On the Use of Jerk for Condition Monitoring of Gearboxes in Non-stationary Operations

Abstract

Diagnostics of rotating machinery is very important to preserve their efficiency. Defects should be detected at an early stage in order to plan maintenance and avoid a stop in production. In this chapter the use of jerk, the derivative of acceleration, is used to diagnose a gearbox with a local tooth defect. A dynamic gearbox model is presented, which includes time-varying mesh stiffness and non-stationary operating conditions modeled as variations of load and speed. In order to model the tooth defect, a reduction in mesh stiffness is introduced proportionally to the severity of the defect. Two case studies are presented. The first one concerns stationary operating conditions and different severity of defects. For this case, the jerk shows good ability to diagnose the presence of the local defect. For the second case, variable loading condition was modeled as a sawtooth shape, whereafter the jerk was used to detect the presence of a defect. The amplitude modulation cause in increase in the vibration level that does not allow the identification of defect using only the acceleration signal. The jerk allows this identification even for significant load variability. The performance of jerk for signals with additive Gaussian noise is discussed, highlighting its limitations.