Skip to Main content Skip to Navigation
Theses

Traitements géométriques et fouille de données pour la reconnaissance d’entités d’usinage complexes

Abstract : In response to competition and new industrial challenges, companies are forced to be more and more efficient, productive and competitive. Managing industrial know-how and the data flow of the manufacturing digital chain must be explored in order to shorten the industrialisation time while ensuring better quality.In this context, this thesis focuses on digital chain data exploration for the capture of good practices in NC machining using a feature-based approach. Several issues related to machining feature characterisation and digital chain data exploitation for machining process knowledge reuse have been identified.To address these issues, a new characterization of multi-level complex machining feature has been proposed. The proposed approach is characterized by a hierarchical structuring of digital chain data and a mapping between the geometrical and machining data. A statistical analysis is then carried out to analyse and exploit this data. Curvature-based segmentation and statistical clustering of machining data were combined to define new machining regions based technological segmentation approach. These regions were then used to characterize the machining feature and thus ensure the reuse of machining data through a feature based and a region based approach exploiting similarity measures a similarity measure. Finally, the developed approach was applied on an industrial case in aeronautics.
Complete list of metadatas

Cited literature [148 references]  Display  Hide  Download

https://tel.archives-ouvertes.fr/tel-02881116
Contributor : Abes Star :  Contact
Submitted on : Thursday, June 25, 2020 - 2:43:41 PM
Last modification on : Friday, October 23, 2020 - 5:03:55 PM
Long-term archiving on: : Wednesday, September 23, 2020 - 4:44:09 PM

File

83171_BENDJEBLA_2020_archivage...
Version validated by the jury (STAR)

Identifiers

  • HAL Id : tel-02881116, version 1

Citation

Soumiya Bendjebla. Traitements géométriques et fouille de données pour la reconnaissance d’entités d’usinage complexes. Génie mécanique [physics.class-ph]. Université Paris-Saclay, 2020. Français. ⟨NNT : 2020UPASN002⟩. ⟨tel-02881116⟩

Share

Metrics

Record views

136

Files downloads

74