These example problems highlight the benefits of nonlinear mesh adaptivity:
Consider a rubber seal simulation, with an initial geometry and mesh as shown:
Downward displacement is applied to the rigid surface on the top of the seal. Following is the program output for the deformed seal:
The shape of the deformed material in the gap is not realistic because the mesh is too coarse. To create a more accurate simulation, nonlinear adaptivity with position-based criteria is used to refine the mesh of the elements which will fill the gap. Two mesh refinements are performed automatically during the solution to create a more detailed mesh in the model. Following is the first mesh refinement:
Following is the second refinement of the part of the rubber seal model that will fill the gap:
When the full load is applied in the simulation, the program outputs the following deformed seal:
With nonlinear adaptivity, the accuracy of the simulation of this type of rubber seal problem is greatly increased.
Another example of the increased accuracy achieved with nonlinear mesh adaptivity involves mesh refinements in regions with high stress concentrations. For example, consider this material cracking model:
Following is the program output for the crack simulation:
The solution and solution time are:
σx | σy | τxy | σ_VM | Elapsed Time |
---|---|---|---|---|
214113 | 91103.2 | 91641.2 | 244606 | 12.00 |
For a more accurate solution at the crack tip, nonlinear adaptivity is applied using energy-based criteria. Following is the output when three refinements are applied during solution:
The solution and solution time are:
σx | σy | τxy | σ_VM | Elapsed Time |
---|---|---|---|---|
262713 | 291816 | 164471 | 361413 | 64 |
Following is the output if the same simulation is performed with a fine mesh and no nonlinear mesh adaptivity:
The solution and solution time are:
σx | σy | τxy | σ_VM | Elapsed Time |
---|---|---|---|---|
293992 | 274278 | 160033 | 365929 | 190 |
As demonstrated, nonlinear adaptivity provides sufficient accuracy, yet with drastically reduced simulation time.
This example problem demonstrates how to apply nonlinear mesh adaptivity to remove mesh distortion and solve problems associated with high deformation. Unlike the two prior problems, general remeshing is necessary with the mesh-quality-based criterion applied in this problem.
Consider a metal forging simulation, with an initial geometry and mesh:
The flexible block has symmetry boundary conditions on three facets. A rigid die is placed on the top of the block with applied displacement in the Y direction. It is intended to compress the block to 80 percent. Without nonlinear adaptivity, the simulation cannot converge because of high mesh distortion, as shown:
The solution diverges at approximately 40 percent load with the following error:
Nonlinear adaptivity with a mesh-quality-based criterion corrects the mesh shape caused by high distortion. To check the mesh quality at every substep, skewness is set to 0.9 (NLADAPTIVE,,,MESH,SKEWNESS,0.9). With several remeshings, the full load is successfully applied, as shown with deformation in this animation:
This example problem demonstrates the benefits of nonlinear mesh adaptivity for simulating cases with very high deformation. The 2D problem uses a generalized plane strain extrusion model.
The flexible block of height = 19.5 will be compressed on the right side by a rigid die (where applied displacements = 30) to reduce its height to 6.5:
The bottom of the block has a symmetric boundary condition, the die is a frictionless rigid, and the material is elastoplastic.
Without nonlinear adaptivity, the solution diverges at approximately 80 percent load due to mesh distortion, generating the following error message:
Nonlinear adaptivity with a mesh-quality-based criterion corrects a highly deformed mesh shape. To check the mesh quality at every substep, the maximum corner angle of the mesh shape is set to 155 (NLADAPTIVE,,,MESH,SHAPE,155). With several remeshings (depending on the platform), the full load is successfully applied:
This example problem using nonlinear mesh adaptivity simulates high mesh-deformation with coupled structural-thermal degrees of freedom. The analysis uses the 2D coupled-field element PLANE222 with structural-thermal coupling and plane strain options.
A U-shape workpiece is compressed between two rigid and frictionless target surfaces. The top target surface is displaced in the negative y direction to compress the workpiece to reach a state of complete self-contact. A temperature difference is applied on the top and bottom surfaces. Temperature-dependent thermal material properties are defined in the model.
Without nonlinear adaptivity, the highly deformed mesh shape leads to a diverged solution. A mesh-quality-based criterion defined by nonlinear adaptivity measures the mesh quality and verifies a maximum corner angle of 150 degrees (NLADAPTIVE,ADD,,MESH,SHAPE,150) at every substep. With refined meshing, the solution can converge with the total displacement load applied.
This example demonstrates the benefits of nonlinear adaptivity with mesh coarsening in addition to mesh refinement and mesh-quality improvement. The 2D problem uses a hyperelastic material (rubber) which is pushed from the left through a narrow channel. The bottom of the block has a symmetric boundary condition. The contact to the channel has friction.
During the analysis, the mesh is adapted to improve the quality of deformed elements and the solution quality in regions of interest (high strain energy). Without nonlinear adaptivity, the solution diverges due to mesh distortion, when the rubber is deformed upon entering the narrow channel (at approximately 30 percent of applied load).
At the end of the analysis, the mesh is adapted to reduce the numerical effort (coarsening the mesh) for elements that have passed the narrow channel and where a refined mesh is no longer required to obtain a high-quality solution.