Experimental evaluation and modelling of the compressive strength of concretes produced with different strength classes of cements

The main aim of the thesis is to propose a prediction model for estimation of compressive strength of concretes with various cements and mixture proportions. The strength of the samples produced with three different types of cement at different rates of water-to-cement ratios and cement richness were investigated experimentally and evaluated statistically. Three type of cement possessing 28-day strengths of 32.5, 42.5, and 52.5 MPa was used in the production of concretes. The concretes were produced at cement richness values of 300, 400, and 500 kg/m3 and w/c rates at changing levels within the interval of between 0.3 and 0.6. By this way, combined influences of cement strength, amount of cement and w/c ratio was experimentally investigated. Totally 36 mixes were cast then the compressive strength values were examined after specified moist curing periods (7 and 28 day). A statistical study were conducted on the experimental results and the significances of the cement strength, w/c values and amount of cement on the compressive strength of the concretes were assessed. Another crucial focus of the current paper is to generate an explicit expression to predict the compressive strength of the concretes tackled with the current study. To derive an explicit formula for estimation, a soft computing method called gene expression programming (GEP) was benefited. The GEP model was also compared with a less complicated estimation model developed by multi linear regression method. The results revealed that compressive strength of the samples were significantly influenced by cement type and aggregate-to-cement ratio. The proposed GEP model indicated a high correlation between experimental and predicted values.