Embedded System for Preventive Maintenance in Industry 4.0

Preventative maintenance strategies are growing in importance in Industry 4.0 thanksto the new possibilities offered by the developing technologies. The possibility of remotelymonitoring the system and take actions based on real time data before seriousfaults occur is also possible by employing wireless sensors in the so called WirelessSensors Networks (WSNs).In the presented work, the state of the art of preventative maintenance in Industry 4.0is discussed. After defining desirable characteristics, sensing technologies for vibration,temperature, magnetic field, audio and humidity are analysed and the suitability of eachof them for industrial environments is discussed. Different communication protocolsfor wireless sensor nodes are described and compared in order to perform the optimalchoice according to the requirements and constraints of the application. Finally, realindustrial applications found in literature are presented for vibration, magnetic field,temperature/humidity and audio monitoring together with characteristics of the sensorsemployed. A brief description of energy scavenging from low frequencies magneticfield is presented, being self-powering a real possibility for future sensors monitoring,increasing reliability and lowering network maintenace costs.Condition Based Maintenance strategy is then applied to the real test case of a manufacturingline based on a Linear Synchronous Motor (LSM). First the vehicle movingon the industrial transport system is characterised in terms of static vibration andmagnetic field. Secondly, the vibration profile is retrieved from the analysis on themoving vehicle and Condition Based Maintenance techniques are applied.Different algorithms were then proposed and analysed for frequency and time domainanalysis. In particular Fast Fourier Transform (FFT) and Short Time Average Energyare also optimised and validated thinking of a long term implementation on anembedded system with reduced computational power.