Performance-based design of self-compacting fibre reinforced concrete

Abstract

The development of self-compacting concrete (SCC) marks an important milestone in improving the product quality and efficiency of the building industry. SCC homogenously spreads due to its own weight, without any additional compaction energy and does not entrap air. SCC improves the efficiency at the construction sites, enhances the working conditions and the quality and the appearance of concrete. Fibres bridge cracks and retard their propagation. They contribute to an increased energy absorption compared with plain concrete. Self-compacting fibre reinforced concrete (SCFRC) combines the benefits of SCC in the fresh state and shows an improved performance in the hardened state compared with conventional concrete due to the addition of the fibres. Due to its special characteristics new fields of application can be explored. This thesis provides tools and models to optimise SCFRC in the fresh and the hardened state. Relevant literature and the experience gained during the experiments are summarised; various experimental studies were performed. The objectives of this research project were to optimise SCFRC in the fresh and the hardened state and to model the behaviour in order to provide reliable design tools; mainly steel fibres were applied. SCFRC can be optimised for various purposes: to apply the highest possible fibre content, to obtain the highest performance-cost ratio, to design the granular skeleton for the highest packing density or to produce with the lowest possible material costs. The effect of the production process on the characteristics of SCFRC was also studied. To introduce into the theoretical background of SCFRC in the fresh state, selected literature on SCC, especially with regard to the packing density and the effects of the fibres on workability is reviewed. Test methods are described; previous experience with SCC and fibre reinforced concrete (FRC) in the fresh state and approaches to model the behaviour are summarised. The packing density of the aggregates and the fibres of SCFRC determine the amount of cement paste that is required to fill the interstices of the granular skeleton. In order to predict the packing density of the granular skeleton, the 'Compressible Packing Model' was used and calibrated with the applied materials. Predictions from five approaches to include the steel fibres were compared with results of experiments to obtain the best accuracy. The accuracy of the predictions depends on the composition of the aggregates. The simulations had an average error close to 2% for optimised mixtures with aggregates up to 8 or 16 mm; predictions of the packing density of mixtures with smaller maximum aggregate sizes were less accurate. Sixteen stable SCCs at defined characteristics in the fresh state were used to study the effect of the type and the content of the steel fibres; in total 121 mixtures were tested. The maximum aggregate size of the reference mixtures was 4, 8 or 16 mm; the volumes of the paste as well as the sand to total aggregate contents were varied. The fibres affect the characteristics of SCC in the fresh state: the slump flow decreases and the yield value, the plastic viscosity (the resistance to flow) and the bar spacing required to avoid blocking increase compared with a reference SCC. For each reference mixture and the applied fibre type the maximum fibre content was determined. Due to the higher density of the steel fibres, segregation might occur even when the aggregates are homogenously distributed. Based on experimental results, models were developed, which quantify the effect of the steel fibres and with which key characteristics of SCFRC in the fresh state can be predicted. Criteria are proposed to design and to characterise SCFRC; basic principles to optimise SCFRC are described. SCFRC was also tested in the hardened state. A summary of the literature discusses mechanical characteristics of conventional and self-compacting concrete reinforced with steel fibres and the effect of the orientation and the distribution of the fibres. Bending tests were performed on seventeen optimised mixtures, which were selected from the studies on the characteristics in the fresh state. The mixtures differed in the compressive strength class, the type and the content of the steel fibres and the manner of manufacturing of the specimens. The variation of the maximum flexural strength was in each case below 12%, which is significantly lower with what is usually obtained for steel fibre reinforced concrete (SFRC). The comparison of the bending behaviour of SCFRC and SFRC indicated significant differences concerning the performance and the variation in the test response: SCFRC performed much better. Two additional studies were performed to determine the origin of the differences. First, the orientation numbers of fibres of the cross-sections of the beams were determined by an image analysis. The fibres in SCFRC were found to be more favourably aligned into the direction of the flow. Second, a comparison between the pull-out behaviour of single fibres from SCC and conventional concrete showed that in most cases higher pull-out forces were obtained with SCC. The single fibre pull-out test might give a better indication of the actual performance of a fibre in SCC than in conventional concrete. Entrapped air and neighbouring fibres affect the performance of a fibre in SFRC more than in SCFRC. By applying the multi-layer procedure of Hordijk an inverse analysis of the bending tests was performed. Based on the experimental results of the bending tests a combined stress-strain/stress-crack width model for SCFRC in tension was developed. The model was calibrated with results from fifteen mixtures, which contained hookedend steel fibres. The proposed stress-strain/stress-crack width model distinguishes three tensile regions: the elastic and the reduced elastic strain ranges and a softening branch. The difference between results of simulations with the combined tensile model and experimental results is in average below 8%. Three full-scale applications with SCFRC are discussed: sheet piles, tunnel segments and large beams. The focus of these studies was on the orientation and the distribution of the steel fibres. Different techniques were applied to quantify 'orientation'. SCFRC was found to be an inhomogeneous material; the fibres are rarely randomly oriented. The preferred orientation of the fibres can be considered as a benefit or, the opposite as an intrinsic weakness of SCFRC. The studies on sheet piles and tunnel segments demonstrated that applications with SCFRC can be economical, offer products with interesting characteristics and present innovative solutions. The production process is an important factor, which affects the performance of SCFRC.