Application of Artificial Neural Network to Predict CBR of Fine-Grained Soil Mixed with Fly Ash

Laboratory testing has been the base for investigating the applicability of soil mixed with any material or additive. Several studies have been reported on fly ash stabilized fine-grained soil. In order to reduce the tedious laboratory method of finding the CBR value of soil-fly ash mixes prior to their field application, attempt has been made to use artificial neural network (ANN) technique for obtaining the CBR by using different influencing parameters of soil and fly ash. The properties of the fly ash and the fine-grained soils have been studied individually and when mixed together in different proportions from the past literatures. Factors affecting the CBR of fine-grained soil mixed with fly ash have been identified, and data have been collected. Using those data, CBR of fine-grained soils mixed with the fly ash has been calculated using ANN technique considering the percentage of soil and fly ash, MDD and OMC as inputs under both soaked and unsoaked condition. The CBR value predicted is found to be nearly close to that of laboratory data, and thus ANN can be used satisfactorily for predicting the CBR of fine-grained soil mixed with fly ash before its field application.