Why is mesh sensitivity so important?

Mesh sensitivity analysis in Abaqus evaluates how the results of an FEA (finite element analysis) change with varying mesh densities. It ensures that the numerical solution converges to an accurate and reliable result as the mesh is refined. This process is critical for achieving dependable simulations without unnecessary computational costs.

In FEA, the solution tends to converge to a unique value as the mesh density increases. A “converged mesh” is one where further refinement produces negligible changes in results, such as stresses, displacements, or energy values.

For example, in a CEL (coupled Eulerian-Lagrangian) study involving a ball-water interaction, different mesh densities produced varying outcomes, such as the number of bounces or whether the ball was submerged. This highlights how mesh density can drastically affect results in dynamic simulations.

While finer meshes improve accuracy, they also increase computation time and resource usage. A balance must be struck between mesh density and computational efficiency by focusing refinement on critical regions (e.g., areas with high stress gradients).

Mesh dependency in CEL studies

In CEL analyses, such as impact or fluid-structure interaction problems, mesh dependency can influence energy dissipation and deformation patterns. Refining the mesh improves convergence, but may also alter physical behaviors observed in simulations.

Practical recommendations

• Perform a mesh convergence study by running simulations with progressively finer meshes until results stabilize.
• Use coarse meshes for trend analysis, but rely on refined meshes for accurate magnitudes of stress or displacement.
• Apply local refinement in regions of interest (e.g., stress concentrations) while keeping coarser meshes elsewhere to optimize performance.