Analysis of the Faults in Ratchet Mechanisms in the Presence of Noise

Bartosz Połok, Piotr Bilski

Abstract

<p>The following paper presents the methodology of monitoring the state of the ratchet mechanisms in the presence of noise. The fault detection is based on the acoustic analysis of the signals generated by the revolving mechanism. Decision is made using machine learning methods, which accuracy is compared. The object of the analysis is the ratchet mechanism installed in the actual BMX-type bicycle. It was shown that the noise suppression approach is suitable for the applied diagnostic framework, leading to the high accuracy in detecting catastrophic faults, such as breaking the tooth inside the mechanism.</p>

Keywords

acoustic analysis, machine learning, fault detection in ratchet mechanisms, denoising

Download

IMEKO-TC10-2023-006.pdf

DOI

10.21014/tc10-2023.006

IMEKO TC

TC10 - Measurement for Diagnostics, Optimization and Control

Back to the proceedings