Improving the local solution accuracy of large-scale digital image-based finite element analyses.

Area of Science:

• Biomechanics
• Computational modeling
• Materials science

Background:

• Digital image-based finite element modeling (DIBFEM) is efficient for meshing biological structures like trabecular bone.
• DIBFEM accurately predicts apparent mechanical properties but struggles with local phenomena due to boundary errors.
• Mesh refinement does not always decrease local maximum errors in DIBFEM.

Purpose of the Study:

• To evaluate a post-processing filtration method for reducing local solution errors in DIBFEM.
• To assess the effectiveness of a 3D, boundary-specific filtering algorithm.
• To improve the simulation of local phenomena such as tissue failure and adaptation.

Main Methods:

• A 3D, boundary-specific filtering algorithm was applied as a post-processing step.
• The study analyzed the mesh size dependency of the filtering algorithm.
• Simulations were performed on representative trabecular bone microstructures.

Main Results:

• The filtering algorithm's effectiveness was mesh size dependent.
• Mean absolute and maximum errors decreased for meshes with >5 elements across the diameter.
• Filtered solutions showed consistently reduced errors with mesh refinement, unlike non-filtered solutions.
• A 30% reduction in maximum stress was observed in a trabecular bone model.

Conclusions:

• Boundary-specific filtering effectively reduces local mean and maximum errors in DIBFEM.
• This method enhances the accuracy of local phenomenon simulations (e.g., tissue failure).
• The technique maintains the efficiency of DIBFEM for model generation and storage.