1. Introduction and Scope of Work
1.1. Introduction
Additive and porous media contain solid matter with interspersed pores. The varying sizes of the pores and their varying density of distribution in the solid matter make such materials highly anisotropic. If the pore size is much smaller than the macro continuum scale, then the pores may be ignored. On the other hand, if the pore sizes approach the macro continuum scale, then in the development of the theory that describes deformation physics of such material, the varying pore sizes, their density and distribution in the solid matter play an important role.
There are many published works that describe how these complex pore spaces within the solid matter can be regularized in three mutually consistent directions so that the material may now be considered as having regularized but different anisotropies in these directions. This problem has been found in the published literature to be more tractable, but the approaches used vary drastically.
Another simple proposition is that we assume in each of three mutually perpendicular (or any other) directions there is a surrogate configuration of pores uniformly distributed in the volume and each pore is of the same size, and they collectively produce the same deformation physics and/or flow physics as the original configuration pore space does. With this assumption, the original porous material is characterized by one in which a characteristic pore length describes the level of anisotropy in each of the three directions. This idealized porous solid matter forms the basis of the foundational concepts (in reference [1]) and derivations presented here in this paper that will eventually help in establishing continuum thermodynamics for additive and porous media.
Since additively manufactured materials are inherently porous, in this paper it suffices to refer to both as porous media. In the following, we present a review of the published works, concentrating just on the fundamental concepts used rather than intricate details of the derivations. The motivation for doing so is to clearly show that our views and concepts presented in this paper on the approach towards establishing a continuum thermodynamic framework for porous media are different than those used currently in the published works, and the principles and the concepts considered in our approach are strictly supported by classical continuum thermodynamics, thus ensuring that the resulting new continuum thermodynamic framework is thermodynamically consistent.
At present, classical thermodynamics is the only valid thermodynamics we know for continuous matter. Thus, in deriving a new thermodynamic framework for newer physics (like porous media), we may require modifications of the existing conservation and/or balance laws and may also require the introduction of new conservation and/or balance laws. We must make sure both of these are supported by classical thermodynamics; if not, then the resulting new thermodynamic framework is not thermodynamically valid based on classical thermodynamics.
Perhaps Mandelbrot [2] and Schroeder [3] were the first to introduce the concept of mass fractional dimension. Tarasov [4] explains how this concept is more practical and can be used to measure fractional dimension as opposed to fractal dimension associated with limits.
In a porous medium, properties like mass obey a power law relation. If
is the mass of the porous media and
is the box size and
is the fractional dimension, then
(1)
Taking the natural logarithm of (1), we get
(2)
When measured data from the different-sized boxes
are plotted using
versus
, the slope of this line gives the value of
. Tarasov [4] showed that relation (1) can be derived using a fractional integral. Derivation of the fractional equation of continuity is also presented in [4]. In another paper [5], Tarasov presents elasticity of fractal material using a continuum model with non-integer dimensional space. This work also intrinsically uses the fractional integral definition of mass. In reference [6] the author considers fractional generalization of phase volume, volume element, and Poisson brackets. Tarasov [7] considers dynamical fractal solids using Euler’s equations and integrals of fractional order. In reference [8] Tarasov uses fractional integrals for hydrodynamic equations of fractal media. Equations describing conservation of mass, balance of linear momentum, and the first law of thermodynamics are derived. The book by Tarasov [9], Fractional Dynamics, has a treatment of dynamics of fractal media using fractional calculus. All these works by Tarasov are based on fractional integral formulation. Works based on fractional integrals can also be found in references [10]-[12]. Nonlocal continuum models and fractional calculus have also been used in references [13] [14] for continuum mechanics. This approach, based on fractional integrals, is widely used and is the subject matter of many other papers.
The second approach that is commonly found is based on the method of dimensional regularization using fractional integrals and then converting the model to the
-frame (Cartesian). Ostoja-Starzewski et al. [15] present what they refer to as homogenization of fractal media via dimensional regularization. The authors point out drawbacks of the approaches based on fractional-order integrals used by Tarasov. Their starting point is the following representation of mass
(3)
followed by a product measure. The authors [15] show that this approach and Tarasov’s work result in the same homogenization of the porous media. This approach eliminates fractional integrals with gamma function used in Tarasov’s approach. The authors introduce coefficients to scale length, followed by scaling of areas and volumes. The paper presents many details using this approach for deriving conservation and balance laws, treatment of elastic solids under finite strain, and many other topics. The approach described in reference [15] is subsequently used by Ostoja-Starzewski [16] for thermomechanics of fractal media, and by Ostoja-Starzewski et al. [17] for beams and fracture mechanics.
Li et al. present fractal solids, product measures and continuum mechanics in Chapter 33 of reference [18]. Ostoja-Starzewski in reference [19] presents continuum of fractal and porous media: integral relations and extremum principles and variational principles for elastic and inelastic media with fractal geometries; in reference [20] Li et al. discuss fractal solids, product measures and fractional wave equations; Ostoja-Starzewski in reference [21] discusses turbulence in porous media, and in reference [22], Joumaa et al. present wave propagation in isotropic fractal media; Balankin [23] [24] also considers fractal media, mechanics of continuum with fractal matrix. Nonsymmetry of the Cauchy stress tensor in the
-frame has been pointed out and used by Ostoja-Starzewski [25] to point out the need for micropolar physics in mechanics of porous media. We do not find any organized thermodynamic approach for addressing porous media physics in the published works.
1.2. Scope of Work
Additively manufactured and porous media are highly anisotropic due to varying pore sizes and their random distribution in the volume of matter. In additive materials, pore sizes and their distribution are not as random and can be controlled by the manufacturing methods employed. In the present work, the additively manufactured materials are treated in the same manner as porous media in general.
Consider a volume
of the porous media in the
-frame, original undeformed state of porous media. Let
-frame with axes
,
,
be a curvilinear coordinate frame in which
, i.e.
are fractions; a value of one is admissible for each of the three. We refer to
-space as fractional curvilinear coordinate space. Fractional indices
are chosen such that when volume
from
-frame is mapped into volume
in the
-frame, the pore space is completely eliminated, and the matter is homogenized through dimensional regularization. Thus, the volume
only has solid matter, homogeneous and isotropic. We use mapping
and its inverse
that are completely general, thus permitting description of highly complex porous media for
; that is, there are no restrictions in the mapping that restrict the presence of porosity in volume
in the
-frame. We show that the relationship between volumes
and
due to a general mapping
reduces to relationship (1) for a specific choice of mapping
, confirming that the general form of the mapping
used here is a more general case of homogenization by dimensional regularization.
When disturbance is applied to volume
, volume
deforms into
, i.e.
, likewise volume
deforms into
, i.e.
. In volumes
and
, the matter is porous, whereas
and
contain undeformed and deformed isotropic, homogeneous solid matter without any pore space. Thus, the conservation and balance laws of classical thermodynamics can be applied to volumes
and
to derive conservation and balance laws in the
-frame (Eulerian description) and
-frame (Lagrangian description). The same holds for constitutive theories.
When the symmetric Cauchy stress tensor
in volume
in
-frame is transformed into Cauchy stress tensor
in the undeformed volume
in
-frame we find that
is not symmetric. The symmetric part of the Cauchy stress tensor
is self-equilibrating, while the antisymmetric components of
generate moments within the deformed volume in the
-frame that must have a mechanism of equilibrating with moments due to some other mechanism so that the deforming volume in the
-frame is in thermodynamic equilibrium. The anisotropy of the porous matter provides resistance to the rotation field in the deforming volume
, and a consequent moment field is generated in the deformed volume. This moment field is balanced by the one created by the antisymmetric components of
, hence preserving thermodynamic equilibrium. The physics of the moment field due to anisotropy of the porous media is exactly parallel to the micropolar physics in microcontinuum theories in which rigid body rotations of the rigid microconstituents interacting with the classical rotation field generate a moment field within the deforming volume. When volume
is mapped into
, this physics in volume
in
-frame must also be mapped into the volume
in the
-frame. This can be accomplished by assuming that the solid matter in volume
is micropolar, isotropic, and homogeneous containing rigid microconstituents that only experience rigid body rotations. The classical rotation field in
is a free field in the absence of microconstituents. The presence of microconstituents interferes with the free field; as a consequence, the classical rotation field becomes the rigid rotations of the microconstituents [1]. Micropolar physics due to the moment field in volume
generated due to microconstituents and the classical rotation field is the map of the micropolar physics in the porous volume in the
-frame. Henceforth we refer to the solid matter in volume
as micropolar medium; therefore, the Cauchy stress tensor
in volume
is not symmetric any more.
For the mathematical models in
- and
-frames to be useful in actual applications, some quantities need to be mapped in
- or
-frames, as our experiments, measurements, and numerical simulations are all in
- and
-frames. First, we note that the laws of thermodynamics are not applicable to volumes
and
in
- and
-frames. But, there are a few facts that help us in arriving at a workable correspondence between the
- and
-frames.
The volume of a material point in
in the
-frame has a map to the corresponding volume
in the
-frame. Both have a center of mass
and
in
- and
-frames. When the matter is disturbed, both
and
have translations (displacements). Between
- and
-frames, the definition of the deformation gradient tensor
exists, but the deformation gradient tensor
between
- and
-frames does not exist (as it requires continuum mechanics); however, the gradients of
with respect to
do exist at the center
in volume
in the
-frame, as these are simply rates of change in the spatial directions (simple calculus). Thus, components of
exist, but
does not. Also, time derivatives of
exist at point
, as these are simply rates of change with respect to time. In the strain measures in
-frame, we need components of
when transforming them to
-frame, we take the liberty to use
for these components for convenience of notation, but keeping in mind that the use of
only implies that we know its components realizing full well that it cannot be defined using deformation gradient tensor
between
- and
-frames as the laws of classical thermodynamics do not hold in
- and
-frames.
The concepts described above toward continuum thermodynamics for porous media are organized in a series of four additional papers. In this paper we present the mapping of the
-frame
-frame and its inverse. We consider the mapping of points, lengths, areas, and volumes between the two frames. This is followed by the kinematics of motion in
-frame and likewise in the
-frame; displacement; deformation gradient tensor; inverse deformation gradient tensor; mapping of points, lengths, areas and volumes between
- and
-frames; covariant and contravariant basis in
-frame; some useful differential relations in
- and
-frames and their map in
-frame; gradient of a scalar; divergence of a tensor of rank one; gradient of a tensor of rank one; divergence of a tensor of rank two; Laplacian of a scalar field and material derivative. This is followed by the derivation of nonlinear strain measures, Cauchy strain tensor
and Green’s strain tensor
(covariant measures) in the Lagrangian description. Derivation of Eulerian strain measure such as Cauchy
and Almansi strain
follows a similar procedure. Derivation of the contravariant Cauchy stress tensor
in
-frame and the moment tensor
is presented followed by first and second Piola-Kirchhoff stress and moment tensors in
-frame based on contravariant Cauchy stress and moment tensors
and
. This is basic material needed for deriving conservation and the balance laws in Eulerian and Lagrangian descriptions in
- and
-frames. These will be presented in forthcoming follow-up papers.
2. General Considerations
We consider regularized porous media by using surrogate pore lengths in three mutually perpendicular or any other convenient directions, thus defining three characteristic pore lengths in three directions. Following Tarasov, let us consider the mass of solid matter in a volume that has anisotropies in three mutual directions. If our objective is to derive conservation and the balance laws for a porous medium that is anisotropic, then we must show how to map this matter into a homogeneous isotropic medium for which we can use classical thermodynamics. In simplistic terms, there are two basic approaches possible: in the first approach, keep the volume the same as
in
-space but change the density such that the solid matter fills the volume and the pore space is eliminated. In the second approach, we keep the density the same as the density of the solid matter but change the volume. In both cases, the end result is that the pores are eliminated. Additionally, we must ensure that in both approaches the matter is homogenized, hence remains isotropic and homogeneous. The first approach requires density change while the volume remains unchanged. In the second approach, preserving the density of solid matter in porous media requires that the volume must change. This little illustration shows that the problem is not mass or density; it is the pore space that must be eliminated by transforming the matter from
-space to another space while preserving isotropic, homogeneous nature of matter through intrinsic homogenization in the mapping. If we are studying transient heat conduction in a porous medium, then in the case of the first approach, keeping the volume the same we must now determine surrogate specific heat and thermal conductivity. Use of a fractional index for mass is obviously not appropriate in this case. This suggests that the second approach, in which the solid material properties are preserved and the volume is changed to achieve pore space free isotropic homogeneous solid matter through homogenization by dimensional regularization, is preferable.
Based on surrogate porosity lengths in three mutual directions, we have three fractional indices that must define the new volume in which the original properties of the solid matter are preserved without pore space.
We have pointed out in the Abstract and the Introduction that the porous medium in
- and
-frames has micropolar physics which must also be mapped in the
-frame in volume
. This is done by assuming that the solid matter in volume
is micropolar, i.e. it is isotropic and homogeneous but contains rigid microconstituents that can only experience rigid body rotations. Interaction of the rigid body rotations of the microconstituents and the classical rotation field (a free field in the absence of microconstituents) provides a map of the micropolar physics in
-,
-frames in volumes
and
into the volumes
and
in
- and
-frames. Thus, the volume
certainly contains isotropic and homogeneous solid matter without void space, but this solid matter is micropolar. Thus the continuum thermodynamics of porous media requires derivation of conservation and balance laws and the constitutive theories for micropolar solid media in curvilinear fractional spaces
and
. In the remainder of this paper, it is always understood that we are considering micropolar theory in curvilinear fractional spaces
and
.
3. Fractional Spaces
Let
-space
be a Cartesian coordinate space. Let
,
and
be fractional indices describing porosity in
-directions. We consider a nonorthogonal curvilinear coordinate space with axes
,
and
and
,
and
. This curvilinear coordinate space
-frame (no sum) or
-frame will be referred to as curvilinear fractional space or fractional space. The surrogate porosities define
,
,
, and the volume
transformed from the
-frame to
in the
-frame eliminates the pore space and preserves the original properties of the solid matter in the porous media. Consideration of micropolar physics in
-space provides a map of micropolar physics present in volumes
and
in
- and
-spaces. Conservation and the balance laws of classical continuum thermodynamics can be applied to the micropolar volume
in the
-frame, as the matter is isotropic and homogeneous in volume
, to derive details of the mathematical model in the
-frame suitable for porous media in the
-frame.
4. Mapping of Points, Lengths, Areas and Volumes in
- and
-Spaces
Let the mapping
be such that
(4)
then
(5)
The mapping
maps
from the
-frame to the corresponding points in the
-frame.
is the inverse of the mapping
that maps points from
-frame to
-frame. This is a mapping of points between the cartesian and curvilinear fractional frames.
4.1. Mapping of Lengths
Consider material points
and
in the
-frame. These map into material points
and
in
-frame. Lengths
and
are not the same, and their directions are not the same either. Using the mapping given by
in (4) we can write
(6)
Using the Taylor series expansion of the first term on the right-hand side of (6) and retaining only up to linear terms in
, we obtain
(7)
Neglecting
term and equating coefficient of
in both sides of (7) we obtain
(8)
or
(9)
Equation (9) maps elemental lengths
,
,
from
-frame to elemental lengths
,
,
in
-frame. Likewise, if we consider
(10)
Then, following the same procedure, we can obtain
(11)
or
(12)
Equation (12) maps
from
-frame to
in
-frame. From (9) and (12), we note that
(13)
Also
(14)
and
are covariant and contravariant vectors.
are tangent to
,
and
at a material point.
and
form nonorthogonal reciprocal basis in
-frame.
4.2. Mapping of Areas
If
,
and
are originating from a material point
, these are mapped into
,
and
in
-frame. Area formed by
,
is
. Likewise the area formed by
,
is
.
(15)
or
(16)
(17)
Similarly
(18)
using
(19)
in (15) and following details in reference [26] we can derive the following
(20)
4.3. Mapping of Volumes
(21)
(22)
using
(23)
Using (23) in (21) and then using (22) (see [26]) we can obtain
(24)
This is a fundamental relationship that shows
in
-space is reduced to volume
in
-space such that the pore space is completely eliminated in the volume
. That is
,
,
in
are such that
has no pore space.
Remarks
1) Mapping
in (4) and its inverse in (5) is the most general case for which
in (24) is a fully populated matrix. The explicit form of the mapping is necessary to further study (24).
2)
,
,
are fractions associated with the porosities that have been used to define the curvilinear fractional space
.
3) We consider a simpler case of mapping
in order to demonstrate that (24) indeed describes the porosity physics correctly. Let
(25)
(26)
Mappings (25) and (26) imply that porosity in one direction does not influence the porosity in the other two directions (could be orthogonal). That is, porosity in each direction is isotropic and is defined by
only. Using mapping
in (25),
becomes a diagonal matrix
(27)
(28)
Based on the isotropic surrogate pore space assumption in each of the three directions, (28) is a fundamental relationship. Consider
(a cube with edge length of
), then (28) reduces to
(29)
a)
implies that
, i.e.
has no pore space, i.e., 1 is the right limit point of the interval for each
, and each
must be greater than zero. This establishes that
(30)
b) Progressively decreasing values of
from one imply progressively increasing pore space.
c) If we choose
(31)
implying completely isotropic pore space in all three directions in the entire volume of matter, then (29) reduces to
(32)
We note:
1) when
, the volume
has no pore space, hence in this case (32) yields
.
2) when
,
and a progressively decreasing value of
defines porous media with progressively increasing pore space.
Thus, with the same assumptions as in references [4]-[9] [15], we are able to show that our approach presented here yields the same results as in references [4]-[9] [15], confirming that for this special case of mapping
, our approach results in the same homogenization by dimensional regularization in volume
as in references [4]-[9] [15].
In the remainder of the paper we go back to
and
in (4) and (5) that yield fully populated
matrix. Simplification in (31) yielding (32) has been used as an illustration to show that the general mapping
and its inverse
for a specific choice considered here yield the same results as in published works. Henceforth, we assume that homogenization by dimensional regularization always exists in the proposed mapping of
.
5. Kinematics of Motion, Deformation and Their Measures
A volume of matter
with material points
deforms into a volume
with material points at
. Likewise
deforms into
. Since
and
contain only isotropic homogeneous solid matter without pore space, the conservation and balance laws must be derived using volumes
and
in
- and
-spaces, considering micropolar physics. Upon deformation, we have
(33)
and
(34)
The material points in
-frame and
-frame experience the same displacement
, the only difference is that in
-frame (Equation (33)) displacement
is a function of
and
whereas in
-frame displacement
is a function of
and
.
5.1. Deformation Gradient Tensors in
- and
-Frames
Material points in
-space map into
-space after deformation. Let
and
be two material points constituting a line segment
. Upon deformation these occupy
and positions in
-space. Let the motion be given by
(35)
we assume that
are continuous and differentiable in their arguments, then
(36)
Expanding the first term on the right-hand side of (36) in Taylor series about
,
(37)
Neglecting
term, we obtain
(38)
In Murnaghan’s notation, we can write (38) as
(39)
On the other hand, if we begin with
(40)
Then, following the same procedure, we can obtain
(41)
or
(42)
is the deformation gradient tensor in
-frame and
is the inverse deformation gradient tensor in
-frame.
(43)
Thus
(44)
Columns of
contain covariant base vectors
, tangent to
-axes at a material point. Rows of
contain contravariant base vectors
.
and
form reciprocal nonorthogonal bases. Using (34) we can express
in terms of displacement gradient tensor
(45)
is the deformation gradient tensor in
-frame. Likewise using
in (34) and differentiating with respect to
we obtain
(46)
(47)
Columns of
are covariant vectors , tangent to the deformed line
,
(48)
Rows of
are contravariant vectors ,
(49)
and
form nonorthogonal reciprocal bases.
5.2. Areas and Volumes in
-,
- and
-,
-Spaces
5.2.1. Areas
First we consider
- and
-spaces. We consider relationships between areas in
- and
-spaces as well as in
- and
-spaces. Let
,
and
be line segments in
-space emanating from material point
and
,
and
be their maps in the
-frame. Then, let the area formed by
and
be
and the area formed by
and
be
, then
(50)
(51)
(52)
Similarly
(53)
using
(54)
(55)
using (54) and (55) in (51) and following details in reference [26]
(56)
5.2.2. Volumes in
- and
-Spaces
(57)
(58)
using
(59)
Using (59) in (57) and then using (58) we can obtain
(60)
6. Differential Relations in
- and
-Frames
In the following, we derive some differential relations in
- and
-frames and express them in
- and
-frames. Let
(61)
(62)
and
(63)
be differential operators (gradient operators) in
-,
- and
-frames.
6.1. Gradient of a Scalar Field in
-Frame
Let
be a scalar field; then the operator
acting on
produces
(64)
also
(65)
Alternatively
(66)
6.2. Divergence of a Tensor of Rank One in
-Frame
Let
be a tensor of rank one, then dot product of
and
is called the divergence of the tensor of rank one,
(67)
(68)
(69)
or
(70)
6.3. Gradient of a Tensor of Rank One in
-Frame
Gradient of
, a tensor of rank one, is the dyadic product of the differential operator
and
and is denoted by
or
.
(71)
or
(72)
or
(73)
Therefore
(74)
6.4. Cross Product of
with a Tensor of Rank One (Curl of a Tensor of Rank One)
If
is a tensor of rank one, then Curl of
in
-space is
(75)
or
(76)
Also
(77)
Therefore
(78)
6.5. Divergence of a Tensor of Rank Two in
-Frame
Let
be a tensor of rank two, then its divergence is denoted by
and is defined as
(79)
(80)
(81)
also
(82)
(83)
Then we have
(84)
6.6. Laplacian of a Scalar Field
It is defined as the divergence of the gradient of a scalar field. Let
be a scalar field, then
(85)
Therefore
(86)
6.7. Material Derivative
While working in
- and
-spaces we often need the material derivative of a quantity defined in
-space. Let
be some quantity of interest, then we need to determine
(87)
(88)
(89)
or
(90)
From (90) we conclude that the material derivative of a quantity defined in
-frame is the same as the material derivative of the quantity defined in the
-frame. When
, that is in Lagrangian description, then
(91)
7. Strain Measures
In this section we derive nonlinear strain measures (finite deformation, finite strain) in Lagrangian and Eulerian descriptions in
- and
-frames.
and
are undeformed and deformed volumes with
and
coordinates of the material points.
7.1. Covariant Cauchy Strain
Consider lengths
and in
- and
-frames. Then
(92)
Since
is fixed, the measure of deformation must be in the term that does not contain
and
. We define Cauchy strain
as
(93)
Since
(94)
(95)
in (95) can be expressed in terms of
.
(96)
Therefore
(97)
Definition of
contains
, gradients of
with respect to
, hence in
-frame. We clearly see that
is symmetric (from (93)). We note
derivation in
-frame is not possible but components of
that are gradients of
with respect to
do exist.
7.2. Covariant Green’s Strain Tensor
Consider
(98)
(99)
Since
are fixed, all the deformation is contained in the terms that do not have
and
, we define Green’s strain
as
(100)
Since
, (100) can be written as
(101)
Using (96) we can write (101) as
(102)
Green’s strain can be expressed in terms of Cauchy strain using (97).
(103)
Remarks
For small strain and small deformation, the nonlinear terms in the strain measures can be omitted, and we obtain
1) Cauchy strain
(104)
or
(105)
2) Green’s strain
(106)
or
(107)
8. Stress Tensors
8.1. Cauchy Stress Tensor
Consider volume
and tetrahedron
such that the oblique plane of the tetrahedron has area
. Upon deformation, the tetrahedron is distorted into
in
-frame. Let
be covariant base vectors to the deformed lines
at a material point, where
are defined by the columns of deformation gradient tensor
. Let
be the edge of the deformed tetrahedron (covariant directions); then the vectors normal to the faces of the tetrahedron
are contravariant base vectors
. We have the following
(108)
(109)
The contravariant stress components
acting on the faces of require covariant bases in
-frame. The components of
in the basis of
-frame are given by
(110)
Substituting from (108) in (110) we obtain
(111)
(112)
(113)
or
(114)
in which
(115)
We note that
(116)
Therefore,
(117)
or
(118)
or
(119)
is a stress tensor with dyads in
-frame.
is contravariant Cauchy stress tensor. It acts on the flat faces of a tetrahedron; the faces are orthogonal to each other and are parallel to the faces of the
-frame, but the oblique plane of this new tetrahedron is the same as that of the actual deformed tetrahedron
.
Cauchy principle can be applied to
, force per unit area on the oblique plane of
and the stress tensor
. We note Cauchy principle is not applicable to the stress tensor
. Furthermore, the areas of the faces of the new tetrahedron are assumed to be the same as those of the original deformed tetrahedron
. Thus
(120)
Following a similar procedure, we can also define the covariant Cauchy stress tensor
for which Cauchy principle similar to (120) holds. As discussed in the references [26] [27],
is only valid for infinitesimal deformations, hence not useful in nonlinear theories.
8.2. Contravariant Cauchy Stress Tensor in
-Frame
This is useful when transforming the mathematical model from
-frame to
-frame in Lagrangian description or from
-frame to
-frame in Eulerian description. Consider the integral of the average stress
on the oblique plane of the deformed tetrahedron with area
and use Cauchy principle.
(121)
Using
(122)
Substituting from (122) in (121)
(123)
Thus,
(124)
or
(125)
Equation (125) shows that Cauchy stress tensor in
-frame is not symmetric, even when
is symmetric.
8.3. First Piola-Kirchhoff Stress Tensor
in
-Frame
Consider
-space. The correspondence rule is
(126)
(127)
(128)
Thus,
(129)
(130)
Using (129) and (130) in (126),
(131)
Therefore,
(132)
Thus,
(133)
is clearly nonsymmetric, even when
is symmetric.
8.4. Second Piola-Kirchhoff Stress Tensor in
-Frame
In this case, we use the following correspondence rule
(134)
(135)
Using (128) in (135),
(136)
and
(137)
Equating (136) and (137),
(138)
Therefore,
(139)
Thus,
(140)
When
is symmetric,
is also symmetric.
8.5. First Piola-Kirchhoff Stress Tensor
in
-Frame
Between
-frame and
-frame we have
(141)
(142)
Substituting from (142) in (141),
(143)
Therefore,
(144)
or
(145)
8.6. Second Piola-Kirchhoff Stress Tensor
in
-Frame
Between
- and
-frames we have
(146)
and
(147)
Using (147) in (146),
(148)
Therefore,
(149)
or
(150)
9. Moment Tensors
Since the map of the porous material in the curvilinear fractional frame is micropolar, in addition to the stress tensors, we also have moments and moment tensors independent of the forces and stresses. For defining the moment measures, we also consider the same deformed tetrahedron
as considered for stress measures, with covariant vectors
forming the edges and the contravariant vectors
being normal to the flat faces of the tetrahedron
formed by the covariant pairs , and .
9.1. Contravariant and Covariant Cauchy Moment Tensors in
-Frame
The contravariant Cauchy moment tensor
or
is defined using tetrahedron
. Let
or
be the contravariant moment components defined using contravariant directions
; then we can express a moment tensor
using the covariant vectors in
basis of the
-frame.
(151)
Using (108), we can write (151) as
(152)
or
(153)
or
(154)
or
(155)
in which
(156)
We note that
(157)
Using (157) in (156),
(158)
or
(159)
or
(160)
is the contravariant Cauchy moment tensor acting on the faces of a tetrahedron in
-frame where faces are flat and orthogonal to each other and are parallel to the planes of the
-frame. The oblique plane of
and this new tetrahedron remains the same. Furthermore the areas of the faces of
and those of the new tetrahedron remain the same. Cauchy principle holds for
and
moment tensors.
(161)
where
or
is the average moment per unit area on the oblique plane of the tetrahedron
. Just like
, the covariant Cauchy moment tensor
is also not valid for nonlinear deformation.
9.2. Contravariant Moment Tensor in
-Frame
This is useful when transforming the mathematical model from
-frame to
-frame in Lagrangian description. Consider the integral of the average moment on the oblique plane of the tetrahedron with area and we use Cauchy principle
(162)
Using
(163)
Using (163) in (162) and equating the corresponding integral in
-frame,
(164)
Therefore,
(165)
and
(166)
9.3. First Piola-Kirchhoff Moment Tensor in
-Frame
Consider
- and
-frames and the correspondence rule
(167)
(168)
(169)
(170)
Using (167) and (168),
(171)
Using (170) and (171) in (167),
(172)
Therefore,
(173)
Thus,
(174)
is non-symmetric even when
is symmetric.
9.4. Second Piola-Kirchhoff Moment Tensor in
-Frame
In this case, we use the correspondence rule
(175)
(176)
(177)
Using (169) in (176),
(178)
Equating (177) and (178),
(179)
Therefore,
(180)
Thus,
(181)
If
is symmetric, then
is also symmetric.
9.5. First Piola-Kirchhoff Moment Tensor in
-Frame
Between
-frame and
-frame we have
(182)
(183)
Substituting from (183) into (182),
(184)
Therefore,
(185)
or
(186)
and
(187)
9.6. Second Piola-Kirchhoff Moment Tensor in
-Frame
Between
- and
-frames we have
(188)
Using
(189)
in (188) we obtain
(190)
Therefore,
(191)
or
(192)
and
(193)
10. Summary and Conclusions
In this paper, we have shown that additive or porous matter volume
in
-space can be mapped into a volume
in the fractional curvilinear coordinate space
with nonorthogonal curvilinear coordinate axes
,
,
such that the volume
contains no pore space and results in homogenization by dimensional regularization [4] [5] [15]. We have considered the most general case of the mapping
,
and
, in which the gradient tensor
mapping elemental lengths
,
,
from
-space to the elemental lengths
,
,
in
-space is fully populated, permitting any desired choice of
based on the porosity of the solid media. It is shown that
in which a value of 1 for all
implies a nonporous medium. Progressively decreasing values of each
from one imply progressively increasing porosity. We have shown that within the limiting choice of mapping
as used in references [4]-[9] [15], the approach presented in this paper is in complete agreement with reported findings.
Besides porosity and homogenization by dimensional regularization, another aspect of the physics of porous media is that when a symmetric Cauchy stress tensor in volume
containing nonpolar, isotropic, and homogeneous matter in the
-frame is mapped or transformed into
-space, it becomes non-symmetric. In the volume
in the
-frame, the symmetric components of the stress tensor are self-equilibrating, while the antisymmetric components generate a moment field in the volume
(or
) which must be balanced by an equal and opposite moment field from some other source. Anisotropy of porous matter and its interaction with the classical rotation field (or a similar rotation field) in the deforming volume
or
generates a moment field that balances the moment field due to antisymmetric stress components. This physics is exactly parallel to the definition of physics in micropolar microcontinua.
Thus, when mapping
in
-frame two things must happen: solid matter in
is mapped in
without porosity and is homogenized by dimensional regularization and secondly the micropolar physics of volume
or
must also be mapped in the volume
. This is easily accomplished by assuming that volume
contains isotropic, homogeneous solid matter with rigid microconstituents that only experience rigid body rotations in volume
i.e. solid matter mapped in volume
is micropolar. Interaction of the rigid rotations of the microconstituents and the classical rotation field in
must generate the same moment field as in volume
or
. We derive conservation and balance laws for micropolar solid matter for
in the
-frame or
in the
-frame using conservation and balance laws of classical thermodynamics with appropriate modification due to new micropolar physics. We may also need to introduce new conservation and/or balance laws. This is all permissible as long as these are supported by classical thermodynamics, because only then the new thermodynamic framework is thermodynamically consistent, i.e., supported by classical thermodynamics, which is essential for its validity. Derivations of conservation and balance laws for micropolar solid media in the
-frame for compressible nonlinear deformation are presented in the follow-up papers. We remark that the assumption of microconstituents in volume
is only used to create the necessary micropolar physics in
or
, which is a map of micropolar physics in volumes
and
.
In this paper, basic material in
-,
- and
-frames that is necessary for deriving conservation and the balance laws and the constitutive theories has been presented. Mapping of points, lengths, areas and volumes in
- and
-spaces; kinematics of motion, deformation and measures; differential relations in
- and
-frames; nonlinear Cauchy and Green’s strain measures in
-frame; contravariant Cauchy stress tensor
and Cauchy principle in
-frame; contravariant Cauchy stress tensor
in
-frame; first and second Piola-Kirchhoff stress tensors in
-frame based on contravariant Cauchy stress tensor
; and the first and second Piola-Kirchhoff stress tensors in
-frame are presented. Similar material has also been presented for the moment tensors that exist due to micropolar physics.
Derivations of the Cauchy moment tensor, first and second Piola-Kirchhoff moment tensors in
-frame are presented followed by their transformation or mapping in the
-frame. The material presented in this paper consists of basic core definitions and derivations that are essential in the derivation of conservation and the balance laws in
- and
-frames and the constitutive theories addressed in subsequent papers.
Acknowledgements
The financial assistance provided to the second author by the Department of Mechanical Engineering and the Department of Mathematics through Graduate Teaching Assistantships is gratefully acknowledged. The second author also gratefully acknowledges the Department of Mechanical Engineering for providing a Summer Research Scholarship. The computing and other facilities provided by the Computational Mechanics Laboratory (CML) of the Department of Mechanical Engineering are also acknowledged.