Basic creep of concrete under compression, tension and bending

Abstract

Investigations on concrete creep are often limited to the compression behavior due to the difficulty of performing tensile tests on cement-based materials. This paper describes the experimental setup developed to achieve direct tensile and bending concrete creep. The precautions taken to obtain relevant data are described. For comparison, tensile, flexural and compressive basic creep test were conducted in parallel. Although the approach is still controversial, the basic creep strain was determined by subtracting the shrinkage strain and instantaneous strain from the total strain. Results available for specimens subjected to 50% of the strength in tension or in compression are presented. The final discussion compares the basic creep under the different types of loading.Because of its poor strain capacity and low tensile strength, concrete is brittle and highly sensitive to cracking. In most cases, cracks can significantly reduce the lifetime of a structure by allowing the ingress of aggressive agents. Cracks have various origins, which are all associated with the existence of extensions: effect of external loading, physicochemical pathologies leading to the formation of new expansive products (Alkali-Aggregate Reaction -AAR, Delayed Ettringite Formation -DEF, corrosion). A non-uniform state of deformation associated with thermal or hydric gradients or restrained strain in bonded cement-based overlays induce built-in tensile stresses which can also initiate cracking. The heterogeneity of concrete also plays an important role in the mechanism of crack formation. Even if a specimen is tested in compression, the first damage that occurs at the paste-aggregate interface is the consequence of tensile stresses resulting from the differences between the two phase moduli (for paste and aggregate) and from the interfacial transition zone (ITZ) [1], which has limited mechanical performance. According to Liners [2], the application of a compressive preload on a specimen can generate damage in tension, resulting in a significant drop in the tensile strength of up to 50% depending on the level and the duration of loading application. Therefore, in order to predict the risk of cracking in concrete elements, it is important to focus on the mechanical properties of concrete in tension, and particularly on its delayed behavior, the effects of which (cracking or stress relaxation) are not yet well understood. Since Freyssinet highlighted the creep of concrete in France in 1912, regardless of Hatt's findings in the United States in 1907 [3], the phenomenon has been extensively studied, as shown by the numerous publications on this topic (see [4][5][6][7][8][9][10] for example).However, due to the difficulties of performing a tensile test on cement-based materials [11], particularly for fixing of the samples to the loading device [12] and measuring the values of the strains, which are too small for most extensometers to cope with, the majority of experimental studies on concrete creep have dealt with compressive creep. When experimen...

Influence of relative humidity on tensile and compressive creep of concrete amended with ground granulated blast-furnace slag

“…There is currently limited literature comparing experimental tensile and compressive creep data for the same test conditions and the few available are not consistent in their findings [3]. This is partly due to the difficulty in carrying out tensile tests on cement-based materials [3,5,11] as a result of their poor tensile strain capacity and brittle nature. Hence, investigations into concrete creep are often limited to the concrete's behaviour in compression.…”

“…As can be seen in Figure 8, for tensile creep tests, shrinkage values of the control specimens were of several magnitudes higher than their corresponding total strain values which at most points were negative. Meanwhile, there is an argument that shrinkage of a loaded specimen could be greater than that of a control sample [11,20] and as such, subtracting the shrinkage strain of a controlled specimen from the total deformation of a loaded sample to get creep values is only a convenient simplification which is not exactly true in reality. This argument is anchored on the existence of load-induced shrinkage and microcracking effect on loaded samples and this is particularly significant in tension samples which have more microcracks and hence more induced shrinkage than compression samples.…”

“…It is well established that concrete cracks even with loading below its load bearing capacity and according to Rossi et al [4] there is a direct relationship between stress level (both in tension and compression) in concrete and the density of microcracks. For a loaded specimen, these loadinduced microcracks which localize at the cement paste-aggregate interface [23] combined with the preexisting cracks created by the restrained shrinkage of the cement paste by the coarse aggregate [24], could cause internal moisture gradients [11] which develop additional internal stresses within the concrete and therefore amplify shrinkage of the loaded specimen. However, the difficulty is in quantifying the magnitude of this additional shrinkage and decoupling it from the actual stress-independent shrinkage.…”

“…The 51% and 68% RH tests were carried out in two separate creep rooms, while for both compressive and tensile creep tests, the 100% RH was obtained by sealing the specimens with double layers of self-adhesive aluminum foil. The sealing was to prevent any moisture exchange with the environment and is a well-established practice in creep and shrinkage experiments [4,11,12]. All compressive creep specimens were cured in a fog room at 99% RH and 20 o C until the beginning of the tests in the controlled environment creep rooms.…”

Time-dependent brittle creep-relaxation failure in concrete

“…However, despite major successes, the phenomenon of creep is still far from being fully understood (Bazant, 2001). Most of the previous studies on the creep behaviour of concrete have been performed at low stresses (Blechman, 2014;Brooks and Neville, 1977;Forth, 2015;Ranaivomanana et al, 2013). In contrast, the brittle time-dependent behaviour of concrete under high and sustained stresses is not well known.…”