Comparative Study of Machine Learning Models Applied to PdM

Thanks to the development of Industry toward digitalization, predictive maintenance (PdM) guidelines are now applied to manage availability, downtime, component utilization rates, equipment lifespan and their maintenance costs, using technologies such as Internet of Things (IoT) devices and artificial intelligence. Data collection through communication networks enables the acquisition of massive amounts of data on equipment conditions and operational processes allowing for automated fault diagnostics after detecting anomalies by analyzing information from pre-installed sensors in the equipment. This paper presents a comparative analysis of machine learning (ML) models extensively applied to PdM. It compares machine learning models based on the most commonly used criteria in PdM and industrial equipment while highlighting the main contributions of researchers, thus offering directions and opportunities for future research.