Advanced Computational Fluid Dynamics Tools for Accurate Rotorcraft Analysis and Design

Overview

Imagine trying to build a new home without power tools! Your job would become much more difficult to complete in a timely and cost-effective manner, and the construction would lack the precision that power tools provide. One of the goals of NASA’s Revolutionary Vertical Lift Technology Project (RVLT) is to develop and validate advanced high-fidelity computational fluid dynamic (CFD) tools with improved physical realism and numerical solution accuracy that can be used in the rotorcraft design process. The recent focus on urban air mobility is one example of innovative concepts that require advanced CFD tools. These designs may use a dozen rotors with electric propulsion. The vision is to improve regional and commute transportation alternatives that are both environmentally responsible and quiet in urban locations. Both low-fidelity, preliminary design tools and advanced vehicle concepts need to be validated with high-fidelity methods.

Project Details

The numerical approach of adaptive mesh refinement (AMR) detects where the solution is changing rapidly and automatically refines the computational mesh to improve flow resolution and solution accuracy. This approach only refines the local mesh where needed and is more computationally efficient than refining the mesh throughout the entire domain. Off-body (OB) AMR has been successfully used in the past to refine the rotor’s blade-tip vortices and vortex wake with embedded Cartesian meshes. Recently, near-body (NB) AMR has been used—for the first time—to refine the flow features on flexible rotor blades with embedded curvilinear meshes. NB-AMR was applied to the very challenging dynamic stall problem, where flow separation on a helicopter rotor blade causes a sudden reduction in thrust, increase in drag, and a nose-down pitching moment. Moreover, this condition also causes large structural vibrations and limits the maximum flight speed of the helicopter.

Results and Impact

In the figure above, NB-AMR is applied to 2D dynamic stall for an airfoil oscillating in angle of attack: (a) shows the flow streamlines about the airfoil at its maximum angle of attack. Note the large, leading-edge vortex passing over the trailing edge, causing a loss in lift and inducing a strong trailing-edge vortex that causes a nose-down pitching moment. Secondary vortices appear near the airfoil’s leading edge. Images (b, c, d) show the dynamic NB-AMR process. The refined curvilinear meshes are automatically embedded into the solution. This greatly improves the resolution of the separated vortical flow and improves the solution accuracy in a cost-effective manner. The figure below shows a Blackhawk helicopter and a CFD solution for the rotor during dynamic stall. Both NB- and OB-AMR are used. You can see the three-dimensional (3D) flow details are much more complex than the simple airfoil problem. In this case, blade-vortex interaction (BVI) causes dynamic stall, a feature that is missing in the simplified 2D problem. There are also small donut-shaped vortex rings, the 3D analog of the 2D vortex in the first figure. The goal is to demonstrate the use of this technology and establish how to best use it for accurate engineering rotorcraft flow analysis and design.

Why HPC Matters

The Blackhawk rotor simulation was carried out on the Electra supercomputer at the NASA Advanced Supercomputing (NAS) facility, using approximately 675 million grid points, 14,500 grids, and 5,600 Intel Xeon Skylake cores. Each rotor revolution required 60 wall-clock hours. HPC resources are needed for these complex processor-intensive simulations.

What’s Next

We are now working to further improve the NB-AMR process by reducing the computational cost and establishing best practices for engineering analysis and design.

More Information

• See the poster for this demonstration (PDF, 2.0MB). As shown in the NASA booth at SC18.