Crank-Nicolson Scheme for Space Fractional Heat Conduction Equation with Mixed Boundary Condition

Abstract

-This paper presents Crank-Nicolson scheme for space fractional heat conduction equation, formulated with Riemann-Liouville fractional derivative. Dirichlet and Robin boundary condition will be considered. To illustrate the accuracy of described method some computational examples will be presented as well.

I. INTRODUCTION

In recent years the applications of mathematical models using the fractional order derivatives are very popular in technical science. Different types of phenomena in physics, biology, viscoelasticity, heat transfer, electrical engineering, control theory, fluid and continuum mechanics can be modeled by using the fractional order derivatives [ 6 ], [ 23 ], [ 10 ], [ 19 ], [ 33 ]. For example in paper [20] mathematical models of supercapacitors are considered. Authors investigated models based on electrical circuits in the form of RC ladder networks. These models are described using fractional order differential equations. Often we are not able to solve these models in an analytical way, so it is important to develop approximate methods of solving differential equations of fractional order.

Murio in paper [21] presents the implicit finite difference approximation for the time fractional diffusion equations with homogeneous Dirichlet boundary conditions, formulated by Caputo derivative. In paper [ 1 ], implicit finite difference method was used to solve time fractional heat conduction equation with Neumann and Robin boundary conditions. In these case also, the Caputo derivative was used. In paper [30] authors describes approximated solution of the fractional equation with

Dirichlet and Neumann boundary conditions.

Paper [ 7 ] describes numerical method for diffusion equation with Dirichlet boundary conditions. To discretization fractional derivative authors used finite difference method and Kansa method. Paper [ 11 ] presents numerical solution of model describe by fractional differential equation with space fractional derivative and Dirichlet boundary conditions. Fractional derivative used in this model was Riemann-Liouville derivative. Authors used finite volume method.

Also Meerschaert and Tadjeran dealt with fractional differential equations [ 13 ], [ 12 ], [ 14 ], [35]. Paper [35] describes Copyright c 2016 held by the author. numerical solution of space fractional diffusion equation with boundary condition of the first kind, and in paper [ 12 ] authors presents finite difference method for two-dimensional fractional dispersion equation. In both papers, as the fractional derivative, the Riemann-Liouville derivative was used.

In this paper we present numerical solution of space fractional heat conduction equation. To the equation the Dirichlet and Robin boundary condition was added. In order to solve the equation, the Crank-Nicolson scheme was used. To illustrate the accuracy of described method some computational examples will be presented as well. The aim to achieve numerical solution of considered model is application it to solve inverse problems. Numerical solution of described model is called solution of direct problem. In the process of solving inverse problem, it is required to solve the direct problem many times. More about inverse problems of fractional order can be found in papers [ 2 ], [ 3 ], [ 4 ], [ 5 ]. In these papers, swarm optimization algorithms are introduced. More about intelligent algorithms and its application can be found in [24], [25], [26], [27], [28].

II. FORMULATION OF THE PROBLEM We discuss the following space fractional heat conduction equation

c % (1) defined in area D = f(x; t) : x 2 [a; b]; t 2 [0; T )g, where (x) is thermal conductivity coefficient, c and % denote the specific heat and density. We assume that (x) > 0 and 2 (1; 2]. The initial condition is also posed u(x; 0) = f (x); x 2 [a; b]; as well as the homogeneous Dirichlet and Robin boundary conditions

u(a; t) = q(t); (x) u1); t 2 [0; T ); t 2 [0; T ); where h describes the heat transfer coefficient and u1 is the ambient temperature. (2) (3) (4)

Fractional derivative with respect to space, which occurs in equation (1), will be the Riemann-Liouville fractional derivative [ 23 ], [ 8 ] determined as follows = (n 1 a x where 2 (n 1; n] and ( ) is the Gamma function [ 34 ]. In case of 2 (1; 2), the equation (1) describes super-diffusive process [ 15 ], [ 16 ], and for = 2, we get the classical heat conduction equation.

III.

NUMERICAL SOLUTION

In this section we describe numerical solution of equation (1) using the finite difference method. Let N; M 2 N be the size of grid in space and time, respectively. We denote grid steps x = (bNa) and t = T =M . Therefore, we get following grid

S = (xi; tk); xi = i x; tk = k t; i = 0; 1; : : : ; N; k = 0; 1; 2; : : : ; M : We assume the following notation i = (xi); gik = g(xi; tk); fi = f (xi); hk = h(tk). Values of approximate function in points (xi; tk), we denoted by Uik.

In order to approximate fractional derivative (5), we used right-shifted Gru¨nwald formula [ 17 ] = ( 1

lim ) N!1 r 1 XN j=0 (j ) (j + 1) where N 2 N and r = x a

N . We denote ! ;j = ( (j ) ) (j + 1)

Discretize the equation (1) and using approximation of fractional derivative, we get U k+1 i t

U k i =

i 2c%( x) i+1 X ! ;j Uik+j1+1 + j=0 i+1 X ! ;j Uik j+1 j=0 gk+ 12 + i c%

Now, we approximate boundary condition of the third kind, so we obtain

UNk =

N UNk 1 +

xhku1 N + xhk u(s; t)(x ds; (5) where

Having regard to both boundary condition, the equation (9) may be written in matrix form (1 + x)x3e t;

U k = [U1k; U2k; : : : ; UNk 1]T ; Gk+ 12 t = h g1k+ 12 t ; : : : ; gNk+ 122 t

; c% c%

N + gNk+ 121 t c% +

N 1 tu1 2( x)

IN 1 is the identity matrix of size N A1 and A2 are defined as follows

i = 1; 2; : : : ; N 1; j = 1; 2; : : : ; N

1; xhk xhk +

xhk+1 N + xhk+1 i ;

Solving systems of equations defined by (11), we obtain approximate values of temperatures in points of grid (6).

In paper [35] authors presents proof of unconditionally stability of presented method in case of homogeneous Dirichlet boundary condition. Doing similar proof it can be proven that described in this paper method is unconditionally stable.

IV. EXPERIMENTAL RESULTS In this section we presents numerical examples of described method.

Example 1: Let consider equation (1), defined in area

D = f(x; t) : x 2 [0; 1]; t 2 [0; 1]g; with following data = 1:7; (x) = (2:2)x2:8;

c = % = 1; u1 = 50;

To equation we added an initial condition

Exact solution of this problem is function

u(x; 0) = x3;

x 2 [0; 1]: u(x; t) = x3e t: Maximal and average absolute errors will be defined by formulas max = 0miaxN juik 1 k M

Uikj;

avg = 1 (N + 1)(M + 1) N M X X juik i=0 k=0

Uikj:

With increase in the first dimension N of the grid, a fixed second dimension M , errors decrease. For example, for M = 100 and N = 100; 200; 300 absolute errors not exceed 6:17 10 3; 3:09 10 3; 2:06 10 3. Increasing the dimension of the grid with respect to time for a fixed dimension N insignificantly impact on errors of approximate solution.

Distribution of errors in points of the grid is presented on figure 1.

Fig. 1. Distribution of errors (N = M = 100) (example 1)

Grid N 10 10 20 50 100 100 100 200 300

M Fig. 4. Distribution of errors for approximate solution (a) and approximate solution (points), exact solution (solid line) (b) in time t = 1 (N = M = 100) (example 2)

Just as it did in the previous examples, we will examine the errors depending on the density of the grid. In table III errors of approximate solutions are presented.

CONCLUSION

In this paper numerical solution of space heat conduction equation is presented. To the equation Robin and Neumann boundary conditions were added. Author used Crank-Nicolson scheme, and right-shifted Gru¨ nwald formula for approximation fractional Riemann-Liouville derivative. To illustrate the accuracy of presented method three examples are presented. In each example results are good. With an increase in the density of the grid, errors of approximate solution decreases.

References

[1] R. Brociek : Implicite finite difference method for time fractional heat equation with mixed boundary conditions . Zesz . Nauk. P S´., Mat. Stosow . 4 ( 2014 ), 72 - 85 .

[2] R. Brociek , D. Słota , R. Wituła: Reconstruction of the thermal conductivity coefficient in the time fractional diffusion equation, in: Advances in Modelling and Control of Non-integer- Order Systems , K.J. Latawiec , M. Łukaniszyn , R. Stanisławski, eds., Lecture Notes in Electrical Engineering, 320 , Springer, 239 - 247 ( 2015 ).

[3] R. Brociek , D. Słota: Reconstruction of the boundary condition for the heat conduction equation of fractional order , Thermal Science 19 , 35 - 42 ( 2015 ).

[4] R. Brociek , D. Słota: Application of intelligent algorithm to solve the fractional heat conduction inverse problem , in: Information and Software Technologies , G. Dregvaite, R. Damasevicius, eds., Communications in Computer and Information Science , 538 , Springer, 356 - 366 ( 2015 ).

[5] R. Brociek , D. Słota: Application and comparison of intelligent algorithms to solve the fractional heat conduction inverse problem . Inf. Technol . Control T. 45 nr 2 ( 2016 ), 184 - 194 .

[6] A. Carpinteri , F. Mainardi: Fractal and Fractional Calculus in Continuum Mechanics . Springer, New York ( 1997 )

[7] W. Chen , L. Ye , H. Sun: Fractional diffusion equations by the kansa method . Comput. Math. Appl . 59 ( 2010 ), 1614 - 1620 .

[8] K. Diethelm : The Analysis of Fractional Differential Equations . Springer, Berlin ( 2010 )

[9] G.H. Gao , Z.A. Sun : A compact finite difference scheme for the fractional sub-diffusion equations . J. Comput. Phys . 230 ( 2011 ), 586 - 595 .

[10] J. Klafter , S. Lim , R. Metzler: Fractional dynamics . Resent advances. World Scientific, New Jersey ( 2012 )

[11] F. Liu , P. Zhuang , I. Turner, K. Burrage , V. Anh: A new fractional finite volume method for solving the fractional diffusion equation . Appl. Math. Modelling 38 ( 2014 ), 3871 - 3878 .

[12] M.M. Meerschaert , H.P. Scheffler , Ch. Tadjeran: Finite difference methods for two-dimensional fractional dispersion equation . J. Comput. Phys . 211 ( 2006 ), 249 - 261 .

[13] M.M. Meerschaert , Ch. Tadjeran: Finite difference approximations for fractional advection-dispersion flow equations . J. Comput. Appl . Math. 172 ( 2006 ), 65 - 77 .

[14] M.M. Meerschaert , Ch. Tadjeran: Finite difference approximations for two-sided space-fractional partial differential equations . Appl . Numer. Math. 56 ( 2006 ), 80 - 90 .

[15] M.M. Meerschaert , D.A. Benson , H.P. Scheffler , P. Becker-Kern : Governing equations and solutions of anomalous random walk limits . Phys. Rev. E 66 ( 2002 ), 102R - 105R .

[16] R. Metzler , J. Klafter: The restaurant at the end of the random walk: recent developments in the description of anomalous transport by fractional dynamics . J. Phys. Rev. A 37 ( 2004 ), R161 - R208 .

[17] K. Miller , B. Ross: An Introduction to the Fractional Calculus and Fractional Differential . Wiley, New York, 1993 .

[18] W. Mitkowski : Approximation of fractional diffusion-wave equation . Acta Mechanica et Aautomatica 5 , no. 2 ( 2011 ), 65 - 68 .

[19] W. Mitkowski , J. Kacprzyk , J. Baranowski: Advances in the Theory and Applications of Non-integer Order Systems . Springer Inter. Publ., Cham ( 2013 ) W. Mitkowski, P. Skruch: Fractional-order models of the supercapacitors in the form of RC ladder networks . Bulletin of The Polish Academy of Sciences Technical Sciences , Vol. 61 , No. 3 , 2013 . pp. 581 - 587 .

D. Murio : Implicit finite difference approximation for time fractional diffusion equations . Comput. Math. Appl . 56 ( 2008 ) 1138 -1145 A. Obra ¸czka: The L2 approximation and shifted Gr u¨nwald methods comparison for RFDE, in: Materiay XV Jubileuszowego Sympozjum ,,Podstawowe Problemy Energoelektroniki, Elektromechaniki i Mechatroniki” . PPEEm 2012 , M. Szczygieł, ed., Archiwum Konferencji PTETiS , 32 , Komitet Organizacyjny Sympozjum PPEE i Seminarium BSE , 2012 , 129 - 132 .

[23] I. Podlubny: Fractional Differential Equations . Academic Press, San Diego ( 1999 ) D. Połap , M. Wozniak: Flexible Neural Network Architecture for Handwritten Signatures Recognition . International Journal of Electronics and Telecommunications , Vol. 62 , No. 2, DOI: 10.1515/eletel-2016-0027, ISSN 2300-1933, De Gruyter Open Ltd 2016 , pp. 197 - 202 .

D. Połap , M. Wozniak , C. Napoli , E. Tramontana: Real-Time Cloudbased Game Management System via Cuckoo Search Algorithm . International Journal of Electronics and Telecommunications , Vol. 61 , No.

4, DOI: 10.1515/eletel-2015-0043, ISSN 2300-1933, De Gruyter Open Ltd 2015 , pp. 333 - 338 .

D. Połap , M. Wozniak , C. Napoli , E. Tramontana: Is swarm intelligence able to create mazes? International Journal of Electronics and Telecommunications , Vol. 61 , No. 4, DOI: 10.1515/eletel-2015-0039, ISSN 2300-1933, De Gruyter Open Ltd 2015 , pp. 305 - 310 .

C. Napoli , G. Pappalardo , E. Tramontana: A Mathematical Model For File Fragment Diffusion And A Neural Predictor To Manage Priority Queues Over Bittorrent . International Journal of Applied Mathematics and Computer Science . Vol. 26 , No. 1, Univ Zielona Gora Press 2016, pp. 147 -160 C. Napoli , G. Pappalardo , E. Tramontana: Improving files availability for BitTorrent using a diffusion model . Proceedings of IEEE International WETICE Conference June , 2014 , pp. 191 -196 C. Napoli , E. Tramontana: Massively parallel WRNN reconstructors for spectrum recovery in astronomical photometrical surveys . Neural Networks . Vol. 83 , Elsevier 2016 , pp. 42 -50 G. Sudha Priya , P. Prakash , J.J. Nieto , Z. Kayar: Higher-order numerical scheme for the fractional heat equation with Dirichlet and Neumann boundary conditions . Numer. Heat Transfer B 63 ( 2013 ), 540 - 559 .

[31] J. Ren , Z.Z. Sun , X. Zhao: Compact difference scheme for the fractional sub-diffusion equation with Neumann boundary conditions . J. Comput. Phys . 232 ( 2013 ), 456 - 467 .

[32] J.P. Roop : Computational aspects of fem approximation of fractional advection dispersion equations on bounded domains in R2 . J. Comput. Appl . Math. 193 ( 2006 ), 243 - 268 .

[33] J. Sabatier , O.P. Agrawal , J.A. Tenreiro Machado: Advances in Fractional Calculus . Theoretical Developments and Applications in Physics and Engineering . Springer, Dordrecht 2007 .

[34] [35] [36] [37] Sz. Rabsztyn , D. Słota , R. Wituła: Gamma and Beta Functions , vol. 1 (in Polish). Wyd. Pol. S´ l., Gliwice ( 2012 ) Ch . Tadjeran, M.M. Meerschaert , H.P. Scheffler : A second-order accurate numerical approximation for the fractional diffusion equation .

J. Comput . Phys. 213 ( 2006 ), 205 - 213 .

Modelling 38 ( 2014 ), 3848 - 3859 .

Y. Zheng , Ch. Li, Z. Zhao: A note on the finite element method for the space-fractional advection diffusion equation . Comput. Math. Appl . 59 ( 2010 ), 1718 - 1726 .