Abstract
Steel-fibre-reinforced concrete (SFRC) has been recognised as an effective solution to resist impact loading on structures. The reliable application and efficient design of SFRC structures depends on the knowledge of its mechanical properties. Since many important factors, including the locations and orientations of fibres and aggregates in concrete and the material properties of concrete matrix, are intrinsically random, the mechanical properties of SFRC present a high level of randomness. To accurately quantify them, effective statistical techniques are indispensable. Using traditional statistical techniques, a large quantity of data, from either experiments or numerical simulations, are needed to derive the correlation between the mechanical properties and the random factors. However, both ways are time-consuming and costly. Therefore, very little information regarding the statistical mechanical properties of SFRC can be found in the current literature. In this study, a kernel-based nonparametric statistical method is proposed to derive the statistical mechanical properties of SFRC with limited number of data. The behaviours of SFRC with randomly distributed spiral-shaped fibres and aggregates under impact loading are simulated using commercial software LS-DYNA. The simulation accuracy is validated by the experimental results. The influences of various volume fractions of fibres on dynamic increase factor (DIF) of the tensile strength of SFRC specimens under dynamic loadings at different strain rates are quantified through a prediction model obtained from kernel regression. The results demonstrate that the proposed method is able to estimate the DIF value of SFRC based on the tensile strength and strain rate, and to derive the statistical mechanical properties of SFRC.