Thesis defense by Hassan CHOUHAD

Summary

In traditional manufacturing, metrology is an essential element in ensuring quality at the end of the production line. The innovative concept of smart manufacturing has led to a repositioning of metrology, which has become proactive at the very heart of production in order to manufacture a compliant first part from the outset. The purpose of this thesis is therefore to propose a methodological approach for the development of a proactive system, augmented by artificial intelligence (AI) models, for monitoring the conformity of a product with specifications during machining and characterizing its defects.  To this end, an initial study of surface appearance was carried out by collecting high-resolution images of coated and cut copper wires that may have defects. The images, taken by a computer vision system based on chromatic confocal imaging, were used to generate different artificial intelligence models. This process involves segmenting and classifying the defects observed. By comparing the accuracy and processing time of the AI models, transfer learning using the mobile-net model showed better performance. In order to broaden the study of surface quality assessment, surface profile measurements were performed on a machine tool using non-contact chromatic confocal sensors. Two approaches were implemented: i) milling aluminum without cutting tool wear signature and ii) milling titanium taking into account the cutting tool wear signature. In both cutting configurations, the machining parameters, surface roughness profiles, and material removal forces were recorded to build a database for training prediction models using machine learning. The results showed that the XGboost model performed best in terms of prediction for both scenarios. Considering the cutting time in titanium milling, the ARIMA time series prediction model was applied to track the evolution of roughness as a function of tool wear. The autoregressive integrated moving average analysis made it possible to track the evolution of roughness as a function of the wear signature.

Keywords

Intelligent machining, proactive quality control, artificial intelligence, chromatic confocal, in-machine metrology.

List of jury members

• Mr. Hassan ZAHOUANI, University Professor, LTDS, Ecole Centrale Lyon (rapporteur)
• Mr. Roberto MARTINS DE SOUZA, University Professor, LFS, Escola Politécnica USP Brazil (rapporteur)
• Mr. Mohamed EL MANSORI, University Professor, MSMP, Arts et Métiers examiner)
• Mr. Ricardo KNOBLAUCH, Lecturer-Researcher, MSMP, Arts et Métiers examiner)
• Mr. Arnaud GOTLIEB, Research Professor (HDR), VIAS, Simula Research Laboratory. Norway (examiner)
• Mr. Satish BUKKAPATNAM, University Professor, TEES, Texas A&M University. USA (examiner)                   
• Mr. Cosimi CORLETO, Engineer, CEO of STIL Marposs Aix-en-Provence (guest)