Macroscopic stress, couple stress and flux tensors derived through energetic equivalence from microscopic continuous and discrete heterogeneous finite representative volumes

Basic Information

• Paper ID: 2412.12616
• Title: Macroscopic stress, couple stress and flux tensors derived through energetic equivalence from microscopic continuous and discrete heterogeneous finite representative volumes
• Authors: Jan Eliáš (Brno University of Technology), Gianluca Cusatis (Northwestern University)
• Classification: cs.CE (Computational Engineering), cond-mat.other
• Submission Date: December 2024
• Paper Link: https://arxiv.org/abs/2412.12616

Abstract

This paper proposes a rigorous derivation method for evaluating macroscopic stress tensors, couple stress tensors, and flux vectors in continuous and discrete heterogeneous systems with independent displacement and rotation degrees of freedom. Unlike classical asymptotic expansion homogenization, this work considers finite-sized representative volume elements (RVEs). The macroscopic quantities are first derived for heterogeneous Cosserat continua, then the continuum equations are discretized to provide macroscopic quantities in discrete heterogeneous systems, and finally the expressions for discrete systems are re-derived directly from discrete properties.

Research Background and Motivation

1. Problem Statement

This paper primarily addresses how to accurately extract macroscopic mechanical and transport properties from microscopic heterogeneous material systems at finite scales, specifically:

• How to derive macroscopic stress tensors from finite-sized representative volume elements
• How to handle couple stress tensors in systems with independent rotational degrees of freedom
• How to establish energetic equivalence relationships between microscopic and macroscopic scales

2. Problem Significance

Many engineering materials (such as concrete, ceramics, granular materials, etc.) exhibit significant heterogeneity at the mesoscopic or microscopic scale. Accurate multiscale modeling is important for:

• Predicting overall mechanical behavior of materials
• Developing reliable numerical models
• Establishing phenomenological constitutive relationships

3. Limitations of Existing Methods

• Asymptotic expansion homogenization: Assumes representative volume elements approach zero, neglecting finite-size effects
• Classical Love-Weber formula: Primarily applicable to stress tensors, with inadequate treatment of couple stress tensors
• Boundary condition handling: Existing methods are insufficient in addressing boundary effects and reference point selection

4. Research Motivation

This work aims to establish a more complete and rigorous theoretical framework capable of:

• Handling finite-sized representative volume elements
• Simultaneously considering both internal force and external force formulations
• Explicitly accounting for the influence of reference point location in couple stress tensors

Core Contributions

1. Established a rigorous derivation framework based on virtual work equivalence, applicable to continuous and discrete heterogeneous systems
2. Derived macroscopic quantity expressions accounting for finite RVE size, including stress tensors, couple stress tensors, and flux vectors
3. Proposed two variant formulations: internal force-based and external force-based expressions
4. Identified additional terms in couple stress tensors, explaining their dependence on reference frames and macroscopic point locations
5. Provided comprehensive verification, including comparisons between analytical solutions and numerical models under steady-state and transient conditions

Methodology Details

Task Definition

The core task of this paper is to establish equivalence relationships between microscopic heterogeneous materials and macroscopic homogeneous continua, specifically including:

• Input: Microscopic stress field σ^mic, couple stress field μ^mic, flux field a^mic
• Output: Macroscopic stress tensor σ^mac, couple stress tensor μ^mac, flux vector a^mac
• Constraints: Satisfying virtual work equivalence principle and equilibrium equations

Theoretical Framework

1. Cosserat Continuum Fundamentals

The main variables in Cosserat continua include:

• Displacement vector u and independent rotation vector θ
• Strain tensor γ = ε - E·θ^m and curvature tensor κ = ∇⊗θ
• Stress tensor σ (non-symmetric) and couple stress tensor μ

The equilibrium equations are:

-ρü + ∇·σ + f = 0                    (linear momentum)
-ρJ_ρθ̈ + ∇·μ + E:σ + z = 0         (angular momentum)

2. Virtual Work Equivalence Principle

The key assumption is that virtual displacements and virtual rotations can be expressed as series in spatial coordinates:

δu_j(x) = α_j^(0) + x_iα_ij^(1) + x_kx_iα_ikj^(2) + ...
δθ_j(x) = β_j^(0) + x_iβ_ij^(1) + ...

Through the virtual work equivalence condition:

δW_int^mac = δW̄_int^mic = δW̄_ext^mic

3. Macroscopic Stress Tensor Derivation

For tensile and shear deformation modes (α^(1) terms), the result is:

σ_ij^mac = ⟨σ_ij^mic⟩ = (1/V)∫_Ω σ_ij^mic dV        (internal force form)
σ_ij^mac = (1/V)[∫_Γ x_it_j^mic dΓ + ∫_Ω x_if_j^mic dV]  (external force form)

4. Macroscopic Couple Stress Tensor Derivation

For bending and torsion deformation modes (β^(1) terms), the result is:

μ_ij^mac = ⟨μ_ij^mic + E_jklx_kσ_il^mic⟩ - E_jklx_k^macσ_il^mac

A key finding is that the couple stress tensor contains an additional term dependent on the location of the macroscopic reference point.

Treatment of Discrete Systems

1. Discretization of External Force Formulation

Direct discretization of continuum expressions:

σ^mac = (1/V)∑_f x_f ⊗ F_f                    (stress tensor)
μ^mac = σ^mac ⊗ x^mac : E + (1/V)∑_z x_z ⊗ Z_z  (couple stress tensor)
a^mac = -(1/V)∑_q x_qQ_q                       (flux vector)

2. Internal Force Formulation (Love-Weber Type)

Through force and moment balance, a generalization of the famous Love-Weber formula is obtained:

σ^mac ≈ (1/V)∑_e A_el_ee_N^e ⊗ t^e
μ^mac ≈ σ^mac ⊗ x^mac : E + (1/V)∑_e A_el_ee_N^e ⊗ [m^e + E:(x_c^e ⊗ t^e)]
a^mac ≈ (1/V)∑_d S_dh_de_λ^d j_d

3. Boundary-Radius Gap Term

The internal force formulation neglects the actual location of external forces, introducing a "boundary-radius gap" error term that can be neglected for systems with large numbers of particles.

Experimental Setup

Verification Strategy

The paper employs two classes of verification methods:

1. Poisson Problem Verification

• Geometry: 1×1 m² square domain
• Discretization: Approximately 48,000 nodes, 72,000 conduit elements
• Test Cases:
  • Constant flux patch test: p = 2(2x-1) + 2(2y-1), a = (-4,-4)^T
  • Nonlinear analytical solution: p = 16x₁x₂(1-x₁)(1-x₂)

2. Mechanical Problem Verification

• Geometry: Cantilever beam, depth D=3m, length S=12m
• Discretization: 1,034,196 degrees of freedom
• Material Parameters: E₀=40GPa, α=0.3, β=0~10⁵
• Comparison Models:
  • Discrete model
  • Homogenized Cosserat model
  • Simplified Cosserat model
  • Cauchy continuum model

Evaluation Metrics

• Accuracy Assessment: Deviation from analytical solutions
• Convergence: Results at different mesh sizes
• Consistency: Comparison between internal and external force formulations
• Boundary Effects: Impact of boundary-radius gap term

Experimental Results

Main Results

1. Poisson Problem Verification

• Variant I (with boundary-radius gap term): Consistently provides exact solutions
• Variant II (neglecting boundary-radius gap term): Accuracy decreases with decreasing mesh size
• Convergence: Results from both variants are similar for large mesh sizes

2. Mechanical Problem Verification

Steady-State Analysis:

• Discrete model aligns well with homogenized model
• As bending stiffness β increases, stress decreases while couple stress increases
• When β<10³, Cosserat effects are negligible and Cauchy continuum can be used

Transient Analysis:

• Displacement responses from all models are highly consistent
• Time evolution of stress and couple stress conforms to expectations
• Validates the formula's effectiveness under transient conditions

Key Findings

1. Boundary-Radius Gap Effect

Experiments confirm the importance of the boundary-radius gap term:

• For small control volumes, this term cannot be neglected
• In systems with large numbers of particles, the term's influence is smaller
• Provides quantitative error assessment

2. Influence of Reference Point Location

Validates theoretical predictions of reference point location effects in couple stress tensors:

• When the macroscopic point is at the domain centroid, the additional term vanishes
• Other locations produce additional couple stress contributions

3. Influence of Material Parameters

• Parameter β significantly affects Cosserat characteristic length ℓ
• When β=0, ℓ≈4.6mm; when β=10⁵, ℓ≈0.7m
• Lamé parameters λ,μ remain relatively stable

Related Work

Classical Theoretical Foundations

• Love-Weber Formula: Classical expression for stress tensors in discrete systems
• Asymptotic Expansion Homogenization: Forest et al.'s infinitesimal RVE theory
• Virtual Work Principle: Chang and Kuhn's energetic equivalence method

Innovations of This Work

1. Finite-Size RVE: Considers size effects compared to asymptotic theory
2. Complete Couple Stress Treatment: Includes contributions from microscopic stress over distance
3. Unified Framework: Simultaneously handles continuous and discrete systems
4. Dual Formulations: Complete correspondence between internal and external force expressions

Conclusions and Discussion

Main Conclusions

1. Theoretical Contribution: Establishes rigorous homogenization theory for finite RVEs, generalizing the classical Love-Weber formula
2. Methodological Innovation: Proposes a complete method for handling reference point dependence in couple stress tensors
3. Practical Value: Provides more accurate tools for multiscale modeling

Limitations

1. Virtual Deformation Mode Selection: Certain higher-order virtual work balances are neglected
2. Boundary Condition Restrictions: Requires all boundaries to have static boundary conditions
3. Computational Complexity: The boundary-radius gap term increases computational burden

Future Directions

1. Nonlinear Extension: Generalization to nonlinear constitutive relationships
2. Coupled Problems: Extension to mechanical-transport coupled problems
3. Higher-Order Theory: Consideration of higher-order virtual deformation modes

In-Depth Evaluation

Strengths

1. Theoretical Rigor: Rigorous mathematical derivation based on virtual work principle
2. Strong Practicality: Directly applicable to existing discrete and continuous models
3. Comprehensive Verification: Includes analytical solutions, numerical comparisons, and parameter studies
4. Clear Presentation: Logical structure and accurate mathematical expressions

Weaknesses

1. Limited Scope: Primarily focused on linear elastic problems; nonlinear extensions require further research
2. Computational Efficiency: Complete formulas have relatively high computational cost
3. Parameter Sensitivity: Some results are highly sensitive to material parameter β

Impact

1. Academic Value: Provides important theoretical foundation for multiscale mechanics
2. Engineering Application: Can be directly used for modeling heterogeneous materials such as concrete and geomaterials
3. Software Implementation: Open-source code provided for easy dissemination

Applicable Scenarios

• Multiscale modeling of granular materials
• Homogenization of heterogeneous composite materials
• Material systems with microscopic rotations
• Engineering problems requiring consideration of size effects

References

The paper cites 63 related references, primarily including:

• Classical literature on Cosserat continuum theory
• Discrete element methods and homogenization theory
• Related work on multiscale modeling
• Literature on numerical methods used for verification

Overall Assessment: This is an excellent paper with rigorous theory and high practical value, providing important theoretical tools and numerical methods for multiscale modeling of heterogeneous materials. The derivation process is clear, verification is comprehensive, and the work has significant academic and practical value for related fields.