Comportement au gel interne de bétons sous contraintes mécaniques

The literature relative to the freezing of concrete is based on freezing-thawing cycle data obtained from specimens which are not subjected to mechanical stresses. In the present study, the initial tensile stresses of concrete are taken into account. Data were developed on control mortars, containing a smaller amount of cement than standard ones in order to represent the actual content of cement in concrete. Two binders, both with and without an air-entraining agent, were utilized: ordinary Portland cement and metakaolin-blended cement, in which the ordinary Portland ciment/metakaolin ratio was 80/20. Prismatic specimens of mortars (40×40×160 mm³) were subjected, in a novel manner, to freezing-thawing cycles in the presence of variable flexural stress rates: 0 to 85% of the rupture load. In order to separate the freezing effect from the load effect, measurements were conducted both in a cold-heat chamber and in a room where the temperature was maintained at 20°C. The strength evolution was analysed and some microstructural properties were assessed (porosity, hydrates). The results obtained show that this new method of testing allows the acceleration of damage when the rate of flexural stress reaches about 60% of the rupture load. The mortar containing metakaolin performs as well as mortars containing air-entraining agents (ordinary Portland ciment-air-entraining agent; metakaolin-air entraining agent). The microstructural analyses explain the behaviour of the different mortars.