Effects of Steel Fibres on Cracking in Reinforced Concrete

Doktorsavhandling, 2011

ABSTRACT Although it is well known that fibre reinforcement acts as a crack arresting agent, there is still a need for deeper knowledge of the actual cracking behaviour, especially regarding cracks with widths smaller than 0.3mm. Today major fibre applications are as a replacement for the welded mesh in industrial floors, and as reinforcement in sprayed concrete. However, other applications exist and are investigated. By combining experiments with finite-element analyses, the effects of fibres on cracking in conventionally reinforced, self-compacting, steel-fibre-reinforced concrete (SCSFRC) were studied. When studying the beginning of the cracking process, the tensile softening behaviour (-w relationship), and the bond stress-slip behaviour, which are the ones mainly affecting the cracking, are clearly of interest. Contradictory information on the effect of fibres on bond behaviour was found in the literature. Pull-out tests with short embedment length were thus carried out. The -w relationship may be obtained indirectly by inverse analysis, e.g. from wedge-splitting tests, or directly, from uniaxial tension tests (UTT); both approaches were used in this work. To investigate the cracking process, tension tests of tie elements were carried out, where, in addition to the load-deformation curves, a full-field strain measuring technique using Digital Image Correlation (DIC) was used to monitor the surface cracking. It was found from the pull-out tests that, for the type and amount of fibres used here, the bond properties at the interface layer were neither reduced nor improved. There were indications, however, that the initial stiffness of the bond stress-slip curves was increased by the self-compacting concrete. The UTT and the tie element testing showed that the scatter was quite high regarding the number of fibres in a cut cross section. It was seen that fibre reinforcement markedly improves tension stiffening and, at a given load, the characteristic crack width is greatly reduced compared with plain concrete. The DIC gave good insight into the surface crack initiation and enabled the tension stiffening to be quantified by relating it to the characteristic crack widths. In addition, it was seen that the cracking load and first-peak tensile stress increased with an increasing amount of fibres. The Finite element analyses of the beams and the tie elements revealed that the methodology used was versatile. It was found that the smeared crack model did not yield crack localization for materials with high fibre content (Vf > 0.5%) if homogenous material properties were assumed. Instead a semi-meso approach was used; properties for plain concrete were assigned to randomly designated parts of the elements, while the remaining elements were assigned modified tensile properties. The modified properties were increased so that the average -w curve of one cross section corresponded to the average curve from the UTT. With the new approach, the load-elongation response agreed better with the experiments; crack localization was obtained and crack widths could be reasonably reproduced.