Abstract:Mahalanobis distance can identify effectively for the initial crack using acoustic emission signal, the text acquisited acoustic emission signal of drawing bucket under different lubricating condition , then using Matlab software program and local wave decomposition to solve the collected acoustic emission signal for adaptive intrinsic mode function component, effectively raised each IMF acoustic energy, transmited signal characteristic parameter, calculated Mahalanobis distance of under three states no crack, crack and obvious crack as standard value, and the other Mahalanobis distance compared with standard values, got accurately the crack status identification, also reflected in the deep drawing process of sheet metal in the friction performance.