Predicting CBR Value of Stabilized Pond Ash with Lime and Lime Sludge Using ANN and MR Models

Abstract

In the present study, a multilayer perception-artificial neural network and multiple regression model is developed for predicting the California bearing ratio (CBR) value of stabilized pond ash. Pond ash collected from Panipat thermal plant is stabilized with lime (2, 4, 6 and 8%) alone and in combination with lime sludge (5, 10 and 15%). Total 51 datasets of experimentally observed CBR value were used in the development of models. Fitness of the model was observed through three statistical parameters i.e. coefficient of correlation (CC), root mean square error (RMSE) and mean absolute error. Both the models predict CBR value with high degree of accuracy having CC more than 0.96. From the sensitivity analysis, it is observed that curing period is the most significant parameter affecting the CBR value of stabilized pond ash.