Mother Wavelet Selection for Control Valve Leakage Detection using Acoustic Emission

Abstract

This paper discusses the method used by the author to select suitable mother wavelet for analysis of acoustic emission (AE) signal. The AE signal were obtained from faulty control valve that experienced three stages of leakage; minor, critical and damaged. Analysis using Wavelet Packet Transform is expected to reveal AE characteristics at different leakages. The selection criteria for mother wavelet to be used in WPT is based on Energy to Shannon entropy ratio. Four mother wavelets were used for selection: Daubechies, Symlet, Biorthogonal and Coiflet. Mother wavelet with the highest Energy to Shannon entropy ratio is chosen for further analysis.