Automatic Hex-Dominant Mesh Generation for Complex Flow Configurations

Paper #:
  • 2018-01-0477

  • 2018-04-03
Predictive accuracy and solution convergence of computational fluid dynamics (CFD) simulations are highly dependent on the quality of computational meshes in terms of cell types, orthogonality, and cell density. Hexagonal mesh elements are preferable to tetrahedral elements for lower memory requirements, lesser computational time, and improved solution accuracy. Though desirable, creating a hexagonal mesh for a complex geometry is often a time-consuming, complicated and manual task. In this paper, bubble mesh algorithm is used to automatically generate hex-dominant meshes for complex flow configurations, such as internal combustion engines. Two important regions of the mesh are the interior mesh and the boundary layer mesh. The proposed method uses directionality and sizing control functions which offer a flexible control over interior mesh sizing. Boundary layer mesh elements are split into multiple divisions with first division having smallest thickness while rest of the layers expand linearly in thickness until they match in value with the interior mesh size. Initial simulation was performed on a flow over backward facing step problem. Additionally, results with two different complex flow configurations are presented. OpenFOAM solver was used. For the first case, fully transient large-eddy simulation (LES) predictions of flow through a sudden expansion valve using bubble mesh are compared against predictions using a fully hex mesh created using the ICEM CFD meshing tool. Mean velocity values at a constant distance from the cylinder head are compared to available experimental measurements. For the second case, steady-state Reynolds-Averaged Navier-Stokes (RANS) predictions of compressible flow through a flow bench with various static valve lifts are discussed. The goal is to predict flow separation observed in the experiments and a resulting reduction in the mass flow rate at large valve lifts. For both cases, all meshes are generated fully automatically, and numerical predictions show good agreement with experimental measurements.
SAE MOBILUS Subscriber? You may already have access.
Attention: This item is not yet published. Pre-Order to be notified, via email, when it becomes available.
Members save up to 36% off list price.
HTML for Linking to Page
Page URL

Related Items

Training / Education
Technical Paper / Journal Article
Training / Education