INTEGRATING SENSOR DATA AND MACHINE LEARNING FOR PREDICTIVE MAINTENANCE IN INDUSTRY 4.0

Abstract

The availability of manufacturing machinery is crucial for having a productive production line. So, for industrialists, being successful in the field of maintenance is crucial if they want to make sure that key equipment is performing as it should and that unscheduled downtime is kept to a minimum. Predictive maintenance skills are viewed as being essential with the rise of complex industrial processes. The assistance that contemporary value chains may provide for a company's maintenance role is another area of focus. The development of sensors and Industry 4.0 technologies has greatly improved access to data from equipment, processes, and products. Electric motor condition monitoring and predictive maintenance help the industry avoid significant financial losses brought on by unforeseen motor breakdowns and significantly increase system dependability. This research offers Enhanced Nave Bayes Artificial Neural Network-based machine learning architecture for Predictive Maintenance. The system was tested in an industrial setting by building a data collection and analysis system using sensors, analyzing the data with a machine learning approach, and comparing the results to those generated by a simulation tool. With the help of the Azure Cloud, the Data Analysis Tool may access information collected by a wide variety of sensors, machine PLCs, and communication protocols. Preliminary results show that the method correctly predicts a wide range of machine states.

Keywords

• predictive maintenance
• machine learning
• sensor data
• enhanced naïve bayes artificial neural network (enbann)
• industry 4.0 technologie