The Cattaneo-Vernotte equation is a generalization of the heat and particle diffusion equations; this mathematical model combines waves and diffusion with a finite velocity of propagation. In disordered systems the diffusion can be anomalous. In these kinds of systems, the mean-square displacement is proportional to a fractional power of time not equal to one. The anomalous diffusion concept is naturally obtained from diffusion equations using the fractional calculus approach. In this paper we present an alternative representation of the Cattaneo-Vernotte equation using the fractional calculus approach; the spatial-time derivatives of fractional order are approximated using the Caputo-type derivative in the range (0,2]. In this alternative representation we introduce the appropriate fractional dimensional parameters which characterize consistently the existence of the fractional space-time derivatives into the fractional Cattaneo-Vernotte equation. Finally, consider the Dirichlet conditions, the Fourier method was used to find the full solution of the fractional Cattaneo-Vernotte equation in analytic way, and Caputo and Riesz fractional derivatives are considered. The advantage of our representation appears according to the comparison between our model and models presented in the literature, which are not acceptable physically due to the dimensional incompatibility of the solutions. The classical cases are recovered when the fractional derivative exponents are equal to 1.
1. Introduction
Fourier’s law satisfies the heat conduction induced by a small temperature gradient in steady state. In steady state, the heat transfer through a material is proportional to the negative gradient of the temperature and to the area. However, there are some cases in which the Fourier equation is not adequate to describe the heat conduction process. More precisely, Fourier law is diffusive and cannot predict the finite temperature propagation speed in transient situations, in this context, the Cattaneo-Vernotte equation corrects the nonphysical property of infinite propagation of the Fourier and Fickian theory of the diffusion of heat, and this equation also known as the telegraph equation for the temperature is a generalization of the heat diffusion (Fourier’s law) and particle diffusion (Fick’s laws) equations. Processes where the traditional Fourier heat equation leads to inaccurate temperature and heat flux profiles are known as non-Fourier type processes [1]; these processes can be Markovian or non-Markovian [2]. In the Markovian processes case, the mean-square displacement of the diffusing particle is proportional to time, while, in the disordered systems or non-Markovian process, the diffusion can be anomalous; in this case the mean-square displacement is proportional to a fractional power of time not equal to one; when anomalous diffusion occurs, the probability density for a diffusing particle is not the usual Gaussian distribution. The mechanism of diffusion is Brownian motion, and this motion is the simplest continuous-time stochastic process. Continuous-time random walks can be coupled with Brownian motion and fractional calculus (FC) to provide an improved estimator in the modeling of anomalous diffusion. A random walk is a mathematical formalization of a path that consists of a succession of random steps [3]. A Lévy flight, also referred to as Lévy motion, is a random walk in which the step-lengths have a heavy-tailed probability distribution. When a random walk is defined as a walk in a space of dimension greater than one, the steps are defined in terms of a probability distribution, and steps move with isotropic random directions [4], and continuous-time random walk schemes are considered in the derivation of time-fractional differential equations. Recently, the subject of FC has attracted interest of researches; this mathematical concept involves nonlocal operators which can be applied in physical systems yielding new information about its behavior, fractional derivatives with respect to coordinates describe power-law nonlocal properties of the distributed system, and there are several papers about the recent history of the FC; see [5–7]. Several approaches have been used for investigating anomalous diffusion, Langevin equations [8, 9], random walks [10, 11], or fractional derivatives, based on FC several works connected to anomalous diffusion processes which may be found in [12–30]. Using phenomenological arguments, Compte and Metzler [31] generalize the Cattaneo-Vernotte equation by introducing fractional derivatives with a continuous-time random walk scheme. The authors of the works presented in [32–34] studied the generalized Cattaneo-Vernotte equation with fractional space-time derivatives, and the order of the spatial and temporal fractional derivatives are β,γ∈(0;2]. Lewandowska and Kosztołowicz in [35] investigate the subdiffusive impedance phenomena of a spatially limited sample for large pulsation of electric field. Tarasov in [36] based on the Liouville equation obtained the fractional analogues of the classical kinetic and transport equations. Qi and Jiang in [37] derived the exact solution of the Cattaneo-Vernotte equation by joint Laplace and Fourier transforms. Other applications of FC to Cattaneo-Vernotte equation are given in [38–41].
The aim of this work is to contribute to the development of a new version of fractional fundamental Cattaneo-Vernotte equation applying the idea proposed in the work [42]; the order considered is (0,2] for the fractional equation in space-time domain; this representation preserves the dimensionality of the equation for any value taken by the exponent of the fractional derivative.
The paper is structured as follows: in Section 2 we explain the basic concepts of the fractional calculus; in Section 3 we present the fractional Cattaneo-Vernotte equation and give conclusions in Section 4.
2. Basic Definitions of Fractional Calculus
The most commonly used definitions in FC are Riemann-Liouville (RL), Grünwald-Letnikov (GL), Caputo fractional derivative (CFD), and Riesz fractional derivative (R) [43–46].
The RL definition of the fractional derivative for (φ>0) is(1)DaRLtφft=1Γm-φdmdtm∫atfηt-ηφ-m+1dη,m-1<φ<m.
For function f(t) the CFD is given by(2)DaCtφft=1Γn-φ∫atfnη-η+tφ-n+1dη,n-1<φ≤n,where dφ/dtφ=DaCtφ is a CFD with respect to t, φ∈R is the order of the fractional derivative, n=1,2,…∈N, and Γ(·) represents Euler’s gamma function.
In the present paper, we would use the CFD definition, since the former is more popular in real applications. For the CFD definition we need to specify additional conditions in order to produce a unique solution, these additional conditions are expressed in terms of integer-order derivatives [46], and this definition is used mainly for the problem with memories. In the case of the RL definition there exist physically unacceptable initial conditions [47].
The Laplace transform of Caputo’s derivative (2) has the form [48](3)LD0Ctφft=sφFs-∑k=0n-1sφ-k-1fk0,where F(s) is the Laplace transform of the function f(t) and n=[R(φ)]+1. From this expression we have two particular cases:(4)LD0Ctφft=sφFs-sφ-1f00<φ≤1,(5)LD0Ctφft=sφFs-sφ-1f0-sφ-2f′01<φ≤2.
The Mittag-Leffler function has gained extensive interest among physicists due its vast potential of applications in the solution of fractional differential equations [48]:(6)Eα,θt=∑m=0∞tmΓαm+θ,α>0,θ>0;when α=1 and θ=1, from (6) we obtain the exponential function.
erfc(α) denotes the complementary error function [48] and it is defined as(7)erfcα=2π∫0αexp-t2dt.
Some common Mittag-Leffler functions are described in [48]:(8)E1/2,1±α=expα21±erfcα,(9)E1,1±α=exp±α,(10)E2,1-α2=cosα,(11)E3,1α=12expα1/3+2exp-12α1/3cos32α1/3,(12)E4,1α=12cosα1/4+coshα1/4.
The Mittag-Leffler function Es/2,r is defined by Miller in [49]:(13)Es/2,rz=z2κ1-rs∑j=0s-1αj1-s/2+rexpαjz2καjs/2+erfcαj1/2zκ-z-2n∑k=02n-1zkΓsk/2+μ,where κ=1/s, r=ns+μ, n=0,1,2,3,…, μ=1,2,3,…. For the calculation of generalized Mittag-Leffler functions at arbitrary precision, see [50, 51].
The Riesz fractional derivative for (φ>0) is [43–46](14)DRxφfx=-12cosπφ/2·1Γn-φ∂∂xn∫-∞xfχ,tdχx-χφ+1-n+-1nΓn-φ∂∂xn∫x∞fχ,tdχχ-xφ+1-n,n-1<φ≤n,where dφ/dxφ=DRxφ is a Riesz fractional derivative with respect to x, φ∈R is the order of the fractional derivative, n=1,2,…∈N, and Γ(·) represents Euler’s gamma function.
3. Fractional Cattaneo-Vernotte Equation
In previous studies of the fractional Cattaneo-Vernotte equation the authors did not consider the physical dimensionality of the solutions. The authors of the work [42] proposed a systematic way to construct fractional differential equations for the physical systems. To keep the dimensionality of the fractional differential equations a new parameter σ was introduced in the following way:(15)∂∂x⟶1σx1-φ·∂φ∂xφ,n-1<φ≤n,n∈N=1,2,3,…,(16)∂∂t⟶1σt1-φ·∂φ∂tφ,n-1<φ≤n,n∈N=1,2,3,…,where φ is an arbitrary parameter which represents the order of the derivative, σx has dimension of length, and σt has the dimension of time. These new parameters maintain the dimensionality of the equation invariant and characterizes the fractional space or fractional temporal structures (components that show an intermediate behavior between a conservative system and dissipative one) [42]; when φ=1 the expressions (15) and (16) reduce to the ordinary derivative. In the following we will apply this idea to generalize the case of the fractional Cattaneo-Vernotte equation.
In this work, we consider generalized Cattaneo-Vernotte equation in the x direction of the form [32–35](17)∇2T-1DT˙-τDT¨=0,where τ is a characteristic relaxation time constant (or the non-Fourier character of the material) and D is the generalized thermal diffusivity; (17) is a hyperbolic diffusion equation; when the parameter τ=0, (17) recovers a parabolic form; in this limit, one has to replace Cattaneo-Vernotte equation by Fourier’s heat transfer equation.
Considering the CFD (2) and (15) and (16), the fractional representation of (17) is(18)1σx21-φD0Cx2φTx,t-1D·1σt1-φD0CtφTx,t-τD·1σt21-φD0Ct2φTx,t=0.The order of the derivative considered is φ∈(0,2] for the fractional Cattaneo-Vernotte equation in space-time domain.
3.1. Fractional Space Cattaneo-Vernotte Equation
Considering (18) and assuming that the space derivative is fractional equation (15) and the time derivative is ordinary, the spatial fractional equation is(19)D0Cx2φTx,t-1Dσx21-φ·∂Tx,t∂t-τDσx21-φ·∂2Tx,t∂t2=0.Suppose the solution(20)Tx,t=T0·expiωtux;substituting (20) into (19) we obtain(21)d2φuxdx2φ+τDω2-i1Dωσx21-φux=0,where(22)kx2=τDω2-i1Dω,is the dispersion relation in the x direction and(23)k~x2=τDω2-i1Dωσx21-φ=kx2σx21-φ,is the fractional dispersion relation; from the fractional dispersion relation (23), we can expect the fractional wave number k~ in the x direction to have real and imaginary parts, δx and βx, respectively. Let us write(24)k~x=δx-iβx;substituting (24) into (23) we have(25)δx-iβx2=δx2-2iδxβx-βx2,where(26)δx2-2iδxβx-βx2=τDω2-i1Dωσx21-φ;solving for βx we obtain(27)βx=ωD·12δxσx21-φ,and for δx(28)δx=ωτD12±121+1τ2ω21/2σx1-φ;substituting (28) into (27) we have(29)βx=12Dτ/D1/2±1/21+1/τ2ω21/2σx1-φ.
Now the fractional wave number is k~x=δx-iβx, where δx and βx are given by (28) and (29), respectively,(30)k~x=ωτD12±121+1τ2ω21/2σx1-φ-i12Dτ/D1/2±1/21+1/τ2ω21/2σx1-φ;equation (30) describes the real and imaginary part of the fractional wave number in terms of the frequency ω, the relaxation time τ, and the generalized thermal diffusivity D, in presence of fractional space components σx.
Considering (23), (21) gives(31)d2φuxdx2φ+k~x2ux=0,the solution of (31) can be obtained applying direct and inverse Laplace transform [47], and the solution of the above equation is given by(32)ux=E2φ,1-k~x2x2φ,where E2φ,1(-k~x2x2φ) is the Mittag-Leffler function.
Therefore the general solution of (21) is given by(33)Tx,t=T0·expiωt·E2φ,1-k~x2x2φ.
Next, we will analyze the case when φ takes different values.
Case 1.
When φ=2, we have(34)k~x=ωτD12±121+1τ2ω21/2σx-1-i12Dτ/D1/2±1/21+1/τ2ω21/2σx-1,and equation (34) represents the fractional wave number in presence of fractional space components σx.
In this case, (33) is written as follows:(35)Tx,t=T0·expiωt·E4,1-k~x2x4,where E4,1 is given by (12) and solution (35) is(36)Tx,t=T02·expiωt·cos-k~x1/2x+cosh-k~x1/2x.
Case 2.
When φ=3/2, we have(37)k~x=ωτD12±121+1τ2ω21/2σx-1/2-i12Dτ/D1/2±1/21+1/τ2ω21/2σx-1/2,and equation (37) represents the fractional wave number in presence of fractional space components σx.
In this case, (33) is written as follows:(38)Tx,t=T0·expiωt·E3,1-k~x2x3,where E3,1 is given by (11) and solution (38) is(39)Tx,t=T02·expiωt·exp-k~x2/3x+2expk~x2/32x·cos-32k~x2/3x.
Case 3.
When φ=1, we have k~x=kx(40)kx=ωτD12±121+1τ2ω21/2-i12Dτ/D1/2±1/21+1/τ2ω21/2,and equation (40) represents the classical wave number kx. From (33) we have(41)Tx,t=T0·expiωt·E2,1-k~x2x2,where(42)Tx,t=RT0·expiωt·exp-ik~xx,and, in (42), k~x=kx, R indicates the real part, and kx=δx-iβx is wave number (40); substituting kx in (42) we have(43)Tx,t=RT0·expiωt-δxx·exp-βxx.Equation (43) represents the classical case for the space Cattaneo-Vernotte equation. The first exponential exp(i(ωt-δxx)) gives the usual plane-wave variation of the thermal field with position x and time t. The second exponential exp(-βxx) gives and exponential decay in the amplitude of the thermal wave.
Case 4.
When φ=1/2, from (33) we have(44)Tx,t=T0·expiωt·E1,1-k~x2x,where k~x is(45)k~x=ωτD12±121+1τ2ω21/2σx1/2-i12Dτ/D1/2±1/21+1/τ2ω21/2σx1/2,and equation (45) represents the fractional wave number in presence of fractional space components σx.
The solution for (44) is(46)Tx,t=RT0·expiωt+1Dσxx·exp-τDω2σxx,where R indicates the real part.
Case 5.
When φ=1/4, from (33) we have(47)Tx,t=T0·expiωt·E1/2,1-k~x2x1/2,where k~x is(48)k~x=ωτD12±121+1τ2ω21/2σx3/4-i12Dτ/D1/2±1/21+1/τ2ω21/2σx3/4,and equation (48) represents the fractional wave number in presence of fractional space components σx.
In this case, (33) is written as follows:(49)Tx,t=T0·expiωt·E1/2,1-k~x2x1/2,where E1/2,1 is given by (8) and solution (49) is(50)Tx,t=T0·expiωt·expk~4x·1-erfc-k~2x1/2.erfc(α) denotes the error function defined in (8). Equation (50) represents the space evolution of the temperature and the amplitude exhibits an algebraic decay for x→∞.
For this case there exists a physical relation between the auxiliary parameter σx and the wave number kx given by the order φ of the fractional differential equation(51)φ=kxσx=σxλ,0<σx≤λ,where λ is the wavelength. We can use this relation in order to write (33) as(52)Tx~,t=T0·expiωt·E2φ-φ21-φx~2φ,where x~=x/λ is a dimensionless parameter. Figures 1(a) and 1(b) show the simulation of (52) for φ values 0.7<φ≤1 and 1.7<φ≤2, respectively.
Simulation of (52) for 0.7<φ≤2.
Table 1 shows the different solutions of (52). The order of the fractional differential equation is φ=2, φ=3/2, φ=1, φ=1/2, and φ=1/4.
Solutions for fractional space Cattaneo-Vernotte equation (52) for different values of φ; R indicates the real part and erfc(α) denotes the error function defined in (8).
Considering (18) and assuming that the time derivative is fractional equation (16) and the space derivative is ordinary, the temporal fractional equation is(53)D0Ct2φTx,t+1τ·σt1-φD0CtφTx,t-Dτσt21-φ∂2Tx,t∂x2=0;suppose the solution(54)Tx,t=T0expik~xxut,where k~x is the wave number in the x direction. Substituting (54) into (53) we obtain(55)d2φutdt2φ+1τσt1-φdφutdtφ+Dτk~x2σt21-φut=0;the solution of (55) can be obtained applying direct and inverse Laplace transform [47]. Taking solution (54) we have(56)Tx,t=T0·expik~xx·Eφ,1-12τσt1-φtφ·E2φ,1-Dτk~x2-14τ2σt21-φt2φ,and solution (56) represents a temporal nonlocal thermal equation interpreted as an existence of memory effects which correspond to intrinsic dissipation characterized by the exponent of the fractional derivative φ in the system.
For underdamped case, we have (D/τk~x2-1/4τ2)=0, ω0=D/τk~x is the undamped natural frequency expressed in radians per second, and α=1/2τ is the damping factor expressed in meters per second. Next, we will analyze the case when φ takes different values.
Case 1.
When φ=2, from (56) we have(57)Tx,t=T0·expik~xx·E2,1-12τσtt2·E4,1-Dτk~x2-14τ21σt2t4,where E2,1 is given by (10) and E4,1 by (12); in this case solution (57) is(58)Tx,t=T02·expik~xx·cos12σtτt·cos-Dτk~x2-14τ21σt21/4t+cosh-Dτk~x2-14τ21σt21/4t.
Case 2.
When φ=3/2, from (56) we have(59)Tx,t=T0·expik~xx·E3/2,1-12τσt1/2t3/2·E3,1-Dτk~x2-14τ21σtt3,where E3/2,1 is given by (13) and E3,1 by (11); in this case solution (59) is(60)Tx,t=T02·expik~xx·-t3/2/2τσt1/22κ1-r3∑j=02αj1-3/2+r·expαj-t3/22τσt1/22κ·αj3/2+erfcαj1/2-t3/22τσt1/2κ--t3/22τσt1/2-2n·∑k=02n-1-t3/2/2τσt1/2kΓ3k/2+μ·exp-Dτk~x2-14τ22/31σt2/3t+2expD/τk~x2-1/4τ21/σt2/32t·cos-32Dτk~x2-14τ21σt2/3t,where κ=1/3, r=3n+μ, n=0,1,2,3,…, μ=1,2,3,….
Case 3.
When φ=1, from (56) we have(61)Tx,t=T0·expik~xx·exp-t2τ·cosDτk~x2-14τ2t,and (61) represents the classic case and the well-known result; from (61) we see that there is a relation between φ and σt given by(62)φ=Dτk~x2-14τ21/2σt,0<σt≤1D/τk~x2-1/4μ21/2.
Then solution (56) for the underdamped case α<ω0 takes the form(63)Tx,t=T0·expik~xx·Eφ,1-12τD/τk~x2-1/4τ2φ1-φt~φ·E2φ,1-φ21-φt2φ,where t^=(D/τk~x2-1/4τ2)1/2t is a dimensionless parameter.
Due to the condition α<ω0 we can choose an example(64)12τD/τk~x2-1/4τ2=3,0≤12τD/τk~x2-1/4τ2<∞.
So, solution (56) takes its final form:(65)Tx,t=T0·expik~xx·Eφ,1-3φ1-φt^φ·E2φ,1-φ21-φt^2φ.
Case 4.
When φ=1/2, from (56) we have(66)Tx,t=T0·expik~xx·expσt4τ2t·1-erfc-σt1/22τt1/2·exp-Dτk~x2-14τ2σtt,and erfc(α) denotes the error function defined in (8). Equation (66) represents the time evolution of the temperature and the amplitude exhibits an algebraic decay for t→∞. Plots for different values of φ are shown in Figures 2(a) and 2(b).
Simulation of (65) for 0.5<φ≤2. The crossover time occurs at t^=1.0 independent of φ, but the crossover time increases with decreasing φ.
Table 2 shows the different solutions of (65). The order of the fractional differential equation is φ=2, φ=3/2, φ=1, and φ=1/2. The change of the order of the derivative describes the crossover from ballistic transport to the diffusion behavior.
Solutions for the fractional time Cattaneo-Vernotte equation in underdamped case (65) for different values of φ; erfc(α) denotes the error function defined in (8).
In the overdamped case, α>ω0 or η>2k~xϵ/μ, the solution of (56) has the form(67)Tx,t=T0·expik~xx·Eφ,1-σt1-φ2τtφ·Eφ,1-14τ2-Dτk~x21/2σt1-φtφ.
Next, we will analyze the case when φ takes different values.
Case 1.
When φ=2, we have(68)Tx,t=T0·expik~xx·E2,1-σt-12τt2·E2,1-14τ2-Dτk~x21/2σt-1t2,where E2,1 is given by (10); in this case solution (69) is(69)Tx,t=T0·expik~xx·cos12τσtcos14τ2-Dτkx21/21σtt.
Case 2.
When φ=3/2, we have(70)Tx,t=T0·expik~xx·E3/2,1-12τσt-1/2t3/2·E3/2,1-14τ2-Dτk~x21/2σt-1/2t3/2,where E3/2,1 is given by (13); in this case solution (70) is(71)Tx,t=T02·expik~xx·-t3/2/2τσt1/22κ1-r3∑j=02αj1-3/2+rexpαj-t3/22τσt1/22κ·αj3/2+erfcαj1/2-t3/22τσt1/2κ--t3/22τσt1/2-2n·∑k=02n-1-t3/2/2τσt1/2kΓ3k/2+μ·-1/4τ2-D/τk~x21/2σt-1/2t3/23∑j=02αj1-3/2+r·expαj-14τ2-Dτk~x2σt-1/2t3/22k·αj3/2+erfcαj1/2-14τ2-Dτk~x2σt-1/2t3/2k--14τ2-Dτk~x21/2σt-1/2t3/2-2n·∑k=02n-1-1/4τ2-D/τk~x21/2σt-1/2t3/2kΓ3k/2+μ,where κ=1/3, r=3n+μ, n=0,1,2,3,…, μ=1,2,3,….
Case 3.
When φ=1, from (67) we have(72)Tx,t=T0·expik~xx·exp-12τ1+1-4τk~x2Dt,and solution (72) represents the classic case.
Taking into account, the relation between φ and σt is(73)φ=14τ2-Dτk~x21/2σt,0<σt≤11/4τ2-D/τk~x21/2.
Solution (67) takes the form(74)Tx,t=T0·expik~xx·Eφ,1-12τ1/4τ2-D/τk~x2φ1-φtφ·Eφ,1-φ1-φtφ,where t^=(1/4τ2-D/τk~x2)1/2t is a dimensionless parameter.
Due to the condition α>ω0 we can choose an example(75)12τ1/4τ2-D/τk~x2=12,1<12τ1/4τ2-D/τk~x2<∞.Then, solution (67) can be written in its final form:(76)Tx,t=T0·expik~xx·Eφ,1-12φ1-φt^φ·Eφ,1-φ1-φt^φ.
Case 4.
When φ=1/2, from (67) we have(77)Tx,t=T0·expik~xx·expσt4τ2t·1-erfc-σt1/22τt1/2·exp14τ2-Dτk~x2σtt·1-erfc14τ2-Dτk~x21/2σt1/2t1/2,and erfc(α) denotes the error function defined in (8). Equation (77) represents the time evolution of the temperature and the amplitude exhibits an algebraic decay for t→∞. Plots for different values of φ are shown in Figures 3(a) and 3(b).
Simulation of (76) for 0.5<φ≤2. The crossover time occurs at t^=1.0 independent of φ, but the crossover time increases with decreasing φ.
Table 3 shows the different solutions of (76). The order of the fractional differential equation is φ=2, φ=3/2, φ=1, and φ=1/2. The change of the order of the derivative describes the crossover from ballistic transport to the diffusion behavior.
Solutions for the fractional time Cattaneo-Vernotte equation in overdamped case (76) for different values of φ; erfc(α) denotes the error function defined in (8).
Now considering (18) and assuming that the space and time derivative are fractional, the order of the time-space fractional differential equations is φ∈(0,1]; for this example we consider, for t>0, x=0 and x=L with Dirichlet condition T(0,t)=T(L,t)=0 and initial conditions 0<x<L, t=0:T(t,0)=T0>0, and 0<x<L, t=0:∂T/∂t|t=0=0.
Applying the Fourier method of the variable separation, the full solution of (18) is(78)Tx,t=T0πτ·Eφ-12τσt1-φtφ·∑m=1∞12m-1IEiφ2m-1πLσx1-φxφ·-1+2τz2m-1z2m-1·Eφ-z2m-1σt1-φtφ+1+2τz2m-1z2m-1·Eφz2m-1σt1-φtφ,where I indicates the imaginary part and zm=L2-4mπτD/2Lτ. In the case when φ=1, we have the classical solution(79)Tx,t=T0πτ·exp-12τt·∑m=1∞12m-1sin2m-1πLx·-1+2τz2m-1z2m-1·exp-z2m-1t+1+2τz2m-1z2m-1·expz2m-1t.
Now considering (18) with the Riesz space fractional derivative, the order of the space fractional differential equations is φ∈(0,1]; for this example we consider, for 0<t≤T, 0<x<L, and Dirichlet condition T(0,t)=T(L,t)=0 and initial conditions T(x,0)=h(x),(80)∂2Tx,t∂t2+1τσt1-φ∂Tx,t∂t=-Dτk~x2σt21-φ·12cosπφ/2·1Γn-φ∂∂xn∫-∞xfχ,tdχx-χφ+1-n+-1nΓn-φ∂∂xn∫x∞fχ,tdχχ-xφ+1-n,and the solution is given by(81)Tx,t=∑n=1∞AntsinnπxL,which satisfies the boundary condition; substituting this condition into (81) we obtain(82)∑n=1∞d2Andt2+1τσt1-φddtAn+Dτk~x2σt21-φλnφ/2AnsinnπxL=0,and the problem for An becomes(83)d2Andt2+1τσt1-φddtAn+Dτk~x2σt21-φλnφ/2An=0which has the general solution(84)Tx,t=An0·exp-12τ1+1-4τk~x2Dt·expDτk~x2σt21-φλnφ/2t;to obtain An(0), we use the initial condition(85)Tx,t=∑n=1∞Tn0sinnπxL=gx,from which we deduce that (86)Tn0=2L∫0LgχsinnπχLdχ=Bn.
Hence, the solution is given by(87)Tx,t=∑n=1∞BnsinnπxL·exp-12τσt1-φ--14τ2-Dτk~x21/2σt1-φt·exp-Dτk~x2σt21-φλnφ/2t,where λn=n2π2/L2.
4. Conclusions
In this paper we introduced an alternative representation of the fractional Cattaneo-Vernotte equation. In particular, a one-dimensional model was considered in detail. We showed that the fractional Cattaneo-Vernotte equation inherits some crucial characteristics; in particular, the parameters σx and σt need to be introduced in order to characterize the behavior of the physical system, which is located into an intermediate state between a conservative and dissipative system presenting anomalous relaxations. This combination of stored and dissipated energy is conveniently based on the representation of linear thermoviscoelasticity theory. Usually this dissipation is known as internal friction. Some special cases are also discussed. Our results indicate that the fractional order φ has an important influence on the temperature. Considering the Dirichlet conditions, in the range 0<φ≤1 for the spatial case, we observe the non-Markovian Lévy flights and, in the temporal case for the range 0<φ<1, the subdiffusion phenomena. In the spatial case when φ=3/2 and φ=2, the diffusion exhibits an increment in the amplitude and the behavior presents anomalous dispersion (the diffusion increases with increasing order of φ), in the range 1<φ≤2 we observe the Markovian Lévy flights. Finally, in the temporal case, when φ=3/2 and φ<2, the diffusion exhibits an increment in the amplitude and presents the superdiffusion, and the case φ=2 represents the ballistic diffusion. A crossover from a power-law behavior for short times to an exponential decay for long times has been found. When memory effects are incorporated using fractional time derivatives, the crossover dynamics is richer. The alternative model and results in this paper provide a new theoretical perspective of the non-Fourier heat conduction. Furthermore, since the solutions are given in terms of the multivariate Mittag-Leffler functions depending only on a small number of parameters, the universality concept (when the class of behavior does not depend on the details of the physical system) can be considered through this methodology since the analytic solutions presented only need a few parameters to describe their behavior; in all cases the solutions preserve the physical units of the system studied.
Among problems for further research we mention the problem of thermal convection of non-Fourier fluids and the non-Newtonian effects in thermal convection (some situations exist where the non-Fourier and non-Newtonian effects are simultaneously present such as rarefied gases with high Knudsen numbers [52–54]; in this case it is important to describe the interaction between the thermal relaxation and viscous (stress) relaxation); another problem is the two- and three-dimensional fractional wave equations considering fractional variational calculus (see [54] and the references therein) with different initial or/and boundary conditions; of course, it would be interesting to consider the fractional thermal wave equations with fractional derivatives defined in different ways. Furthermore, the methodology proposed in this work can be applied in the critical phenomena theory, self-similarity, scale-invariance, propagation of energy in dissipative systems, theory of viscoelastic fluids and solids, relaxing gas dynamics, irreversible thermodynamics, theory of thermal stresses, thermoelasticity, cosmological models, finance modeling, theory of diffusion in crystalline solids, and the description of anomalous complex processes.
Competing Interests
The authors declare no competing interests.
Acknowledgments
The authors would like to thank Mayra Martínez for the interesting discussions. J. F. Gómez Aguilar acknowledges the support provided by CONACYT: Cátedras CONACYT para Jovenes Investigadores 2014.
JosephD. D.PreziosiL.Heat waves1989611417310.1103/RevModPhys.61.41MR977943ZBL1129.80300LenziE. K.EvangelistaL. R.LenziM. K.RibeiroH. V.de OliveiraE. C.Solutions for a non-Markovian diffusion equation2010374414193419810.1016/j.physleta.2010.08.049MR26848662-s2.0-77956393861Ben-AvrahamD.HavlinS.2000Cambridge, UKCambridge University Press10.1017/cbo9780511605826MR1881970MetzlerR.ChechkinA. V.KlafterJ.Levy statistics and anomalous transport: levy flights and subdiffusion200952185239MachadoJ. T.KiryakovaV.MainardiF.Recent history of fractional calculus20111631140115310.1016/j.cnsns.2010.05.027MR2736622ZBL1221.260022-s2.0-77957362117Tenreiro MachadoJ. A.GalhanoA. M.TrujilloJ. J.Science metrics on fractional calculus development since 1966201316247950010.2478/s13540-013-0030-yMR30337012-s2.0-84879647927Tenreiro MachadoJ. A.GalhanoA. M. S. F.TrujilloJ. J.On development of fractional calculus during the last fifty years201498157758210.1007/s11192-013-1032-62-s2.0-84891741140SandevT.MetzlerR.TomovskiZ.Velocity and displacement correlation functions for fractional generalized Langevin equations201215342645010.2478/s13540-012-0031-2MR29441092-s2.0-84869153006EabC. H.LimS. C.Fractional Langevin equations of distributed order20118331003113610.1103/physreve.83.031136MR27882332-s2.0-79961097638De AnnaP.Le BorgneT.DentzM.TartakovskyA. M.BolsterD.DavyP.Flow intermittency, dispersion, and correlated continuous time random walks in porous media20131101818450210.1103/PhysRevLett.110.1845022-s2.0-84877767153MagdziarzM.MetzlerR.SzczotkaW.ZebrowskiP.Correlated continuous-time random walks in external force fields201285505110310.1103/physreve.85.0511032-s2.0-84861917468ZebA.ZamanG.JungI. H.KhanM.Optimal campaign strategies in fractional-order smoking dynamics2014695-622523110.5560/zna.2014-00202-s2.0-84901507666JesusI. S.Tenreiro MachadoJ. A.Fractional control of heat diffusion systems200854326328210.1007/s11071-007-9322-2MR24429442-s2.0-52549095901ZingalesM.Fractional-order theory of heat transport in rigid bodies201419113938395310.1016/j.cnsns.2014.04.004MR32127892-s2.0-84901653384Gómez-AguilarJ. F.BaleanuD.Fractional transmission line with losses20146910-1153954610.5560/zna.2014-00492-s2.0-84910054646Gómez AguilarJ. F.BaleanuD.Solutions of the telegraph equations using a fractional calculus approach20141512734MR3181953MishraV.VishalK.DasS.OngS. H.On the solution of the nonlinear fractional diffusion-wave equation with absorption: a homotopy approach2014693-413514410.5560/zna.2013-00842-s2.0-84900548404SibatovR. T.UchaikinV. V.Fractional differential approach to dispersive transport in semiconductors200952101019104310.3367/ufne.0179.200910c.10792-s2.0-76649096343YangX.-J.BaleanuD.SrivastavaH. M.Local fractional similarity solution for the diffusion equation defined on Cantor sets201547546010.1016/j.aml.2015.02.024MR33396382-s2.0-84939986878MetzlerR.KlafterJ.The random walk's guide to anomalous diffusion: a fractional dynamics approach2000339117710.1016/s0370-1573(00)00070-3MR1809268WuG.-C.BaleanuD.ZengS.-D.DengZ.-G.Discrete fractional diffusion equation2015801-228128610.1007/s11071-014-1867-2MR33242602-s2.0-84925507033DefterliO.D'EliaM.DuQ.GunzburgerM.LehoucqR.MeerschaertM. M.Fractional diffusion on bounded domains201518234236010.1515/fca-2015-0023MR3323906ZBL06418052SchumerR.MeerschaertM. M.BaeumerB.Fractional advection-dispersion equations for modeling transport at the earth surface20091144F00A0710.1029/2008jf0012462-s2.0-77954316974MeerschaertM. M.BensonD. A.SchefflerH.-P.BaeumerB.Stochastic solution of space-time fractional diffusion equations2002654404110310.1103/physreve.65.041103MR19179832-s2.0-41349099804LiuF.MeerschaertM. M.McGoughR. J.ZhuangP.LiuQ.Numerical methods for solving the multi-term time-fractional wave-diffusion equation201316192510.2478/s13540-013-0002-2MR30166392-s2.0-84871790575KullbergA.del-Castillo-NegreteD.MoralesG. J.MaggsJ. E.Isotropic model of fractional transport in two-dimensional bounded domains201387505211510.1103/physreve.87.0521152-s2.0-84878387975GorenfloR.LuchkoY.StojanoviM.Fundamental solution of a distributed order time-fractional diffusion-wave equation as probability density201316229731610.2478/s13540-013-0019-6MR30336902-s2.0-84879674063MainardiF.LuchkoY.PagniniG.The fundamental solution of the space-time fractional diffusion equation200142153192MR1829592AguilarJ. F. G.HernándezM. M.Space-time fractional diffusion-advection equation with Caputo derivative20142014828301910.1155/2014/283019MR32535752-s2.0-84907402909GorenfloR.MainardiF.MorettiD.ParadisiP.Time fractional diffusion: a discrete random walk approach200229112914310.1023/a:1016547232119MR19264702-s2.0-0036650850CompteA.MetzlerR.The generalized cattaneo equation for the description of anomalous transport processes199730217277728910.1088/0305-4470/30/21/006MR16034382-s2.0-0041107241PostvenkoY.Theories of termal stresses based on space-time-fractional telegraph equations201264103321332810.1016/j.camwa.2012.01.066AtanackovicT.KonjikS.OparnicaL.ZoricaD.The Cattaneo type space-time fractional heat conduction equation2012244–629331110.1007/s00161-011-0199-4MR29928362-s2.0-84868574230AtanackovicT. M.PilipovicS.ZoricaD.A diffusion wave equation with two fractional derivatives of different order200740205319533310.1088/1751-8113/40/20/006MR23201452-s2.0-34247636164LewandowskaK. D.KosztołowiczT.Application of generalized cattaneo equation to model subdiffusion impedance2008395121112202-s2.0-55249115118TarasovV. E.Transport equations from Liouville equations for fractional systems200620334135310.1142/S0217979206033267MR2206188ZBL1130.823352-s2.0-32644440042QiH. T.JiangX. Y.Solutions of the space-time fractional Cattaneo diffusion equation2011390111876188310.1016/j.physa.2011.02.010MR27882872-s2.0-79953328485XuH.-Y.QiH.-T.JiangX.-Y.Fractional Cattaneo heat equation in a semi-infinite medium2013221601440110.1088/1674-1056/22/1/0144012-s2.0-84873384919YuY. J.TianX. G.LuT. J.Fractional order generalized electro- magneto-thermo-elasticity20134218820210.1016/j.euromechsol.2013.05.006MR31074792-s2.0-84880051116VongS.-W.PangH.-K.JinX.-Q.A high-order difference scheme for the generalized Cattaneo equation20122217018410.4208/eajam.110312.240412aMR3119537ZBL1287.650682-s2.0-84894856330ZhaoX.SunZ.-Z.Compact Crank-Nicolson schemes for a class of fractional Cattaneo equation in inhomogeneous medium201562374777110.1007/s10915-014-9874-5MR33043312-s2.0-84922338101Gómez-AguilarJ. F.Miranda-HernándezM.López-LópezM. G.Alvarado-MartínezV. M.BaleanuD.Modeling and simulation of the fractional space-time diffusion equation2016301–311512710.1016/j.cnsns.2015.06.014MR34200002-s2.0-84939486494MillerK. S.RossB.1993New York, NY, USAJohn Wiley & SonsMR1219954PodlubnyI.1999New York, NY, USAAcademic PressMR1658022BaleanuD.DiethelmK.ScalasE.TrujilloJ. J.2012World ScientificComplexity, Nonlinearity and ChaosDiethelmK.FordN. J.FreedA. D.LuchkoY.Algorithms for the fractional calculus: a selection of numerical methods20051946–874377310.1016/j.cma.2004.06.006MR21056482-s2.0-10644238068DiethelmK.2010Berlin, GermanySpringerHauboldH. J.MathaiA. M.SaxenaR. K.Mittag-Leffler functions and their applications201120115129862810.1155/2011/298628MR28005862-s2.0-80052654574MillerK. S.Some simple representations of the generalized Mittag-Leffler functions2001111132410.1080/10652460108819297MR1828253GarrappaR.PopolizioM.Evaluation of generalized Mittag-Leffler functions on the real line201339120522510.1007/s10444-012-9274-zMR30686012-s2.0-84879234882StruchtrupH.TaheriP.Macroscopic transport models for rarefied gas flows: a brief review201176567269710.1093/imamat/hxr004MR28376142-s2.0-80053238550BrennerH.Steady-state heat conduction in a gas undergoing rigid-body rotation. Comparison of Navier-Stokes-Fourier and bivelocity paradigms201370294510.1016/j.ijengsci.2013.03.009MR30795772-s2.0-84878148158StruchtrupH.Resonance in rarefied gases2012244–636137610.1007/s00161-011-0202-0MR29928412-s2.0-84868543634TarasovV. E.Fractional vector calculus and fractional Maxwell's equations2008323112756277810.1016/j.aop.2008.04.005MR2463217ZBL1180.780032-s2.0-53049088436