A parametric ROM for the linear frequency domain approach to time-accurate CFD
|
|
- Blaze Daniels
- 6 years ago
- Views:
Transcription
1 Platzhalter für Bild, Bild auf itelfolie hinter das Logo einsetzen A parametric ROM for the linear frequency domain approach to time-accurate CFD Ralf Zimmermann, AG Numerik, U-BS Model Reduction of Complex Dynamical Systems ModRed, MPI Magdeburg Dec
2 A parametric ROM for the LFD approach to CFD Motivation: Parametric reduced-order models Interpolating data on manifolds he Linear Frequency Domain (LFD approach to CFD Offline, semi-offline and online ROM stages A practical example Conclusions & Outlook Dec 12, 213 Ralf Zimmermann A parametric ROM for the LFD approach to CFD Page 2
3 Motivation: Parametric reduced-order models Key idea of subspace based reduced order modeling: Construct low-dimensional subspace Restrict computations to low-dimensional subspace (e.g. via projection Examples: Reduced Bases methods Krylov subspace methods Proper Orthogonal Decomposition Dec 12, 213 Ralf Zimmermann A parametric ROM for the LFD approach to CFD Page 3
4 Motivation: Parametric reduced-order models In this talk: Parameter dependent large-scale linear systems:,,,, Parameters determine operating points Parameters are considered as examination parameters Dec 12, 213 Ralf Zimmermann A parametric ROM for the LFD approach to CFD Page 4
5 Motivation: Parametric reduced-order models Standard order reduction ansatz: use subspace of dim,, Determine coefficient vector via Orthogonal residual condition Minimum residual condition inner product:, = ( S( Dec 12, 213 Ralf Zimmermann A parametric ROM for the LFD approach to CFD Page 5
6 Motivation: Parametric reduced-order models Standard order reduction ansatz: use subspace of dim,, Determine coefficient vector via Orthogonal residual condition Minimum residual condition inner product: A ortha = b orth, = ( S( Dec 12, 213 Ralf Zimmermann A parametric ROM for the LFD approach to CFD Page 6
7 Motivation: Parametric reduced-order models Standard order reduction ansatz: use subspace of dim,, Determine coefficient vector via Orthogonal residual condition Minimum residual condition inner product: A mina = b min, = ( S( Dec 12, 213 Ralf Zimmermann A parametric ROM for the LFD approach to CFD Page 7
8 Motivation: Parametric reduced-order models Suppose that at r operating points, we have computed bases,, Objective: Construct ROM at new operating point,, Required: Parametrized trajectory of projection bases : Idea: Interpolate given bases [1] [2] [1] D. Amsallem, C. Farhat, 28 [2] Nguyen.S., 212 Dec 12, 213 Ralf Zimmermann A parametric ROM for the LFD approach to CFD Page 8
9 Motivation: Parametric reduced-order models Challenges: ROM subspace spanned by orthogonal reduced-order basis (ROB Ensure that parametric variation always gives orthogonal bases Reduced bases matrices must be considered as points in the Grassmann manifold of k-dimensional subspaces or the Stiefel manifold of tall-skinny orthonormal matrices [3] [3] A. Edelman,.A. Arias, S.. Smith, 1998 Dec 12, 213 Ralf Zimmermann A parametric ROM for the LFD approach to CFD Page 9
10 A parametric ROM for the LFD approach to CFD Motivation: Parametric reduced-order models Interpolating data on manifolds he Linear Frequency Domain (LFD approach to CFD Offline, semi-offline and online ROM stages A practical example Conclusions & Outlook Dec 12, 213 Ralf Zimmermann A parametric ROM for the LFD approach to CFD Page 1
11 Interpolating data on manifolds (A: basic principle Matrix manifolds Stiefel manifold,! Grassmann manifold ", # dim => each point is a subspace of dimension m, equivalence class of matrices spanning a given subspace may be represented by point in Stiefel manifold Dec 12, 213 Ralf Zimmermann A parametric ROM for the LFD approach to CFD Page 11
12 Interpolating data on manifolds (A: basic principle G( n, m U ( q 3 ( q 1 U ( q 1 U ( q ( q 2 U ( q 2 U ( q 3 U ( q * U ( q 2 ( q 3 U ( q U q ( G( n, m U( q 1 Dec 12, 213 Ralf Zimmermann A parametric ROM for the LFD approach to CFD Page 12
13 Interpolating data on manifolds (B: Matrix Manifolds and here numerical linear algebra comes in Grassmann exponential mapping '( '( [ ] [ ] G( n, m, U = exp [ U WΣQ ] U ( [ ] = SVD = [ U Q cos( Σ m m : + W sin( Σ] G( n, m, Isometric mapping: tangent space to manifold Dec 12, 213 Ralf Zimmermann A parametric ROM for the LFD approach to CFD Page 13
14 Dec 12, 213 Ralf Zimmermann A parametric ROM for the LFD approach to CFD Page 14 and here numerical linear algebra comes in Grassmann logarithmic mapping Interpolating data on manifolds (B: Matrix Manifolds ( 1 1 ] [ ] [ ( (, ( tan,, ( ] [ ] [ log :,, ( ] [ = Σ Σ = = = U U U U U I Q W W m n G U I U U m n G U n n SVD U U m m Isometric mapping: manifold to tangent space '( '(
15 Interpolating data on manifolds (C: A word of caution Challenges It cannot be taken for granted that the interpolation s result stays in the tangential space! here are interpolation schemes, that are not admissible in this regard! angential vectors are represented by * +-matrices: Entry-by-entry interpolation may require millions of interpolation procedures Dec 12, 213 Ralf Zimmermann A parametric ROM for the LFD approach to CFD Page 15
16 Interpolating data on manifolds (C: A word of caution rivial example: Nearest-neighbor interpolation leaves the data subspace y(x = 5x x Dec 12, 213 Ralf Zimmermann A parametric ROM for the LFD approach to CFD Page 16
17 Interpolating data on manifolds (C: A word of caution,,,,, Dec 12, 213 Ralf Zimmermann A parametric ROM for the LFD approach to CFD Page 17
18 Interpolating data on manifolds (C: A word of caution Conclusion Let - -,,.. be vectors in a Grassmann tangent space, represented by matrices An admissible interpolator must be of the form * ( q = k l = 1 f l ( q l In 1D, e.g. linear interpolation and Lagrange interpolation fulfill this requirement In multiple D, radial basis function interpolation is feasible Dec 12, 213 Ralf Zimmermann A parametric ROM for the LFD approach to CFD Page 18
19 A parametric ROM for the LFD approach to CFD Motivation: Parametric reduced-order models Interpolating data on manifolds he Linear Frequency Domain (LFD approach to CFD Offline, semi-offline and online ROM stages A practical example Conclusions & Outlook Dec 12, 213 Ralf Zimmermann A parametric ROM for the LFD approach to CFD Page 19
20 he linear frequency approach to CFD Underlying assumption Flow to be simulated is periodic and approximately harmonic Valid for pitching motions of small amplitude Flow state vector given by mean flow + time-dependent fluctuations / 1 2 / 3, 4 #(flow vars #(grid points Same applies to grid coordinates and grid cell volumes Dec 12, 213 Ralf Zimmermann A parametric ROM for the LFD approach to CFD Page 2
21 he linear frequency approach to CFD First-order aylor and transition to Fourier space give complex linear system: : ; < = 3 (: = temporal mean of grid cell volumes, 9 = frequency Dec 12, 213 Ralf Zimmermann A parametric ROM for the LFD approach to CFD Page 21
22 he linear frequency approach to CFD First-order aylor and transition to Fourier space give complex linear system: : ; < = 3 9, : ; 9, : = 3 (: = temporal mean of grid cell volumes, 9 = frequency Parametric dependency: - Mach number determines operating point (mean value : - Frequency varies at each operating point 9 ; ; *9, :+ = 3 Dec 12, 213 Ralf Zimmermann A parametric ROM for the LFD approach to CFD Page 22
23 A parametric ROM for the LFD approach to CFD Motivation: Parametric reduced-order models Interpolating data on manifolds he Linear Frequency Domain (LFD approach to CFD Offline, semi-offline and online ROM stages A practical example Conclusions & Outlook Dec 12, 213 Ralf Zimmermann A parametric ROM for the LFD approach to CFD Page 23
24 Offline, semi-offline and online stage Offline stage Select operating points (OPs -,, > At each OP, compute projection basis? via snapshot POD Dec 12, 213 Ralf Zimmermann A parametric ROM for the LFD approach to CFD Page 24
25 Offline, semi-offline and online stage Offline stage Select operating points (OPs -,, > At each OP, compute projection basis? via snapshot POD For any OP of interest, assess the trajectory : via Grassmann Interpolation to obtain * + = 3 Dec 12, 213 Ralf Zimmermann A parametric ROM for the LFD approach to CFD Page 25
26 Offline, semi-offline and online stage Semi-offline stage At operating point, project full order system onto Exploit separable parameter 9, : : A 1 9 < " 1 89B* + 9, 1 9 < 1 * + = Dec 12, 213 Ralf Zimmermann A parametric ROM for the LFD approach to CFD Page 26
27 Offline, semi-offline and online stage Semi-offline stage At operating point, project full order system onto Exploit separable parameter 9, : : A 1 9 < " 1 89B* + = Similar for orth res 9, 1 9 < 1 * + = Dec 12, 213 Ralf Zimmermann A parametric ROM for the LFD approach to CFD Page 27
28 Offline, semi-offline and online stage Semi-offline stage At operating point, project full order system onto Exploit separable parameter dependency: precompute A, ", B* + = * + = Dec 12, 213 Ralf Zimmermann A parametric ROM for the LFD approach to CFD Page 28
29 Offline, semi-offline and online stage Online stage For any frequency 9 compute reduced system and right hand 9, A 1 9 < " 1 89B* + 9, 1 9 < 1 * + Solve the (m x m 9, Effort: E D independent of full scale dimension = real time online stage! Dec 12, 213 Ralf Zimmermann A parametric ROM for the LFD approach to CFD Page 29
30 A parametric ROM for the LFD approach to CFD Motivation: Parametric reduced-order models Interpolating data on manifolds he Linear Frequency Domain (LFD approach to CFD Offline, semi-offline and online ROM stages A practical example Conclusions & Outlook Dec 12, 213 Ralf Zimmermann A parametric ROM for the LFD approach to CFD Page 3
31 A practical example Grassmann interpolation of ROBs for LFD approach to CFD NACA64A1 airfoil: Snapshots in transonic Regime Shocks 2-parameter model: Computational grid, size: 1,727 points Ma {.8,.82,.84, }.86,.88,.81, κ {.1,.3,.5,.7,.9} α, Re 7.5mio fixed Dec 12, 213 Ralf Zimmermann A parametric ROM for the LFD approach to CFD Page 31
32 A practical example Snapshot and OP-bases sample plan (reduced frequency Sampled operating points defined by Mach number Dec 12, 213 Ralf Zimmermann A parametric ROM for the LFD approach to CFD Page 32
33 Ma.85, κ =.2 metric: 1/vols prom LFD Ref proj LFD Ref Ma.85, κ =.4 metric: 1/vols prom global ROM LFD Ref proj LFD Ref.5.5 A practical example Freqdom_drho_re Freqdom_drho_re , * + :G- * + Freqdom_drho_re X Ma.85, κ =.6 metric: 1/vols prom LFD Ref Proj LFD Ref Freqdom_drho_re X Ma.85, κ =.8 metric: 1/vols prom LFD Ref proj LFD Ref Dec 12, 213 Ralf Zimmermann A parametric ROM for the LFD approach to CFD Page X X
34 Ma.87, κ =.2 metric: 1/vols prom LFD Ref proj LFD Ref Ma.87, κ =.4 metric: 1/vols prom global ROM LFD Ref proj LFD Ref.5.5 A practical example Freqdom_drho_re Freqdom_drho_re , * + :G- * + Freqdom_drho_re X Ma.87, κ =.6 metric: 1/vols prom LFD Ref Proj LFD Ref Freqdom_drho_re X Ma.87, κ =.8 metric: 1/vols prom LFD Ref proj LFD Ref Dec 12, 213 Ralf Zimmermann A parametric ROM for the LFD approach to CFD Page X X
35 Ma.85, κ =.2 metric: Euclid prom LFD Ref proj LFD Ref Ma.85, κ =.4 metric: Euclid prom LFD Ref proj LFD Ref.5.5 A practical example Metrics matter! Freqdom_drho_re -.5 Freqdom_drho_re X X, * + * + Freqdom_drho_re Ma.85, κ =.6 metric: Euclid prom LFD Ref Proj LFD Ref Freqdom_drho_re Ma.85, κ =.8 metric: Euclid prom LFD Ref proj LFD Ref Dec 12, 213 Ralf Zimmermann A parametric ROM for the LFD approach to CFD Page X X
36 A practical example Metrics matter! W.r.t. this metric, even the projection of the reference is quite off!, * + : * + Freqdom_drho_re Freqdom_drho_re Ma.85, κ =.2 metric: L2 prom LFD Ref proj LFD Ref X Ma.85, κ =.6 metric: L2 prom LFD Ref Proj LFD Ref Freqdom_drho_re Freqdom_drho_re Ma.85, κ =.4 metric: L2 prom LFD Ref proj LFD Ref X Ma.85, κ =.8 metric: L2 prom LFD Ref proj LFD Ref Dec 12, 213 Ralf Zimmermann A parametric ROM for the LFD approach to CFD Page X X
37 A practical example he orthogonal residual approach is non-competitive Ma.85, κ =.4 orthogonal res LFD Ref prom metric: 1/vols prom metric: Euclid prom metric: L2 Ma.87, κ =.8 orthogonal res LFD Ref prom metric: 1/vols prom metric: Euclid prom metric: L2.5.2 Freqdom_drho_re Freqdom_drho_re X X Dec 12, 213 Ralf Zimmermann A parametric ROM for the LFD approach to CFD Page 37
38 A practical example he global 2D POD basis approach is non-competitive Ma.85, κ =.4 minimum res, metric:1/vols Global ROB Ma.85, κ =.4 orthogonal res, metric: 1/vols Global ROB.5.5 Freqdom_drho_re Freqdom_drho_re -.5 LFD Ref dim-3 grob dim-5 grob -.5 LFD Ref dim-3 grob X X, * + :G- * + Dec 12, 213 Ralf Zimmermann A parametric ROM for the LFD approach to CFD Page 38
39 Ma.85, κ =.4 minimum res, metric:1/vols Global ROB he reduced global 2D POD basis approach is competitive, if only the 5 most dominant POD modes are kept, but Freqdom_drho_re LFD Ref dim-3 grob dim-5 grob X Dec 12, 213 Ralf Zimmermann A parametric ROM for the LFD approach to CFD Page 39, * + :G- * +
40 1 Ma.84, metric: 1/vols grob prob 1 Ma.85, metric 1/vols grob prob.8.8 any global basis approach lacks the interpolation property! Error κ Ma.86, metric: 1/vols grob prob Error κ Ma.87, metric: 1/vols grob prob Error Error κ Dec 12, 213 Ralf Zimmermann A parametric ROM for the LFD approach to CFD Page κ
41 Computational effort Full order dimension: 5*2*#(grid points = 17,27 Reduced order dimension: 5(complex = 1 (real at each OP iming results Offline Mean flow Full-order LFD Complex POD total CPU time ~3s 5*35s ~.26s ~475s (7min Semi-offline Grassmann interp. Reading system Order reduction CPU time ~.66s ~1.s ~.15s ~11s Online Set up ROM Solve ROM Build flow state CPU time Below measure acc. Below measure.1s ~.1s Dec 12, 213 Ralf Zimmermann A parametric ROM for the LFD approach to CFD Page 41
42 Computational effort Full order dimension: 5*2*#(grid points = 17,27 Reduced order dimension: 5(complex = 1 (real at each OP iming results Offline Mean flow Full-order LFD Complex POD total CPU time ~3s 5*35s ~.26s ~475s (7min Semi-offline Grassmann interp. Reading system Order reduction CPU time ~.66s ~1.s ~.15s ~11s Online Set up ROM Solve ROM Build flow state CPU time Below measure acc. Below measure.1s ~.1s Dec 12, 213 Ralf Zimmermann A parametric ROM for the LFD approach to CFD Page 42 Online speed up: factor ~35
43 A parametric ROM for the LFD approach to CFD Motivation: Parametric reduced-order models Interpolating data on manifolds he Linear Frequency Domain (LFD approach to CFD Offline, semi-offline and online ROM stages A practical example Conclusions & Outlook Dec 12, 213 Ralf Zimmermann A parametric ROM for the LFD approach to CFD Page 43
44 Conclusions & Outlook Parametric ROM for LFD solver introduced Parameters of major interest (Mach, frequency allow for a natural division into operation points and examination parameters ROM splits into three stages Offline (provide snapshot data, compute projection bases Semi-Offline Online (precompute frequency-independent system parts, separable parameter dependency renders an interpolation of reduced systems as in [4,5] unnecessary (solve reduced system, real-time performance Dec 12, 213 Ralf Zimmermann A parametric ROM for the LFD approach to CFD Page 44 [4] Degroote J. et al., 21 [5] Panzer, H. et al., 21
45 Conclusions & Outlook Metric proves to be essential in order to obtain good results (requires minor technical modifications, not shown here Minimum residual approach outperforms orthogonal residual approach Reduced global basis approach gives satisfactory results, but underperforms when compared to Grassmann interpolation, (Necessarily lacks interpolation property Next step: geometrically optimized reduced order basis Dec 12, 213 Ralf Zimmermann A parametric ROM for the LFD approach to CFD Page 45
46 hank you for your attention! Dec 12, 213 Ralf Zimmermann A parametric ROM for the LFD approach to CFD Page 46
47 Dec 12, 213 Ralf Zimmermann A parametric ROM for the LFD approach to CFD Page 47 An admissible interpolation scheme: radial basis functions Radial basis functions (RBF in a nutshell Given: Sample points Sample values RBF interpolator: R p y y p y y R p p k k d k = = (,..., (,..., (,..., ( ( 1 1 * p y y p y k ψ Ψ = ( ( k k k j i j i R p p r = Ψ, ( ( k k R p p r p p r = M 1 ψ
48 An admissible interpolation scheme: radial basis functions Application to Grassmann tangent vectors Given: Sample points Sample values p 1 1,..., p k = ( p R 1 d,..., k = ( p k R n m Dec 12, 213 Ralf Zimmermann A parametric ROM for the LFD approach to CFD Page 48
49 Dec 12, 213 Ralf Zimmermann A parametric ROM for the LFD approach to CFD Page 49 ( ( ( = = = Ψ = Ψ Ψ = k l l l l m k l l l k l l p e p p 1 1 :, 1 1 :,1 1 1 ( (,, ( ψ ψ ψ K An admissible interpolation scheme: radial basis functions Application to Grassmann tangent vectors RBF interpolator: ( ( ( m j k j j m j n i k j i j i m j n i j i p p p Ψ = Ψ = (,..., ( (,..., ( ( 1 :, 1 :,, 1, 1,, *, ψ ψ = = k l l l p f 1 (
50 An admissible interpolation scheme: radial basis functions Conclusion RBF interpolation is admissible (so is e.g. linear interpolation Entry-by-entry RBF interpolation tangent matrices interpolating coefficients of linear expansion Associated computational effort: IJ interpolation procedures, each for K sample points vs. K interpolation procedures, each for K sample points Dec 12, 213 Ralf Zimmermann A parametric ROM for the LFD approach to CFD Page 5
Geometric adaptation and parameterization of subspacebased reduced models
Platzhalter für Bild, Bild auf Titelfolie hinter das Logo einsetzen Geometric adaptation and parameterization of subspacebased reduced models Ralf Zimmermann, AG Numerik, ICM, TU Braunschweig Workshop:
More informationAn Interpolation Method for Adapting Reduced-Order Models and Application to Aeroelasticity
An Interpolation Method for Adapting Reduced-Order Models and Application to Aeroelasticity David Amsallem and Charbel Farhat Stanford University, Stanford, CA 94305-3035, USA Reduced-order models (ROMs)
More informationUNSTEADY TAU AND ROM FOR FLUTTER COMPUTATIONS OF TRANSPORT AIRCRAFT. Reik Thormann, Ralph Voß. Folie 1
UNSTEADY TAU AND ROM FOR FLUTTER COMPUTATIONS OF TRANSPORT AIRCRAFT Reik Thormann, Ralph Voß Folie 1 Standardfoliensatz >01.03.2007 Outline Motivation Synthetic Mode Correction (SMC) Linearized CFD Method
More informationNumerical Analysis and Statistics on Tensor Parameter Spaces
Numerical Analysis and Statistics on Tensor Parameter Spaces SIAM - AG11 - Tensors Oct. 7, 2011 Overview Normal Mean / Karcher Mean Karcher mean / Normal mean - finding representatives for a set of points
More informationEfficient O(N log N) algorithms for scattered data interpolation
Efficient O(N log N) algorithms for scattered data interpolation Nail Gumerov University of Maryland Institute for Advanced Computer Studies Joint work with Ramani Duraiswami February Fourier Talks 2007
More informationUnsupervised learning in Vision
Chapter 7 Unsupervised learning in Vision The fields of Computer Vision and Machine Learning complement each other in a very natural way: the aim of the former is to extract useful information from visual
More informationModelling and Visualization of High Dimensional Data. Sample Examination Paper
Duration not specified UNIVERSITY OF MANCHESTER SCHOOL OF COMPUTER SCIENCE Modelling and Visualization of High Dimensional Data Sample Examination Paper Examination date not specified Time: Examination
More informationarxiv: v1 [math.oc] 29 Jul 2014
Progressive construction of a parametric reduced-order model for PDE-constrained optimization Matthew J. Zahr 1 and Charbel Farhat 1,2,3 arxiv:1407.7618v1 [math.oc] 29 Jul 2014 1 Institute for Computational
More informationLarge-Scale Face Manifold Learning
Large-Scale Face Manifold Learning Sanjiv Kumar Google Research New York, NY * Joint work with A. Talwalkar, H. Rowley and M. Mohri 1 Face Manifold Learning 50 x 50 pixel faces R 2500 50 x 50 pixel random
More informationDevelopment of a CFD Capability for Full Helicopter Engineering Analysis
Development of a CFD Capability for Full Helicopter Engineering Analysis George Barakos Department of Engineering University of Liverpool 5/6 April 2006 Loughborough University Collective effort of more
More informationA reduced-order model for NS equations from boundary matching and OT
A reduced-order model for NS equations from boundary matching and OT Angelo Iollo Michel Bergmann Andrea Ferrero Institut de Mathématiques de Bordeaux Inria Memphis Team Haysam Telib Optimad Engineering,
More informationState of the art at DLR in solving aerodynamic shape optimization problems using the discrete viscous adjoint method
DLR - German Aerospace Center State of the art at DLR in solving aerodynamic shape optimization problems using the discrete viscous adjoint method J. Brezillon, C. Ilic, M. Abu-Zurayk, F. Ma, M. Widhalm
More informationUNDERBODY BLAST (project 1b)
UNDERBODY BLAST (project 1b) PI: Mingjun Wei, New Mexico State University (overall project lead: Charbel Farhat, Stanford University) Graduate Students: HaoQan Gao, Mehdi Tabandeh Khorshid Post- Doctoral
More informationA GPU-based Approximate SVD Algorithm Blake Foster, Sridhar Mahadevan, Rui Wang
A GPU-based Approximate SVD Algorithm Blake Foster, Sridhar Mahadevan, Rui Wang University of Massachusetts Amherst Introduction Singular Value Decomposition (SVD) A: m n matrix (m n) U, V: orthogonal
More informationExploiting a database to predict the in-flight stability of the F-16
Exploiting a database to predict the in-flight stability of the F-16 David Amsallem and Julien Cortial December 12, 2008 1 Introduction Among the critical phenomena that have to be taken into account when
More informationIntroduction. To fully understand the capabilities and the limits of the POD, you just have to know a few main points. The main ideas behind the POD :
This documentation is made for the engineer who would like to leverage Proper Orthogonal Decomposition for his work, bringing some intelligence in the use of Scilab Model Reduction Toolbox. Introduction
More informationData mining with sparse grids
Data mining with sparse grids Jochen Garcke and Michael Griebel Institut für Angewandte Mathematik Universität Bonn Data mining with sparse grids p.1/40 Overview What is Data mining? Regularization networks
More informationFunction approximation using RBF network. 10 basis functions and 25 data points.
1 Function approximation using RBF network F (x j ) = m 1 w i ϕ( x j t i ) i=1 j = 1... N, m 1 = 10, N = 25 10 basis functions and 25 data points. Basis function centers are plotted with circles and data
More informationOnline Subspace Estimation and Tracking from Missing or Corrupted Data
Online Subspace Estimation and Tracking from Missing or Corrupted Data Laura Balzano www.ece.wisc.edu/~sunbeam Work with Benjamin Recht and Robert Nowak Subspace Representations Capture Dependencies Subspace
More informationIntroduction to ANSYS DesignXplorer
Lecture 4 14. 5 Release Introduction to ANSYS DesignXplorer 1 2013 ANSYS, Inc. September 27, 2013 s are functions of different nature where the output parameters are described in terms of the input parameters
More informationSupplementary Figure 1. Decoding results broken down for different ROIs
Supplementary Figure 1 Decoding results broken down for different ROIs Decoding results for areas V1, V2, V3, and V1 V3 combined. (a) Decoded and presented orientations are strongly correlated in areas
More informationData fusion and multi-cue data matching using diffusion maps
Data fusion and multi-cue data matching using diffusion maps Stéphane Lafon Collaborators: Raphy Coifman, Andreas Glaser, Yosi Keller, Steven Zucker (Yale University) Part of this work was supported by
More informationChapter 3 Numerical Methods
Chapter 3 Numerical Methods Part 1 3.1 Linearization and Optimization of Functions of Vectors 1 Problem Notation 2 Outline 3.1.1 Linearization 3.1.2 Optimization of Objective Functions 3.1.3 Constrained
More informationPlanes Intersecting Cones: Static Hypertext Version
Page 1 of 12 Planes Intersecting Cones: Static Hypertext Version On this page, we develop some of the details of the plane-slicing-cone picture discussed in the introduction. The relationship between the
More informationComputing and Processing Correspondences with Functional Maps
Computing and Processing Correspondences with Functional Maps SIGGRAPH 2017 course Maks Ovsjanikov, Etienne Corman, Michael Bronstein, Emanuele Rodolà, Mirela Ben-Chen, Leonidas Guibas, Frederic Chazal,
More informationComputer Vision 2. SS 18 Dr. Benjamin Guthier Professur für Bildverarbeitung. Computer Vision 2 Dr. Benjamin Guthier
Computer Vision 2 SS 18 Dr. Benjamin Guthier Professur für Bildverarbeitung Computer Vision 2 Dr. Benjamin Guthier 1. IMAGE PROCESSING Computer Vision 2 Dr. Benjamin Guthier Content of this Chapter Non-linear
More informationLecture 2 September 3
EE 381V: Large Scale Optimization Fall 2012 Lecture 2 September 3 Lecturer: Caramanis & Sanghavi Scribe: Hongbo Si, Qiaoyang Ye 2.1 Overview of the last Lecture The focus of the last lecture was to give
More informationReduced-Order Modelling strategies for the finite element approximation of the Incompressible Navier-Stokes equations
Reduced-Order Modelling strategies for the finite element approximation of the Incompressible Navier-Stokes equations Joan Baiges,1,2, Ramon Codina 2,1 and Sergio Idelsohn 3,1 1 Centre Internacional de
More informationUsing Subspace Constraints to Improve Feature Tracking Presented by Bryan Poling. Based on work by Bryan Poling, Gilad Lerman, and Arthur Szlam
Presented by Based on work by, Gilad Lerman, and Arthur Szlam What is Tracking? Broad Definition Tracking, or Object tracking, is a general term for following some thing through multiple frames of a video
More informationLinear Methods for Regression and Shrinkage Methods
Linear Methods for Regression and Shrinkage Methods Reference: The Elements of Statistical Learning, by T. Hastie, R. Tibshirani, J. Friedman, Springer 1 Linear Regression Models Least Squares Input vectors
More informationFast Radial Basis Functions for Engineering Applications. Prof. Marco Evangelos Biancolini University of Rome Tor Vergata
Fast Radial Basis Functions for Engineering Applications Prof. Marco Evangelos Biancolini University of Rome Tor Vergata Outline 2 RBF background Fast RBF on HPC Engineering Applications Mesh morphing
More informationData mining with sparse grids using simplicial basis functions
Data mining with sparse grids using simplicial basis functions Jochen Garcke and Michael Griebel Institut für Angewandte Mathematik Universität Bonn Part of the work was supported within the project 03GRM6BN
More informationModel Reduction for Variable-Fidelity Optimization Frameworks
Model Reduction for Variable-Fidelity Optimization Frameworks Karen Willcox Aerospace Computational Design Laboratory Department of Aeronautics and Astronautics Massachusetts Institute of Technology Workshop
More informationMachine Learning / Jan 27, 2010
Revisiting Logistic Regression & Naïve Bayes Aarti Singh Machine Learning 10-701/15-781 Jan 27, 2010 Generative and Discriminative Classifiers Training classifiers involves learning a mapping f: X -> Y,
More informationLecture 3: Camera Calibration, DLT, SVD
Computer Vision Lecture 3 23--28 Lecture 3: Camera Calibration, DL, SVD he Inner Parameters In this section we will introduce the inner parameters of the cameras Recall from the camera equations λx = P
More informationInverse and Implicit functions
CHAPTER 3 Inverse and Implicit functions. Inverse Functions and Coordinate Changes Let U R d be a domain. Theorem. (Inverse function theorem). If ϕ : U R d is differentiable at a and Dϕ a is invertible,
More informationSDLS: a Matlab package for solving conic least-squares problems
SDLS: a Matlab package for solving conic least-squares problems Didier Henrion 1,2 Jérôme Malick 3 June 28, 2007 Abstract This document is an introduction to the Matlab package SDLS (Semi-Definite Least-Squares)
More informationcalibrated coordinates Linear transformation pixel coordinates
1 calibrated coordinates Linear transformation pixel coordinates 2 Calibration with a rig Uncalibrated epipolar geometry Ambiguities in image formation Stratified reconstruction Autocalibration with partial
More information10701 Machine Learning. Clustering
171 Machine Learning Clustering What is Clustering? Organizing data into clusters such that there is high intra-cluster similarity low inter-cluster similarity Informally, finding natural groupings among
More informationContents. Implementing the QR factorization The algebraic eigenvalue problem. Applied Linear Algebra in Geoscience Using MATLAB
Applied Linear Algebra in Geoscience Using MATLAB Contents Getting Started Creating Arrays Mathematical Operations with Arrays Using Script Files and Managing Data Two-Dimensional Plots Programming in
More informationFFD, mesh morphing & reduced order models: Enablers for efficient aerodynamic shape optimization
FFD, mesh morphing & reduced order models: Enablers for efficient aerodynamic shape optimization F Salmoiraghi, G Rozza SISSA mathlab A Scardigli, H Telib OPTIMAD engineering srl Outline 1. Prac'cal)problems)in)shape)op'miza'on)
More informationUnit II-2. Orthogonal projection. Orthogonal projection. Orthogonal projection. the scalar is called the component of u along v. two vectors u,v are
Orthogonal projection Unit II-2 Orthogonal projection the scalar is called the component of u along v in real ips this may be a positive or negative value in complex ips this may have any complex value
More informationHuman Gesture Recognition on Product Manifolds
Journal of Machine Learning Research 13 (2012) 3297-3321 Submitted 5/12; Revised 10/12; Published 11/12 Human Gesture Recognition on Product Manifolds Yui Man Lui Department of Computer Science Colorado
More informationLTI Thesis Defense: Riemannian Geometry and Statistical Machine Learning
Outline LTI Thesis Defense: Riemannian Geometry and Statistical Machine Learning Guy Lebanon Motivation Introduction Previous Work Committee:John Lafferty Geoff Gordon, Michael I. Jordan, Larry Wasserman
More informationDevelopment of a Maxwell Equation Solver for Application to Two Fluid Plasma Models. C. Aberle, A. Hakim, and U. Shumlak
Development of a Maxwell Equation Solver for Application to Two Fluid Plasma Models C. Aberle, A. Hakim, and U. Shumlak Aerospace and Astronautics University of Washington, Seattle American Physical Society
More informationMotivation. Parametric Curves (later Surfaces) Outline. Tangents, Normals, Binormals. Arclength. Advanced Computer Graphics (Fall 2010)
Advanced Computer Graphics (Fall 2010) CS 283, Lecture 19: Basic Geometric Concepts and Rotations Ravi Ramamoorthi http://inst.eecs.berkeley.edu/~cs283/fa10 Motivation Moving from rendering to simulation,
More informationAlternative Statistical Methods for Bone Atlas Modelling
Alternative Statistical Methods for Bone Atlas Modelling Sharmishtaa Seshamani, Gouthami Chintalapani, Russell Taylor Department of Computer Science, Johns Hopkins University, Baltimore, MD Traditional
More informationQuaternions and Exponentials
Quaternions and Exponentials Michael Kazhdan (601.457/657) HB A.6 FvDFH 21.1.3 Announcements OpenGL review II: Today at 9:00pm, Malone 228 This week's graphics reading seminar: Today 2:00-3:00pm, my office
More informationSurfaces, meshes, and topology
Surfaces from Point Samples Surfaces, meshes, and topology A surface is a 2-manifold embedded in 3- dimensional Euclidean space Such surfaces are often approximated by triangle meshes 2 1 Triangle mesh
More informationSemi-Lagrangian kinetic and gyrokinetic simulations
Semi-Lagrangian kinetic and gyrokinetic simulations M. Mehrenberger IRMA, University of Strasbourg and TONUS project (INRIA) Marseille, Equations cinétiques, 11.11.2014 Collaborators : B. Afeyan, A. Back,
More informationAlmost Curvature Continuous Fitting of B-Spline Surfaces
Journal for Geometry and Graphics Volume 2 (1998), No. 1, 33 43 Almost Curvature Continuous Fitting of B-Spline Surfaces Márta Szilvási-Nagy Department of Geometry, Mathematical Institute, Technical University
More informationsmooth coefficients H. Köstler, U. Rüde
A robust multigrid solver for the optical flow problem with non- smooth coefficients H. Köstler, U. Rüde Overview Optical Flow Problem Data term and various regularizers A Robust Multigrid Solver Galerkin
More informationOptimising MPI Applications for Heterogeneous Coupled Clusters with MetaMPICH
Optimising MPI Applications for Heterogeneous Coupled Clusters with MetaMPICH Carsten Clauss, Martin Pöppe, Thomas Bemmerl carsten@lfbs.rwth-aachen.de http://www.mp-mpich.de Lehrstuhl für Betriebssysteme
More informationSpace Filling Curves and Hierarchical Basis. Klaus Speer
Space Filling Curves and Hierarchical Basis Klaus Speer Abstract Real world phenomena can be best described using differential equations. After linearisation we have to deal with huge linear systems of
More informationDimension Reduction CS534
Dimension Reduction CS534 Why dimension reduction? High dimensionality large number of features E.g., documents represented by thousands of words, millions of bigrams Images represented by thousands of
More informationIntroduction to ANSYS CFX
Workshop 03 Fluid flow around the NACA0012 Airfoil 16.0 Release Introduction to ANSYS CFX 2015 ANSYS, Inc. March 13, 2015 1 Release 16.0 Workshop Description: The flow simulated is an external aerodynamics
More informationFunded by the European Union INRIA. AEROGUST Workshop 27 th - 28 th April 2017, University of Liverpool. Presented by Andrea Ferrero and Angelo Iollo
INRIA AEROGUST Workshop 27 th - 28 th April 2017, University of Liverpool Presented by Andrea Ferrero and Angelo Iollo Aero-elastic study of a wind turbine subjected to a gust Development of high-fidelity
More informationAdaptive sampling for nonlinear dimensionality reduction based on manifold learning Franz, Thomas; Zimmermann, Ralf; Goertz, Stefan
Syddansk Universitet Adaptive sampling for nonlinear dimensionality reduction based on manifold learning Franz, Thomas; Zimmermann, Ralf; Goertz, Stefan Published in: Model Reduction of Parametrized Systems
More informationExponential Maps for Computer Vision
Exponential Maps for Computer Vision Nick Birnie School of Informatics University of Edinburgh 1 Introduction In computer vision, the exponential map is the natural generalisation of the ordinary exponential
More informationThe Pre-Image Problem in Kernel Methods
The Pre-Image Problem in Kernel Methods James Kwok Ivor Tsang Department of Computer Science Hong Kong University of Science and Technology Hong Kong The Pre-Image Problem in Kernel Methods ICML-2003 1
More informationApproximation of a Fuzzy Function by Using Radial Basis Functions Interpolation
International Journal of Mathematical Modelling & Computations Vol. 07, No. 03, Summer 2017, 299-307 Approximation of a Fuzzy Function by Using Radial Basis Functions Interpolation R. Firouzdor a and M.
More informationTAU mesh deformation. Thomas Gerhold
TAU mesh deformation Thomas Gerhold The parallel mesh deformation of the DLR TAU-Code Introduction Mesh deformation method & Parallelization Results & Applications Conclusion & Outlook Introduction CFD
More informationThe end of affine cameras
The end of affine cameras Affine SFM revisited Epipolar geometry Two-view structure from motion Multi-view structure from motion Planches : http://www.di.ens.fr/~ponce/geomvis/lect3.pptx http://www.di.ens.fr/~ponce/geomvis/lect3.pdf
More informationRelative Constraints as Features
Relative Constraints as Features Piotr Lasek 1 and Krzysztof Lasek 2 1 Chair of Computer Science, University of Rzeszow, ul. Prof. Pigonia 1, 35-510 Rzeszow, Poland, lasek@ur.edu.pl 2 Institute of Computer
More informationRecognition, SVD, and PCA
Recognition, SVD, and PCA Recognition Suppose you want to find a face in an image One possibility: look for something that looks sort of like a face (oval, dark band near top, dark band near bottom) Another
More informationR3DG Features: Relative 3D Geometry-based Skeletal Representations for Human Action Recognition
R3DG Features: Relative 3D Geometry-based Skeletal Representations for Human Action Recognition Raviteja Vemulapalli 1, Felipe Arrate 2, Rama Chellappa 1 1 Center for Automation Research, UMIACS University
More informationNonlinear Noise Reduction
Chapter 3 Nonlinear Noise Reduction Many natural and engineered systems generate nonlinear deterministic time series that are contaminated by random measurement or dynamic noise. Even clean time series
More informationRadial Basis Function Networks: Algorithms
Radial Basis Function Networks: Algorithms Neural Computation : Lecture 14 John A. Bullinaria, 2015 1. The RBF Mapping 2. The RBF Network Architecture 3. Computational Power of RBF Networks 4. Training
More informationAllScale Pilots Applications AmDaDos Adaptive Meshing and Data Assimilation for the Deepwater Horizon Oil Spill
This project has received funding from the European Union s Horizon 2020 research and innovation programme under grant agreement No. 671603 An Exascale Programming, Multi-objective Optimisation and Resilience
More informationGEOMETRIC TOOLS FOR COMPUTER GRAPHICS
GEOMETRIC TOOLS FOR COMPUTER GRAPHICS PHILIP J. SCHNEIDER DAVID H. EBERLY MORGAN KAUFMANN PUBLISHERS A N I M P R I N T O F E L S E V I E R S C I E N C E A M S T E R D A M B O S T O N L O N D O N N E W
More informationCS 664 Segmentation. Daniel Huttenlocher
CS 664 Segmentation Daniel Huttenlocher Grouping Perceptual Organization Structural relationships between tokens Parallelism, symmetry, alignment Similarity of token properties Often strong psychophysical
More informationRecent developments for the multigrid scheme of the DLR TAU-Code
www.dlr.de Chart 1 > 21st NIA CFD Seminar > Axel Schwöppe Recent development s for the multigrid scheme of the DLR TAU-Code > Apr 11, 2013 Recent developments for the multigrid scheme of the DLR TAU-Code
More informationEfficient Imaging Algorithms on Many-Core Platforms
Efficient Imaging Algorithms on Many-Core Platforms H. Köstler Dagstuhl, 22.11.2011 Contents Imaging Applications HDR Compression performance of PDE-based models Image Denoising performance of patch-based
More informationChapter 6. Curves and Surfaces. 6.1 Graphs as Surfaces
Chapter 6 Curves and Surfaces In Chapter 2 a plane is defined as the zero set of a linear function in R 3. It is expected a surface is the zero set of a differentiable function in R n. To motivate, graphs
More informationNotes on point set topology, Fall 2010
Notes on point set topology, Fall 2010 Stephan Stolz September 3, 2010 Contents 1 Pointset Topology 1 1.1 Metric spaces and topological spaces...................... 1 1.2 Constructions with topological
More informationCMSC 714 Lecture 6 MPI vs. OpenMP and OpenACC. Guest Lecturer: Sukhyun Song (original slides by Alan Sussman)
CMSC 714 Lecture 6 MPI vs. OpenMP and OpenACC Guest Lecturer: Sukhyun Song (original slides by Alan Sussman) Parallel Programming with Message Passing and Directives 2 MPI + OpenMP Some applications can
More informationGEOMETRIC MANIFOLD APPROXIMATION USING LOCALLY LINEAR APPROXIMATIONS
GEOMETRIC MANIFOLD APPROXIMATION USING LOCALLY LINEAR APPROXIMATIONS BY TALAL AHMED A thesis submitted to the Graduate School New Brunswick Rutgers, The State University of New Jersey in partial fulfillment
More informationCS 450 Numerical Analysis. Chapter 7: Interpolation
Lecture slides based on the textbook Scientific Computing: An Introductory Survey by Michael T. Heath, copyright c 2018 by the Society for Industrial and Applied Mathematics. http://www.siam.org/books/cl80
More informationModeling the Spectral Envelope of Musical Instruments
Modeling the Spectral Envelope of Musical Instruments Juan José Burred burred@nue.tu-berlin.de IRCAM Équipe Analyse/Synthèse Axel Röbel / Xavier Rodet Technical University of Berlin Communication Systems
More informationLecture 15: More Iterative Ideas
Lecture 15: More Iterative Ideas David Bindel 15 Mar 2010 Logistics HW 2 due! Some notes on HW 2. Where we are / where we re going More iterative ideas. Intro to HW 3. More HW 2 notes See solution code!
More informationHistograms of Sparse Codes for Object Detection
Histograms of Sparse Codes for Object Detection Xiaofeng Ren (Amazon), Deva Ramanan (UC Irvine) Presented by Hossein Azizpour What does the paper do? (learning) a new representation local histograms of
More informationIntroduction to Geometric Algebra Lecture V
Introduction to Geometric Algebra Lecture V Leandro A. F. Fernandes laffernandes@inf.ufrgs.br Manuel M. Oliveira oliveira@inf.ufrgs.br Visgraf - Summer School in Computer Graphics - 2010 CG UFRGS Lecture
More informationCluster Analysis. Mu-Chun Su. Department of Computer Science and Information Engineering National Central University 2003/3/11 1
Cluster Analysis Mu-Chun Su Department of Computer Science and Information Engineering National Central University 2003/3/11 1 Introduction Cluster analysis is the formal study of algorithms and methods
More informationHomeomorphic Manifold Analysis (HMA): Generalized Separation of Style and Content on Manifolds
Homeomorphic Manifold Analysis (HMA): Generalized Separation of Style and Content on Manifolds Ahmed Elgammal a,,, Chan-Su Lee b a Department of Computer Science, Rutgers University, Frelinghuysen Rd,
More informationBasis Functions. Volker Tresp Summer 2017
Basis Functions Volker Tresp Summer 2017 1 Nonlinear Mappings and Nonlinear Classifiers Regression: Linearity is often a good assumption when many inputs influence the output Some natural laws are (approximately)
More informationParameter based 3D Optimization of the TU Berlin TurboLab Stator with ANSYS optislang
presented at the 14th Weimar Optimization and Stochastic Days 2017 Source: www.dynardo.de/en/library Parameter based 3D Optimization of the TU Berlin TurboLab Stator with ANSYS optislang Benedikt Flurl
More informationDimension Reduction of Image Manifolds
Dimension Reduction of Image Manifolds Arian Maleki Department of Electrical Engineering Stanford University Stanford, CA, 9435, USA E-mail: arianm@stanford.edu I. INTRODUCTION Dimension reduction of datasets
More informationLecture 7: Linear Regression (continued)
Lecture 7: Linear Regression (continued) Reading: Chapter 3 STATS 2: Data mining and analysis Jonathan Taylor, 10/8 Slide credits: Sergio Bacallado 1 / 14 Potential issues in linear regression 1. Interactions
More informationDigital Geometry Processing
Digital Geometry Processing Spring 2011 physical model acquired point cloud reconstructed model 2 Digital Michelangelo Project Range Scanning Systems Passive: Stereo Matching Find and match features in
More informationSupport Vector Machines
Support Vector Machines RBF-networks Support Vector Machines Good Decision Boundary Optimization Problem Soft margin Hyperplane Non-linear Decision Boundary Kernel-Trick Approximation Accurancy Overtraining
More informationA second order algorithm for orthogonal projection onto curves and surfaces
A second order algorithm for orthogonal projection onto curves and surfaces Shi-min Hu and Johannes Wallner Dept. of Computer Science and Technology, Tsinghua University, Beijing, China shimin@tsinghua.edu.cn;
More information2008 IEEE. Reprinted with permission.
Traian E. Abrudan, Jan Eriksson, and Visa Koivunen. 2008. Steepest descent algorithms for optimization under unitary matrix constraint. IEEE Transactions on Signal Processing, volume 56, number 3, pages
More informationMULTI-OBJECTIVE OPTIMISATION IN MODEFRONTIER FOR AERONAUTIC APPLICATIONS
EVOLUTIONARY METHODS FOR DESIGN, OPTIMIZATION AND CONTROL P. Neittaanmäki, J. Périaux and T. Tuovinen (Eds.) CIMNE, Barcelona, Spain 2007 MULTI-OBJECTIVE OPTIMISATION IN MODEFRONTIER FOR AERONAUTIC APPLICATIONS
More informationSELECTIVE ALGEBRAIC MULTIGRID IN FOAM-EXTEND
Student Submission for the 5 th OpenFOAM User Conference 2017, Wiesbaden - Germany: SELECTIVE ALGEBRAIC MULTIGRID IN FOAM-EXTEND TESSA UROIĆ Faculty of Mechanical Engineering and Naval Architecture, Ivana
More informationVideo event detection using subclass discriminant analysis and linear support vector machines
Video event detection using subclass discriminant analysis and linear support vector machines Nikolaos Gkalelis, Damianos Galanopoulos, Vasileios Mezaris / TRECVID 2014 Workshop, Orlando, FL, USA, November
More informationData Mining. CS57300 Purdue University. Bruno Ribeiro. February 1st, 2018
Data Mining CS57300 Purdue University Bruno Ribeiro February 1st, 2018 1 Exploratory Data Analysis & Feature Construction How to explore a dataset Understanding the variables (values, ranges, and empirical
More informationChapter 7: Computation of the Camera Matrix P
Chapter 7: Computation of the Camera Matrix P Arco Nederveen Eagle Vision March 18, 2008 Arco Nederveen (Eagle Vision) The Camera Matrix P March 18, 2008 1 / 25 1 Chapter 7: Computation of the camera Matrix
More informationParallel and Distributed Sparse Optimization Algorithms
Parallel and Distributed Sparse Optimization Algorithms Part I Ruoyu Li 1 1 Department of Computer Science and Engineering University of Texas at Arlington March 19, 2015 Ruoyu Li (UTA) Parallel and Distributed
More informationConstrained Aero-elastic Multi-Point Optimization Using the Coupled Adjoint Approach
www.dlr.de Chart 1 Aero-elastic Multi-point Optimization, M.Abu-Zurayk, MUSAF II, 20.09.2013 Constrained Aero-elastic Multi-Point Optimization Using the Coupled Adjoint Approach M. Abu-Zurayk MUSAF II
More informationTraining of Neural Networks. Q.J. Zhang, Carleton University
Training of Neural Networks Notation: x: input of the original modeling problem or the neural network y: output of the original modeling problem or the neural network w: internal weights/parameters of
More information