M&S. EDISON. e-science EDISON_CFD[2] EDISON. (geometry kernel software),,, EDISON_CFD.,.. 4. e-science EDISON M&S.., EDISON , b f. Find to min

Size: px

Start display at page:

Download "M&S. EDISON. e-science EDISON_CFD[2] EDISON. (geometry kernel software),,, EDISON_CFD.,.. 4. e-science EDISON M&S.., EDISON , b f. Find to min"

길호 옹
6 years ago
Views:

1 권형일, 조영민, 이슬기, 김사지 1, 최성임 2* DEVELOPMENT OF e-science BASED AERODYNAMIC DESIGN OPIMIZATION FRAMEWORK FOR AN AIRFOIL H.I. Kwon, Y.M. Jo, S.G. Yi, S.J. Kim and S.I. Choi Design optimization is to find optimum of design space which is defined by design variables, using optimization algorithm. Recently, numerical design optimization including computational fluid dynamics (CFD) has been focused in not only aerospace engineering, but general engineering field because of many research and development for high performance computing. However it could hardly find the design optimization softwares and contents of which educational or research purposes in aerospace engineering. In this study, as one framework of EDISON DESIGN OPTIMIZATION, aerodynamic design optimization framework for an airfoil based on EDISON (EDucation-research Integration through Simulation On the Net) portal. As well as the softwares development, essential contents are also developed for the lecture associated with design optimization in the aerospace engineering. Software and contents on aerodynamic design optimization could be expected to be assisted in the lecture or lab (Washington Accord). (ABEEK)..[1] Corresponding author schoi1@kaist.ac.kr...,,.., (high fidelity) M&S (modeling & simulation).

2 M&S. EDISON. e-science EDISON_CFD[2] EDISON. (geometry kernel software),,, EDISON_CFD.,.. 4. e-science EDISON M&S.., EDISON , b f. Find to minimize f (b) Subject to h i = 0 b L b b U g j 0 h g,. ( ).. (finite difference), (complex derivative method), (automatic difference), Adjoint (adjoint variable method). X (1) n n+1 S. (1) (steepest descent), (conjugate gradient method ; Fletcher-Reeves method), (quasi-newton method), (modified feasible directions algorithm), (SLP ; Sequential Linear Programing), (SQP ; Sequential Quadratic Programing).[3][4] n n+1.,., Simplex [5], [6]. EDISON ( ). (2)

3 . (2) f, x (Step Size)... (3). (subtraction),. (3) Adjoint. Algorithmic [7][8].. Adjoint Adjoint Adjoint,. Adjoint., Adjoint ,,,.. 2 (Hessian Matrix) H (4). (4) H (n=1).. D n. p y (5) (6) (7).., (8).. (8) (9),.,., SLP SQP.[3][4][5] S ( ) (1). Maclaurin Series.

4 , (10) i (10) S (11) (11). (12) mf (, 0<m<1), (13). (12) (13) EDISON m 0.1.,,..[3][9] 2.3. (Noise).. Simplex[5]. [6].,.. [4][10]. (design of experiment) ,,.. / 2. EDISON_CFD. Fig.1 EDISON design framework

Fig.2 Data flow chart in EDISON design framework 3.

TFI (trans-finite interpolation) [14]. O-type i = 1 i = imax 3.

5 Fig.2 Data flow chart in EDISON design framework 3.2 (Geometrical Kernel SW) (mapping).,,.., (profile fitting). NACA, PARSEC[11], NURBS[12]. (14). (14). Hicks-Hennen [13].. EDISON PARSEC, NURBS NACA4, Hicks-Hennen TFI (trans-finite interpolation) [14]. O-type i = 1 i = imax 3. 4 undulatory airfoil[15]. Fig. 4. Mesh Deformation Software Fig.3 Hicks-Hennen Bump Function(bump=7) ( ),.,,

, Adjoint.. Adjoint EDISON_CFD. S. EDISON 5,. (1) n n+1.. - - - - - Kuhn-Tucker 3.4 EDISON EDISON_CFD[2]. EDISON_CFD 2 / N-S /Euler. Roe, RoeM, AUSM, AUSMPW+ 2. Euler, Runge-Kutta, LU-SGS.

6 , Adjoint.. Adjoint EDISON_CFD. S. EDISON 5,. (1) n n Kuhn-Tucker 3.4 EDISON EDISON_CFD[2]. EDISON_CFD 2 / N-S /Euler. Roe, RoeM, AUSM, AUSMPW+ 2. Euler, Runge-Kutta, LU-SGS. Standard k-e, Wilcox s k-w, Mentor s k-w SST. 3.5 EDISON.., ( / ),,, (, ).,,. 4. EDISON. 1. Table 1. EDISON airfoil design framework contents list # I II III IV Fig.5 Design Optimization Methodology

7 I.,,. II..[4]. III II..,.,.. I~III IV.. IV.,.,. 6 NACA , [16] Fig.6 Drag minimization design result (left : airfoil shape, right : wall pressure coefficient distribution) 5. e-science EDISON_CFD EDISON.,,... [17],., Adjoint. EDISON GUI ( , ) [1]

8 [2] 2012,,,,,,, e-science EDISON_CFD,, 2011, pp [3] 1997, Jan A. Snyman, Practical Mathematical Optimization : An Introductino to Basic Optimization Theory and Classical and New Gradient-Based Algorithms, Springer. [4] 2010,,,,. [5] 2005, Ashok D. Belegundu and Tirupathi R. Chandrupatla, Optimization Concepts and Applications in Engineering 3th Ed., Pearson Education. [6] 2004, Randy L. Haupt and Sue Ellen Haupt, Practical Genetic Algorithm 2th Ed., Wiley. [7] [8] [9] 1989, G. N. Vanderplaats and S.R. Hansen, DOT User's Manual, VMA Engineering. [10] 1995, Raymond H. Myers and Douglas C. Montgomery, Response Surface Methodology, Wiley inter. science. [11] 1988, Sobieczky, H., "Parametric Airfoils and Wings", Notes on Numerical Fluid Mechanics, Vol. 68, pp [12] 1995, Les Piegl and Wayne Tiller, The NURBS Book, Springer. [13] 1978, Hick, R. M. and Henne, P. A., Wing Design by Numerical Optimization, Journal of Aircraft, Vol.15, No.7, pp [14] 2000, L. Dubuc, F. Cantariti, M. Woodgate, B. Gribben, K.J. Badcock and B.E. Richard, "A grid deformation technique for unsteady flow computations," International Journal for numerical methods in fluid, Vol.32, pp [15] 2007, Chang Shu, Nigyu Liu, Yongtian Chew and Zhiliang Lu, Numerical Simulation of Fish Motion by Using Lattice Boltzmann Immersed Boundary Velocity Correction Model, Journal of Mechanical Science and Technology, Vol.21, pp [16] 2012,,,,, e-science, [17] 2010, Hyung Il Kwon, Ju Yeol You and Oh Joon Kwon, Enhancement of wind turbine aerodynamic performance by a numerical optimization technique, Journal of Mechanical Science and Technology, Vol26. pp

<31325FB1E8B0E6BCBA2E687770>

<31325FB1E8B0E6BCBA2E687770> 88 / 한국전산유체공학회지 제15권, 제1호, pp.88-94, 2010. 3 관내 유동 해석을 위한 웹기반 자바 프로그램 개발 김 경 성, 1 박 종 천 *2 DEVELOPMENT OF WEB-BASED JAVA PROGRAM FOR NUMERICAL ANALYSIS OF PIPE FLOW K.S. Kim 1 and J.C. Park *2 In general,