Machine and industrial monitorization system by analysis of acoustic signatures

Abstract

Prevention and diagnosis of potential malfunctions in industrial machines is an important factor in the total costs of operation. In recent years several new signal processing techniques have appeared, that under very specific conditions, allow the separation of mixed signals into their original components. This paper describes the development of a monitoring system using acoustic analysis in industrial environments in order to detect defects or malfunctions. We describe the conception and implementation of a system that uses a technique known as Blind Source Separation (BSS). It was used to monitor and diagnose industrial machines with mobile components in which each component has a particular acoustic signature.