Numerical Computation Tema tteceevoleds cseteat ear ubetes SECOND EDITION C. Pozrikidis —Contents Preface 1 Computers and Computing Al 12 13 1d 15 Lé Lz 18 19 1.10 al 11 113 Analytical and mumerical computation Hardware and software . 1.24 Hardware . 1.2.2. Software The binary system 0. ‘The IP address. . Binary system arithmetic»... 0... Computer memory and addresses Computer programming LT.L Assembly code Sete 1.7.2. High-lovel languages. 0.2.0.2 ee 17.3. Binary executables L74 Recursive function calling 1.7.5 Evaluation of mathematical functions Plots, graphs, and data. visualization . Compiling in Unix ‘The floating point representation 1.10.1 Precision . 1.10.2 Round-off error : 1.10.3 Floating point operations . The hexadecimal and octal systems . Algorithms : 1.12.1 Maximnm and minjsnum of an array 112.2 Indexing . 112.3. Greatest cominon divisor . . 1.124 Numerical instability Algorithms and numerical efficiency 1.13.1 Nested multiplication by Horner's algorithin. 1.13.2 Sorting... 0... eee eae 13 16 18 20, 20 21 23 2 26 32 35 36 36 38 40 42 43 44 AG 48 48 491.13.3 Binary search . . 1.43.4 The AbKashi algorithm 1.18.5 Computing + eee eee 1.13.6 From easy to difficult probleins ee 1.14 Computer simulation . . 1.14.1 Dynamic and Monte Carlo simulation . 114.2 A Monte Carlo method for computing 2 wn 1.14.3 Buffon's needle . . LAA Assossing whether point lies inside a region in a plane L145 Assessing whether a point lics inside a volume in space 1.14.6 Numerical evidence . 1.15 Simulation of stochastic and random processes 4.15.1 Random walkers with constant steps 1.15.2. Random walkers with continuous steps 1.15.8 Random sequences with a specified pdf . 1.15.4 Box-Muller transformation»... 0.0... 1.16 Numerical error and its reduction see 1.16.1 Absolute and relative ertor 66... 1.16.2 Order of a method . + 116.3 Assessing the order of a method... . - 1.16.4 Error reduction by Richatdson oxtrapolation . Iterations, numerical sequences, and their convergence LITA Acecleration of convergence Numerical Matrix Algebra and Matrix Calculus 24 Matrices... 5. eee ee 2.11 Transpase of a matrix Lea 212 Adjoint of amotrix .. 0.0.2... 21.3 Vectors are slender matrices... 2... 214 Matrix algebra . 2.1.5 Vector transformations . . 2.41.6 Gram-Sclnnidt orthogonalization ». 2.1.7 Matrix functions and their derivatives 2.2 Square matrices . 2.2.1 Symmetric and skew- symmetric ‘ natrices: oe 2.2.2 Hermitian or self-adjoint matrices 2.2.3 Skew-Hermitian matrices... . 2.2.4 Matrices with a special structure 2.2.5 Positive-definite matrices... 2.2.6 Trace 2.3 Matrix inverse, eofactors, and the determinant 2.3.1 Matrix inverse 23.2 The Woodbury and Shetman-Morrison formulas 3 Inverse of a diagonal matrix 69 82 83 84 86 88 88 94 94 95 96 97 97 104 103 107 109 110 11 112 142 7 ig 120 120 121 12224 25 2.6 27 29 2.10 2.3.5 2.3.6 Orthogonal matrices... 2 : 2.3.7 Unitary matrices 238 Plane rotation with Givens matrices 2.3.9 Reflection with Houscholder matrices 3.10 Matrix functions Computation of the determinant Computation of the matrix inverse es boas 2.5.1 Inverse by columns : 25.2 Inverse from the LU decomposition : . . 2.5.3 Inverse from the QR. decomposition . vee vee 2.5.4 Improving an approximation to the inverse LU and related decompositions . . 2.6.1 Doolittle decomposition . , 26.2 LDU decomposition... ... . 26.3 Tridiagonal matrices 2.6.4 Crout decomposition . vee 2.65 Pivoting and Gauss climination ... . enna 2.6.6 Cholesky decomposition... . . : QR decomposition. . eee cence 2.7.1 QR decomposition with Gram-Schmidt orthogonalization, 2.7.2 QR decomposition by reflection 2.7.3 QR decomposition by Givens rotation Systems of linear algebraic captions : see 28.1 Cramer's rule . eee 2.82 Singular matrices and linear mapping . ee 2.8.3 Singular systems... 2... ces en 2.84 Homogeneous systems and null space Matrix eigenvalus bees 29.1 The characteristic polynomials... . boven eee 2.9.2 Relationship betwoen the eigenvalues, the trace, and the determinant 2.9.3 Powers, mverse, and functions of a matrix toe 294A Hermitian matrices 00. eee 2.9.5 Approximate location of eigenvalues . . . -. 2.9.6 Coefficients of the characteristic polynomial 2.9.7 Diagonal matrix of eigenvalues Bigenvectors and principal vectors... . « bee 2.10.1 Propertics of eigenvectors 2.10.2 Principal vectors . fee voces 2.103 Left eigenvectors... eae bv e eee 2.10.4 Matrix of eigenvectors 2.10.5 Normal matrices . . 2.10.6 Eigenvalues and cigenvectors of te adjoint Minors, cofactors, and the determinant : Inverse in terms of the cofactors and the matrix adjoint 124 125 126 126 127 128 130) 135) 138 138 140 wi a4 144 145 “ii 147 150 151 152 156 158 160 160 163 162 163 164 165 165 166 166 \67 a7 169 169 169 170 471 72 17224 igenvalues of positive-definite Hermitian matric: 172 2.11 Siailarity transformations, diagonalization, and the Jordan canonical form 173 2.11.1 Schur normal form bec e eee 175. 2.41.2 Similarity of tridiagonal matrices 175 2.11.3 Matrix diagonalization bette teens 176 2AL4 Jordan canonical form 6 es U7 2.41.5 Jordan projection matri 178 DULG The Jordan form is a vero of the characteristic polynomial... 180 2.12 Wielandt deflation... 2.43 Linear mapping . 2.13.1 Bvolution of a vector subject to successive mapping... . « : 186 2.44 Matrix functions... : 188 2141 The Cayley-Hamilton theorem sees 189 3.142 Characteristic polynomial based on a Krylov sequetice . +. 190 2.14.3 The exponential function Sees 191 Linear Algebraic Equations 193 3.1 Significance and applications 193 3.1.1 Forces on the elements of a crane truss . 194 3.1.2 Mass balance around a dual separation process « 195 3.4.3 Temperature distribution, in a plate 198 3.2 Diagonal and triangular systems 2. eee 206 3.2.1 Diagonal systems 207 3.2.2 Lower triangular systems and forward substitution... .... 0.00. 207 3.2.3 Upper triangular systems and backward substitution... . 208 3.24 Block-diagonal and block-triangular systems... 2... 209 3.3. General procedures and overview ones bones 210 3.3.1 Iterative methods... 0... 22 3.3.2 Solution improvement : 212 3.3.3 Sparse systems bees 213 3.3.4 Overdetermined systems : ait 335 Singular systems with an infinity of solutions... 24 3.36 Sensitivity and the condition number... 0 eee 216 3.3.7 Systems with complex numbers 27 3.4 Tridiagonal systems 218 341 Heat conduction distribution along a pipe... . eens 218 3.4.2 ‘Thomas algorithm Seven ee 220 343. Simplified Gauss elimination . « 224 3.5 Pentadiagonal systems 225 35.1 Beam bending 352 Generalized Thomas algorithm. . . 3.6 Cyclic rednetion 3.7 Nearly easy systems eee . . . 243.10 3.8.7 Iterative methods . Bat 3.9.2 3.9.3 3.9.4 3.9.5 3.9.6 3.9.7 3.9.8 3.9.9 3.9.10 Minimization and residual methods . . . 3.101 3.10.2 10.3 3.10.4 3.10.5 Sherman-Morrison decomposition Provisional Thomas algorithm jimination and related methods Pivoting Algorithm with row pivoting. . Preconditioning . Pivoting and permutation matrices Mnitiple right-hand sides Computation of the inverse . Gauss-Jordan reduction Fixed-point iterations for one equation Pixed-point iterations for a system of equations Jacobi method Gauss-Seidel method Weighted residual correction Weighted Jacobi iterations Successive overrelaxation (SOR) Richardson iteratioms . . . Operator and grid-based splitting : Acceleration of iterations by deflation . . . . Steepest-descent search... 02... ee Directional search . Conjugate gradients Biconjugate gradients Other methods . . Nonlinear Algebraic Equations ‘Mathematical and physical context. 41 42 43 44 4. 412 413 One equation . Two equations . Many equations Formalism and general procedures 4.2.1 4.2.2 Estimating the location of a root... . Estimating the roots of polynomials Bracketing methods Fixod-point iterations 6. e Aad 42 44.3 Iterated mapping, Fixed-point iterations for one equation Fixed-point iterations for a system of equations Newton’s and related methods for one equation 454 452 Secant method . . Muller's method 235 237 238 239 240 247 248, 249, 249 249 252 252 254 257 258 259 260 260 262 262 263 267 268 270 272 277 280 281 281 281 285, 286 288 290 291 298 294 294 296 302 306 310 3124.61 4.6.2 4.6.3 4.7 Roots of polynomials . ATA 472 47.3 Regula-falsi method Clubie and high-order convergence. - vton's and related methods for a system of equations... oe ees Newton iteration function Quasi-Newton methods Broyden’s method Bairstow’s method . . « Laguerre’s method bee bees Other methods... . Sees 4.8 Difficult problems, continuation, and embedding cee Eigenvalues of Matrices 5.1 Mathematical and physical context... 6... 3.1L Inclination of an elongated body 5.1.2 Vibration of bodies connected by springs... « 5.1.3. Eigenvalue problems from differential equations 5.4 Periodic solutions of Mathien’s equation... 6.6... 5.1.5 Complex matrices . : 5.18 Generalized eigenvalue problems . . 5.1.7 Nonlinear eigenvaite problems . 5.2. Diagonal, triangular, tridiagonal, ci culant, ad Hessenberg matrices onal and triangular matrices ‘Tiidiagonal matrices with constant diagonal elements ‘Tridiagonal matrices whose eigenvalues are roots of orthogonal polynomials Circulant matrices . Hessenberg matrices... 5.3 Roots of the characteristic polynomial . 5.34 32 SAL 5.5.1 5.5.2 Evaluation of the characteristic polynomial Hydrodynamic stability 5.4 The power method ...... Computing further eigenvalues imilarity transformations Reduction of symmetric matrices Reduetion of arbitrary matrices Function Interpolation and Differentiation 6.1 Interpolation condition and interpolation types 6. Vandermonde interpolation rir 6.3. Newton interpolation 64 Neville and Aitken interpolation, 641 64,2 Neville interpolation . . Aitken interpolation, 3i4 315 317 317 320 321 326 328 331 332 333 335 335 335 336 337 338, 342 BAT 349. 349 349) 352 B54 355 BST 358, 358 361 369 371 385 391 301 392 396, 398 398 40067 68 6.9 6.10 611 6.12 613, G14 6.16 617 Lagrange interpolation . . 65.1 Yauchy relations . 6.5.2 Representation in terms of the Lagrange generating polynomial 6.5.3 Evenly spaced data... . 6.5.4 — averse Lagrange interpolation 6.5.5 Cardinal interpolation finetions Properties of Lagrange polynomial 6.6.1 Detivatives 66.2 Node differentiation matrix 6.6.3 Representation in terms of the Vandermonde matri Inverse of the Vandermonde matrix Lagrange projection matrices : Error and convergence of polynomial interpolation 6.8.1 Interpolation error 68.2 Convergence Lagrange polynomials corresponding to polynomial roots Representation in terms of the generalized Vandermonde matrix 683 ‘The Lobesgue constant 2... s vette eee e es Optimal polynomial interpolation Lobatto interpolation . Scleetion of the interpolation variable Local polynoraiat interpolation 6.12.1 Lincar interpolation . . 6.12.2 Quadratic interpolation... . 6.12.3 Cubic interpolation Cubic-spline interpolation : 613.1 One interval... bones 6.13.2. Two intervals 6.13.3. Arbitrary number of intervals 6.13.4 Applications Lagrange interpolation for high-multiplity data. 6.14.1 Generalized Lagrange polynomials 6.14.2 Dual multiplicisy data : Lagrange projection matrices with multiple eigenvalues 6.15.1 Diagonalizable matrices... 00 ee ee 6.15.2 Nondiagonalizable matrices Hermite interpolation. . Leet e eee 6.16.1 Interpolating polynomial with matching slopes 6.16.2 High-order Hermite interpolation .......- Porametric representation of lines... 2... Lee 6.17.1 Global representation and interpolation... . 6.17.2 Local interpolation, cubic splines, and circular ares 6.17.3 Which method? 402 405 106 407 407, 408 a0 410 40 413 ald AS 416 AT 421 421 422 424 427 432 435 136 436 438 440 0 440 ddd 4a AB 456 456 459 460 AG2 462 465, 466 467 473 473 475 4816.18 Numerical differentiation 6.19 6.20 6.21 6181 6.18.2 6.18.3 6.18.4 618.5 6.19.1 6.19.2 6.19 Numerical partial differentiation 6.20.1 6.20.2 Surface interpolation 6.21.1 Formulas from linear interpolation Formulas from quadratic interpolation Formulas from high-order interpolation Tables of finite-difference formulas Finite-difference formulas from the node. differentiation matrix . Interpolation in two variables Bivariate interpolation through a Cartesian grid. Interpolation through a triangular grid Interpolation from scattered data Cartesian grids fee Totropic triangular grids... Representation with triangles Numerical integration 7 72 73 75 One-dimensional integrals . . Integration by local polynomial interpolation 721 Gauss-Legendire quadrature . 7.3.1 7.3.2 7.3.3 Trapezoidal rule... . Buler-Maclaarin formula : ‘Trapezoidal rule for periodic funetions Simpson's I/3tule.. 2... Simpson's 3/8, Boole’s, and higher-order rules. . Seemingly singular integrands beens Midpoint rule and open forrmulas . Integrands with singular derivatives Canonical form . . Lagrange interpolation Optimization . . . Gaussian integration quadratures TAL 7A2 748 TAA TAS 746 TA? TAs 74.9 TAA 74AL Computation of quadrature weights Quadrature weights based on optimality Formulas and tables of base points and weights Gauss-Legendre quadrature Gomss-Chebyshev quadrature. . « Ganss-log quadrature Gaugs-Laguerre quadrature... . . Generalized Gause-Laguerre quadrature Gauss-Hermite quadrature. . Ganss-Chebyshev quadrature of the Advenced quadratures Constrained integration quadratures < 482 483 484 485, 487 490 497 497 500, 504 507 507 Bui 512 Bld 517 BIT 523 52d 528 530 530 533 534 535 535, 538 538 539 339 546 548 549 549 549 550 550 552 555, 556 559 559 5607.5.1 Radan quadrature . . 7.5.2 Lobatto quadrature... 7.5.3 Chebyshev second-kind quadrature 7.6 Singular integrands . 7.6.1 Product-integration rules 162 Gauss Chebyshey quadrature for integrands with an x 7.6.3 Integrands with a logarithmic singularity 7.7 Integrals over infinite domains 7.8 Rapidly varying and oscillatory integrands 7.9 ¥redholm integral eanations : 79.1 N¥strom method . 2 Fredholm method»... . 79.3. Heat conduction ftom an isothermal body 7.10 Areal integrals fee 710.1 Product-integration rules... sy. scene 7.10.2 Integral over a rectangle 7.10.3 Integral over a quadrilateral /? singularity TAL Integral over a triangle TALL Flat triangles . 7.11.2 Curved triangles 7.12 Quadratures for miscellaneous shapes... es Function Approximation 8.1 Polynomial approximation 8.1.1 Taylor series approximation 8.4.2 Weierstrass approximation 8.2 Least-squares mininization beeen eee 8.2.1 Least-squares polynomials a Approximation with orthogonal polynomials. . . er approximation Bédior segments Bévier splines Bézier surfaces 8.4 B-spline approximation 8.4.1 Influence functions 8.4.2 Linear B-splines vee 84.3 Quadvatic Besplines .. 2... 8.44 Cubic B-splines 84.5 High-order B-splines . 8.4.6 B-spline represeutation of lines and surfaces 8.5. Padé approximation 8.6. ‘Dsigonometric approximation and interpolation. 8.6.1 Fourier projection 560 563 567 569) S74 S75 576 B77 580 582 583 584 587 588 589 592 593 593 602 603 606 e068 606 608 612 613 616 620 620 623, 623 624 625, 626 626 628 630 632 636, 640, 64187 88 8.9 Exponential fitting with Prony's method Fourier coefficients by the trapezoidal rule 8.6.5 Fast Fourier trans B71 B72 R72 874 8.7, KFT for N=2.., FFT for J FFT for } -. FFT for N=2" Implementation . . Fourier approximation of a fimotion of two variables BSL 8.8.2 Evenly spaced points Periodie fanctions Ordinary Differential Equations aL O38 94 25 Problem formulation and physical context... es 9.1.1 Vector fornmlation 9.1.2 Equations of higher order 9.1.3. Bvolution equations and dynamical systerns . 914 Start-up of a chemical reactor. . vee 9.1.5 Bvolution of an ecosystem 9.1.6 Phase space and solution orbits . fee 9.1.7 Phe Lorenz system 2 ee Initial-valne problems 9.2.1 Stability of an equilibrium point . 9.2.2 Ikconditioned or inherently unstable systems 9.2.3 Chaos . Linear autonomous systems Sees 9.3.1 Diagonal systems oe ee 9.3.2 ‘Triangular systems 9.3.3. Systems with diagonelizable matrices 9.3.4 Systems with nondiagonalizable matrices . . Numerical approximations 9.4.1 Buler method 9.4.2 lnplicit Buler meshod 9.4.3 Crank-Nicolson method bones 9.4.4 Modified Euler method (RK2)-. 0.0. 94.5 Third-order Runge-Kutta method (RK3) 94.6 Fourth-order Runge-Kutta method (RK4) ) 9.4.7 Midpoint method 9.4.8 Second-order Adams-Bashforth method ‘Time stepping and successive mapping . . « 95.1 Numerical diffusion or damping 643, AT 619 652 652 654 a4 637 660 662 665, 666 670 671 678 678 679 680 683 683 685 686 688 692 692 693 694, 694 695 695 695 702 709 710 Tu 7 n2 713 715 75 716 m7 799.5.2 Numerical instability 9.5.3 Stability graphs 9.5.4 Stiff equations 9.6 Integration of nonlinear equations... 2... 9.6.1 Euler's incthod 9.6.2 Iruplicit and semt-implieit Buler methods 9.6.3 Crank-Nicolson method... 0... 9.6.4 Adams-Bashforth methods ee 9.6.5 Adats-Moulton methods... 2... 9.6.6 Midpoint methods... -. 2... Stérmer-Verlet algorithm for second-order systems Other methods... 2. 9.7 Time stepping and algebraic mapping 9.8 Predictor-corrector methods 9.81 Multipoint and Runge-Kutta methods . : 736 9.8.2 Adams-Bashforth-Moulton methods eee 736 98.3 Runge-Kutta methods... . veces 737 9.84 Socond-order Runge-Kutta method (RK2). . . eee 738 9.8.5 Third-order Runge-Kutta method (RK3) .. . . bones 739 9.8.6 Fourthorder Runge-Kutta method (K4) be nee 744 9.9 A particle straddling a fluid interface oe . 749 9.10 Error estimate and adaptive control. . . . 760 9.10.1 Runge-Kutte-Fehlberg scoond and third order methods (RKF28) . 760 9.10.2 Ruge-Kutta-Feblberg fourth and fifth order methods (RKF45) 762 9.11 Stiff equations . 9.11.1 Numerical methods for stiff systems |... eee coe 64 9.12 Boundary-value, free-boundary, and eigenvalue problems... 6... « 766 9.12.1 Two-point bomdary-value problems 766 9.12.2. Separated boundary conditions . . . 767 9.12.3 Nonseparated boundary conditions 770 9.12.4 Numerical methods 77 9.12.5 Free-boundary problems eens TT 9.12.6 Bigenvalue problems... 2.0... ee cee 773 9.18 The shooting method... . . Sec e eee ee TTA 9.13.1 A bubble or liquid drop on @ swface nT) 9.13.2 Implementation of Newton's method bees 783 913.3 Sealing 2.0... : 787 9.134 Linear systems and superposition beeen eee co 788 9.13.5 Partial inverse shooting bocce ee 789 9.13.6 Unseparatcd boundary conditions . 790 10 Discretization Methods tor Ordinary Differential Equations 792 10.1 Finite-difference methods bee e eee 792 W.L.1 Implementation10.1.2 To beat a dram with a. finite-difference method cee e e796 10.1.3 Nonlinear cquations 6.2... . . 799 10.2 Multigrid methods 800 10.2.1 Two-grid scheme . sul 10.2.2. Poisson equation in one ditnension 802 10.2.3 Smoothing efficiency . 804 16.2.4 Eficetive projection inatrix of the two-grid scheme 810 Multigrid V cycle 815, Explicit V cycle multigrid code « See ee S16 Recursive V cycle multigrid code . 322 Woyeles oo. ccc 825 Full multigrid cycle (PMG) 825 10.3 Finite-volume motheds oe 826 10.3.1 Element discretization . 826 10.3.2 Finite-volune formulation : beets 832 10.4 Weightod-residual methods fee +. 834 10.4.1 Galerkin and Lanczos’s tau methods fees eee 835 10.4.2 Spectral metisods . bees : B39 10.4.3. Collocation methods foe bee St 10.5 Finive-element methods... 0... 00.50. beeen S45 10.5.3. Linear element interpolation voces 846 19.5.2. Element grading Seve MT 10.5.3. Galerkin projection . sees Lo 8H7 10.5.4 Formulation of a linear algebraic system sees S48, 10.5.5 Relation to the finite difference method tone 851 10.5.6. Flux at the Dirichlet end 852 105.7 Galerkin finite-element equations by way of Dirac’s dehta function 853 10.5.8 Element diffusion and mass matrices bees 8d 10.5.9 Assembly cbt e eee 858 10.5.10 Finite-element eode occas 858 10.5.L1 Quadratic and spectral clements 2 ee BGM 11 Finite-Difference Methods for Partial Differential Equations 865 11.1 Convection diffusion -reaction equation... . . 865 11.1.1 Classification Sees 867 11.1.2 Initial and boundary conditions... os eee . 867 1.3 Hea 868 1L14 Vectorial convection-diffusion 869 11.1.5 Origin of the Laplacian 369 nite-difference formulation . - tee tee vee 87h 121 Conservative form... 2s . 874 11.3 Grid generation 876 11.3.1 Conforined mapping . 879 113.2 Transformation of governing equations . +. 8884 115 16 8 11.3.3. Numerical grid generation... . . Explicit methods for one-dimensional unsteady diffusion 114.1 Reduction to ODEs... -. 2... 11.4.2. Time-space discretization 114.3. PLCS discretization . Ld CTCS (leap-trog) discretization 114.5 The Du Fort and Frankel method. 114.6 Summa: 114.7 Heat conduction across two layers. : Implicit methods for one-dimensional unsteady diffusion 11.5.1 Implicit BI'CS or Laasonen method . 11.5.2. The Crank-Nicolson method 11.5.8. Three-level implicit methods . . Explicit methods for unsteady diffusi 11.6.1 Fully explicit FTCS discretization 11.6.2. Temperature distribution in a chimney wall 11.6.3. Temperature distribution across a cooling rod in mieroclectronies 11.6.4 Assessment. of explicit methods : Implicit methods for unstendy diffusion in two and three dimensions 11.7.1 BCS discretization tee 11.7.2 Crank-Nicolson discretization. . 11.7.8. ADI method in two dimensions . 11.7.4 ADI method in three dimensions Operator splitting and fractional steps Finite-volume method m in two and three dimensions 12 Finite-Difference Methods for the Poisson and Laplace Equations 12.1 12.2 12.3 12.4 12.5 12.6 12.7 128 12.9 Finite-differenec implementation 12.1.1 Direct solution 121.2 ‘Transient approach to steady state. Poisson equation in a rectangular domain Mollifed color function for an immersed boundary Steady temperature distribution across a chimney wall Fluid flow in a cavity Solution in curvilinear coordinates Shear flow over a buropy wall. Nonlinear Poisson equation Multigrid methods . . 12.9.1 ‘Two-grid method 12.9.2 Prolongation 12.9.3. Restriction Multigrid V cycle 940 040 940. wa 944 958 967 973 982 989 997 1006 1006 1008 1008 100913 Finite-Difference Methods for Convection—Diffusion 1017 13.1 One-dimensional convection oa see ee ee 107 13.41 Particle method... : ras eee ORL 13.1.2 Shocks 2... . cee 1922 13.2. Finite-difference methods seca eo 1025 13.2.1. Linear convection - Stee eee 1026 13.2.2 Explicit FTCS discretization |. . cee . 1026 13.2.3 FTBS method . ee 1029 13.24 FTVS method : 1082 13.2.5 Upwind differencing, numerical diffusivity, and the OFL condition... 1032 13.2.6 Numerical diffusion . 13.2.7 Numerical cone of influence... « 13.28 Lax’s modification of the PTOS method . 13.2.9 The Lax-Wendroff method... . 13.2.10 Explicit CTS or leapfrog method . 13.2.11 Implicit BTCS 13.2.12 Crank-Nicolson method . . 13.218 Summary... ees 13.214 Modified dynamics and explicit numerical diffusion 13.2.15 Multistep methods... . 13.3 Nonlinear convection bese 18.8.1 Burgers equation... ++... Finite-volume method... . . « 13.4 Convection in two and three dimensions. 134.1 Fully explicit FTCS discretization 134.2 Lax'smethod . 00.6.6 6s 134.3. Implicit methods 13.4.4 Operator splitting and fractional steps 13.5 Explicit methods for convection-diffusion in one dimension 13.5.1 FTCS discretization . 13.5.2 Upwind differencing 13.5.3 Generalized upwind differencing 13.5.4 Higher-order methods 13.5.5 Explicit CTCS and the Du Fort and Frankel method 13.6 Implicit methods for convection -diffusion in one dimension 13.6.1. BCS discretization 13.6.2. Crank-Nicolson method 13.6.3. ‘Two-level implicit methods... . beeen 13.6.4 ‘Three-level fully implicit methods o.oo eae 13.7 Multistep and relpted methods for convection-diffusion in one dimension 13.7.1 Operator splitting and fractional stops 13.7.2. Hopscotch 12.7.8 Nonlincar convection diffusion13.8 Convection-diffusion in two and three dimensions : 1062 13.8.1. Unstendy flow in a eavity : : 1064 13.9 Computational fluid dynamics . . ve ae wee . 1066 13.9.1 Alternative system of governing equations : . 1067 13.9.2. Pressure boundary conditions toe ce 1068 13.9.3 Compatibility condition for the pressure... ee 1068 13.94 Operator splitting and solenoidal projection. . cone 1068 13.9.5 Convection-diffusion step 1070 13.9.6 Projection step. beens 1071 13.9.7 Boundary conditions for the intermediate velocity... 0.2.0.0... 1073 13.9.8 Flow in acavity .... . Sees 1073 14 Boundary-Element Methods 1081 141 Laplace's equation in two dimensions . 1082 14.1.1 Green's first and second identities and the reciprocal relation... ss... 1082 14.1.2 Green’s functions. . Pec e 1083 14.1.3 Froe-space Green's function . . bees sees 1084 14.1.4 Integral properties of the Groen’s functions... sss... se ee L085 14.1.5 Integral representation . oe see 1085 14.1.6. Green’s third identity vee 1086 14.1.7 Integral equations . . . . bee eee a. 1086 14.8 Boundary-element collocation method... . . 1088 14.2 Contour discretization eee : - 1088 14.2.1 Linear elements . . . cee ee ees 1089 14.2.2 Circular elements... bees eee . 1092 14.2.3 Discretization code Secs 1096 14,3 Influence coefficients. eee an . 1099 14.4 Temperature distribution across a chimney wall... 6... . 104 15 Finite-Element Methods for Partial Differential Equations 1113 15.1 General framework beeen eee seve ee UB 15.2 Convection-diffusion in two dimensions fee / 1d 15.2.1 Galerkin projection . : 1116 15.2.2 Domain discretization and inte: cctv eee eee HIT 6.28 GFEM equations... . a Stee ee HLS 5.2.4 Implementation of the Dirichlet boundary condition se 1120 168 Then triangles 6. eee : vee TERR 15.3.1 Element matrices ee ee bee eee ee 1123 15.4 Grid generation nn : 1127 15.4.1 Successive subdivisions 5... 6... teen ee LBP 15.4.2 Delaunay triangulation. : bees oe. 1132 Was Generalized connectivity mattices cee 1135 15.4.4 Element and node numbering... « : : 1136 15.5 Code for Laplace's equation . : -.. 118A Calculus Primer 1 Functions of one real variable 2 Functions of two variables . AS Dirac delia function B Orthogonal Polynomials B.1_ Definitions and basic properties Bad B12 Clenshaw's algorithm B13. Orthonormal polynomials B14 ChristoffelDarboux formula. B.1.5 Summary of orthogonal families. BG Gram-Schmidt orthogonalization B.2 Legendre polynomials... . 5 Sees B.3 Radaw polynomials Bd Lobatto polynomials B.5. Chebyshev polynomials B.6 Chebyshev polynomials of tie second kind . . , 33.7 Jacobi polynomials B.S Laguerre polynomials B.9 Hermite polynomials . . . B.i0 Gram polynomials Unix Primer C++/Matlab/Fortran 77 Dictionary ASCH Code “mo 0 Matlab Primer FA tnvoking Matlab oo. 0.00.00. F.2. Matlab programming . . . V3 Grammar and syntax... . FA Precision F.5. Matlab commands F.6 Matlab examples ¥.7 Matlab functions F.8 User-defined fimetions . F.9 Matlab graphics References index ‘valuation as the determinant of a tridiagonal matrix 1143 1143, 1146 1146 1150 1150 116 1155 1156 1156 1157 LES. 1165 L167 1169 17 1173 1175 177 1178 1179 1181 1185 1195 1199, 1199 1200 1200 1204 1204 1205 1208 1212 1213" 1220 1231