COMPRESSIVE STRENGTH PREDICTION IN GGBS-BASED CONCRETE USING REGRESSION MODEL EQUATION

Abstract

Concrete exposed to a conventional environment exhibits some levels of predictability under serviceability, while highly uncertain in the deterioration behavior of concrete is when concrete is in marine environment. Based on a laboratory based data-driven approach, this study considered the possibility of developing a regression model equation to predict the compressive strength over long curing age and provide an excellent opportunity to understand the complexity of concrete strength development dynamics in real practice. The 7-days measured early compressive strength value and the absolute density respectively were modeled as the input parameters. The 28-days strength value were predicted and compared with a laboratory measured actual 28-days strength values. The resulting two-parameter interaction based on regression model analysis is given as Y_{28th day Strength} = -55.1664 + 0.2146*(7-day strength) + 0.0387*(absolute density). The validation findings using Multiple R = 0.92, R^{2 =} 0.85, Adjusted R² = 0.55 and the Standard Error = 0.83, shows that both the 7-day strength and absolute density have a positive effect in predicting 28^th day strength as the dependent variable. Specifically, a one-unit increase in 7-day strength improves the predicted value of the dependent variable by 0.2146 units, while a one-unit increase in absolute density increases it by 0.0387 units, while holding other factors constant. Thus, the application of regression model equation in the prediction of compressive strength is effective and the results can be employed in the design satisfactorily and the model can be employed in the assurance of quality control in construction sites

KEYWORDS Compressive Strength, Regression Model, Prediction and Concrete