Shrinkage behaviour of self-compacting concrete with a high volume of fly ash and slag experimental tests and analytical assessment

•Free and restrained shrinkage of SCC under different curing conditions was tested.•The high slag and fly ash volume was used in SCCs with the same paste volume.•Fly ash replacement was found as the most effective in reducing free shrinkage.•Slag replacement was found as less effective in reducing free shrinkage.•The difference in the shrinkage at the sample thickness was calculated from the bending drain. This study assesses the effectiveness of the high-volume replacement of Ordinary Portland cement (OPC) by non-clinker binders in terms of shrinkage behaviour. Fly ash (FA) or granulated blast furnace slag (GBFS) was used as a partial replacement for OPC at 60% by volume of the paste in self-compacting concrete (SCC). SCC having both FA (30%) and GBFS (30%), as well as SCC only with OPC, were also investigated. In all tested self-compacting concretes, the volume of the paste, which determines the shrinkage strains, was preserved. The shrinkage behaviour of SCCs was tested in shrinkage drains, bending drains and shrinkage rings. The total effect of autogenous and drying shrinkage under different drying conditions, also with the additional heating of the concrete samples was analysed. The strains were registered in the early stage of concrete curing, directly after mixing all concrete components, as well as in the later period, up to 28 days. The results indicate that the applied high volume of supplementary cementitious materials in self-compacting concrete (SCC) significantly reduces the shrinkage strains in comparison to SCC only made of OPC. Simultaneously, fly ash replacement is found as the most effective in reducing the free shrinkage of self-compacting concrete. The registered shrinkage development also shows the significant role and the level of very early shrinkage, which decided the values of 28-day strains. In the analytical part, the results from three used testing devices were compared with each other to confirm the effect of the non-clinker components on the shrinkage behaviour. Additionally, the results have been related to the Eurocode 2 and ACI shrinkage models, as well as the method for calculation of the differences in the shrinkage strains between the bottom and top surface of the sample based on measurements in the bending drain was presented.