Induction motor´s broken bars detection by using Stockwell transform

This paper presents a Stockwell transform-based method for processing starting angular speed data from an induction motor and transforming it into some forms that are useful for fault detection and diagnostics of rotor broken bars. The proposed method generates angular speed frequency-time spectrum that is dependent to the number of rotor broken bars and of the motor load level. The angular speed signal energy, the amplitude of its spectral component with the highest frequency and the amplitude-time diagrams of its spectral components are analyzed as number of broken rotor bars indicators. The results are obtained by simulation in Matlab/Simulink of a faulty induction motor.