Variability of silica fume concrete and its effect on seismic safety of reinforced concrete buildings

Design of structures made using silica fume (SF) concrete to an acceptable level of safety requires the probabilistic evaluation of its mechanical properties. An extensive experimental program was carried out on compressive strength, flexural strength, and tensile splitting strength of SF concrete. Seven concrete mixes with different proportions of SF were designed to produce 490 concrete samples. The probabilistic models to describe the variability of the mechanical properties of SF concrete were proposed. Two-parameter probability models such as Weibull, normal, lognormal, and gamma distribution were considered for the representation of variability. The probability distribution models were selected based on goodness-of-fit tests such as the Kolmogorov-Sminrov (KS), chi-square (CS), and log-likelihood (LK) tests. The results obtained from the models are useful for description of the variability of selected mechanical properties of SF-incorporated concrete. This study proposes the lognormal distribution function as the distribution model that most closely describes the variations of different mechanical properties of SF concrete from a practical point of view. Further, the performance of typically selected buildings using SF concrete was evaluated through fragility curves and reliability indices incorporating the proposed probability distributions and variability of compressive strength property. It was found that 15%-25% of partial replacement of cement with SF may yield better performance of the frames.