Characterization and matched -field processing localization of photoacoustic signals.

This dissertation presents the results of an investigation performed to characterize photoacoustic sound from gases in an open environment and to determine its utility for localizing small gas clouds. Photoacoustics is the generation of acoustic waves due to unsteady heating from a light source. It is well understood for trace gas detection and spectroscopy when the gases are placed in chambers. However, it is poorly understood in an open environment. Leak detection and localization are critical quality control processes because many industrial and domestic machines use or convey pressurized gases or liquids. Unintended leaks from machine components may be detrimental to consumers, manufacturers, and the environment. Current leak testing methods are either subjective, time consuming, or lack automated localization capability. The use of photoacoustic signals measured with multiple microphones for the localization of leaks is examined to address the shortcomings of the current leak testing methods. Scaling laws for photoacoustic sound pressure are developed with dimensional analysis and verified with experiments using a carbon dioxide laser and sulfur hexafluoride as the tracer gas to generate the photoacoustic sound. A photoacoustic signal model based on first principles is developed and takes in to account gas cloud shape and realistic gas absorption. For acoustically distributed gas clouds, the model and experiments agree to within 3 dB in a 10--120 kHz bandwidth. For acoustically compact gas clouds, the model and experiments agree to within 3 dB in a 30--120 kHz bandwidth. Matched-field processing is applied to photoacoustic measurements made by a four-microphone array. The photoacoustic sound is generated by scanning a carbon dioxide laser beam over a calibrated leak source of sulfur hexafluoride. The results of this study indicate that measured photoacoustic signals processed using matched-field processing can be used to accurately localize gas clouds from leak sources that leak at a rate of 1.19 x 10-5 CM3/S to within +/-1 mm Different processing techniques are demonstrated and acoustic propagation model robustness studies are performed.PhDAcousticsApplied SciencesMechanical engineeringPure SciencesUniversity of Michigan, Horace H. Rackham School of Graduate Studieshttp://deepblue.lib.umich.edu/bitstream/2027.42/132503/2/9963925.pd