evenT-based prEdictive analytics for Control and Health MAnagement of INdustrial sysTems – TECHMAINT

Prognostics and Health Management (PHM) has been widely used in industrial systems to provide real-time data management, processing and control. PHM aims to predict the evolution of the system state in the future based on the current and past condition monitoring data, in the way to anticipate system failures.
So, PHM is referred to Predictive Analytics frame encompassing a set of artificial intelligence tools to create predictive models. Indeed, collected real-time data allow to develop models but continuous monitoring is usually expensive or even impossible. An alternative approach would be to employ discrete event data representative of events (e.g. alarms, modes). The main objective of using discrete data for prognostics is to recognize, at least, patterns constructed from a sequence of discrete events, their occurrences rules, and their interactions. We aim at developing novel probabilistic approaches for discrete event data analysis combined with machine-learning approaches. We will also incorporate mixed and multidimensional methods to handle discrete and continuously measured information. Such prognostics based on this is not scientifically mature today leading to submit this proposal TECHMAINT to contribute to its enough foundations.