Development of Void Prediction Models for Kansas Concrete Mixes Used in PCC Pavement

Abstract

Permeability of the concrete material used in Portland Cement Concrete (PCC) pavement structures is a major factor for long-term durability assessment. To properly characterize the permeability response of a PCC pavement structure, the Kansas Department of Transportation (KDOT) generally runs the Boil Test (BT) to determine the % void response. The BT typically measures the volume of permeable pore space within the concrete samples over a period of five hours at a concrete age of 7, 28, and 56 days. In this study, backpropagation Artificial Neural Network- (ANN) and Regression-based % void response prediction models for the BT are developed by using the database provided by KDOT in order to reduce the duration of the testing period or ultimately eliminating the need to conduct the BT. The noted excellent prediction accuracy of the developed models proved that the ANN and the Regression models have efficiently characterized the BT response. Therefore, they can be considered as effective and applicable models to predict the permeability (% void response) response of concrete mixes used in PCC pavements.