Prediction of Efficiency Factor of Ground-Granulated Blast-Furnace Slag of Concrete Using Artificial Neural Network

The relative performance of various supplementary cementitious materials (SCMs) can be compared with that of portland cement using the practical concept of the efficiency factor (or χ value). This study describes the use of artificial neural networks (ANNs) for the prediction of the efficiency factor of ground-granulated blast-furnace slag (GGBFS) in concrete based on published test results. Feed-forward back-propagation neural networks have been used. The ANN model was established by the incorporation of a large experimental database and by appropriately choosing the architecture and training process. The introduced ANN model provided a more accurate tool to calculate χ and capture the effects of five main parameters: concrete composition (the water-binder ratio [w/b]; cement dosage; and the GGBFS replacement level); age; and curing temperature, confirming the reported experimental data. This study shows that the use of this model for numerical investigations on the parameters affecting the efficiency of SCMs in concrete is successful. A mathematical model was also developed based on the ANN model's results for predicting the χ value of GGBFS in terms of percentage replacement (from 0 to 80%) and concrete age (from 2 to 90 days), as these are considered the most important factors affecting concrete strength. This evaluation makes it possible to design GGBFS concretes for a desired strength at any given age and replacement level. [PUBLICATION ABSTRACT]