A Tacholess Angle-time Cyclostationary Analysis Method for Bearing Fault Diagnosis

Abstract

Cyclostationary analysis has now become a basic analysis method for vibration signals of rotating machinery. During the start and stop phases of the machine, the vibration signals with varying rotational speeds include both time-dependent components and angle-dependent components. The angle-time cyclostationary technique proposes an effective approach that can simultaneously analyze signals in both time and angle domain. However, the method requires rotational speed as a necessary condition, which is not available in the absence of a tachometer. To overcome this issue, a novel speed curve estimation method combined with time-angle cyclostationary analysis for bearing diagnosis is proposed. This method uses the instantaneous frequency of the vibration signal to obtain the rotational speed, and then analyzes the rotational speed information and the vibration signal through the angle-time cyclostationary tool. The proposed method is validated by numerical and experimental analysis.