Genetic algorithm-based yield stress equations for concrete at high temperature and prolonged mixing time

Abstract

Experiments were designed to investigate the flow behavior of portland cement paste and concrete incorporating superplasticizers. The paste and concrete mixtures were subjected to prolonged mixing for up to 110 min at high temperature. The yield stress values of concrete and that of the corresponding cement paste were measured using a rotating rheometer and viscometer, respectively. The results reveal a weak linear correlation between the yield stress of concrete mixtures and that of the corresponding cement pastes. Results also indicate that the yield stress of concrete varies in a linear fashion with the elapsed time, while its variations with the temperature and superplasticizer dosage follow power and inverse power functions, respectively. In this study, the genetic algorithms (GA) technique was used to predict the yield stress of concrete considering various parameters, such as the mixing time, ambient temperature, and superplasticizer dosage. A sensitivity study was conducted to evaluate the ability of the GA equations thus developed to capture the effects of test parameters on the yield stress of concrete. It was found that the GA equations were sensitive to the effects of test parameters and provided yield stress predictions that compared well with corresponding experimental data.