Preliminary results from the application of automated adjoint code generation to CFL3D

Carle, Alan; Fagan, Mike; Green, Lawrence L.

doi:10.2514/6.1998-4807

Cited by 38 publications

(23 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…6 this difficulty was addressed by development of the "iterated reverse-mode" scheme, where only the log files for the final forward-pass iteration are stored and used during the subsequent iterative solution for the derivatives.…”

Section: Adh-av Method_ Model Problemmentioning

confidence: 99%

Some advanced concepts in discrete aerodynamic sensitivity analysis

Taylor

Green

Newman

et al. 2001

15th AIAA Computational Fluid Dynamics Conference

Self Cite

View full text Add to dashboard Cite

Introductionhas been proven an effective tool for extracting aerody-namlc SDs from these modern CFD codes" . The present study will essentially build on earlier studies" m an effort to exploit the full potential of the latest version of ADIFOR 3107 fol_ obtaining SDs from CFD codes. and Eq. (3) One objective of this particular development of the AV method for aerodynamic SDs is to ensure that the relationship given by Eq. (8) is clearly understood. The I-I method for solving the forward-mode, FO SD Eq. (4) iswhere superscript M is the FO SD iteration index, andWith the I-I methodology, the CFD flow solution operator Pint is also used to solve the SD equations; this operator in Eq. (12) is evaluated and fixed using the steady-state solution for the nonlinear flow. The requisite terms of Eq. (13) are constructed either by hand differentiation (i.e., the HDII method, which is very tedious and time consuming to complete with accuracy for advanced CFD codes) or by AD, which is the forward-mode ADII method of previous studies.In contrast with the ADII method, a straightforward BB application of AD to the CFD code, which is the ADBB method, is represented symbolically asClearly ADII (Eq. (12) and (13)) and ADBB (Eq. (14)) yield the same result at steady-state convergence of each (recall Eq. (2) Second-Order Sensitivity DerivativesThe SO SD methods are presented subsequently using the index notation and beginning with the following preliminary definitions:_= r°_

show abstract

Section: Adh-av Method_ Model Problemmentioning

confidence: 99%

Some advanced concepts in discrete aerodynamic sensitivity analysis

Taylor

Green

Newman

et al. 2001

15th AIAA Computational Fluid Dynamics Conference

Self Cite

View full text Add to dashboard Cite

show abstract

“…Automatic differentiation [3][4][5] is a technique for augmenting computer programs with statements for the computation of derivatives. It relies on the fact that every function, no matter how complicated, is executed on a computer as a (potentially very long) sequence of elementary operations such as additions, multiplications, and elementary functions such as sine and cosine.…”

Section: Automatic Differentiationmentioning

confidence: 99%

“…The reverse mode is used in another tool, ADJIFOR 5 (Automatic Adjoint Generation in FORTRAN). The ADIFOR and ADJIFOR tools have been developed jointly by the Center for Research on Parallel Computation at Rice University and the Mathematics and Computer Sciences Division at Argonne National Laboratory.…”

Section: Automatic Differentiationmentioning

confidence: 99%

Determination of stability and control derivatives using computational fluid dynamics and automatic differentiation

Park

Green

Montgomery

et al. 1999

17th Applied Aerodynamics Conference

Self Cite

View full text Add to dashboard Cite

With the recent interest in novel control effectors there is a need to determine the stability and control derivatives of new aircraft configurations early in the design process. These derivatives are central to most control law design methods and would allow the determination of closed-loop control performance of the vehicle. Early determination of the static and dynamic behavior of an aircraft may permit significant improvement in configuration weight, cost, stealth, and performance through multidisciplinary design. The classical method of determining static stability and control derivatives-constructing and testing wind tunnel models-is expensive and requires a long lead time for the resultant data. Wind tunnel tests are also limited to the preselected control effectors of the model. To overcome these shortcomings, computational fluid

show abstract

“…This lengthy software development cycle has been the primary impediment to widespread use of either approach in conjunction with Euler-and Navier-Stokes-based simulation tools. Tools aimed at automating this process have been under development for some time; 7,9,25,38,39 however, these applications are seldom "hands-off" and frequently fail to produce code that rivals the speed and low storage requirements of hand-developed implementations.…”

Section: Introductionmentioning

confidence: 99%

Efficient Construction of Discrete Adjoint Operators on Unstructured Grids Using Complex Variables

Nielsen

Kleb

2006

AIAA Journal

View full text Add to dashboard Cite

A methodology is developed and implemented to mitigate the lengthy software development cycle typically associated with constructing a discrete adjoint solver for aerodynamic simulations. The approach is based on a complex-variable formulation that enables straightforward differentiation of complicated real-valued functions. An automated scripting process is used to create the complex-variable form of the set of discrete equations. An efficient method for assembling the residual and cost function linearizations is developed. The accuracy of the implementation is verified through comparisons with a discrete direct method as well as a previously developed handcoded discrete adjoint approach. Comparisons are also shown for a large-scale configuration to establish the computational efficiency of the present scheme. To ultimately demonstrate the power of the approach, the implementation is extended to high temperature gas flows in chemical nonequilibrium. Finally, several fruitful research and development avenues enabled by the current work are suggested.

show abstract

Preliminary results from the application of automated adjoint code generation to CFL3D

Cited by 38 publications

References 20 publications

Some advanced concepts in discrete aerodynamic sensitivity analysis

Some advanced concepts in discrete aerodynamic sensitivity analysis

Determination of stability and control derivatives using computational fluid dynamics and automatic differentiation

Efficient Construction of Discrete Adjoint Operators on Unstructured Grids Using Complex Variables

Contact Info

Product

Resources

About