Quadratic-nonlinearity power-index spectrum and its application in condition based maintenance (CBM) of helicopter drive trains

This paper introduces a quadratic-nonlinearity powers-index spectrum (QNLPI(f)) measure that describes quantitatively how much of the mean square power at certain frequency f is generated by nonlinear quadratic interaction between different frequencies inside signal spectrum. The proposed index QNLPI(f) is based on the bicoherence spectrum, and the index can be simply seen as summary of the information contained in the bicoherence spectrum in two dimensional graph which makes it easier to interpret. The proposed index is studied first using computer generated data and then applied to real-world vibration data from a helicopter drive train to characterize different mechanical faults. This work advances the development of health indicators based on higher order statistics to assess fault conditions in mechanical systems.