New Results in Numerical and Experimental Fluid Mechanics V

Present aircraft design methods must be continuously improved. This is due to environmental and future transport market requirements. Promising results are expected from development of advanced high lift systems for civil transport aircraft. At present the validation of the numerical design of a high lift / low speed system is done with subscale aircraft models. The final performance of the new aircraft is derived by extrapolation. This extrapolation to a full-scale aircraft poses a high risk in time and money. The accurate prediction of low speed / high lift performance early in the design process will, therefore, be a key asset in the development of competitive future aircraft. Consequently the Airbus approach for future design verification concentrates on earlier design verification by the use of high-Re facilities, especially the ETW. The experience gained in the qualification and development phase of low-speed testing in ETW via R&T programmes (as EUROLIFT) will be exploited in future in aircraft development work by early use of high-Re benchmark testing in close combination to the design work with advanced CFD-tools.

For the simulation of low Reynolds- and Mach number flows, a boundary layer code in combination with a database method is coupled with an unstructured RANS code. The technique is validated in 2D using a glider airfoil and in 3D for a glider; showing good agreement with experiments and a well-validated 2D design code.

Flow field measurements with a stereoscopic PIV-system were performed behind the DLR-ALVAST half-model in a high lift configuration with and without an ultrahigh bypass ratio engine simulator in the DNW-NWB Low-Speed Wind Tunnel Braunsehweig at Maeh numbers of 0.18 and 0.22 (≈ 62 m/s and ≈ 76 m/s). In the plane behind the engine, the vortex position and strength strongly depended on thrust and angle of attack.

Wind tunnel investigations have been performed to study the effect of different span loadings, obtained by means of differential flap settings, on the structure of the wake vortex behind a half-model of a four-engine large transport aircraft. Advanced hot-wire anemometry technique has been used to measure the wake vortex velocities at five cross stations up to 4.5 span distances downstream of the model. Results obtained are analysed to identify possible benefits in terms of alleviation of the wake vortex hazard.

This work presents results obtained from PlY measurements done in the wake of a swept and tapered transport-type aircraft wing model. The measurements are condueted in a water tunnel up to 4 spans behind the model. A fin is mounted on the upper side of the wing just in front of the outer flap edge. The influence of this fin on the structure of the vortex wake and the hazard posed to a following aircraft is investigated. The fin alters the structure of the vortices which may lead to an accelerated decay further downstream. In the investigated cases the fin also stimulates the meandering of the vortices, which alters the dynamic effects when a following aircraft enters the wake.

Results of pressure field measurements on the A400m half model in the low speed wind tunnel of Airbus Bremen, Germany, are presented. The model is equipped with two propeller simulators driven by pressurized air. The great challenge was to compensate the influence of the warm and cold air pipes which are guided inside the wing, creating large temperature differences on the models surface. The PSP (pressure sensitive paint) measurements were carried out with the mobile DLR PSP system using the DLR02 paint formulation. The influence of propellers with different thrust levels on the pressure distribution was determined suec, essfully. An assessment of the accuracy of the results can be performed by comparison of PSP data with conventional pressure tap data.

The design of a generic wind tunnel model for tailplane stall investigations to setup an experimental database for code validation is presented. The configuration is optimised to obtain large Reynolds numbers at the horizontal tailplane in a wind tunnel of limited size. Fuselage and wing are used to create representative downwash conditions at the tailplane inside the wind tunnel, compared to free flight. The presented strategy doubles the achievable Reynolds number of the tailplane, while simulating the spanwise trends of the downwash in sufficient accuracy. Numerical simulations preliminary to forthcoming wind tunnel investigations, using the unstructured TAU code, show a separation of the boundary layer starting at the trailing edge with high cross flow velocities at the outer tailplane. A deflected elevator shifts flow separation towards lower incidence angles of the tailplane.

Numerical solutions of the Navier Stokes equations for a modem transport configuration with three different wing tip devices are presented. Solutions are obtained at high speed for the cruise configuration and at low speed for the corresponding take-off configuration. This study is performed within the European Union project M-DAW (

M

odelling and

D

esign of

A

dvanced

W

ing Tip Devices). The numerical results are analysed to characterise the performance of the wing tip devices using global force coefficients, local distributions of force and moment, surface contours and skin friction lines. In addition the vortex wake was analysed in the near field tip region for take-off. Numerical results are compared with experimental data obtained in cryogenic wind tunnel tests at high speed in the ETW (European Transonic Wind tunnel) and at low speed in the DNW-KKK (Cryogenic Wind Tunnel Cologne).

An approach to optimise the aerodynamic shape of a flying wing transport aircraft is presented. It is split into 2D airfoil design and 3D twist and chord length opfimisation. The applied finite volume flow solver is the DLR code FLOWer. The methods are described and results for a flying wing configuration are presented. The paper represents parts of DLR’s contribution to the aerodynamics work package of the VELA project, supported by European Commission.

The design, construction, manufacturing and geometric properties of a generic 65° Delta Wing for the International Vortex Flow Experiment 2 are described. Briefly, the measurement equipment implemented in the model and the external sensors are specified. A wind tunnel investigation has been performed and the results for two different leading edge geometries (sharp and rounded) at three angles of attack are discussed in detail. Further, the results are compared with the ones obtained by NASA and also with computational results. Partly developed (α = 13°) and fully developed (α = 18°) leading edge vortices as well as vortex breakdown (α = 23°) are considered.

A selection of steady and unsteady validation results for the Reynolds-averaged Navier-Stokes solver FLM-NS with respect to an experimental delta wing test case is presented and compared to other numerical results. Having put the numerical method’s validity into evidence, the unsteady flow induced by an oscillating flap is investigated for a Fighter Type Delta Wing.

The turbulent flowfield above the wing of a delta—canard—configuration at moderate (α = 15°) and high (α = 24°) angle of attack was measured at a Re—number of 0.97.106 in a wind tunnel by hotwire anemometry. Leading edge flap settings of η

l.e.

= 0° and η

l.e.

= −20° were used. At moderate angle of attack and deflected leading edge flap a strong vortex originates from the side edge of the non-deflected inboard wing leading edge part. This inboard wing vortex is located close to the fuselage. It is a dominant flow feature and forms the center of the vortical flow separating from the wing surface. The separation line is clearly different from the leading edge flap hinge line. At high angle of attack the flow separates at the leading edge for both the non—deflected and deflected leading edge ease. The resulting leading edge vortex is subject to breakdown dose to the apex. In case of the deflected leading edge, the interaction of inboard wing vortex and leading edge vortex remits in decreased downstream expansion of the burst vortex, also reducing turbulence intensity levels.

An overview about recent results of the DLR-Project SikMa-“Simulation of Complex Maneuvers” is presented. The objective of the SikMa-Project is to develop a numerical tool to simulate the unsteady aerodynamics of a free flying aeroelastic combat aircraft, by use of coupled aerodynamic, flight-mechanic and aeroelastic computations. To achieve this objective, the unstructured, time accurate flow-solver TAU is coupled with a computational module solving the flight-mechanic equations of motion and a structural mechanics code determining the structural deformations. By use of an overlapping grid technique (chimera), simulations of a complex configuration with movable control-surfaces are possible.

Experimental simulation was conducted in the Transonic Wind-Tunnel GSttingen (

TWG

) of

DNW

with the 65 degrees cropped delta wing

AEROSUM

model, which was managed in the DLR project SikMa (Simulation of complex maneuvers). Besides unsteady force and surface pressure measurements an optical deformation measurement technique was applied. In order to a better understanding of the instability behavior of free-to-roll rolling motions of a delta wing flow a high-time resolution optical technique was applied to get hints i.e. to coupling effects of aerodynamic flow with elastic wing model which influence the stability of the delta wing flow. Therefore defined periodic rolling motions were investigated. Discontinuous jumps of normal deformation amplitudes of the rear flaps are recognized while the harmonic rolling motion of the wing body. Although technical or frictional effects cannot be excluded, it seems possible to explain this behavior by aerodynamic-elastic coupling effects of the delta wing flow with the elastic wing-flap structure.

Supersonic commercial transport aircraft (SCT) are known for their short travel times and also for their low-speed noise.

This paper focuses on the aerodynamic design of low-drag SCT high-lift systems contributing significant low-speed noise reductions by enabling steeper climbing. The high-lift system design - relying on separated flow- and validation work at DLR will be reported.

The EC-145 helicopter of Eurocopter Deutschland has been numerically investigated. How a Chimera grid system for the fuselage and both rotors, modelled as actuator discs, can be assembled is commented on. The analysis of the flow solution shows the wakes of both discs by means of sheets of total pressure gains and how they develop. Also the highly vortical flow pattern behind the cell is evidenced using stream ribbons. Average values for the lift and drag coefficients could be obtained from the steady-state computation, which converged only down to a certain level since the flow proves to be inherently highly unsteady.

A method has been developed how to integrate numerical fluxes of velocity fields of rigid body motion exactly in the context of an edge based data structure to ensure free stream consistency for rotating flows. For the unstructured DLR TAU code this extends the field of applications to rotor and helicopter aerodynamics. This is demonstrated with a calculation of the 4-bladed rotor HELI7A in hover and a comparison with results from the structured DLR FLOWER code and experimental data.

Unsteady CFD simulations have been conducted for generic isolated- and installed-propeller configurations at low-speed flight conditions. The propeller is a four-bladed design typical for modem regional turboprop airerafl. The computations were performed with the DLR TAU-eode and the numerical results are compared with experimental data. The results of the Euler and Navier-Stokes computations agree well with the wind tunnel wake data and surface pressure distributions. Additionally, an analysis of the forces acting on the wing and the propeller is performed.

The paper describes the setup and validation of a process-chain based on the DLRTAU code with adaptation and flow solver for the determination of the jet-wake behind jet engines for viscous flows is shown. This jet-wake (often called jet-blast) plays an important role for the design and the operation of airports. As an example of the process-chain the jet-blast impact on a hangar by an AIRBUS A380 configuration is simulated using this technique.

The paper present a numerical/experimental investigation of the flow field resulting around a generic vehiole controlled by means of a lateral jet in supersonic motion. The experimental data are acquired in the DLR wind tunnel TMK in Cologne. To obtain the numerical results, the 3-D viscous, turbulent, Reynolds-averaged hybrid Navier-Stokes Code TAU of DLR is used. Caloulations are made for free stream Maeh number of M∞=2.8 and various angles of attack. Several grids of varied density and structure and different Inrbulenee models are investigated. The CFD results in terms of surface pressure, normal force coefficients, and jet interaction are in good agreement with the results from wind-tunnel tests. On hand the numerical results it is shown important features of the interaction mechanisms between the lateral jet and the flow field originated by the motion of the vehicle.

Steady and unsteady interactions of supersonic overexpanded free jets with plates are visualized by standard shadowgraphy and modeled numerically using the DLR Tan Code. Unsteady behaviors of jet-plate interactions are studied by the method of multi-exposure photography that has been combined with a synchronized pressure measurement on the plate. It is shown that there exists a distinct correlation between shock motions and pressure fluctuations on the plate.

In this basic study the interaction of a lateral jet with supersonic external flow was examined. The study consists of wind tunnel experiments at Mach 5 in the Ludwieg Tube Facility of the German Aerospace Center (DLR) as well as of numerical simulations with the DLR-TAU-code. The efficiency of two flow control devices (upstream mounted needle and weak jet) for stabilization and control of flow separation by means of bow shock attenuation was investigated. The results show that the investigated flow control devices are able to influence the shape and size of the main separation zone but are not capable to suppress it completely. A comparison of wall pressures and limiting streamline patterns showed reasonable agreement between numerical simulations and experimental data.

This paper presents analytical methods for calculating the heating of a hypersonic projectile equipped with a conical forebody. Calculated heat fluxes are compared with data obtained from ISL shock tube experiments. Two theoretical approaches based on the classical boundary layer theory are described for the laminar as well as for the turbulent boundary layer formation. In both cases, a coordinate transformation is applied that enables flat plate solutions to be adapted to conical coordinates. The heat flux on the cone’s surface is measured with special thin film gauges. The heat flux measurements performed at a flight altitude of 15 km show a laminar to turbulent boundary layer transition. In contrast to this, the boundary layer at a flight altitude of 21 km develops fully laminar.

The hybrid DLR RANS solver TAU coupled to a transition prediction module was successfully applied to a single-element airfoil automatically taking into account the locations of laminar-turbulent transition. The experimentally measured transition locations could be reproduced with very high accuracy. A sensitivity study of the parameters of the coupling procedure was performed in order to investigate the behaviour of the coupled system with respect to the accuracy and robustness of the iteration procedure for the transition locations. The transition prediction coupling structure and the underlying algorithm are described. The functions of the coupling parameters and their impact on the transition location iteration and the convergence of the simulations are described and documented.

This investigation shows that the properties of the two-dimensional flow around a high-lift multi-element airfoil obtained by solving the Reynolds-averaged Navier-Stokes equations can be used for the prediction of attachment-line transition (ALT) by the criterion of Pfenninger. Flow calculations are performed for three spanwise sections of a three-dimensional swept and tapered high-lift wing for which the occurrence of ALT is assumed at higher Reynolds numbers. It is shown that the onset of ALT is predicted reliably for changes of the angle of attack. The method is also applicable to indicate the spanwise location where ALT occurs first.

Measurements and simulations are presented of the flow past a tailplane research airfoil which is designed to show a mixed leading-edge trailing-edge stall behaviour. The numerical simulations were carded out with two flow solvers that introduce transition prediction based on linear stability theory to RANS simulations for cases involving laminar separation bubbles. One of the methods computes transition locations across laminar separation bubbles whereas the other assumes transition onset where laminar separations occur. For validation of the numerical methods an extensive measurement campaign has been carded out. It is shown, that the methodology mentioned first can simulate the size of laminar separation bubbles for angles of attack up to where the separation bubble and the turbulent separation at the trailing edge are well behaved and steady in the mean. With trailing edge separation involved, the success of the new numerical procedure relies on the diligent choice of a turbulence model. Finally, for flows with increased unsteady behaviour of both, separation bubble and turbulent separation, which were observed at higher angles of attack in the experiment between maximum lift and leading-edge stall, steady state prediction methods for transition can no longer be applied and time-accurate methods have to be developed in a further step.

Flow separation and transition to turbulence in a smooth axisymmetric diffuser at

Re

D

1, ≈ 7800 were investigated both numerically and experimentally.

The inlet flow is an incompletely developed laminar pipe flow with a typical boundary layer thickness (δ

99

/

D

1

≈ 0.3). The smooth diffuser contour causes a pressure-induced laminar separation. Due to the inflection point within the shear layer, instabilities cause a transition of the separated laminar flow. Further downstream, the flow reattaches turbulent and recovers slowly to a turbulent equilibrium boundary layer. Periodic disturbances are introduced upstream of the separation point in order to control the breakdown of the separated shear layer. In the present study, two different perturbation modes are tested experimentally and compared in detail with numerical investigations.

Measurements of the temporal and spatial development of natural and controlled boundary layer instabilities in the linear and weakly nonlinear stage of transition were carded out using a laminar wing glove in-flight and in a wind tunnel. An 74 piezo-sensor-array and a spanwise hot-wire array together with several independent point sources for the controlled experiments were used. Depending on the excitation case, typical structures which characterize fundamental and the oblique breakdown were observed. The results of natural transition show two-dimensional Tollmien- Schlichting waves in the linear stage as well as three-dimensional wave packets in the beginning of nonlinear stage, and were observed during flight and in the wind tunnel. Comparing in-flight and wind tunnel measurements, the disturbance structures are qualitatively similar, but the amplification of 3D instabilities occurs later and more rapidly in the wind tunnel.

A laminar separation bubble is formed in a region of adverse pressure gradient on a flat plate by a separating boundary layer that undergoes transition, finally leading to a reattached turbulent boundary layer. Linear amplification of steady three-dimensional disturbances in the flow before separation and in the laminar part of such a separation bubble is studied by means of direct numerical simulation and an adjoint-based optimization technique suited to study spatial optimal transient growth. The steady disturbances develop as streaks following their excitation in the region of favorable pressure gradient. At separation and inside the bubble, numerical and experimental results show good agreement with theoretical predictions for the optimal disturbance. The growth rate of the steady distutrbance is seen to possess a maximum around the spanwise wave length that was found to be a preferred one in the corresponding experiment.

A family of laminar separation bubbles (LSB) in a swept boundary layer flow - hereafter referred to as “swept LSB”- is used to study the effect of sweep and of the propagation direction of disturbance waves on the quality of linear stability theory (LST) and solutions of the parabolized stability equations (PSE). To this end spatial LST and linear PSE solutions are qualitatively and quantitatively compared to highly resolved results of direct numerical simulations (DNS). The sweep angle of the base flow is systematieally varied between 0° and 45° and a variety of Tollmien-Schlichting waves as well as the most amplified stationary cross-flow mode are investigated. It turns out, that even though LST works satisfactory in the presence of sweep, flow separation and back flow, PSE is clearly preferable in terms of accuracy.

The experimental investigations described in the present paper deal with the reduction of the total aerodynamic drag of a generic car model (Ahmed-Body) by means of periodic forcing. The experiments carried out in this study focus on a unique approach to separation control using fundamental frequencies for local forcing of the shear layer separated from the rear end of the car model. The excitation of large scale vortex structures by periodic forcing intensifies the primary momentum transfer between the separation region and the outer flow, resulting in a substantial reduction of the separation length. A total drag reduction of 27% was achieved using the flow control method described in this study.

This paper demonstrates the use of active separation control on a high-lift configuration in order to enhance the aerodynamic performance in terms of lift and drag. The aim is to delay boundary layer separation on the flap’s upper surface by periodic excitation using a pulsed wall jet The experimental and numerical results show a massive improvement of almost all aerodynamic coefficients over a wide range of angles of attack and flap deflection angles. By actuating with correct excitation parameters, the jet formed by the single slotted flap can be reattached or kept attached, depending on the conditions, to the surface. Lift is increased by up to 12% while drag is reduced by the same amount. As a result, the lift to drag ratio defining the aerodynamic quality is improved by up to 25%. A numerical calculation on the basis of unsteady Reynolds-averaged Navier Stokes equations is also presented to determine the influence of different excitation parameters.

Pulsed jet actuators are studied in a low-speed wind tunnel by means of phase-locked stereoscopic particle image velocimetry (PIV) to examine the interaction of the jet with a turbulent boundary layer flow along a flat plate. The aim is the transport of high-momentum fluid from the outer part of the boundary layer flow towards the wall to actively delay or avoid flow separation. It is shown that a properly arranged actuator jet produces a strong streamwise vortex, which is well suited to enhance the desired mixing process. The strength and position of this streamwise vortex is of primary importance for the efficiency of the actuator concept. Different jet-exit-hole geometries, their impact on the vortex-structure and their ability to suppress or delay separation are discussed.

Surface smooth plasma actuators were used to control leading-edge flow separation on the flying wing airfoil Eppler E338 for angles of attack of up to 12° past stall at low Reynolds numbers. The plasma actuators were operated over a range of free-stream speeds from 2.2 to 6.6 m/s giving chord Reynolds numbers from 26K to 79K. The plasma actuators produced a 2-D wall jet in the flow direction along the surface of the airfoil and thus added momentum to the boundary layer. Each actuator consisted of two metal electrodes separated by a dielectric layer which was part of the airfoil surface. At all five free-stream speeds from 2.2 to 6.6 m/s, the maximum lift coefficient could be reached and the stall regime relaxed. The power to achieve this was approximately 17 watts per meter or 8.6 watts per actuator over the span width. It was found that the application of low power plasma actuators could simplify the design of mini and micro air vehicles (MAVs) by calculating with the maximum possible lift coefficients obtained from simplified fluiddynamic equations.

Supersonic lateral jets are a convenient method for aerodynamic steering of bodies flying at hypersonic speeds. An accurate prediction of the resulting aerodynamic force is difficult, because the flowfield around such a jet in a hypersonic cross-flow is quite complex. In addition to the jet thrust a considerable pressure load on the body earl be generated by the jet/cross-flow interaction. Several flow and configuration parameters have an effect on the flow interaction pressure force. In this study, the influence of the body’s boundary layer upstream of the side jet is studied experimentally. A shock tunnel is used to generate a Mach-6-flow around a cone-cylinder model with a side jet. By means of flow visualization and unsteady pressure measurements the flowfield around the side jet is characterized for different conditions.

In the present study a numerical coupling tool for fluid-thermal structure interaction is validated in terms of a generic body-flap model with strong radiation effects on the flap leeside. Coupling effects such as flow topology changes on the flap leeside occur. Special emphasis is placed on the numerical rebuilding of the flow condition entering the test section which has a large impact on the coupling effects. The numerical results found are in good agreement with the experimental data gained in the arc heated facility.

In the presented work the influence of the new defined parameter

d = D/δ

on the shock system of a double ramp configuration was investigated, with

D

describing the distance between the two ramp kinks. Oil the basis of the experimental and numerical investigations the functional dependency of the most important parameter of the Free Interaction Concept, the upstremn interaction length

l

0

on the parameter

d

, could clearly be shown. Therefore an enlargement of the free interaction concept is suggested with respect to the double ramp parameter

d

. Moreover, through the investigations of the flat plate/double ramp configuration a critical minimum of the first ramp length was found for the investigated

R/m-ranges

, to realise the wished two shock compression.

Fundamental requirements for future launcher technologies are a cost-efficient access to space as well as the improvement of safety and reliability. In the presented contribution an unsteady turbulent flow simulation of the Ariane-5 launcher during ascent will be carried out by detached eddy simulation (DES). In a first step the nature of separation shocks in turbulent over-expanded axisymmetric nozzles is simulated. Different implementations of detached eddy simulation models are investigated for a compressible wake flow as a validation case. Finally steady and unsteady Ariane-5 simulations are carried out at Mach 0.8 wind tunnel conditions including jet flow and the wake of the boosters.

The main mission critical part of planned reusable re-entry vehicles is its thermal protection system. The surface material has to withstand high heat loads and a chemically aggressive environment in the upper atmosphere. In case of re-usability the candidate material has to withstand these loads several times. Therefore, the mass loss during one re-entry mission has to be as small as possible. In order to predict the mass loss, the material is investigated in plasma wind tunnels and meanwhile numerical simulation of surface processes is possible at Institut für Raumfahrtsysteme (IRS). This paper describes for the first time a quantitative comparison of the specific mass loss estimated on the one hand by plasma wind tunnel experiments and on the other hand by numerical simulation. Results for pressureless sintered silicon carbide (SSiC) and realistic C/C-SiC material are presented. Finally, an attempt is made to interpret the occurring differences.

The electromagnetic railgun technology appears to be an interesting alternative to launch small payloads into Low Earth Orbit (LEO), as this may introduce lower launch costs. A high-end solution, based upon present state of the art technology, has been investigated to derive the technical boundary conditions for the application of such a new system. This paper presents the main concept and the design aspects of such propelled projectiles with special emphasis on flight mechanics, aero-/thermodynamics, materials and propulsion characteristics. Launch angles and trajectory optimisation analyses are carried out by means of 3 degree of freedom simulations (3DOF). The aerodynamic form of the projectile is optimised to provoke minimum drag and low heat loads. The surface temperature distribution for critical zones is calculated with DLR developed Navier-Stokes codes TAU, HOTSOSE, whereas the engineering tool HF3T is used for time dependent calculations of heat loads and temperatures on project surface and inner structures. Furthermore, competing propulsions systems are considered for the rocket engines of both stages. The structural mass is analysed mostly on the basis of carbon fibre reinforced materials as well as classical aerospace metallic materials. Finally, this paper gives a critical overview of the technical feasibility and cost of small rockets for such missions.

This paper presents an overview of numerical simulations of unsteady wing nacelle interference for the generic WIONA configuration. The geometry of the model

Aerodynamic simulations were performed in order to compute the unsteady load response of the NLR7301 airfoil for heave and pitch oscillations in the transonic range. The data were used to determine the aeroelastic flutter boundary of a rectangular two degree of freedom wing. On the one hand the DLR TAU code was used to solve the Reynolds averaged Navier-Stokes (RANS) equations applying the Spalart-Allmaras and κ-ω turbulence models. On the other hand computations were made with the DLR TDLM code, which solves the time linearised transonic small disturbance (TSD) equation. The settings were chosen according to experiments by Dietz et al. [1] in the Transonic Wind Tunnel Göttingen (TWG). The comparison between the computed and the experimental results showed that the simulation codes are capable of computing reliable results. Characteristic transonic aeroelastic phenomena were found. The appropriate modeling of viscous effects in transonic flow came out to be an indispensable condition.

This paper gives a survey of Airbus’ strategy about the low noise high-lift design. The different

airframe noise

sources on a commercial aeroplane are discussed. It will be completed by a physical description of especially those noise sources which typically occur on wings with deployed high-lift devices. Moreover the paper presents the concept to embed the noise criteria basically in the multidisciplinary high-lift design of Airbus. Finally, experimental techniques as well as theoretical tools suitable to describe the process of noise source generation and the identification of low-noise design parameters will be discussed.

To minimize slat noise according to the guidelines of “A Vision for 2020” brush-like devices were installed on a slat trailing edge to “soften” it. In experiments slat noise was considerably reduced by such devices. But their effect on aerodynamics is yet unknown.

This study shows that the investigated brush-like devices on slat trailing edges influence the aerodynamic properties of high-lift systems adversely. Because of increasing boundary layer thickness with length and diameter of the brushes the suction peaks are reduced so that

C

C

decreases.

These are preliminary results. An optimization of slat chord length and brush length should be conducted to minimize

C

L

, reduction.

Within an experimental trailing edge noise reduction study in the Aeroacoustic Wind Tunnel Braunschweig (AWB) both acoustic and aerodynamic effects of trailing edge brush devices were examined. Directional microphone and hot wire measurements were undertaken on a zero-lift generic plate model (Re = 2.1 to 7.9 x 10

6

). Various brush concepts were tested to clarify the functional relationship between design parameters and the ensuing aeroacoustic properties. First results of this ongoing work indicate a significant source noise reduction in excess of 10 dB, thereby revealing two relevant noise reduction mechanisms. In addition to broadband turbulent boundary layer trailing edge noise also narrow band contributions due to vortex shedding from the edge were alleviated.

The noise generation of turbulent flows near the edges of airplanes and automobiles is a general design problem and its importance increases in times of growing traffic. Turbulent boundary layers being convected past the trailing edge into the wake are known to generate an intense, broadband scattering noise. In this feasibility study the high-speed PIV technique was applied to a generic trailing edge noise experiment as performed on a flat plate model in the Aeroacoustic Wind Tunnel Braunschweig (AWB). Trailing edge noise sources have been measured simultaneously with instantaneous velocity vector fields to relate the generated sound to the ensuing aeroacoustic source quantities. The first step towards a new procedure for trailing edge noise prediction, combining numerical methods with the high-speed PIV measurement, is presented.

Trailing edge noise can be predicted with the help of a synthetic turbulent velocity field. Aasuming isotropic conditions this field may be generated via an energy spectrum of turbulence like the modified von Kármán spectrum (MVKS). In this work one-dimensional wavenumber spectra of turbulence obtained from hot-wire measurements at the trailing edges of a thin, flat plate and a NACA0012 airfoil were compared to the respective spectra extracted from the MVKS. Good agreement at all measuring positions is possible with a modified form of the MVKS. The remaining discrepancies can be attributed to the anisotropy of boundary layer turbulence.

Combustion noise of unconfined turbulent flames has been investigated using a hybrid CFD/CAA Method. A large-eddy simulation (LES) of a turbulent non-premixed flame is used to determine the source terms for the computational aeroacoustics (CAA) simulation. The governing CAA equations, namely the Acoustic Perturbation Equations (APE), have been extended to take into account noise generated by reacting flow effects. The right-hand side of the pressure-density relation within the APE system shows that the major source term, the heat release per unit volume, is encoded in the density fluctuation. Therefore the total time derivative of the density is used as source term to simulate combustion noise.

A fully implicit relaxation technique is developed to accelerate convergence of current multigrid solvers. The flux Jacobians are efficiently recomputed during iteration. The implicit operator allows increasing CFL numbers of the basic explicit scheme to O(100), and it properly addresses the stiffness in the discrete equations associated with highly stretched meshes. Compared to a well tuned, standard reference code, computation times are more than halved.

Direct numerical simulation (DNS) of incompressible turbulent pipe flow was carried out on unstructured grids for a Reynolds number based on the friction velocity and the pipe diameter of

Re

τ

= 360 using the spectral/hp element method (SEM) by Karniadakis and Sherwin [3]. The main objective was to investigate the computational aspects of this DNS with respect to accuracy, CPU time and memory requirements. DNS results of evaluated statistical moments of up to fourth order agree well with data from the literature. A conducted performance study reveals the computational requirements for DNS of turbulent flows using this SEM.

The Speziale-Sarkar-Gatski (SSG) Reynolds stress model is implemented into DLR’s Navier-Stokes solver FLOWer blended with the Wilcox stress-ω model in the near wall region. The length scale is supplied by Menter’s ω-equation. Results for 2D flows are presented for the transonic flow around the RAE 2822 airfoil, Cases 9 and 10, and the Aérospatiale A airfoil at β = 13.3°. Results for 3D flows are shown for the transonic flow around the ONERA M6 wing and the DLR-ALVAST wing-body configuration. Improvements are achieved with respect to predictions with the Wilcox κ-ω model concerning shock positions, trailing edge separation and the pressure distribution near the wing tip due to an improved resolution of the wing tip vortex.

Turbulent flows at very low Mach numbers are numerically investigated using largeeddy simulation (LES). The numerical computations are carried out by solving the viscous conservation equations for compressible fluids. An implicit dual time stepping scheme combined with low Mach number preconditioning and a multigrid acceleration technique is developed for LES. The method is validated for turbulent channel flow at

Re

τ

= 590 and flow across a cylinder at Re = 3900 and different low Mach numbers. The data are compared with numerical and experimental findings from the literature. The computations show an efficiency increase by a factor of up to 60.

In the CFD community structured and unstructured codes are in use and under further development. Both types of codes have their specific advantages. Unstructured codes enable a higher geometric complexity in the simulation, whereas structured codes are usually characterized by better performance properties. From developer and user side it would be desirable to have one code being able to handle structured and unstructured meshes as well as grids containing both types of regions. At DLR a first prototype of a mixed structured/unstructured code has been developed, in order to explore the potential of such a combined procedure. The prototype is based on the unstructured DLR TAU code and includes ingredients from the block-structured FLOWer code like implicit smoothing techniques, which are applied in the structured regions of the computational domain. Within this paper the prototype will be described and first promising results are presented for 2D applications.

The numerical analysis of aerodynamic flows is in general limited to steady geometries. Depending on the flow conditions steady or transient flow solutions in the relative frame of the body are computed. In order to take into account the flexibility of the body (e.g. fluttering wing) and the motion of the body (manoeuvre flight), moving computational meshes are required. The CFD method has to take into account meshes with moving nodes and deforming control volumes. The present paper shows computational results of different applications with moving grids, e.g. an oscillating airfoil, a fluttering wing and a guided manoeuvre flight of an airplane.

The simulation quality of numerical flow simulations depends on the choice of physical modelling as well as an appropriate numerical treatment. In this study, a standard two-equation turbulence model has been extended for compressible, rotational flow as it occurs in turbomachinery and subsequently applied to different turbomachinery relevant flows of varying complexity. A number of different numerical schemes has been employed to evaluate their impact on the solution.

The represented investigations are concerned with simulations of unsteady aircraft aerodynamics and thus belongs to the research field of computational fluid dynamics (CFD) for aerospace applications. The calculations are based on the surface singularity panel method VSAERO for simulation of quasi-steady motions and on the solution of the Time-dependent Reynolds-averaged Navier-Stokes (TRANS) equations using the finite volume parallel solution algorithm with an unstructured discretization concept (DLR TAU-code). The analysis consists of two parts, in order to obtain a more detailed understanding of the unsteady aerodynamical and flight mechanical behaviour of an airplane. For this purpose systematic investigations have been performed with basic configurations which have NACA0012 profiles (e.g. wing, wing + horizontal tail, wing + vertical tail). The second part of the investigation is the calculation of the dynamic derivatives of a modem transport aircraft configuration (DLR-F 12). The objective of this investigation is to adapt the numerical methods for the calculations of dynamic derivatives and to validate the numerical results against wind-tunnel data.

In this paper, the results of a Detached-Eddy Simulation (DES) of the burst vortex system over a delta wing using a structured, locally refined mesh is described. For evaluation of the simulation quality, averaged velocities are compared directly with the experiment, and an investigation of the instantaneous structure of the spiral burst of the primary vortex is presented. To assess the modelling characteristics, a comparison with an equivalent RANS technique is conducted, in which the topology of the time-averaged surface flow is discussed in relation to the global structure of the vortex system. The simulation of the single wing presented here constitutes the basis of ongoing investigations, and is thereby intended to represent the minimum simulation quality which could be obtained for a complete configuration.

The objective of this numerical investigation is the evaluation of the small disturbance Naviex-Stokes method FLM-SD.NS for test cases of two-dimensional transonic flow. For this reason the results of a NLR7301 flap-oscillation and a pitching NACA 64A010 airfoil are compared to experimental data. The influence of viscosity is shown by comparison to results of Euler computations while the time effort is judged by comparing with results of unsteady full Navier-Stokes computations.

Turbulent mixed convection in two generic configurations, the geometry of which was deduced from a passenger cabin and a sleeping bunk of a modem long distance passenger aircraft, has been investigated numerically by means of Reynolds averaged Navier-Stokes (RANS) computations. In the same configurations Particle Image Velocimetry (PIV) measurements were conducted to validate the numerical results. The performed comparison indicates that results obtained using RANS with low Reynolds number turbulence models agree considerably better with experimental data than those of RANS with high Reynolds number turbulence models.

This paper is a summary of an ongoing numerical investigation dealing with the influence of modelling parameters on numerical simulation of 2-D unsteady transonic flows. The focus of research lies on a simulation of self-sustained shock oscillation (shock-buffet) since this mechanism can play an important role in the aeroelastic behaviour of modern large-span aircrafts. Three different 2-D profiles have been used in the investigation and the influence of following modelling parameters has been analysed: (a) turbulence modelling, (b) flow solver spatial discretization schemes, and (c) temporal resolution parameters.

Turbulent boundary layers subjected to strong acceleration (favourable pressure gradient - FPG) with reference to the Direct Numerical Simulation (DNS) of Spalart (1986) were examined computationally using a differential near-wall Second-Moment Closure (SMC) model within the RANS (Reynolds-Averaged Navier-Stokes) framework, accounting separately for viscous and kinematic wall blocking. Besides the three accelerating eases treated by DNS, characterized by the acceleration parameter K = 1.5 • 10-6; 2.5 • 10

−6

and 2.75 ∂ 10

−6

, tile laminarizing turbulent boundary layer was also investigated. The value of K = 3.2 - 10

−6

was found to be sufficiently high to cause complete relaminarization of the initially turbulent flow, agreeing well with experimental findings. Integral flow parameters, mean velocity and turbulence quantities are computed in close agreement to available DNS database. The analysis of the anisotropy evolution reveals a continuous tendency of the flow turbulence to reach the one-eomponental isotropic state with an increase of the acceleration intensity.

This paper presents a Direct Numerical Simulation (DNS) of turbulent channel flow with periodic hill constrictions at

Re

= 2808. For the DNS, special attention is paid to the grid design by analysis of the Kolmogorov length scale and the wall shear stress in order to be well resolved over the entire numerical domain. Therefore the DNS simulation provides data for a detailed study of physical flow phenomena and near-wall studies. Results for instantaneous and averaged flow fields are presented. An investigation of the near wall behaviour reveals the applicability of explicit wall models for Large Eddy Simulation (LES) in regions with high wall shear stress and a complete failure of conventional wall scaling in separation and reattachment regions where the wall shear stress is small.

Various methods for unsteady flow computations: LES (Large Eddy Simulation), DES (Detached Eddy Simulation) and URANS (Unsteady Reynolds-Averaged Navier-Stokes) were used to study the effects of boundary-layer forcing on the mean flow and turbulence. Two flow configurations were considered: a periodically perturbed flow over a backward-facing step (with fixed separation point) at a low Reynolds number (

Re

c

= 3700, Yoshioka et al., 2001 [10]) and a high Reynolds number (

Re

c

= 9.36·10

5

) flow over a wall-mounted hump with steady suction (flow separation at the smooth surface), Rumsey et al., 2004 [5]. Whereas LES and DES reproduced all important flow characteristics observed in the experiments, RANS (S-A and k-ω SST models were employed) method exhibited a weaker sensitivity to the shear layer oscillations i.e. boundary-layer suction. It is shown that DES method, representing a hybrid RANS/LES approach, is capable of reproducing the mean flow and turbulence features of a quality comparable with the LES method, but employing significanfly coarser spatial resolution.

Several improved or new techniques for propeller integration testing have been implemented at the German-Dutch Wind Tunnels (DNW). Presented here is an air-return line bridge system along with the applicable correction methodology. The implementation path followed is illustrated by the air-return line bridges systems concept validation and the systems performance validation during wind tunnel operation.