Date of Award


Document Type


Degree Name

Doctor of Philosophy (PhD)

Legacy Department

Civil Engineering


Rangaraju, Prasad

Committee Member

Putman , Bradley

Committee Member

Poursaee , Amir

Committee Member

Bridges , William


High performance concrete mixtures often contain multiple cementitious components. Among these, cement is the most expensive in addition to having a higher carbon footprint. Life cycle assessment of cement production reveals that the cement content is the most important factor in determining a concrete mixture's embodied energy and carbon footprint. Compressive strength, an important property of concrete, is directly related to the quantity of cement used in the mixture. However, higher quantities of cement lead to durability issues. The increased concerns about the durability of concrete over the past decade have increased focus on improving the long-term performance of concrete structures. The goal of reducing the quantity of cement has led the use of supplementary cementitious materials (SCM) such as slag, fly ash, silica fume and others as a replacement.
The traditional method for optimizing high performance concrete mixtures involves systematically varying the individual proportions of the components in small increments and studying the resultant effect. In this method, the basis for selecting SCM dosage is arbitrary and often focuses on a specific set of requirements such as strength or durability. Optimizing the component proportions in the traditional way to achieve the desired properties is time-consuming, requiring a large number of trial batches, making this process expensive and inefficient. The use of statistical mixture design techniques has the potential to reduce the number of test runs needed, especially when multiple cementitious components are used and multiple requirements have to be simultaneously satisfied.
The research reported here investigates the use of a statistical design of experiments approach, specifically the simplex-centroid mixture design, using three cementitious components and a minimum of seven design points representing specific mixture proportions. In this study, a ternary blend of portland cement, slag and Class F fly ash was used. The total cementitious content of the concrete was kept constant although the individual proportions were varied. Fresh and hardened properties of concrete were evaluated, including mechanical properties such as compressive strength and split tensile strength and durability indicators such as rapid chloride ion permeability and expansion due to alkali-silica reaction. With the use of statistical design software (JMP), strength and durability prediction equations were developed and subsequently validated using an additional five concrete mixtures. These prediction equations investigated here generated a response surface for a given property as a function of the proportions of the three cementitious components using the seven concrete mixtures. Multiple response surfaces were superimposed on the simplex design region, and optimum cementitious mixtures were identified. The ternary blends were also used to evaluate mortars for alkali-silica reaction potential in mortar bars, and fundamental studies on cementitious paste systems involved pore solution extraction analysis and electrical resistivity.
The results obtained from this study showed that the properties of concrete such as compressive strength and rapid chloride ion permeability had a good correlation between the actual and predicted values whereas properties such as split tensile strength did not a show good correlation. The deleterious effects of alkali-silica reaction in mortar and concrete were evaluated using a threshold expansion value. These evaluations indicated that the mixtures below the threshold expansion contour in the simplex region did not show any alkali-silica reaction distress. The results from the cementitious paste studies showed that the electrical resistivity of the cementitious paste systems increased with decreasing ionic concentrations in the pore solution due to the replacements of cement with SCMs. In addition, the pore solution analysis showed that because of the pozzolanic reaction of SCMs, the alkali ions become trapped in the secondary C-S-H gel and the pore solution alkalinity is reduced with age. At elevated temperatures due to the instability of the calcium sulfo-aluminate phases, the sulfate ions (SO4-2) dissolved back into the pore solution. Using the simplex centroid design technique the pore solution results can be used to generate response surface for ionic concentrations of cementitious paste systems.
Results from this research suggest that the simplex-centroid design could be a valuable tool for minimizing the number of trial batches needed to identify the optimal concrete proportions for achieving the desired properties. As an outcome of this research, guidelines were developed for using the simplex-centroid method for concrete mixture design applications. The optimum mixtures obtained for various concrete applications within the simplex region yielded optimum cement dosages, in turn reducing the cost of concrete and its carbon footprint.
Future work in this area should include using different SCMs to optimize desired properties of concrete. In addition, this concept can be extended to include the w/c ratio, another important property of concrete. Various statistical mixture designs techniques can also be explored to improve the predictability power. Ultimately, the research in this area should lead to more cost-effective concrete with a smaller carbon footprint that can be adopted for use in the field.