• Guibert, Robin

Abstract

Surface roughness is linked to numerous physical phenomena, involved both in manufacturing (friction, visual aspect ...) and in day-to-day life (walking, object gripping ...). Roughness multiscale analysis is notably a powerful tool allowing the isolation of physical phenomena by their scale of application and the determination of their behavior laws. However, multiscale analysis requires a wider observable scale range than what is proposed by 2D or 3D profilometers to detect the relevant roughness scales. Stitching is an assembly technique combining elementary topographies, which overcomes the intrinsic limits of topography measurement apparatus and offers high precision 3D topographies over a large field of measurement. Although stitching algorithms are often used in surface metrology, their study is a true challenge as it is a fundamentally multidisciplinary problem, requiring knowledge in topography, materials science, metrology, mathematics, optimization process and computer science. This multidisciplinary aspect explains the collaboration between the mechanical and computer science departments during this research work. To answer the challenges of stitching, an approach in two axes is proposed. The first axis focuses on the study of the stitching procedure and on its integration in the workflow of the MorphoMeca team from the LAMIH, via the design of a software suite. Novel evaluation methods for repositioning errors and the quality of stitching are developed, thanks to the creation of a database dedicated to the testing of stitching algorithms. New stitching algorithms are also proposed, through the use of multimap and optimization processes. Finally, this first axis is concluded by the benchmarking of the performance of stitching algorithms regarding their algorithmic complexity and their ingestion capability for huge stitchings. A second axis proposes the application of the stitching to real-case studies. Multiscale analysis methods are then compared, thanks to the stitching and the study of polymer abrasion. Finally, a multiscale analysis of physical phenomena involved in the abrasion of nine polymers allows the identification of four wear mechanisms. Both axes highlight the interest of 3D topography stitching, either from an academic or an industrial perspective.

Topics

• impedance spectroscopy
• surface
• polymer