Prediction of Unconfined Compressive Strength of a Stabilised Expansive Clay Soil using ANN and Regression Analysis (SPSS)

Industrial waste stabilization is one of the common techniques used to improve the strength of weak soils as it is very difficult to find suitable soil everywhere for foundation. The strength developed by this stabilization technique depends on percentage addition of additives, soil properties and curing period. In order to reduce the time consuming process of laboratory investigation and for effective utilization of industrial waste additives, a mathematical model is developed relating the gain in strength (UCS) with the variables responsible. In this study, Artificial Neural Network (ANN) and Regression analysis (SPSS) is used for prediction of unconfined compressive strength (UCS) of clay stabilized with industrial waste namely bottom ash and Ecosand. The predictor variables used for developing the models are Curing Period, Bottom Ash and Ecosand dosages, Liquid Limit (LL), Plasticity Index (IP), and Free Swell. Based on laboratory investigations, the data's are divided into three groups based on the additives added. Regression analysis and artificial neural networks prediction showed that acceptable correlations exist between soil properties and unconfined compression strength of the stabilized clay soil. Based on analysis, artificial neural network (R2=0.978) showed closer prediction results when compared to the regression analysis (R2=0.939).