서브메뉴
검색
Control Related Metrics for Multidisciplinary Design Optimization
Control Related Metrics for Multidisciplinary Design Optimization
상세정보
- 자료유형
- 학위논문 서양
- 최종처리일시
- 20250211152956
- ISBN
- 9798384043249
- DDC
- 629.1
- 서명/저자
- Control Related Metrics for Multidisciplinary Design Optimization
- 발행사항
- [Sl] : University of Michigan, 2024
- 발행사항
- Ann Arbor : ProQuest Dissertations & Theses, 2024
- 형태사항
- 170 p
- 주기사항
- Source: Dissertations Abstracts International, Volume: 86-03, Section: B.
- 주기사항
- Advisor: Cesnik, Carlos Eduardo;Kolmanovsky, Ilya Vladimir.
- 학위논문주기
- Thesis (Ph.D.)--University of Michigan, 2024.
- 초록/해제
- 요약The push for cleaner aviation to meet the net-zero carbon emissions by 2050 goal will require introducing innovative technologies in future aircraft. Active control is a key enabler of these technologies. For example, actively reducing gust and maneuver loads allows the design of lighter and higher-aspect-ratio wings that reduce fuel burn.Traditionally, airframe and controller design are done sequentially. However, due to the couplings between these disciplines, this approach cannot guarantee that the final design is optimal on a system level. Multidisciplinary design optimization (MDO) offers a solution by simultaneously considering all disciplines and their cross-couplings. However, doing a detailed controller design early in the design cycle may still be undesirable due to the additional complexity brought into the MDO process, the fact that many state-of-the-art control design techniques are not suitable for inclusion in gradient-based optimization, and the presence of uncertainties about the final control architecture. Thus, approaches that introduce meaningful control-related constraints for airframe design without controller co-design are of interest.This work develops a method to estimate the performance of the closed-loop system, without controller design, based on parameterizing and designing the sensitivity transfer function of the closed-loop system. This approach efficiently approximates the best achievable closed-loop performance under stochastic disturbances, quantified by a class of risk metrics based on the power spectral density of signals in the system. It also enforces a technological limit on the bandwidth of the closed-loop system via a constraint on the Bode sensitivity integral relation. Analytical formulas for the derivatives of these metrics are derived in direct and adjoint modes and verified against finite difference approximations. These formulas allow fast and accurate derivative calculation, enabling the inclusion of the proposed metrics in gradient-based optimization problems.This approach is demonstrated in the MDO design of the wingbox of an aircraft with a gust load alleviation (GLA) system including a fatigue constraint, and in the analysis of the glideslope tracking performance of a supersonic configuration aircraft on the approach to land phase of flight, considering constraints on the risk of glideslope deviation, stall, and saturation of the elevator.The inclusion of maneuver load alleviation (MLA) in the design is considered through a strategy that is minimally intrusive to a pre-existing MDO problem. This strategy relies on a nested (bi-level) optimization approach and includes a model order reduction process to improve computational performance. It is demonstrated in the aerostructural design of a cantilevered wing, which includes a flutter constraint.The implementation of these methods in aeroelastic and MDO software is discussed, emphasizing the efficient use of algorithmic differentiation (AD). This enables the definition of new metrics without direct interaction with the AD tool, saving programming effort.
- 일반주제명
- Aerospace engineering
- 일반주제명
- Astronomy
- 일반주제명
- Mechanics
- 키워드
- Risk metrics
- 기타저자
- University of Michigan Aerospace Engineering
- 기본자료저록
- Dissertations Abstracts International. 86-03B.
- 전자적 위치 및 접속
- 로그인 후 원문을 볼 수 있습니다.
MARC
008250123s2024 us c eng d■001000017164391
■00520250211152956
■006m o d
■007cr#unu||||||||
■020 ▼a9798384043249
■035 ▼a(MiAaPQ)AAI31631205
■035 ▼a(MiAaPQ)umichrackham005598
■040 ▼aMiAaPQ▼cMiAaPQ
■0820 ▼a629.1
■1001 ▼aBahia Monteiro, Bernardo.
■24510▼aControl Related Metrics for Multidisciplinary Design Optimization
■260 ▼a[Sl]▼bUniversity of Michigan▼c2024
■260 1▼aAnn Arbor▼bProQuest Dissertations & Theses▼c2024
■300 ▼a170 p
■500 ▼aSource: Dissertations Abstracts International, Volume: 86-03, Section: B.
■500 ▼aAdvisor: Cesnik, Carlos Eduardo;Kolmanovsky, Ilya Vladimir.
■5021 ▼aThesis (Ph.D.)--University of Michigan, 2024.
■520 ▼aThe push for cleaner aviation to meet the net-zero carbon emissions by 2050 goal will require introducing innovative technologies in future aircraft. Active control is a key enabler of these technologies. For example, actively reducing gust and maneuver loads allows the design of lighter and higher-aspect-ratio wings that reduce fuel burn.Traditionally, airframe and controller design are done sequentially. However, due to the couplings between these disciplines, this approach cannot guarantee that the final design is optimal on a system level. Multidisciplinary design optimization (MDO) offers a solution by simultaneously considering all disciplines and their cross-couplings. However, doing a detailed controller design early in the design cycle may still be undesirable due to the additional complexity brought into the MDO process, the fact that many state-of-the-art control design techniques are not suitable for inclusion in gradient-based optimization, and the presence of uncertainties about the final control architecture. Thus, approaches that introduce meaningful control-related constraints for airframe design without controller co-design are of interest.This work develops a method to estimate the performance of the closed-loop system, without controller design, based on parameterizing and designing the sensitivity transfer function of the closed-loop system. This approach efficiently approximates the best achievable closed-loop performance under stochastic disturbances, quantified by a class of risk metrics based on the power spectral density of signals in the system. It also enforces a technological limit on the bandwidth of the closed-loop system via a constraint on the Bode sensitivity integral relation. Analytical formulas for the derivatives of these metrics are derived in direct and adjoint modes and verified against finite difference approximations. These formulas allow fast and accurate derivative calculation, enabling the inclusion of the proposed metrics in gradient-based optimization problems.This approach is demonstrated in the MDO design of the wingbox of an aircraft with a gust load alleviation (GLA) system including a fatigue constraint, and in the analysis of the glideslope tracking performance of a supersonic configuration aircraft on the approach to land phase of flight, considering constraints on the risk of glideslope deviation, stall, and saturation of the elevator.The inclusion of maneuver load alleviation (MLA) in the design is considered through a strategy that is minimally intrusive to a pre-existing MDO problem. This strategy relies on a nested (bi-level) optimization approach and includes a model order reduction process to improve computational performance. It is demonstrated in the aerostructural design of a cantilevered wing, which includes a flutter constraint.The implementation of these methods in aeroelastic and MDO software is discussed, emphasizing the efficient use of algorithmic differentiation (AD). This enables the definition of new metrics without direct interaction with the AD tool, saving programming effort.
■590 ▼aSchool code: 0127.
■650 4▼aAerospace engineering
■650 4▼aAstronomy
■650 4▼aMechanics
■653 ▼aMultidisciplinary design optimization
■653 ▼aFundamental limitations of control
■653 ▼aRisk metrics
■653 ▼aController agnostic
■653 ▼aManeuver load alleviation
■653 ▼aGust load alleviation
■690 ▼a0538
■690 ▼a0346
■690 ▼a0606
■71020▼aUniversity of Michigan▼bAerospace Engineering.
■7730 ▼tDissertations Abstracts International▼g86-03B.
■790 ▼a0127
■791 ▼aPh.D.
■792 ▼a2024
■793 ▼aEnglish
■85640▼uhttp://www.riss.kr/pdu/ddodLink.do?id=T17164391▼nKERIS▼z이 자료의 원문은 한국교육학술정보원에서 제공합니다.
