Prediction of properties of sand mixture with HDPE by MLR, EPR and Stepwise models

Authors

1 Department of Civil Engineering, Faculty of Engineering, University of Guilan, Guilan,

2 PHD

10.22034/ceej.2024.55984.2242

Abstract

Due to the increasing waste plastics and the lack of landfill places, researchers are investigating the different methods of reusing and recycling them. One of the proposed methods is reusing them in engineering projects, but the combination of these waste plastics with materials including soil significantly affects their properties. The present study investigates and compares the statistical methods of Multiple Linear Regression (MLR), Evolutionary Polynomial Regression (EPR) and Stepwise Regression in order to predict the geotechnical properties of sand-shredded waste HDPE (High Density Polyethylene) mixture including E and mv and K0. Input data have been obtained from a series of large scale oedometer tests on HDPE-sand mixture considering different content of HDPE (0, 2, 4, 6,8%), two relative densities (70% and 40%) and three different normal stresses (200 kPa, 300 and 100). The performance of the models provided by EPR, MLR and Stepwise was evaluated using the root mean square error (RMSE), coefficient of determination (R2) to obtain the best fitting model. The results showed that the EPR and MLR model can be used as a powerful tool for modeling the geotechnical properties of sand-HDPE and other soil mixtures and waste materials with similar conditions. By using these methods, in addition to reducing the cost of tests, the time to access the results of the large-scale oedometer test is also significantly accelerated. Comparing the model presented in the Stepwise method for predicting the behavioral model of mixtures, it didn’t provide appropriate and accurate results and its use is not suggested.

Keywords

Main Subjects

Journal of Civil and Environmental Engineering

Articles in Press, Accepted Manuscript
Available Online from 07 May 2024

Files

Share

How to cite

Statistics