Induction motor fault diagnosis based on analytic wavelet transform via Frequency B-Splines

In this paper a new methodology of transient motor current signature analysis (TMCSA) is proposed. The approach consists on obtaining a 2D time frequency plot representing the time-frequency evolution of all the harmonics present on an electric machine transient current. Identifying characteristic patterns in the time-frequency plane, produced by some of the fault related components, permits the machine diagnosis. Unlike other CWT based methods, this work uses complex frequency B-splines wavelets. It is shown that these wavelets enable high detail in the time-frequency maps and an efficient filtering in the region neighbouring the main frequency. These characteristics make easy the identification of the patterns related to the fault components. As an example, the technique has been applied to no load startup currents of healthy motors and motors with broken bars, showing the Complex FBS Wavelets capabilities. The diagnosis has been done via the identification of the upper sideband harmonic.