Journal of Applied Mathematics and Physics, 2017, 5, 1301-1319 http://www.scirp.org/journal/jamp

ISSN Online: 2327-4379 ISSN Print: 2327-4352

DOI: 10.4236/jamp.2017.56109 June 27, 2017

A Second Order Characteristic Mixed Finite Element Method for Convection Diffusion Reaction Equations

Tongjun Sun

School of Mathematics, Shandong University, Jinan, China

Abstract A combined approximate scheme is defined for convection-diffusion-reaction equations. This scheme is constructed by two methods. Standard mixed finite element method is used for diffusion term. A second order characteristic fi-nite element method is presented to handle the material derivative term, that is, the time derivative term plus the convection term. The stability is proved and the L2-norm error estimates are derived for both the scalar unknown va-riable and its flux. The scheme is of second order accuracy in time increment, symmetric, and unconditionally stable.

Keywords Mixed Finite Element Method, Characteristic Method, Second Order Accuracy, Convection Diffusion Reaction Equations

1. Introduction

Let Ω be a bounded domain in ( )2,3d d =R with Lipschitz boundary

D RΓ = Γ ∪Γ , where D RΓ ∩Γ =∅ . Let T be a positive constant. In this paper, we will consider the following linear convection-diffusion-reaction equations: find ( ): 0,Tφ Ω× → R such that

( ) ( ) ( ) ( ) ( )( )( ) ( ) ( ) ( )

( ) ( )( ) ( ) ( ) ( ) ( ) ( )

( ) ( )0

,, , ,

, , , in 0, ,

, 0, on 0, ,

, , , , on 0, ,

,0 , in .

D

R

x tx t x t div x x t

tr x x t f x t T

x t T

x t x x t x g x t T

x x

φφ φ

φ

φ

αφ φ

φ φ

∂+ ⋅∇ − ∇ ∂

+ = Ω×

= Γ × + ∇ ⋅ = Γ × = Ω

v A

A n

(1)

How to cite this paper: Sun, T.J. (2017) A Second Order Characteristic Mixed Finite Element Method for Convection Diffusion Reaction Equations. Journal of Applied Ma- thematics and Physics, 5, 1301-1319. https://doi.org/10.4236/jamp.2017.56109 Received: May 27, 2017 Accepted: June 24, 2017 Published: June 27, 2017 Copyright © 2017 by author and Scientific Research Publishing Inc. This work is licensed under the Creative Commons Attribution International License (CC BY 4.0). http://creativecommons.org/licenses/by/4.0/

Open Access

T. J. Sun

1302

Here, [ ]: 0, dTΩ× →v R is the convection vector field; [ ]: 0,r TΩ× → R is the reaction function; [ ]: 0,f TΩ× → R and [ ]: 0,Rg TΓ × → R are given scalar functions; 0α > is a constant and n is the outward unit normal vector to Γ ; : dΩ→ A denotes the diffusion matrix function, where d is the space of symmetric d d× matrices, such that

2 20 1

, 1, ,

dd

ij i ji j

a aξ ξ ξ ξ ξ=

≤ ≤ ∀ ∈∑A R (2)

where 0 1,a a are positive constants. Typically, these Equations (1) express the general mathematical model incor-

porating different types of transport phenomena in engineering and applied sciences, such as the dispersal of a pollutant through a moving viscous medium (e.g., a river or the atmosphere, [1]), currents in semiconductor devices [2], and airflow past an aerofoil (see [3], for example). When the diffusive term is smaller than the convective one, these equations are the so-called convection dominated problems (see [4]).

In many practical convection-diffusion processes, convection essentially do-minates diffusion (e.g., in some financial models [5]), and although the govern-ing differential equation is parabolic, it displays several characteristics of hyper-bolic problems. When applied to these problems, standard finite element and fi-nite difference methods usually exhibit some combination of nonphysical oscil-lation and excessive numerical dispersion [6] [7]. It is therefore logical to design numerical procedures that incorporate the parabolic/hyperbolic nature of these problems. One such method is the modified method of characteristics (MMOC) which was first formulated for a scalar parabolic equation by J. Douglas and T. F. Russell in [8] and then extended by Russell [9] to nonlinear coupled systems in two and three spatial dimensions. Similar schemes had been defined by Piron-neau [10] for the incompressible Navier-Stokes equations and by Süli [11] and Morton, Priestley, and Süli [12] for first-order hyperbolic equations, with the latter technique being referred to as the Euler characteristic Galerkin method. The intent of the method is to obtain accurate approximations to convection- dominated problems. Basically, in the modified method of characteristics, the time derivative and the convection term are combined as a directional deriva-tive. In other words, the procedure involves time stepping along the characteris-tics, allowing us to use large, accurate time steps.

Mixed finite element method has been proven to be an effective numerical method for solving fluid problems. It has an advantage to approximate the un-known variable and its diffusive flux simultaneously. There are many research articles on this method ([13] [14] [15] [16]). An algorithm combining the mixed finite-element method and the modified method of characteristics was first ap-plied to the miscible displacement problem in porous media by Ewing, Russell, and Wheeler [17]. Then, the scheme had been extended by Wheeler and Dawson [18] to advection-diffusion-reaction problems. Numerical results have verified that large, accurate time steps are possible, and sharp fronts have been resolved

T. J. Sun

1303

(without oscillations or numerical diffusion) by coarser grids that standard pro-cedures can use.

Arbogast and Wheeler [19] defined a characteristics-mixed method to ap-proximate the solution of an advection-dominated transport problem. It used a characteristic approximation that is similar to that of MMOC-Galerkin method to handle advection in time and a lowest order mixed finite element spatial ap-proximation for diffusion term. Piecewise constants were in the space of test function, so mass is conserved element-by-element. It was proved finally that the method was optimally convergent to order 1 in time and at least suboptimally convergent to order 3/2 in space. In [20], we have considered a combined nu-merical approximation for incompressible miscible displacement in porous me-dia. Standard mixed finite element was used for Darcy velocity equation and a characteristics-mixed finite element method was presented for approximating the concentration equation. Characteristic approximation was applied to handle the convection term, and a lowest order mixed finite element spatial approxima-tion was adopted to deal with the diffusion term. Thus, the scalar unknown concentration and the diffusive flux can be approximated simultaneously. This approximation conserves mass globally. The optimal L2-norm error estimates were derived. Then, we extended this method to the slightly compressible misci-ble displacement problem in [21].

It should be pointed out that the works mentioned above which involved the characteristic method all gave one order accuracy in time increment t∆ . That is to say, the first order characteristic method in time was analyzed. As for higher order characteristic method in time, Rui and Tabata [22] used the second order Runge-Kutta method to approximate the material derivative term for convec-tion-diffusion problems. The scheme presented was of second order accuracy in time increment t∆ , symmetric and unconditionally stable. Optimal error esti-mates were proved in the framework of L2-theory. Numerical analyses of con-vection-diffusion-reaction problems with higher order characteristic/finite ele-ments were analyzed in [23] [24], which extended the work [22]. The ( )2l L∞ error estimates of second order in time increment t∆ were obtained.

The goal of this paper is to present a second order characteristic mixed finite element method in time increment to handle the material derivative term of (1). It is organized as follows. In Section 2, we formulate an approximate scheme that combines the second order characteristic finite element method for the material derivative term and mixed finite element method for the diffusion term. In Sec-tion 3, we prove the stability of the combined approximate scheme. In Section 4, we derive the L2-norm error estimates for both the unknown scalar variable and its flux. The scheme is of second order accuracy in time increment, symmetric, and unconditionally stable.

2. Formulation of the Method

In this paper, we adopt notations and norms of usual Sobolev spaces. Moreover, we adopt some notations for the functional spaces involved, which were used in

T. J. Sun

1304

[22] [23] [24]. For a Banach space X and a positive integer m, spaces [ ]( )0, ,mC T X and ( )( )0, ,mH T X are abbreviated as ( )mC X and ( )mH X ,

respectively, and endowed with norms

( ) [ ]( ) ( ) ( )

( ) ( )122

00, 0, , 0: max max , : d ,m m

mTj jC X H Xt T j m X Xj

t t tϕ ϕ ϕ ϕ∈ = =

= =

∑∫

where ( )jϕ denotes the j-th derivative of ϕ with respect to time. The Banach space mZ is defined by

( )( ) ; 0, , ,m j m jZ f C H j m−= ∈ Ω =

equipped with the norm ( ) : max ;0 .m j m jZ C H j mϕ ϕ −= ≤ ≤ Similar spaces

and norms are considered for the boundary sets ΓR and ΓD. Denote by ( )1D

HΓ Ω the closed subspace of ( )1H Ω defined by

( ) ( ) 1 1: , 0 ,D D

H Hϕ ϕΓ ΓΩ = ∈ Ω =

and [25]

( ) ( )( ) ( ) 2 2; : , .d

H div L div LΩ = ∈ Ω ∈ Ωv v

2.1. The Characteristic Lines

Now, we define the characteristic lines associated with vector field v and recall some classical properties satisfied by them. Thus, for given ( ) [ ], 0,x t T∈Ω× , the characteristic line through ( ),x t is the vector function ( ), ;eX x t ⋅ solving the initial value problem

( ) ( )( ) ( ), ; , ; , , , ; .ee e

X x t X x t X x t xτ τ τ ττ

∂= =

∂v (3)

Next, assuming they exist, we denote by eF (respectively, by L) the gradient of eX (respectively, of v ) with respect to the space variable x, i.e.,

( ) ( ) ( ) ( ) ( ) ( ), ; : , ; , , : , .e r re rsrs

s s

XF x t x t L x t x t

x xτ τ

∂ ∂= =

∂ ∂v (4)

We adopted some propositions and lemmas from [23]. Proposition 1. If ( )( )0 nC C∈ Ωv for 1n ≥ an integer, then eX ∈

[ ] [ ]( )0 0, 0,C T TΩ× × and it is nC with respect to the x variable. In order to compute second order approximations of matrices eF and 1

eF − , we need the following equations:

( ) ( )( ) ( )

( ) ( )( ) ( )2

2

, ; , ; , , ; ,

, ; , ; , , ; .

ee e

ee e

F x t L X x t F x t

F x t L X x t F x tt

τ τ τ ττ

τ τ τ ττ

∂ = ∂∂ ∂ = ∇ + ∂∂

v v (5)

Proposition 2. If ( )( )0 1C C∈ Ωv , then

( ) ( )( ) [ ]0 1

, ; , , , 0, .C C t

eF x t e x t Tτ

τ τΩ −

≤ ∀ ∈Ω ∈v

(6)

T. J. Sun

1305

Proposition 3. If ( )( ) ( )( )0 2 1 1C C C C∈ Ω ∩ Ωv , then eF satisfies the Taylor expansion

( ) ( ) ( )

( ) ( )( ) ( )

, ; ,

, ; , , ; d ,

e

te es

F x t s I s t L x t

s L X x t F x tt

τ τ τ τ τ

= + −

∂ + − ∇ + ∂ ∫v v

(7)

and its inverse, 1eF − , satisfies the Liouville’s theorem

( ) ( ) ( )( )

( ) ( )( ) ( )( )

1 , ; , ; ,

, ; , , ; , ; d .

e e

te e es

F x t s I t s L X x t s s

t L X x t F X x t s st

τ τ τ τ τ

− = + −

∂ − − ∇ + ∂ ∫v v

(8)

By using the Liouville’s theorem and the chain rule, we obtain

( ) ( ) ( )( )

( ) ( ) ( ) ( )( )(( )( ) ( ) ( )( ))

1 1

221 1

2

T

, ; , ; , ; , ,

, ; , ; , ; ,

, ; , , ; , .

e e e

e e e

e e

detF x t detF x t div X x t

detF x t detF x t div X x t

div L X x t L L X x tt

τ τ τ ττ

τ τ τ ττ

τ τ τ τ

− −

− −

∂ = −∂∂ =∂

∂ ⋅ + − ⋅ ∂

v

v

v v

(9)

Proposition 4. If ( )( )0 1C C∈ Ωv , then

( ) ( )( ) [ ]0 1

1 , ; , , , 0, .C C t

edetF x t e x t Tτ

τ τΩ −

− ≤ ∀ ∈Ω ∈v

(10)

Proposition 5. If ( )( ) ( )( )0 2 1 1C C C C∈ Ω ∩ Ωv , then 1edetF − satisfies

( ) ( ) ( )

( ) ( )( )

1

21

2

, ; 1 ,

, ; d .

e

tes

detF x t s s t div x t

s detF x tτ τ ττ

−

−

= − −

∂+ −

∂∫

v (11)

2.2. Variational Formulation

From the definition of the characteristic lines and by using the chain rule, it fol-lows that

( )( ) ( )( )( )( ) ( )( )

d , ; , , ; ,d

, ; , , ; , .

e e

e e

X x t X x tt

X x t X x t

φ φτ τ τ ττ

τ τ φ τ τ

∂=∂

+ ⋅∇v (12)

By introducing the flux σ φ= ∇A and using (12), we rewrite equation (1.1) at point ( ), ;eX x t τ and time τ as follows

( )( ) ( )( )( )( ) ( )( ) ( )( )

( )( ) ( )( ) ( )( )

d , ; , , ; ,d

, ; , ; , , ; , ,

, ; , , ; , ; , .

e e

e e e

e e e

X x t div X x t

r X x t X x t f X x t

X x t X x t X x t

φτ τ σ τ τ

ττ φ τ τ τ τ

σ τ τ τ φ τ τ

− + =

= ∇

A

(13)

Before giving a week formulation of (13), we adopted a lemma from [23], which can be considered as a Green’s formula.

Lemma 6. Let ( ) ( )2: ,X X X CΩ→ Ω ∈ Ω be an invertible vector valued function. Let F X= ∇ and assume that ( )1 1 .F C− ∈ Ω Then

T. J. Sun

1306

( )( ) ( ) ( ) ( ) ( )( ) ( )( ) ( )( ) ( ) ( ) ( )( ) ( )

T

1 T

d d

d d ,

div X x x x F x x X x x

F x X x x x div F x X x x x

ψ ψ

ψ ψ

−

Ω Γ

− −

Ω Ω

= ⋅ Γ

− ⋅∇ − ⋅

∫ ∫∫ ∫

w n w

w w (14)

with ( )( )1H X∈ Ωw a vector valued function and ( )1Hψ ∈ Ω a scalar func-tion.

Now, we can multiply (13) by test functions ( )1D

Hψ Γ∈ Ω and ( );H divχ ∈ Ω , integrate in Ω respecetively, and apply the usual Green’s formula and (14) with ( ) ( ), ;eX x X x t τ= , obtaining

( )( ) ( ) ( ) ( )( ) ( )

( ) ( )( ) ( )( )( ) ( )( ) ( )( ) ( ) ( )( ) ( )

( )( ) ( )

( )( )( ) ( )( ) ( )( )( )

1

T

T

d , ; , d , ; , ; , dd

, ; , ; , d

, ; , ; , d

, ; , ; , d

, ; , d ,

, ; , , , ; , ; , , .

R

e e e

e e

e e

e e

e

e e e

X x t x x F x t X x t x x

divF x t X x t x x

r X x t X x t x x

F x t x X x t x

f X x t x x

X x t X x t X x t

φ τ τ ψ τ σ τ τ ψτ

τ σ τ τ ψ

τ φ τ τ ψ

τ σ τ τ ψ

τ τ ψ

σ τ τ χ τ φ τ τ χ

−

Ω Ω

−

Ω

Ω

−

Γ

Ω

+ ⋅∇

+ ⋅ + − ⋅ Γ= = ∇

∫ ∫

∫∫∫∫

n

A

(15)

Lemma 7. [23] Let ( ) ( )2: ,X X X CΩ→ Ω ∈ Ω be an invertible vector valued function satifying ( ) ,X x x x= ∀ ∈Γ . Let F X= ∇ and assume that

( )1 1 .F C− ∈ Ω Then

( ) ( ) ( ) ( ) ( ) ( )T 1d d ,F x x x x x x detFψ ψ− −

Γ Γ⋅ Γ = ⋅ Γ∫ ∫n w n w (16)

with ( )( )1w H X∈ Ω and ( )1Hψ ∈ Ω , where n is the outward unit normal vector to Γ .

Now, replacing in (15) formula (16) with ( ) ( ), ;eX x X x t τ= , and replacing the Robin condition, we have

( )( ) ( ) ( ) ( )( ) ( )

( ) ( )( ) ( )( )( ) ( )( ) ( )( )( ) ( ) ( )

( )( ) ( ) ( )( ) ( ) ( )

( )( )( )

1

T

1

1

d , ; , d , ; , ; , dd

, ; , ; , d

, ; , ; , d

, ; , , ; d

, ; , d , ; , , ; d ,

, ; , , ,

R

R

e e e

e e

e e

e e

e e e

e e

X x t x x F x t X x t x x

divF x t X x t x x

r X x t X x t x x

X x t x detF x t

f X x t x x g X x t x detF x t

X x t X x t

φτ τ ψ τ σ τ τ ψ

ττ σ τ τ ψ

τ φ τ τ ψ

αφ τ τ ψ τ

τ τ ψ τ τ ψ τ

σ τ τ χ

−

Ω Ω

−

Ω

Ω

−

Γ

−

Ω Γ

+ ⋅∇

+ ⋅

+

+ Γ

= + Γ

=

∫ ∫

∫∫∫∫ ∫

A ( )( ) ( )( )( ); , ; , , .eX x tτ φ τ τ χ

∇

(17)

2.3. The Combined Approximate Scheme

We now present our time-stepping procedure for (17). Let N be the number of time steps, t T N∆ = be the time step, and nt n t= ∆ for 0,1 2,1,3 2, ,n N= . We will use the notation ( ) ( ): ,n nx x tψ ψ= for a function. Moreover, for

0,1,n = , we define

( ) ( ) ( ) ( )

( ) ( )

1 1

1 1 1 11 12 2 2 2

: , ; , : , ; ,

: , ; , : , ; .

n n n n n ne e e e

n n n nn ne e e e

X x X x t t F x X x t t

X x X x t t F x X x t t

+ +

+ + + ++ +

= =

= =

(18)

T. J. Sun

1307

By fixing 1, 0,1, , 1nt t n N+= = − in (17) and using a Crank-Nicolson me-thod [26] with respect to τ . Thus, from (18) we have

( ) ( )( )( ) ( ) ( )

( ) ( ) ( )( ) ( )

( ) ( ) ( )( ) ( )

( ) ( ) ( ) ( )( ) ( )( ) ( )

( ) ( ) ( ) ( )( ) ( )

( ) ( )( )( ) ( )

( ) ( ) ( ) ( )( ) ( )

11

1

T

1

11

1

11

1d d2

1 d21 d21 1d d2 21 d2

1 d2

1 d ,2

R

R

n n ne n

n n ne e

n n ne e

n n n ne e

n n ne

n n ne

n n ne

x X xx x x x x

t

F x X x x x

div F x X x x x

r x x x x r X x X x x x

x x det F x x

f x f X x x x

g x g x det F x x

φ φψ σ ψ

σ ψ

σ ψ

φ ψ φ ψ

α φ φ ψ

ψ

ψ

++

Ω Ω

−

Ω

−

Ω

+

Ω Ω

−+

Γ

+

Ω

−+

Γ

−+ ⋅∇

∆

+ ⋅∇

+ ⋅

+ +

+ + Γ

= +

+ + Γ

∫ ∫

∫

∫

∫ ∫

∫

∫

∫A ( ) ( )( ) ( )( )1 1 1, , .n nx x xσ χ φ χ− + +

= ∇

(19)

By using (8) and (11), we see that

( ) ( ) ( ) ( )( ) ( )( )( ) ( ) ( ) ( )( )( ) ( ) ( )( ) ( )( )

1 2

1 21

T 2

,

1 ,

.

n n ne e

n ne

n n ne e

F x I x tL X x O t

det F x tdiv x O t

div F x t div X x O t

−

− +

−

= + ∆ + ∆ = + ∆ + ∆ = ∆ ∇ + ∆

v

v

(20)

Taking (20) into (19), we can obtain

( )( ) ( )( )

( ) ( )( )

( )( ) ( )( )( )( ) ( )( )

( )

( )( ) ( )

( ) ( )

1 1

1

1 1

1 1 1

1 1 1

d d2

d2

d2

d2

1d

21

d d ,2 2

, , .

R

R

n n n n n ne e

n n n ne e

n n n ne e

n n n ne e

n n n

n n n n n ne

n n

X x X xx x

tt L X X x x

t div X x X x x

r r X x X xx

tdiv

f f X x g g tdivx

φ φ σ σψ ψ

σ ψ

σ ψ

φ φψ

φ φα ψ

ψ ψ

σ χ φ χ

+ +

Ω Ω

Ω

Ω

+

Ω

+ +

Γ

+ + +

Ω Γ

− + +

− ++ ⋅∇

∆∆

+ ⋅∇

∆+ ∇ ⋅

++

+ + ∆+ Γ

+ + + ∆= + Γ

= ∇

∫ ∫

∫

∫

∫

∫

∫ ∫

v

v

v

A

(21)

We propose two explicit numerical schemes to approximate ( )neX x :

( ) ( ) ( )

( ) ( ) ( )

1

112

: Euler scheme ,

: Runge-Kutta scheme .2

n nE

nn nRK

X x x t x

tX x x t x x

+

+ +

= − ∆

∆ = − ∆ −

v

v v (22)

A similar notation to the one in Section 2.2 is used for the Jacobian of nEX ,

T. J. Sun

1308

namely,

( ) ( ) ( ) ( )1: .n n nE EF x X x I x tL x+= ∇ = − ∆

Now, we restate three lemmas concerning properties of the characteristic line approximations. For this, we require the time step to be bounded and the veloc-ity to satisfy the following assumption.

Claim 1. The velocity field ( )( )0 1,C W ∞∈ Ωv and satisfies 0=v on Γ . Lemma 8. [23] Under Claim 1, if ( )( )0 1, 1 2C W t∞ Ω

∆ <v , we can see that

( ) ( ) = .n nE RKX XΩ = Ω Ω (23)

Lemma 9. [23] Under Claim 1, if ( )( )0 1, 1 2C W t∞ Ω∆ <v , we have

( ) ( ) ( ) ( ) ( )( ) ( )( )1 22 31 1 .n n nEF x I tL x t L x O t

− + += + ∆ + ∆ + ∆ (24)

Corollary 1. Under the assumptions of Lemma 9, x∀ ∈Ω , we have

( ) ( ) ( ) ( )( )( ) ( ) ( )( ) ( )( )0 1,

1 21

1 2

1 ,

1 .

n nE

nE C W

det F x tdiv x O t

det F x t O t∞

− +

−

Ω

= + ∆ + ∆ ≤ + ∆ + ∆

v

v (25)

Lemma 10. [22] Under Claim 1, if ( )2Lψ ∈ Ω and ( )( )0 1, 1 2C W t∞ Ω∆ <v ,

then there exists a positive constant c such that

( )2 21 , for 0, , and , ,n

iX c t n N i E RKψ ψ≤ + ∆ = = (26)

where ( )n ni iX Xψ ψ= .

Thus, in the case where the characteristic lines and their gradients are not ex-plicitly known, we propose the following time approximation of (21)

( ) ( )

( ) ( )

( )

( ) ( )

1 1

1 1 1

1 11

1 1 1

d d2

d d2 2

1d d

2 2

1d d ,

2 2, , .

R

R

n n n n n nRK E

n n n n n nE E

n n n n n nE

n n nn n nE

n n

X Xx xt

t tL X x div X x

r r X tdivx

g g tdivf f X x

φ φ σ σψ ψ

σ ψ σ ψ

φ φ φ φψ α ψ

ψ ψ

σ χ φ χ

+ +

Ω Ω

Ω Ω

+ + +

Ω Γ

+ ++

Ω Γ

− + +

− ++ ⋅∇ ∆

∆ ∆+ ⋅∇ + ∇ ⋅

+ + + ∆ + + Γ

+ + ∆+= + Γ

= ∇

∫ ∫

∫ ∫

∫ ∫

∫ ∫

v

v

v

A

(27)

The time difference approximation (27) will be combined with a standard Ga-lerkin finite element and mixed finite element in the space for ( )xφ and ( )xσ , respectively [27] [28]. We discrete ( )xφ in space on a quasi-uniform

finite element mesh hφ of Ω with maximal element diameter hφ . For ( )xσ ,

we denote as 0h >σ and hσ similarly. Let ( ) ( )1 ;

D

k lh hV W H H divφ Γ× ⊂ Ω × Ω

σ

be finite element spaces with index k and l, respectively. We define a bilinear form 1 2n

h+ on k l k

h h hV W Vφ σ φ× × and a linear form 1 2n

h+

on khVφ

for 0, , 1n N= − by

T. J. Sun

1309

( )

( )( ) ( )( )( ) ( )

( )

1 11 2

1 1 1

11 2

, , , ,2

, ,2 2

1, , ,

2 2

,2

R

n n n n n nn h h RK h h Eh h h h h h

n n n n n nh E h h E h

n n n n n nh h E h h

h h

n n nn E

h h h

X Xt

t tL X div X

r r X tdiv

gf f X

φ φ σ σφ σ ψ ψ ψ

σ ψ σ ψ

φ φ φ φψ α ψ

ψ ψ

+ ++

+ + +

Γ

++

− +≡ + ∇

∆ ∆ ∆

+ ∇ + ∇ ⋅

+ + + ∆ + +

+≡ +

v

v

( )1 11, ,

2R

n n n

h

g tdivψ

+ +

Γ

+ + ∆

v

(28)

where , ,k l kh h h h h hV W V

φ σ φφ σ ψ∈ ∈ ∈ .

Then, the fully discrete scheme reads: Given 0 kh hVφ ∈ , find

1,

Nn nh h nφ σ

=∈

k lh hV Wφ σ× such that

( ) ( )

( ) ( )

1 2 1 2

1 1 1

, , , ,

, , , .

n n kh h h h h h h h

n n lh h h h h h

V

Wφ

σ

φ σ ψ ψ ψ

σ χ φ χ χ

+ +

− + +

= ∀ ∈

= ∇ ∀ ∈

A (29)

Throughout the analysis, K will denote a generic positive constant, indepen-dent of , ,h h tφ σ ∆ . Similarly, ε will denote a generic small positive constant.

3. Stability of the Approximate Scheme

In this section, we derive the stability of the approximate scheme (29). In order to develop the stability, some assumptions on the different terms of (1) are re-quired.

Claim 2. The velocity field ( )( )0 2,C W ∞∈ Ωv and satisfies 0=v on Γ . Remark. Throughout this paper 1c denotes the maximum between the

positive constant appearing in Lemma 10 and the norm of the velocity in ( )( )0 2,C W ∞ Ω .

Claim 3. The diffusion matrix coefficients, ijA , belong to ( )1,W ∞ Ω . More- over, A is a positive definite symmetric d d× matrix and there exists a strictly positive constant δ which is a uniform lower bound for the eigenvalues of

1−A . As a consequence of Claim 3, there exists a unique positive definite symmetric

d d× matrix function C , such that 1 2− =A C and ( )1,ijC W ∞∈ Ω . Let us in-

troduce the constant ( ) 1,

22 ,

: max ij Wi jc C ∞ Ω

= . Clearly, Claim 3, we have

( )2 2 2120 0 0, , .dw w w w c w wδ −≤ = ≤ ∀ ∈A C R (30)

Claim 4. The reaction function, ( )1,r W ∞∈ Ω , satisfies ( )0 r xγ< ≤ in Ω , where γ is a constant.

Under the previous claims, let ( )1,

2

3 : .W

c r∞ Ω

=

Claim 5. The source function ( )( )0 2f C L∈ Ω . In Robin boundary condition, ( )( )0 2

Rg C L∈ Γ and 0α > . For a given series of functions 0

Nnψ=

, we define the following norms

T. J. Sun

1310

( ) ( )

( ) ( )

2 2 2

1

1 22

0 =0

21 1

max , ,

, , .

Nn n

l L l Ln N n

n n nt

t

t

ψ ψ ψ ψ

ψ ψ ψ ψ ψ ψ

∞

−

≤ ≤

+ −

= = ∆

∂ = − ∆ =

∑

A A

Similar definitions are considered for functional spaces ( )( )2Rl L∞ Γ and

( )( )2 2Rl L Γ associated to the Robin boundary condition. Moreover, we define

( )( )2 2

1 21 2

0,0.

R R

Nn

t tl Ln

tψ ψ−

Γ Γ=

∂ = ∆ ∂ ∑

Lemma 11. Let the above Claims 2 - 5 and 1 1 2c t∆ ≤ be assumed. If

1,

Nn nh h nφ σ

= be the solution of (29). Then it holds that

( )

( ) ( )

( )

1 2 1

22 2 2

0,

2 21 1

2 21 1 1

0,

2 2 2 2 21 1

2

, ,

12 4 4 4

12 41 14 4

2

R

R

n nh h h h

n n n nt h h h h

n n n n n nh h RK h h E

n n n n n nh h E h h

n n n n nh h h h h

nh

t t tr

X Xt

r r X tdiv

c c t c t

c t r r

φ

φ σ φ

δ αφ σ φ φ

δφ φ σ σ

αφ φ φ φ

φ φ σ σ φ

φ

+ +

Γ

+ +

+ + +

Γ

+ +

∆ ∆ ∆ ≥ ∂ + + +

+ − + +∆

+ + + + + ∆

− + + + ∆ + ∆ ∇

+ ∆ +

v

( )2 21

0,,

R

n nh hφ α φ+

Γ+

(31)

the constant c is given by 1 2 1 3max 1, , 2 , .4

c c c c cδγ = −

Proof. Substituting 1nh hψ φ += into (28), we have

( )

( )( ) ( )( )

( ) ( )

1 2 1

1 11 1 1

1 1

1 1 11 1

1 2 3

, ,

, ,2

, ,2 2

1, ,

2 2R

n nh h h h

n n n n n nn nh h RK h h Eh h

n n n n n n n nh E h h E h

n n n n n nh h E h hn n

h h

X Xt

t tL X divv X

r r X tdivv

I I I

φ σ φ

φ φ σ σφ σ

σ φ σ φ

φ φ φ φφ α φ

+ +

+ ++ − +

+ +

+ + ++ +

Γ

− += +

∆

∆ ∆+ ∇ + ∇ ⋅

+ + + ∆ + +

= + + +

A

4 5 6 .I I I+ +

(32)

Lemma 4.1 in [23] implies that

( )2 2 21 11

2 2 211

1 12 2

1 1 ,2 2 2

n n n n n nh h RK h h RK

n n n n nt h h h h RK

I X Xt t

c Xt

φ φ φ φ

φ φ φ φ

+ +

+

= − + −∆ ∆

≥ ∂ − + − ∆

(33)

and

T. J. Sun

1311

( )( )( )

1 1 1

1

1

2 2 21 12

2 2 21 11 2

2 2 212

1 14 41 114 4

12 .4 4 4

n n n n n nh h E h h E

n n n n nh h h h E

n n n n nt h h h h E

I X X

c t c X

t c X

σ σ σ σ

δ σ σ σ σ

δ δσ σ σ σ

− − −

−

−

+ +

+ +

+

= − + +

≥ − + ∆ + +

∆ ≥ ∂ + − + +

A A A

A

A

(34)

Next, by using 1 1c t∆ < , we obtain

2 21 113 11 .

2 2n n n n n

h h h hc ttI c t L σ φ σ φ+ +∆∆

≤ + ∆ ∇ ≤ + ∇

Then when 3 0I ≥ and 3 0I < , we have

2 2113 .

2n nh h

c tI σ φ +∆≥ − + ∇ (35)

Similarly, for 4I we obtain the estimate 2 211

41 .

2 4n nh h

c tI σ φ +∆≥ − − (36)

Analogous computations to term 2I give

( )

( )

22 15

2 211 1 3

14 4

max , .2

n n n nt h h h E

n nh h

tI r r r X

tc c c r r

φ φ φ

γ φ φ

+

+

∆ ≥ ∂ + +

∆− +

(37)

For the boundary integral term 6I , we first use some properties of the inner product in the space ( )2

RL Γ and the inequality ( )21 11 1 3c t c t+ ∆ ≤ + ∆ to get

the estimate

( ) ( ) ( )2 2 221

1 10, 0,0,1 1 1 3

R RR

ntdiv c t c tψ ψ ψ+Γ ΓΓ

+ ∆ ≤ + ∆ ≤ + ∆v

for ( )2RLψ ∈ Γ . Thus, we obtain

( ) 22 1 16 0, 0,

2

1 0,

14 4

3 .4

R R

R

n n n nt h h h

nh

tI tdiv

c t

α αφ φ φ

α φ

+ +

Γ Γ

Γ

∆ ≥ ∂ + + + ∆

− ∆

v (38)

Then, by summing up from (33) to (38), inequality (31) follows. By using Lemma 11 and following the arguments to Lemmas 5.6, 5.7 and

Theorem 5.8 in [23], we can get the following stability theorem: Theorem 12 (Stability Theorem) Let the above Claims 2 - 5 assumed, and

1

,Nn n

h h nφ σ

= be the solution of (29) subject to the initial value 0

hφ . Then there exist two positive constants c and ( )1 2 3, , , ,d d c c cδ γ= , such that, for t d∆ < , we have

( ) ( ) ( ) ( )( )

( ) ( )( ) ( )( )

2 2 22

2 2 2 2 2 2

0 0 0 0

0,

14 4 162

12 4 4 16

.

R

R

R R

h h h hl L l L l Ll L

h h h h

tl L l L l L

t t tr

t t tc r

f g t g

δ αφ σ φ φ

δ αφ σ φ φ

∞ ∞ ∞∞ Γ

Γ

Γ Γ

∆ ∆ ∆+ + +

∆ ∆ ∆≤ + + +

+ + + ∆ ∂

(39)

T. J. Sun

1312

4. Error Estimate Theorem

Now, we turn to derive a priori error estimate in L2-norm for the solutions of (29). In order to state error estimates, we need two following Lagrange interpo-lation operators ([29] [30]) ( )0: k

h hC Vφ

Π Ω → , and ( )0:c

lh hP C WΩ → .

Lemma 13. There exist positive constants 1K and 2K , independent of hφ and hσ respectively, such that

( ) ( )1 1 01 1 , 0,1, ,k s k

h ks K h s H Cφψ ψ ψ ψ+ − ++

Π − ≤ = ∀ ∈ Ω ∩ Ω (40)

and

( ) ( )1 1 02 1 , 0,1, .l m l

h lmP K h m H Cσσ σ σ σ+ − ++

− ≤ = ∀ ∈ Ω ∩ Ω (41)

Let , ,n n n n n nh h h h he φ φ η φ φ= −Π = −Π , .n n n n n n

h h h h hP Pθ σ σ ρ σ σ= − = − Corresponding to 1 2n

h+ and 1 2n

h+ , we introduce a bilinear form 1 2nA +

on ( ) ( ) ( )1 1;D D

H H div HΓ ΓΩ × Ω × Ω and a linear form 1 2n+ on ( )1D

HΓ Ω for 0, , 1n N= − as follows

( )1 11 1 1 121 2 2 2 2 2

T1 1 1 1 12 2 2 2 2

11 12 2

1

d, , , ,d

, ,

, ,

R

nn n n nne e e

n n n n n

e e e

n n

e

n

X F Xt

div F X r X

det F

φφ σ ψ ψ σ ψ

σ ψ φ ψ

α φ ψ

−++ + + ++

−+ + + + +

−+ +

Γ

+

≡ + ∇ + ⋅ + +

( )11 1 1 1

2 2 2 2 2, , .

R

n n n n

e ef X det F gψ ψ ψ−

+ + + +

Γ

≡ +

(42)

If φ and σ φ= ∇A are the solutions of (1), we have for 0, , 1n N= −

( ) ( ) ( )1 2 1 2 1, , , .D

n n Hφ σ ψ ψ ψ+ +Γ= ∀ ∈ Ω (43)

We decompose ,h he θ as

( ) ( )( ) ( )

( )( )

1 2 1 2 1 2 1 2

1 2 1 2

, , , ,

, , ,

n n n nh h h h h h h h h h

n nh h

e θ ψ ψ η ρ ψ

φ σ ψ

+ + + +

+ +

= − +

+ −

(44)

where kh hV

φψ ∈ .

In order to estimate the terms on the right-hand side of (44), we adopt the following lemmas.

Lemma 14. [23] Assume the above Claims 2 - 5 hold, and that the coefficients of the problem (1) satisfy ( )( ) ( )( ) ( )( )0 3, 1 2, 2C W C W C L∞ ∞ ∞∈ Ω ∩ Ω ∩ Ωv ,

0,Γ=v ( )3, ,W ∞∈ ΩA ( )2,r W ∞∈ Ω and that ( )( )0 1, 1 2C W t∞ Ω

∆ <v . Let the

solution of (43) satisfy 3Zφ ∈ , 3Zφ∇ ∈ , ( )( )2 2

R RC LφΓ∈ Γ . Then, for each

0,1, , 1,n N= − there exist two functions 12

1 :n

Rξ+

Ω→ and 12

2 :n

R Rξ+

Γ → , such that

T. J. Sun

1313

( )( )1 1

1 2 1 2 2 21 2, , , , ,

R

n nn nh φ σ ψ ξ ψ ξ ψ

+ ++ +

Γ

− = +

(45)

( )1D

Hψ Γ∀ ∈ Ω . Moreover, ( )1

221

nLξ

+∈ Ω , ( )

122

2

n

RLξ+∈ Γ and the following

estimates hold:

( ) ( )

( ) ( )( )

3 3 2

2 2 2

122

1 1

122

2 1 ,0,

,

,R R

R

n

Z Z Z

n

Z C L

c t r

c t

ξ φ σ φ

ξ σ α φ

+

+

Γ ΓΓ

≤ ∆ + +

≤ ∆ ⋅ +

n (46)

where 1c denotes a constant independent of t∆ . Lemma 15. [23] Assume the above Claims 2 - 5 hold, and that the coefficients

of the problem (1) satisfy ( )( ) ( )( )0 2, 1 1,C W C W∞ ∞∈ Ω ∩ Ωv , 2 ,f Z∈ g ∈ ( )( )2 2

RC L Γ and ( )( )0 1, 1 2C W t∞ Ω∆ <v . Then, for each 0,1, , 1,n N= − there

exist function 12 :

n

f Rξ+

Ω→ and 12 :

n

g R Rξ+

Γ → , such that

( )( ) ( )1 1

1 2 1 2 12 2, , , .R

n nn nh f g Hψ ξ ψ ξ ψ ψ

+ ++ +

Γ

− = + ∀ ∈ Ω

(47)

Moreover, ( ) ( )1 1

2 22 2,n n

f g RL Lξ ξ+ +∈ Ω ∈ Γ and the following estimates hold:

( )

( ) ( )( )

2

2 2

122

1

122

10,

,

,R

R

n

f Z

n

g C L

c t f

c t g

ξ

ξ

+

+

ΓΓ

≤ ∆

≤ ∆

(48)

where 1c denotes a constant independent of t∆ . Now, we turn to bound the third term on the right-hand side of (44). Lemma 16. Assume the above Claims 2 - 5 hold, and that the coefficients of

the problem (1) satisfy ( )( ) ( )( ) ( )( )0 0 0 1 1k kC C C H H Hφ +∈ Ω ∩ Ω ∩ Ω and

1 1 2c t∆ < . There exists

( )( ) ( )

( ) ( )

( ) ( )

( )

1 2 1

21

21

21 1

0,

2 22 2 2 21 1 1

0,

21

, ,

1 ,8 2

1 ,8 2

1 ,8 2 RR

R

n nh h h h

n n n n nh h E t h h

n n n n nh h E t h h

n n n n nh h t h h

n n n n n nh h h h

e

tC C X C C

tre re X r re

te e tdiv e

c e c t re re e

cK hφ

η ρ

θ θ ρ θ

η

α αη

θ θ α

+ +

+

+

+ +

ΓΓ

+ + +

Γ

∆ ≤ + + ∂

∆ + + + ∂

∆ + + + ∆ + ∂

+ + ∆ + + + +

+

v

( ) ( )2 1 12 1

2222

, ; 1, ;

2 22 2 12 1

1 1

,

n n kn n k

k nL t t H k

L t t H

l n nhl

tt t t

cK h c t eσ

φφ φ

σ

+ ++ +

+

+

∂ + + ∆ ∆ ∂ ∆

+ + ∆

(49)

T. J. Sun

1314

with ( ) 1 2 1 3 1max 1, 1 4 , 4,c c c c c cδ= + , c a positive constant.

Proof. From the definition of 1 2nh+ , we have

( )

( )( ) ( )( )( ) ( )

1 2 1

1 11 1 1

1 1

1 1 11 1

1 2 3

, ,

, ,2

, ,2 2

1, ,

2 2R

n nh h h h

n n n n n nn nh h RK h h Eh h

n n n n n n n nh E h h E h

n n n n n nh h E h hn n

h h

e

X Xe

t

t tL X e div X e

r r X tdive e

I I I

η ρ

η η ρ ρθ

ρ ρ

η η η ηα

+ +

+ ++ − +

+ +

+ + ++ +

Γ

− += +

∆

∆ ∆+ ∇ + ∇ ⋅

+ + + ∆ + +

= + + +

A

v

v

4 5 6 .I I I+ +

(50)

From the arguments of Lemma 4.1 in [24], we get similarly the bounds of the above terms 1 5 6, ,I I I as the followings:

( )( )( ) ( )2 1 1

2 1

2 2 2 221 1 1

1 , ;, ;

21

1 1 22

1 ,2

n n kn n k

k

L t t HL t t H

nh

c t c K hI

t t

e

φ φ φ + ++

+

+ ∆ + ∂ ≤ + ∆ ∂

+

(51)

( ) ( )

( ) ( )( )

22 21 35 1 3 1 1

22 21 1 3 1

21

, 12 4

2

81 ,8

n n k nt h h k

n nh h

n n nh h E

c c ttI r re c t c K h

c c t c tre re

re re X

φη φ+

+

+

∆∆ ≤ ∂ + + + ∆

+ ∆ ∆+ +

+ +

(52)

( ) ( )

( ) ( )

( )

21 16 0,

2221 1 1 1 2 2

1 1

21 1

0,

, 12 8

2 1 24 2

2.

4

R R

R

n n n n nt h h h h

k nk

nh

tI e e e tdiv

c t c t c t c tc K h

c t c te

φ

α αη

α φ

α

+ +

Γ Γ

Ω +

Γ

∆ ≤ ∂ + + + ∆ ∆ + ∆ + ∆ + ∆ + + ∆ + ∆

+

v

(53)

To estimate 2I , we divide it into three parts

( ) ( )1 12 21 22

1 , , ,2

n n n nh h h hI C C C C I Iρ θ ρ θ+ + = − + + (54)

where

( ) ( ) ( )( )

( )( )( )

21

122

1 1, , ,2 21 , .2

n n n n n nh h h E h E

n n n n n nh E E h h E

I C C C X C X

I C X C X X

ρ θ ρ θ

ρ θ θ+

= −

= +

Using the transformation ( )nEy X x= , we have

T. J. Sun

1315

( ) ( )( ) ( )( ) ( )( )

( )

1

21

2 22 21 22 1

1 1 d2

.4

n n nE h h

l n nhl

I det F x C x C x x

c c t K hσ

ρ θ

σ θ

−

Ω

+

= − ⋅ ∆

≤ +

∫ (55)

Now, we replace in 22I equality ( )( ) ( ) ( )n nEC X x C x D x= − , where

( ) ( )( ) ( )1

1: , d , . . ,n

n

tn n nij ij Et

D x C Y x s x s a e x+

+= ∇ ⋅ ∈Ω∫ v (56)

with ( ) 1 2nijD x c c t≤ ∆ , and the function ( ) 1, : ,n n n

EY x t t + ⋅ →Ω is defined by

( ) ( ) ( )1 1, : .n n nEY x s x t s x+ += − − v Thus, we get

( ) ( )( )( ) ( )

1 122

2 2 21 1

1 ,2

1 1 .8 8

n n n n n nh E h h h E

n n n n n nh E h h E h

I C X C D C X

C X C C X D

ρ θ θ θ

ρ θ θ θ

+ +

+ +

= − +

≤ + + +

(57)

Moreover, we have

( ) ( )

( )

2 22 21 2 2 1

2 221 2 11 2

1 ,

.

n n l nh E l

n nh h

C X c t c K h

D c c t

σρ σ

θ θ

+

+ +

≤ + ∆ ≤ ∆

(58)

By considering together from (54) to (58), we can state

( ) ( )

( ) ( ) ( )

22 21 22 1 2 2 1

222 21 1 2 1 1

, 12 4

2 1 .8 8

n n l nt h h l

n n n n nh h h h E

c c ttI C C c t c K h

c c t c tC C X

σρ θ σ

δ θ θ θ θδ

+

+ +

∆∆ ≤ ∂ + + + ∆

+ ∆ ∆+ + + +

(59)

Similar arguments, i.e. Lemma 5.4 in [23] lead to

( ) ( )

( ) ( )

2 22 2 2 21 1 2 2 1

2 22 2 2 21 1 2 2 1

1 ,

1 .

n n n l nh E l

n n n l nh E l

L X c t c c K h

div X c t c c K h

σ

σ

ρ σ

ρ σ

+

+

≤ + ∆ ∇ ⋅ ≤ + ∆

v (60)

Using these inequalities, 3I and 4I can be bounded as follows

( )

( )

2 2 2 22 21 1 2 2 1

3 1

2 2 2 22 21 1 2 2 1

4 1

1,

4 41 1 .

4 4

ln n

hl

ln n

hl

c t c c K h tI

c t c c K hI

σ

σ

σ δ θδ

σ θ

+

+

+

+

+ ∆ ∆≤ +

+ ∆ ≤ +

(61)

Gathering (50), (51), (52), (53), (59), and (61) together, we complete the proof.

We now turn to estimate 1nhe + and 1n

hσ+ . From the definition of 1 2n

h+ and

1 2n+ , we see

( ) ( ) ( )( )( )( )

1 2 1 1 2 1 1 2 1 2 1

1 2 1 2 1

, , , ,

, , .

n n n n n n nh h h h h h h h h h

n n nh h

e e e e

e

σ η ρ

φ σ

+ + + + + + +

+ + +

= + −

+ −

(62)

From Lemma 11, we obtain the lower bound for (62)

T. J. Sun

1316

( )

( ) ( )

( )

1 2 1

22 2 2

0,

2 21 1

2 21 1 1

0,

2 2 2 2 21 1

2

, ,

12 4 4 4

12 41 14 4

2

R

R

n nh h h h

n n n nt h h h h

n n n n n nh h RK h h Ee

n n n n n nh h E h h

n n n n nh h h h h

nh

e e

t t te re e

e e X Xt

re re X e e tdiv

c e e c t c t e

c t re r

σ

δ αθ

δ θ θ

α

θ θ

+ +

Γ

+ +

+ + +

Γ

+ +

∆ ∆ ∆ ≥ ∂ + + +

+ − + +∆

+ + + + + ∆

− + + + ∆ + ∆ ∇

+ ∆ +

v

( )2 21

0,.

R

n nh he eα+

Γ+

(63)

By jointly considering the lower bound (63), the upper bounds given in Lemmas 14 - 16, we deduce

( )( ) ( )

( ) ( )

22 2 2

0,

2 2 2 2 21 1

2 2 21

0,

22 21

12 4 4 4

2

,2

, ,2 2

R

R

R

n n n nt h h h h

n n n n nh h h h h

n n n n nh h h t h h

n n n nt h h t h h

n n nh h

t t te re e

c e e c t c t e

tc t re re e C C

t tr re e

c t re re

δ αθ

θ θ

α ρ θ

αη η

θ θ

Γ

+ +

+

Γ

Γ

+

∆ ∆ ∆ ∂ + + + ≤ + + + ∆ + ∆ ∇

∆ + ∆ + + + ∂

∆ ∆ + ∂ + ∂

+ ∆ + + +( )

( ) ( )2 1 12 1

2 21

0,

2222 2

1 , ; 1, ;

2 2 22 2 1 12 1

2 2 2 21 1 1 12 2 2 2

1 20, 0,

1 1

.

R

n n kn n k

R R

n nh

k nL t t H k

L t t H

l n n nh hl

n n n n

f g

e

cK h tt t t

cK h c t e c e

c

φ

σ

α

φφ φ

σ

ξ ξ ξ ξ

+ ++

+

Γ

+

+ +

+

+ + + +

Γ Γ

+

∂ + + + ∆ ∆ ∂ ∆

+ + ∆ +

+ + + +

(64)

Analogous computations to those developed in Lemma 16 give

( )

( )( )

2 2 2 22 21

2 2223 1

1

2 22 21 1 0,

, ,2 2 8

, ,2 2 8

, , 0, .2 16 RR

lj j j j

h h hl

kj j j j

h h hk

j j k j jh h hk

c K h tt tC C C

c K h tt tr re re

t te c K h t e j N

σ

φ

φ

ρ θ σ θ

η φ

α αη α φ

+

+

Ω + ΓΓ

∆∆ ∆≤ +

∆∆ ∆ ≤ + ∆ ∆

≤ ∆ + ∀ =

(65)

Putting (65) into (64), multiplying (64) by t∆ , summing it about time form 0t = and Nt t= , taking the initial values 0 0 0 0,h h h hPφ φ σ σ= Π = and using dis-

crete Gronwall’s inequality, we can derive the following error estimate: Theorem 17 (Error Estimate) Let Claims 2 - 5 be assumed. Let 3Zφ ∈ ∩

( )( ) ( )( )0 1 1k kC H H H+ Ω ∩ Ω be the solution of (1) with 3 ,Zφ∇ ∈

( )3

R RZφΓ∈ Γ ,

1,

Nn nh h nφ σ

= be the solution of (29) subject to the initial values

T. J. Sun

1317

0 0 0 0,h h h hPφ φ σ σ= Π = . Then there exists a positive constant c independent of ,h hφ σ and t∆ such that

( ) ( )

( ) ( )( )

( ) ( ) ( )

( ) ( )( ) ( )

2 2

22

1 0 1 0 1

3 3 2 2 2 2

3 3

2, ,

52

, ,

12 8

8 16

.

R

k k l

R R

R R

h hl L l L

h h l Ll L

k lH H C H C H

Z Z Z Z Z Z

Z Z

t

t tr r

ch ch

c t r f g

c t g

φ σ

δφ φ σ σ

αφ φ φ φ

φ φ σ

φ σ φ φ

φ

∞ ∞

∞∞

+ +

Γ

Γ Γ

Γ Γ

∆− + −

∆ ∆+ − + −

≤ + +

+ ∆ + + + + +

+ ∆ +

(66)

Acknowledgement

This research was supported by the NSF of China (Grant Nos. 11271231 and 11301300).

References [1] Frolkovic, P. and De Schepper, H. (2000) Numerical Modelling of Convection

Dominated Transport Coupled with Density Driven Flow in Porous Media. Ad-vances in Water Resources, 24, 63-72. https://doi.org/10.1016/S0309-1708(00)00025-7

[2] Markowich, P.A. and Szmolyan, P. (1989) A System of Convection-Diffusion Equa-tions with Small Diffusion Coefficient Arising in Semiconductor Physics. Journal of Differential Equations, 81, 234-254. https://doi.org/10.1016/0022-0396(89)90122-8

[3] Morton, K.W. (1996) Numerical Solution of Convection-Diffusion Problems. Chapman & Hall, London.

[4] Ewing, R.E. and Wang, H. (2001) A Summary of Numerical Methods for Time- Dependent Advection-Fominated Partial Differential Equations. Journal of Com-putational and Applied Mathematics, 128, 423-445. https://doi.org/10.1016/S0377-0427(00)00522-7

[5] Wilmott, P., Dewynne, J. and Howison, S. (1993) Option Pricing, Mathematical Models and Computation. Oxford Financial Press, Oxford.

[6] Ewing R. (1983) The Mathematics of Reservoir Simulation. SIAM, Philadelphia.

[7] Johnson, C. (1987) Numerical Solution of Partial Differential Equations by the Fi-nite Element Method. Cambridge University Press, Cambridge.

[8] Douglas, Jr.J. and Russell, T.F. (1982) Numerical Methods for Convection-Domi- nated Diffusion Problems Based on Combining the Method of Characteristics with Finite Element or Finite Difference Procedures. SIAM Journal on Numerical Analy-sis, 19, 871-885. https://doi.org/10.1137/0719063

[9] Russell, T.F. (1985) Time-Stepping along Characteristics with Incomplete Itera-tionfora Galerkin Approximation of Miscible Displacement in Porous Media. SIAM Journal on Numerical Analysis, 22, 970-1013. https://doi.org/10.1137/0722059

[10] Pironneau, O. (1982) On the Transport-Diffusion Algorithm and Its Application to the Navier-Stokes Equations. Numerische Mathematik, 38, 309-332. https://doi.org/10.1007/BF01396435

[11] Süli, E. (1986) Convergence Analysis of the Lagrange-Galerkin Method for the

T. J. Sun

1318

Navier-Stokes Equations, Report 86/3. Oxford University Computing Laboratory, Oxford.

[12] Morton, K.W., Priestley, A. and Süli, E. (1986) Stability Analysis of the Lagrange- Galerkin Method with Non-Exact Integration, Report 86/14. Oxford University Computing Laboratory, Oxford.

[13] Johnson, C. and Thomée, V. (1981) Error Estimates for Some Mixed Finite Element Methods for Parabolic Type Problems. RAIRO Analyse Numérique, 15, 41-78. https://doi.org/10.1051/m2an/1981150100411

[14] Raviart, P.A. and Thomas, J.M. (1977) A Mixed Finite Element Method for Second Order Elliptic Problems, Mathematical Aspects of the Finite Element Method. Lec-ture Notes in Mathematics 606, Springer, Berlin.

[15] Nedelec, J.C. (1980) Mixed Finite Elements in R3. Numerische Mathematik, 35, 315-341. https://doi.org/10.1007/BF01396415

[16] Douglas, Jr.J. and Roberts, J.E. (1985) Global Estimates for Mixed Methods for Second Order Elliptic Equations. Mathematics of Computation, 44, 39-52. https://doi.org/10.1090/S0025-5718-1985-0771029-9

[17] Ewing, R.E., Russell, T.F. and Wheeler, M.F. (1984) Convergence Analysis of an Approximation of Miscible Displacement in Porous Media by Mixed Finite Ele-ments and a Modified Method of Characteristics. Computer Methods in Applied Mechanics and Engineering, 47, 73-92. https://doi.org/10.1016/0045-7825(84)90048-3

[18] Wheeler, M.F. and Dawson, C.N. (1984) An Operator-splitting Method for Advec-tion Diffusion-Reaction Problems. The Mathematics of Finite Elements and Appli-cations, V, 122-144.

[19] Arbogast, T. and Wheeler, M.F. (1995) A Charcteristics-Mixed Finite Element Me-thods for Advection-Dominated Transport Problems. SIAM Journal on Numerical Analysis, 32, 404-424. https://doi.org/10.1137/0732017

[20] Sun, T.J. and Yuan, Y.R. (2009) An Approximation of Incompressible Miscible Dis-placement in Porous Media by Mixed Finite Element Method and Characteris-tics-mixed Finite Element Method. Journal of Computational and Applied Mathe-matics, 228, 391-411. https://doi.org/10.1016/j.cam.2008.09.029

[21] Sun, T.J. and Yuan, Y.R. (2015) Mixed Finite Element Method and the Characteris-tics-mixed Finite Element Method for a Slightly Compressible Miscible Displace-ment Problem in Porous Media. Mathematics and Computers in Simulation, 107, 24-45. https://doi.org/10.1016/j.matcom.2014.07.005

[22] Rui, H.X. and Tabata, M. (2002) A Second Order Charcteristic Finite Element Scheme for Convection-Diffusion Problems. Numerische Mathematik, 92, 161-177. https://doi.org/10.1007/s002110100364

[23] Bermúdez, A., Nogueiras, M.R. and Vázquez, C. (2006) Numerical Analysis of Convection-Diffusion-Reaction Problems with Higher Order Characteristic/Finite Elements, Part I: Time Discretization. SIAM Journal on Numerical Analysis, 44, 1829-1853. https://doi.org/10.1137/040612014

[24] Bermúdez, A., Nogueiras, M.R. and Vázquez, C. (2006) Numerical Analysis of Convection-Diffusion-Reaction Problems with Higher Order Characteristic/Finite Elements, Part II: Fully Discretized Scheme and Quadrature Formulas. SIAM Jour-nal on Numerical Analysis, 44, 1854-1876. https://doi.org/10.1137/040615109

[25] Brezzi, F. (1974) On the Existence, Uniqueness and Approximation of Saddle-Point Problems Arising from Lagrangian Multipliers. RAIRO Analyse Numérique, 2, 129-151. https://doi.org/10.1051/m2an/197408R201291

T. J. Sun

1319

[26] Wheeler, M.F. (1973) A Priori L2 Error Estimates for Galerkin Approximations to Parabolic Partial Differential Equations, SIAM Journal on Numerical Analysis, 10, 723-759. https://doi.org/10.1137/0710062

[27] Arnolds, D.N., Scott, L.R. and Vogelus, M. (1998) Regular Inversion of the Diver-gence Operator with Dirichlet Boundary Conditions on a Polygonal. Annali della Scuola Normale Superiore di Pisa-Classe di Scienze, 15, 169-192.

[28] Ciarlet, P.G. (1991) Basic Error Estimates for Ellptic Problems, Handbook of Nu-merical Analysis, Volume II. North-Holland Publishing Company, Amsterdam, 19-351.

[29] Russell, T.F. (1986) Finite Elements with Characteristics for Two-component In-compressible Miscible Displacement, SPE 10500. Proceedings of 6th SPE Sympo-sium on Reservoir Simulation, New Orleans, 123-135.

[30] Ciarlet, P.G. (1978) The Finite Element Method for Ellptic Problems. North-Hol- land Publishing Company, Amsterdam.

Submit or recommend next manuscript to SCIRP and we will provide best service for you:

Accepting pre-submission inquiries through Email, Facebook, LinkedIn, Twitter, etc. A wide selection of journals (inclusive of 9 subjects, more than 200 journals) Providing 24-hour high-quality service User-friendly online submission system Fair and swift peer-review system Efficient typesetting and proofreading procedure Display of the result of downloads and visits, as well as the number of cited articles Maximum dissemination of your research work

Submit your manuscript at: http://papersubmission.scirp.org/ Or contact jamp@scirp.org