Fire Safety Research Part 2: Optical Fire Detection

Haven’t been writing a lot about the actual work I’m doing at the moment, let’s fix this…

Working in a fire safety engineering project, I recently had the opportunity to take part in the writing of other people’s research proposals, by finding references about the technology of detecting unintentional fires from video material or infrared/visible/UV spectral measurements. As is the case with many natural phenomena and patterns, a human can intuitively recognize an accidental fire immediately when he/she sees it, but it’s a lot more difficult to teach a computer or similar apparatus to do the same, despite the ability of technical instruments to see wavelengths of light that a person cannot see. This is a result of human perception being a product of over a billion years of evolution and having developed an excellent pattern recognition ability.

On a closer look at the subject, it became apparent that devices that simply attempt to detect the abnormally quick temperature rise related to fire, are not fast enough for early-warning purposes. A smoke alarm can work better, but it doesn’t necessarily react to a fire if it happens in a space that has large openings to ambient or a very large room volume in cubic meters (think about an aircraft hangar), as the smoke has to get quite thick before an alarm is set off. Devices that detect UV radiation should be calibrated to ignore wavelengths that are emitted by electric arcs, welding torches or the Sun. An infrared detector should not react to ordinary hot objects like a heated electric stove.

Fortunately, an unintentional fire has some features that are relatively simple for a man-made device to notice. The radiation emitted by a fire from burning cellulose or artificial polymers contains infrared spectral peaks that correspond to the vibration normal modes of carbon dioxide and water vapor. The same wavelengths are not especially pronounced in the radiation from a glowing hot metal object or similar. A fire that is not intended to take place is practically always a diffusion flame, not a premixed one like that of a propane or oxyacetylene torch. Diffusion flames have an unstable flickering behavior that is not seen in premixed gas flames and can be detected by an oscillation of average pixel whiteness in a grayscale video image, taking place at frequencies of 0.5 to 20 Hertz (applying a temporal bandpass filter at that frequency range to the video data will make it apparent).

Fires can also take place in the wilderness, especially in dry and warm weather, and satellite imaging can sometimes make them noticeable before the harmful effect on the outdoor air quality of nearby cities. Burning biomass is special for containing the telltale potassium spectral lines, as all living cells have a considerable intracellular potassium concentration, which remains in the material even after drying. Most people who have studied more chemistry than the high school level, remember the flame test that gives a typical purple color for a sample of potassium ions that doesn’t at the same time contain too much impurities like sodium or calcium.

As an additional research topic, the proposal I had to find material for contained an idea of a protective suit that effectively reflects IR radiation typical for fires. As some may have heard, many surfaces that appear diffuse in visible wavelengths, start reflecting in a mirror-like way when IR radiation is directed on them (this is related to the relative length scales of the surface coarseness and the radiation wavelength). A large fire can emit enough thermal radiation to cause burns on unprotected skin even as far as 100 meters away, and even a person in a full protective suit can’t withstand flashover-stage fire conditions for a time of over a minute or so.

So, that’s what kind of stuff I’m working on at the moment. The subjects that I handled in some other posts have some connections to this material, though – for instance the harmonic and anharmonic oscillator energy states that I calculated in the imaginary time Schrödinger post, have connections to the detailed IR spectral line positions of gases such as carbon dioxide and water vapor.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s