Detection and classification of failures as an emergent behavior in a machinery system modelled as a system of systems

Abstract

In Industry 4.0 context predictive maintenance is a hot topic with several challenges that can be investigated by a system of systems engineering approach where the emergent behaviour can be related to possible system failures. The Internet of Things (IoT) can be also used to monitor and control the different system components. This work proposes a simple approach to detect, monitor, and control emergent behavior by IoT sensors. Starting from the raw data which were daily extracted from sensors, an analysis of the collected signals is performed to help decision-making concerning its maintenance. In the case study related to a door system in a bus, the proposed method can recognize the different damage states analysing the input parameters: temperature, pressure, and humidity of door systems. Some recent reliability performance indicators are used to evaluate the damage classification.