An Improved Finite Element Modeling Method for Triply Periodic Minimal Surface Structures Based on Element Size and Minimum Jacobian

Siqi Wang, Chuangyu Jiang, Xiaodong Zhang, Yilong Zhang, Baoqiang Zhang, Huageng LuoarXiv 2025 · DOI · arXiv

Abstract

Triply Periodic Minimal Surface (TPMS) structures, such as the Gyroid, are increasingly utilized in additive manufacturing for their superior mechanical properties. However, accurately simulating these complex geometries using traditional voxel-based finite element modeling (FEM) often requires excessive computational resources due to the "staircase effect." This paper presents an improved two-parameter voxel method that jointly controls element size ($L$) and the Minimum Jacobian (MJ). By allowing controlled element distortion, the method achieves higher geometric fidelity and numerical convergence with fewer elements. Results indicate a 50% reduction in solving time and significantly improved alignment with experimental data, establishing a more efficient framework for lattice structure analysis.

Introduction

Triply Periodic Minimal Surface (TPMS) structures are characterized by their zero-mean-curvature surfaces and periodic nature in three dimensions. Their high strength-to-weight ratios and energy absorption capabilities make them ideal for high-performance engineering applications. Despite these advantages, simulating TPMS geometries remains a challenge.

Standard voxel-based modeling is frequently employed due to its simplicity in discretizing complex volumes into a 3D grid of cubic elements. However, representing smooth, curved TPMS surfaces with rigid cubes leads to a "staircase effect," where the model overestimates the actual volume and stiffness. To mitigate this, researchers typically reduce the element size, leading to models with millions of elements that are computationally expensive to solve. Furthermore, a lack of consensus on mesh convergence criteria for these structures complicates the validation of simulation results.

Methodology

The proposed research introduces a two-parameter voxel method designed to balance geometric accuracy and computational efficiency. Unlike traditional methods that rely solely on shrinking element size ($L$), this approach introduces the Minimum Jacobian (MJ) as a second control parameter.

Comparison of traditional voxel method (MJ=1.0) and the proposed two-parameter voxel method across varying element sizes (L) and Minimum Jacobian (MJ) values.

By lowering the MJ threshold, elements are permitted to deviate from perfect cubes, allowing the mesh to "warp" and better conform to the curved boundaries of the TPMS geometry. The workflow involved generating TPMS geometries in MATLAB and using the "Shrink Wrap" module in Hypermesh for discretization. To ensure rigorous validation, the Grid Convergence Index (GCI) based on Richardson extrapolation was employed to estimate the true numerical solution and assess convergence across different mesh densities.

Results and Discussion

The study identified a critical threshold for the Minimum Jacobian. Simulation results, particularly the equivalent Young’s modulus, were found to stabilize when the MJ reached approximately 0.3. This "sweet spot" allows for significant improvements in mesh fidelity without the exponential increase in element count associated with traditional voxel refinement.

Relationship between the number of elements and (a) element size L, and (b) Minimum Jacobian MJ, illustrating the convergence behavior of the proposed method.

Key performance metrics include:

  • Computational Efficiency: For a Gyroid structure with a relative density of 0.45, the two-parameter method reached a solution in 58 seconds, compared to 121 seconds for a traditional voxel model of similar accuracy—a 52% improvement in efficiency.
  • Convergence Accuracy: The two-parameter model exhibited a higher order of convergence ($p = 2.756$) than the standard model ($p = 1.358$), indicating more robust numerical behavior.
  • Geometric Fidelity: Traditional voxel models (MJ=1.0) tend to "bloat" the geometry, leading to an overestimation of stiffness. The proposed method corrects this by better approximating the intended relative density.

Interestingly, the method showed even greater performance gains when applied to graded structures, where density varies spatially, demonstrating its versatility for complex lattice designs.

Conclusion

The integration of Minimum Jacobian regulation alongside element size control provides a powerful tool for the finite element analysis of TPMS structures. While highly effective for implicit solvers, the method faces limitations in explicit time integration, where distorted elements can reduce the stable time step. Future research will focus on optimizing MJ parameters specifically for explicit dynamics to maintain the balance between accuracy and simulation speed.

Summary updated

BibTeX
@misc{wang2025improved,
  author = {Siqi Wang and Chuangyu Jiang and Xiaodong Zhang and Yilong Zhang and Baoqiang Zhang and Huageng Luo},
  title = {{An Improved Finite Element Modeling Method for Triply Periodic Minimal Surface Structures Based on Element Size and Minimum Jacobian}},
  year = {2025},
  eprint = {2506.04028},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2506.04028},
}
Figure