Data mining methodology for on-line process control

Description de l'offre

Country:

FRANCE

Region:

LORRAINE

Sector:

ArcelorMittal Europe - Long Carbon

Unit:

LCSE BARS

Site:

ArcelorMittal Gandrange

Field of activity:

Research and Development / Metallurgy Innovation

Title of internship:

Data mining methodology for on-line process control

Subject of the internship:

The quality of the as-cast semi-products is determinant for the properties of the final products. In this framework, the R&D centre of Gandrange works in close collaboration with several casting plants, to define appropriate process parameters ensuring quality and productivity. Thus, deviations or drifts of process conditions can have detrimental consequences on production performances and even safety… The objective is to set up a statistical methodology dedicated to detect such deviations/drifts and anticipate production issues. This work is based on the process records that are continuously taken by the plants during production. The mission will be divided in several main aspects, to be developed in parallel: - To get familiar with the process of continuous casting by a basic training - To perform a literature review of statistical methods compatible with the present industrial context (online / offline) - To define the most appropriate methodology of analysis for casting records - To define indicators for process characterization (ability to reach target conditions, stability of the regulation, repeatability of the process conditions…) - To transfer the methodology for live monitoring of process records and identification of drifts All along the mission, the results must be regularly capitalized in formal reports and the developed knowledge must be regularly spread within the local R&D team. The trainee will join the cluster “Process Modelling & Instrumentation”, which is developing activities in statistical analysis. His/Her expertise will contribute to strengthen the skills of the R&D centre of Gandrange (dedicated to long products) and to propose innovative solutions to plants. He/She will work with the statistical software STATISTICA. Some exchanges will be possible with ArcelorMittal plants and/or R&D centres.

Profil recherché

Applicant Profile:

The trainee owns an advanced knowledge in the fields of data analysis and statistics. He/She has good communication skills, to facilitate his/her integration in the cluster Process Modelling & Instrumentation and to collaborate with other teams of Arce

Education:

Master degree in Mathematics and statistical analysis

Level of education:

MASTER (Bac+4/5)

Language:

French: fluent - English: operational

Duration of the internship:

6 months

Period of the internship:

From September/October

Managers of the internship:

Julien COLOMBEAU
Tel:
Email: ************@*******.**

© 2007 -