Formulation and Performance of Model Concrete in Reduced-Scale Physical Model Tests

Abstract

The utility of geotechnical centrifuge tests depends on how correctly they predict the physical and mechanical behaviour of concrete. In this study, a model concrete material that consisted of α-gypsum plaster, fine silica sand, and water was developed. An orthogonal test design was used to evaluate the effect of the mix proportion on the model concrete performance. The physical (i.e., flowability and bleeding rate) and mechanical (i.e., compressive and flexural strength) characteristics were considered as indices. Various mix ratios resulted in remarkable relative contributions to model concrete performance, and each raw material dosage exhibited positive or negative synergy. The water–plaster ratio (W/P) and aggregate–plaster ratio (A/P) strongly influenced the mechanical and physical characteristics, respectively. Multiple linear regression analysis (MLRA) was carried out to determine a forecast model for various small-scale test demands. Finally, the applicability and outlines of the presented forecasting method in proportioning design were evaluated by typical use of model concrete in small-scale model tests.