Concrete Strength Prediction using Multi-Linear Regression Model: A case study of Nairobi Metropolitan
Kiambigi1, Maina, Gwaya, A.O2, Koteng, D.O3

1Kiambigi, MAINA, Department of Construction Management, JKUAT.
2GWAYA, A.O, Department of Construction Management, JKUAT.
3Koteng, D.O, Department of Civil and Construction Engineering, TUK.

Manuscript received on January 02, 2019. | Revised Manuscript received on January 05, 2019. | Manuscript published on January 30, 2019. | PP: 11-20 | Volume-8 Issue-5, January 2019. | Retrieval Number: E3177018519
Open Access | Ethics and Policies | Cite | Mendeley
© The Authors. Published By: Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC BY-NC-ND license (

Abstract: Early prediction of strength is key in effective and efficient planning for concrete construction projects. There are several empirical correlations that have been developed to determine concrete strength estimation from early age results though each model has its own limitations when applied. A multistaged evaluation of the existing prediction models (BS modification factors, German model, Abrams model, Bolomey’s model and ACI model) was performed for concrete strength data obtained from experimental work conducted under standard conditions in the laboratory. The data on compressive strength was obtained from concrete made from 6 different samples of fine aggregates whose physical and chemical properties had been determined. The limitations for each model was noted which then gave a basis for need for a statistical method that could predict strength more accurately. A multiple linear regression technique was used. The variables used to predict were watercementations ratio, quantities of mix design constituents, physical and chemical properties of the fine aggregates. Multiplelinear regression models developed for this study yielded coefficients of determination (CODs) for concrete strength prediction at 7, 14, 28, 56, 112 and 180-days curing. The regression models were then validated using a different set of samples that were not included in the formulated models. The predicted values of compressive strength obtained using the regression models were found to be in agreement with the experimental results obtaining CODs of 0.7821, 0.7186, 0.8416, 0.755, 0.7695 and 0.8444 for 7, 14, 28, 56, 112 and 180 days respectively.
Keywords: Concrete Mix Design, Concrete Strength, Fine Aggregates.