Development of a multiple linear regression model for predicting the 28-day compressive strength of Portland pozzolan cement

Abstract

A multiple linear regression model was developed for the prediction of the 28-day compressive strength of Portland Pozzolan Cement (PPC) containing Santorin Earth as an admixture. The independent variables of the model were: (1) the compressive strength at the age of 2 days, (2) the compressive strength after autoclave hydrothermal treatment at 2.07 N/mm2 (300 psi) and 214°C for 3 h, (3) the C3S/C2S ratio of the clinker, (4) the insoluble residue of the cement, (5) the pozzolanic activity factor and (6) the C3A content of the clinker. The evaluation of the proposed model was performed by various statistical tests, all of which were successful. These tests included: multiple correlation, test of the significance of coefficients (t-test), estimation of confidence intervals for coefficients, test for outliers and unusual residuals, test for influential points, conditional sums of squares, R-squared and analysis of variance. There was very good agreement between the strength predicted by the multiple regression model and experimental results.