Catálogo de publicaciones - libros

It is demonstrated that the unstructured, hybrid method of the MEGAFLOW project is capable to compute the engine-airframe installation drag for several engine types although drag differences for varying engine positions can be very small. Results are presented for the DLR-F6 and ALVAST configurations with through-flow nacelles and powered engines. Verification by grid refinement as well as validation with wind tunnel data completes this investigation.

This part of the MEGAFLOW project is concerned with the numerical simulation of the viscous compressible flow around transport aircraft high lift configurations and its validation against wind tunnel experiments. The investigations are based on the solution of the Reynolds-averaged Navier-Stokes equations (RANS) using the MEGAFLOW code system with a finite volume parallel solution algorithm. Both the block-structured (FLOWer) and the unstructured (TAU) code modules are used for this task. Based on the high lift configurations DLR-ALVAST and DLR-F11 lift polars are computed with both codes and compared against each other and against measurements. Further on the structured Chimera technique implemented in FLOWer is applied to a high lift configuration.

Detailed numerical shape optimization will play a strategic role for future aircraft design. It offers the possibility of designing or improving aircraft components with respect to a pre-specified figure of merit subject to geometrical and physical constraints. However, the extremely high computational expense of straightforward methodologies currently in use prohibits the application of numerical optimization for industry relevant problems. Optimization methods based on the calculation of the derivatives of the cost function with respect to the design variables suffer from the high computational costs if many design variables are used. However, these gradients can be efficiently obtained by solution of the continuous adjoint flow equations.

The present paper aims at describing the potential of the adjoint technique for aerodynamic shape optimization. After a brief description of the aerodynamic optimization process developed at DLR, specific requirements for an optimization framework combined with the adjoint technique are introduced. The drag reduction by constant lift and pitching moment for the RAE 2822 airfoil in transonic flow is then presented as validation case. An extension to multi-point optimization demonstrates the capability of the methodology to solve more complex problems. At the end, the body optimization and the wing optimization of a supersonic commercial aircraft confirm the flexibility of the framework and the efficiency of the adjoint technique.

Shape parametrization has been identified as an import issue in aerodynamic design optimisation based on high-fidelity CFD-methods. For given shapes, which are available as CAD-models, post-parameterization method, based on freeform deformation, has been established to simplify and to automate the generation of geometrical variants to be used for CFD analyses. To create the necessary deformation lattices, structured grid generation techniques of a grid generation system, developed at DLR, are utilized. As this grid generation system has the salient feature to store and to replay a sequence of processes with different parameter settings, modifications of shapes, given by polygonal curves and surfaces, can be performed instantly.

The present freeform deformation method has reached a state, where it can be integrated into design loops to handle a variety of shape optimisation tasks. In two examples the applicability of the method for aerodynamic wing design and detailed design of a wing tip is demonstrated.

In the aerodynamic industrial design process, the use of numerical simulation is of ever increasing importance. In order to adequately capture flow features such as pressure-induced separation or shock-boundary-layer interaction, an appropriate representation of turbulence is needed. This contribution summarizes the efforts undertaken at TU Berlin to develop, implement and validate advanced linear and non-linear models in the aerodynamic flow solvers FLOWer and TAU in the framework of MEGAFLOW and related projects. The accuracy of the approaches is discussed on various cases and statements with respect to their computational performance are given. The results indicate that improved predictive accuracy can be obtained from advanced Eddy-Viscosity Models at a moderate computational surplus.

A Large-eddy simulation version of the FLOWer code is introduced to compute attached flow around a quasi three-dimensional airfoil at a Mach number = 0.088, Reynolds number = 8 × 10, and an angle of attack of 3.3°. For the treatment of the subgrid-scale stresses the MILES approach is chosen. The visualization of instantaneous flow fields shows the typical flow features such as the streaky structures in the near-wall region to be well resolved and the overall agreement of the computational results with experimental data to be satisfactory.

The 3D Navier-Stokes solver TAU is coupled with linear stability analysis methods in order to predict flows including transition due to Tollmien-Schlichting (TS) and crossflow (CF) instabilities. The new simulation capability is investigated for an airfoil and compared with data of 2D boundary layer methods that include transition prediction based on a well-known envelope method and with experiments. The results indicate the levels of grid and residual convergence needed for accurate transition prediction. First applications of transition prediction in 3D for a 1:6 prolate spheroid are discussed. It is shown that transition calculations for fully 3D flows are numerical feasible and yield physically reasonable results for moderate angles of attack.

The present contribution outlines several selected applications of the MEGAFLOW software at DLR, roughly according to the time-frame of the MEGAFLOW II project from 1998 – 2002. The majority of the applications is based on 3-dimensional viscous computations featuring the solution of the compressible Reynolds-averaged Navier-Stokes equation in combination with one or two transport equation turbulence models. The examples highlight a quite broad range of applications. This refers to onflow speed as well as to the mode of application, covering analysis as well as design and optimization tasks. In general a clear trend in the use of the MEGAFLOW system becomes visible. On the one hand the analyses of very specific flow and/or aircraft details on overall configurations of increasing complexity is carried out. On the other hand more design and optimization applications are requested. For both types of applications the MEGAFLOW software has become an indispensable tool for the aerodynamic and also multi-disciplinary tasks of DLR.

The MEGAFLOW projects — I and II — are key drivers for CFD (Computational Fluid Dynamics) technology research within Germany, possibly in Europe as a whole. MEGAFLOW combines the development of most modern CFD tools and suites with the daily application of these tools for aircraft design and data processes. Due to aircraft industries being forced to optimize their designs and to give performance predictions in highest accuracy terms, while the time for design cycles are reduced, high level CFD becomes more and more a basic design tool. However, in many cases CFD reaches its limits due to the ongoing increase of complexity with respect to geometry and flow conditions. Further development of physical flow modeling and further validation of new application cases need to be continued. This paper describes AIRBUS-D requirements on methods and tools development in MEGAFLOW, validated by AIRBUS-D test cases. It outlines which requirements are fulfilled, which are still open, and which have had to be added over the years.