Offre en lien avec l’Action/le Réseau : – — –/– — –

Laboratoire/Entreprise : LS2N/IMT Atlantique
Durée : 5-6 mois
Contact : samir.loudni@imt-atlantique.fr
Date limite de publication : 2022-02-28

Contexte :
Recently, with the emergence of Industry 4.0 (I4.0), predictive maintenance (PdM) based on data-driven methods has become the most effective solution to address smart manufacturing and industrial big data, especially for performing health perception (e.g. fault diagnosis and remaining useful life (RUL) estimation). Here, maintenance corresponds to the process that deals with equipment or system components to ensure their normal operating under any circumstance. PdM relies on the continuous monitoring of the equipment or the machine to predict when maintenance actions are necessary; hence the maintenance can be scheduled. Detecting and preventing failures is thus essential, and industries seek to minimise the number of operational failures, minimise their operational costs, and increase their productivity.

Failure Prediction is one of the critical components of PdM for which the main goal is to predict the approximate moment when some failure could occur. Recent works have addressed anomaly detection for PdM in order to predict incipient failures from historical data.

In the last decade, new research have began connecting data mining to symbolic Artificial Intelligence (AI). Such fertilization leads to a number of algorithms that have been proposed within Constraints Programming (CP) and Satisfiability (SAT) for mining sequences, frequent item-
sets, association rules, clustering, classification, etc. The main advantage
of symbolic AI approaches for pattern mining is their declarativity and flexibility, which include the ability to incorporate new user-specified constraints without the need to modify the underlying system.

Sujet :
The objective of this internship is to use constraint programming to apply symbolic data mining techniques on historical data to characterise the healthy behaviour of equipment. We will consider especially symbolic data mining techniques applicable to time series data where data are generated in streams. The internship will address the two following principal tasks:
• Knowledge discovery process about normal behaviour;
• The anomaly detection in new data.

Profil du candidat :
– Étudiant M2 ou 3ème ingénieur en Informatique
– bonnes compétences en programmation (Java, Python)
– connaissances en programmation par contraintes (la maîtrise des outils associés comme la bibliothèque Choco serait un vrai plus)
– une compétence en fouille de données et des méthodes associées
– goût pour la recherche et le travail collaboratif intra-équipe.

Formation et compétences requises :
– Étudiant M2 ou 3ème ingénieur en Informatique
– bonnes compétences en programmation (Java, Python)
– connaissances en programmation par contraintes (la maîtrise des outils associés comme la bibliothèque Choco serait un vrai plus)
– une compétence en fouille de données et des méthodes associées
– goût pour la recherche et le travail collaboratif intra-équipe.

Adresse d’emploi :
IMT Atlantique campus de Nantes

Document attaché : 202202010855_TASC_internship_2021.pdf