Non-stationary statistical fault indicators estimation applied on IAS machine surveillance

In the domain of machinery surveillance, most part of condition monitoring techniques have been developed for stationary operation. Recently, analysis of the Instantaneous Angular Speed (IAS) has been proven to effectively detect bearing mechanical faults, even in early stages. Nevertheless, the impact of torque and operating speed macroscopic variations over the IAS spectral indicators does not seem to have been deeply discussed. The following work presents a methodology for the estimation of IAS perturbation amplitude indicators as a function of the operating conditions. The approach can be used as a surveillance tool, and also to reproduce non-stationary conditions on machine numerical models by knowing the indicator response on stationary conditions. The methodology was validated through an experimental test rig by the observation of the shaft unbalance frequency on the IAS angular spectra.