Accepted Manuscript
Title: Thermodynamic consistency of objective rate equations
Author: A. Morro
PII:
DOI:
Reference:
S0093-6413(16)30348-2
http://dx.doi.org/doi:10.1016/j.mechrescom.2017.06.008
MRC 3176
To appear in:
Received date:
Revised date:
Accepted date:
15-12-2016
6-6-2017
9-6-2017
Please cite this article as: A. Morro, Thermodynamic consistency of objective
rate equations, <![CDATA[Mechanics Research Communications]]> (2017),
http://dx.doi.org/10.1016/j.mechrescom.2017.06.008
This is a PDF file of an unedited manuscript that has been accepted for publication.
As a service to our customers we are providing this early version of the manuscript.
The manuscript will undergo copyediting, typesetting, and review of the resulting proof
before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that
apply to the journal pertain.
Thermodynamic consistency of objective rate equations
A. Morro
DIBRIS, University of Genoa, Via Opera Pia 13, 16145 Genoa, Italy
t
Highlights
Ac
ce
pt
ed
M
an
us
cr
ip
Deformable and heat-conducting solids are modelled by objective rate equations for stress
and heat flux.
Objective time derivatives are those of Jaumann, Green-Naghdi, Oldroyd, and Truesdell.
Thermodynamic consistency is established for the Jaumann, Green-Naghdi, Oldroyd
derivatives in the spatial description.
Thermodynamic consistency of equations involving the Truesdell rate is shown in the
material description.
1
Page 1 of 8
Thermodynamic consistency of objective rate equations
us
cr
ip
t
A. Morro∗
Abstract
M
an
The paper addresses the modelling of solids via rate equations for the heat flux and the stress. As with any
constitutive property, the rate equations are required to be both objective and consistent with the second law
of thermodynamics. Upon a review of a connection between objective time derivatives, attention is restricted
to rate equations where the time derivatives are those named after Jaumann, Green and Naghdi, Oldroyd,
and Truesdell. Equations involving the Truesdell rate are investigated within the material description thanks
to the identity between the time derivative of the material fluxes and the Truesdell rate in the current
configuration. The remaining equations are examined within the spatial description. The occurrence of a
skew tensor in the objective derivative results in a further restriction on the constitutive properties. The
thermodynamic requirements are found to be satisfied and the corresponding free energy is determined by
direct integrations.
Ac
ce
pt
ed
Keywords: Rate equations; Objectivity; Compatibility with thermodynamics; Internal variables
∗ Corresponding author at: DIBRIS, University of Genoa,
Via Opera Pia 13, 16145 Genoa, Italy. Email address: angelo.morro@unige.it
Preprint submitted to Mech. Res. Commun.
June 6, 2017
Page 2 of 8
4
5
6
7
8
9
10
11
12
of the material field (of vectors and tensors) is the
material field of the corresponding Truesdell rate.
The modelling of solid behaviour is often performed by means of internal variables, namely
scalar, vector, or tensor fields governed by rate
equations. Internal variables trace back to Duhem
[1], §I.8, and Bridgman [2]. Thermodynamic analyses within continuum approaches were given by
Eckart [3], and next by Green and Naghdi [4], and
Coleman and Gurtin [5]. It is customary to describe
the constitutive properties related to the internal
variables via rate equations providing the evolution
by means of a function of the state of the material.
There are many examples of models based on
rate equations of internal variables. The so-called
Maxwell-Cattaneo equation relates the heat flux q
to the temperature gradient g, in the form
40
2. Connection between rate equations
q + τ q̇ = −κg,
τ, κ > 0.
(1)
t
3
39
38
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
M
We can then view q as an internal variable, such
that q̇ is a (linear) function of g and q itself. The
Cauchy stress T is often modelled by a rate equation, mainly on the basis of rheological models. Let
E be the infinitesimal strain tensor. The rate equation
1
Ṫ = k Ė − T,
k, τ > 0,
τ
characterizes the Maxwell model. Similar rate
equations describe the standard linear solid, the
Wiechert model, and the Jeffreys model.
Two questions arise in connection with the rate
equations. The time derivative, of vectors and tensors, is required to be objective namely it has to
transform like vectors and tensors under a change
of frames of reference. The well-known derivatives
named after Jaumann, Green and Naghdi, Oldroyd
and Truesdell are objective and moreover are connected to each other in a sense that is shown in
§2.
The second question is related to compatibility of
rate equations, as constitutive equations, with thermodynamics. This is the main topic of the paper.
The subject is of interest because objective time
derivatives involve appropriate spin tensors. The
analysis of Jaumann, Green-Naghdi, and Oldroyd
time derivatives is developed in the spatial description and the occurrence of a spin tensor is shown
to provide an additional restriction relative to the
corresponding equations in non-objective form. Instead, the Truesdell rate is more naturally framed
and investigated within the material description.
This procedure is allowed since the time derivative
We let a body occupy a three-dimensional region.
The points of the body are labelled by their position
vector X in a reference configuration R while x
is the position in the current configuration. We
denote by F the deformation gradient, relative to
R, FiK = ∂XK xi , and by C = FT F the CauchyGreen tensor. Moreover, v is the velocity, L is the
velocity gradient, Lij = ∂xj vi , D is the stretching
tensor and W is the spin tensor so that L = D+W.
∇ is the gradient in the current configuration while
∇X is the gradient relative to the reference position
X. The superscript T means transpose. We let θ
be the absolute temperature and hence g = ∇θ.
Also, sym denotes the symmetric part while Sym is
the set of symmetric tensors. For any vector p, p2
stands for p · p.
Let F , F ∗ be two frames of reference. The vector
positions of a point x and x∗ , relative to F and F ∗ ,
are related by ([6], §17)
us
cr
ip
2
1. Introduction
an
1
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
ed
pt
15
58
59
60
ce
14
57
x∗ = c(t) + Q(t)x,
(2)
where c(t) is an arbitrary vector-valued function,
of time t, while Q(t) is a proper orthogonal tensor
function, det Q = 1. Moreover we denote by Ω =
Q̇QT = −ΩT the spin tensor associated with Q.
Let K be any objective tensor so that the values
in F ∗ and F are related by K∗ = QKQT . The
material time derivative gives
K˙ ∗ = ΩK∗ + K∗ ΩT + QK̇QT .
Ac
13
56
(3)
Let A be a tensor which transforms to A∗ such that
A∗ = QAQT + Ω.
(4)
Substitution of Ω from (4) into (3) results in
K˙ ∗ − A∗ K∗ − K∗ A∗ T = Q(K̇ − AK − KAT )QT .
This means that, for any tensor A satisfying (4)
and any objective tensor K, the derivative
◦
K= K̇ − AK − KAT
61
(5)
is objective.
If p is an objective vector, p∗ = Qp, then
p˙∗ = ΩQp + Qṗ.
2
Page 3 of 8
82
3. Entropy inequalities
75
p˙∗ − A∗ p∗ = Q(ṗ − Ap).
76
77
This means that, subject to (4),
78
◦
(6)
p= ṗ − Ap
79
80
is an objective derivative.
The velocity gradient L and the spin W obey the
transformation law (4) (see [7], §20.3)
L∗ = QLQT + Ω,
63
W∗ = QWQT + Ω
Let ε be the energy density (per unit mass) and ρ
the mass density. The balance of energy is written
in the form
whereas D is objective and ∇ · v is invariant.
Let R be the rotation provided by the polar decomposition theorem, so that F = RU, U2 = FT F.
Since F∗ = QF and R∗ = QR then
ρε̇ = T · D − ∇ · q + ρr,
an
62
67
the objective time derivatives (5) and (6) become
the Jaumann, Green-Naghdi, and Oldroyd time
derivatives of tensors and vectors [7, 8].
Let A△ and A▽ be two tensors which transform
to A∗△ and A∗▽ according to (4). Hence the two
derivatives
ed
66
Hence ṘRT too satisfies (4).
Letting
A = W, ṘRT , L,
△
▽
K= K̇ − A▽ K − KAT▽
ce
K= K̇ − A△ K − KAT△ ,
pt
65
83
84
are objective and the two rate equations
△
K △ = K ▽ − (A△ − A▽ )K − K(AT△ − AT▽ )
69
70
71
72
73
(11)
where ψ = ε − θη is the Helmholtz free energy density.
Let J = det F and ρR = J ρ. Moreover let
TRR = JF−1 TF−T .
(12)
Hence ρR and qR are the mass density and the heat
flux in the reference configuration. Also, TRR is the
second Piola-Kirchhoff stress tensor. In the material description the Clausius-Duhem inequality can
be written in the form [12, 13]
(7)
holds. The analogous statement is true for vectors.
It then follows that, subject to the connection
(7), the Jaumann, Green-Naghdi, and Oldroyd
derivatives describe the same material response. Of
course, the two material responses are different if
K △ and K ▽ do not satisfy (7).
Look at the Truesdell rate given by [9]
1
−ρR (ψ̇ + η θ̇) + 21 TRR · Ċ − qR · ∇X θ ≥ 0. (13)
θ
85
86
87
88
p := ṗ − Lp + (∇ · v)p,
(8)
K := K̇ − LK − KLT + (∇ · v)K.
(9)
1
−ρ(ψ̇ + η θ̇) + T · D − q · g ≥ 0,
θ
qR = JqF−T ,
▽
are equivalent if only
68
Substitution of ρr −∇·q from the balance of energy
allows inequality (10) to be written in the form
K= K ▽ ,
Ac
K= K △ ,
r being the heat supply. Let η be the entropy density. For the present purposes there is no loss in
generality by letting the entropy flux be equal to
q/θ so that the entropy inequality is assumed in
the Clausius-Duhem form
q ρr
ρη̇ + ∇ · −
≥ 0.
(10)
θ
θ
M
R˙ ∗ R∗ T = Ω + QṘRT QT .
64
t
81
Equations (8)-(9) take the forms (6)-(5) by letting
A = L − (∇ · v)1 and A = L − 21 (∇ · v)1.
In [10] an objective re-formulation of the
Maxwell-Cattaneo equation is derived in a form
that, in the present notation, reads τ q +q = −kg.
The Truesdell rate p is referred to as Oldroyd’s upper convected derivative in [10] and a particular
Lie-Oldroyd derivative in [11].
74
us
cr
ip
Substitution of Ω from (4) provides
Hence we can state the second law of thermodynamics by saying that inequality (11), or (13), holds
for any thermodynamic process compatible with the
balance equations.
Sometimes the second law of thermodynamics is
assumed to be expressed by, or to comprise,
q · g ≤ 0,
(14)
3
Page 4 of 8
which is true for Fourier’s law and ensures that heat
flows from hot to cold regions at any time t (see
[14, 15, 16]). Now, by (10) we have
The linearity and the arbitrariness of Ḋ, ġ, θ̇ imply
1
1
1
−η̇ + (ε̇ − T · D) + 2 q · g ≤ 0.
θ
ρ
ρθ
whence ψ = ψ(C, θ, q, T ). The remaining inequality suggests that we make the identification
T = T − 2ρ F ∂C ψ FT .
105
93
94
The linearity and the arbitrariness of g, D, and A
require that
an
92
1
ρβ∂q ψ − q = 0,
θ
which reduces to (14) if q · g has a constant sign.
That the approximation ψ = ψ(θ) provides (14) is
consistent with the derivation of (14) in [16] where ρ
is constant and ε̇ = θη̇ is regarded to hold (locally).
4. Thermodynamic consistency of spatial descriptions
ed
Denote by T a dissipative stress to be identified
by the thermodynamic analysis. To fix ideas we let
A = −AT and consider the rate equations
T + T A = −γT
T + µD,
T˙ − AT
97
98
99
100
101
102
103
104
T − ρµ∂T ψ = 0,
(19)
T − T A) + ρ∂q ψ · Aq = 0,
ρ∂T ψ · (AT
(20)
ργ∂T ψ · T + ρν∂qψ · q ≥ 0.
(21)
Substitution for ∂T ψ and ∂q ψ from (19) into (21)
gives
γ
ν 2
T ·T +
q ≥ 0.
µ
βθ
Since γ, ν > 0 it follows that
µ > 0,
β > 0.
(22)
Moreover, in view of (19), eq. (20) becomes
in the spatial description, where ν, β, γ, µ are allowed to depend on temperature. The left-hand
sides of (15) and (16) become the Jaumann or the
Green-Naghdi time derivative according as A = W
or A = ṘRT . Equations (15) and (16) reduce to
the Maxwell-Cattaneo equation and the Maxwell
model when A = 0 and D is replaced by Ė. We
assume ν > ν0 > 0 and γ > γ0 > 0 so that q and
T are bounded in time.
Let Γ = (C, θ, g, D, q, T ) be the set of independent variables so that ψ and η are functions of Γ
T are given by (15) and (16). Moreover
while q̇ and Ṫ
let ψ(Γ) be differentiable. Since
1
1
T − T A) + q · Aq = 0.
T · (AT
µ
β
ce
96
(16)
(23)
The skew symmetry of A makes each term vanish,
T = 0,
T · AT
Ac
95
(15)
pt
q̇ − Aq = −νq − βg,
(18)
Hence we have
M
91
(17)
T − T A − γT
T ) + (T
T − ρµ∂T ψ) · D
−ρ∂T ψ · (AT
1
(ρβ∂q ψ − q) · g − ρ∂q ψ · (Aq − νq) ≥ 0.
θ
If no work is done, D = 0, and ψ = ψ(θ), so that
ε̇ = θη̇, then the inequality simplifies to
Z t2
1
q · g dt ≤ 0,
2
ρθ
t1
90
η = −∂θ ψ.
us
cr
ip
If the system undergoes a cyclic process as t ∈
[t1 , t2 ], and hence η(t2 ) = η(t1 ), then
Z t2
1
1
1
(ε̇ − T · D) + 2 q · g dt ≤ 0.
θ
ρ
ρθ
t1
89
∂g ψ = 0,
t
∂D ψ = 0,
T · T A = 0,
q · Aq = 0.
Finally, by (19) two direct integrations give
ψ = Ψ(C, θ) +
1
1
q2 +
T ·T .
2ρβθ
2ρµ
∂C ψ · Ċ = 2(F ∂C ψ FT ) · D
Equations (17)-(21) are then necessary and sufficient for the thermodynamic consistency of (15)
and (16), subject to the identification of T . Now,
then, upon evaluation of ψ̇ and substitution in (11),
T = 2ρ F ∂C Ψ FT + T ,
−ρ(∂θ ψ + η)θ̇ + (T − 2ρF ∂C ψ FT ) · D
1
− q · g − ρ∂g ψ · ġ − ρ∂D ψ · Ḋ
θ
−ρ∂q ψ · (Aq − νq − βg)
T − T A − γT
T + µD) ≥ 0.
−ρ∂T ψ · (AT
106
107
108
109
110
where T is given by (16), may be viewed as the
generalized stress of the standard linear solid, with
Ė replaced by D. Accordingly, while the Maxwell
model is said to describe viscoelastic fluids, the
whole stress T describes a viscoelastic solid.
4
Page 5 of 8
T = 0, A = 0)
In stationary conditions (q̇ = 0, Ṫ
116
Since ν, γ > 0, the thermodynamic requirements
(22) result in the positive value of the heat conductivity β/ν and of the shear viscosity µ/γ.
The thermodynamic requirement (23) holds for
every skew tensor A. Hence the conclusions hold
for the Green-Naghdi and Jaumann derivatives.
117
4.1. A model with the Oldroyd derivative
111
112
113
114
115
t
T = (µ/γ)D.
τ (θ)q̇ = −κ(θ)g − q
thus showing a relaxation time τ (θ). This may indicate that we look for models where the appropriate
internal variable, p say, is just τ (θ)q [18]. To account for deformation and hence for objectivity of
the time derivative we may replace (15) by
The occurrence of D in A = L leads to different
thermodynamic results. For definiteness we investigate the rate equation
ṗ − Ap = ν̂p − β̂g,
ν̂ and β̂ being functions of θ. We then let Γ =
(C, θ, g, D, p, T ), with T subject to (16), and repeat the previous procedure to obtain again (17)
and
1
ρβ̂∂p ψ −
p = 0, T − ρµ∂T ψ = 0,
(27)
θτ
T − T A) + ρ∂p ψ · Ap = 0,
ρ∂T ψ · (AT
(28)
(24)
⋄
119
120
121
ργ∂T ψ · T + ρν̂∂p ψ · p ≥ 0.
ed
122
for q. So we replace q̇−Aq in (15) with q := q̇−Lq
whereas no rate equation for the stress is considered. Accordingly we let Γ = (C, θ, g, D, q) be the
set of independent variables. Upon substitution of
ψ̇, the entropy inequality (11) becomes
M
118
1
−ρ(∂θ ψ + η)θ̇ − ρ∂g ψ · ġ − ρ∂D ψ · Ḋ − q · g
θ
T · D − ρ∂q ψ · (Wq + Dq − νq − βg) ≥ 0,
+T
ce
pt
where T is given by (18). The linearity and the
arbitrariness of ġ, Ḋ, and θ̇ imply that ∂g ψ = 0,
∂D ψ = 0, and η = −∂θ ψ. The remaining inequality
holds if and only if
T̂ · D ≥ 0, T̂ := T − ρ∂q ψ ⊗ q,
Ac
1
βρ∂q ψ − q = 0,
θ
128
∂q ψ ⊗ q ∈ Sym
∂q ψ · q ≥ 0.
(25)
1
ψ = Ψ(C, θ) +
q2 ,
2βρθ
129
β > 0.
130
131
132
Hence ∂q ψ ⊗ q ∈ Sym holds identically. As a consequence
T = 2ρ F ∂C ψ FT +
124
125
126
(29)
Hence we find that
ψ = Ψ(C, θ) +
1
2ρβ̂θτ
p2 +
1
T ·T ,
2ρµ
where µ > 0, β̂ > 0.
Remark.
It is sometimes claimed that the
Maxwell-Cattaneo equation is not compatible with
the second law of thermodynamics. This objection
stems from the heat conduction inequality (14) as
the content of the second law for undeformable bodies [14, 15, 16]. It is true in fact that, by (1),
1
(q · q + τ q · q̇)
κ
and the right-hand side is not positive definite because of the time derivative q̇. Even worse would be
the case with objective rate equations. No inconsistency though holds because, for time-dependent
fields, the second law is expressed by (11) or (13).
−q · g =
By (25) it follows that
123
(26)
an
q̇ − Lq = −νq − βg
4.2. A model for second sound in solids
The appropriate internal variables need not be
the physical variables q and T . As an example, in
[17] the heat flux equation is found to fit data on
wave speed at cryogenic temperatures by taking the
rate equation in the form
us
cr
ip
q = −(β/ν)g,
127
1
q ⊗ q + T̂,
βθ
133
134
135
5. Thermodynamic consistency of material
descriptions
Rate equations involving the Truesdell derivative
are naturally framed within a material description.
This is so since the definition (12) and use of
where T̂ is the dissipative part of T, possibly T̂ =
µD, µ > 0. Hence the rate equation for q via the
Oldroyd derivative implies that the stress embodies
the q-dependent term q ⊗ q/βθ.
J˙ = J∇ · v,
˙
F−1 = −F−1 Ḟ F−1
5
Page 6 of 8
make the material heat flux qR and the second
Piola-Kirchhoff stress TRR satisfy
−1
ṪRR = JF
144
145
146
147
148
149
150
151
152
153
154
155
156
157
ẇ + λw = −α∇X θ,
(32)
Ẏ + ΛY = ΦĊ,
(33)
for the invariant vector w and the invariant tensor
Y within the material description. We assume that
λ, Λ, α, and Φ are invariant scalars so that the rate
equations (32) and (33) are invariant. We assume
that λ ≥ λ0 > 0, Λ ≥ Λ0 > 0. The field w may be
taken as a scalar, ξ say, times qR . The tensor Y
will be identified in a moment by thermodynamic
arguments.
Let
ψ, η, qR , TRR
be
continuous
functions of the set of independent variables
Γ = (C, θ, w, Y, ∇Xθ, Ċ). Moreover we let ψ
be continuously differentiable while ξ, ..., Φ are
allowed to depend on θ. Evaluation of ψ̇ and
substitution in (13) provides
+( 12 TRR − ρR ∂C ψ − ρR Φ∂Y ψ) · Ċ
1
−ρR (∂θ ψ + η)θ̇ + (ρR α ∂w ψ − qR ) · ∇X θ
θ
+ρR λ∂w ψ · w + ρR Λ∂Y ψ · Y ≥ 0.
1
Y − ρR Φ∂Y ψ = 0,
2Ξ
158
∂Ċ ψ = 0,
η = −∂θ ψ.
ρR α∂w ψ −
(36)
1
w = 0,
ξθ
ρR λ∂w ψ · w + ρR Λ∂Y ψ · Y ≥ 0.
(37)
The direct integrations with respect to w and Y
provide
ψ = Ψ(θ, C) +
1
1
Y·Y+
w2 . (38)
4ρR ΦΞ
2ρR αξθ
Substitution in (37) shows that the inequality holds
if and only if
159
160
161
162
α > 0,
164
165
166
167
(34)
The remaining inequality,
Φ > 0.
(39)
The requirements (34), (38), and (39) are necessary
and sufficient for the thermodynamic consistency of
the rate equations (32) and (33) and the assumptions (35), (36).
Equations (32) and (33) are now integrated. Subject to λ ≥ λ0 > 0 and Λ ≥ Λ0 > 0 we can integrate
on (−∞, t] to obtain
Z t
t
Y(t) =
Φ(τ ) exp(− ∫ Λ(s)ds)Ċ(τ )dτ, (40)
−∞
The linearity and the arbitrariness of the values
∇X θ̇, C̈, θ̇ imply that
∂∇X θ ψ = 0,
ξ = ξ(θ) > 0.
The linearity of Ċ, ∇Xθ and the arbitrariness of
Ċ, ∇X θ, w, Y require that
163
−ρR ∂∇X θ ψ · ∇X θ̇ − ρR ∂Ċ ψ · C̈
(35)
us
cr
ip
142
143
1
Y.
Ξ
an
141
M
140
1
w,
ξ
qR =
ed
139
Moreover we let
Hence the effect of the Truesdell derivative in the
spatial description can be determined by considering the time derivative in the material description.
The material fields qR and TRR are invariant
under a change of frame. As a consequence, the
time derivatives are also invariant under a change
of frame. Indeed, if λ, ξ, and Λ, Ξ are invariant then
λqR , (ξqR )˙, ΛTRR , and (ΞTRR )˙ are invariant too.
The temperature gradient ∇X θ, in the reference
configuration, and the Cauchy-Green tensor C, and
hence Ċ, are invariant. Borrowing again from the
Maxwell-Cattaneo equation for the heat flux and
the equation for a solid modelled by the Maxwell
element we consider the rate equations
pt
138
TRR = 2ρR ∂C ψ +
. (31)
ce
137
(Ṫ−LT−TL +(∇ · v)T)F
Ac
136
(30)
−T
T
t
q̇R = J(q̇ − Lq + (∇ · v)q)F−T ,
indicates that we identify Y with a positive scalar
function 2Ξ(θ) times 21 TRR − ρR ∂C ψ so that
τ
and the like for w. Hence by (35) the second PiolaKirchhoff stress TRR is the sum of an elastic part,
2ρR ∂C Ψ, and a memory functional, Y/Ξ, of the
history of Ċ. By (36) the reference heat flux qR , is
a memory functional, w/ξ, of the history of ∇Xθ.
We now look for the rate equations, in the spatial
description, associated with (32) and (33). First
we observe that, by viewing ψ as a function of F
through C, the chain rule gives ∂F ψ = 2 F ∂C ψ. As
a consequence ρ ∂F ψ FT = 2 ρ F ∂C ψ FT and hence
2 ρR ∂C ψ = ρ J F−1 ∂F ψ = J F−1 (ρ ∂F ψ FT )F−T .
( 21 TRR − ρR ∂C ψ − ρR Φ∂Y ψ) · Ċ
1
+(ρR α∂w ψ − qR ) · ∇Xθ
θ
+ρR λ∂w ψ · w + ρR Λ∂Y ψ · Y ≥ 0,
This in turn implies that
TdRR := Y/Ξ = J F−1T F−T ,
T := T−ρ ∂F ψ FT ,
(41)
6
Page 7 of 8
where T is the (Cauchy) stress T, deprived of the
elastic part ρ ∂F ψ FT , in the spatial description.
As with (31), since TdRR = J F−1T F−T we find
that
is thermodynamically consistent if and only if so is
◦
p = fO (p, g) + Dp.
d
Let w = ξqR and Y = ΞTRR . From (32) and
(33) we have
(ξqR )˙+λξqR = −α∇X θ,
182
References
178
179
180
(ΞTdRR )˙+ΛΞTdRR = ΦĊ.
Hence by (30) and (42) we obtain
α
˙
q̇ − Lq + ∇ · v q + (λ + ξ/ξ)q
= − F ∇Xθ,
Jξ
183
184
185
Φ
T −T
T LT +∇·v T +(Λ+ Ξ̇/Ξ)T
T =
T˙ −LT
F Ċ FT .
JΞ
To determine T we go back to (41) and find that
186
187
188
189
190
170
171
1
(t)F(t)Y(t)F−T (t)
JΞ
191
192
193
where Y(t) is given by (40). The analogue holds
for q with obvious changes.
194
195
196
ed
197
172
198
6. Conclusions
199
pt
The rate equations involving objective time
derivatives (for a vector p) are relations of the form
ṗ − Ap = fA(p, g),
A = −AT ,
(43)
ce
for the Jaumann and Green-Naghdi time derivatives, and
p = fT (p, g),
(44)
Ac
for the Truesdell derivative. The thermodynamic
consistency is shown in the spatial description for
(43), with fA (q, g) = −νq − βg, and in the material description for (44) with fT (p, g) = −λp −
(α/Jξ)Bg. The crucial role of the Truesdell rate is
related to the identity
˙
JF−1 KF−T = JF−1 K F−T ,
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
173
174
175
176
[1] C. Truesdell, Rational Thermodynamics, Springer, New
York (1984), ch. 5.
[2] P.W. Bridgman, The Nature of Thermodynamics, Harvard University Press, Cambridge, MA, 1961.
[3] C. Eckart, The thermodynamics of irreversible processes. IV. The theory of elasticity and anelasticity,
Phys. Rev. 73 (1948) 373-382.
[4] A.E. Green, P.M. Naghdi, A general theory of an
elastic-plastic continuum. Arch. Rational Mech. Anal.
1 (1965) 251-281.
[5] B.D. Coleman, M.E. Gurtin, Thermodynamics with internal state variables, J. Chem. Phys. 47 (1967) 597613.
[6] C.A. Truesdell, W. Noll, The non-linear field theories of
mechanics. In Encyclopedia of Physics, ed. S. Flügge,
vol. III/3, Springer, Berlin, 1965.
[7] M.E. Gurtin, E. Fried, L. Anand, The Mechanics and
Thermodynamics of Continua, Cambridge University
Press, 2011.
[8] J.G. Oldroyd, On the formulation of rheological equations of state, Proc. Roy. Soc. A 200 (1950) 523-541.
[9] C.A. Truesdell, The simplest rate theory of pure elasticity, Comm. Pure Appl. Math. 8 (1955) 123-132.
[10] C.I. Christov, On frame indifferent formulation of the
Maxwell-Cattaneo model of finite-speed heat conduction, Mech. Res. Comm. 36 (2009) 481-486.
[11] I.C. Christov, P.M. Jordan, On an instability exhibited
by the ballistic-diffusive heat conduction model of Xu
and Hu, Proc. R. Soc. A 470 (2014) issue 2161.
[12] A. Morro, Evolution equations and thermodynamic restrictions for dissipative solids, Math. Comp. Modelling
52 (2010) 1869-1876.
[13] A. Morro, Evolution equations for non-simple viscoelastic solids, J. Elasticity 105 (2011) 93-105.
[14] F. Franchi, B. Straughan, Continuous dependence on
the relaxation time and modelling, and unbounded
growth, in theories of heat conduction with finite propagation speeds, J. Math. Anal. Appl. 185 (1994) 726-746.
[15] M.B. Rubin, Hyperbolic heat conduction and the second law, Int. J. Engng Sci. 30 (1992) 1665-1676.
[16] A. Barletta, E. Zanchini, Hyperbolic heat conduction
and local equilibrium: a second law analysis, Int. J.
Heat Mass Transfer 40 (1997) 1007-1016.
[17] D.D. Coleman, D.C. Newman, Implications of a nonlinearity in the theory of second sound in solids, Phys.
Rev. B37 (1988) 1492-1498.
[18] A. Morro, T. Ruggeri, Second sound and internal energy
in solids, Int. J. Non-Linear Mech. 22 (1987) 27-36.
M
T (t) =
t
181
Yet the occurrence of Dp makes the analysis of
⋄
the thermodynamic consistency of p = fO (p, g)
quite different from, or inequivalent to, that for
◦
p = fO (p, g) + Dp. For definiteness this is shown
via the example in §4.
177
T −T
T LT )F−T . (42)
(TdRR )˙ = JF−1 (T˙ +∇·v T −LT
us
cr
ip
169
an
168
for any tensor K and the analogue for vectors. The
thermodynamic restrictions so established are all
satisfied and the corresponding free energy is determined.
⋄
Since the Oldroyd derivative p is related to the
◦
⋄
◦
Jaumann derivative p = ṗ − Wp by p = p −Dp
then the rate equation
222
223
224
225
226
227
228
229
230
⋄
p = fO (p, g)
7
Page 8 of 8