Multivariate Regression Model to Predict Geotechnical Properties of Fly Ash-Stabilized Clayey Soil

Abstract

Fly ash is a waste product from coal-fired power plants; it can be used for stabilization of soils, road sub-base constructions, embankment material, and other geotechnical fields. Substantial amount of time, resource, and energy is required to obtain experimental data like CBR values and other parameters in road sub-base constructions. Application of multivariate regression models can be used in predicting the CBR values, and other properties of the stabilized soil have a potential of reducing the total time and cost of construction. In this study, fly ash is used as stabilizing agent for clayey soils in road sub-base construction; then, multivariate models are developed to predict the CBR values and other properties of soil. Different percentages of fly ash were added to clayey soil, and then, CBR test and direct shear test were performed. With the obtained values, multivariate linear and non-linear regression models were developed to check the usefulness in prediction of the properties. In this paper, addition of 20% fly ash by dry weight of soil is recommended for clayey soil stabilization. Based on the regression analysis, the independent variables selected for prediction are very good at predicting unsoaked CBR values with an adjusted R2 value of 97%.KeywordsFly ashMultivariate regressionCBR valuePredictionModelsClayey soils