The Application of Artificial Intelligence Methods on the Optimizing Improvement Depth of Dynamic Compaction

Abstract

Abstract Dynamic Compaction (DC) with tamping is one of the soil improvement methods and has been significantly accepted due to its advantages over other soil improvement methods. In this paper, an attempt has been made to use the fuzzy logic and Sugeno inference system to investigate the effect of the parameters involved in soil improvement operations using the DC method on the relative improvement depth. A correlation has been proposed to estimate the effective depth for the granular soils and compared to the fuzzy results. The Particle Swarm Optimization (PSO) algorithm has been used to achieve maximum improvement depth. The results indicate that the interaction between tamper weight and height of tamping plays the most important role in the design procedures. The results reveal that by using PSO, the maximum depth of improvement is increased by 33%. Studies show that the optimal tamper radius for most of the compaction patterns with medium to high applied energies is equal to 1.5 to 2 meters, the optimal numbers of drops are equal to 25, and the optimal grid spacing is equal to 6 to 7 meters. It should be noted that using PSO dynamic compaction pattern; the maximum improvement depth has been achieved.