The objective of this study is to detect a cavity and estimate its size using sound waves in a laboratory model chamber filled with dry sand. One side of the chamber is covered with an acrylic plate, and a cavity is placed between the plate and sand. Sound waves are generated by impacting the plate with an instrumented hammer, and are measured using a microphone. The measured sound waves are analyzed with four comprehensive analyses including the measured area under the rectified signal envelope (MARSE) energy, flexibility, peak magnitude of wavelet transform, and frequency corresponding to the peak magnitude. The test results show that the accuracy of cavity detection using the MARSE energy is higher for thicker plates, whereas that using flexibility is higher for thinner plates. The accuracies of cavity detection using the peak magnitude of wavelet transform, and frequency corresponding to the peak magnitude are consistently high regardless of the plate thickness. Moreover, the cavity size may be under- or overestimated depending on the plate thickness and the selected analysis method. The average of the cavity sizes estimated by these methods, however, is slightly larger than the actual cavity size regardless of the plate thickness. This study demonstrates that microphones may be effectively used for the identification of a cavity and the estimation of its size.
Keywords: MARSE energy; cavity; flexibility; microphone; sound waves; wavelet transform.