This dataset includes 1918 finite element models of pile-net composite foundation (PNCF) with varying embankment loads and geological conditions. The above model is based on the Abaqus platform. The dataset can serve as a benchmark for theoretical validation and the development of deep learning prediction models in this field.
In this dataset, a full row of piles and half of the spacing between adjacent piles on both sides along the embankment extension direction are selected as the finite element modeling analysis range. All samples in this dataset are pile-net composite foundations, each with pile caps and geogrids. Since the stiffness of the piles is much greater than that of the soil, the piles are defined as elastic bodies, while the other components are defined as elastoplastic bodies using the MC constitutive model. Normal contacts between the rigid piles, cushion, soil, and embankment are defined as hard contact, and tangential contacts are defined as penalty contact (Coulomb friction). Parameters for piles, cushion, embankments, soil, geogrids, loads, and interfaces were selected based on literature, engineering surveys, and computational efficiency, as shown in Table 1. This model dataset includes not only settlement field data but also stress and strain field data. However, this study only extracted the surface displacement data from all PNCF models.
Links to the dataset of PNCF finite element models：https://pan.baidu.com/s/1D21muf-alAEL4uMbO64vjg?pwd=nggd and https://pan.baidu.com/s/1Yk4Z0ADgcWfNA83ipFpdVw?pwd=dxqv