Comparative investigation on tensile behaviour of UHPFRC

The recent advent in the field of cementitious composites with the emergence of high and ultra-high performance fiber reinforced concretes (HPFRC and UHPFRC) is considered a great opportunity in infrastructure accelerated construction, strengthening and repair. These materials draw their exceptional mechanical performance from two characteristics: the densely packed microstructure, and the ability of the materials to sustain their tensile resistance after cracking and to strain-harden over a significant range of tensile strain when the appropriate number of fibers is used. In structural design, however, the tensile resilience of these materials can be considered only if it can be characterized by standards and repeatable tests. Four-point bending tests (FPBT) on prism samples are used widely due to their perceived simplicity, particularly for quality control. Although these tests have been introduced in standards as the recommended option, experience shows that in practice they lead to experimental scatter and variability owing to spurious interactions between the specimen and the testing setup. This study explores experimentally, through a round-robin test, the effect of these interactions on the dispersion of the results. Three preblended commercial mixes were considered; identical specimens were tested in three different laboratories using their available hardware to conduct the quality control tests. The variability in the custom details of the FPBT setups and the uncertainties introduced by the operator, as well as the effect of casting methodology on flexural strength were investigated. Bilinear tensile stress-strain and stress-crack mouth opening relationships of the tested materials were derived using the inverse analysis procedures as prescribed in Annex 8.1 of CSA-S6 (Canadian highway bridge design code-annex 8.1 fibre reinforced concrete, Canadian Standards Association, Toronto, 2019) and Annex U of CSA-A23.1 (Ultra-high-performance concrete-UHPC, Canadian Standards Association, Toronto, 2019). Results indicate that cracking strength of HPFRC and UHPFRC derived from inverse analysis was generally higher than the conservative empirical lower bound estimates, whereas a significant difference is evident between results from the three laboratories participating in this round robin testing program. Despite the scatter, in all cases the inverse analysis indicated that the materials exhibited tension hardening behaviour.