Abstract:This paper introduces a novel approach that combines Proper Orthogonal Decomposition (POD) with Thermodynamics-based Artificial Neural Networks (TANN) to capture the macroscopic behavior of complex inelastic systems and derive macroelements in geomechanics. The methodology leverages POD to extract macroscopic Internal State Variables from microscopic state information, thereby enriching the macroscopic state description used to train an energy potential network within the TANN framework. The thermodynamic consistency provided by TANN, combined with the hierarchical nature of POD, allows to reproduce complex, non-linear inelastic material behaviors as well as macroscopic geomechanical systems responses. The approach is validated through applications of increasing complexity, demonstrating its capability to reproduce high-fidelity simulation data. The applications proposed include the homogenization of continuous inelastic representative unit cells and the derivation of a macroelement for a geotechnical system involving a monopile in a clay layer subjected to horizontal loading. Eventually, the projection operators directly obtained via POD, are exploit to easily reconstruct the microscopic fields. The results indicate that the POD-TANN approach not only offers accuracy in reproducing the studied constitutive responses, but also reduces computational costs, making it a practical tool for the multiscale modeling of heterogeneous inelastic geomechanical systems.

What problem does this paper attempt to address?

The problem this paper attempts to address is how to effectively capture the macroscopic behavior of complex inelastic systems in multi-scale modeling and derive macroelements in geomechanics. Specifically, the paper proposes a new method that combines Proper Orthogonal Decomposition (POD) and Thermodynamics-based Artificial Neural Networks (TANN) to extract macroscopic internal state variables (ISVs) from microscopic state information to train energy potential networks within the TANN framework. This method not only reproduces the complex nonlinear inelastic material behavior and macroscopic geomechanical system response but also reduces computational costs, making it an effective tool for practical applications. ### Main Issues 1. **Complexity in Multi-Scale Modeling**: - Multi-scale modeling of complex inelastic systems requires considering interactions at both microscopic and macroscopic levels, and traditional methods struggle to accurately capture these complex nonlinear behaviors. - The vast amount of information at the microscopic scale, when directly used for macroscopic scale modeling, leads to prohibitively high computational costs. 2. **Thermodynamic Consistency and Adherence to Physical Laws**: - Machine learning-based methods show potential in multi-scale modeling, but ensuring these models comply with fundamental physical laws (such as the second law of thermodynamics) is a challenge. 3. **Identification of Internal State Variables**: - In multi-scale modeling, effectively identifying and selecting internal state variables (ISVs) to accurately describe the system's state at the macroscopic scale is a key issue. ### Solution - **POD-TANN Method**: - **POD**: Used to extract macroscopic internal state variables (ISVs) from microscopic state information, retaining the maximum amount of information through dimensionality reduction techniques. - **TANN**: An energy potential network constructed based on thermodynamic principles, ensuring the model adheres to physical laws and accurately captures the complex behavior of inelastic materials. ### Application Examples - **Homogenization of Continuous Inelastic Representative Volume Elements**: Validated the effectiveness of the POD-TANN method in handling multi-scale problems of continuous inelastic materials. - **Horizontal Loading of a Single Pile in Clay Layers**: Derived a macroelement with a single degree of freedom using the POD-TANN method, demonstrating its application in complex 3D finite element models. ### Main Contributions - Proposed a new multi-scale modeling method that combines POD and TANN, capable of efficiently capturing the macroscopic behavior of complex inelastic systems. - Reduced computational costs through dimensionality reduction techniques, making the method feasible for practical engineering applications. - Validated the effectiveness and accuracy of the method in applications of varying complexity.

A POD-TANN approach for the multiscale modeling of materials and macroelement derivation in geomechanics

Bridging The Micro And Macro For Granular Media: A Computational Multi-Scale Paradigm

A Thermodynamics-Informed Neural Network for Elastoplastic Constitutive Modeling of Granular Materials

A Coupled SPFEM/DEM Approach for Multiscale Modeling of Large‐deformation Geomechanical Problems

Parallel Hierarchical Multiscale Modelling of Hydro-Mechanical Problems for Saturated Granular Soils

Multiscale modeling of inelastic materials with Thermodynamics-based Artificial Neural Networks (TANN)

Multiscale Simulation of the Responses of Discrete Nanostructures to Extreme Loading Conditions Based on the Material Point Method

A reduced order model for geometrically parameterized two-scale simulations of elasto-plastic microstructures under large deformations

Multiscale Thermodynamics-Informed Neural Networks (MuTINN) towards fast and frugal inelastic computation of woven composite structures

A macroelement stabilization for multiphase poromechanics

Physics-driven neural networks for nonlinear micromechanics

An efficient hierarchical Bayesian framework for multiscale material modeling

Modelling of particle flow code geotechnical material parameter relationships based on orthogonal design and back propagation neural network

A multi-field decomposed model order reduction approach for thermo-mechanically coupled gradient-extended damage simulations

A Multi-Field and Fluid–solid Coupling Method for Porous Media Based on DEM-PNM

A deep material network for multiscale topology learning and accelerated nonlinear modeling of heterogeneous materials

Learning the physics-consistent material behavior from experimentally measurable data via PDE-constrained optimization

A micromechanical scheme with nonlinear concentration functions by physics-guided neural network

A user material approach for the solution of multi-field problems in Abaqus: Theoretical foundations, gradient-enhanced damage mechanics and thermo-mechanical coupling

A Stochastic Multi-scale Approach for Numerical Modeling of Complex Materials - Application to Uniaxial Cyclic Response of Concrete

Neural networks meet anisotropic hyperelasticity: A framework based on generalized structure tensors and isotropic tensor functions