A Sonar Experiment to Study Sound Propagation through Flames

Mustafa Z. Abbasi – mustafa_abbasi@utexas.edu
Preston S. Wilson – pswilson@mail.utexas.edu
Applied Research Laboratories
Department of Mechanical Engineering
The University of Texas at Austin

Ofodike A. Ezekoye – dezekoye@mail.utexas.edu
Joelle I. Suits – jsuits@utexas.edu
Department of Mechanical Engineering,
The University of Texas at Austin

Popular version of paper 2pPA14 presented at the 164th ASA Meeting, 2012 in Kansas City, Missouri.


Tragedies like the Charleston, SC., sofa warehouse fire (June, 2007), in which several firefighters perished after becoming disoriented and trapped in the structure [1], motivated our research group to look for improved technologies to aid firefighter navigation and rescue on the fireground. The interior of a burning building is an extremely hostile environment. In addition to dealing with the challenging thermal and toxic chemical environment, firefighters can neither see nor hear as well as in normal conditions, as illustrated by the photograph in Fig. 1 that shows the heavy smoke and fire associated with the sofa warehouse fire.
Fig 1. Photograph of the June 2007 Charleston, South Carolina sofa warehouse fire, showing heavy smoke and fire rolling out the windows. This photo was taken less than a minute after rescue crews were forced out of the showroom by the interior conditions. [Figure adapted from: http://www.cdc.gov/niosh/fire/reports/face200718.html.
Currently, firefighters use tools such as thermal imaging cameras (TIC) to detect obstacles, passageways and judge distances inside a burning building. It has been found through research on TIC that heat and open flame significantly degrade TIC performance [2]. Our group seeks to develop a complementary imaging technique based on the sonar paradigm (similar in concept to systems used for parking assistance in cars) to augment TIC technology. Our initial goal is to locate an open doorway or hallway through flames and smoke within a burning building, but even this simple goal could greatly enhance a firefighters ability to navigate through a burning building and out of danger. From a more fundamental physical acoustics perspective, we are interested in understanding how fire, heat and smoke between the device and a target effect acoustic propagation.

Early in this research, following the automotive parking assist example, we experimented with an ultrasonic device that operated at 40 kHz and generated a narrow beam of sound. Unfortunately, we found that fire strongly distorted the sound beam and prevented it from making it through even small-sized laboratory fires. Experimental evidence and finite-element numerical modeling indicated that it was not the temperature and sound speed contrast of the fire that caused the distortion, as one might expect. Instead it was scattering of sound from the turbulent structure of the fire that caused the distortion. The wavelength of sound at 40 kHz is about 8 mm, which is smaller than the characteristic size of the turbulent structure of the fire, and hence the turbulence can easily scatter sound at that frequency.

Subsequently, we have experimented with a more advanced acoustic source, called a parametric array [3], that can create a much lower frequency sound in a narrowly focused beam. The lower frequencies (on the order of a few kiloHertz) are not as readily scattered by the turbulent structure of fire, yet the parametric source can still maintain a narrow beam at low frequencies, which greatly increases the spatial resolution possible from a sonar system using the source. In the initial results reported here, we found that acoustic signals from the parametric array can effectively penetrate moderate sized flames, reflect off of wall surfaces on the opposite side of the flames, and return back through the flames to a co-located acoustic receiver (a microphone) positioned near the source. Tracking these echoes, one can determine the distance to the wall and also create an image of the location of wall by scanning the source spatially, as shown in Fig. 2, along with a corresponding TIC image of the same fire.

Fig. 2. The image on the left is a front view of a flame located 1.5 meters in front of a wall, taken with a thermal imaging camera. The flame dominates the image and the presence of a wall behind the flame is obscured. The figure on the right is a top or plan view of the same room obtained by sonar imaging through the flame. The flame itself is not visible in the sonar image, but the wall is visible, and the image provides distance information missing from the infrared image. The sonar image of the wall is far less affected by the presence of the flame. Combining information from both images would be better than either image alone
These initial results indicate that acoustic sensing can potentially be used to help firefighters navigate in a burning building either as a stand-alone sonar system or combined with a TIC system. The currently work is basic in nature and further development is underway, however we can imagine a future where the technology developed in this study can be directly incorporated into standard firefighter gear or onto future mobile platforms, for example robots used to assist fire fighters.

Acknowledgements
This work was supported by internal R&D funds at Applied Research Laboratories at The University of Texas at Austin [http://www.arlut.utexas.edu/], a University Affiliated Research Center with broad experience in sonar development and signal processing.

References

[1] N.P. Bryner, S.P. Fuss, B.W. Klein, and A.D. Putorti Jr, “Technical Study of the Sofa Super Store Fire, South Carolina, June 18, 2007, Volume I,” National Institutes of Standards and Technology NIST SP – 1118, 2011.

[2] F. Amon, N. Bryner, A. Lock, and A. Hamins, “Performance Metrics for Fire Fighting Thermal Imaging Cameras – Small- and Full-Scale Experiments,” National Institute of Standards and Technology NIST Technical Note 1499, 2008.

[3] M.B. Bennett and D.T. Blackstock, “Parametric array in air,” J. Acoust. Soc. Am. 57 562–568 (1975).