Modelling the Effect of High Temperatures on Concrete Made with Discarded Bottle Powder (DBP) and Metakaolin (MTK)

Abstract

This study models the effect of high temperatures of 100°C, 200°C, 300°C, 400°C and 500°C on blended concretes made with discarded bottle powder (DBP) and metakaolin (MTK). The blended concrete of two mixes containing ten (10%) of discarded bottle powder only and blended concretes of three mixes containing ten (10%) of discarded bottle powder with 5%, 10%, 15% and 20% metakaolin were produced. Six different mixes were produced for the concrete cubes and beams. Seventy-two (72) concrete cubes of sizes 150mm x 150mm x 150mm were produced to determine the compressive strengths (before and after heating the hardened samples) and Seventy-two (72) concrete beams of sizes 150mm x150mm x 500mm were also produced to determine the flexural strengths (before and after heating the hardened samples). Slump test was conducted on each of the six mixes produced to determine its workability. The concretes were cured for 28 days by complete immersion in water. The concrete cubes and beams were heated at the above-mentioned temperatures before testing using Electrical Motorized compression Machine and Universal Testing Machine respectively. The results showed that the strengths generally decreased as the percentage of metakaolin (MTK) increased, with increased temperature. Concrete cubes made with 5% MTK achieved the 28days target strength of 25N/mm² at temperature of 100°C. The optimum replacement level is MTK 5% DBP 10% at100°C. The data obtained were analyzed using response surface regression and Analyses of Variance (ANOVA) in the Minitab 16 Statistical Software. Models were developed to predict the compressive and tensile strengths with percentage mix and temperature as independent predictor variables. The t-Test indicates that there is no statistical significant difference between the experimental and predicted strength value at 0.05 level of significant. The correlation coefficients, R² of 96.57% and 91.89% for compressive and flexural strengths model, indicate that the model prediction gave a good correlation factor which implies that the model predictions are accurate.