you known, right?

 

[Phrak]

Sprinklers detect heat directly, they have a low melting point plug.

No! That's what I thought. We must have heard the same wife's tale. (Did you catch it from me? ) See post 13.

 

[f95toli]

You've piqued my curiosity. In some sprinkler heads there's is a low melting point solder joint between two mechanical elements.

Indeed. And there are fire detection systems (the kind that that trigger alarms) that work by detecting heat. Where I work they are used in most of the labs where we handle cryogenic liquid since we tend to generate a lot of "smoke" (vapour from the dewars) when transferring liquid helium. We know from experience that the vapour will trigger conventional smoke detection based fire alarms so heat-detection based systems are really the only sensible option.
However, Heat detection systems are -as far as I understand less reliable- and react much more slowy than smoke detection systems which probably explains why they are not very common in normal rooms.

 

[Jimmy Snyder]

Is there a way that I could create a fire detecting thermometer that would detect a fire if the temperature went above a certain point? For example, if a thermometer went above 200 degrees C this would signal a fire? How would I create this?

When the building collapses into a pile of cinders, it signals a fire.

 

[turbo]

Instead of re-inventing the wheel, look up flame-sensors for natural-gas/propane ovens and water heaters. If there is no pilot flame, the sensor will not allow the main gas valve to open and power up the unit. Should be easy to strip one out of an old appliance and adapt the output to trigger an alarm instead of opening the main gas valve.

 

[mgb_phys]

However, Heat detection systems are -as far as I understand less reliable- and react much more slowy than smoke detection systems which probably explains why they are not very common in normal rooms.

In a regular house fire the fumes from burning foam furniture would kill you long before there was any visible flame - hence smoke detectors rather than heat detectors

ps. I once had a little accident with pressurized liquid nitrogen dewer that resulted in my spraying LN2 over all the ceiling mounted smoke detectors,followed by a visit from the fire services, then a visit from the chemical spill response team (once I had explained to the fire service that there was no fire) and then a visit from the police who were there to cordon off the chemical spill (explaining to the police who had been called to a physics lab in the middle of the night, that the chemical I had spilled was air and that it had all gone - was the tricky part)

 

[cronxeh]

(explaining to the police who had been called to a physics lab in the middle of the night, that the chemical I had spilled was air and that it had all gone - was the tricky part)

That is hilarious

 

[mgb_phys]

The police arrive an hour later, it was very late - it had been a long day
There is also a bit of a town-vs-gown thing, the locals have a distrust of the university roughly the same as the villagers in a Frankenstein movie.

Police: Where this chemical spill then ?

Me: It's ok, I opened the windows it's all gone now

Police: Looking toward the sleeping town in the distance with a worried expression- (Fortunately UK police don't carry guns and this was before they regarded everyone with a cell phone camera as a terrorist )

By some miracle I didn't get arrested,
We didn't even get into trouble with the fire people - as far as they were concerned it was a genuine chemical spill (I don't think they got to do many of them)
The real miracle was I finished the PhD!
My supervisor got told the story by everyone and assumed it must have been exaggerated in the telling!

 

[Moonbear]

(explaining to the police who had been called to a physics lab in the middle of the night, that the chemical I had spilled was air and that it had all gone - was the tricky part)

I can just imagine that one. I definitely feel for you on that, but am going to laugh anyway.

I think our animal quarters at one university had heat sensors instead of smoke detectors in the actual animal rooms, because of all the dust from bedding that would trigger conventional sensors. But, IIRC, it wasn't based on reaching any particular temperature threshold, but was set off based on a particular rate of increase in temperature, which it was also possible to set off when they emptied a room and hosed it with hot water for cleaning (but it's also possible they were conventional smoke detectors and the steam set them off). I think they started to just call ahead and notify whoever monitored the alarms whenever they were doing cleaning.

 

[T.O.E Dream]

mgb_phys you've provided a truly funny story. Sorry to inturupt it.

Let me be more specific once again. I want to detect when a fire starts in a cooking fire, more specifically, a grease fire like one that happens in a pot over the stove. I'm thinking (but not sure) maybe I can use a photo-diode. This is because a fire gives out a lot of light energy and a not so sensitive one so that a light bulb or other regular light sources don't trigger it.

Can someone give feedback to this idea? Again, I'm an idiot towards this stuff. Well, I'm an idiot in general (for example, I almost spelled "idiot" wrong before the red line appeared underneath it).

 

[edward]

There are dozens of flame sensors on the market. Do you need something that you will build yourself for a project or assignment??

http://www.appliancezone.com/ShowProduct.aspx?id=16302&AspxAutoDetectCookieSupport=1

 

[T.O.E Dream]

Do you need something that you will build yourself for a project or assignment?

Yes, sorta. I'm trying to create something for a project which requires a flame detector for cooking fires. It's one part of it. I need something really small that will detect flames.

 

[cronxeh]

Maybe its time to move this to Mechanical Engineering forum.

I think your problem boils down to thermal detector vs photon detector, and your budget. Play around with near infrared cameras, bolometers, and pyrometers.