Experimental optimization of high-strength self-compacting concrete based on D-optimal design

Because of the high variability in the material content of high-strength self-compacting concrete (HSSCC), a huge number of trial sets (experiments) are often essential to find the optimum mixture proportions needed to achieve desired HSSCC properties. In the current investigation, an effort was made to considerably reduce the number of such experiments and to study the influence of different mixture parameters on the properties of HSSCC using a response surface method coupled with D-optimal design of experiment (DOE). In this study, water-binder ratio, cement content, fine-aggregate percentage, fly ash content and superplasticizer content were chosen as input variables, with compressive strength, passing ability, segregation, and manufacturing cost as responses. For each response, a full quadratic model and second-order polynomial equation was developed using central composite design (CCD) and D-optimal design. Analysis of variance (ANOVA) was carried out to check adequacy. Responses were analyzed in blocks where two mixture parameter vary and others stay constant. To determine the efficiency of the approach, the D-optimality of the CCD and D-optimal design were studied and compared. It was found that the D-optimal design reduces the number of trials 40.38-50% compared with standard design with less significant change on responses (0.32-0.54% change for compressive strength, 0.93-0.415% change for passing ability, 0.99-1.35% change for segregation and 3.9-4.35% change for cost), and provides dependable predictions as demonstrated in this investigation.