Skip to Main content Skip to Navigation
Theses

An end-to-end approach of statistical tolerancing under industrial constraints: Contribution to the design/industrial virtual twin

Abstract : In the manufacturing process of a product, various assembly steps are necessary. Several types of requirements have to be met at each level and involve considerations about dimensional uncertainties on the parts to be assembled. Tolerancing is the activity in charge of the management of these uncertainties and takes place both in the product development phase and in the series production phase. In the context of the aeronautics industry, in particular with regards to tolerancing on aerostructures, specificities have to be taken into account in the development of adequate methods and tools. Prior to production, one of the main issues of tolerancing amounts to allocate tolerance limits suited to a given acceptable scrap rate. The aim is to allow the actors concerned with tolerance intervals to agree on a consistent and robust tolerance value. A statistical methodology based on a Chernov bound approach applied to a sum of uniform distributions is proposed. In the production phase, the availability of measurement data allows to refine the statistical tolerancing approach. The linear model often considered can be corrected to serve new approaches. A methodology to manage acceptance criteria on tolerance values is proposed, basing the decision support on risk concepts pertinently defined for industrial actors. Within the framework of the revision of tolerance sharing in an assembly, an optimization problem is formulated with appropriate industrial costs in order to propose the optimal tolerance re-sharing in a stack chain. Finally, the proposed methodologies are implemented in tools allowing industrial processing and end-to-end management of tolerances from elementary parts to final product assembly, thus contributing to the elaboration of the product virtual twin.
Document type :
Theses
Complete list of metadata

https://hal.archives-ouvertes.fr/tel-03258192
Contributor : Ambre Diet <>
Submitted on : Friday, June 11, 2021 - 1:38:51 PM
Last modification on : Wednesday, June 16, 2021 - 3:27:18 AM

File

these_pdf.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : tel-03258192, version 1

Citation

Ambre Diet. An end-to-end approach of statistical tolerancing under industrial constraints: Contribution to the design/industrial virtual twin. Mathematics [math]. IMT - Institut de Mathématiques de Toulouse UMR5219, 2021. English. ⟨tel-03258192⟩

Share

Metrics

Record views

34

Files downloads

22