LES Workflow with Turnaround Time of Less Than 10 Hours

Large Eddy Simulation (LES) has recently gained importance in computational fluid dynamics (CFD). This surge is primarily due to the inherent trade-offs involved in traditional Reynolds-averaged Navier Stokes (RANS)-based CFD, such as limited design spaces, long run times, and reduced physics. Code-design and computing architecture advancements have made high-fidelity simulations such as LES attainable for complex industrial problems like boundary layer separation, aeroacoustics, and combustion. Not only do these advancements provide more confidence in our simulation results, but they also significantly enhance LES solver performance through the utilization of GPU compute architectures. With these improvements, LES workflows can now be applied to real-world engineering tasks, achieving turnaround times of less than 10 hours, thereby making LES a practical choice for production-level CFD environments.

In this blog post, we will focus on how users can benefit from an LES workflow using Fidelity Pointwise and Fidelity LES Solver, formerly CharLES, specialized meshing tool, and GPU-enabled wall-modeled LES code, respectively.

CFD Prediction for High-Lift Aerodynamics (Slotnick, 2019)

Why Fidelity LES Solver?

Four compelling reasons why Fidelity LES Solver is the game changer for high-fidelity LES simulations -

• Industry-first Voronoi diagram-based massively parallel mesh environment
• Robust, non-linearly stable numerical schemes and advanced physical models
• Rapid visualization and interrogation for large datasets
• Scalable, GPU-resident multi-physics flow solvers

Preprocessing is a crucial phase in the CFD workflow, as it significantly affects overall accuracy and typically takes up about 75 to 80% of the total workflow time. To assist CFD users during this stage, Fidelity LES has developed an advanced meshing tool called Fidelity Stitch. This tool is designed to improve accuracy and decrease the time needed to enhance mesh quality metrics, thereby making the preprocessing workflow much more efficient.

Fidelity Stitch is a Voronoi diagram-based volume meshing tool for LES. Voronoi diagrams are a unique partition based on the Euclidean distance. This meshing process has two inputs.

1. The first input is for the user to import a watertight and manifold surface mesh that ‘Stitch’ will use to clip the diagram.
2. The second input is generating sites. The topology is an outcome of generating site placement and the intersection of that site stencil with the surface mesh. Stitch will then generate arbitrary polyhedral cells directly.

Lloyd's algorithm is used to smooth the mesh iteratively. This smoothing procedure produces favorable near-wall alignment and distributes cell volume more uniformly at interfaces where high resolution is essential.

Why Fidelity Pointwise?

The four points that make Fidelity Pointwise the preferred choice for meshing in the LES workflow are as follows –

• Quick generation of high-quality surface meshes on complex geometry for Stitch.
• Process is robust, generic, and can be automated using the Flashpoint suite of tools
• Customizable workflow
• Export the SBIN format native to Stitch, directly

About Flashpoint: ASM in Fidelity Pointwise

Automatic Surface Mesh (ASM) is a goal-driven tool designed to encapsulate “best practices” in surface meshing. Starting with a watertight geometry model, ASM creates seamless, smoothly matched, and properly clustered domains. While the tool is completely “automated,” the user can fine-tune and control all meshing parameters to tailor the surface mesh to their specific application.

The ASM Command panel has three tabs:

1. Global to set all the global parameters and “goals” for the surface meshing, such as mesh algorithm, curvature resolution, growth rate, etc.
2. Surfaces to define surface resolution or mapping filters to directly control the local mesh resolution (per quilt/surface regions)
3. Boundaries to define boundary stretching filters for anisotropic meshing (2D T-Rex). The user can manually enable/disable/control stretching on automatically classified boundaries as well as user-defined filters. 

Fidelity Pointwise to Fidelity LES Solver Workflow

The Fidelity Pointwise to  Fidelity LES Solver workflow begins with the importing of a CAD file into Fidelity Pointwise. The user can then prepare and arrange the model surfaces to suit their engineering needs. A high-quality surface mesh can be swiftly generated on the model by utilizing Flashpoint's ASM feature in Fidelity Pointwise. This is succeeded by the creation of the far-field and symmetry boundaries within the fluid domain using the automatic volume meshing (AVM) feature in Fidelity Pointwise. Subsequently, boundary conditions are established, and the setup is exported in the native SBIN format.

The entire process is designed for full automation, enabling the generation of a high-quality mesh with minimal manual adjustment. Further, utilizing the built-in tools within Fidelity Pointwise, additional tweaks can be implemented to enhance the quality of the surface mesh and streamline the workflow more effectively. Ultimately, the SBIN file is employed to generate the volume mesh in Stitch, facilitating the setup and execution of simulations using the Fidelity LES Solver. The seamless integration of these steps demonstrates the efficiency and capability of this comprehensive solution.

In conclusion, co-design and computing architecture advancements have significantly enhanced the feasibility and efficiency of LES workflows for complex industrial applications. These improvements make LES a practical option for production-level CFD environments, reducing turnaround time. Users can harness the power of GPU-enabled wall-modeled LES codes combined with robust, automated meshing tools to achieve high-fidelity simulations by utilizing Fidelity Pointwise and Fidelity LES Solver. These capabilities highlight why Fidelity LES stands out as an industry leader, providing rapid visualization for large datasets and scalable, multi-physics flow solvers that address a wide array of engineering challenges.

References

Slotnick, J., "CFD Prediction for High Lift Aerodynamics Recent Progress and Emerging Opportunities," October 2019. Retrieved from: https://www.aerosociety.com/media/12473/15-jeffrey-slotnick.pdf

Watch the on-demand video to learn how to generate high-fidelity LES simulations with Fidelity Pointwise to  Fidelity LES Solver workflow.