Elasticity $\mathscr{M}$-tensors and the Strong Ellipticity Condition

Weiyang Ding, Jinjie Liu, Liqun Qi, Hong YanarXiv 2017 · DOI · arXiv

Abstract

In this paper, we establish two sufficient conditions for the strong ellipticity of any fourth-order elasticity tensor and investigate a class of tensors satisfying the strong ellipticity condition, the elasticity $\mathscr{M}$-tensor. The first sufficient condition is that the strong ellipticity holds if the unfolding matrix of this fourth-order elasticity tensor can be modified into a positive definite one by preserving the summations of some corresponding entries. Second, an alternating projection algorithm is proposed to verify whether an elasticity tensor satisfies the first condition or not. Besides, the elasticity $\mathscr{M}$-tensor is defined with respect to the M-eigenvalues of elasticity tensors. We prove that any nonsingular elasticity $\mathscr{M}$-tensor satisfies the strong ellipticity condition by employing a Perron-Frobenius-type theorem for M-spectral radii of nonnegative elasticity tensors. Other equivalent definitions of nonsingular elasticity $\mathscr{M}$-tensors are also established.

Introduction

In the study of how materials deform (linear elasticity), fourth-order tensors are used to describe the relationship between stress and strain. For a physical material model to be valid, it must satisfy the strong ellipticity (SE) condition. Mathematically, this condition ensures that the governing equations are well-behaved and that solutions to boundary-value problems actually exist.

The challenge is that verifying the SE condition is notoriously difficult. It is equivalent to checking if a tensor is "rank-one positive definite," which is a non-convex problem. While necessary and sufficient conditions exist, they are often computationally expensive or difficult to apply to complex, anisotropic materials. This paper seeks to find "checkable" sufficient conditions—rules that are easier to verify and, if met, guarantee the material is stable.

Methods and Theoretical Framework

Matrix Unfolding and Sum-Preservation

A core technique used is matrix unfolding, where a complex fourth-order tensor is converted into a large 2D matrix ($n^2 \times n^2$). While a positive definite unfolding matrix implies strong ellipticity, the reverse isn't always true. The authors introduce a "sum-preserving" modification: strong ellipticity holds if the unfolding matrix can be transformed into a positive definite form by redistributing values among specific entries while keeping their total sums constant.

Example of a matrix representation ($A_x = A_y$) used in the numerical verification of strong ellipticity for $n=3$.

Alternating Projection Algorithm (POCS)

To practically verify this condition, the paper proposes a numerical Alternating Projection Algorithm. This method iteratively checks if a tensor can be modified into a positive definite form through the aforementioned sum-preserving shifts. If the algorithm converges, the original tensor is proven to be strongly elliptic.

Elasticity $\mathscr{M}$-tensors

The authors define a new category of tensors inspired by the well-known "M-matrices" in computer science and economics. These elasticity $\mathscr{M}$-tensors are defined based on their off-diagonal entries and their $\mathscr{M}$-spectral radius, a specialized eigenvalue concept for tensors.

Key Insights

  • Bridging M-Matrices and Elasticity: The authors successfully extended the theory of M-matrices—prized for having "nice" properties like stable inverses—to the world of fourth-order elasticity tensors.
  • Diagonal Dominance: They proved that if an elasticity tensor (with non-positive off-diagonal entries) is "diagonally dominant" in a specific way, it is guaranteed to be strongly elliptic. This provides a simple "back-of-the-envelope" check for certain material models.
  • Perron-Frobenius Extension: The paper extends the Perron-Frobenius theorem (a cornerstone of matrix theory regarding non-negative matrices) to the $\mathscr{M}$-spectral radius of elasticity tensors.

Key Results

  • Sufficient Conditions: The paper establishes two new sufficient conditions for strong ellipticity that are easier to compute than the standard definition.
  • The $\mathscr{M}$-Tensor Guarantee: The authors prove that any "nonsingular elasticity $\mathscr{M}$-tensor" is M-positive definite, meaning it automatically satisfies the strong ellipticity condition.
  • Characterization Table: The paper concludes by providing 13 equivalent conditions (labeled C1 through C13) for a tensor to be a nonsingular elasticity $\mathscr{M}$-tensor, giving researchers a versatile toolkit for verification.
  • Algorithm Convergence: They demonstrate that their alternating projection algorithm is a reliable way to identify M-positive definiteness in practice, as shown in numerical examples for $n=2$ and $n=3$ dimensions.

Limitations and Future Work

While the conditions provided are sufficient, they are not necessary. This means a material could be perfectly stable (strongly elliptic) even if it fails the tests proposed in this paper. Additionally, the theory of $\mathscr{M}$-tensors specifically requires off-diagonal entries to be non-positive, which does not apply to all possible elastic materials. Future work may focus on numerical efficiency for higher-order cases as the field moves toward more complex material simulations.

Summary updated

Abstract

In this paper, we establish two sufficient conditions for the strong ellipticity of any fourth-order elasticity tensor and investigate a class of tensors satisfying the strong ellipticity condition, the elasticity $\mathscr{M}$-tensor. The first sufficient condition is that the strong ellipticity holds if the unfolding matrix of this fourth-order elasticity tensor can be modified into a positive definite one by preserving the summations of some corresponding entries. Second, an alternating projection algorithm is proposed to verify whether an elasticity tensor satisfies the first condition or not. Besides, the elasticity $\mathscr{M}$-tensor is defined with respect to the M-eigenvalues of elasticity tensors. We prove that any nonsingular elasticity $\mathscr{M}$-tensor satisfies the strong ellipticity condition by employing a Perron-Frobenius-type theorem for M-spectral radii of nonnegative elasticity tensors. Other equivalent definitions of nonsingular elasticity $\mathscr{M}$-tensors are also established.
BibTeX
@misc{ding2017elasticity,
  author = {Weiyang Ding and Jinjie Liu and Liqun Qi and Hong Yan},
  title = {{Elasticity $\textbackslash\{\}mathscr\{M\}$-tensors and the Strong Ellipticity Condition}},
  year = {2017},
  eprint = {1705.09911},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/1705.09911},
}
Figure