| EPSRC Reference: |
GR/R21882/01 |
| Title: |
Near Optimal Automatic Differentiation For CFD Fluxes |
| Principal Investigator: |
Forth, Dr SA |
| Other Investigators: |
|
| Researcher Co-Investigators: |
|
| Project Partners: |
|
| Department: |
Engineering Systems |
| Organisation: |
Cranfield University |
| Scheme: |
Standard Research (Pre-FEC) |
| Starts: |
01 April 2001 |
Ends: |
31 March 2003 |
Value (£): |
100,800
|
| EPSRC Research Topic Classifications: |
| Fluid Dynamics |
Numerical Analysis |
|
| EPSRC Industrial Sector Classifications: |
| Aerospace, Defence and Marine |
|
|
| Related Grants: |
|
| Panel History: |
|
|
Summary on Grant Application Form |
|
Computing derivatives of functions defined by computer code is of increasing importance in CFD due to the requirement for fast solution by Newton Solvers and optimisation via adjoint methods. We shall restrict attention to functions defined by code with: about 10 inputs and outputs; branching; BUT no 'unrollable' loops. Such code is typical of that used for inter-cell flux functions in finite-volume flow solvers.AD by source text translation converts code for a vector function y=f(x) into code for calculating the Jacobian matrix or vector-JaCObian products. Current tools (ADIFOR, Odyssee, TAMC) using the classical forward and reverse AD algorithms are well proven but far from optimal in terms of the number of floating point calculations taken to compute the Jacobian.Recent theory shows that we can improve the performce of classical forward and reverse AD by hybridising the two either through use of sparsity (Coleman & Verma), or by regarding AD as obtaining derivatives by solution of a sparse linear system (Naumann) and using Gaussian elimination to minimise the ensuing fill-in and hence reduce the operation count. We shall bring much of Nauman's theoretical work to the stage where it maybe used in Engineering problems described by Fortran code and in so-doing extend the work to cope with branching.
|
|
Key Findings |
|
This information can now be found on Gateway to Research (GtR) http://gtr.rcuk.ac.uk
|
|
Potential use in non-academic contexts |
|
This information can now be found on Gateway to Research (GtR) http://gtr.rcuk.ac.uk
|
|
Impacts |
|
| Description |
This information can now be found on Gateway to Research (GtR) http://gtr.rcuk.ac.uk |
| Summary |
|
| Date Materialised |
|
|
|
|
Sectors submitted by the Researcher |
|
This information can now be found on Gateway to Research (GtR) http://gtr.rcuk.ac.uk
|
| Project URL: |
|
| Further Information: |
|
| Organisation Website: |
http://www.cranfield.ac.uk |