I am not sure is any flame a fluid in a sense you ascribe by way you have posed the problem. Never seen anything like this done to date an I conduct fire-fighting experiments. Large fires you are looking at makes this measurement a mission impossible, I believe. See if you could introduce radioactive compounds into your flame and track these somehow. Please keep0 me posted, I have an interest in this.

I thank you for your contribution. I believe that opportunities may exist by using double thermocouple probes but a quantitative study of their time response must be performd to ensure their ability to capture fast fluctuations in large scale fire flames.

Xavier, I am not quite sure how any doubling of thermocouples could take care of the influence of the flame temperature fluctuations which are there owing to the fact the flame is turbulent? It all reverts to the issue you would like to look at the flame structure as it was a turbulent flow. What about density, if you insist on a flow-like interpretation? Come to think of it, shouldn't you simply add something to the fuel that would emanate extra glow or radiation when drawing energy from the flame? Same as in the 'cold fire' Christmas tree ornaments. Some 'cute powder' that could be read by Geiger counter, or another detection technique?

I'm not sure to follow.

First, u seem to point out that the flame structure of a real scale fire is not turbulent.

How could it not be? Whatever the Reynolds number you built, it is large caomparing to unity.

About density, fire is commonly interpreted as a barotropic flow. Density is supposed to be linearly dependant on temperature. The TC double probe could lead to a turbulent scalar flux.

Second, you plan to capture the velocity field by mapping the glowing region of the flow. What you call glow is the thermal radiation of soot concentration. Are soot natural tracers of the flow? Not sure. The glow is a volumetric information, it integrates a whole contribution of radiant sources over a optical depth. How could it lead to a planar measurment for instance? Finally, u purpose to seed the flow with radioactive particles: what is the concentration, the flow mass rate, the size distribution of these particles ? What is their cost? Could we throw radioactive quantities away in the open like that?

That is not obvious.

Xavier, I am not sure where did you see me stating the flame structure is laminar? Interpretation of a flame, not the fire, though, as a barometric flow seems odd to me. This could be the case, I do not deny this, but how would you reconcile very turbulent variations of temperature with a constant density? It seems you, guys, are looking at the flame as it was a jet? Than there is a problem, there could be a linear fire front, but turbulent flame(s) it is made of is/are always three-dimensional. So you need tracers, of one kind or another and a 3-D mapping. When you refer to the optical length, you seem to be mixing apples and oranges (sorry, I'd rather be candid about this). Flame front is on the order of tenths of meters, viscous turbulent scale, say, micrometers.... it looks as if you have jumped from a micro-scale to the mega-scale w/o having contemplated any model to bridge the gap. Tracers would let you you have a glimpse into the general turbulent nature of the fire, or flame front, but certainly not into the minute details of a single flame. The latter you do in a lab, small scale, so other requirements on light- or radiation emitting tracers have to be met. Do you have a definition of the larger project you are working on? Could be in French, 'j'en connais assez bien. Let me think about the gap you seem to be grappling with. I could take this and check it experimentally at the Central School for Firefighting in Warsaw, Poland. Where, we started experiments on extinguishing large fires using water mist guns (l'eau atomisee, or brouiliard a l'eau), so we may have some insights to share with you coming in shortly.

Hello Xavier,

An easy technique to measure the 2D velocity of a flame is to make a high-speed video of the flame, and to apply PIV processing algorithms on it. You will obtain the velocity field of the flame, but not of the surroundings, but it is really easy and you get already nice results really fast.

Ah, bon? This is excellent news to me too. Experimental insights are priceless. Now the tough question, how big a flame could you process using this PIV contraption of yours? Xavier is interested in fires, rater than flames. This is why I have objected. One flame engulfed in a huge fire means huge convection eddies on top of the turbulent-scale eddies. It would be hard to decipher 2-D flame flow, from an overlaid 3-D structure of the fire. But here I am but a pupil and an eager listener.

I've used this technique for flames coming from Boilover phenomenon, so a flame going from a few 10s of centimeters to more than 2 meters. And the results I obtained were consistent with what I was expecting. Concerning the 3D aspect of the fire, you can try to plot the out-of-plane strain dW/dz = -dU/dx - dV/dy to have an idea of the importance of the out of plane flow and to see if this technique can still be applicable to the type of fire you are interested in or not. If you want more details, I can also put you in contact with my colleague who developed this technique for large scale car park fires (send me a message).

Hi Delphine, now you are talking. I think I would very much like to talk to this colleague of yours. I have designed a new way of producing fire watermist guns, and the relations between the throw, the water density of the watermist plume and the droplet size weigh heavily on the fire-combating efficiency achieved using any given gun. I went from small portable fire extinguishers, through CRMs, diesel engine containers, RR engine (locomotive) and wagon cabins, and now am contemplating mining corridors, highway tunnels etc. There is a chance I could tune my guns a lot better once I know anything about the small v. large scale eddies for sizable fires. The largest fires I have tackled in earnest and studied in some detail to date were cooking oils bursts in kitchen ranges and there the flames were 2m, or a bit less. Thanks for your help. Isn't boilover phenomenon related to self-ignition of liquid fuels?

Thank you a lot for your contributions. I will read these with a great attention and trying to answer ASAP.

I would personally implement a schlieren/shadow graph technique, and couple it to a high speed camera. Gary Settles descibes a simple set-up with a field of view of 2m by 2m. The technique would permit you to resolve the regions of rapid change in density. You could then follow these "fluid particles" and track their velocity. You can even try implementing a stereo technique and reconstruct the 3D flow field.

Add some micron scale metal particles (iron or aluminum), and these will act as tracers, as thy will be advcted by the flow. They will light up very brightly, and will trace out the streamlines. You can also resolve their motion in 3D with 2 high speed cameras.