On Stability Criteria of Explicit Difference Schemes for Certain Heat Conduction Problems with Uncommon Boundary Conditions

Similar documents
EXPANSION OF ANALYTIC FUNCTIONS IN TERMS INVOLVING LUCAS NUMBERS OR SIMILAR NUMBER SEQUENCES

I9 Lateral deflections of circular plates

A L A BA M A L A W R E V IE W

Week 7: Integration: Special Coordinates

EFFECTIVE CONDUCTIVITY, DIELECTRIC CONSTANT AND PERMEABILITY OF A DILUTE SUSPENSION*)

MATH 2050 Assignment 6 Fall 2018 Due: Thursday, November 1. x + y + 2z = 2 x + y + z = c 4x + 2z = 2

LARGE AMPLITUDE OSCILLATIONS OF A TUBE OF INCOMPRESSIBLE ELASTIC MATERIAL*

(4) cn = f f(x)ux)dx (n = 0, 1, 2, ).

I N A C O M P L E X W O R L D

A NOTE ON A BASIS PROBLEM

IMPROVEMENT OF AN APPROXIMATE SET OF LATENT ROOTS AND MODAL COLUMNS OF A MATRIX BY METHODS AKIN TO THOSE OF CLASSICAL PERTURBATION THEORY

A CONVERGENCE CRITERION FOR FOURIER SERIES

Math 234. What you should know on day one. August 28, You should be able to use general principles like. x = cos t, y = sin t, 0 t π.

n r t d n :4 T P bl D n, l d t z d th tr t. r pd l

ALGEBRAIC PROPERTIES OF SELF-ADJOINT SYSTEMS*

18 BESSEL FUNCTIONS FOR LARGE ARGUMENTS. By M. Goldstein and R. M. Thaler

Ht) = - {/(* + t)-f(x-t)}, e(t) = - I^-á«.

Math 353, Practice Midterm 1

4 4 N v b r t, 20 xpr n f th ll f th p p l t n p pr d. H ndr d nd th nd f t v L th n n f th pr v n f V ln, r dn nd l r thr n nt pr n, h r th ff r d nd

Gauss s Law. Name. I. The Law: , where ɛ 0 = C 2 (N?m 2

4 8 N v btr 20, 20 th r l f ff nt f l t. r t pl n f r th n tr t n f h h v lr d b n r d t, rd n t h h th t b t f l rd n t f th rld ll b n tr t d n R th

Exam 1 Solutions. Note that there are several variations of some problems, indicated by choices in parentheses. Problem 1

Numerical solution of hyperbolic heat conduction in thin surface layers

Chapter 4: Transient Heat Conduction. Dr Ali Jawarneh Department of Mechanical Engineering Hashemite University

Turán s problem and Ramsey numbers for trees. Zhi-Hong Sun 1, Lin-Lin Wang 2 and Yi-Li Wu 3

E. not enough information given to decide

AP Physics C. Electric Potential and Capacitance. Free Response Problems

Practice Problems for Exam 3 (Solutions) 1. Let F(x, y) = xyi+(y 3x)j, and let C be the curve r(t) = ti+(3t t 2 )j for 0 t 2. Compute F dr.

I ƒii - [ Jj ƒ(*) \'dzj'.

0 t b r 6, 20 t l nf r nt f th l t th t v t f th th lv, ntr t n t th l l l nd d p rt nt th t f ttr t n th p nt t th r f l nd d tr b t n. R v n n th r

PERIODIC SOLUTIONS OF SYSTEMS OF DIFFERENTIAL EQUATIONS1 IRVING J. EPSTEIN

ON THE NUMERICAL RANGE OF AN OPERATOR

Numerical Quadrature over a Rectangular Domain in Two or More Dimensions

1.1 GRAPHS AND LINEAR FUNCTIONS

GROUP OF TRANSFORMATIONS*

Last Time. Magnetic Torque Magnetic Dipole in a B-Field: Potential Energy

Th n nt T p n n th V ll f x Th r h l l r r h nd xpl r t n rr d nt ff t b Pr f r ll N v n d r n th r 8 l t p t, n z n l n n th n rth t rn p rt n f th v

CATAVASII LA NAȘTEREA DOMNULUI DUMNEZEU ȘI MÂNTUITORULUI NOSTRU, IISUS HRISTOS. CÂNTAREA I-A. Ήχος Πα. to os se e e na aș te e e slă ă ă vi i i i i

ALTERNATIVE DERIVATION OF SOME REGULAR CONTINUED FRACTIONS

,. *â â > V>V. â ND * 828.

YALE UNIVERSITY DEPARTMENT OF COMPUTER SCIENCE

The Helically Reduced Wave Equation as a Symmetric Positive System

QUESTION ANSWER. . e. Fourier number:

1.2 LECTURE 2. Scalar Product

Mathematica Slovaca. Jaroslav Hančl; Péter Kiss On reciprocal sums of terms of linear recurrences. Terms of use:

Chemistry Higher level Paper 1

f b a 2 Bay - az V /m. Section 4.2- Coulomb's Law and Field Intensity

13 PDEs on spatially bounded domains: initial boundary value problems (IBVPs)

46 D b r 4, 20 : p t n f r n b P l h tr p, pl t z r f r n. nd n th t n t d f t n th tr ht r t b f l n t, nd th ff r n b ttl t th r p rf l pp n nt n th

22 t b r 2, 20 h r, th xp t d bl n nd t fr th b rd r t t. f r r z r t l n l th h r t rl T l t n b rd n n l h d, nd n nh rd f pp t t f r n. H v v d n f

(a) Consider a sphere of charge with radius a and charge density ρ(r) that varies with radius as. ρ(r) = Ar n for r a

APSIDAL ANGLES FOR SYMMETRICAL DYNAMICAL SYSTEMS WITH TWO DEGREES OF FREEDOM

18.02 Multivariable Calculus Fall 2007

ALGEBRAIC PROPERTIES OF THE LATTICE R nx

ON THE CONVERGENCE OF APPROXIMATE SOLUTIONS OF

Recursive Computation of the Repeated Integrals of the Error Function


828.^ 2 F r, Br n, nd t h. n, v n lth h th n l nd h d n r d t n v l l n th f v r x t p th l ft. n ll n n n f lt ll th t p n nt r f d pp nt nt nd, th t

Questions Chapter 23 Gauss' Law

Polynomial expansions in the Borel region

ON PERIODIC SOLUTIONS OF NONLINEAR DIFFERENTIAL EQUATIONS WITH SINGULARITIES

Introduction to Heat and Mass Transfer. Week 8

Second Year Electromagnetism Summer 2018 Caroline Terquem. Vacation work: Problem set 0. Revisions

ASYMPTOTIC HEAT CONDUCTION IN ARBITRARY BODIES

ON RESTRICTED SYSTEMS OF HIGHER INDETERMINATE EQUATIONS * E. T. BELL

THE PRESERVATION OF CONVERGENCE OF MEASURABLE FUNCTIONS UNDER COMPOSITION

Chapter III. The bivariate Gini-index

Chapter 3: Steady Heat Conduction

NARAYANA IIT/PMT ACADEMY

PROCEEDINGS. The present investigation will show that these two line elements, together

The goal of this chapter is to study linear systems of ordinary differential equations: dt,..., dx ) T

MATH20411 PDEs and Vector Calculus B

Math 20C Homework 2 Partial Solutions

Optical Imaging Chapter 5 Light Scattering

Problem Set 1. This week. Please read all of Chapter 1 in the Strauss text.

P a g e 5 1 of R e p o r t P B 4 / 0 9

(1-2) fx\-*f(x)dx = ^^ Z f(x[n)) + R (f), a > - 1,

3 Chapter. Gauss s Law

Chapter 15 Appendix Moment of Inertia of a Spherical Shell

[2 marks] Show that derivative of the angular velocity. What is the specific angular momentum j as a function of M and R in this Keplerian case?

The solution of linear differential equations in Chebyshev series

MINIMAL.4-SETS, INFINITE ORBITS, AND FIXED ELEMENTS

E & M Qualifier. August 16, To insure that the your work is graded correctly you MUST: 5. number every page starting with 1 for each problem,

Applications of Gauss Law

Massachusetts Institute of Technology Instrumentation Laboratory Cambridge, Massachusetts

Dielectric Polarization, Bound Charges, and the Electric Displacement Field

Physics Electricity & Op-cs Lecture 8 Chapter 24 sec Fall 2017 Semester Professor

MATH 18.01, FALL PROBLEM SET # 6 SOLUTIONS

176 5 t h Fl oo r. 337 P o ly me r Ma te ri al s

INFINITE RADICALS IN THE COMPLEX PLANE1

DISTRIBUTIONS FUNCTIONS OF PROBABILITY SOME THEOREMS ON CHARACTERISTIC. (1.3) +(t) = eitx df(x),

ELEC9712 High Voltage Systems. 1.2 Heat transfer from electrical equipment

Stability investigation for the inverse heat conduction problems Krzysztof Grysa & Artur Maciqg Kielce University of Technology, AL 1000-leciaPP 7,

Vector analysis. 1 Scalars and vectors. Fields. Coordinate systems 1. 2 The operator The gradient, divergence, curl, and Laplacian...

A DUALITY THEOREM FOR NON-LINEAR PROGRAMMING* PHILIP WOLFE. The RAND Corporation

J2 e-*= (27T)- 1 / 2 f V* 1 '»' 1 *»

FUSION TECHNOLOGY INSTITUTE

Colby College Catalogue

( ) PROBLEM C 10 C 1 L m 1 50 C m K W. , the inner surface temperature is. 30 W m K

Chapter 5 Time-Dependent Conduction

Transcription:

On Stability Criteria of Explicit Difference Schemes f Certain Heat Conduction Problems with Uncommon Boundary Conditions By Arnold N. Lowan Abstract. Stability criteria are derived f the explicit difference schemes appropriate to the following problems: 1) heat conduction in a slab in contact with a well stirred liquid; 2) heat conduction in a slab radiating to one face of a thin slab with infinite thermal conductivity, the other face of which radiates into a medium at prescribed temperature; 3) heat conduction in a cylinder radiating to the inner surface of a thin coaxial cylindrical shell with infinite thermal conductivity, the outer surface of which radiates into a medium at prescribed temperature. Although the exact analytical solutions of certain problems in heat conduction involving complicated boundary conditions are known, thé complexity of the analytical expressions is often such as to make them impractical f the numerical evaluation of the solutions. This is f instance the case of the writer's solution of the problem of "heat conduction in a solid in contact with a well stirred liquid" [1]. It is also the case of the problems treated by Walter P. Reid and dealing with the heat conduction in a semi-infinite solid ( cylinder) when the boundary surface radiates to one boundary surface of a thin slab ( thin cylindrical shell), with infinite thermal conductivity, the other boundary surface of which radiates into a medium at prescribed temperature [2], [3]. To obtain numerical answers to the above problems it is expedient to evaluate the solutions of the appropriate explicit difference analogs. The object of this rept is to derive the stability criteria f the difference schemes appropriate to the problems above-mentioned. Consider first the problem of heat conduction in a slab one face of which is in contact with a well-stirred liquid. F the sake of concreteness we shall first assume that the other face is kept at 0 C. The mathematical fmulation of the problem is as follows: (1) ^-*g Oáxáa, t>0 (2) T(x, 0) = f(x) (3A) r(o, o=o (4) -di=\-k-)wr-dt < *>*» *-«In (4) we have the boundary condition appropriate to the case where the face x = a is in contact with a layer of a well-stirred liquid of width d, density po and specific heat Co. The constants K and /; are the thermal conductivity and thermal diffusivity of the slab, respectively. Received August 17, 1959; revised November 4, 1960. 179

180 ARNOLD N". LO WAN We shall consider the problem of stability of the explicit difference scheme (5) r...+1 =- rt^.n + (1-2r)Tm,n + rtm+1,n; m = 1, 2, 3, M where Tm,n - T(mAx, nàt); r - fcai/(ar)*, and a - (M + l)ax. The difference analog of (4) is (6) whence T.n - T.V+l.n M+l.n+l - T Ṁ+l.% Ax AÍ (7) 31+l.n+l «TV,, + (1 - s)tm+1.n ; Since Ai = r(ax)*/fc m r(ax)lpc/ic where p and c are the density and specific heat of the slab, it follows that s «r(ax/d)(pc/poco). Since usually Ax <5C d and c «Co it is reasonable to assume that Ar/d-pc/poCo < 1 and a ftii s < 1, if we anticipate the fact that the criterion f the stability of the explicit difference scheme under consideration is r ^ \. The system of M + 1 equations consisting of (5) and (7) may be written in the matrix-vect fm (8) T,+l = A T where T» and T +i are the (il/ + 1) dimensional vects whose components are the temperatures Tm, and r,,,+i, respectively, with m 1, 2, 3, M 4-1 and where ~1-2r r oax (9) r 1-2r r r 1 2r 1 - s Examination of the matrix A shows that if r g $ so that 1 2r ^ 0, then the largest of the sums of the absolute values of the elements of each of the first M rows of A is equal to 1. Furtherme, we have seen that it is reasonable to assume that s á 1 1 - s ^ 0; accdingly, the sum of the absolute values of the elements of the last row of A is again «1. The condition f the stability of the difference scheme under consideration is thus satisfied [4]. In the above treatment we assumed that the face x 0 is kept at 0 C. If instead the temperature at x = 0 is prescribed, so that (3B) no, t) = *(o it is readily seen that the stability criterion r _ we found above is valid, since the err vect E satisfies the difference equation (8) and since it vanishes on x = 0. We define the err vect E» as the vect T, T *. where T is the "true" solution of (8) and T * is the vect generated by successive application of

ON STABILITY CRITERIA OF EXPLICIT DIFFERENCE SCHEMES 181 (8) when during some time step an inaccurate vect Tt* (where k can of course be = 0) is used in lieu of the true T*. It may be briefly mentioned that the above stability criterion is equally valid in case of boundary conditions of the fm (3C) lx" = f * = (3D) T~= ht f x = - In the case of the boundary condition (3C) the first element of the first row of the matrix A is replaced by 1 r; in the case of the boundary condition (3D) the element in question is replaced by 1 2r 4- ocr where a = 1/(1 4- AAr). In either case the largest of the sums of the absolute values of the elements of the rows of A is still equal to one and therefe the stability condition is satisfied if r ^ J. F a discussion of convergence the reader is referred to [4], Section VI. Consider now the problem of heat conduction in a slab subject to the conditions stipulated in the first of Reid's articles mentioned above. The mathematical fmulation of the problem is as follows: (10) ^- = kjt 0^1^,00 dt dx2 ill) T(x,0) "fix) (12) T(0, i) = 0 -*( ) (13) -#\je) =hatia,t) -vit)] (14) PoCod^ = hl[tia,t) -vit)]-hvit) (15)»(0) = V In the above: K = thermal conductivity of the thick slab k = thermal diffusivity of the thick slab hi = coefficient of heat transfer between the two slabs ht coefficient of heat transfer between the thin slab and the surrounding medium (with temperature zero) t>(i) = temperature of thin slab p», Co and d are the density, specific heat and width of the thin slab. If we put (16) ^-b, _ = C, ~^=<r equations (13) and (14) assume the fm (13<) ~ (fr)«= b itia'l) ~ v{t)] (14*) a^. = b[tia,t) -vit)]-cvit).

182 ARNOLD N. LO WAN We shall investigate the stability of the explicit difference scheme 1 m,n+l I m.n " (TOT_i, 2Tm,n 4" Tm+\, ) ~W~ (Ar)2 (17) Tm,n+l = rtm^.,+ (1-2r)Tm,n + rtm+l.n, m = 1, 2, 3, M where 7V = T(mAx, nm), r = ^ and Ax = ^^. The difference analogs of (13*) and (14*) are 3V, - rw,. = &(r _ j Ar (18) ÏW, =,,\ A TM.» + r^r-». = ptm.n 4- qvh (say) 1 4- oar 1 4-»Ax and (19) Vn+\> "' = bt " - (6 + c>y" Ai P.+, - b-f TM+1,n + [l - (6 +gc) *] f» «OÍW. + ßvK (say) where we have written vn and i'n+i f vinat) and v[(n + 1) Af], respectively. (18) and (19) we eliminate T^+x,,, we get If from (20) t;b+l = aptu.n + iß + aq)vn. If we rewrite (18) in the fm (18*) Tu+i.n+i = ptm,h+i + qvn+i and eliminate t' and vn+i from (18*), (18), and (19), we get (21) Tm+i.u+i = iß + aq)t +lin + ptm,k+l - pßtm.n. If in (21) we replace T u,n+\ by its expression from (17) with h = M, we ultimately get,. (22) T*+i.,+1 = prtu-x,» 4- p(l - 2r - 0)7V + (pr 4- ß + «3)7^,,, = PTm-í., 4- Q7V,, 4- ST*+1.. (say). Starting with the values of Tm,n f n = 0 equations (17) and (22) will generate in succession the values of Tm,n f n = 1, 2, and rre = 1, 2, 3, M 4-1. Similarly, starting with i;(i) = V f í = 0 equation (19) will generate in succession the values of vn = f(nai).

ON STABILITY CRITERIA OF EXPLICIT DIFFERENCE SCHEMES 183 The system of M 4-1 equations consisting of (17) and (22) with the errs Em,n written f the rm, 's may be written in the matrix-vect fm (23) En+i = AEn where E and E +i are the (3/4-1) dimensional vects whose components are the errs in the values of the temperature at the M 4-1 lattice points max, m = 1, 2, 3, 3/ 4-1 at the times nat and (n + l)at and where l-2r l-2r (24) 1-2r If we assume (25) 0 < r á \ r 1 2r r P Q S it is clear from (24) that the sum of the absolute values of the elements of the first 3/ rows of A is equal to 1. If in addition we assume that (26) 1-2r - 0 0, then, it is seen from (22) that pr + ß 4- aq ^ 0 (27) P\ + \Q\ + \S\=p-pß + ß + aq=l cat ^ Thus, if (25) and (26) are satisfied, the largest of the sums of absolute values of the elements of the rows of matrix A is equal to 1, and therefe the difference scheme under consideration is stable. The inequalities (25) and (26) may be written whenr*» (28Ï (29} nat < 1 (Ax)2 na^ < ib + c)at (Ax)2 h^ln^ + h^h*0 (Ax2) > 2n<r b + c At ^ min Vax2 a L 2n ' b + c\ ' In conclusion, choosing the intervals Ax and At in accdance with (28) and (29) will insure the stabilitv of the difference scheme under consideration.

184 ARNOLD N. LOWAN Finally, consider the problem of heat conduction in a cylinder subject to the boundary conditions stipulated in the second of Reid's articles mentioned above. If, f the sake of convenience, we denote by x the radial distance dinarily denoted by r, the mathematical fmulation of the problem is as follows: (30) (31) dt dt. (d't. 1 dt\ T(x, 0) = fix) (32) dx = 0 f x = 0 (33) (34) \dx )x-a mi [Tia, t) - vit)] dv 2irpo c0 d = mi [Tia, t) vit)] m2 vit) dt The only essential difference between the above problem f the cylinder and the cresponding problem f the slab is the differential equation (30) which takes the placeof( 10) and the new condition (32). The difference analogs of (30) and (32) are m.fl+l At T ftm-l. 2lm,B 4" lm.n+l (Ax)2 4- max ' m+l.n T *] 2Ax (35) Tm,B+1 = r Í1 - J Tm-!.n 4- (1-2r)T.» + r(l + ) Tm+l.n and m = 1, 2,3, 3/ (36) To,n T\,n The difference equation (35) takes the place of the difference equation ( 17). In view of (35) and (36) it is readily seen that the previous developments f the slab may be carried out with the sole exception that in lieu of (24) the matrix A is now (37).4 = 1-2r H> 2" 1-2r M> 1-2r ('- > 0$ 0+ > 1-2r Q s where the symbols P, Q, and S have the same significance as befe. Since f the first M rows of A the largest of the sums of the absolute values of the elements is

ON STABILITY CRITERIA OF EXPLICIT DIFFERENCE SCHEMES 185 equal to 1 (provided r ; ^) we reach the conclusion that the criteria of stability are identical with those of the previous problem and are given in the inequalities (28) and (29). It may be briefly mentioned that the above analysis of stability may be readily extended to the case where the cylinder and the coaxial thin cylindrical shell are replaced by a sphere and a concentric thin spherical shell. Yeshiva University New Yk, New Yk 1. A. N. Lowan, "Heat conduction in a solid in contact with a well-stirred liquid," Phil. Mag., v. 17, 1934, p. 849-854. 2. Walter P. Reíd, "Heat flow in a half space," Quart. Appl. Math., v. 14,1956, p. 206-208. 3. Walter P. Reíd, "Heat flow in a cylinder," Quart. Appl. Math., v. 16, 1958, p. 147-153. 4. A. N. Lowan, The Operat Approach to Problems of Stability and Convergence of Solutions of Difference Equations and the Convergence of Various Iteration Procedures, Scripta Mathematica, New Yk, 1957, p. 79.