Abstract
Pervious concrete (PC) pavement is a sustainable type of concrete pavement that can protect and restore natural ecosystem. The permeability coefficient is the most important characteristic of PC. The purpose of this experimental study was to investigate the effect of mixture design parameters, particularly water-to-cement ratio (W/C) and size of aggregate on the permeability coefficient of PC. The thirty six mixtures were made with W/C in range of 0.28 to 0.34, 350 kg/m3 cement content and 9.5 to 19.5 mm maximum size of aggregate. In this study the feasibility of using the artificial neural networks (ANN) in predicting the effect of aggregate size and W/C on amount of permeability coefficient of PC was investigated. For modeling, 65% of data was used for model training and remaining 35% was used for model testing. Based on the lowest root mean squared error (RMSE), the best ANN model was chosen. The results showed that the W/C and aggregate size are key parameter, which significantly affect the performance of PC. The ANN modeling was developed in this study can facilitate prediction permeability coefficient of PC. This approach can reduce the number of trial batches for target performance of samples.
Key words: Pervious concrete, artificial neural network, permeability coefficient.