Mise en place d’une solution Machine Learning pour la maintenance prédictive dans l’industrie 4.0.

Abstract #

The growing amount of data accessible in nearly every sector necessitates the use of

algorithms for automated data analysis. This requirement is underlined in predictive

maintenance, where the ultimate goal is to forecast hardware component failures by continually

monitoring their state in order to schedule maintenance activities ahead of time.

These observations, which are often in the form of time series and event logs, are created

by monitoring systems and cover the lifespan of the respective components. The major

difficulty of data driven predictive maintenance is analyzing this history of observation in order

to create predictive models.

Machine Learning has grown common in this field since it allows for the extraction of

information from a multitude of data sources with minimal human interaction.

The objective of this project is the realization of a Machine Learning solution for

predictive maintenance in industry 4.0. This project is divided into 3 parts. First, we started

with a presentation of AMEE, the host organization. Then, we presented the methodologies and

algorithms of Machine Learning using a documentary approach based on scholarly articles on

the same subject. Finally, we started the realization of the predictive solution using the Machine

Learning algorithms.

Keywords :

Data, Machine Learning, Predective Maintenance, Industry 4.0, Analysis algorithm,

Predictive Model, Data preparation, Algorithmic model.