Towards a better understanding of the ultimate behaviour of lightweight aggregate concrete in compression and bending
MetadataVis full innførsel
Ordinary normalweight concrete has a low strength-to-weight ratio compared to steel. This may prevent the use of concrete for the construction of tall buildings, floating structures, longspan bridges, etc. One way of overcoming this problem is by reducing the weight of the concrete by replacing the normal gravel aggregates by lighter types. Unfortunately, there is a general scepticism regarding the use of lightweight aggregate concrete for structural applications. This concern is attached to the more brittle post-peak material behaviour and smoother crack surfaces of these concretes compared to ordinary gravel concrete. However, in this work, the post-peak material behaviour and the force transfer across cracks are considered to be unimportant regardless of the weight of the concrete. The working hypothesis has been that the three key material characteristics generally dictating the ultimate response of concrete structures are: the large effect small secondary stresses have on the compressive strength; the abrupt increase of the transverse expansion at a stage close to, but not beyond, the peak stress level; and the rapid unloading of the material beyond the peak stress level. As a consequence of these features, the strength and especially the ductility of structural concrete members depend on local triaxial stress conditions that inevitably develop in the compressive zone just prior to failure rather than stress-redistributions owing to post-peak material characteristics as commonly believed. This hypothesis has previously been used with success to predict and explain the behaviour of normalweight concrete in the ultimate limit state. In this work, it is applied to lightweight aggregate concrete. In its verification process, results from experimental programs reported in the literature have been carefully examined. In conclusion, the test results seem to support the working hypothesis. A three-dimensional nonlinear finite-element code, with a novel failure criterion accounting for the density of the concrete, was also developed for verification purpose. The results from the analyses demonstrated that structural collapse always occurred before the strength criterion in compression was exceeded anywhere within the structure. Since only the compressive region of the failure envelope varied with the weight of the concrete, there was no effect of reducing the density of the concrete in the analyses. This was in accordance with experiments, where usually no dramatic differences in the strength of the members were observed. The somewhat lower ductility of the members with decreasing density can be explained by a lower degree of stress triaxiality in the compressive zone. This was considered to be a result of the often quite modest transverse expansion of the lightweight aggregate concretes prior to failure. However, since the lack of experimental data led to the assumption that the deformational behaviour was the same as that of their normalweight counterparts, this effect could not be catered for by the analyses. Nonetheless, the theoretical foundation, examination of test results and numerical analyses in this study seem to collectively support the working hypothesis. The research work presented herein is therefore considered to contribute to a better understanding of the ultimate behaviour of lightweight aggregate concrete at both the material and the structural level.