Abstract

Noise measurement is one of the technologies for health monitoring and diagnosis of rotating machines such as gearboxes. Although, significant research has been undertaken in understanding the potential of noise measurement in monitoring gearboxes, this has been solely applied on any types of gears (spur, helical, etc.). The condition monitoring of a lab-scale, single stage, gearbox, represents the real vehicle’s gearbox, using non-destructive inspection methodology and the processing of the acquired waveform with advanced signal processing techniques is the aim of the present work. Acoustic emission was utilized for this purpose. The experimental setup and the instrumentation are present in detail. Emphasis is given on the signal processing of the acquired noise measurement signal in order to extract conventional as well as novel parameters of potential diagnostic value from the monitoring waveform. The evolution of selected parameters/features versus test time is considered, evaluated and the parameters with most interesting diagnostic behavior are highlighted. The present work also describes the results concluded by long term (~6.0 h) experiments to a defected gear system, with a transverse cuts ranged from 0.75 to 3.0 mm to simulate the artificial tooth crack. Different parameters, related to the analysis of the recording signals coming from acoustic emission are presented and their diagnostic value is discussed for the development of a condition monitoring system.

Key words: Diagnostic, geared system, sound pressure level, stationary signal, faulty gear, measuring devices, condition of gear, monitoring, maintenance action.