Numerical Method for Blasius Equation on an infinite Interval

Similar documents
A) is empty. B) is a finite set. C) can be a countably infinite set. D) can be an uncountable set.

62. Power series Definition 16. (Power series) Given a sequence {c n }, the series. c n x n = c 0 + c 1 x + c 2 x 2 + c 3 x 3 +

Numerical Solution of the Two Point Boundary Value Problems By Using Wavelet Bases of Hermite Cubic Spline Wavelets

NUMERICAL METHOD FOR SINGULARLY PERTURBED DELAY PARABOLIC PARTIAL DIFFERENTIAL EQUATIONS

Topics. Homework Problems. MATH 301 Introduction to Analysis Chapter Four Sequences. 1. Definition of convergence of sequences.

A 2nTH ORDER LINEAR DIFFERENCE EQUATION

A NEW CLASS OF 2-STEP RATIONAL MULTISTEP METHODS

3 Gauss map and continued fractions

MAT1026 Calculus II Basic Convergence Tests for Series

Math 257: Finite difference methods

Existence of viscosity solutions with asymptotic behavior of exterior problems for Hessian equations

PAPER : IIT-JAM 2010

A NOTE ON BOUNDARY BLOW-UP PROBLEM OF u = u p

A collocation method for singular integral equations with cosecant kernel via Semi-trigonometric interpolation

NEW FAST CONVERGENT SEQUENCES OF EULER-MASCHERONI TYPE

APPROXIMATION BY BERNSTEIN-CHLODOWSKY POLYNOMIALS

Poincaré Problem for Nonlinear Elliptic Equations of Second Order in Unbounded Domains

PRELIM PROBLEM SOLUTIONS

Complex Analysis Spring 2001 Homework I Solution

Solutions to Homework 7

Chapter 4. Fourier Series

Exact Solutions for a Class of Nonlinear Singular Two-Point Boundary Value Problems: The Decomposition Method

INVERSE THEOREMS OF APPROXIMATION THEORY IN L p,α (R + )

lim za n n = z lim a n n.

Chapter 4 : Laplace Transform

b i u x i U a i j u x i u x j

Math 113 Exam 3 Practice

A General Iterative Scheme for Variational Inequality Problems and Fixed Point Problems

Common Coupled Fixed Point of Mappings Satisfying Rational Inequalities in Ordered Complex Valued Generalized Metric Spaces

METHOD OF FUNDAMENTAL SOLUTIONS FOR HELMHOLTZ EIGENVALUE PROBLEMS IN ELLIPTICAL DOMAINS

Math 113 Exam 3 Practice

Chapter 10: Power Series

Convergence of random variables. (telegram style notes) P.J.C. Spreij

DETERMINATION OF MECHANICAL PROPERTIES OF A NON- UNIFORM BEAM USING THE MEASUREMENT OF THE EXCITED LONGITUDINAL ELASTIC VIBRATIONS.

Sequences and Series of Functions

Integrable Functions. { f n } is called a determining sequence for f. If f is integrable with respect to, then f d does exist as a finite real number

-ORDER CONVERGENCE FOR FINDING SIMPLE ROOT OF A POLYNOMIAL EQUATION

University of Colorado Denver Dept. Math. & Stat. Sciences Applied Analysis Preliminary Exam 13 January 2012, 10:00 am 2:00 pm. Good luck!

Taylor expansion: Show that the TE of f(x)= sin(x) around. sin(x) = x - + 3! 5! L 7 & 8: MHD/ZAH

6.3 Testing Series With Positive Terms

Subject: Differential Equations & Mathematical Modeling-III

The Phi Power Series

TMA4205 Numerical Linear Algebra. The Poisson problem in R 2 : diagonalization methods

CSE 1400 Applied Discrete Mathematics Number Theory and Proofs

Sequences. A Sequence is a list of numbers written in order.

Chapter 6 Infinite Series

On forward improvement iteration for stopping problems

Subject: Differential Equations & Mathematical Modeling -III. Lesson: Power series solutions of Differential Equations. about ordinary points

CS537. Numerical Analysis and Computing

Infinite Sequences and Series

Numerical Solutions of Second Order Boundary Value Problems by Galerkin Residual Method on Using Legendre Polynomials

Inverse Nodal Problems for Differential Equation on the Half-line

THE REPRESENTATION OF THE REMAINDER IN CLASSICAL BERNSTEIN APPROXIMATION FORMULA

f(w) w z =R z a 0 a n a nz n Liouville s theorem, we see that Q is constant, which implies that P is constant, which is a contradiction.

1 Lecture 2: Sequence, Series and power series (8/14/2012)

The Numerical Solution of Singular Fredholm Integral Equations of the Second Kind

Lecture 8: Solving the Heat, Laplace and Wave equations using finite difference methods

CHAPTER 10 INFINITE SEQUENCES AND SERIES

CS321. Numerical Analysis and Computing

LOWER BOUNDS FOR THE BLOW-UP TIME OF NONLINEAR PARABOLIC PROBLEMS WITH ROBIN BOUNDARY CONDITIONS

Sequence A sequence is a function whose domain of definition is the set of natural numbers.

Math 25 Solutions to practice problems

Weighted Approximation by Videnskii and Lupas Operators

Find quadratic function which pass through the following points (0,1),(1,1),(2, 3)... 11

Oscillation and Property B for Third Order Difference Equations with Advanced Arguments

On Exact Finite-Difference Scheme for Numerical Solution of Initial Value Problems in Ordinary Differential Equations.

Time-Domain Representations of LTI Systems

arxiv: v1 [cs.sc] 2 Jan 2018

On general Gamma-Taylor operators on weighted spaces

MAS111 Convergence and Continuity

Math 341 Lecture #31 6.5: Power Series

IJITE Vol.2 Issue-11, (November 2014) ISSN: Impact Factor

DECOMPOSITION METHOD FOR SOLVING A SYSTEM OF THIRD-ORDER BOUNDARY VALUE PROBLEMS. Park Road, Islamabad, Pakistan

Seunghee Ye Ma 8: Week 5 Oct 28

A STUDY ON MHD BOUNDARY LAYER FLOW OVER A NONLINEAR STRETCHING SHEET USING IMPLICIT FINITE DIFFERENCE METHOD

x x x Using a second Taylor polynomial with remainder, find the best constant C so that for x 0,

We are mainly going to be concerned with power series in x, such as. (x)} converges - that is, lims N n

Finite Difference Approximation for Transport Equation with Shifts Arising in Neuronal Variability

Positive Solutions for Singular m-point Boundary Value Problems with Sign Changing Nonlinearities Depending on x

CHAPTER 1 SEQUENCES AND INFINITE SERIES

(A sequence also can be thought of as the list of function values attained for a function f :ℵ X, where f (n) = x n for n 1.) x 1 x N +k x N +4 x 3

MATH4822E FOURIER ANALYSIS AND ITS APPLICATIONS

Dirichlet s Theorem on Arithmetic Progressions

L 5 & 6: RelHydro/Basel. f(x)= ( ) f( ) ( ) ( ) ( ) n! 1! 2! 3! If the TE of f(x)= sin(x) around x 0 is: sin(x) = x - 3! 5!

Measure and Measurable Functions

Chapter 8. Uniform Convergence and Differentiation.

ON WELLPOSEDNESS QUADRATIC FUNCTION MINIMIZATION PROBLEM ON INTERSECTION OF TWO ELLIPSOIDS * M. JA]IMOVI], I. KRNI] 1.

Analytic Continuation

Boundary layer problem on conveyor belt. Gabriella Bognár University of Miskolc 3515 Miskolc-Egyetemváros, Hungary

Sequences of Definite Integrals, Factorials and Double Factorials

V. Adamchik 1. Recursions. Victor Adamchik Fall of x n1. x n 2. Here are a few first values of the above sequence (coded in Mathematica)

} is said to be a Cauchy sequence provided the following condition is true.

Modified Decomposition Method by Adomian and. Rach for Solving Nonlinear Volterra Integro- Differential Equations

Chapter 2 Systems and Signals

An Analysis of a Certain Linear First Order. Partial Differential Equation + f ( x, by Means of Topology

SOLUTIONS TO EXAM 3. Solution: Note that this defines two convergent geometric series with respective radii r 1 = 2/5 < 1 and r 2 = 1/5 < 1.

NTMSCI 5, No. 1, (2017) 26

Solutions to HW Assignment 1

Higher-order iterative methods by using Householder's method for solving certain nonlinear equations

Please do NOT write in this box. Multiple Choice. Total

Transcription:

Numerical Method for Blasius Equatio o a ifiite Iterval Alexader I. Zadori Omsk departmet of Sobolev Mathematics Istitute of Siberia Brach of Russia Academy of Scieces, Russia zadori@iitam.omsk.et.ru 1 Itroductio Blasius problem o a half-ifiite iterval is cosidered. This problem has a place uder mathematical modellig of viscid flow before thi plate. Blasius problem is a boudary value problem for a oliear third order ordiary differetial equatio o a half-ifiite iterval. This problem was ivestigated i may articles. For example, G.I. Shishki [1] studied asymptotic behavior of differetial ad differece solutios to get differece scheme with a fiite umber o odes for eough log iterval. We apply the method of markig of set of solutios, that satisfy the limit boudary coditio at ifiity to trasform a problem uder cosideratio to a problem for a fiite iterval [2]. To use that method we have problems, coected with oliearity of differetial equatio ad with ubouded coefficiet before secod derivative. We offer to do trasformatio of idepedet variable to avoid a problem, coected with boudless of that coefficiet. We cosider a Blasius problem as model oliear problem for applicatio of developig techic. 2 Case of a liear problem Cosider a problem: εu (x) + [a(x) + x]u (x) = f(x), (1) u(0) = A, lim u(x) = 0. (2) Suppose,that a(x) + x 0, ε > 0, fuctios a(x), f(x) are smooth eough, lim a(x) = a 0, lim f(x) = 0. Note, that domai is ubouded ad solutio u(x) ca be ubouded for a bouded fuctio f(x). Itegratig equatio (1), we get, that if a(x) + x α > 0, the u(x) A + 1 α x 0 f(s) ds. Supported by the Russia Foudatio of Basic Research uder Grat 04-01-00578 1

To avoid uboudedess of coefficiet before first derivative, itroduce ew variable t = x2 2. The problem (1)-(2) become a form: εu (t) + b(t)u (t) = F (t), where u(0) = A, lim u(t) = 0, (3) b(t) = a( 2t) 2t + 1 + 1 2t, F (t) = f( 2t)/(2t). So, coefficiet b(t) is bouded for t, separated from zero. We ca do some other replacemet of x to do fuctio b(t) uiformly bouded, but it is ot importat for method, cosidered below. By the ext equatio we mark solutios of differetial equatio (3), that satisfy the limit boudary coditio at ifiity [2, 3, 4]: εu (t) + g(t)u(t) = β(t), (4) where g(t) is solutio of sigular Cauchy problem for Riccati equatio: εg (t) + b(t)g(t) g 2 (t) = 0, lim g(t) = 1, (5) β(t) is solutio of sigular problem: εβ (t) + [b(t) g(t)]β(t) = εf (t), lim β(t) = 0. (6) Thak to coditio lim g(t) = 1 every solutio of equatio (4) teds to zero at ifiity ad equatio (4) picks that solutios of differetial equatio (3), that satisfy the limit boudary coditio at ifiity. Lemma 1 Let 0 < b 1 b(t) b 2. The b 1 g(t) b 2. Proof. At first prove, that g(t) > 0, v(t) = exp Obviously, v(t) is solutio of a problem: It follows from (7), that 0 t <. Defie t 0 ε 1 g(s)ds. εv + b(t)v (t) = 0, v(0) = 1, lim v(t) = 0, (7) v(t) > 0, v (t) < 0, 0 < t <. Takig ito accout that g(t) = εv (t)/v(t), we get g(t) > 0, 0 t <. Prove, that g(t) b 2. Cosider a problem: εg 2 + b 2 g 2 g 2 2 = 0, lim g 2 (t) = b 2 2

with solutio g 2 (t) = b 2. Let z = g 2 g. The z(t) is solutio of a problem: εz (t) [g 2 b(t) + g(t)]z(t) = (b(t) g 2 )g 2 0, lim z(t) = b 2 b 0. (8) If we suppose, that for some s z(s) < 0, the we shall have cotradictio. g(t) b 2. I the same maer we ca prove, that g(t) b 1. Lemma is proved. Usig a problem (6), we ca prove, that It follows, that β g 0 F (s) ds. Usig equatio (4), trasform a problem (3) to a problem for a fiite iterval: εu (t) + b(t)u (t) = f(t), 0 < t < L, u(0) = A, εu (L) + g(l)u(l) = β(l). (8) Prove, that problems (3) ad (8) have a same solutio for 0 t L. Cosider iitial value problem: εu (t) + g(t)u(t) = β(l), u(0) = A. (9) Takig ito accout equatios (5),(6), we prove, that solutio of a problem (9) satisfies to problems (3) ad (8). Problems uder cosideratio have uique solutio, therefor problems (3) ad (8) have a same solutio for 0 t L. So, problem (3) is exactly reduced to a problem (8), formulated for a fiite iterval. We have oly to fid coefficiets g(l) ad β(l). Fuctios g(t), β(t) as solutios of problems (5),(6) we ca sick as asymptotic series: g(t) ε g (t), β(t) ε β (t), (10) or i a form: I case of represetatio (11) we suppose, that g(t) (2t) /2 g, β(t) (2t) /2 β. (11) a( 2t) = f( 2t) = ( (2t) /2 a + O (2t) (N+1)/2), ( (2t) /2 f + O (2t) (N+1)/2). I both cases (10) ad (11) we got recurret formulas o β, g. I particular, i case of represetatio (11) we have: g 0 = 1, g 1 = a 0, g 2 = 1 + a 1, g 3 = εa 0 + 4a 0 + a 2 + 4a 0 a 1, β 0 = 0, β 1 = f 1, β 2 = f 2 β 1 ε 1 (g 3 a 2 ). 3

Make iverse trasformatio of idepedet variable ad get a ext problem for a fiite iterval: εu (x) + [a(x) + x]u (x) = f(x), 0 < x < L 0, ε u(0) = A, u (L 0 ) + g(l)u(l 0 ) = β(l 0 ). (12) L 0 Usig maximum priciple, we ca prove ext lemma. Lemma 2 Let ũ(t) is solutio of problem (12) with perturbed coefficiets g(l 0 ), β(l 0 ), g(l 0 ) g(l 0 ), β(l 0 ) β(l 0 ), g(l 0 ) σ > 0. The u(x) ũ(x) σ 1 ( u(l 0 ) + 1). Cosider a problem for umerical experimets: [ ] 2x u (x) + x + 1 + x u (x) = 0, u(0) = 1, lim u(x) = 0. We compare differet approaches for formulatio of boudary coditio at a fiite poit istead the limit boudary coditio at ifiity. Let = max vh ṽ, h where ṽ h is solutio of the scheme of upwid differeces for eough log iterval [0, L 0 ], L 0 = 100; v h is solutio of the same differece scheme o a shot iterval depedig o boudary coditio. I Table 1 error is represeted for differet approaches ad itervals. 3 Blasius problem Cosider Blasius problem: Let Table 1: Absolute errors for a liear problem L w (L) = 0 g (L) g 0 g 0 + g 1 /L g 0 + g 1 /L + g 2 /L 2 2 0.12e 01 0.95e 02 0.11e 02 0.44e-04 3 0.35e 03 0.20e 03 0.19e 04 0.41e-05 5 0.36e 07 0.13e 07 0.95e 09 0.84e-10 7 0.32e 12 0.86e 13 0.49e 14 0.23e-15 u (x) + u(x)u (x) = 0, u(0) = 0, u (0) = 0, u(x) = v(x) + x, The problem (13) ca be writte i a form: lim u (x) = 1. (13) w(x) = v (x). v (x) = w(x), v(0) = 0, 4

w (x) + [v(x) + x]w (x) = 0, w(0) = 1, Cosider iterative method for a problem (14): Lemma 3 Let The iterative method (15) coverges. lim w(x) = 0. (14) v (x) = w (x), v (0) = 0, w (x) + [v 1 (x) + x]w (x) = 0, 0 < x <, w (0) = 1, lim w (x) = 0. (15) v 0 (x) v(x), v 0 (x) v 1 (x), x > 0. Proof. Prove, that for every value of x sequeces v (x), w (x) are mootoe decreasig ad have low bouds. Low bouds. Let z (x) = v (x) v(x), p (x) = w (x) w(x). Compose a problem: z (x) = p (x), z (0) = 0, p (x) + [v 1 (x) + x]p (x) = z 1 (x)w (x), 0 < x <, p (0) = 0, lim p (x) = 0. By iductio prove, that for all x > 0 z (x) 0, p (x) 0. Accordig to coditios of lemma z 0 (x) 0. Let z 1 (x) 0. Prove, that p (x) 0. Suppose, that for some s p (s) < 0. The there is poit of egative miimum of fuctio p (x), it leads us to cotradictio. So, p (x) 0. It follows, that z (x) 0. Accordig to method of mathematical iductio for all z (x) 0, p (x) 0. It implies, that v (x) v(x), w (x) w(x). Mootoy. Prove, that sequeces v (x), w (x) are mootoe decreasig. Let z (x) = v (x) v +1 (x), p (x) = w (x) w +1 (x). The z (x) p (x) are solutios of a problem: z (x) = p (x), z (0) = 0, p (x) + [v 1 (x) + x]p (x) = z 1 (x)w +1(x), 0 < x <, p (0) = 0, lim p (x) = 0. Usig method of iductio, prove, that z (x) 0, p (x) 0. Accordig to coditios of lemma z 0 (x) 0.. It is follows from equatio o p (x), that if z 1 (x) 0, the p (x) 0 for x > 0. Coditio p (x) 0 implies, that z (x) 0. So, usig method of mathematical iductio, we proved, that for every z (x) 0, p (x) 0. We proved, that for every x > 0 sequeces v (x), w (x) are mootoe decreasig ad have low bouds. It s kow, that i this case sequeces uder cosideratio have the property of covergece. Lemma is proved. Cosider a case v 0 (x) = 0. Prove, that coditios of a lemma 3 are fulfilled. First coditio has a place, because v(x) 0. Verify the coditio v 0 (x) v 1 (x), x > 0. It follows from (15), that w 1 (x) 0 for x > 0. It implies, that v 1 (x) 0 = v 0 (x). So, coditios of lemma 3 are valid. 5

Cosider a questio of reductio Blasius problem to a fiite iterval. For every fixed problem (15) is liear ad we ca use results, obtaied i liear case. Secod equatio i (15) correspods to (1) with a(x) = v 1 (x). Trasformed to a fiite iterval problem (15) has a form: w v (x) = w (x), v (0) = 0, (x) + [v 1 (x) + x]w (x) = 0, 0 < x < L, 1 w (0) = 1, L w (L) + g (L)w (L) = 0. (16) Coefficiet g (L) ca be calculated o base of asymptotic series (11) as it was discussed for liear case. Cosider results of umerical experimets. Write differece scheme for a problem (16): w k+1 v k = v k 1 2w k + w k 1 h 2 w 0 = 1, 1 L + wk 1 + w k h, v 0 = 0, 2 + [v 1 k + x k ] wk+1 w k h w K w K 1 h + g (L)w K = 0, = 0, k = 1, 2,..., K 1. (17) Compare errors, correspodig to differet approaches for formulatio of boudary coditio istead of limit boudary coditio at ifiity. Let = max vh ṽ, h where ṽ h is solutio of problem (17) for eough log iterval [0, L 0 ], L 0 L, whe error of limit boudary coditio trasfer to poit L 0 is iessetial. Let L 0 = 100, h = 0.1 We cotiue iteratios, if max v k v 1 k > δ, δ = 10 13. Iitial iteratio is defied as v 0 (x) = 0. I Table 2 k error is preseted for differet L ad g (L) i compare with coditio w (L) = 0. I this Table g 0 = 1, g 1 = v (L). We have similar results i case h = 0.01. Table 2: Absolute errors for Blasius problem L w (L) = 0 g (L) = g 0 g (L) = g 0 + g 1 /L 3 0.83e 1 0.24e 1 0.19e 1 5 0.76e 3 0.15e 3 0.47e 4 7 0.34e 6 0.53e 7 0.97e 8 9 0.12e 10 0.15e 11 0.18e 12 It follows from umerical experimets, that usig special boudary coditio, based o equatio (4), we get more accurate results i compare with classical approach. Author thaks N.B. Koyukhova for useful proposals, that were took ito accout. 6

Refereces [1] G.I. Shishki Grid Approximatio of the Solutio to the Blasius Equatio ad of its Derivatives, Computatioal Mathematics ad Mathematical Physics, 41, 1, 37-54 (2001). [2] A.A. Abramov ad N.B. Koyukhova Trasfer of Admissible Boudary coditios From a Sigular Poits of Liear Ordiary Differetial Equatios, Sov. J. Numer.Aal. Math. Modellig, 1, 4, 245-265 (1986). [3] A.I. Zadori The Trasfer of the Boudary Coditio From the Ifiity for the Numerical Solutio to the Secod Order Equatios with a Small Parameter Siberia Joural of Numerical Mathematics, 2, 1, 21-36 (1999). [4] A.I. Zadori, O.V. Haria Numerical method for a system of liear equatios of secod order with a small parameter o a semi-ifiite iterval Siberia Joural of Numerical Mathematics, 7, 2, 103-114 (2004). [5] J.D. Kadilarov, L.G. Vulkov ad A.I. Zadori A method of lies approach to the umerical solutio of sigularly perturbed elliptic problems Lecture Notes i Computer Sciece, 1988, 451-458 (2001). [6] A.I. Zadori Reductio from a Semi-Ifiite Iterval to a Fiite Iterval of a Noliear Boudary Value Problem for a System of Secod-Order Equatios with a Small Parameter, Siberia Mathematical Joural, 42, 5, 884-892 (2001). 7