Prediction of unconfined compressive strength of clay subgrade soil stabilized with Portland cement and lime using Group Method of Data Handling (GMDH)

Unconfined Compressive Strength (UCS) test is commonly employed to determine the strength and quality control of the stabilized layers. This method is time consuming due to the needed time for curing of samples. Also, this method can be costly if the number of samples is increased. In this paper, the group method of data handling (GMDH) was used to develop simple models with sufficient accuracy to predict the UCS of clayey subgrade stabilized with Portland cement and lime. To this end, after stabilization of samples with different percentages of Portland cement and lime at three different moisture contents (dry, wet and optimum moisture content) and curing times of 7, 14, 21, 28 and 60 days, the UCS tests were conducted to establish a comprehensive database including 150 records. In the next step, two different UCS prediction models for clayey subgrade soil stabilized with Portland cement and lime were developed by using the GMDH method. The R2 value for training and testing sets for samples stabilized with Portland cement was 0.9294 and 0.94522, respectively, while the R2 value for training and testing sets for samples stabilized with lime was 0.8897 and 0.880, respectively. In addition, the sensitivity analysis of each model showed that the cement percentage and moisture content have the most impact on the predicted UCS, respectively.