Computational Fluid Dynamics 2010

Computational Fluid Dynamics 2010

Alexander Kuzmin Editor

Computational Fluid Dynamics 2010 Proceedings of the Sixth International Conference on Computational Fluid Dynamics, ICCFD6, St Petersburg, Russia, on July 12-16, 2010

With 516 Figures

123

Editor Alexander Kuzmin St Petersburg State University Laboratory of Aerodynamics St Petersburg Russia [email protected]

ISBN 978-3-642-17883-2 e-ISBN 978-3-642-17884-9 DOI 10.1007/978-3-642-17884-9 Springer Heidelberg Dordrecht London New York Library of Congress Control Number: 2011923661 c Springer-Verlag Berlin Heidelberg 2011 This work is subject to copyright. All rights are reserved, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilm or in any other way, and storage in data banks. Duplication of this publication or parts thereof is permitted only under the provisions of the German Copyright Law of September 9, 1965, in its current version, and permission for use must always be obtained from Springer. Violations are liable to prosecution under the German Copyright Law. The use of general descriptive names, registered names, trademarks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. Cover design: eStudio Calamar S.L., Heidelberg Printed on acid-free paper Springer is part of Springer Science+Business Media (www.springer.com)

Fifteen Years of Boeing–Russia Collaboration in CFD and Turbulence Modeling/Simulation Sergey V. Kravchenko, Philippe R. Spalart, and Mikhail Kh. Strelets

Abstract The paper highlights major outcomes of a more than 15-year tight collaboration between scientists and engineers of Boeing Commercial Airplanes (Seattle) and a group of scientists from St.-Petersburg under the aegis of the Moscow Boeing Technical Research Center. The collaborative research covers a wide spectrum of basic and applied aerodynamic problems and has resulted in numerous prominent achievements in Computational Fluid Dynamics, turbulence modeling and simulation, and Computational Aero-Acoustics (CAA). The paper focuses on the turbulence modeling and simulation aspects of the mutual work (the CAA studies are presented in a separate paper of M. Shur et al. in this volume).

1 Introduction In 1993, soon after the foundation of the Boeing Technical and Research Center in Moscow (BTRC), the first author organized a visit of a group of Boeing’s leading experts in CFD and turbulence modeling headed by Dr. W.-H Jou to Moscow, St.-Petersburg, and Novosibirsk aimed at letting them get acquainted with a work of Russian scientists in these areas. All the three authors of this paper participated in a series of meetings in Moscow which took place in the course of this visit and well remember their friendly and creative atmosphere. As a result, a number of bilateral, based on common scientific interests, groups have been formed, thus giving a start to a long-term mutually beneficial collaboration which is continuing for more than 15 years now. As far as CFD and turbulence modeling are concerned, initially this group included several scientists from the State Institute of Applied Chemistry (SIACh) and St.-Petersburg State Polytechnic University (SPbSPU) headed by Prof. Strelets and from the Central Institute of Aviation Motors Building in Moscow headed by S.V. Kravchenko (B) The Boeing Company, Chicago, IL 60606-1596, USA e-mail: [email protected]

A. Kuzmin (ed.), Computational Fluid Dynamics 2010, C Springer-Verlag Berlin Heidelberg 2011 DOI 10.1007/978-3-642-17884-9_17,

155

156

S.V. Kravchenko et al.

Prof. Secundov, from one side, and Dr. Spalart and some other scientists and engineers from Boeing Commercial Airplanes (BCA) in Seattle, from the other side. Later on the major part of the work in these areas moved to St.-Petersburg and was carried out in the framework of a long-term Collaboration Agreement between BCA, SIACh and R & D company “New Technologies and Services” (NTS) under general supervision of P. Spalart, the Technical Projects monitor at BCA. Starting from then, a core of the team remains essentially unchanged, which in itself is far from typical for international R & D activity. A major reason of this “stability” and a high productivity of the joint team is that the relationships between Seattle and St.-Petersburg are not customer-supplier-like, which presumes formulating SOW’s from one side and delivering reports from the other side. They are rather joint-team relationships, with a daily communication between the researches and engineers from both sides. This permits an efficient combination of not only scientific and technical expertise but also cultural backgrounds of both groups and makes their mutual work a great intellectual and human adventure, bridging the two continents. An outcome of the 15 years work, which is outlined in 32 volumes of multi-pages reports, 43 papers in the archived US, Russian, and European Journals and in proceedings of 54 International and National Conferences and Workshops, cannot be, of course, presented in one publication with any reasonable level of completeness. So in this paper we just highlight a few achievements reflecting different aspects of the work which includes a basic research in the area of turbulence modeling and simulation and applied studies directed to solution of specific practical problems. However before doing this we briefly outline the NTS code which is an in-house CFD code developed in the course of the project and used in all the studies which results are presented below.

2 NTS CFD Code This code was designed in St.-Petersburg in the middle of the nineties based on the expertise of the CFD group of SIACh. It is a finite volume code which accepts 2D and 3D structured multi-block grids of the Chimera type and computes compressible (at arbitrary Mach number) and incompressible, steady and unsteady flows within different approaches to turbulence treatment (RANS, DES, SAS, LES, and DNS). A range of numerical schemes implemented in the code includes implicit high order (3rd or 5th) upwind and hybrid (5th order upwind/4th order centered) flux difference splitting scheme of Rogers and Kwak for incompressible flows and of Roe, Van Leer, and Weiss and Smith schemes for compressible flows. Numerical implementation of these schemes is performed by implicit relaxation algorithms (Plane/Line Gauss-Seidel relaxation, LU relaxation, and DDADI algorithm), which may be arbitrarily specified by a user in different grid-blocks. During 15 years passed from the first release of the NTS code aimed at computations on

Fifteen Years of Boeing–Russia Collaboration

157

one-processor PC’s, it has undergone significant enhancements including adding massively parallel capability and can now run on contemporary PC’s clusters and mainframe super-computers. The code has passed extensive code-to-code comparisons with other public, in-house industrial, and commercial CFD codes (CFL3D of NASA, CGNS of Boeing, ELAN of the Technical University of Berlin, CFX and FLUENT) and, as of today, is considered as one of the most reliable and efficient CFD codes for aerodynamic applications. Thanks to this, even with a rather restricted CPU power available to the team, especially in the early stages of the work, it has been always competitive with and often superior to other groups possessing much more powerful computers. As an example, Fig. 1 presents flow visualization from DNS of the flat plate turbulent boundary layer at the momentumthickness Reynolds number 666 with wall-mounted Large-Eddy Break Up (LEBU) device carried out on PC with a grid of about 3 million cells [11].

Fig. 1 Fragment of the computational grid and instantaneous vorticity contours from DNS of turbulent boundary layer with LEBU [11] (flow from left to right)

158


3 RANS Turbulence Modeling RANS studies carried out by the team include validation of existing and newly developed/enhanced RANS models and numerous computations of aerodynamic flows of practical interest. A space limit of the paper does not permit to dwell upon these studies in any detail, but two proposals of the team illustrating its activity in this area still deserve mentioning. The first one is a general approach to sensitization of eddy-viscosity turbulence models to rotation and curvature and a Rotation-Curvature correction to Spalart-Allmaras one-equation eddy viscosity RANS model (SARC) based on this approach. Proposed in the work of Spalart and Shur [10] in 1996 it still remains a leader in this field and is superior even to Reynolds Stress Transport models which, in principle, should account for the RC effects “by definition”. The model is simple and robust and has been proven to provide accurate predictions of a wide range of aerodynamic and industrial flows of great practical importance in which other available RANS models turn out to be helpless. An example of SRAC performance in the vortical wake of an airplane at high lift configuration [5] is presented in Fig. 2 demonstrating an excellent agreement of CFD with experimental data. One more example is CFD-based design and optimization of the Vortex Generators for 737 flight-deck noise reduction carried out with the use of the SARC model in [1] (see Fig. 3). The second of the abovementioned proposals is an original approach to modeling of separation of laminar boundary layers from bluff-bodies and transition to turbulence in the separated shear layers in the framework of RANS [2]. Later on it has

Fig. 2 Comparison of computed wake vortex evolution with experiment [5]


159

Fig. 3 Photographs of VG’s installed on airplane and 1/3-octave band sound pressure level measured in flight test at 35,000 feet and Mach number 0.78 at pilot’s outboard ear with and without VG’s

been proven to be compatible with hybrid RANS-LES turbulence models and now serves as an efficient tool for computing sub-critical flows over bluff bodies with the use of such models.

4 Detached-Eddy Simulation of Turbulence This is may be the most prominent achievement of the joint team in the area of turbulence modeling. Proposed in 1997 [9], Detached-Eddy Simulation (DES) turned the “vector of efforts” of turbulence research to hybrid RANS-LES approaches, which made simulations of extremely complex massively separated flow possible even on PC’s and are currently widely used all over the world. A basic idea of DES is using a unique turbulence model which automatically functions as a RANS model in the whole attached boundary layer and as a sub-grid scale LES model in separation regions away from the solid walls, thus seamlessly coupling both approaches in a way permitting to employ their best features, namely, the computational efficiency and accuracy of RANS in attached boundary layers and universality and affordable CPU of LES in the separation regions. A success of the first applications of DES to flows past airfoil at high (beyond stall) angles of attack

160


[3] and to sub- and super-critical flows past circular cylinder [12], resulted in its rapid dissemination (currently DES is implemented in many academic and most of the industrial and commercial CFD codes) and extensive use.1 This, in turn, has permitted to highlight some issues with the original DES approach, analyzed in detail in a lecture given by Dr. Spalart at ICCFD-3 in 2004 [6]. These issues were addressed in further works of the Boeing-NTS team which resulted in a modified version of DES (Delayed DES or DDES [8]) resolving the issue of Grid-Induced Separation caused by a premature switch from RANS to LES mode within attached boundary layer, the original version of DES suffers from in case of the so-called “ambiguous” grids [6]. Then DDES has been further improved by adding the Wall Modeled LES capability (Improved DDES [4]), which has not been originally viewed as a natural DES application area. A detailed analysis of the current status of DES and its modifications and enhancements is given in the review paper of Spalart [7]. Note that only a list of successful simulations carried out with the use of DES, DDES, and IDDES by their authors and other researches and engineers all over the world would occupy several pages. So here we present only two examples of such simulations. The first one is a kind of “exotic”: it is DES of the flow over raised airport runway of the Santa Catarina Airport (Funchal, Madeira Island, Portugal) which photograph in the course of construction is shown in Fig. 4. The runway was built on stilts, 185 m wide and 58 m high on the island-side. This rare configuration, in case of cross-wind from the ocean, was likely to cause separation off the edge of the runway platform, creating lateral wind shear at some height over the runway and reversed flow at its surface, as well as high relative turbulence intensity, and the objective of DES study carried out in the course of the runway building was to quantify these effects. The flow visualization of DES carried out on a modest grid of 750,000 cells reveals separation of the flow off the edge of that runway platform, shear layer roll-up, and formation of essentially 3D roller/rib structures, as is typical for thin bodies at high angles of attack. In order to assess the runway flow conditions

Fig. 4 Photograph of runway and visualization of wind-flow over the runway

1 This has been to a considerable extent promoted by participation of SPbSPU and NTS in the European projects FLOMANIA, DESider, and ATAAC devoted to development and validation of new turbulence modeling and simulation approaches.


161

quantitatively minima and maxima of the x-velocity component were collected from the simulation over time and the spanwise domain. These data have shown that the range of the velocity variations (the strength of the “wind gusts”) over the runway may reach ±100% of the nominal wind velocity. This information has been used by the airport administration for defining maximum wind velocity for safe operation of the runway. The second example is a very recent application of DDES for the tandem cylinders flow which is of both significant academic and practical interest as a prototype for interaction problems commonly encountered in airframe noise configurations, e.g., a landing gear (the simulation was performed with the use of resources of the Leadership Computing Facility at Argonne National Laboratory). Figure 5 shows the flow visualization, which reveals extremely complex fine turbulence structures resolved by the simulation both in the gap between the cylinders and in the rear cylinder wake and reflects the crucial growth of the computer power compared to the 2000 year simulation shown in Fig. 4. Note however that even with the large grid (about 60 million points) and low-dissipative numerics used in the simulation, the agreement with the experiment of some of the flow quantities remains far from perfect, thus suggesting that far not all the issues associated with a reliable prediction of the considered flow are successfully resolved. First of all, this is related to the accurate representation of the evolution of separated shear layers which remains a great physical and computational challenge.

Fig. 5 Visualization of tandem cylinders flow (λ2 isosurface)

162


References 1. Anderson, B., Shur, M., Spalart, P.R., Strelets, M., Travin, A.: AIAA Paper, 2005-0426 (2005) 2. Shur, M., Spalart, P.R., Strelets, M., Travin, A.: In: Proceedings of the 3rd ECCOMAS CFD Conference, pp. 676–682 (1996) 3. Shur, M., Spalart, P.R., Strelets, M., Travin, A.: In: Proceedings of ETMM 4, 669–678, Elsevier, Amsterdam, Lausanne, New York, Shannon, Singapore, Tokyo (1999) 4. Shur, M., Spalart, P.R., Strelets, M., Travin, A.: A hybrid RANS-LES approach with delayedDES and wall-modelled LES capabilities. Int. J. Heat Fluid Flow. 29, 1638–1649 (2008) 5. Slotnik, J., Czech, M., Yadlin, Y.: Application of OVERFLOW to the evolution of aircraft wake vortices. In: Proceedings of the 9th Overset Composite Grid and Solution Technology Symposium, Pennsylvania State University, State College, PA (2008) 6. Spalart, P.R.: Computational Fluid Dynamics 2004, 3–12, Springer-Verlag, Berlin, Heidelberg (2006) 7. Spalart, P.R.: Detached-eddy simulation. Ann. Rev. Fluid Mech. 41, 181–202 (2009) 8. Spalart, P.R., Deck, S., Shur, M.L., Squires, K.D., Strelets, M.Kh., Travin, A.K.: A new version of detached-eddy simulation, resistant to ambiguous grid densities. Theor. Comp. Fluid Dyn. 20, 181–195 (2006) 9. Spalart, P.R., Jou, W.-H., Strelets, M., Allmaras, S.R.: Advances in DNS/LES. In: Proceedings of the First AFOSR International Conference on DNS/LES, Greyden Press, Columbus, OH (1997) 10. Spalart, P.R., Shur, M.: On the sensitization of turbulence models to rotation and curvature. La Recherche Aerospatiale 3, 465–474 (1996) 11. Spalart, P.R., Strelets, M., Travin, A.: Direct numerical simulation of large-eddy-break-up devices in a boundary layer. Int. J. Heat Fluid Flow. 27, 902-910 (2006) 12. Travin, A., Shur, M., Strelets, M., Spalart, P.R.: Detached-eddy simulations past a circular cylinder. FTAC 63, 293–313 (1999)