مشخصات پژوهش

This paper presents a novel computational procedure for the maximum dry density of mixed soils containing oversize particles. At first, the large-scale compaction test data for mixed soils are analyzed by an artificial neural network to determine the main factors affecting the compaction. These factors are then imposed on a genetic programming method and a new mathematical equation emerges. The new equation has more conformity with the experimental data in comparison with the previous correction methods. Besides, the mixed soil dry density is associated with most base soil and oversize fraction specifications. With regard to the sensitivity analyses, if the mixed soil contains high percentages of oversize fraction, the mixed soil composition is governed by the specification of oversized grains, such as specific gravity and the maximum grain size and by increasing these factors, the mixed soil dry density is increased. In mixed soil with a low content of oversize, the base soil specification mainly controls the compaction behavior of mixed soil. Furthermore, if the base soil is inherently compacted with greater dry density, adding the oversize slightly improves the mixed soil dry density. In contrast, adding oversized grains to the base soil with a lower dry density produces a mixed soil with greater dry density. By increasing the maximum grain size difference between the oversize fraction and base soil, the dry density of mixed soil is enhanced.