Accelerating Speed And Accuracy In Aeroacoustic Predictions

Simulating aeroacoustics is a tedious task that blends the complexities of fluid dynamics with the nuances of acoustic phenomena. This dual-domain challenge is marked by numerous technical obstacles, ranging from the wide range of scales involved to the subtler distinctions of acoustic signals amidst turbulent flows. To navigate these technical difficulties and advance our understanding of aeroacoustic behaviors, engineers and researchers must employ sophisticated tools like the Cadence Fidelity LES. This blog post explores the multifaceted challenges of aeroacoustic simulations and showcases case studies where Fidelity LES has proven helpful in solving these intricate problems.

Challenges simulating aeroacoustics

Simulating aeroacoustics presents significant challenges. Below is a list of some of these shortcomings:

• Wide Range of Scales: Aeroacoustic phenomena cover extensive spatial and temporal scales. Capturing every scale demands fine grid resolutions and long simulation times.
• Acoustic Wave Amplitudes: Aeroacoustic signals are often subtler than turbulent flow pressure fluctuations, making it tough to distinguish them from dominant flow structures.
• Far-Field Propagation: Sound generated by local aerodynamic sources can propagate over large distances, making full-domain simulations computationally prohibitive.
• Complex Geometries: Real-world problems often involve intricate geometries, such as aircraft engines or vehicle exteriors, which complicate fluid flow and sound propagation modeling.
• Boundary Conditions: Choosing and implementing appropriate boundary conditions is critical to avoid spurious reflections or non-physical behaviors.
• Transient Nature: Many problems are unsteady and require transient simulations, increasing computational effort and complicating statistical analysis.
• Nonlinear Interactions: High sound levels can involve nonlinear aerodynamic and acoustic interactions, requiring additional computational resources.
• Multiphysics Interactions: Some simulations need to account for other physical effects, like heat transfer or combustion, complicating the setup.
• Numerical Dissipation: Numerical methods can introduce artificial dissipation, dampening or suppressing acoustic signals of interest.

Simulation of a supersonic jet traveling faster than the speed of sound (Mach 1).

Faster and accurate aeroacoustics simulation

Choosing the right simulation software is crucial for tackling real-world engineering challenges and fundamental research in aeroacoustics. Fidelity LES is a premier tool for high-fidelity flow analyses, including aeroacoustics. This software leverages large eddy simulation (LES) by integrating advanced numerical techniques and models to simulate unsteady flows with reduced dissipation and dispersion. Employing various solver formulations based on the finite volume method, Fidelity LES effectively captures different flow conditions, including low-speed, high-speed, and reacting flows.

Fidelity LES excels across varying grid resolutions thanks to advanced sub-grid and wall modeling. It accurately captures flow phenomena even on coarse grids, demonstrating impressive efficiency and scalability on both CPUs and GPUs. Notably, one V100 GPU matches the processing power of nearly 400 Intel Skylake 2018 CPUs for its implicit low-Mach solver. Engineers can seamlessly manage their entire simulation workflow with Fidelity LES, a user-friendly application that covers everything from preparing geometries to analyzing results.

Intel Skylake 2018 CPU core equivalent per NVIDIA A100 and V100 GPUs for various flow scenarios using Fidelity LES.

Fidelity LES offers detailed 3D views of intricate engineering models, including real-time adjustment of simulation settings. Additionally, users benefit from a command glossary with automatic suggestions and graphs tracking key metrics.

Aeroacoustic case studies with Fidelity LES

1. NASA’s fan noise source diagnostic test

A wall-modeled LES of NASA’s fan noise SDT was simulated using Fidelity LES to scrutinize the effects of varying outlet guide vane (OGV) configurations on the aerodynamics and acoustic profile of the fan. The computational setup included the fan, OGVs, nacelle, and full test section, which minimized numerical inaccuracies. The fan operated at a reduced rotational speed of 7,809 rpm, or about 61.7% of its design speed.

SDT fan with three different OGV configurations.

Results indicated that the low noise OGV design lowered noise levels by roughly 2 dB compared to standard and low count OGV setups, corroborated by real-world tests. Despite suboptimal mesh resolutions in areas like the fan blade, OGV surfaces, and the tip gap, the simulations demonstrated strong alignment with experimental data. Fidelity LES achieved high accuracy, with errors as low as 0.5% in certain aerodynamic efficiencies, showcasing its robust predictive capabilities for future aeroacoustics research.

Flow Mach number and surface shear stress for the NASA fan SDT with low noise OGV configuration at approach condition (61.7% design speed) from the GPU-accelerated Fidelity LES simulation. 

2. Full-scale car with Honda R&D

Honda R&D used Fidelity LES for simulating the aeroacoustic profile of a Sedan at 120 kph, comparing 2 mm and 1 mm mesh resolutions. The finer mesh uncovered detailed turbulence (vortex shedding), especially near the A-pillar, affecting aerodynamics and noise. Validation against wind tunnel data showed a close match. Additionally, performance on GPUs versus CPUs was assessed, with 32 Nvidia V100 GPUs completing the task in 1.5 hours, significantly faster than 2560 AMD EPYC CPUs, which took 6 hours. This efficiency is vital for modern vehicle design, including in the eVTOL aircraft industry.

Time average (top images) and rms (bottom images) of the surface pressure on the top and right-hand side of the full-scale sedan car for the LES cases Acs118M and Acs253M.

3. Multiple blade VTOL rotors with Honda R&D

In a study leveraging Fidelity LES, researchers investigated the aeroacoustic performance of VTOL rotors for drones and urban air mobility vehicles, focusing on the accuracy and efficiency of high-frequency aeroacoustic simulations. They examined rotors with 2 to 5 blades, comparing their findings with data from Honda’s wind tunnel tests. The study found that noise levels increased with the number of rotor blades, aligning with empirical data. Additionally, Fidelity LES was used to simulate a full-scale eVTOL aircraft with eight rotors and two propellers, demonstrating the benefits of GPU acceleration. This research highlights Fidelity LES’s potential to develop quieter, more efficient aerial vehicles for urban use.

Surface shear stress on Honda’s full-scale eVTOL vehicle (color scale) and instantaneous pressure in two horizontal and vertical planes (greyscale).

In the complex realm of aeroacoustic simulations, overcoming challenges such as scale disparity, subtle acoustic wave distinctions, and non-linear interactions is critical for advancing both engineering applications and fundamental research. The Cadence Fidelity LES stands out as an exemplary tool, demonstrated by its proven capabilities in various demanding scenarios, including NASA’s fan noise diagnostics and Honda R&D’s vehicle and VTOL rotor simulations. By providing high accuracy, efficiency, and scalability, particularly through GPU acceleration, Fidelity LES enables researchers and engineers to push the boundaries of aeroacoustic understanding and develop quieter, more efficient designs. Download the Engineer’s Guide to Simulating Aeroacoustics to learn more.