This experimental study precisely identifies the soil layer interface using sound signals collected during cone penetration tests (CPTs). Using the high sensitivity of the sound sensor at the cone tip, this study effectively increases the accuracy in identifying the soil interface location. Shiluo sand and Jhongda Lateritic soil were used to construct large-scale soil layers. An acoustic CPT was applied to determine sound pressure and its root mean square (RMS). The Fourier spectrum and Hilbert-Huang transform (HHT) method were utilized and compared in their ability to analyze sound signals. The HHT method with a local energy-instantaneous frequency-time distribution performed well. According to analytical results, both the RMS and HHT methods are more effective than cone resistance in locating the soil interface. The major contribution of this study is that it uses the RMS of sound wave as a more accurate, effective and easy method for exploration of soil distribution and detection of soil interfaces.
