Connect with us

How to detect radiant surface area?

Discussion in 'Electronic Design' started by Chris Carlen, Oct 9, 2003.

Scroll to continue with content
  1. Chris Carlen

    Chris Carlen Guest


    I have a scientist who needs to determine the area of the emitting
    surface of a radiant soot cloud in the combustion chamber of an
    optically accessible internal combustion engine. The measurement needs
    to be done at a rate of at least 4kHz, preferrably 15kHz, with a
    resolution of at least 1 part in 256, preferrably 1024 or so.

    We have considered a high frame rate low resolution camera, but these
    are somewhat expensive, and we would like to investigate cheaper
    alternatives before attempting to justify the purchase of such a camera.

    We plan to image the soot cloud onto some suitable detector or camera.
    The individual pixel data would be compared against a threshold in
    post-processing, to reveal the area of the image.

    What other methods might be able to return the same information?

    Thanks for input.

    Good day!

    Christopher R. Carlen
    Principal Laser/Optical Technologist
    Sandia National Laboratories CA USA
    -- NOTE: Remove "BOGUS" from email address to reply.
  2. R.Legg

    R.Legg Guest

    Sounds like a sampling problem. Does he really need sequential images
    of 4000 clouds?

    I would have thought that unstable conditions in an engine, designed
    to do things repeatedly and reliably, should show up just as easily
    looking at periodic images. The same with dynamic conditions, if
    properly synchronized.

    As an alternative - why not slice the image, looking at adjacent scan
    lines of a slow-scan camera? (~4Hz total scan with single-line images
    from 1000 clouds). It might tell you as much as you want to know, when
    compared to known references of single cloud events.

  3. Ian Stirling

    Ian Stirling Guest

    At some wavelengths, the soot should emit as a black body (assuming for
    the moment that it's isothermal).
    If you can take a spectrum of the emission, you may be able to isolate this
    black body emission, and simply measuring a grain of soots emissivity will
    get you the rest.
  4. Chris Carlen

    Chris Carlen Guest

    Actually, what is needed is only 10 images (or more if possible), but
    only 250us or less apart.
    Things aren't quite so ideal. What is being attempted is a measure of
    the effective soot temperature, or "bulk" soot temp. This must be
    determined at several crank angles per cycle, then over several other
    parameter variations.

    This requires determining the radiant surface area so as to be able to
    normalize an intensity reading, which will be factored in with two color
    intensity measures to fit to an emission curve to get temp.
    Unfortunately, the soot doesn't behave as a black body, so additional
    info is needed than just two colors. Hence the desire to make an
    absolute intensity per area determination.

    Now what I am unsure about, is if it would be possible to simply
    characterize the average soot radiant surface area at a given crank
    position, for a given set of conditions. This could be done once, then
    the desired experiments performed using that data to make sense of the
    two color data on the fly.

    But I don't think this will work. The scientist has spend considerable
    time on this, and he is refining still more efforts on the part of many
    others as reported in literature. I don't think there is such an easy
    way out that has been overlooked. I think the reason has to do with the
    fact that the soot radiant surface area is just not consistent enough
    from cycle to cycle at any given crank position to be able to make such
    a one time characterization. But I am guessing a bit here since I am
    not the scientist, just helping the fellow to find out what techniques
    might be available that he may not be aware of.

    Hmm. I really think 2D spatial resolution is needed.

    Thanks for the input!

    Good day.
  5. Tim Shoppa

    Tim Shoppa Guest

    My gut feeling is that measuring at many more than 2 wavelengths will
    give you a better handle on the temperature than just at 2 wavelengths.
    Probably good enough that you won't have to know radiant surface area
    at all.

    Nothing is a perfect black body, but if you look at a restricted part
    of the spectrum (away from emission lines/absorption lines) it will
    probably be good enough. Lots of people make their livings by looking
    at spectra and turning them into temperature/density measurements.

  6. Jim Meyer

    Jim Meyer Guest

    The area of a cloud is a particularly ill defined quantity.

    Could it be that he wants to measure the total amount of soot and
    that he assumes that if he measures the total emission from the soot
    cloud that he can derive the total soot from that?

    Or is he more concerned with the shape of the cloud?

  7. Jim Meyer

    Jim Meyer Guest

    Does this paper describe what you want to do?

  8. Chris Carlen

    Chris Carlen Guest

    This is a paper from my department, and the "Sandia Cummins heavy Duty
    Engine" is the experiment that this post is about.

    The paper is a bit old. The lab changed hands since then. I'll try to
    persuade the scientist to join the discussion, as some of the questions
    folks are responding with might better be considered by him.

    Thanks for the input.

    Good day!

    Christopher R. Carlen
    Principal Laser/Optical Technologist
    Sandia National Laboratories CA USA
    -- NOTE: Remove "BOGUS" from email address to reply.
  9. Tim Jackson

    Tim Jackson Guest

    If what you really want is emission per unit area then just measure emission
    from "one unit" of area: mask off a known small area of the cloud, say 5mm
    square, and measure the radiation characteristics of that.

    As the radiation is presumably not uniform then knowing the total area and
    total emission won't allow you to normalise the intensity anyway, you.would
    need to map your measurements and integrate over the visible area.

    You can get plenty of time resolution with a photocell (or two) and an
    oscilloscope on one point. If you need a spatial map then you can move your
    mask mechanically (step and repeat) from cycle to cycle to progressively map
    the whole chamber. You can take several traces at each point to determine
    the relevance of instabilities and get an average temperature for each

    If you really -have- to do it in parallel and in one cycle then what you are
    doing is to all intents and purposes building a high speed camera

    Tim Jackson
Ask a Question
Want to reply to this thread or ask your own question?
You'll need to choose a username for the site, which only take a couple of moments (here). After that, you can post your question and our members will help you out.
Electronics Point Logo
Continue to site
Quote of the day