Evolution of Geometric Sensitivity Derivatives from Computer Aided Design Models

Jones, William T.; Lazzara, David S.; Haimes, Robert

doi:10.2514/6.2010-9128

Cited by 7 publications

(3 citation statements)

References 18 publications

(19 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…al. ] 11 while interfacing with its external CAD family of products. All of the above BRep schemes were developed to support meshing and computational modeling.…”

Section: Boundary Representations (Breps)mentioning

confidence: 99%

The Development of a Geometry Engine with Analytic Sensitivities

Bhagat

Blair

Shih

et al. 2012

50th AIAA Aerospace Sciences Meeting Including the New Horizons Forum and Aerospace Exposition

View full text Add to dashboard Cite

The objective of this work is to develop and demonstrate geometry based sensitivity analysis capability that uses gradient-based optimization for conceptual designs. This work uses an object oriented environment with overloaded operators to generate analytic sensitivity trails to support automated gradient-based optimization solvers. It uses an established geometry engine to create and manipulate NURBS-based topologies. The preliminary results of the integration are presented using an aircraft wing with substructures, which demonstrate the ability to interactively synthesize geometric concepts.

show abstract

“…al. ] 11 while interfacing with its external CAD family of products. All of the above BRep schemes were developed to support meshing and computational modeling.…”

Section: Boundary Representations (Breps)mentioning

confidence: 99%

The Development of a Geometry Engine with Analytic Sensitivities

Bhagat

Blair

Shih

et al. 2012

50th AIAA Aerospace Sciences Meeting Including the New Horizons Forum and Aerospace Exposition

View full text Add to dashboard Cite

show abstract

“…Last but not least, proprietary rights protecting CAD kernels make it close to impossible to compute analytical sensitivities through say the adjoint formulation, an extremely limiting factor for gradient-based ASO and MDO. While external interfaces can partly alleviate those issues, [3][4][5] customizing them can be just as difficult as developing a brand new, tailored geometry modeler.…”

Section: Introductionmentioning

confidence: 99%

Geometry Generation of Complex Unconventional Aircraft with Application to High-Fidelity Aerodynamic Shape Optimization

Gagnon

Zingg

2013

21st AIAA Computational Fluid Dynamics Conference

View full text Add to dashboard Cite

This paper describes a geometry generator capable of providing a high-quality model of unconventional aerospace vehicles with fast turnaround. These desirable but contradictory properties are made possible by striking the right balance between user interaction and modeling automation. The former is achieved through a popular Python-based command line interface and the latter through a component-based approach. Final geometries are defined analytically with watertight networks of non-uniform rational B-spline surfaces that are topologically four-sided. Such a tool is expected to leverage the high-fidelity multidisciplinary design analysis and optimization packages that are being increasingly incorporated in earlier stages in the design of novel aircraft. As an example, a drag minimization problem applied to a regional jet geometry generated by the proposed tool concludes the paper.

show abstract

“…Unfortunately, many difficulties related with interfacing a high-end CAD package with external optimizers persist, mainly because the sensitivity of a surface mesh to the geometric design variables cannot be retrieved by usual means such as automatic differentiation [8,9]. This forces costly and potentially inaccurate finite-difference approximations [10], or, if at all possible, the daunting task of identifying and differentiating individual geometric entities implicitly defined by the associativity tree of a master model [11].…”

mentioning

confidence: 99%

Two-Level Free-Form and Axial Deformation for Exploratory Aerodynamic Shape Optimization

Gagnon

Zingg

2015

AIAA Journal

View full text Add to dashboard Cite

An intuitive shape parameterization and control technique suitable for high-fidelity aerodynamic shape optimization is presented. It relies on the principles of free-form and axial deformation, enabling thorough exploration of the design space while keeping the number of design variables manageable. Surface sensitivities to the design variables are readily available; their inclusion in a highly efficient and robust adjoint-based optimization methodology involving linearly elastic volume mesh deformation and a Newton-Krylov solver for the Euler equations is described. The flexibility of the proposed approach is demonstrated through the exploratory shape optimization of a three-pronged feathered winglet, leading to a span efficiency of 1.19 under a height-to-span ratio constraint of 0.1, and an optimization of a regional jet wing at transonic speed where a winglet is allowed to develop starting from a planar wingtip extension, leading to an 18.8% reduction in drag.

show abstract

Evolution of Geometric Sensitivity Derivatives from Computer Aided Design Models

Cited by 7 publications

References 18 publications

The Development of a Geometry Engine with Analytic Sensitivities

The Development of a Geometry Engine with Analytic Sensitivities

Geometry Generation of Complex Unconventional Aircraft with Application to High-Fidelity Aerodynamic Shape Optimization

Two-Level Free-Form and Axial Deformation for Exploratory Aerodynamic Shape Optimization

Contact Info

Product

Resources

About