Rommel Stiward Prieto, Diego Alberto Bravo Montenegro and Carlos Rengifo
The purpose of this paper is to approach predictive maintenance (PdM) of brushless direct current (BLDC) motors using audio signal processing and extracting statistical and…
Abstract
Purpose
The purpose of this paper is to approach predictive maintenance (PdM) of brushless direct current (BLDC) motors using audio signal processing and extracting statistical and spectral features to train classical machine learning (ML) models.
Design/methodology/approach
The proposed methodology relies on classification predictive model that shows the motors prone to failure. To verify this, the model was implemented and tested with audio data. The trained models are then deployed to an Industrial Internet of Things (IIoT) application built using Django.
Findings
The implementation of the methodology allows for achieving performance as high as 92% accuracy, proving that spectral features should be considered when training ML models for PdM.
Originality/value
The proposed model is an effective decision-making tool that provides an ideal solution for preventive maintenance scheduling problems for BLDC motors.