l45 - Lecture 45: Sensitivity Analysis 1 Last Time...

Info iconThis preview shows pages 1–9. Sign up to view the full content.

View Full Document Right Arrow Icon
1 Lecture 45: Sensitivity Analysis
Background image of page 1

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
2 Last Time … Completed consideration of multigrid methods Solved an example problem
Background image of page 2
3 This Time … We will: Look at emerging methods to perform sensitivity analysis using CFD
Background image of page 3

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
What inputs yield desired outputs ? Adjoint Problem: How is a specific output affected by all inputs? Tangent Problem: How does a specific input affect all outputs? Determine effect of uncertainties and tolerances Plan experiments better Improve single-point simulations Mesh sensitivity Sensitivity to iteration parameters Why Sensitivity and Adjoints? Can gain a lot more information than conventional single- point simulations!
Background image of page 4
Approaches Continuous approach » Derive sensitivity/adjoint versions of pde » Discretize and solve » Need to repeat for each physical model » Generally more complex Discrete approach » Start with discrete pde and add sensitivity/adjoints » Easier to automate » Adjoints more difficult 5 x m n mn y x    J = m outputs n inputs Jacobian
Background image of page 5

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
6 Continuous Approach: Sensitivity Equation Governing equations are differentiated with respect to each parameter of interest New PDEs are derived for the sensitivity coefficients, as well as appropriate boundary conditions These are solved using numerical methods similar (but not identical) to the original PDE CPU times scale sub-linearly with the number of parameters
Background image of page 6
7 Example: Transient Heat Conduction in a 1D Slab L x   44 0 | s r x q T T     | xL h T T    0 0 0 0 | | | | | | ( ,0) Parameters: C, k, q , , , , , x x s r x x L x L x L i s r i T q T C q k t x x T q k q T T x T q k h T T x T x T T h T T                           Courtesy K. Dowding, Sandia
Background image of page 7

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
8 Sensitivity Equation for Heat Capacity Differentiate energy equation wrt C: + + 0 Multiply by nominal value of C: +C + 0 Scaled sensitivity of q wrt C: C T q T T q CC C t x t C t x C T T q C C C t C t x C q qC C                                                              PDE for new variable : C C C T k C k T x C x T TT T Ck t x C t                         Courtesy K. Dowding, Sandia
Background image of page 8
Image of page 9
This is the end of the preview. Sign up to access the rest of the document.

This note was uploaded on 12/29/2011 for the course ME 608 taught by Professor Na during the Fall '10 term at Purdue University-West Lafayette.

Page1 / 31

l45 - Lecture 45: Sensitivity Analysis 1 Last Time...

This preview shows document pages 1 - 9. Sign up to view the full document.

View Full Document Right Arrow Icon
Ask a homework question - tutors are online