Posts Tagged "mesh quality"


It is a well-known fact in the CAE community that the efficiency and accuracy of finite element models are directly dependent on the quality of the underlying meshes in the model. The various quality parameters associated with elements are element size, aspect ratio, skew angle, jacobian, warp, and many more. Yet another parameter of concern is element topology, which means triangular/quadrilateral elements in case of shell meshes and tetrahedral/hexahedral elements in case of solid meshes. Each of these element topologies has its own advantages and disadvantages; for example, tetrahedral elements are easy to create on complex geometries but they have slower convergence, while hexahedral elements are very much desired in computational expensive simulations such as crash due to better convergence and accuracy but cannot be created easily.

Due to specific meshing requirements arising from the increasing complexity of part geometries, meshing techniques are becoming more important across all industry verticals. Transportation & mobility is primarily concerned with hexahedral meshes of pre-defined quality for very complex geometries. This industry has more focus on using Hypermesh and Ansa as a dedicated meshing tools. However these tools are primarily known for good meshing capabilities only. When there is a need to create input decks for advanced non-linear simulations such as with Nastran solution sequences 600/700 or for Abaqus multiphysics or acoustics, many of the solver features are not supported by Hypermesh or Ansa and have to be entered manually into the deck. Aerospace industry has almost always a requirement for composites modeling. They prefer a user interface that can either create or import composite plies and layups. The need for high quality meshes on complex geometries is rather rare. Due to these reasons Aerospace industry has been relying on MSC Patran since many years due to its composites modeling capabilities. However industry is now looking at alternate tools as Patran is losing its competitive edge on CAD import, CAD repair as well as meshing techniques. The CAD repair features are very minimal, there is no CAD associative interface to propagate design changes on FE side and meshing techniques offered are still at very basic level as well.

The objective of this blog is to highlight the meshing techniques in Abaqus CAE that makes CAE a tool of choice in situations where a decent quality mesh, tight integration with multiple CAD platforms as well as tight integration with Abaqus solver are topics of concern for the analyst. It’s worth mentioning for Aerospace industry audience that Abaqus CAE has basic composite modeling capabilities. For advanced composite modeling and visualization capabilities, there is an add on module called composite modeler for Abaqus CAE and there is tight integration between CATIA composites workbench and Abaqus CAE for transfer of FE meshes as well as ply layup information.


 There are primarily four meshing techniques available in Abaqus CAE, both for solid meshes as well as for shell meshes.

Free meshing: This is the easiest of all the techniques as it almost always works with a single click. It primarily generates quadrilateral or triangular elements on surfaces and tetrahedral elements on solids, even on very complex geometries. The downside is that user has very minimal control on elements quality except controlling the mesh density using global and local seeding options.

Sweep meshing: This technique is useful when hexahedral elements are needed on solids with minimal geometry editing though this technique is applicable on surfaces as well. The meshing algorithm automatically identifies a source side and a target side on the geometry, it creates a quadrilateral shell mesh on the source side and sweeps those elements to the target side thereby converting them to hexahedral or bricks. The underlying shell mesh is automatically deleted. The downside is some geometry restrictions with respect to source and target side.

Structured meshing: This meshing techniques is useful when high quality hexahedral or near to perfect shell elements are required on solids or surfaces. This technique offers a better mesh control to the user compared to sweep meshing technique. It works by partitioning the complex solids into smaller six or eight sided parametric solids that can be brick meshed. The nodes at the boundaries are automatically fused to ensure connectivity.

Bottom’s up meshing: This is the last approach when all the other meshing techniques fails. It works on the concept of divide and rule. To some extent it resembles sweep meshing but the underlying geometry restrictions are removed.


This is one feature that sets Abaqus CAE apart from other meshing tools available in the market. While doing meshing, user can see either entire part or regions associated with part (in case of partitions) in pre-defined colors. These colors helps in determining which region of the part would be meshed with which meshing technique if the mesh algorithm is executed. The color cold is as follows:











The process is quite interactive. The orange color is most undesirable as these regions are non-meshable and require further partitions. Once the region is correctly partitioned and subdivided regions become meshable, the color code is updated instantly. What the user needs to see is the combination of greens, yellows and pinks with peach at certain times before executing meshing operation. During meshing, user has option to either mesh one region at a time or the entire part having multiple regions. In case of interfaces having different element topologies on each side such as green with pink or yellow with pink, tie constraints are automatically created at the boundary to ensure mesh connectivity.

Below is an example of a part that has been partitioned to create certain sweep meshable yellow regions where brick elements are needed. The other region is pink with tetrahedral elements associated to it.



Transition of orange region to either yellow, green or peach requires intelligent partitioning of surfaces or solids. While there are many such partitioning tools available, achieving desired results with minimum partitions requires some practice in using these tools. Let’s highlight few of these partitioning methods:

Solid partitions:  Six options are available.


























This is an optional process that may be needed before partitioning and prior to meshing. This is a way to fix bad CAD data. Many times CAD data has more details than needed by the meshing algorithm. This includes very short edges and very small surfaces. Virtual topology offers certain tools to combine such small faces and edges. There is also an option to suppress these small features so that meshing algorithm does not recognize them.


© Tata Technologies 2009-2015. All rights reserved.