Etude du comportement deformationnel des betons de reparation

The majority of concrete road structures within Canada are exposed to aggressive environments that contribute to increase the rate of concrete deterioration. Often, rehabilitation is necessary to extend the service life of these structures. In most cases, the roads are repaired by replacing the damaged concrete surface with a new layer of concrete (overlay). The rigid concrete substrate restrains partly early-age deformations of the new concrete overlay and may cause the apparition of tensile stresses. This contributes to the formation of harmful cracks in the new overlay or debonding between the substrate and the overlay. To ensure the durability of these repairs, the first goal of this project was to study the creep and shrinkage behaviour of repair concretes. Whereas shrinkage and compressive creep tests are standardized and prevalent, only a few research laboratories can carry out more complicated tensile creep tests. Therefore, a flexural creep test was developed at the Ecole Polytechnique de Montreal and an analysis method was proposed to evaluate the tensile creep from the flexural and compressive creep test results. The second goal of this project was to study the deformation in a repaired hybrid element and ensure that its behaviour under loading is monolithic, i.e. that there is no debonding of the overlay.;For the second experimental phase, three reinforced concrete beams were made and left free of restraint for two months. The beams were then placed on two supports. One beam was used as a control reference while the surface of the other beams is hydrodemolished and repaired. One beam was repaired with an OPC overlay and the other with a UHPFRC overlay. After four months, neither cracking nor debonding was observed on the overlays. The three beams were then tested under cyclic and static bending loads. The beam repaired with an OPC overlay demonstrated a behaviour identical to the reference beam. However, the beam repaired with an UHPFRC overlay was more rigid than the reference beam and exhibits a strain-hardening behaviour. Both repaired beams behaved as monolithic elements. There was no debonding between overlay and substrate layers, even without the use of reinforcement or studs in the overlays.;Finally, a finite element model of a repaired beam was created and analyzed with CESAR. The calculation tools used for the analysis did not account for drying shrinkage or viscoelastic behaviour of the concrete. Therefore, the deformations of the repaired beam were not perfectly reproduced. However, if the choices made for this model are considered correct, the behaviour of UHPFRC with and without the contribution of viscoelasticity can be analysed. It can then be concluded that 79% of the deformations of UHPFRC overlay were relaxed at early age.;The first experimental phase of the project consisted in characterizing the shrinkage and creep behaviour of two concretes: an ordinary Portland cement concrete (OPC) and an ultra high performance fiber reinforced concrete (UHPFRC). Shrinkage, compressive and flexural creep tests were performed at Polytechnique. Total shrinkage deformation was larger for UHPFRC than for OPC. However, drying shrinkage was more significant than autogenous shrinkage in the case of OPC and the opposite was true for UHPFRC. Compressive and flexural creep tests confirmed the small impact of drying on the behaviour of UHPFRC. In fact, the difference between basic and total creep of this concrete was almost inexistent, in both compression and flexion. Basic tensile creep tests were then performed on the UHPFRC at the Laboratoire Central des Ponts et Chaussees (LCPC). With the compression, tensile and flexural creep test results for the UHPFRC, an analysis method was developed to obtain tensile creep data. The method was also applied to the OPC and then verified with a finite element model of the flexural creep test. The model was analysed using CESAR-LCPC, a finite element program.