🔢 mathematics

A Staggered Discontinuous Galerkin Method for linear elasticity problem on Polytopal Meshes

This paper introduces a novel, locking-free staggered discontinuous Galerkin method for linear elasticity on polytopal meshes based on the Hellinger-Reissner principle, which ensures symmetric stress with normal continuity, satisfies local momentum balance, and features a hybridizable structure for simplified implementation and rigorous theoretical guarantees.

Original authors: Long Chen, Xuehai Huang, Ruishu Wang, Ran Zhang

Published 2026-06-16

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Long Chen, Xuehai Huang, Ruishu Wang, Ran Zhang

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine you are trying to build a perfect model of a rubber band being stretched or a bridge being loaded. In the real world, two main things happen: the material pushes back (stress), and the material moves (displacement). To simulate this on a computer, mathematicians use a set of rules called "Linear Elasticity."

This paper introduces a new, clever way to solve these rules on a computer, especially when the computer's map of the world is made of weird, irregular shapes (like a puzzle with jagged pieces) rather than perfect squares or triangles.

Here is the breakdown of their new method, using simple analogies:

1. The Problem: The "Perfect Fit" Puzzle

Traditionally, when computers try to simulate how materials stretch, they have to force the "stress" (the internal pressure) and the "displacement" (the movement) to fit together perfectly at the edges of every single piece of the puzzle.

The Old Way: It's like trying to glue two pieces of a puzzle together so tightly that they can't move at all relative to each other. If the puzzle pieces are weird shapes, this is incredibly hard to do. It often leads to "locking," where the computer thinks the material is made of steel when it's actually rubber, making the simulation fail.
The Requirement: The math needs to ensure two things:
1. Balance: The forces must balance out (Newton's laws).
2. Symmetry: The internal pressure must be balanced in all directions (like a balloon pushing out equally).

2. The Solution: The "Staggered" Dance

The authors created a method called Staggered Discontinuous Galerkin (SDG). Think of this as a dance where two partners (Stress and Displacement) are allowed to step out of sync, but in a very specific, organized way.

The Mesh (The Floor): Imagine the floor is covered in irregular tiles (polytopes).
The Dual Mesh (The Mirror): They create a second, invisible layer of tiles by connecting the center of each tile to its corners. Now, every tile has a "partner" tile.
The Staggered Move:
- Stress is allowed to be "jumpy" (discontinuous) inside the original tiles, but it must flow smoothly across the edges of the original tiles.
- Displacement is allowed to be "jumpy" inside the partner tiles, but it must flow smoothly across the edges of the partner tiles.

It's like two groups of people walking on a dance floor. Group A (Stress) holds hands with their neighbors on the left, while Group B (Displacement) holds hands with their neighbors on the right. They don't need to hold hands with each other directly; they just need to follow the rhythm of their own specific grid. This "staggered" arrangement makes the math much more stable and prevents the "locking" problem.

3. The "No-Vertex" Trick

Most old methods required knowing exactly what was happening at the sharp corners (vertices) of the puzzle pieces. This made the computer code very heavy and slow because every corner had to talk to every other corner.

This new method removes the corners.

The Analogy: Imagine a relay race. Old methods required every runner to stop and high-five the person at the exact corner of the track. This new method says, "Just pass the baton along the edge of the track."
Why it matters: By ignoring the corners and focusing only on the edges, the computer can solve the problem much faster. It uses a technique called "hybridization" to break the big, scary math problem into tiny, independent problems that can be solved easily and then stitched back together.

4. The "Locking-Free" Guarantee

One of the biggest headaches in simulating materials is "locking." This happens when the math gets so stiff that it can't simulate soft, squishy materials (like rubber or biological tissue) accurately.

The Paper's Claim: Their new method is "locking-free." This means whether you are simulating a stiff rock or a soft jelly, the math works perfectly. It doesn't get confused when the material becomes very hard to compress.

5. The Results

The authors tested this method on computers using:

Triangles: Standard shapes.
Polygons: Weird, multi-sided shapes (like a pentagon or hexagon).

They proved mathematically that the method is stable and accurate. Their computer tests showed that as they made the puzzle pieces smaller, the answer got more accurate at the expected speed, regardless of how "squishy" or "stiff" the material was.

Summary

In short, this paper presents a new, robust way to simulate how materials stretch and bend. It does this by:

Using a "staggered" grid system where stress and movement follow different but complementary paths.
Getting rid of the need to calculate values at sharp corners, which speeds up the computer.
Ensuring the simulation works perfectly for both hard and soft materials without breaking.

It's a new set of instructions for computers to build better, faster, and more accurate models of the physical world, even when the world is made of messy, irregular shapes.

Technical Summary: A Staggered Discontinuous Galerkin Method for Linear Elasticity on Polytopal Meshes

Problem Statement
This paper addresses the numerical approximation of the linear elasticity problem using the Hellinger-Reissner mixed variational principle. The goal is to simultaneously approximate the symmetric stress tensor $\sigma \in H(\text{div}, \Omega; S)$ and the displacement field $u \in L^2(\Omega; \mathbb{R}^d)$ . The primary challenges in discretizing this mixed formulation are threefold: (i) satisfying the inf-sup condition to ensure stability; (ii) enforcing both the symmetry of the stress tensor and its normal continuity across element boundaries; and (iii) achieving optimal convergence rates. Existing methods often rely on vertex degrees of freedom (DoFs), which hinder hybridization, or employ weak symmetry, which may not strictly satisfy the balance of angular momentum locally. Furthermore, many methods struggle with "locking" in the incompressible limit ( $\lambda \to \infty$ ).

Methodology
The authors propose a novel Staggered Discontinuous Galerkin (SDG) method defined on arbitrary polytopal meshes. The approach utilizes a primal-dual mesh structure:

Mesh Construction: A primal mesh $K_h$ consists of general polytopal elements. A dual mesh $K^*_h$ is constructed by connecting the barycenter of each primal element to its vertices, creating a refined triangulation $T_h$ .
Staggered Continuity: The method reverses the continuity assignments found in previous SDG elasticity works.
- Stress ( $\sigma_h$ ): Defined as piecewise polynomials on the refined triangulation $T_h$ . Crucially, the stress is strongly symmetric and possesses normal continuity across the boundaries of the primal elements ( $K \in K_h$ ).
- Displacement ( $u_h$ ): Approximated by piecewise polynomials on the primal elements $K_h$ . The displacement is continuous across the boundaries of the dual elements ( $\omega_F \in K^*_h$ ).
Finite Element Spaces:
- The stress space $\Sigma_h$ is constructed to be $H(\text{div})$ -conforming with strong symmetry. The local DoFs are defined via normal fluxes on primal faces and interior moments, avoiding vertex-based unknowns.
- The displacement space $U_h$ consists of vector-valued polynomials on primal cells.
- For the lowest order ( $k=0$ ), the stress space is enriched with rigid motion spaces to ensure well-posedness.
Variational Formulation: The scheme is derived from the first-order system of linear elasticity. It includes stabilization terms involving the divergence of the stress and jumps in normal fluxes across internal faces of the refined triangulation. These terms are designed to prevent numerical locking rather than to enforce continuity.
Hybridization: The method admits a hybridizable structure. By introducing Lagrange multipliers on the element boundaries, the stress variable can be statically condensed, resulting in a global system involving only the boundary DoFs of the displacement.

Key Contributions

Strong Symmetry on Polytopes: The paper constructs symmetric stress spaces with normal continuity on arbitrary polytopal meshes without relying on vertex DoFs. This is achieved through a specific geometric decomposition of the stress space and the careful selection of DoFs.
Reversed Continuity: Unlike previous SDG elasticity methods [31, 41] that enforce weak symmetry, this work enforces strong symmetry by interchanging the continuity patterns of stress and displacement relative to the primal and dual meshes.
Locking-Free Property: The method is proven to be locking-free. The stability and convergence remain robust in the incompressible limit ( $\lambda \to \infty$ ) because the stabilization terms control the trace of the stress independently of the Lamé parameters.
Local Conservation: The formulation preserves the local balance of linear momentum (via the divergence equation) and angular momentum (via the strong symmetry of the stress tensor).
Hybridizable Structure: The absence of vertex DoFs allows for a hybridization procedure that significantly reduces the size of the global linear system, simplifying implementation and reducing computational cost.

Results

Theoretical Analysis: The authors provide a comprehensive theoretical analysis, proving the well-posedness of the scheme via discrete inf-sup conditions and coercivity. Optimal a priori error estimates are derived, showing convergence rates of $O(h^{k+1})$ for the stress in the energy norm and the displacement in a discrete $H^1$ -like norm, and $O(h^{k+2})$ for the displacement in the $L^2$ norm.
Numerical Experiments:
- Locking-Free Validation: Numerical tests on triangular meshes with varying Lamé parameters ( $\lambda = 1, 10^2, 10^4, 10^6$ ) demonstrate that the error remains independent of $\lambda$ , confirming the locking-free nature of the method.
- Convergence Rates: Experiments on both triangular and general polygonal meshes confirm the theoretical convergence orders for $k=0$ and $k=1$ .
- Polytopal Flexibility: The method successfully handles general polygonal meshes, demonstrating its applicability beyond standard simplicial or quadrilateral grids.

Significance
The paper claims that this work extends the SDG framework, previously applied to the Poisson equation [16], to the more complex field of linear elasticity. The significance lies in providing a method that combines strong symmetry, normal continuity, and hybridizability on general polytopal meshes without vertex degrees of freedom. This addresses a long-standing difficulty in mixed finite element methods for elasticity, where satisfying symmetry and continuity simultaneously often leads to complex elements with vertex DoFs or requires weak symmetry. The proposed method offers a robust, locking-free alternative that is theoretically sound and computationally efficient due to its hybridizable structure.