Intelligent diagnostic system for the rachet mechanism faults detection using acoustic analysis

Abstract

The paper presents the diagnostic procedure for the ratchet mechanisms' fault detection based on the acoustic signal analysis. The diagnostic framework was proposed, consisting in three steps. First, the symptoms are extracted from audio recordings using the proposed measurement system. Next, training and testing data sets, representing various faults of the mechanism are created. Finally, the selected Artificial Intelligence-based classifier is used to extract knowledge about the relation between the internal mechanism’s state and symptoms observed in measurement data. The classifier is used to automatically evaluate state of the actual mechanism. Experiments were performed to verify the system’s efficiency (especially ability to detect and locate faults related with the pawl degradation) during the analysis of the selected gears in the Bicycle Motocross (BMX). The selected classifiers were proven to be suitable for the task. The system’s accuracy close to 100% makes it a perfect tool for the analysis of real-world objects.