SUMMARYWe investigated propagation properties of Lamb waves in a gas-insulated switchgear (GIS) tank to diagnose insulation performance of GIS. The acoustic signals excited by a free metallic particle colliding with the tank sheath were measured using AE sensors. The wavelet transform was applied to decompose the wave data into its time-frequency components. As a result, difference of propagation properties of Lamb waves is clearly seen with different sizes of GIS tank. Based on the characteristics of Lamb waves, algorithms for location identification of a free metallic particle were examined in model GIS using two AE sensors. Herein, we propose a new system for location identification of a free metallic particle in GIS. Moreover, it is verified that the new identification system is suitable as a diagnostic technique for GIS.