Discrete element method to investigate flexural strength of pervious concrete

Abstract

Pervious concrete (PC) is a sustainable pavement material that has been widely used for low volume roads, walkways, and bicycle paths. Past literature studies have proposed several mix design methods to prepare PC with desired properties. However, the current mix design procedure involves substantial consumption of materials and is labor-intensive iterative process. Several numerical simulation studies used discrete element method (DEM) in simulating PC as a granular material (aggregates) coated with cement paste. Even though several researchers investigated mechanical properties such as compressive, flexural, and tensile strengths in the laboratory, and simulated the behavior of PC under compression and tensile loads, no known study has attempted simulating flexural response of PC. Given the significance of flexural strength in the pavement design process, it is essential to simulate the flexural behavior of PC. Therefore, the objective of this study was to develop a DEM framework to simulate PC mixtures and predict their flexural behavior by simulating the flexural strength test in a DEM based software YADE. A laboratory experimental program was conducted to obtain the experimental data required for calibration and validation of the proposed PC simulation model. In addition, a DEM program was developed to simulate the PC mixtures with the desired gradation and porosity levels, as achieved in the laboratory. A flexural strength test similar to ASTM C78 (1 MPa/s loading rate) was devised to simulate the flexural behavior of the prepared numerical PC specimens. The cohesive particle model in YADE was calibrated using experimental data and then used as the constituent model to simulate the interaction between cement coated aggregates in these flexural strength simulations. Overall, the proposed DEM approach could simulate the PC specimens with specified porosities and was able to predict the flexural response of the PC mixtures at 97% accuracy. It is envisioned that the proposed model could be further used to design PC mixtures with desired porosity and flexural strength properties instead of preparing trial mixtures, which would substantially reduce wastage of materials and human effort.