Practical formulation for estimating the compressive strength of self-compacting fly ash concrete using gene-expression programming

Abstract

This study aims to develop a gene expression programming (GEP)-based model for estimating the strength of self-compacting concrete (SCC) using fly ash (FA). The model considers the effects of six input variables, including the binder content, the FA proportion, the water/binder ratio, the fine aggregate content, the coarse aggregate content, and the superplasticizer dosage. The 28-day compressive strength of 114 concrete samples was used to generate the prediction model. The trial runs indicate that the GEP model with four genes and 120 chromosomes demonstrates strong performance, achieving a high coefficient of correlation and low errors (e.g., RMSE and MAE). The selected model is reliable, transparent, and easy to use in practice in designing the mix proportion for the SCC. The analysis of variable contributions demonstrates that the water/binder ratio and proportion of FA have the most significant influence on the strength of the SCC, while the fine aggregate content shows a comparatively minor effect. Thus, the strength of SCC could be increased significantly by reducing the water/binder ratio with a low proportion of FA content. The novel model from this study could help engineers in estimating the strength of SCC with reasonable FA content and choosing the appropriate mix proportion to achieve the design strength.