We consider a body that at some instant occupies the region
\({\mathcal {B}}\) of the Euclidean three-dimensional space and is bounded by the piecewise smooth surface
\(\partial {\mathcal {B}}\). The motion of the body is referred to the reference configuration
\({\mathcal {B}}_0\) and a fixed system of rectangular Cartesian axes
\(Ox_i\) (
\(i = 1, 2, 3\)). We shall employ the usual summation and differentiation conventions: Latin subscripts are understood to range over the integers (
\(1, 2, 3\)), summation over repeated subscripts is implied and subscripts preceded by a comma denote partial differentiation with respect to the corresponding Cartesian coordinate. In what follows we use a superposed dot to denote partial differentiation with respect to time
\(t\). We consider the linear theory of thermopiezoelectricity of non simple materials established in [
1] adopting the entropy production inequality proposed by Green and Laws [
3] (in particular, Sect.
2), in local form
$$\rho \phi \dot{\eta } \ge - q_{{i,i}} + \rho h + \frac{1}{\phi }q_{i} \phi _{{,i}} ,$$
where the function
\(\phi\) is a suitable defined strictly positive thermal function. The fundamental system of field equations consists of the equations of motion
$$\begin{aligned} \tau _{ji,j}-\mu _{kji,kj}+\rho f_{i}=\rho \ddot{u}_{i}, \end{aligned}$$
(1)
the equation for the quasi-static electric field
$$\begin{aligned} \sigma _{i,i}-Q_{ji,ji}=g, \end{aligned}$$
(2)
the energy equation in the linear approximation
$$\begin{aligned} \rho T_0{\dot{\eta }}=-q_{i,i}+\rho h, \end{aligned}$$
(3)
the constitutive equations
$$\begin{aligned} \tau _{ij}&= a_{ijkl}^{(11)}e_{kl} + a_{ijklh}^{(12)}\kappa _{klh} + a_{ijk}^{(13)}E_k \\& \quad + a_{ijkl}^{(17)}V_{kl} + a_{ij}^{(14)} ( \theta + \beta {\dot{\theta }} ), \\ \mu _{ijk}&= a_{lhijk}^{(12)}e_{lh} + a_{ijklhm}^{(22)}\kappa _{lhm} + a_{ijkl}^{(23)}E_l\\& \quad + a_{ijklh}^{(27)}V_{lh} + a_{ijk}^{(24)} ( \theta + \beta {\dot{\theta }} ), \\ \quad -\sigma _i&= a_{jki}^{(13)}e_{jk} + a_{jkli}^{(23)}\kappa _{jkl} + a_{ij}^{(33)}E_j\\&\quad + a_{ijk}^{(37)}V_{jk} + a_i^{(34)} ( \theta + \beta {\dot{\theta }} ), \\ -Q_{ij}&= a_{klij}^{(17)}e_{kl} + a_{klmij}^{(27)}\kappa _{klm} + a_{kij}^{(37)}E_k\\& \quad + a_{ijkl}^{(77)}V_{kl} + a_{ij}^{(47)} ( \theta + \beta {\dot{\theta }}), \\ \quad -\rho \eta&= a_{ij}^{(14)}e_{ij} + a_{ijk}^{(24)}\kappa _{ijk} + a_i^{(34)}E_i\\& \quad + a_{ij}^{(47)}V_{ij} + c ( \theta + \beta {\dot{\theta }} ) +\frac{1}{\beta }\!\left( \gamma {\dot{\theta }} + a_i^{(56)}g_i \right) , \\ \quad -\frac{q_i}{T_0}&= \frac{1}{\beta } \left( a_i^{(56)}{\dot{\theta }} + a_{ij}^{(66)}g_j \right) , \end{aligned}$$
(4)
where
\(c=a^{(44)}\), and the geometric equations
$$\begin{aligned} \begin{alignedat}{2}&e_{ij}=\frac{1}{2}(u_{i,j}+u_{j,i}),&\kappa _{ijk}=u_{k,ij},\\&E_i=-\varphi _{,i}, \qquad V_{ij}=-\varphi _{,ij},&g_i = \theta _{,i}, \end{alignedat} \end{aligned}$$
(5)
where
\(\tau _{ji}=t_{ji}+\mu _{kji,k}\) and
\(\sigma _{i}=D_{i}+Q_{ji,j}\).
We denote with
\(t_{ij}\) the stress tensor,
\(\mu _{ijk}\) the hyperstress tensor,
\(\rho\) the reference mass density,
\(f_i\) the external body force per unit mass,
\(u_i\) the displacement vector field,
\(D_i\) the displacement electric vector field,
\(q_i\) the heat flux vector,
\(Q_{ij}\) the electric quadrupole,
\(g\) the density of free charges,
\(\varphi\) the electric potential,
\(\eta\) the entropy per unit mass,
\(h\) the heat supply per unit mass and unit time and
\(\theta\) the difference of the absolute temperature
\(T\) from the absolute temperature in the reference configuration
\(T_{0}>0\) (i.e.
\(\theta =T-T_{0}\)). Following [
29,
30], for the traction vector
\(t_i\), the hypertraction vector
\(\mu _{ji}\), the heat flux
\(q\) and the the generalized surface charge density
\(Q_{i}\) we have
$$\begin{aligned} \begin{alignedat}{2}&t_i = t_{ji} n_j, \quad&\mu _{ji}&= \mu _{kji}n_k,\\&q = q_i n_i, \quad&{\dot{Q}}_i&= {\dot{Q}}_{ji}n_j, \end{alignedat} \end{aligned}$$
where
\(n_i\) is the outward unit normal vector to the boundary surface
\(\partial {\mathcal {B}}\). The constitutive coefficients satisfy the following symmetry relations:
$$\begin{aligned} \begin{alignedat}{2}&a_{ijkl}^{(11)} = a_{jikl}^{(11)} = a_{klij}^{(11)},\quad{} & {} a_{ijk}^{(13)} = a_{jik}^{(13)}, \\&a_{ijklh}^{(12)} = a_{jiklh}^{(12)} = a_{ijlkh}^{(12)}, \quad{} & {} a_{ij}^{(14)} = a_{ji}^{(14)},\\&a_{ijklhm}^{(22)} = a_{jiklhm}^{(22)} = a_{lhmijk}^{(22)}, \quad{} & {} a_{ij}^{(15)} = a_{ji}^{(15)},\\&a_{ijkl}^{(17)} = a_{jikl}^{(17)} = a_{klij}^{(17)}, \quad{} & {} a_{ijkl}^{(23)} = a_{jikl}^{(23)},\\&a_{ijkhl}^{(27)} = a_{jikhl}^{(27)} = a_{ijklh}^{(27)},\quad{} & {} a_{ijk}^{(24)} = a_{jik}^{(24)},\\&a_{ijkl}^{(77)} = a_{jikl}^{(77)} = a_{klij}^{(77)}, \quad{} & {} a_{ijk}^{(25)} = a_{jik}^{(25)},\\&a_{ij}^{(57)} = a_{ji}^{(57)},{} & {} a_{ijkl}^{(26)} = a_{jikl}^{(26)},\\&a_{ij}^{(66)} = a_{ji}^{(66)}, \quad{} & {} a_{ij}^{(33)} = a_{ji}^{(33)},\\&a_{ijk}^{(67)} = a_{ikj}^{(67)},\quad{} & {} a_{ijk}^{(37)} = a_{ikj}^{(37)},\\&\quad{} & {} a_{ij}^{(47)} = a_{ji}^{(47)}. \end{alignedat} \end{aligned}$$
(6)
For sake of simplicity we set
\(k=\gamma /\beta\),
\(k_{i}=\frac{1}{\beta }a_{i}^{(56)}\) and
\(k_{ij}=\frac{1}{\beta }a_{ij}^{(66)}\). From the entropy inequality the following restriction has been obtained
$$\begin{aligned} k{\dot{\theta }}^2+2k_i{\dot{\theta }}g_{i}+k_{ij}g_ig_j\ge 0,\qquad \forall \,{\dot{\theta }},g_i. \end{aligned}$$
(7)
Then it is necessary that
\(k\) be non negative and the thermal conductivity tensor
\(k_{ij}\) be positive semi-definite.
Following [
6,
7], we consider
\(P_i\),
\(R_i\),
\(\Lambda\) and
\(H\) defined in such a way that the total rate of work of the surface forces over the smooth surface
\(\partial {\mathcal {P}}\) can be expressed in the form
$$\begin{aligned} \begin{aligned} \int _{\partial {\mathcal {P}}}&\left( t_{ki}{\dot{u}}_i + \mu _{kji}{\dot{u}}_{i,j} - {\dot{D}}_k\varphi - {\dot{Q}}_{ki}\varphi _{,i} \right) n_k~da\\&=\int _{\partial {\mathcal {P}}} \left( P_i{\dot{u}}_i + R_i{\mathcal {D}}{\dot{u}}_i - {\dot{\Lambda }}\varphi - {\dot{H}}{\mathcal {D}}\varphi \right) da \end{aligned} \end{aligned}$$
where
\({\mathcal {P}}\subseteq {\mathcal {B}}\) is a part of the body and
\(\partial {\mathcal {P}}\) its boundary. We have
$$\begin{aligned}&P_i = ( \tau _{ji} - \mu _{kji,k} ) n_j - {\mathcal {D}}_j ( \mu _{kji} n_k ) + ( {\mathcal {D}}_l n_l ) \mu _{kji} n_k n_j, \\&\Lambda = ( \sigma _j - Q_{kj,k} ) n_j - {\mathcal {D}}_j ( Q_{kj} n_k ) + ( {\mathcal {D}}_l n_l ) Q_{kj} n_k n_j, \end{aligned}$$
and
$$\begin{aligned} R_i = \mu _{kji} n_k n_j, \qquad H = Q_{kj} n_k n_j, \end{aligned}$$
where
\({\mathcal {D}} \equiv n_i \partial / \partial x_i\) is the normal derivative and
$$\begin{aligned} {\mathcal {D}}_i \equiv ( \delta _{ij} - n_i n_j ) \frac{\partial }{\partial x_j} \end{aligned}$$
is the surface gradient. Now, we denote with
$$\begin{aligned} {\mathcal {U}} = ( u_i, \varphi , \theta ) \end{aligned}$$
the solutions of the mixed initial-boundary value problem
\(\Pi\) defined by Equations (
1)–(
5) with the restrictions (
7), where the coefficients satisfy the symmetries (
6), the following initial conditions
$$\begin{array}{*{20}l} {u_{i} (0) = u_{i}^{0} ,} \hfill & {\dot{u}_{i} (0) = v_{i}^{0} ,} \hfill \\ {\theta (0) = \theta ^{0} ,} \hfill & {\dot{\theta }(0) = \vartheta ^{0} ,} \hfill \\ \end{array}$$
(8)
in
\(\bar{{\mathcal {B}}}\) and the following boundary conditions
$$\begin{array}{*{20}l} {u_{i} = \hat{u}_{i} \;{\text{on}}\;S_{1} \times I,} \hfill & {P_{i} = \hat{P}_{i} \;{\text{on}}\;\Sigma _{1} \times I,} \hfill \\ {{\mathcal{D}}u_{i} = \hat{d}_{i} \;{\text{on}}\;S_{2} \times I,} \hfill & {R_{i} = \hat{R}_{i} \;{\text{on}}\;\Sigma _{2} \times I,} \hfill \\ {\varphi = \hat{\varphi }\;{\text{on}}\;S_{3} \times I,} \hfill & {\Lambda = \hat{\Lambda }\;{\text{on}}\;\Sigma _{3} \times I,} \hfill \\ {{\mathcal{D}}\varphi = \hat{\xi }\;{\text{on}}\;S_{4} \times I,} \hfill & {H = \hat{H}\;{\text{on}}\;\Sigma _{4} \times I,} \hfill \\ {\theta = \hat{\theta }\;{\text{on}}\;S_{5} \times I,} \hfill & {q = \hat{q}\;{\text{on }}\Sigma _{5} \times I.} \hfill \\ \end{array}$$
(9)
In the following, instead of
\(\vartheta ^0\) we will use
\(\eta ^0\) defined as the initial value of the entropy obtained from (
4)
\(_5\) evaluated at
\(t=0\). In the previous initial and boundary conditions
\(u_i^0\),
\(v_i^0\),
\(\theta ^0\),
\(\eta ^0\),
\({\hat{u}}_i\),
\({\hat{d}}_i\),
\({\hat{\varphi }}\),
\({\hat{\xi }}\),
\({\hat{\theta }}\),
\({\hat{P}}_i\),
\({\hat{R}}_i\),
\({\hat{\Lambda }}\),
\({\hat{H}}\) and
\({\hat{q}}\) are prescribed functions and the surfaces
\(S_i\) and
\(\Sigma _i\) such that
$$\begin{aligned} {\bar{S}}_i \cup \Sigma _i = \partial {\mathcal {B}}, \quad \quad S_i \cap \Sigma _i = \varnothing , \qquad i=1, \ldots , 5, \end{aligned}$$
where the closure is relative to
\(\partial {\mathcal {B}}\). The (external) data of the mixed initial-boundary value problem in concern are
$$\begin{aligned} \begin{aligned} \Gamma = \big \{&f_i, g, h, u_i^0, v_i^0, \theta ^0, \eta ^0,\\&{\hat{u}}_i, {\hat{d}}_i, {\hat{\varphi }}, {\hat{\xi }}, {\hat{\theta }}, {\hat{P}}_i, {\hat{R}}_i, {\hat{\Lambda }}, {\hat{H}}, {\hat{q}} \big \}. \end{aligned} \end{aligned}$$
Following [
2], we define an ordered array of functions
$$\begin{aligned}\begin{aligned} \pi =(&u_{i},\varphi ,\theta ,e_{ij},\kappa _{ijk},E_{i},V_{ij},\\ {}&g_i,\tau _{ij},\mu _{ijk},\sigma _{i},Q_{ij},\eta ,q_{i}) \end{aligned} \end{aligned}$$
as an admissible process on
\(\bar{{\mathcal {B}}}\times I\) with the following properties
1.
\(u_{i}\in C^{4,2}(\bar{{\mathcal {B}}}\times I)\), \(\varphi \in C^{4,0}(\bar{{\mathcal {B}}}\times I)\), \(\theta \in C^{2,2}(\bar{{\mathcal {B}}}\times I)\), \(e_{ij},V_{ij}\in C^{2,1}(\bar{{\mathcal {B}}}\times I)\), \(\eta \in C^{0,1}(\bar{{\mathcal {B}}}\times I),\) \(\kappa _{ijk},E_{i},\mu _{ijk},Q_{ij}\in C^{2,0}(\bar{{\mathcal {B}}}\times I)\), \(g_i,\tau _{ij},q_{i}\in C^{1,0}(\bar{{\mathcal {B}}}\times I)\);
2.
\({e_{ij}=e_{ji}}\), \(\kappa _{ijk}=\kappa _{jik}\), \(\,V_{ij}=V_{ji}\), \(\tau _{ji}=\tau _{ij}\), \(\mu _{kji}=\mu _{jki}\), \(Q_{ji}=Q_{ij}\) on \(\bar{{\mathcal {B}}}\times I\).
We say that
\(\pi\) is a process corresponding to the supply terms
\((f_{i},g,h)\) if
\(\pi\) is an admissible process that satisfies the fundamental system of field Eqs. (
1)–(
5) with the restriction (
7) on
\({\mathcal {B}}\times I\), where the coefficients must satisfy the symmetries (
6) and where
\(I=[0,t_1]\) is a bounded time interval. Then, if a process
\(\pi\) satisfies the initial conditions (
8) and the boundary conditions (
9), we identify it as a solution of the mixed initial-boundary value problem
\(\Pi\). Henceforth, the dependence on
\({\textbf{x}}\) is implicit, while the dependence on the time variable might be explicit. We consider the following notations for any function
\(f\)$$\begin{aligned} Af=f+\beta {\dot{f}},\qquad {\bar{f}}(t)=\int _{0}^{t}f(\tau )d\tau ,\qquad \forall t\in I. \end{aligned}$$
The proof of this lemma is immediate.
We consider the body
\({\mathcal {B}}\) subjected to two different external data systems
$$\begin{aligned}\begin{aligned} \Gamma ^{(\alpha )} = \biggl (&f_{i}^{(\alpha )}, g^{(\alpha )}, h^{(\alpha )}, u_{i}^{0(\alpha )}, v_{i}^{0(\alpha )}, \theta ^{0(\alpha )}, \eta ^{0(\alpha )}, \\&{\hat{u}}_{i}^{(\alpha )}, {\hat{d}}_{i}^{(\alpha )}, {\hat{\varphi }}^{(\alpha )}, {\hat{\xi }}^{(\alpha )}, {\hat{\theta }}^{(\alpha )},\\&{\hat{P}}_{i}^{(\alpha )}, {\hat{R}}_{i}^{(\alpha )}, {\hat{\Lambda }}^{(\alpha )}, {\hat{H}}^{(\alpha )}, {\hat{q}}^{(\alpha )} \biggr ), \end{aligned} \end{aligned}$$
with
\(\alpha =1,2\), and we denote the corresponding solutions of the mixed initial-boundary problem value with
$$\begin{aligned}\begin{aligned} \pi ^{(\alpha )} = (&u_{i}^{(\alpha )}, \varphi ^{(\alpha )}, \theta ^{(\alpha )}, e_{ij}^{(\alpha )}, \kappa _{ijk}^{(\alpha )}, E_{i}^{(\alpha )}, V_{ij}^{(\alpha )},\\&g_i^{(\alpha )}, \tau _{ij}^{(\alpha )}, \mu _{ijk}^{(\alpha )}, \sigma _{i}^{(\alpha )}, Q_{ij}^{(\alpha )}, \eta ^{(\alpha )}, q_{i}^{(\alpha )}). \end{aligned} \end{aligned}$$