 | Li Ch. A surrogate-based framework with hybrid refinement strategies for aerodynamic shape optimization: Diss. … Dr.-Ing. / Deutsches Zentrum für Luft- und Raumfahrt, Institut für Aerodynamik und Strömungstechnik, Braunschweig. - Köln: DLR, Bibliotheks- und Informationswesen, 2013. - vi, 100 p.: ill., graph., tabl. - (Forschungsbericht; 2013-16). - Ref.: p.95-99. - ISSN 1434-8454
|
Nomenclature ................................................... II
1 Introduction ................................................. 1
1.1 Problem description ..................................... 2
1.2 Literature survey ....................................... 4
1.3 Thesis motivation ....................................... 4
1.4 Thesis overview ......................................... 4
2 Aerodynamic shape optimization ............................... 7
2.1 Model construction of optimization problems ............. 7
2.2 Aerodynamic shape parameterization ...................... 8
2.3 Mesh generation & flow simulation ...................... 11
2.4 Optimization algorithms ................................ 13
2.5 Examples for optimizing a deceptive function for a
4-digit NACA profile ................................... 17
3 Surrogate-based optimization ................................ 21
3.1 Overview ............................................... 21
3.2 Design of experiment ................................... 22
3.3 Surrogate models ....................................... 25
3.4 Refinement strategies .................................. 27
3.5 Imputation of missing data ............................. 30
3.6 Convergence criteria ................................... 31
3.7 Examples for optimizing a deceptive function for a
4-digit NACA profile ................................... 32
4 Hybrid strategies for global optimization based on
surrogate model ............................................. 37
4.1 Hybrid refinement strategies ............................ 37
4.2 Enhanced hybrid sample refinement strategies ........... 42
4.3 Refinement for problems in multiple-dimensional
design space ........................................... 47
4.4 Experience in setting refinement process ............... 51
5 Framework ................................................... 53
5.1 Introduction ........................................... 53
5.2 Structure of the framework ............................. 53
5.3 Application of the framework ........................... 54
6 Test cases .................................................. 61
6.1 Minimization of the drag for a NLF0416 profile with
10 design variables .................................... 61
6.2 Minimization of the drag for.a RAE2822 profile with
many design variables .................................. 65
6.3 Optimization of the take-off performance for a high-
lift profile with constraints .......................... 69
6.4 Optimization design of a laminar profile at two
design points .......................................... 79
7 Concluding remarks .......................................... 91
7.1 Achievements ........................................... 91
7.2 Conclusions ............................................ 92
7.3 Outlook ................................................ 93
References ..................................................... 95
|
|
