Abstract:Cavity expansion is a canonical problem in geotechnics, which can be described by partial differential equations (PDEs) and ordinary differential equations (ODEs). This study explores the potential of using a new solver, a physics-informed neural network (PINN), to calculate the stress field in an expanded cavity in the elastic and elasto-plastic regimes. Whilst PINNs have emerged as an effective universal function approximator for deriving the solutions of a wide range of governing PDEs/ODEs, their ability to solve elasto-plastic problems remains uncertain. A novel parsimonious loss function is first proposed to balance the simplicity and accuracy of PINN. The proposed method is applied to diverse material behaviours in the cavity expansion problem including isotropic, anisotropic elastic media, and elastic-perfectly plastic media with Tresca and Mohr-Coulomb yield criteria. The results indicate that the use of a parsimonious prior information-based loss function is highly beneficial to deriving the approximate solutions of complex PDEs with high accuracy. The present method allows for accurate derivation of solutions for both elastic and plastic mechanical responses of an expanded cavity. It also provides insights into how PINNs can be further advanced to solve more complex problems in geotechnical practice.

What problem does this paper attempt to address?

The problem that this paper attempts to solve is: how to use Physics - Informed Neural Networks (PINNs) to solve the stress field in elastic and elastoplastic cavity expansion problems. Specifically, the author proposes a new Parsimonious Universal Function Approximator, aiming to balance the simplicity and accuracy of the model, so as to solve complex partial differential equations (PDEs) and ordinary differential equations (ODEs) efficiently and accurately. ### Problem Background and Motivation The cavity expansion problem is a classic problem in geotechnical engineering and can be described by partial differential equations (PDEs) and ordinary differential equations (ODEs). Traditional analytical, semi - analytical and numerical methods have limitations in dealing with complex problems, such as high computational cost and difficult meshing. In recent years, PINNs, as an emerging solution method, have performed well in solving various PDEs/ODEs, but its application in elastoplastic problems is still uncertain. ### Main Contributions of the Paper 1. **Propose a parsimonious loss function**: A new parsimonious loss function structure is introduced to balance the simplicity and accuracy of the model, and the effectiveness of this principle is verified by analyzing the back - propagation gradient. 2. **Apply PiNet to solve elastoplastic problems for the first time**: A PiNet framework capable of solving forward elastoplastic geological problems is proposed, taking into account the Tresca and Mohr - Coulomb yield criteria, demonstrating the generalization ability of the model. 3. **Verify the effectiveness of the method**: By comparing with the analytical solution, the accuracy and effectiveness of the proposed method in solving elastic and elastoplastic cavity expansion problems are verified. ### Main Research Contents - **Elastic analysis**: Based on the assumed conditions, the governing equations, compatibility equations and constitutive relations applicable to elastic materials are derived. - **Elastoplastic analysis**: The material is divided into elastic and plastic regions, and the respective governing equations and boundary conditions are derived, and the yield conditions are considered. - **Design of parsimonious loss function**: By simplifying the governing equations, the output dimension is reduced, the prediction accuracy is improved, and the computational complexity is reduced. - **Numerical experiments**: Numerical experiments are carried out for different types of material behaviors (isotropic, anisotropic, ideal elastoplastic) to verify the effectiveness of the method. ### Conclusion This research proposes a new parsimonious PiNet framework and successfully solves the problem of stress field solution in elastic and elastoplastic cavity expansion problems. By introducing the parsimonious loss function, not only the solution accuracy is improved, but also the computational complexity is significantly reduced, providing a new solution for more complex geotechnical engineering problems in the future.

Parsimonious Universal Function Approximator for Elastic and Elasto-Plastic Cavity Expansion Problems

M‐PINN: A mesh‐based physics‐informed neural network for linear elastic problems in solid mechanics

A physics-informed neural network technique based on a modified loss function for computational 2D and 3D solid mechanics

Finite-PINN: A Physics-Informed Neural Network Architecture for Solving Solid Mechanics Problems with General Geometries

Finite element-integrated neural network framework for elastic and elastoplastic solids

Modeling finite-strain plasticity using physics informed neural network and assessment of the network performance

Exact Dirichlet boundary Physics-informed Neural Network EPINN for solid mechanics

Physics-informed Neural Networks (PINN) for computational solid mechanics: Numerical frameworks and applications

Physics informed deep learning for computational elastodynamics without labeled data

Physics-Informed neural network solver for numerical analysis in geoengineering

Development of a cascaded multitask physics-informed neural network (CM-PINN) to construct the muti-physical field model of rubber bushing press fitting

A stepwise physics‐informed neural network for solving large deformation problems of hypoelastic materials

Physics-informed neural network simulation of multiphase poroelasticity using stress-split sequential training

Physics-informed Neural Networks for Heterogeneous Poroelastic Media

A data-driven physics-constrained deep learning computational framework for solving von Mises plasticity

A novel solution for seepage problems using physics-informed neural networks

A PINN-based modelling approach for hydromechanical behaviour of unsaturated expansive soils

Data-driven forward and inverse analysis of two-dimensional soil consolidation using physics-informed neural network

A mixed formulation for physics-informed neural networks as a potential solver for engineering problems in heterogeneous domains: Comparison with finite element method

Transfer learning based physics-informed neural networks for solving inverse problems in engineering structures under different loading scenarios

Enhanced physics‐informed neural networks for hyperelasticity