Predictive Maintenance in Industry 4.0: A Review of Data Processing Methods

Abstract

Industries require a revolution with the aim to improve productivity, sustainability, reduce costs, achieve customer satisfaction, and adopt emerging technologies. The current industrial revolution, known as Industry 4.0, is defined by smart technologies, which include robotics, big data, artificial intelligence (AI), and the internet of things (IoT). A breakdown in industrial process has consequences in terms of safety, productivity, and efficiency. Predictive maintenance is a cornerstone of Industry 4.0, which predicts the remaining useful life of equipment’s, devices, and machines using condition monitoring, data analytics, and machine learning (ML). This article reviews a number of methods, including signal analysis, statistical analysis, and artificial intelligence-based methods, that can be used for performing predictive maintenance on the collected data from machines or equipment’s. The signal processing and statistical analysis methods require domain knowledge, but the AI methods require labelled data (most of the data is healthy and lacks faulty data), which is a challenging task.