Statistical investigation of the effects of different origin aggre-gate properties on the mechanical properties of concrete

Abstract

Aggregates are one of the most important components of the concrete mix and the aggregate properties have a great influence on the properties of the hardened concrete. From this point of view, the contribution of the nature of the aggregate to the general mechanical behavior of concrete should be better understood and investigated. The main purpose of this study is to present practical and useful equations for the rapid evaluation of the basic properties of concrete, especially during the preliminary design phase. A series of laboratory studies were conducted to determine the effects of twenty aggregates with very different origins and properties on various properties of concretes. Laboratory tests with the same or similar principles as aggregate rocks have been repeated for concrete samples produced using these aggregates. Nine SRA equations were generated to directly predict similar concrete properties with the help of aggregate properties. Equations with very high coefficients of determination were produced between various aggregate properties and concrete properties except for the SHH property. In addition, nonlinear multiple regression analysis (NMRA) was used for twenty-four equations to predict some basic properties of concrete (strength, abrasion, tensile, and impact). The performances of the derived equations were evaluated with a statistical tool developed by the author. Accordingly, the models with the strongest prediction capacity were obtained for STSC, LAC, CSC, and, IEC respectively. Results from NMRA showed that equations with the highest coefficient of determination were obtained with Model-3 for CSC (Equation 15), and STSC (Equation 18), and Model-1 for LAC (Equation 22), and IEC (Equation 31). Concrete has a complex structure that is affected by many parameters. In accordance with this complex nature, the approach to predict the main concrete properties with nonlinear multiple methods by including similar aggregate properties as well as non-destructive methods representing concrete/aggregate (hardness, ultrasonic pulse velocity and physical properties) has been successful. The extremely high coefficients of determination, ranging from 0.81 to 0.96, obtained with NMRA, indicate that basic concrete properties can be strongly predicted by aggregate properties and some concrete properties that can be tested non-destructively.