Department News
Multi-Fidelity and Multi-Disciplinary Design Optimization of Low-Boom Supersonic Jets
Seminar Date
2005-10-11
Author
빈종훈
Date
2005-10-06
Views
1864
1. 제 목 : Multi-Fidelity and Multi-Disciplinary Design Optimization of Low-Boom Supersonic Jets
2. 연 사 : Seongim Choi
3. 일 시 : 2005년 10월 11일 (화) 오전 11:00-12:00
4. 장 소 : 서울대 301동 117호
5. 내 용 :
Supersonic jets have been drawing great attention after the end of service for the Concorde was announced on April of 2003. It is believed, however, that supersonic aircraft may make a viable return in the business jet market. This study focuses on the design optimization of feasible supersonic business jet configurations. Preliminary design techniques for mitigation of ground sonic boom are investigated while ensuring that all relevant disciplinary constraints are satisfied (including aerodynamic performance, propulsion, stability & control and structures.) In order to achieve reasonable confidence in the resulting designs, high-fidelity simulations are required, making the entire design process both expensive and complex. In order to minimize the computational cost, surrogate / approximate models are constructed using a hierarchy of different fidelity analysis tools including PASS, A502/Panair and Euler/NS codes. Direct search methods such as Genetic Algorithms (GAs) and a nonlinear SIMPLEX are employed to find optimum designs in searches of large and noisy design spaces. A local gradient-based search method can be combined with these global search methods for small modifications of candidate optimum designs. The Mesh Adaptive Direct Search (MADS) method can also be used to explore the design space using a solution-adaptive grid refinement approach. These hybrid approaches, both in search methodology and surrogate model construction, are shown to result in viable supersonic jets that demonstrate reductions of sonic boom and improved aerodynamic performance.
6. 약 력 :
2005 ph.D, Aeronautics &Astronautics, Stanford University
1998 M.S., Aerospace Engineering, Seoul National University
1996 B.S., Aerospace Engineering, Seoul National University
7. 문 의 : 기계항공공학부 김 규 홍 (☏ 880-9346)
2. 연 사 : Seongim Choi
3. 일 시 : 2005년 10월 11일 (화) 오전 11:00-12:00
4. 장 소 : 서울대 301동 117호
5. 내 용 :
Supersonic jets have been drawing great attention after the end of service for the Concorde was announced on April of 2003. It is believed, however, that supersonic aircraft may make a viable return in the business jet market. This study focuses on the design optimization of feasible supersonic business jet configurations. Preliminary design techniques for mitigation of ground sonic boom are investigated while ensuring that all relevant disciplinary constraints are satisfied (including aerodynamic performance, propulsion, stability & control and structures.) In order to achieve reasonable confidence in the resulting designs, high-fidelity simulations are required, making the entire design process both expensive and complex. In order to minimize the computational cost, surrogate / approximate models are constructed using a hierarchy of different fidelity analysis tools including PASS, A502/Panair and Euler/NS codes. Direct search methods such as Genetic Algorithms (GAs) and a nonlinear SIMPLEX are employed to find optimum designs in searches of large and noisy design spaces. A local gradient-based search method can be combined with these global search methods for small modifications of candidate optimum designs. The Mesh Adaptive Direct Search (MADS) method can also be used to explore the design space using a solution-adaptive grid refinement approach. These hybrid approaches, both in search methodology and surrogate model construction, are shown to result in viable supersonic jets that demonstrate reductions of sonic boom and improved aerodynamic performance.
6. 약 력 :
2005 ph.D, Aeronautics &Astronautics, Stanford University
1998 M.S., Aerospace Engineering, Seoul National University
1996 B.S., Aerospace Engineering, Seoul National University
7. 문 의 : 기계항공공학부 김 규 홍 (☏ 880-9346)