Towards Machine Learning and Low Data Rate IoT for Fault Detection in Data Driven Predictive Maintenance

Abstract

While predictive maintenance is a concept that has been around for several decades, it is only due to the relatively recent arrival and expeditious development of fourth industrial revolution technologies, such as the internet of things and machine learning, that it has become more of a reality. Rural communities face several challenges in their day to day lives and while several development projects have been enacted to address these problems, many have failed due to a multitude of factors. One of the contributing factors to these rural development projects failing is the lack of or insufficient maintenance. The aim of this study was to show how fault detection in data driven predictive maintenance in remote and rural locations could be achieved using the one-class support vector machines algorithm and low data rate (bandwidth) internet of things. The results of this study show how fault detection in predictive maintenance can be achieved using the one-class support vector machines algorithm and low bandwidth internet of things sensors, for rural applications. The outcome of this study provides a steppingstone to implementing data driven predictive maintenance in remote and rural locations.