Optimization of concrete mix design based on three-level separation distance of particles

Abstract

This study proposed a modified mix design approach for producing concrete with reduced cement content and CO2 emissions. Three key design parameters: BSD (binder separation distance), FSD (fine aggregate separation distance) and CSD (coarse aggregate separation distance), from paste, mortar and concrete level are suggested based on the particle packing theory. The potential synergistic effects of BSD, FSD and CSD on fresh behavior, compressive strength and binder efficiency of designed concrete are evaluated. The results show that by considering the effect of particle specific area and particle size distribution, three proposed parameters exhibit better eco-efficiency and high correlations with concrete properties. The experiments together with numerical analysis show that the correlations coefficient of BSD, FSD and CSD together on slump, flow rate and compressive strength are 0.979, 0.923 and 0.913, respectively, indicating that these three values can be used together to tailor the properties of concrete. It is found that BSD is the most important parameter in determining the mechanical property, with a high correlation of 0.97. While FSD and CSD together are more sensitive on fresh behaviors. In addition, a prediction model based on machine learning was established, the errors between the measured and predicted values are within 10% and the binder efficiency in mixtures designed with three scale separation distances were increased by 7% in general, which validates the feasibility of using particle separation distances as key design parameters, and the applicability of this prediction method.