A Breakthrough in Rotorcraft Prediction Accuracy Using Detached Eddy Simulation

Overview

Rotorcraft design engineers continually strive to improve the performance and safety of helicopters and tiltrotors, and to reduce their impact on the environment. However, rotor blades generate tip vortices and encounter those of other blades, resulting in blade vortex interaction (BVI). This BVI can affect the rotorcraft's performance, noise, and vibration levels, and is difficult to accurately predict using computational fluid dynamics (CFD) methods. Moreover, the multidisciplinary nature of rotorcraft simulation means that designers must consider interactions between rotor blade aerodynamics, blade flexibility, and blade motions to achieve trimmed, steady flight.

Project Details

A report for NASA's Rotary Wing (RW) Project stated that the figure of merit (FM), the chief hover performance parameter, was underpredicted by 2-6%, depending on the numerical approach. To bring this seemingly small discrepancy into perspective, every half-percent error results in one less passenger that a vehicle can carry. Additionally, underresolved rotor wake simulations result in inaccurate interactions between the rotor blades and the rest of the vehicle.

One goal of the RW Project is to develop physics-based CFD tools to improve the predictive accuracy of these flowfields. This is accomplished using recent improvements to NASA's OVERFLOW CFD code to solve the time-dependent Navier-Stokes equations for the V22 Osprey tiltrotor and UH-60 Black Hawk helicopter rotors in hover and forward flight.

Results and Impact

For the first time, the FM has been predicted within experimental error for the V22 Osprey and Black Hawk helicopter rotors in hover, over a wide range of flow conditions. This dramatic improvement is due to the use of CFD algorithms with high-order spatial accuracy, refined grid resolution on the rotor blades, and an improved detached eddy simulation (DES) hybrid turbulence model. High-resolution accuracy on the rotor blades was important for all cases studied, but the DES hybrid turbulence model was crucial in the BVI cases.

Dynamic adaptive mesh refinement (AMR) is used to efficiently resolve the rotor wakes by detecting vortices and locally added blocks of higher-resolution Cartesian grids. This allows the vortices and rotor wake to be resolved by grids four times as fine as, and at half the computational cost of a uniform (non-AMR) grid refinement approach. The AMR resolution revealed remarkable details of the blade vortices, and turbulent flow structures that had never before been observed either computationally or experimentally. These flow simulations have been quantitatively validated with experimental measurements.

The use of AMR greatly improved the prediction of vortex core size and growth with wake age (distance along the tip vortex). The error in vortex core growth rate has been dramatically reduced from 60% to 4% error. These groundbreaking results will impact the next generation of rotorcraft analysis and design.

Role of High-End Computing Resources

These flow simulations were made possible by NASA's Pleiades supercomputer. Solutions typically required 1,500-4,600 cores for 1-2 weeks. Grid sizes varied from 60-750 million grid points, and required 5-24 hours of wall-time per rotor revolution.