ReferenceAvailable
Mesh Adaptation
Adjust mesh resolution to improve accuracy or reduce the number of degrees of freedom.
Adjust mesh resolution to improve accuracy or reduce the number of degrees of freedom.
Where to find it
Mesh workspace; System → Mesh
How to use it
Choose the adaptation method and mode, set the error or size limits, and inspect each adapted mesh and solution. Choose Build Mesh to apply the sequence.
Settings and initial values
| Setting | Initial value | Meaning |
|---|---|---|
| Selection | all | all selects the applicable domains; otherwise use a named selection or the selection editor. The current estimator uses the available published nodal state fields; Selection does not restrict error recovery or uniform resizing to a local domain mask. |
| Adaptation method | Naive | Naive rescales sizes without an error estimate. Solution-driven uses recovered gradients of published nodal state fields and repeated remesh/solve steps. Only one enabled Mesh Adaptation feature is supported. Solution-driven currently requires a Stationary study, Netgen (Python), and an enabled Free Tetrahedral generator without Swept / Mapped. |
| Mode | Improve accuracy | Improve accuracy uses smaller target sizes. Reduce DOFs requests a coarser mesh to reduce degrees of freedom, meaning the number of unknowns. In solution-driven mode, coarser trial solutions are checked against the target estimated error; fewer unknowns are not guaranteed. |
| Maximum adaptation cycles | 3 | Use a positive integer. Naive applies the size factor raised to this cycle count in one sizing plan; this is not a sequence of error-checked solves. Solution-driven limits remesh/solve attempts and may stop earlier when its estimate meets the target. |
| Refinement factor | 1.5 | Use a finite dimensionless factor greater than 1. Naive Improve accuracy divides the initial sizes by this factor raised to Maximum adaptation cycles; 2 with two cycles gives one quarter of the initial size. Solution-driven divides marked local target sizes by the factor. |
| Coarsening factor | 1.25 | Use a finite dimensionless factor greater than 1. Naive multiplies initial sizes by this factor raised to Maximum adaptation cycles. Solution-driven uses coarser trial meshes and checks their estimated error. |
| Target relative error | 0.05 | Enter a positive fraction less than 1: 0.05 means 5%. This is a normalized recovered-gradient estimate from nodal state fields, not a certified bound on stress, flux or another chosen output. Naive adaptation does not use it for error control. |
| Bulk marking fraction | 0.5 | Use a positive fraction at most 1. Elements are sorted by squared error indicator and marked until their cumulative contribution reaches this fraction of the total; 0.5 targets half the squared error, not a percentage of elements. This is used by Solution-driven adaptation. |
| Maximum elements | 1000000 | Use a positive integer count of tetrahedral elements, not nodes or degrees of freedom. Solution-driven adaptation checks the current and trial mesh against this cap; it is not a target mesh density. |
Result
Inspect the element distribution and counts before running a study.
Troubleshooting
Solution-driven adaptation needs a suitable solved result. Keep a reference solution and compare the quantity you actually need.