Application of Physics-Informed Neural Networks for forward and inverse analysis of pile-soil interaction

The application of the Physics-Informed Neural Networks (PINNs) to forward and inverse analysis pile-soil interaction problems is presented. main challenge encountered in Artificial Network (ANN) modelling presence abrupt changes material properties, which results large discontinuities gradient displacement solution. Therefore, a domain-decomposition multi-network model proposed deal with strain fields at common boundaries regions soil layers. parametric study single piles embedded both homogeneous layered formations demonstrated under axisymmetric plane conditions. performance parameter identification (inverse analysis) particularly investigated. It shown that by using PINNs, localized data acquired along pile length –possibly obtained via fiber optic sensing–can be successfully used for inversion parameters formations.