Chaos theory using density of maxima applied to the diagnosis of three-phase induction motor bearings failure by sound analysis

Abstract

Bearing failures in the industry are a recurring problem that can cause permanent damage to machines and interrupt production in important sectors of a factory. For this reason, over the past few decades, different studies have been carried out to develop techniques for diagnosing this failure. The main challenge found in the diagnosis is to identify the fault and its severity with the machine operating under dynamic load conditions, generally requiring a large acquisition window. This paper presents an approach based on quantification of the chaotic behavior for the characterization of rigid ball bearing failure of a three-phase induction motor through the method called signal analysis based on chaos using density of maxima (SAC-DM) using the sound signal emitted by the engine. This technique is based on an algorithm that counts peaks of the motor sound signal in the time domain to detect faults using only a sensor and an algorithm with a low computational cost. Experimental results show that the SAC-DM technique is sensitive to bearing failure and allows the diagnosis to be made even under load variation and with an acquisition window of just 0.28s.