EPS Daneciklerinin ve/veya Cam Tozunun Killi Zeminlerin Kıvam Limitlerine Etkisi ve Limitlerin YSA ve Regresyon ile Tahmin Edilmesi.

The consistency of fine soils is an essential tool in the classification and estimation of their parameters. Firstly, the changes in the consistency limits of the clayey soil improved with the waste material were investigated experimentally. Bentonite, which was classified as a high plasticity clay according to the unified soil classification system, was improved by using only waste glass powder, only waste EPS beads and both additives at different rates. Liquid limit and plastic limit tests were carried out. Secondly, a dataset (65 data) was gathered for the consistency limits of treated soils with similar waste materials in this study and literature. 80% of the data was selected to be used as training and 20% as a test dataset. Empirical correlations were obtained with the multiple linear regression method. The same dataset was used in the artificial neural network method (ANN) and the algorithm was trained. Both methods were run with the testing dataset and the results were compared. In both methods, the determination coefficients obtained from the training and testing data sets are satisfactorily high, and it is thought that the consistency limits of the treated clays will be estimated close to the actual values. In order to check that the results obtained by the ANN method are independent of the selected data sets, cross-validation tests were performed. As a result of this study, empirical correlations and ANN methods are proposed to be used in estimating the consistency limits of clayey soils improved with mentioned waste materials. [ABSTRACT FROM AUTHOR]