REGRESSION MODELS FOR PREDICTION OF COMPRESSIVE AND TENSILE CONCRETE STRENGTH

Obam, Sylvester Ogah, Jagba,Aondona Shadrack, and Adeke, Paul T.

ABSTRACT

Prediction of concrete strength helps to fast-track completion time of construction job and reduces waste of materials. This study was undertaken to determine linear regression models for the prediction of compressive and tensile strengths of concrete. Laboratory experimental methods were used to carry out the tests. Five concrete grades 20, 25, 30, 35, and 40 were adopted for the tests. Five model equations were developed using compressive and tensile strength results at 28-day with the aid of MATLAB Software. The specific gravity of river sand, granite, and river gravel were found to be 2.60, 2.61, and 2.72 respectively. The grading of the aggregates shows that, coefficient of uniformity (Cu) of river gravel and granite are 1.9 and 1.64 respectively. The slump for granite concrete varies from 50 to 59 mm and that of river gravel varies from 51 to 59 mm. Compressive strength of granite concrete are between 20.1 and 40.7 N/mm². The compressive strength of the river gravel concrete is between 19.3 and 38.8 N/mm². Compressive strength of the granite-admixture concrete is between 20.26 and 41.12 N/mm². The maximum compressive strengths of 40.2, 38.8, and 41.1N/mm² at 28-day of curing, was observed for granite-concrete, river-gravel, and granite-admixture grade 40 respectively. The tensile strength at 28-day for granite concrete is between 1.55 and 3.74 N/mm² while river-gravel concrete strength values are between 1.20 and 3.41 N/mm². The results revealed that, the tensile strength of concrete cylinders produced with granite are higher than the ones made with river gravel. The developed models can predict compressive and tensile strength of concrete with high degree of accuracy. The models were validated with concrete produced using randomly selected mix ratios. Statistical t-test shows that there is no significant difference between the observed and predicted strengths. It is recommended that the models could be used for prediction of concrete compressive and tensile strengths.

Keywords: Linear Regression, Models, Prediction, Concrete Strength