ENGINEERING AND CFD

Introduction

Several research activities are currently in progress at the Chair of Aerodynamics and Fluid Mechanics (AER) of the Technical University of Munich (TUM) [1]. This report provides an overview of the SuperMUC project pr86fi entitled ‘Aerodynamic Investigations of Vortex Dominated and Morphing Aircraft Configurations with Active and Passive Flow Control’ and consists of three parts.

In the first part, a diamond wing configuration with low-aspect ratio and high leading-edge sweep angle is investigated. For a safe and efficient flight of such a configuration, a sound knowledge about the control surface efficiency and the static and dynamic aerodynamic characteristics is indispensable. The investigated vehicle is equipped with three separated pairs of control surfaces, which enable roll, pitch and yaw maneuvers by deflection. Due to the changes in the overall flow field with increasing angle of attack and sideslip angle (leading-edge vortices, large scale flow separation), linear theories like the potential theory are not valid for the investigated case. Using high-fidelity numerical tools enables a realistic computation of the flow field and a representation of non-linear effects. In addition to the controllability by means of control surfaces, the dynamic stability in consequence of unsteady free stream conditions is investigated.

Investigations of flow control on a 65° sweptback half delta-wing are comprised in the second subproject. The flow around delta wings is dominated by two large counter rotating leading-edge vortices that provide additional lift in comparison to conventional wings. At very high angles of attack, the vortices break-up or even disappear leaving a dead-water region above the wing, thus limiting the flight envelope. By using active flow control at the leading edge through pulsed blowing or oscillating flaps the aerodynamic performances, especially the maneuverability, controllability and stability of delta-wing configurations can be enhanced.

Furthermore, in the context of a DFG project (Strömungs-Struktur-Eigenschaften von flexiblen Tragflächen für Windrotoren, DFG-BR1511-8), flexible wing structures for wind turbine blades are investigated. The blade structure concept is flexible and adapts its geometry with respect to the local free stream conditions. The lift polar can be shifted towards higher maximum angles of attack and higher maximum lift coefficients due to the adaptivity of the structure. This way, the aerodynamic of such a concept is significantly improved over a wide range of angles of attack.

Results and Methods

The control surface efficiency and dynamic characteristics of a low aspect-ratio diamond-wing configuration are investigated by means of steady state and time-accurate simulations. Control surfaces responsible for roll, pitch and yaw are considered in experimental and numerical analyses. The dynamic characteristics are exploited by means of harmonic rigid body motions. Both, the control surfaces and dynamic characteristics are numerically investigated by solving the compressible (U)RANS equations using the SA-turbulence model with the DLR TAU-Code. Due to the complex flow field (vortex formation, large scale flow separations), Figure 1, well resolved hybrid grids (50-90 M elements) and a large number of iterations (steady: 40 000, unsteady harmonic motion: up to 200 000) per simulation are necessary. The boundary layer is resolved by a prismatic grid with a first cell height ensuring y+<1. Depending on the configuration, the simulations have been run on SuperMUC with 544 up to 1204 cores. [2, 3]

The investigations of the active vortex-flow manipulation at very high angles of attack comprise wind tunnel testing and complementary numerical simulations. The latter investigations are conducted with the commercial flow solver ANSYS-CFX by computing the incompressible Navier-Stokes equations on a discretized computational domain (40.6 M cells) through the finite volume method. The Unsteady Reynolds Average Navier-Stokes (URANS) approach with the Shear Stress Transport (SST) turbulence model is used. To enhance the spectral content, the Scale Adaptive Simulation scheme and a hybrid RANS/LES approach are employed. The latter method combines the accuracy of LES and the efficiency of RANS. Advection and transient schemes are second order accurate. Convergence is assured by 13 inner coefficient loops per time-step whose size is correlated to the spatial refinement. Figure 2 shows the flow field above the delta wing with and without pulsed leading edge blowing in the post-stall flight regime.

The simulations predict well the flow reattachment on the wing’s upper surface [4]. The problem was computed by several simulation runs on 196 cores on SuperMUC. Each run computed about 1500 time steps resulting in 9222 used CPUh. In total, 10 such runs have been calculated for different cases and turbulence models.

The investigation of the flow around a morphing wing includes experiments and numerical investigations. The morphing wing is made of a rigid inner structure as leading- and trailing edge and a membrane wrapped around the two spars. This membrane has the capacity to adapt itself to the incoming flow by equilibrating the pressure around its upper and lower sides. This capacity results in a deflection of the membrane (camber/geometry change), which leads to a higher lifting capacity and a delay in the stall phenomenon, which appears additionally smoother. The configuration was numerically investigated by the means of Fluid Structure Interaction simulations using two different FSI couplings: on the one side the coupling CFX-ANSYS/APDL-ANSYS and on the other side the coupling TAU/CARAT++. In both cases, the coupling was investigated for a quasi 2D case, which corresponds to an airfoil extruded in the third direction, with the extrusion length equal to 1% of the chord length.

Concerning the experiments, a quasi 2D model was investigated with end-plates in order to avoid the tip vortices. The polar Cl-alpha is plotted in Figure 3 and shows the characteristics of the elasto-flexible membrane airfoil/wing and for its rigid counterpart geometry [5]. The lift coefficient of the elasto-flexible membrane airfoil is higher or similar to its rigid case in the linear domain of the polar. The deflection leads to a higher maximal lift coefficient and later to a smoother and delayed stall phenomenon. Nevertheless, the results of the experiments are different with the FSI as the maximal lift coefficient is achieved for 17° whereas it appears around 12°-14° in the simulations. This supposes that the 3D effects are too significant during the experiments, which could be avoided by simulating a 3D model.

The next FSI simulations, which are currently in preparation, will be the next step of the project.