This research has investigated the behaviour of joints and cracks under single and multiple cycles of load. This provides an increased understanding of concrete slab on grade performance, enabling more effective design and monitoring procedures. Examination of the geometry of cracks and joints within concrete slabs on grade has demonstrated that the commonly assumed parallel formation is erroneous. Measurements using embedded strain gauges, coring and surface profile levelling have uncovered that a high percentage of joints will contain larger crack widths at the surface than at the base, caused by differential shrinkage. The opening itself is relatively linear; however, the top 50mm of the slab is prone to a higher gradient of movement due to the increased drying effect towards the surface. A series of deflection tests using a Falling Weight Deflectometer and Prima dynamic plate enabled slab response under load to be evaluated. Four sites were examined in total and correlations found between: load transfer, load step, edge cantilever and crack geometry. This produced valuable information regarding the influence of load transfer and crack width on the overall slab behaviour. Foundation voiding and crack face free slip was also shown to influence deflection magnitude. A small-scale test facility was developed for the assessment of deterioration in various 'V' shaped and parallel crack widths under high cycle loading. The data demonstrated that joint/crack failure contains four distinct phases of deterioration, each of which is controlled by a different mechanism. 'V' shaped cracks produced a much greater load transfer than that of a parallel crack with the incorporation of A142 mesh and steel fibres reducing differential displacement. Load magnitude and aggregate size were also shown to have significant effects. The value of reinforcement was found to assist with serviceability requirements, keeping displacement within acceptable levels and preventing the onset of serious degradation A finite element model was developed to enable the load transfer mechanism results from the laboratory test to be used in the assessment of full slab response. Simulations of field testing produced a series of lower bounds in respect to deflections and the associated response calculations. Theoretical behaviour of a typical slab was assessed with subbase support, joint stiffness, slab thickness and the incorporation of a subbase, found to be highly influential in reducing slab deflections. The three main sections of work comprising site data collection, laboratory testing and Finite Element modelling have been used together to provide a much greater understanding of the influence of cracks and joints. This has included the deterioration of cracks over time and an examination of how this and other site-based factors affect overall slab behaviour.

Cette contribution, en s’appuyant sur expérimentation et modélisation numérique vise à une meilleure compréhension du comportement des dalles en béton armé sous sollicitations de cisaillement. Une campagne expérimentale a été réalisée sur des dalles épaisses à pleine échelle de centrales nucléaires. Ces dalles sans armatures d’effort tranchant sont soumises à une sollicitation de cisaillement en chargement quasi-statique. Les essais sont réalisés en faisant varier différents paramètres qui peuvent influencer le comportement au cisaillement. Sont ainsi étudiés : résistance en compression du béton, épaisseur, taux d’armatures longitudinales et transversales, taille des granulats, longueur de la plaque de chargement. L’influence des efforts de membrane, de compression ou de traction, sur le comportement au cisaillement a également été analysée. Les résultats des essais sont ensuite comparés aux prédictions des codes de calcul. Ces résultats ont d’abord permis d’apporter une réponse aux divergences qui existent entre l’Eurocode 2 et l’Annexe Nationale Française quant à la prédiction du cisaillement. Ont également été évalués le niveau de précision donné par d’autres normes de dimensionnement au cisaillement: la norme américaine ACI 318-14, le code nucléaire AFCEN ETC-C 2010, le fib-Model Code 2010 et l’approche par la théorie de la fissure critique de cisaillement CSCT. Ensuite est évalué la possibilité d’analyses non-linéaire par élément finis (EF) pour reproduire le phénomène du cisaillement dans les dalles. Un modèle de béton élastoplastique avec endommagement est combiné à une analyse quasi-statique à schéma de résolution explicite. Des lois de comportement non linéaires appropriées du béton avec des comportements post-pic associés à un critère énergétique ont été considérées. La bonne concordance entre le modèle proposé et les résultats expérimentaux en termes de résistance au cisaillement et de modes de rupture permet de valider la modélisation proposée. Une étude paramétrique a été réalisée sur la base du modèle proposé avec les mêmes propriétés mécaniques de béton. Des lois simplifiées permettant d’estimer les capacités en cisaillement en fonction des différents paramètres étudiés sont finalement proposées. / This study, based on experiments and numerical modeling, aims at a better understanding of the shear behavior of reinforced concrete slabs. An experimental campaign was carried out on full-scale thick slabs typical of nuclear power plant slabs. These slabs without shear reinforcement are subjected to a quasi-static shear loading. The tests are carried out by varying different parameters that can influence the shear behavior: the concrete compressive strength, the slab depth, the bottom longitudinal and transverse reinforcement ratio, the concrete aggregate size, the loading plate length. The influence on shear behavior of compression or tension membrane forces has also been analyzed. The results of tests are then compared with the predictions of the calculation codes. These results first of all helped to answer the differences between the Eurocode 2 and the French National Annex concerning the prediction of the shear capacity of reinforced concrete slabs. The level of accuracy given by other shear dimensioning standards was also assessed: The American standard ACI 318-14, the AFCEN ETC-C 2010 code used for nuclear buildings, the fib-Model 2010 and the Critical Shear Crack Theory. Next, we evaluate the possibilities of a non-linear finite element analysis (EF) to reproduce the phenomenon of shear in slabs. An elastoplastic concrete model with damage was used and combined with a quasi-static analysis using an explicit resolution scheme. Appropriate nonlinear behavior laws of concrete with post-peak behaviors associated with an energy criterion were considered. The good agreement between the proposed model and the experimental results in terms of shear strength and failure modes allowed validating the proposed modeling. A parametric study was conducted based on the numerical proposed model with the same mechanical properties of concrete. Simplified laws allowing estimating the shear capacities according to the different parameters studied are proposed.