A Novel Optimization Demodulation Method for Gear Fault Vibration Overmodulation Signal and Its Application to Fault Diagnosis

Abstract

Many demodulation methods are devoted to obtain the gear fault modulation frequency as the fault diagnosis indicator instead of the precise amplitude and frequency demodulation. It is because the precise demodulation is a challenge especially for the over-modulation signal potentially caused by the gear fault, which further hinders the gear fault degree diagnosis. This paper proposes a novel optimization demodulation method for over-modulation signal. The Hilbert transform demodulation and Jacobi-Anger expansion form are respectively combined with the trust region reflection optimization algorithm to achieve the precise amplitude and frequency demodulation. The objective function about AM parameters considers the zero-crossing feature of the time domain AM in over-modulation signal. The iterated Hilbert transform demodulation is innovatively applied to remove the singularities effect in the residual signal caused by the over-modulation. The simulations and experiments demonstrate that the proposed method has the advantages of high accuracy and strong anti-noise performance in the demodulation of over-modulation signal. The quantitative FM and AM signals obtained by the proposed method can be applied to detect the gear fault degree, which provide an approach for establishing linkages between the modulation intensity and fault degree.