%0 Journal Article %T Utilization of Soil Stabilization with Cement and Copper Slag as Subgrade Materials in Road Embankment Construction %J International Journal of Transportation Engineering %I Tarrahan Parseh Transportation Research Institute %Z 2322-259X %A Shahiri, Jaber %A Ghasemi, Mojtaba %D 2017 %\ 07/01/2017 %V 5 %N 1 %P 45-58 %! Utilization of Soil Stabilization with Cement and Copper Slag as Subgrade Materials in Road Embankment Construction %K copper slag %K Cement stabilization %K Unconfined compressive strength %K Elastic modulus %K Artificial Neural Network %R 10.22119/ijte.2017.45834 %X In this study, unconfined compression tests have been conducted to investigate the impacts of copper slag on mechanical characteristics for stabilized cement and un-stabilized soil. Dozens of specimens were prepared at four percentages of cement (i.e. 0%, 2%, 4% and 6%) and five percentages of copper slag (i.e. 0%, 5%, 10%, 15% and 20%) by weight of dry soil. The samples compacted into a cylindrical specimen and processed for the curing periods of 28, 60 and 90 days. The test results indicated that the inclusion of copper slag had a significant effect on the unconfined compressive strength (UCS). For cement stabilized specimens, the improvement impacts of the copper slag on the UCS was more tangible than un-stabilized ones. Furthermore, an increase in the UCS was most apparent in the 2% cemented specimen wherein the UCS increased more than 78% as the copper slag increased up to 20%. Moreover, it was evident that the more amount of copper slag increased, the more optimum moisture content (OMD) declined and additionally maximum dry density (MDD) of soil was on the rise, while the results of the increase in cement was quite the reverse. Moreover, an artificial neural network (ANN) model has been developed using eight input parameters including: copper slag content, cement content, water content, dry density, liquid limit, plastic limit, PH and curing age. An ANN network, composed of 10 neurons in a hidden layer, was considered as the appropriate architecture for predicting the elastic modulus of mixtures, and an excellent conformity was acquired between the observed test data and the predicted ones. The results was proven that the proposed model can be efficiently applied to predict the elastic modulus of stabilized soils. %U http://www.ijte.ir/article_45834_c1f9c4c33bd701b940aa4425e068bc5d.pdf