Application of the Teager-Kaiser energy operator in bearing fault diagnosis

► Normal and fault bearing vibration signals are transformed to Teager-Kaiser signal. ► Statistical and energy-based features are extracted from the Teager-Kaiser signal. ► Same features are extracted from time and AM bearing signals for comparison. ► Neural network classifier and least-square SVM classifier are used to obtain diagnosis success rates. ► Features extracted from Teager-Kaiser signal shows better performance.