Explicit soft computing model to predict the undrained bearing capacity of footing resting on aggregate pier reinforced cohesive ground

摘要

Aggregate pier has been frequently suggested as a traditional solution to enhance the undrained bearing capacity and reduce settlement of footings resting on cohesive ground. However, there is no widely accepted method that can be utilized to accurately estimate the ultimate undrained bearing capacity of this system due to the complexity of the interaction between the soil and the aggregate pier. In addition, all of the available empirical correlations for this problem have been proposed using traditional regression analysis with no attempts to use advanced soft computing techniques to propose an explicit mathematical formulation. In this research, this gap in knowledge has been addressed by employing the genetic algorithm multi-objective evolutionary regression analysis to propose a novel explicit mathematical model to predict the undrained bearing capacity of footing on aggregate pier reinforced ground. The new soft computing model scored mean absolute error, root mean square error, mean, percentage of predictions with accuracy of 80% and coefficient of determination of 71.08, 97.00, 1.00, 87, and 0.91, respectfully for the training subgroup data and 92.81, 105.33, 1.1, 86, and 0.93, respectfully for testing subgroup data. In addition, the new model was found to predict the ultimate undrained bearing capacity with lower error compared to the available empirical correlations. Finally, a parametric study has been conducted using the new model to provide an insight into the factors that influence the undrained bearing capacity of footing reinforced with aggregate pier and to give additional confidence in the model's robustness.