Smoother Transitions with Mosaic Meshing

By Harsh Vardhan, Director of Software, Development – Meshing, ANSYS

Transitioning between various types of mesh elements in complex geometries and flow regimes has long been a major CFD simulation challenge. ANSYS Mosaic technology automatically and conformally connects any type of mesh to any other type with general polyhedral elements. The result is faster simulations with greater solution accuracy while using less RAM.

Save PDF
Smoother Transitions with Mosaic Meshing

"Mosaic technology ensures that the best type of mesh element is used in every section of the geometry for optimal results."

Mosaic meshing turbulent airflow

Detailed aerodynamics SBES simulation of DrivAer model shows details of Mosaic meshing and turbulent airflow around the wheel hub and the external side mirror.

Accuracy and performance are two of the most critical concerns in computational fluid dynamics (CFD) simulation, and both are highly dependent on the characteristics of the mesh. To realize the power of simulation in early design stages and test many variant designs, the meshing process should be automated. For an automated meshing process, diverse types of elements are needed to deliver optimal performance in resolving different geometries and flow features. But transitioning between varying types of elements poses significant challenges. The transition zone has typically relied on non-conformal interfaces or on pyramids/tetrahedra, but these could come with issues regarding solver performance, mesh quality and excessive cell count.

ANSYS has developed the new, patent-pending Mosaic technology, which conformally connects any type of mesh to any other type of mesh, to eliminate the need to compromise. This ensures that the best type of mesh element is used in every section of the geometry for optimal results.

Mosaic technology automatically connects different types of meshes with general polyhedral elements. Poly-hexcore, the first application of Mosaic technology, fills the bulk region with octree hexes, keeps a high-quality, layered poly-prism mesh in the boundary layer and conformally connects these two meshes with high-quality general polyhedral elements. The resulting simulation is faster with greater solution accuracy while using less RAM.

A NEW PARADIGM IN TURBULENCE MODELING

GEKO turbulence model

GEKO turbulence model

The GEKO (generalized k-omega) turbulence model is suitable for all flow applications. It has the flexibility to tailor turbulence models to specific applications. In this example, GEKO’s free and tuneable parameters were adjusted to match simulation to the physical measurements.

CASE STUDY: USING MOSAIC MESHING TO MODEL AUTOMOBILE AERODYNAMICS

When designing a new automobile, the aerodynamics of the vehicle are of major importance. To understand fundamental flow phenomena to great effect, and to openly share/publish/compare simulation and physical testing practices and results, researchers at the Institute of Aerodynamics and Fluid Mechanics at Technical University of Munich (TUM) proposed a customizable DrivAer model in 2011. The DrivAer body, with options for 18 different parameters (for example, fastback, estate-back or notchback rear-end configurations; a detailed underbody or a smooth one; with mirrors or without, etc.) enables engineers to investigate many more detailed and complex aerodynamic phenomena and share their findings widely. The DrivAer geometry is available for free download to interested parties and is being used extensively throughout the automotive industry.

MESHING THE DRIVAER MODEL

Surface Meshing

The surface mesh was created with isotropic triangles using advancing-front meshing technology. The sizing definition is very simple and allows automatic refinement of the mesh in areas of high curvature, small features and close proximity to surfaces. Users can add specific sizing as required. Surface mesh refinement is also applied on surfaces where wake volumetric refinement has been specified using a series of scaled offset shapes from the car geometry.

Volume Meshing Using Mosaic

Engineers applied Mosaic meshing to the volume including the separate, conformal volume regions inside the wheel hubs. This allows the solver to apply the correct rotational body forces in the simulation. In such simulations, accurate prediction of boundary layer separation is crucial, so the resolution of the mesh close to car surfaces with extruded prisms is important for accuracy. For this reason, 20 highly anisotropic poly-prism elements (y+ approximately equal to 1) were used to capture the boundary layer accurately with smooth growth into the bulk mesh. Away from walls, perfect hex elements were automatically generated to efficiently fill the volume and capture gradients and vorticity in off-body flow. To connect the prism and the hex elements a polyhedral “transition layer” was generated by Mosaic to give high quality and low mesh count transition between the two.

For this external aerodynamics simulation, engineers generated an extremely high-quality mesh with 63 million elements. This provides a huge saving in number of elements when compared to the traditional hexcore with trianglebased prisms and tetrahedral transition layer, which consisted of 106 million elements. The Mosaic polyhexcore mesh also utilized parallel meshing and took less than 16 minutes to generate on 32 cores requiring less than 115 GB of RAM. ANSYS Fluent users can mesh on distributed memory HPC clusters directly from the software without need for additional HPC licenses.

CFD Simulation

Fluent users embracing the new Mosaic technology are reporting up to 2-times solver speedup with similar or better levels of accuracy than was observed using a purely polyhedral mesh approach. External aerodynamics innovations in the solver are also leading to better predictions than ever before using new proprietary turbulence-modeling capabilities. The GEKO (GEneralized K-Omega) model is a tunable RANS model, which allows safe modification of parameter sets to better match experimental data for a wide range of turbulence phenomena, including separation and vorticity. Stress-blended eddy simulation (SBES), on the other hand, is an enhanced scale-resolving unsteady turbulence model which gives unparalleled resolution of fine turbulence structures within a hybrid RANSLES framework. Both innovative models combine well with Mosaic meshing technology to deliver a leap in productivity and accuracy for external aerodynamics simulation.

Mosaic mesh-connecting technology has the potential to deliver exciting new combinations of meshing elements that will help meet the challenge of increasing complexity and accuracy requirements for years to come. Since Mosaic is an enabling technology, you can expect to see it appearing in other ANSYS meshing workflows soon.

poly-hexcore meshing for simulation

Mosiac-enabled mesh was used to capture details of turbulent flow around the rear of this car.

click below to start a conversation with ANSYS

Contact Us
Contact Us
Contact