Lors de la mise en oeuvre du béton dans les coffrages ou lors du pompage du béton, une couche, appelée couche limite, constituée d’eau et de fines, se forme près de l’interface. Diverses études ont montré que le frottement à l’interface béton-paroi dépend essentiellement des propriétés rhéologiques de la couche limite, de sa composition et de la rugosité de la paroi. L’objectif principal de cette étude est de comprendre les phénomènes se produisant à l’interface à l’échelle des grains. Toutefois, compte tenu de la complexité de deux corps en contact : couche limite et paroi, le système étudié est simplifié. La complexité du matériau béton a ainsi été simplifiée, en considérant que ce milieu est constitué de grains sphériques monodisperses et sans cohésion. Par ailleurs, la rugosité de la paroi formée en réalité par des aspérités irrégulières, est représentée par des géométries simples. L’étude repose sur l’utilisation de deux modèles, un modèle analogique et un modèle numérique.L’approche analogique est constituée d’un dispositif expérimental monté sur le tribomètre. Le cisaillement s’effectue en déplaçant, une baguette crénelée placée sous un empilement de billes confiné par une masse. Trois baguettes représentant trois rugosités différentes sont utilisées. Les résultats expérimentaux montrent que les conditions à l’interface (vitesse, rugosité, pression) influent sur le frottement et sur les mécanismes à l’interface.L’approche numérique est basée sur la méthode des éléments distincts. La rugosité de la paroi est modélisée par un assemblage de sphères collées sur un plan mobile. Les résultats obtenus montrent une bonne concordance entre les approches numérique et expérimentale. De même, l’étude numérique réalisée en conditions périodiques et pour différentes rugosités montre que les conditions à l’interface influent sur le frottement et sur le comportement du milieu près de l’interface (profils de vitesses, épaisseur de la couche cisaillée, etc.). / When fresh concrete is placed into formworks or during concrete pumping, a layer, called boundary layer, consisting of water and fines, is formed at the interface. Various studies have shown that the friction between the concrete and the wall (form or pipe) is closely linked to the rheological properties of the boundary layer, its composition and the wall roughness. This study aims at understanding the wall-concrete interface phenomena at a grain scale. However, due to the complex nature of the two bodies in contact: boundary layer and wall, a simplified system has been considered. The concrete has been simplified by assuming that it consists of monodispersed spherical grains and without cohesion. Moreover, the irregular asperities of the wall roughness are represented by simple geometries. The study is based on the use of two models, an analog model and a numerical model.The analog approach consists of an experimental device fitted on the tribometer. The test consists in moving rough plates under stacks of confined monodispersed beads. The roughness of the plate is represented by crenels. Three plates representing different roughnesses are used. The experimental results show that the friction and the interface mechanisms are affected by the interface conditions (velocity, roughness, pressure).The numerical approach is based on the Discrete Element Method (DEM). The wall roughness is modeled by an assembly of spheres fixed on a moving plane. The results obtained by numerical simulations compared with those of the analog model show a good confrontation. As well, the numerical study carried out with periodic boundary conditions and for different roughnesses shows that the interface conditions affect the friction and the behavior of the medium near the interface (velocities profiles, thickness of the shear layer, etc.).

The large majority of New Zealand's road network is constructed from thin surfaced unbound flexible pavements where a granular layer provides the main structural strength of the pavement. The current New Zealand empirical design theory states that permanent deformation should largely be attributed to the subgrade and that shape loss in the granular layers is simply a consequence of a previously deformed subgrade. However, recent research and field trials have indicated that basecourse shear strains may be a large contributor to rutting in unbound granular layers. The purpose of this investigation was to determine whether the shear strains induced under heavy vehicle loads can be accurately measured using in-situ induction coils and whether the shear strains are related to permanent pavement deformation. In this investigation a rosette configuration of free floating induction coils was designed to measure principal basecourse shear strains. The principal strains were then used to construct Mohr's circle of strain in order to calculate the maximum shear strain occurring in the granular layer. The rosettes were installed in two full scale test pavements at the Canterbury Accelerated Pavement Testing Indoor Facility (CAPTIF). The pavements were loaded with an 8 tonne dual wheel axle load for 1 million and 600,000 load applications respectively and strain and rut depth testing occurred periodically throughout the test life. The research showed that the rosette coil arrangement was a feasible and accurate device for measuring in-situ shear strains in granular pavement layers. Finite element modelling confirmed the accuracy of the system. The results from the two CAPTIF pavements showed that there was a strong linear relationship between the magnitude of the basecourse shear strain and the rut depth at the end of the post construction compaction period. The investigation also showed that shear strain magnitudes in the region of 5000µƐ result in rapid shear failure in the granular layer. In addition, after the post construction compaction period had finished, the rate of change of shear strain was proportional to the rate of change of rut development. The results indicated that there was approximately a 4:1 ratio between the rate of change in rut depth and the rate of change in shear strain after the initial post construction period. Investigations into the effect of load magnitude on the magnitude of the basecourse shear strain showed that a linear relationship existed between the two parameters. Further to this, load location testing revealed that for a dual wheel configuration, 50mm of lateral wheel variation either side of a point of interest was the maximum allowable movement that would result in similar strain measurements. The research highlighted the dominance of the longitudinal tensile strain and shear strain over the vertical compressive strain within granular layers. As a result, these pavement responses should be considered in further granular pavement research in addition to the commonly used vertical compressive strains.