Application of Artificial Intelligence Techniques for Rolling Dynamic Compaction

Rolling dynamic compaction (RDC), involving non-circular modules towed behind a tractor, is now widespread and accepted among many other soil compaction methods. However, to date, there is no accurate method to reliably predict the increase in soil strength after the application of a given number of passes of RDC. This paper presents the application of artificial intelligence (AI) techniques in the form of artificial neural networks (ANNs) and genetic programming (GP) for a priori prediction of the density improvement by means of RDC in a range of ground conditions. These AI-based models are developed by using in situ soil test data, specifically cone penetration test (CPT) and dynamic cone penetration (DCP) test data obtained from several ground improvement projects that employed the 4- sided, 8-tonne ‘impact roller’. The predictions of ANN- and GP-based models are compared with the corresponding actual values and they show strong correlations (r > 0.8). Additionally, the robustness of the optimal models is investigated in a parametric study and it is observed that the model predictions are in a good agreement with the expected behaviour of RDC.