UNCLASSIFIED

AD 664 271

THEORY OF MICROPOLAR ELASTICITY

A.C. Er ingen
Princeton University
Princeton, New Jersey

June 1967

Processed for...

DEFENSE DOCUMENTATION CENTER

DEFENSE SUPPLY AGENCY

FOR FEDERAL SCIENTIFIC AND TECHNICAL INFORMATION

U S DEPARTMENT OF COMMERCE / NATIONAL BUREAU OF STANDARDS / INSTITUTE FDR APPLIED TECHNOLOGY

UNCLASSIFIED
THIS DOCUMENT IS BEST
QUALITY AVAILABLE. THE COPY
FURNISHED TO DTIC CONTAINED
A SIGNIFICANT NUMBER OF
PAGES WHICH . DO NOT
REPRODUCE LEGIBLY.
THEORY OF MICROPOLAR
ELASTICITY

by A, Cemal Bringen

for Office of Naval Research

Department of the Navy
Contract
tract N%14J7'A-0I5140004
D n n
Technical Report No. 1
June 1967
fl ^N Z 3 1968 //[
PRINCETON UNIVERSITY
Department of Aerospace and Mechanical Sciences
G

Thia doauaoat hca B5 affw» J

for nul lie roLtro cni aokc te
^r^hnMen la tmlinitod.

Reproduction in whole or in part is permitted for any purpose

of the United States Government
Reproduced by the
CLEARINGHOUSE
(or Federal Scientific & Technical
Intormation Springfield Va 22151
A
V^
 THEORY OF MICROPOLAR ELASTICITY

A. Cemal Eringen

for

Office of Naval Research

Department of the Navy
Contract N-0014-67-A-0151-OOO4

Technical Report No. 1

June 1967

PRINCETON UNIVERSITY

Department of Aerospace and Mechanical Sciences

Reproduction in whole or in part is permitted for any purpose of the

United States Government
 -1-

TABLE OF CONTENTS

Abstract 2

1. Introduction 4

2. Deformation and Microdeformation 9

3. Strain and Microstrain Tensors 15

4. Micropolar Strains and Rotations ...... 20

5. Geometrical Meaning of Micropolar Strains and Rotations, . . 28

6. Invariants of Strain Tensors 33

7. Volume Changes 37

8. Compatibility Conditions 40

9. Some Special Deformations 43

10. Motion, Micromotion, Material Derivative of Tensors 53

11. Velocity, Acceleration, Microrotation, Spin 56

12. Material Derivative of Arc Length 61

13. Races of Strain Measures 67

14 External and Internal Loads. 70

15. Mechanical Balance Laws 73

16. Stress and Couple Stress 79

17. Local Balance Laws , 83

18. Conservation of Energy , . 88

19. Principle of Entropy - . . 91

20. Theory of Micropolar Elasticity 94

21. Restriction on Micropolar Elastic Moduli . , . . .... 100

22. Field Equations, Boundary and Initial Conditions 103

23. Indeterminate Couple Stress Theory , , . 107

24. Propagation of Waves in an Infinite Micropolar Elastic Solid 113

 -la-

25. Reflection of a Longitudinal Displacement Wave 120

26. Micropolar Surface Waves 124

27. Stress Concentration Around a Circular Hole 128

28. Galerkin and Papkovitch Representations 134

29. Micropolar Infinite Solid Subject to Concentrated Force and

Concentrated Couple • • • • 139

Summary 143

Possible Directions for Future Research , 144

Acknowledgement 145

References 146

List of Symbols . 150

Figure Captions

Figures
 -2-

ABSTRACT

In this article we present a self-contained account of the recent

theory of micropolar elasticity. Micropolar elastic materials, roughly

speaking, are the classical elastic materials with extra independent degrees

of freedom for the local rotations. These materials respond to spin inertia

and body and surface couples, and as a consequence they exhibit certain new

static and dynamic effects, e.g., new types of waves and couple stresses-

The theory is fully deterministic as against the background of the recently

popular indeteiministic couple stress theory (cf. Art 23). The mechanics

of certain classes of materials with fiberous and elongated grains (e.g ,

dumbbell types ' 0rains) represents a potential field of application of the

theory.

The geometry of deformation and its measures are introduced on a more

effects
general background of materials exhibiting granular and microstructure^Cmicro-

morphic materials). Various types of microstralns and microrotations are

discussed. Compatibility conditions for the micropolar strains are derived

The kinematics of strains, microstralns and rotations are presented Basic,

laws of motion, conservation of mass, conservation of microlnertia, balance

of momentum, balance of moment of momentum, conservation of energy are postu-

lated and their local forms are obtained. The thermodynamics of micropola-

solids Is formulated and the consequences of the entropy inequality are

discussed. Constitutive equations are found for the linear theory of micro-

polar elasticity. The basic field equations and initial and boundary conditions

are given.

The indeterminate couple stress theory is shown to result as a special

case of the theory when the motion is constrained. Several static and dynamic

problems are solved to reveal seme ot the new physical phenomera exhibited by
 -3-

the theory. These include propagation of waves in infinite micropolar

elastic solids, reflection of various types of micropolar waves in a

half space, surface waves, the stress concentration around a circular

hole in a tension field, and force and moment singularities in an infinite

solid. The Papkovitch and Galerkin representations are presented.

The article is based mostly on the works of Eringen and his co-workers

published during the past several years. Many parts, however, contain new

compositions and several other results are presented for the first time.
 -A-

1. INTRODUCTION

Classical continuum mechanics is based upon the fundamental idea

that all material bodies possess continuous mass densities, and that the

laws of motion and the axioms of constitution are valid for every part

of the body no matter how small they may be. Accordingly, a small volume

AV enclosed within a surface AS possesses a mass density p defined by

(1.1) lim Am
P
' AV -0 AV

where Am is the total mass contained in AV. Here p is Independent of the

size of AV and depends only on the position vector x of a point in AV and

time t. Consider the following experiment for the measurement of p: The

mass density of a homogeneous material may be calculated approximately by

weighing a large number of pieces having different volumes and calculating

the ratio Am/AV for each piece. If the resulting numbers p are plotted

against AV, one finds that this ratio is nearly constant when AV is greater

than certain critical volume AV and begins to show dependence on AV when

AV < AV . The size of AV depends on the constitution of the material. As

AV approaches zero, this dependence becomes violent. Fig. 1.1. This situa-

tion is well-understood when we remember the granular and molecular nature

of materials. The classical continuum theory may therefore not represent

a good mathematical model for the approximation of a physical theory In the

range AV < AV ,

The loss of accuracy in classical continuum mechanics may stem from

another important reason. If the response of the body is sought to an external

physical effect in which the length scale is comparable to the average grain

or molecular size contained in the body, the granular or molecular constituents

 -5-

of the body are exited individually. In this case, the intrinsic motions

of the constituents (microelements) must be taken into account. This point

becomes clear especially in connection with the propagation of waves having

high frequencies or short wave lengths. When the wave length is of the same

order of magnitude as the average dimension of the microelements, the intrinsic

motions of the microelements of AV with respect to the center of mass of AV

can affect the outcome appreciably. This situation previals in practical

applications when the material under consideration is a composite material

containing macromolecules, fibers, and grains. For such materials, the critical

volume AV is of the order of magnitude of the cube of a fraction of an inch

to several inches. Solid propellant grains, polymeric materials, and fiberglas

are but a few examples for such materials.

Another example is the anomalous behavior of blood when flowing through

capillaries. Blood consists of a fluid (plasma) in which are dispersed elements

of microscopic dimensions (corpuscles). When blood flows through capillaries

whose diameter is comparable to that of the corpuscles, the flow characteristics

differ from those for large vessels. As a further example, experiments have

shown that the resistance of a solid to surrounding fluid can be reduced by as

much as 1/3 when a minute amount of additives are cast into fluid.

It may be conjectured that a rational treatment of surface tension, micro-

cracks, microfracture, and the mechanics of granular media and composite materials

ultimately will have to be based on the theory of microcontinua. The nature of

experimental work on the properties of such materials, no doubt, will be affected

with these developments.

Presently there exist several approaches to the formulation of micro-

mechanics. Some of these theories are very general in nature but incomplete

and not closed. Others are concerned either with special types of material

structure and/or deformations. Fundamental ideas contained in some of these

theories can be traced all the way to Bernoulli and Euler in connection with
 -6-

their work on beam theories. In the elementary beam theory, with each section

of the bar there is associated two sets of kinematical quantities, namely,

a deformation vector and a rotation vector and two types of Internal loads,

namely, the tractions and couples. In plate theory, we have a similar

situation. Bar and plate theories involving these independent quantities

were recorded by Kelvin and Tait[1879]. The existence of the stress couples

independent of tractions is essential to these theories For three-dimensional

bodies, this concept is found in the work of MacCullagh [1839] in connection

with his work on optics. Lord Kelvin went as far as building a model of what

he called "quasi-rigid" ether which is supposed to provide a mechanical model

for Maxwell's theory of electromagnetism. The existence and basis of couple

stress in elasticity was also noted by Voigt [1887] in connection with polar

molecules.

In a remarkable monograph, E. and F. Cosserat [1909] gave a unified

theory for the deformable bars, surfaces, and bodies. A Cosserat continuum

is defined as a three-dimensional continuum, each point of which is supplied

1 2
with a triad . By the use of a principle which they call "euclidean action"

and by calculating the variation of the internal energy density, they gave

the equations of local balance of momenta for stress and couple stress and

the expressions of surface tractions and couples. In the work of the Cosserat

brothers, we find that the effect of couple stress en the motion of deformable

bodies is fully taken into account.

Some fifty years elapsed after the work of the Cosserats with very little

activity in this field. The idea of a Cosserat continuum was revived in various

In the terminology of Truesdell and Noll [1960, Art 256] "directors", the
same terminology and similar ideas were used by Tcapin [1964], Green, Rivlin
and Naghdi [1965].
2
Equivalent to the principle of objectivity, ct. Enngen [1962, Art. 27].
 •7-

special forms by GUnther [1958], Grioli [I960], Aero and Kuvshinskli [1961],

and Schäfer [1962] of whom GUnther also remarked on the connection to the

theory of dislocations. The question of couple stress was reopened with

an incomplete theory of Cosserat bars and surfaces included in Truesdell

and Toupin [1960]. Mindlin and Tiers ten [1962], Toupin [1962], Eringen

[1962, Arts. 32, 40] recapitulated a special Cosserat continuum now known

as the (indeterminate) couple stress theory. In these theories, the rotation

vector is not an independent vector, consequently the antisymmetric part of

the stress and symmetric part of the couple stress remains undetermined

(cf. Art. 23 below). Eringen and Suhubi [1964a,b] and bringen [1964 ] intro-

duced a general theory of a nonlinear microelastic continuum in which the

balance laws of continuum mechanics are supplemented with additional ones,

and the intrinsic motions of the microelement contained in macrovolume AV

are taken into account. This theory, in special cases, contains the Cosserat

continuum and the indeterminate couple stress theory. Independently, a micro-

structure theory of elasticity was published by Mindlin [1964] and a multipolar

continuum theory by Green and Rivlln [19641. Botn of these theories appear

to have contacts with these of Eringen and Suhubi, in special situations,

Following these works, an intense activity began and literature now contains

several hundred papers in this and in related fields A proper assessment

of these works with appropriate references is beyond the scope of this article.

This article is concerned, basicially, with special types of continua

called micropolar continua. The thecry was initiated by Eringen and Suhubi

[19b4b, Art. 6 ] as a special case of their work on the microelastic solid

and was named couple stress theory. Later, Eringen [1966a,b] recapitulated

and renamed it micropolar theory and proved several uniqueness theorems. A

similar theory appears to be given, independently, by Palmov [1964] for the

 -8-

linear elastic solid. While the theory is fresh and no experimental work

has been published as yet, we believe that the results obtained so far are

sufficient to strengthen the future of the theory.

The theory of micropolar elasticity is concerned with material media

whose constituents are dumbbell molecules. These elements are allowed to

rotate independently without stretch The theory is expected to find appli-

cations in the treatment of mechanics of granular materials with elongated

rigid grains and composite fiberous materials.

The first eight sections (Arts. 2 - 9) of this article give a treat-

ment of the geometry of deformation and microdeformation, strain and rotation

measures, compatibility conditions, and some illustrative examples of deformation.

Sections 10 - 13 are devoted to kinematics and rate measures. External and

internal loads and the balance laws are discussed in Sections 14 to 17, and

energy and entropy in Sections 18 and 19. The constitutive equations of the

theory of micropolar elasticity and restrictions on the coefficients are

derived in Sections 20 and 21. The field equations, boundary and initial

conditions are prescribed and discussed in Section 22. In Section 23, we

show how the indeterminate couple stress theory arises as a special case of

the micropolar theory. Sections 24 through 29 are devoted to the solutions

of various problems.

Micropolar continuum mechanics is in the stage of its infancy. The

linear theory is reasonably simple and it lends itself to the solution of

some important boundary and initial value problems. A large class of unsolved

problems and experimental work offer a challenge to future workers.

 -9-

2. DEFORMATION AND MICRODEFORMATION

A material point P of a body B having volume V and surface S in its

undeformed and unstressed state may be located by its rectangular coordinates

m
X , X and X (or simply \, * M.S), Fig. 2.1. If the body is allowed

to move and deform under some external loads, it will occupy a region

having volume V and having surface S. Referred to the same rectangular frame

of reference, the new position of the point P will be x1, x , x (or simply

x, , k ■ 1,2,3). Under the assumption of Indestructibility and impenetrability

of matter, each material point in the undeformed body B will occupy a unique

position in the deformed body B.

Conversely, each point in t' can be traced back, to a unique point in B.

Thus, the deformation of the body at time t may be prescribed by a one-to-one

mapping

(2.1) x, ■ x. (X^ ,X2 ,X_ , t), 1,2,3

or its inverse motion

(2.2) L ■ X^(x.,X2,X-,t), 1,2,3

We assume that (2.2) is a unique inverse of (2.1) for all points contained in

the body except possibly some singular surfaces, lines,and points. For this

to be valid, the three functions x. (X.,X„,X_,t) must possess continuous partial

derivatives with respect to X., X.^and X. for all times, and the jacobian

is äx„ «3

(2.3) J = det Ü ^
ax.
3X,

ax.
3X-

ax.
\

»X,

ax"
!5
•X,
3xq

ax.
 -10-

must not vanish. Henceforth, we assume this to be the case.

The partial derivatives

(2,4) =
VK 3L ' h,k ' RL

are called deformation gradientst and they are basic in the study of continuum

mechanics.

We now consider a volume element AV enclosed within its surface AS in

the undeformed body. Let the center of mass of AV have the position vector X.

All materials possess certain granular and fibrous structure with different

sizes and shapes. If the physical phenomena under study has a certain charac-

teristic length (such as wave length) that compares with the size of prains

in the body, then the microstructure of the material must be taken into consid-

eration. In such situations, classical continuum mechanics must be modified

by considering the effect of the granular character of ti^e medium. Suppose

that the element AV + AS contains N discrete micromaterial elements AV + AS ,

(a ■ 1,2,...N), each with a mass density p . The position vector of a

material point in a microelement may be expressed as

(2.5) X(a) - X+ E(a)

-(a)
where = is the position of a point in the microelement relative to the

center of mass of AV + AS, Fig. 2.2. Upon the deformation of the body,

AV + AS goes into Av + As with the microelement displaced with respect to

the mass center. Because of the rearrangements and the relative deformations

of the microelements, the center of mass P may now move to a position p and

the material point Q to a new position q in the deformed body. The final

position of the a particle will therefore be

(2.6) x(a) - x+ C(a)

 -11-

where x is the new position vector of the center of mass of Av and £ is

the new relative position vector of the point originally located at X .

The motion of the center of mass P of AV is expressed as usual by (2.1) or

simply

(2.7) x - x(X,t)

The relative position vector C however depends not only on X and t but

also on r , i.e. ,

(2.S) E^ . C(a) (X.E(a).t)

A microstructure theory must lean heavily on the assumption charac-

terizing the nature of the functional character of i . Eringen and Suhubi

[1964a,b] and Eringen [1964] have constructed a general cneory in which (2.8)

is linear in E . The theory so constructed was later called by Eringen

the theory of micromorphic materials. The basic assumption underlying this

theory is the

Axiom of Affine Motion. The material points in AV + ^S undergo ji

homogeneous deformation about the center of mass, thus,

C(a) - x1(X,t)E(ia) + x2(X.t)E(2a) + X3(X,t)5(3a) , a - 1,2,...N

or simply

(2.9) t(a) . XK(X.t)-.K(a)

where sununation over repeated indices is understood This assumption is justi-

fiable on the physical grounds that when AV + AS is small enough, its motion

consists of a^ translation. £ rotation about its center of mass, and a^ homogeneous

deformation. Note that in all classical theories of continuous media, this last

assumption is missing (cf. Eringen [1962, Art. 10 1).

 -12-

Clearly, the theory can be extended to include quadratic and higher order

terms in H . However, it quickly takes a very complex form, losing its

simplicity and usefulness. In this article, in fact, we shall be concerned

mostly with a much simpler case, namely, the theory of micropolar elasticity .

This latter theory admits only rigid raicrorotations of the micro-

volume elements about the center of mass of the volume element. In other

words, the basic assumption (2.9) is further simplified by placing further

restrictions on the three vector functions x,,- As we shall see later, this
'is.

will amount to reducing the number of the microdeformation functions x,, from

three to one. In fact, we shall also be dealing mainly with the linear theory.

In classical continuum mechanics, the problem is centered around the determina-

tion of the spatial position x of all material points of the body at a given

instant. This means that when we are through witii all calculations, we will

have the three functions x^(X,t) determined. In the theory of micromorphic

materials, in addition the three vector functions x,,(X,t) (equivalently

nine scalar functions) •nust also be determined. The complicated nature of

the problem and the necessity of additional physical concepts and laws are

now apparent.

In coordinate form for the spatial position of a material point in a

microelement, we have

(2.10) x^ - ^(X.t) ♦ xkK(X,t)^a), k,K = 1,2,3

It is now clear that we have 3 + 9 = 12 functions, x^(X,t) and x. ^(X.t), to

determine the spatial position of the a material point x^

J iSt as in (2.2), we introduce the inverse nucromutions X such that

The theory was developed in the original paper ot Eringen and Suliubi [l%4b,
Sec. 6] as a special case of the general theory and was called the couple
stress theory. Eringen [1965] later named it the micropolar theory.
 -13-

Here and throughout this article, sununation is understood to be over the

repeated indices, e.g.,

X
VlU klXU + X
k2X2i + X
k3X3i

The symbols 4 and 6 are the Kronecker deltas which are 1 when

the indices take tue same numerical value and zero otherwise.

In component form (2.9) reads

(2.12) ^a) - xkK(X,t)5^0l)

Upon multiplication of both sides by >C. and using (2.11), we also get

(1.13) >' . V».«^

In vector form this reads

(2.14) E(a) - ^(».t)^

The rectangular components of X^ are denoted by X and those of xK

by XkK, i.e. ,

(2.15) XK- XkK(X.t)ik . ^-^(«.t)!,

where I and i, are, respectively, the unit base vectors for the material

coordinates JC, and the spatial coordinates x^.

The motion and the Inverse motion of a material point in a micro-

element are therefore expressed by

(2.16) x^a) - «jjUit) ♦ XkK(X,t)^0)

(2.17) X^a) - ^(K.t) + X

Kk
(x tK a)
' k
 -14-

In vectorlal form these read

(2.18) x(Ql) - x(X,t) + xKa,t)4a)

(2.19) X(0l) - X(x,t) ♦ Xk(x.t)^a)

It is clear that we may employ either representations (2.16) or

or
(2.17), .»nd that whenever either set of functions (ViXi^) (^»■^vi,)
are determined, the problem is completed since the other set is soluble

from the one that is found.

 -15-

3. STRAIN AND MICROSTRAIN TENSORS

The differential line element in the deformed body is calculated

through (2.18)

(3.1) ■'?(a)-<?.K+*LiK^>)dXK+*KdH'<0

where an index followed by a comma denotes partial differentiation. This

convention will be used throughout this article. Thus, for example,

9x _ ^L
(3.2)
-'K ' -W^ ' ^L,K ' 3^

3X »fl

are used for brevity. Note that (3.2). and (3.3). are the classical defor-

mation gradients, and (3.2)„ and (3.3). are the microdeformation gradients

of the present theory.

The square of the arc length is now calculated by forming

,. (a) .2 . (a) . (a)

(ds ) = dx • dx

Upon using (3.1) and forming the scalar product, we find

O.*) (ds(a))2 - (cKL ♦ 2 r^E,^ ♦ xkM>K xkN(LVN) «K^L

+ 2( +
V \LXkM,K=M) dX d5
K L

+ X
kKXkLd::KdrL

where we also dropped the superscript a on - and d= for brevity, since

this is understood whenever we use the letter ;(and also ()• In (3.4)
 -16-

we introduce the notations

Of these, C,,, is the classical Green deformation tensor and V

and T^, w are new ^^__^—^_—-^^

1CLM mlcrodeformation .^__^_
tensors of the present theory.

We now introduce the displacement vector u as the vector that

extends from X(a) to * , Fig. (3.1). Thus we write

(3.8) u(a) -x-X+C-E-u+S-E

where

(3.9) u . x - X

is the classical displacement vector, the components of which in X^ and x,

are, respectively,

(3 10)
- \'u- h' \\K - h

where

(3 12) 6 6
- kK ^ Kk 'h'h

Since the spatial and material frames are taken to be the same rectangular

frame of reference, 6, is none other than the Kronecker "elta which has
IcK
 -17-

the value one when the two indices take the same numerical value and zero

otherwise. It is possible to write x. ö, i Xj. and X^ I )C. But we keep

the convention of majuscule indices for the material frame and miniscule

indices for the spatial frame of reference. This convention is aspec ^v

useful in the nonlinear theory.

From (3.10) and (3.11), by partial differentiation, we obtain

(3 13) (6 + lJ
- \,K' LK L,K)6kL

Similarly, we introduce the microdisplacement tensors «t (X,t) and tfi , (x,t)

by

X (6 +
(3.13) kK " LK *LK)6kL

(3 16)
- V- (\k- V6Ki

By use of (3.9) and (3.15) and (3.16), we see that (3.8) may also be expressed

(3.17) u(a) - u+C- E- O^+VlA^K

(3.18) u(a) - U ♦ C - g - («,, ♦ ^Vi,

Upon substituting (3.13) and (3.15) into (3.5) to (3.7), we find that

(3 19) 6 + U + U +
- Sa" KL K,L L.K VKV

(3 20) 6 + + U + U
- ^KL " KL *KL L,K M,K*M1

(3,21) r + U
KLM " ^KL.M N.KV,M
 ■18-

So far all these expressions are exact. For a linear theory, one assumes

that the product terms are negligible so that

(3 22) C 6 + + U
- KL KL Vl L.K

(3 23)
- VL ' 6KL +
Vl + U
L.K

(3,24) r
KLM ^KL.M

In this case, the difference between the spatial and material representations

disappears so that one may use u in place of ü and $ in place of ■t' , etc.,

a fact which is well known in the classical continuum theory (cf. Eringen

[1962, Art. 14]). For the microdeformation, this may be seen as follows: If

we use (3.15) and (3.16) in (2.11), we obtain

(3.25)
W * - (6 + * U 4 6
' KL ^ KM KM mÄMmiX

Neglecting the product terms, we see that

(3 26)
- *KL • ^6KAL

which is the proof of our statement.

Since we will be dealing with the linear theory, we shall not distin-

guish between material and spatial representations except when it becomes

necessary for clarity in the development

The material (or lauranuian) strain tensot E and the materi al nicro-

strain tensors t,, and i", „ are defined in the 1 Lrear theory by
K.L KLM ' ^

(3 27)
- ^ ' 7(CKL " 6KL) " K,L +
\J

(J 2Ö) £ E + U
- KL V- V " \L L,K
(3 29) r
- KLM"JKL,M
 -19-

In the light of what has been shown above, we may also introduce

the spatial (or eulerian) strain tensor e, and spatial microstrain

tensors. e, „ and u „ in a similar fashion:

' k lc k S-m

(3 30)
- \t i<\tt**i9J
(3 31)
- «ki; ^k^^.k

(3 32)
- \.m " ' \t,m

Clearly, when these tensors are known, c'.ianges in arc length and angles

during any deformation of the body can be calculated.

For the linear theory, the difference between the squares of arc

length in the Reformed and undeformed body follows from (3.4) and using

(3.15) and (3.27) to (3.29)

(ds(a))2 - (ds(a))2 « 2^ + r^dx^

(3 33) +2(E
- KL+rLMKVdXKdHL

+ (E + E 2E
KL U - KL)d5KdEL

In classical continuum mechanics, only the first term on the right involv-

ing E survives.

From (3.33) it is clear that when E,,, , E... and r__u vanish, there
KL K.L KLM

will be no change in the arc length after a deformation. In such a situa-

tion, the body is said to undergo a rigid motion.

 -20-

4. MICRÜPOLAR STRAINS ANÜ ROTATIONS

We now consider a special class of materials in which the state

of the microdeformation can be described by a local rigid motion of the

microelements. A large class of materials exist in which the micro-

material elements are dumbbell type molecules. Materials consisting of

rigid fibers or elongated grains fall into this category For example,

wood, certain rocks, and minerals contain elongated molecular elements.

Among fluids, blood possesses dumbbell-like molecules. For such media,

the micromorphic material theory becomes much simpler. Mathematically,

this specialization is obtained by setting

(4.1)J
K * - - *
KL LK

or in the spatial notation $ " ~ *oi,* 1° three-dimensional space, every

skew-symmetric second-order tensor $ can be expressed by an axial vector

KM

* defined by

(4 2)
- *K"2eKLMV

where c is the alternating tensor defined as

K.LM

e E E
l23 ' 231 " 312 "' "E213 " "C132 " "e321 ': ^

c
„.w " 0 otherwise
KLM

Expression (4.2) is a compact expression of

*1 ' *32, *2 " *13, *3 ' *21

The solution of (4.2) for ^ is given by

KL

K(4.3) * ■ -££ •*
' KL KLM M
 •21-

Substitutlng this into (3.15), we sei that

(4 4) X 6
- kK " lcK " WM

In the classical theory we have the rotation tensor

(4 5)
- \L--\K ^VL-VK*
The axial vector R, corresponding to this is given by

\ ' 1 CKLMRML " I ^LlAl.L'

(A.6)

'ScL ' ~ eKLMRM

Using (3.27) and (4.6) , we find that

(4 7) U E + R
- K.L- KL KL-EKL- ^KLlAl

When this and (4.2) are substituted into (3.28) and (3.29)., we get

(A 8) £ E + £
' KL " KL KLM(RM " V

r
(4.9) KLM " "eKLM*N,M

When ^ - •M, wee see that E^ - E^ and r^ - 1^^ and the

microstrains are no longer independent of the classical strains and

rotations. Thus the micropolar theory assumes that the classical rota-

tion R^ is_ different from the microrotation. In the micropolar theory,

we have therefore six functions to determine, namely U (X,t) and 1 (X,t)

Once this is done, length and angle changes can be fully calculated.

For the micropolar theory, the spatial position of the a point

is obtained through (3.8), (3.18), and (4.3), i.e..

 -22-

(4.10) x(a) -X+E + u-E"*

From this it is apparent that $ represents an angular rotation of a

microelement about the center of mass of the deformed macrovolume element

(4 is the moment arm from this centroid), Fig. 4.1. Accordingly, we also

have

(4.11) § - E - E x $

which shows that aside from a rigid body translation the relative position

of E of a materiax point after deformation is obtained by translating S

parallel to itoelf to the center of mass x of the deformed macrovolume

element, and then rotating it an amount E * $. Since we are dealing with

linear theory, we also have

(4.12) E-4+^*<>

where i)) = «t» is the spatial microrotation. In fact, we have the complete dual

to equations (4.1) to (4.10) which we record here for future convenience.

(4 13)
- *k"2Wm* ' ♦ki'-W.

(4 14) X
- K. - \i ' Wm

(4,15) r
k " I £iamrnU ' r
ki " '^klufm ' 2 e
k)imUm,i

(4.16) u^ - e. . - e, . r
K,i ki kdm m

]f.i ki kimm m

(4 18) Y
- kilm " ^k^n^n^
 -23-

and

(4.19) x(a) -X+H+u-5x$

Now consider f.ie deformation of an Infinitesimal vector

(a)
dX - dX + d= at . -. Upon deformation, this vector becomes

(4.20) dx(a) - dx + dC - dX + dH + ^JOL, - dE x $ - H x tt.AL

By use of (4.7) and (4.6) we may write

(4.21) u.KdXK - UL>KdXKIL - ELKdXKIL ♦ ^d^^

E.,. dX I. - dX x R
KL 1C-L

Similarly using (4.9)

X
« ^NdXN - C
KLMHL*M.NdXN^ " ^KLnh^h

For convenience we now Introduce the notation

(4 22) S r
- W KLM\

so that

(4.23) HX *.NdxN.-r:KM)dxMiK-r[KM]dxMiK

where Indices In parenthesis (and brackets) Indicate the symmetric (and

antisymmetric) parts. Carrying (4.21) and (4.23) Into(4,20) we rearrange it into

(4.24) dx^ - dX + d= - (dX ■< R + dH x $ + dX x r)

+ (E
KL+ r
(KL))dVL

where we also defined a new microrotation vector T by

 -24-

(4,25) r E C
K 2 KLMrML ' r
[MK] " 'eKLMrM

We name this vector mlnirotatlon for distinction from the microrotation

#. If we carry (4.9) into (4.25K, we also find

(4 26) r
- K - 7 (-\.LEK +
VV
Equation (4.24) reveals that the deformation of the vector
(a)
dX ' - dX + d= may be achieved by the following three operations:

(a) A rigid translation of dX + dH from the material centroid

X to the spatial centroid x.

(b) Rigid rotations of dX and d= by the amounts dX » (R + F) and

x
dH *, respectively.

(c) Finally, a stretch represented by the strains E and T.

in (4.24).

The following special cases help to visualize these deformations:

(i) When H ^ 0 we have

(4.27) dx(a) - dx - dX - dX * R + E^dX^

This, of course, is a well-known theorem in classical continuum mechanics

attributed to Helmholtz (cf. Eringen [1962, Art. 10]).

(ii) When H ■ constant vector

(4.28) dx(a) - dX - dX « (R + [) + (1^ + T^^dXj^^

Here, ul course, we have no rotation of dr : 0.

(iii) # ■ constant microrotation. In this case we have

V)" V] " rK"0 and we obtain

 -25-

(4.29) dx(0t) - dX + d£ - dX x R + dH x * + EKLdXKl

In this case, the rotation consists only of a macrorotation of dX and a

microrotatlon of d=. We can also write it as

(4.3U) d^ - dx(ül) - dx = d: - dX * r - d5 ■ K + r .dX,.!.

(a)
in this form we see the difference between the deformation of dX and

that of dX of which the latter is known to us from the classical theory.

This difference, therefore, is the result of the composition of a mini-

rotation of dX, a macrorotation of di- , and the ministraining of dX

characterized by T. . The terminology of minirotation is being used

(.K.L.)
for [ and ministraining for T .. üf course, R is the classical rotation

for which we use the terminology macrorotation.

For the spatial representation the dual of (4.10) is

(4.31) X(a) -X+4-U+C**

From this, in the same way as in the case of (4.24), we obtain

(4.32) dX(Ql) = dx + d4 -•- (dx - r) + d£, * J + dx * y)

" {\i ' '(W^^i

where r is the spatial macrorotation vector defined by (4.15) and e, is

the spatial macrostrain tensor. The spatial minirotation vector Y -Li- given

by

(4 33)
' \ " 2 W«* "2 ^t.lV VV
and

(A 34) Y
' km 'WiJl ' ^in^n^l
 -26-

Finally, below we record the component form of various strains and

rotations in spatial rectangular coordinates x, y, z. For the components

of the displacement vector u and micropolar rotation vector ^ we write,

respectively, u, v, w and <Ji , $ , $ .

Macrostrain tensor:

3u 1 ,3u . 3v,
e = T— , e " — (— + r—)
xx dx xy 2 dy 3x

/, Q CK dv 1 .dv 9w,
(♦•35J e • -r- , e ' - (— + r-)
yy 3y yz 2 9z 3x

3w 1 /3w , 3u,
zz 9z * zx 2 3x 3z

Micropolar strain tensor:

3u 3v 3w
xx 3x yy 3y zz 3z

3v 3u ,
xy 3x z yx 3y z
(4,36)
3w
e =T <)> , c =—+^
yz 3y x zy 3z x

3u 3w .
e ' $ , c -r—+4)
zx 3z y xz 3x y
 -27-

Mlcropolar strain tensor of third order:

3$ 3$T 3*
x x X
yzx zyx 3x ' yzy zyy 3y yzz zyz 3z

3(J> 3$ 34i
(4.37) Y zxx " "Y xzx ■ "T^3x . Y zxy = "Y xzy ■ "T^
3y • Y zxz ■ "Y xzz " ~r^
3z

3d> 3(t 30
z z z
xyz yxx 3x * xyy yxy 3y ' xyz yxz 3z

all other Y, „ - 0,

Macrorotation vector;

!_ ,_3w 3v.
r
x " 2 ^3y " 3z;

(4-38) ^'Hu'W
1 ,3V 3uN
z 2 3x 3y

Microrotation vector:

(4.39) 4>-iM+(M
x- v* + ((ik
z~

Minirotation vector:

30 30 30 30
\'i^ lX--^y--^] +

30 30 30 30
i £ 1
(4.40) Y -7[(^ + -T U - -T ^ --S«.!
y 2 3x 3z y 3x x 3z z

30 30 30 30
YY - - u
f (—- + —^) E ~L -L
z 2 3x iy'^z 3x^x 3y>
 -28-

5. GEOMETRICAL MEANING OF MICROPOLAR STRAINS AND ROTATIONS

The geometrical significance of various strain and rotation measures

can be understood quickly if we consider the change dx in x as the

vector sum of three changes, namely,

(5.1) dx(0l) - dx + dy + dx

where

(5.2) dx - dX - dX * R + E^dX^

(5.3) dy - -dX x r+ r^dx^

(5.4) dz - dE - dH x *

Here dx is that known to us from classical continuum theory (cf. Eringen

[1962, Art. 6]). Accordingly, on the right-hand side of (5.2) the first

term dX represents translation of dx from X to x; the second is a rotation;

and the last term represents the straining of the body. More specifically,

consider the vector dX at the point X of the undeformed volume element dV.

This vector after deformation becomes dx.

Writing (5.2) in another form

(5.5) dx - C^

where

(5 6)
- ^E^r ^V^M
we see that a parallelepiped with side vectors I.dX , 1 dX„, and I dX

after deformation becomes a rectilinear parallelepiped with side vectors

 -29-

C.dX. , C„dX_ and C.dX_, Fig. 5.1. The stretch A.„, and extension £,.., are

defined by

dx - |dX|
(5.7) Iw 5A(W) -1 I
dX

Now calculate the extension of one of the sides of the undeformed parallele-

piped at X, e.g., l^dX :

|dx|2 = c1dx1 • c1dx1 - c11(dx1)2

Hence

E A 1 1
(l) " (l) " " ^-

But we have

C 1 + 2E
ll - 11

Thus

(5.8) E(1) - A(1) - 1 - /I + 2E11 - 1

From this it follows that

2E (1 + E
11 " (l))2 - 1

<<
For small extensions ^f-,^ 1, and this approximates to

(5.9) E11 » E(1)

an 0 t e
which provides a meaning for the normal components t^.,. ^OT* ^ ^33 ^ ^
E E we
infinitesimal strain tensor. For the shear strains E,?» 23» 3i ^inc^

a geometrical meaning by calculating the change of angle between two side

 -30-

vectors such as li^X. and I-dX-. The angle 6..-.. between these two vectors

after deformation is calculated by

^ldXl • ^dX2 C
12
C08 e
(12)- IcJdX^C idX2-^==

2E
12
/I + 2E11 /I + 2E22

The change of angle T,^. " T" ^riP^ between the original and the final

angle follows from this

2E,
12
(5.10) sin f
y
^ /I + 2E1] /I + 2E22

r so t at we
For small strains E << 1, E - << 1 and sin ?(-.j) ' n2^ ^ ^ave

the approximation

(5.11) r(l2) - 2E12

which provides the geometrical meaning for the shear strain E . Similar

results are, of course, valid for E „ and E^, •

For the rotation vector R we have the following geometrical inter-

pretation: Let N^ be a unit vector in the X., X„-plane at X. After defor-

mation, N- becomes a vector n^ at x. Bring n^. to X and obtain its projection

n_ on X., X^-plane, Fig. 5.2. Let the angle between N and n- be denoted by

6 . We can show that (cf. Eringen [1962, Art. 10]) the average * tan e>

over-all angles * that N_ can make with the X-axis is related to R_ by

(5.12) < tan ö,>

/H + E..)(l + E,-) - E2
11 22 l2
 -31-

an are t e
where E ., B«»« ^ '13 ^ infinitesimal strain components and R. it

the rotation component in the X.-direction, i.e.,

dU
1 3U2 l
3 12 2 ^Xj^ 3X2;

For small strains (5.12) can be approximated to give

(5.13) < e3> R3

which provides a geometrical meaning for R... Similar interpretations are,

of course, valid for R1 and R„. Accordingly, for small deformations and

rotations, R represents the average rotation angle of the diagonal vector

dX of the undeformed parallelepiped.

The above discussion concerning the deformed parallelpiped with

diagonal dx at x may be used to provide a geometrical meaning for the

deformed parallelepiped with diagonal dy at x. To this end, from (5.3)

we observe that r takes the place of the rotation R and I". , that of E .

Accordingly, the undeformed parallelep'ped with diagonal vector at X + E

with 5 fixed, undergoes an additional rotation I and length and angle

changes represented by the strains T, , at the spatial point x + ^ that

the point X + H occupies after deformation. This deformation takes place

with S being displaced parallel to itself. This deformation emanates pri-

marily from the parallelepiped with diagonal dX being constructed at the

point X + E with E fixed imtead of being at X, as in the first case.

Finally, a parallelepiped with diagonal dE constructed at X + E

with X fixed is rigidly displaced to a spatial point x + 4 and rotated with

the microrotation $ at that point. Thus the final rectilinear parallelepiped

with diagonal dx at x + C is constructed as a result of these translations,

rotations, and strains.

 -32-

The picture of the deformation in these latter cases is clarified

further if we write (cf. (3.1))

(5.14) dx(a) - (CK + CydX^ + xKdSK

where

(5 15) £
- ?K ?L.K=L

and \v is related to (^ by (3.15).

In the form (5.1A), the deformation of a parallelepiped with diagonal

dX at X + = and another one with d: are shown on Figure 5.3. According to

this picture, dX at X + | with fixed E becomes (C ♦ C )dX^, Fig 5.4, and

the one with diagonal dE at X + i with X fixed becomes x^dX^, Fig. J.5. The

resulting deformation of dX + dE at X + E is the vector sum of these two

deformations. Fig. 5.3.

 -33-

b. INVARIANTS OF STRAIN TENSORS

The state of local deformation at a point X + E of a micromorphic

material is fully determined when the three material strain tensors

{bA) L r
hh * KL ' KLM

are given. With knowledge of these tensors, we can calculate length

(cf. (3.30)) and angle changes (Art. 5) and construct the spatial locations

of various material points in the body. Instead of the list (6.1), we may

of course employ the spatial tensors £,„, E. and y, . An important

question in continuum mechanics is: If the material (spatial) coordinates

are rigidly rotated at X (at x), are there some functions of the material

(spatial) strain measures which remain unchanged? The answer to this ques-

tion is provided by the theory of invariants. In fact, the theory of invariants

is concerned with a more difficult question, namely: To determine the complete

set of invariants (called integritv basis) of a given set of vectors and

tensors which are unchanged under an arbitrary group of transformations of

coordinates. The minimal basis is a subset of these invariants which can

be employed to express all other invariants of a given set. The answer to

this question is important on two accounts:

(i) Constitutive equations rausL be form-invariant under rigid motions

of the spatial frame of reference. This is known as the principle of objec-

tivity . Investigation of this restriction on the constitutive equations often

requires knowledge of the invariants of constitutive variables such as strain

measures.

(ii) The Material symmetry places restrictions on the form of the con-

stitutive equations when the material frame of reference is transformed

according to some group of transformations. For example, when the material

 -34-

has a plane of symmetry, the constitutive equations should not change their

forms when a rpflection of axes is performed according to the plane of

symmetry. Similarly, for Isotropie materials, the constitutive equations

remain form-invariant under the full group of orthogonal transformations

of the material frame of reference. For anisotropic materials with some

symmetry axes, the group of transformations is less restrictive.

The invariants of a symmetric second-order tensor (such as E ) in

three dimensions are:

(6 2) E E II E (E
- ^ KK ' E I KKELL ' E
KLELK) ' lll
E * det E
KL

Instead of this set, one may also employ

trE EE
KL

tr
(6.3) ^EKLELK

3
tr E ::E
KLELMEMK

The above two sets are related to each other. In fact

2 2
tr E - 1. . tr E = I - 211 ,
(6.4)

trE3- I3- JIEI1E+3IIIE

The determination of the minimal integrity basis of (61) is much more

complicated. In fact, to our knowledge the basic invariants of third and

higher-order tensors have not been studied to date Fortunately, in the

present theory T always occur in the form T 5 = r so that we may

instead search for the invariants of

(6 5) r
' hi ' ^L ' KL
 -35-

For the micropolar theory, the situation Is simplified further

since E, , « E . The Invariants of two symmetric and two antisymmetric

second-order tensors

(6 6) E E E r r
' (KL) " KL ' [KL1 ' (KL) ' [KL]

would be sufficient for this purpose. The Integrity basis for the proper

orthogonal group for such a set has been studied by various authors. Below

we give a table for the construction of these Invariants- For the sake

of simplicity, we introduce the symbols a, b for the symmetric tensors and

u, v for the antisymmetric tensors. In Table 6.1 we give the invariants

of these tensors in ascending order of the integrity basis of various sub-

sets of a, b, u and v. The integrity basis of the quantities listed in each

entry of the first column includes all entries on this row and the integrity

basis of all subsets of these quantities. Thus, for example, the integrity

basis for a, b in the second row includes that of a and b, namely, tr a,

2 3 2 3
tr a , tr a , and tr b, tr b , tr b . Also an asterisk (*) placed on the

products indicates that we include in this list all other products obtained

from this by cyclic permutations of the symmetric matrices. A dagger (t)

indicates the inclusion of all quantities obtained by cyclic permutation of

the skew-symmetric matrices Thus, for example, ab* means the inclusion of

the set

ab , ba

2
Similarly, u vat means the inclusion of the set

2 2
u va+ , vu t+

For other details and more extensive studies on invariant theory see Spencer

[1966]. Since we will be interested, generally, in a linear theory, many

entries in this list will not be needed in the construction of the constitutive

equations.
 -36-

TABLE 6.1

Matrix products whose traces form the

integrity basis in the proper orthogonal group

Matrices Matrix Products

2 3
a: a : a
2* 2 2
a, b ab; ab ; a b

u
2 2 2 2 2
u, a u a; u a ; u aua
2* 2 2 2 * 2 2* 2
u, a, b uab; ua b ; ua b ; ua ba ; ua b a ; u ab;
2 2*2 22*
u a b ; u aub; u aub

y» Y uv
2 2 t 2 2+ 2 2T
u, v, a uya; uya ; u ya ; u ya ; u aya
2 * 2* 2 2 2 *
u, v, a, b uvab; uvba; uva b ; uvba ; uva b ; uva ba ;
2 ,t 2 . t 2, *t
u yab ; u ayb ; u bya
 -37-

7. VOLUME CHANGES

Here we calculate the change of volume with deformation. A volume

element dVL - dX dX^dX at X + H with fixed 5 will be called a material

macrovolume element and one dt/ ds.dE.d5. with fixed X will be called a

material minivolume element. After deformation, dV becomes dv and dU

becomes dv given

(7,1) dv = JdX dX dX

(7.2) dv - jd=1d52d53

whero J and j are the jacobians of deformation with 5 and X fixed respec-

tively, i.e. ,

(7 3) JEdet (
- \,K+ «kM.KV

(7.4) j det (x^)

To obtain the ratios of deformed volume elements to those of undeformed

ones, we need to calculate the jacobians J and j. Since the determinant of

the product of matrices is equal to the product of the determinants of matrices,

we have

J = Met [(x^ ♦ XkM>K.M)(xk^L + XkNfL^)]}1/2

1/2
!det (x X + + X X + }
k,K k,L ^.L^kM.K M k,K kN,L-N ^M.K^N.ITMV

Upon using (3,5) and (3.6) this becomes

(7.5) J - Uet (CKL ♦ :LMKEM + .^^ + r^r^'^H^) )112

 •38-

where

(7 6) ä
- S VkV

is the Inverse tensor to C . In the calculation of the last term on the

right-har.d side, we employed

X r
kL,M ' KLMXK,k

which follows from (3.7).

The above result (7.5) is valid for a general micromorphic material

and can be further simplified. We are, however, interested only in a

linear theory. In this case (7.5) can be approximated by

J. (det [\L^2EKL+2r(ia)]}1/2

Expansion and linearization of this gives

(7 7) J1+E
- KK+rKK

Thus the macrovolume change with fixed S, to a linear approximation, is

given by

(7.8) ~ - I - tr E + (V x ♦) ' 5
dv
o

where we used (4.9). Here the first term on the right-hand side is the

classical expression of the dilatation,and the second term is the addi-

tional volume change due to microdeformation.

Similarly, we can calculate the mimvolume element dv by determining

j given by (7.4). In this case

(7.9) J- [det (X,^)]1'"-

 -39-

For the linear theory from (3.15) we have

X = (6
kK MK+ V6KM

Hence (7.9) to a linear approximation is

(7.10) j ■ [det (6KL + *KL + *LK)]1/2 - ! +

^KK
+ 0
(*2)

For the micro^olar theory 4> « 0 so that we have

3^ - i . o<.2,

Hence, in^ the linear micropolar theory there will be no minivolume change.

A minivolume (for X fixed) rigidly rotates without altering its value.

 -40-

8. COMPATIBILITY CONDITIONS

The strain tensor e. and micropolar strain tensors e. . and y

are expressed in terms of the displacement field u. and microrotation

field 4» by

(u +
^ V - 2 k.i V^

(8 2) £ U + £
- k£ " £.k £km*m

(8 3) V
- kUi " £k*nVm

Of these,(8.1) is the eulerian linear strain tensor known to us from the

classical theory and (8.2) follows from (A.17) by use of (8.1) and (4.15)

and (8.3) is identical to (4.18). When the six quantities u and L are

prescribed, these strain fields are determined uniquely through (8.1) to

(8.3) by mere substitution. If instead the six strains e, ., nine micro-

strains c.g , and nine non-vanishing components of y are prescribed,

then the determination of the displacement and microrotation fields requires

the solution of twenty-four partial differential equations (8.1) to (8.3)

for the six unknowns u, and $ . Such a system is over-determined and

restrictions must be imposed on e. „, c. . , and y, . . These conditions are

kit' kü. kÄ.m

known as the compatibility conditions. For the classical strain tensor e. ,

the compatibility conditions are (cf. Sokolnikoff [1956]):

(8.4/ e +e -e -e ■ 0
k£,mn mn.kJ. kni,«,n An,km

If we notice that

(8 5) e = e (F + E
- k (U) 2 -k£ £k)
 -41-

then (8.4) can also be written as

(8.6) G/. „x +
C/ •, , „ - £/, \ „ - £/„ si ■ 0
(kll),mn (mn),k£ (kni),«,n («,n),km

An alternative derivation of (8.6) is instructive: To this end through

(8.2) we calculate the displacement

r field u, from
k

(8 7) U
- k ' \ + V^k + Wm)dx£
where u, is the value of u, at one end x' of an open curve C in the body.

Integrating by parts the second term under the integral, we also write

(8 8)
- \ " \ + W.^i - ^ + Veik + "^^ ~ ^u^]6xi
For the displacement field u, to be independent of the path C followed

between the points x and x , the integrand must be expressible as a

mm
total differential, i.e.,

£
ik- WX* - X>m,i = F
k,i

rfhere F. (x,t) are single-valued and possess continuous partial derivatives

with respect to x, through second order. From this it follows that

F = Ff
k,ij k,ji

Consequently

Skj " t'ttm^i - ^Vi'.J " £jk.i + [

WXÄ " x2Um,jj.i " ü

txpanding the second term and using (8.3) we obtain

ik,j jk,i ikj jki

 -42-

These are the necessary and sufficient conditions for the displacement field

u to be single-valued and continuous in a simply-connected region. Elimina-

ting Y from (Ö.9) by differentiating and using (8.2) and (8.3), we obtain

(H.6) again.

A similar method can be applied

rr to Y. „ as follows. First we solve for
k*m
♦ by multiplying both sides of (8.3) by t , Hence

(••10) ♦ r,m " "T e

2 bkit
n ^1ki.m
n

Integration of (8.10) along a smooth open curve C gives

= +
^r *r 2 {, ZkiryUmd\

The condition of single-valuedness for $ now reads

(8 U) e
- klr(vkt.,n ' >
Un.m) = 0

Thus we have proved the following.

Theorem: The necessary and sufficient conditions for the integrability

of the system (8.1) _to (8.3) for a^ simply-connected domain is the satisfaction

of the compatibility conditions (8.4) and (8.9) and (8.11)*.

We note that the terms outside the line integral in (8.8) represent

rigid body deformation.

*
Condition (8.9) was obtained by Sandru [1^66]. For the general nonlinear
theory of micromorphic materials, the proof was given by Bringen [1967).
 -43-

9. SOME SPECIAL DEFORMATIONS

In this section we present a lew special deformations for illustrative

purposes.

(i) RiRicl Deformation. The deformation of a body is called rigid if

the distance between every pair of points X and Y in the body remains

unchanged. From (3.33) it is clear that the necessary and sufficient condi-

tion for the ri^ld deformation of a linear nicromorphic body is

(9.1) E-l-O , [ - 0

Alternatively, in terms of the spatial measures of strains

(9.2) e=e=ü , y-O

For a micropolar body E * Ü implies

(9 3) U + B
- K " ^SoA K

where R is an arbitrary skew-symmetric tensor and B is an arbitrary vector,

both of which are independent of X. The condition f ■ 0 implies that •_ be

KL K
independent of X. Finally t = 0 gives

(9.4) *K - ^

where

(9 5)
- \ ' 1 ^LH^L

is a rotation vector independent of X.

(ii) Isochoric Deformations. The deformation will be called macroisochoric

if the material macrovolume remains unchanged. It will be called miniisochoric

if the minivolume is unchanged. The necessary and sufficient condition for

 -44-

macroisochorlc deformations according to (7.7) or (7.8) is

(9 6) E
- KK+rKK"0

The condition for miniisochoric deformation follows from (7.10). Since,

for the linear micropolar bodies b K.K - 0, we see that the ————
linear mlcro-

polar bodies undergo only miniisochoric deformation. We note that, in

general, this is not true for general micromorphic materials.

For the condition (9.6) to be valid for all ;, it is necessary and

sufficient that

(9.7)
1 E « 0 . r = 0
' K.K ' KK

(iii) Homogeneous Strain. The state of strain in a body will be called

homogeneous when the deformation is linear and homogeneous in the position

vectors X and H of the material points, i.e. ,

(9.8) x(a) - D^ + PKEK

where Ü and 0 are constant vectors. In terms of the components of these

vectors, (9.8) is equivalent to the system

(9 lü) =
- ^k 'WK

where I) and i), are constants for static deformations and functions of time

only for the dynamical motions and

i9 n
- > 9k "Wk • ?K"Wk
 -45-

Equation (9.9) is the expression of homogeneous strain in classical continuum

mechanics. The deformation described by this set carries straight lines into

straight lines, ellipses into ellipses, ellipsoids into ellipsoids. The

microhomo^eneous deformation (9,10) is new, and it possesses the same kind of

properties with respect to '=.. For the micropolar body, 0 can be replaced

by a single vector V given through the solution of

(9 12) P
- kL - 6kL " WM

namely

(9.13) P • -• e L>
^ ' M 2 kLM kL

Equation (9.12) results from equations (3.15) and (4.3) with * ■ P. Thus the

microhomogeneous deformation can also be expressed by

(9.14) | - H - E x p

where the vector P is independent of X and of course H.

The material deformation tensors follow from (3.5) to (3.7) and (9.9)

and (9.1Ü) with V - x^- Thus

C
KL " \KDkL
(9.15)
P
4'„. - D. „ ,. . r - 0
'KL " "kK kL ' 'KLM

It is now clear that the strain measures are homogeneous. For a micropolar

solid we use (9.12) so that

(9 16) y £
- KL - ^K " LMNVM

where we put
 -46-

Ü 6
LK - kLDkK

Below we give several special cases:

(iiia) Uniform Macrodilatation In this case, D is a diagonal matrix

having the same entries, i.e..

r
D 0 o'
(9.17)
0 D i) 0 < D < ^
0 0 D

The deformation tensors in this case take the forms

(9.18) C D 6 DP
KL " KL ' \L ' kL6kK ' r
KLM = 0

The deformation carries a parallelepiped having edge vectors f^dX^ at X + E

to one with edge vectors DdX.l , Fig. 9.1. The ratio of the length of an

edge to its original value (the macrostretch) L, , is therefore given by

L D
(K) =

The angle between any two edge vectors of the deformed macroelement is 90°.

The deformation carries a raacrocube of unit volume to a macrocube of volume

D . The microelement changes according to the values of V , From (9.16),

for this case, we have

(9.19) * ■ ü(6
V - e V )
KL KL KLM M'

From this it is cxear that

(9.20) 4'11 - t22 t « Ü f

33 lCL " "W^M ' ^"^

Thus, the microstretch ^/^x» the ratio of the edge vector of the deformed

microelement to that of the deformed element, is given by

 -47-

(9.21)
V) ■D
Hence, the edges of the microelement are stretched the same amount as those

of the macroelement. For L) ■ 1 we have no macrodeformation or mlcrodeforma-

tion. However, the microelement undergoes a rotation described by V, For

the general micromorphic materials, the situation is much more complicated

and it is possible to have microdeformations even when the macrovolume remains

unchanged. This situation is, of course, a familiar one in molecular theories

of crystal lattices.

(iiib) Liniaxial Strain. Consider the homogeneous strain characterized

by
— -
D 0 0
(9.22) D = 0 1 II 0 < D < ^

Q 0 1

In this case, for a micropolar solid through (9.15) and (9 16), we get

D2Ü 1) D -P,D P,tf

3 2
(9.23) 0 1 0
0,1) 1
3 "P.1
0 0 1
-P,D P. 1
2 i

Macrostretch L/KN and microstretch fc, . are

L ,= L s 1
L D
(l) " (2) (3)
(9.24)
ü
V) ' (2) (3) i

For P. ■ P_ ■ P_ ■ 0, w« have the classical uniaxial strain according to

which a bar of length dX. after deformation becomes a bar of length DdX ,

Fig. 9.2. For non-vanishing 9 we see that the microelement is stretr'.ed

i y the amount Ü in the X direction in addition to having a rotation. The

 -48-

geometry becomes particularly simple for the two-dimensional case for which

we have

(9.25) D
D 0 D 0 D -P3D]
. t
0 1 0 1
.D3D 1
J

A sketch of the deformation is shown on Fig. 9.2. The macroelement OACB

after deformation becomes OA'C'B elongated an amount (D-l)dX along the

X.-axis vith no change in the lateral directions. The microelement Cacb

is stretched by the amount DdH in X direction becoming Da'c'b' After-

wards, a rigid microrotation occurs about the X-axis at C, The final

shape of Cacb is marked by Ca"c"b". For the general micromorphic materials,

it is possible also to have microstretches independent of the macrodeforraa-

tions.

Generally, when a bar is stretched in one direction without any

constraints on its sides it will also change its lateral dimensions. This

situation is characterized more realistically by

D 0 0

(9.26) D = 0 DO

0 0 D,

In this case the state of strain is called simple extension. For simple

extension we have

D^ Ü 0 D1 -0^3 D^

(9.27) C 0 u2 Ü D
V3 2 -D2P1
o 0 u: D P
-V2 2 1
 -49-

Frora (9.27) 1 it is clear that a macroelement in the shape of a rectangular

parallelepiped after deformation becomes another rectangular parallelepiped

with its sides elongated proportionally to D., D_, and D_. A microelement

in the shape of a rectangular parallelepiped changes its sides in the same

proportions, however, it also rotates.

(iiic) Simple Shear. In classical continuum mechanics a homogeneous

strain characterized by

ISO
(9.28) 0 10 — 00 < S < '*>

0 0 1

is called a simple shear. Here S is independent of X^. The spatial position

(a)
x . of any material point X + = after deformation is given by

4^ = X
l + SX
2 + H
l +
^3 " V3-2
(a)
(9.29) X2 + .2 + V3h - P1S3

«<a) = X3 + .3 + P1E2 - V2h

In the case of 0 ■ 0, simple shear rotates X, = const planes rigidly about

their lines of intersection with X« ■ 0-plane, by an amount equal to the

angle of shear y given by

(9.30) arc tan S

The X- ■ const, and > ■ const, planes are unchanged (cf. Eringen [1962, Art. 15])

The deformation tensors C, f, and 1 are given by

 -50-

(9.31)
-p.
sV
S 1+S 0 (S+1)P3 1-SP 2~Ui

0 0 -p.

I - Q

From (9.31) we see that even when there is no macroshear, that is when S » 0,

we shall have a microrotation prescribed by the microdeformation tensor

1 -V V
3 2

P.
(9.32)

-P 0 1
2 1

The picture for a plane microdeformation, in this case, is similar to the

one described by Fig. 9.2.

(iv) Plane Strain. In classical continuum mechanics when the defor-

mation of a body is identical in a family of parallel planes and vanishes

in the directions of their common normal, we say that a state of plane strain

exists. This plane strain is thus characterized by

(9.33) 5^ - xk(X1,X2) , (k = 1,2) X X

3 ' 3

For the micropolar deformations, we define the state of plane-micropolar strain

similarly. Using (4.11) and setting «t. 5 * z 0, t_ ^ * we have

h'h' *(X1.X2)52

(9.34) C2 - H2 + *(X1,X2)53

,,
Si - i
 -51-

Thus, in a plane macrostrain and microstrain the deformation is fully

prescribed when the two unknown displacements

ül(xl,x2) -= «j - x2
(9.35)
U2(X1,X2) - x2 - X2

and the microdisplacement <i(X. ,X?) is determined. The field equations

should, therefore, consist of three partial differential equations

replacing the two equations of the classical theory. This, as we shall

see, is the case (cf. Arts. 25 and 26).

The spatial deformation teneors for the plane strain are

C C 0 t f 0
ll 12 11 12
(9.36) C21 C 0 * TT 0
U
22 21 22
0 0 1 0 0 1

r - -r ■ -• , (M - 1,2)
12M 21M 'M

Equations (3.27) to (3.29) and (A.3) provide the relations between defor-

mations, tensors, and strains (or displacement vectors):

C 6 + 2E 6 + U +
KL - KL KL - KL K.L "L.K

(9.37)
t «£ +6 -6 -e «t+U
KL KL KL KL KLM M L,K

KLM KLN N,M

Thus, for the strain tensors E and £ we have

KL K.L

E E 0 £ L 0
ll 12 ii n
(9.38) E21 E ü , £ E21 £ 0
22 22
0 0 0 0 0 0
 -52-

when

31^ 1 aUj^ 3U2 3U2

E
ll " "äx^ ' E
12 " E
21 " 2 ^Xj +
■33^" )
' E
22 " ^Xj

aUj^ 3u2
9 39 E f +
( - ) ii"^; • -i2--* —;

3U. 3U0
F - 4 + —^ FC . i-
21 * 3X2 ' 22 3X2
 -53-

10. MOTION, MICROMOTION, MATERIAL DERIVATIVE OF TENSORS

A material point at X + H at time t ■ 0 is carried to a spatial

point x + 4 at time t. The motion of this material point in a body

is described by one parameter family of transformations.

(10.1) x(a) - x(X,t) + ax.H.t)

where x(X,t) is the place occupied by the center of mass X of a macro-

volume element dV + dS at time t and C(X,5,t) is the relative position

of the point X + 5 at time t with respect to the center of mass. For

micromorphic bodies we have (3.9) or

(10.2) x(X,t) - X + u(X,t)

for x(X,t) and (2.9) for C(X,H,t). For micropolar bodies, (10.2) remains

valid but (2.9) is replaced by (A.11), i.e.,

(10.3) S(X(|»t) ■ 5 - a « *(X,t)

In these expressions U(X,t) and l(Xft) are respectively the macrodisplace-

ment vector and microdisplacement vector. The parameter t is real repre-

senting time.

According to the axiom of continuity and indestructibility of matter,

the inverse motions X(x,t) and H(x,C,t) are assumed to exist. Thus we

may also write

(10.4) X(a) - X(x,t) + =(x, ,t)

where g is given by (? 13) for micromorphic bodies. We also have

(10.5) ¥(x,t) - x - u(x,t)

 -54-

However, for micropolar bodies (2.13) is replaced by (4.12), i.e.,

(10.6) § + § x $(x,t)

For displacement vectors, we employ the same symbol u and ♦ both in material

and spatial descriptions. However, in the former case u and i* are assumed

to be functions of X and t, and in the latter case functions of x and t

since we may substitute X - X(x,t) for X in u(X,t) and $(X,t) to pass from

material description to spatial descriptions. Single-valued inverses X(x,t)

to x(X,t) and =(X,4,t) to § - ^(X,5,t) are assumed to exist at a neighborhood

of X at all times except possibly some singular points, lines, and surfaces

in the body. A sufficient condition for this is the continuity of partial

derivatives of these functions with respect to X and

(10.7) det (^ K) + 0 det ($k KW 0

in some neighborhood of X,. at all times. We assume that such is the case.

In the kinematics of continuous media, the time rates of vectors and

tensors associated with material points play an important role.

Üef. 1. The material derivative of any tensor is defined as the

partial derivative of that tensor with respect to time with the material

coordinates X and = held constait. The material derivative is indicated

either by placing a dot on the letters or by D/Dt • Thus, for example,

DF 3F (X,t) |

Df. 3f (x,t)
(10.8)

D
*k ^k
Dt ■i " *K(X't)=K
X.

where subscripts attached to a bar indicate that those variables are held
 -55-

constant during the differentiation. Since through the centroidal motion

we have

(10.9) xk - x^X.t) or X^- ^(x.t)

we have

fk(x,t) - fk(x(X,t),t)

so that

3xe(X,t)
f
k " at It
X I

In short, without ambiguity we write

(10.10) f, + f
Dt " dt k,lXi

The first term on the extreme right of this equation represents the time
x
rate of change that occurs at a place/Jat time t. The second group of

terms is known as the convective change. These arise from the motion of

the material point X through the place x.

In the case when the tensors involve the macromotion in the material

differentiation, we also consider the relative location vector z held

constant, cf. (10.8).. Another example is provided by differentiation of

(4.11).

(10.11)
^M(X,t)
x
-E <t> or -e
kLM-L >t
 •

-se-

il. VELOCITY, ACCELERATION, MICROROTATION, SPIN

Def. 1^ Velocity is the time rate of change of the position vector

of ji material point. Thus

• •
v
(11.1) v = x(X,t) or k " ^

where X is held constant. Since x ■ x(X,t) we have

ax(X,t)
x(X,t)
at

Note that in a body with microstructure, X is the position of the center

of mass of a macrovolume element and it may or may not be actually occupied

by a material point. Nevertheless, in defining the time rates of vectors

and tensors associated with the body, we refer to X as the material point.

Upon replacing X by (10.9)? we also write

• •
(11.2) v - x(X(x,t),t) - xk(x,t)ik - vk(x,t)ik

where i, are the spatial rectangular unit base vectors. This equation

defines the velocity field v at a spatial point x at time t. This is the

eulerian concept of the velocity field which is prominent in hydrodynamics.

For the lagrangian viewpoint, we express the velocity vector in the material

frame of reference X^. Thus

* ^(X^)
(11.3) Y.VK(X.t)IK , VK.—^p-6^

Def. 2^. Acceleration is the time rate of change of the velocity

vector of ji material point. Thus

Dr.
(11.4) a £ v or \ ' vk ° —
 -57-

For the lagrangian viewpoint we have

3V (X.t)
(11.S) a - AK(X,t)IK , AK £ —~

and for the eulerian viewpoint

Dv 3v
(11 6) a
- k -— —+\yi
m

Here we notice the appearance of the convective terms v, v

The velocity v and acceleration a defined above are the kinematical

quantities describing the motion of the center of mass X in a macromaterial

element V + S. We now proceed to obtain the relative velocity and accelera-

tion of a material point X + 5 with respect to the center of mass X. For

these, we take the time rates of the relative motion for a micropolar body

given by (2.9), namely,

(11.7) § - XK(X.t)5K

Thus

• •
i ' XK(X.t)5K
(11.8)

§ - XK(X,t)EK

Alternative expressions are obtained by replacing E by its expression

K

(U.9) ^-V^k

Thus, for example .

(11.10) i - vk(x,tKk or ir v^
 •

•58-

where

(11.11) vk(x.t) I XK(X.t)XKk(x.t) . v^ I x,KXKk

It is understood that X appearing in the argument of x is also replaced

by (lü.9)2.

De f. !_. The three vectors v defined by (11.11) are called the

gyration vectors. and their components v form the gyration tensor.

When the gyration tensor is given, we can calculate the eulerian

microvelocity § by (11.10). For the microacceleration in a similar fashion,

we obtain

I ■ Vk +
Vk ■ Vk +
Vk»ci

where we used (11.10)„. Thus

(11.12) { - ak(x,t)Ck or ^ - a^fo

where

(11.13) öL (x,t) i V. ♦ V V . or a., = v.. ■¥ v. \J .

~K ~ -k -m mk ik ik im mk

Def. ^ The three vectors a. (x,t) defined by (11.13) are called

the spin tensor.

The total velocity v and acceleration a of a material point

X + = can now be calculated by

(11.14) v(a) - x + £ - v + v, r
k k

(11.15) a(0l) - a + a - v + o^C

 -59-

For a micropolar body, these expressions are modified by use of (4.A) and

(4.14). Thus, for example,

(llll6)
\i - -WM+ ekKMC^mVm

For the linear theory this is simplified to

(11.17) v. „ = -e. „ i

Upon introducing an axial vector v. , called microgyration vector, by

(11.18) v, ■ T Ci „ v . , v, . ■ -e, . v
k 2 kim ml ' kl klm m

we see that

(11.19) vk - *k

and (11.10) now reads

x
(11.20) £ - -^ v

Similarly, one can calculate £. An alternative approach that may be instruc-

tive is through taking the time rate of (11.20)

• *
(11.21) C--^xy-£xu--§'<v-»-Uxv) xv

If we recall a vector identity

(11.22) (a ^ b) ' c - (a • c)b - (b • c)a

the above expression can be written as

(11.23) £; - -C x v + U • v)v - (v • vH
 ■60-

whose component form is

(11.24) Ck - ak^

where

(11.25) a. , = -e, . v + v v - v v 6

It can be seen that this is identic11 to (11.13)».

The total velocity and acceleration vectors of a material point

X + H in a micropolar body can now be expressed, respectively, by

(11.26) v(Ql) - x(a) -x-l-^-y-^xv

(11.27) a(a) - v(a) -a+'i-v-^xv+^xv) xv

•
Here v and v refer to the centroidal point of the macrovolume element,

and the remaining terms on the extreme rights of (11.26) and (11.27) are

the relative velocity and acceleration about the centroid. Equation

(11.27) can be linearized further by dropping the triple vector product

term on the extreme right.

 -61-

12. MATERIAL DERIVATIVE OF ARC LENGTH

In continuum mechanics, the time rates of arc length, elements of

surface, and volume in the deformed configuration are often required.

Here we prepare the groundwork for this, while at the same time introducing

certain new concepts essential to the study of motion.

Fundamental Lemma 1^. The material derivative of the displacement

gradient is given by

(u iJ
' ft ^y - "k.K - Vt*!,!

Th e proof of this is immediate since D/Dt and 3/3X^ can be exchanged, i.e,

Dx
D k

where we used x. ■ v. (x,t) and the chain rule of differentiation. Another

useful expression that follows from (12.1) by multiplying it by dX is

(12.2) dxk . vk>A

A corollary to Fundamental Lemma 1 is

(i2 3)
- ihv^ ■ mhti*i.k
which is proved by differentiating x X^ - *I,IJ' Thus

Now multiply this by )L ^. This gives (12.3).

 -62-

Theorem 1^ Tlie material derivative of the square of the arc length

2
ds Li given b^

(12.4) els2 - 2dkidxkdxi

where

(12 5 d
- ) k£Ev(M) 'l<\,i +
\^

Is called the deformation rate tensor.

2
To prove (12.5) we take the time rate of ds ;

,2 D
- - - (d^dx^ - 2 H k d^ - 2vk>£dxkdxi
DC

(v +V
k,£ £,k)d\dX£

Hence the proof.

In the material description, (12.4) can be written as

2
^ • \l\,K\,Ldhd\

If we use

ds2 C
- KLdXKdXL - (6
KL + 2E
KL)dXKdXL

then

(12.6) d.2 . ^d^ - 2EKLdXKdXL

By comparing this with the foregoing expression, and since d, and C and
KK KL

E are synunetric tensors, we find that

 -63-

(12 7) C
- KL-2EKL"Zdk.\.KXÄ.L

This is the material derivative of the lagranglan strain measures C and E.

2
When d ■ 0 we have D(ds )/Dt - 0. Conversely, when for arbitrary
2
dx, D(ds )/Dt - 0 we must have d ■ 0. Hence we have

Theorem 2^ (Kllllng). The necessary and sufficient condition for

the macromotlon x(X, t) to^ be^ rigid Is d - 0.

Note that macrorlgld motion does not Imply mlcrorlgld motion. As

we shall see helow, the microelements may undergo non-rlgld motions even

though macroelements may be moving rigidly.

Fundamental Lemma 2^. The material derivative of_ the mlcromotlon Is

given by

(12.8) x^CX.t) - vkix,K

This result follows from (11.ll). by multiplying It by x., and using (2.11) .

A corollary to (12,8) Is

which Is obtained by taking the material time rate of (2.11) and multiplying

the result by X .

Theorem 3^. The material derivative of the mlcrodlsplacement gradient

X is given
kK,L ^L

(12 lü) V
- Dt ^kK.^ " ^kK.L ' MX«,l +
\^mXZKXmfL

To prove this, we take the partial derivative of (12.8) with respect to X^

and exchange D/Dt and 9/3)L since this is permissible.

 -64-

A corollary to this theorem is

(12.11) if, - v. cU + v. 0 C0dx

km m kli,,iii £ m

which is proved by taking the material derivative of

(12 12)
- ^k - \K,Lhdh + X
kKdEK

and using (12.8), (12.10), and Ck - XMAP«

Theorem 4^. The material derivative of the square of the arc length
a 2
tA
(as ( )^) •
is^ •
given k-l
by

^[(ds(a))2] - I*M ♦ *%k ♦ (V^ *V)k)Cr]dxkdxi

(12.13)
+ 2 + +
<^,k \i \r,k^r)dxkd^ + (
\£ +
^k)dVCl

To prove thi'j, we take the material derivative of

(12.14) (d8(0l)) - dx dx
k k
+ 2dx d
k ^
+ d
^kd4k

and use (12.2) and (12.11).

If we now introduce the microdeformation rates

(12 15) £ + V
- \i \i £,k

/n WN - (a) - f(
a
)
(12 16) a V a
- k£m ' U,m ' k£ - V,»V

1
Eringen [1964c]
 -65-

equatlon (12.13) can be expressed in the form

£ ((ds(a)) ] - 2[dk£ ♦ a^^Jd^dx^ ♦ 2(bu ♦ alk)dxkd^

(12.17)

In a region of a micromorphic body when

(12.18) d - 0 , b - 0

we have the general solution

(12 19) V + m
- k " \l\ \ ' \i \i

where UL . is an angular velocity and b is a velocity, both of which are

independent of x, and

(12.20) ^ «• -^ - •

Upon substituting (12.19) into (12.16), we see that a, ■ 0. Conversely,

2
(a)
we can show that the vanishing d and b is also necessary for D[(ds ) ]/Dt"0.

Hence we have proved

Theorem 5^. The necessary and sufficient conditions for ji micromor-

phic body to undergo microrigid motion are

(12.21) b - d - 0

This theorem replaces the well-known Killing's theorem (Theorem 2 above) for

micromorphic bodies.

For linear micropolar bodies, considerable simplification is achieved

in the foregoing results. To this end we recall (4.4) and (11.18)., namely,

v
(12.22) X, „ " ^ i/ - Ci „w*w . i 0 " -e, „ v
kK kK kKM M kS, kHm m
 -66-

Substitution of these into (12.8) and (12.10) and linearization gives,

respectively,

(12.23) xkK -e^^ , XkK>L . -Cj^V^^^

Hie expression of (12.11) for a linear micropolar body is

(12.24) H; ■ -e, , d[,snv - c, „ v 4„dx

at
-li klm «, ■ klm m,r i r

or in vector notation

(12.25) dt - -dC * v - C * v.
- r dx r

The material derivative of the square of the arc length for this case is

(12.26) A [(ds(a))2] - 2[dk£ ♦ a(k0]dxkdx{ ♦ 2^ ♦ «^^C,

where

(12.27) b, „ vn , - f, „ v , aln(0l) 1 -e, v f

ke e,k k£m m ' ki krm in,! r
 -67-

13. RATES OF STRAIN MEASURES

Def. 1. The i:ime rates of various strain measures are thü same

their material derivat ives. Th1US , for example,

DS •
_ KL
E
KL Dt » Dt

.D£k.
'•" EKL Dt > \i Dt

• Dr • D
. KLM _ \£m
r >
KLM Dt Dt

We now proceed to give explicit expressions for these quantities.

Theorem _1. The lagrangian strain rates are given by

(13 2) E
- KL " -k*S,K*t,l

(13 :3) £
- KL- \i\,KXlL

(13.4) r b
KLM " kJlXk,KX£L,M +
^Hm^.-^ilL^.M

The proof of (13.2) has already been given in Art. 12. To prove (13.3),

we calculate the material derivative of

V(13.5) E ■ iV -6 •x.y -6
' KL KL KL nc,KXkL KL

Hence

• •
E X x + X X
KL " ^KL ' k,K kL k,K kL

Upon using (12.1) and (12.8) we obtain (13.3). The proof of (13.4) is

constructed similarly by taking the time rate of

 -68-

(13 6) r
- KLM ^ "k.K^L.M

and using (12.1) and (12,10).

Theorem 2_. The eulerlan strain rates are given by

(13.7) e, . = d. „ - (e. v . + e .v )
ki ki. kmm.i mlrn.k

(13.8) e. „ - b, „ - (e. v „ + e „v . )
kl kl km ml ml m,k

(13 9) + e
- \im ' -\lm krar£m " (Y
k£rVr.m +
\rm\l + Y
r£mVr.k)

The proofs of these are somewhat lengthy and will not be given here. They

are obtained by differentiating the expressions for the strains and using

various results obtained in Art. 12, For the proof of (13.7) see Eringen

[1962, Art. 22], and for (13.8) and (13.9) see Eringen [1967].

Equations (13.2) to (13.4) and (13.7) to (13.9) indicate that the

eulerian strain rates are not the same as the deformation rates. If at

time t the medium is unstrained and the motion is just beginning, we can

set x * X and 5 ■ H so that

E
KL(X'0) " d
k£6kK6a

E
(13.10) KL(X'0) * b
k*6kK6*L

r
KI.M(X'0) = a
k£m-kK6£LÖmM

and

e
k>0) ■ d
k£

(13.11) eu(x.O) - bki

>. . (x,0) ■ -a, .

k£m K£m
 -69-

For the infinitesimal deformation theory, the terms enclosed in parentheses

on the right-hand side of (13.7) to (13.9) can be neglected. Therefore,

in this case (13.11) should be valid approximately for all times, i.e..

•ktW ' \i
(13.12) ^.t) - bkt

VktM^'V ' -ak£m

 -70-

14. EXTERNAL AND INTERNAL LOADS

A material body subject to the external and internal forces undergoes

a deformation. These forces may be of mechanical, electrical, chemical,

and other origin. Here we are only concerned with the mechanical forces.

In the particle mechanics of Newton, the force F acting on a particle is

considered to be a function of the position of the particle x, its velocity

v, and time t, i.e.,

(14.1) F - F(x,v,t)

When we have a collection of particles, then for each particle we may write

(14.2) Fa - Fa(xa,va,t) , (a - 1,2,..,)

In a volume element AV of a continuum, we have a large number of particles

interconnected with such forces. If the particles of a continuum are not

free to move independently, then the interparticle forces are balanced among

themselves in pairs. This then places restrictions on (14.2) so that the

number of independent forces is much smaller than the free collection of

particles. In conformity with these restrictions and with the basic postu-

lates of continuum mechanics, the forces acting on a body are resolved into

a resultant force F and a resultant couple M giv^n by

(14.3) F ■ I*• f-a ', M - y^ x-a xF-a

a a

The first of these equations gives the vector sum of all forces acting on

each material point with position x , and the second gives the vector sum

of the moments of these forces about a point which constitutes the origin

of x . In a continuum, the force field is usually considered to be continuous

and (14.3) may be replaced by

 -71-

(14.4) F - / dF , M - / (x » dF + dM)
V 1/
where dF is the force density at a point x and dM is a couple density.

This latter terms arises in classical mechanics from coupling various

particles in the form of doublets or rigid blocks so that some of the

applied forces on particles produce a couple also. This physical pic-

ture can be used in the construction of the theories of micromechanics

From a continuum viewpoint, whatever the origin may be, the forces

and couples may be divided into three categories.

a. Extrinsic Body Loads. These are the forces and couples that

arise from the external effects. They act on the mass points of the

body. They appear in the form of body forces and body couples per unit

mass of the body. The force of gravity is an example of a body force,

and an electromagnetic moment in a polarized medium is an example of a

body couple. A body couple can also arise from the uneven distribution

of the mass among microvolume elements. Fig. 14,1 and Fig. 14,2.

b. Extrinsic Surface Loads (Contact Loads) . These loads arise from

the action of one body on another through the contacting surfaces. At a

small macrosurface they are equipollent to a force and a couple. Thus,

for example, forces acting on a macrosurface of Fig. 14.3 are equipollent

to a force and a couple placed at the centroid of the marosurface element

Aa, Fig. 14,4.

When a macrovolume Av is allowed to tend to dv, in general the body

couple L vanishes since the moment arms cf forces tend to zero while the

forces are assumed to remain bounded. Similarly, when Aa is allowed to

tend to da, the surface couple M will approach to zero. This is the classical

picture in continuum mechanics

 •72-

Because of the granular nature of the bodies, the mathematical limits

dv and da for the surface and volume elements are approximations which may

not be admissible for various physical phenomena in which the applied loads

produce effects with some typical lengths (e.g., wave len^-h) that are

comparable to distances and sizes of the microelements. In such situations,

Av and Aa are not infinitesimals, and the granular nature of the bodies must

be taken into account in some form. This then requires that we consider the

existence of both forces and couplesfor macrovolume and surface elements.

c. Internal Loads. Internal loads arise from the mutual action of

pairs of particles that are located inside the body. According to Newton's

third law, the interparticle forces cancel each other so that the resultant

force is zero.

In continuum mechanics, the internal effects are found by isolating .1

small macroelement from the body and considering the effect of the rest of

the body as forces and couples on the surface of the macroelement as illus-

trated in Figs. 14.1 to 14.4. Internal forces give rise to the stress and

couple stress hypotheses,as we shall see below.

Let the surface force per unit area at a point x on the surface of a

body having exterior normal n be denoted by t, ,, and the surface couple

per unit area by m, ,. Let the body force and body couple per unit mass at

an interior point of the body be respectively represented by f and I. The

total force F and the total couple M about a point Ü acting on the body

can be calculated by, Fig. 14.5

(14.5) F - fet, .da + / pfdv

(14.6) M - ic[m, , + x » t, Jda + /p(l + x « f)dv

'o (n) - (n) ^ -

Concentrated loads are imagined as resulting from a limiting process in which

the surface loads or body loads are distributed over a very small region.
 -73-

15. MECHANICAL BALANCE LAWS

The mechanical balance laws - conservation of mass, balance of

momentum, and balance of moment of momentum - are obtained by a process

of averaging applied to a macrovolume element containing N microelements

for which the classical balance laws are postulated to be valid. Each

of the microelements is considered to have a uniform mass density,

I. Principle of Conservation of Mass. The total mass of each micro-

element remains constant during any deformation. Thus, ii and p

respectively denote the mass densities of a microelement a before and after

deformation, and AV and Av their volume. Fig. 15.1, then

(15.1) pJa)AvJa) - p(a)Av(a) , (a not summed)

The total mass of a macrovolume before and after deformations is respectively

given by

(15.2) p0AV0 . I p0 AVo

a-1

(15.3) pA i I p(a)AV(a)

These equations in effect define the mass densities PQ(X) and p(x,t) of the

undeformed and deformed macrovolume elements. In view of (15.1) we see that

(15.4) p AV - oAv

If we let AV and AV approach their limiting values dV and dV, then

PodVo - pdv
(15.5)
or
P
0 dv
-- — - J - det (x^)
 -74-

which are the equivalent expressions of the principle of conservation of

mass for the macrovolume element.

In Section 2 we said that X is the position vector of the center of

mass of a macroelement. Accordingly,

jp(.)?(a)A?(a) = 0
a

Upon using (15.1) and (2.14) this gives

(aMaK (a) - 0
Xjp^'Av^

Since X ^ 0, this shows that the position vector x is the center of mass

of the deformed macrovolume. Consequently,

Theorem 1_. The motion carries the center of mass of the undeformed

macrovolume to the center of mass of the deformed macrovolume .

Nect, we calculate the second moments

(15..) p^v . I^s^'X"

upon substituting (15.1) and (2.13), this may be written as (cf. Eringen

[1964c])

(15 7)
- hi - WWt

where

M« in ^
D A,. - ? (oi)r£(a).(a)AAv (a)
(15.8) Vi " 4p ^
a

Quantities I and i, are respectively called the material and spatial

microinertia tensors. Equations (15.7) may be stated as

Eringen [19640]
 -75-

Theorem 2. The mlcroinertla Is conserved, i.e.,

(15 9) 0
- £ ^vw -
Using (12.9), this may also be expressed as (Bringen [1964c])

- + i, „ v - i, v„ - i .v, =0
3t V.l,m m km Im ml km

In a micropolar continuum, a combination of i and I occur more

frequently. These are

J 1
KL " MM6KL " Ha
(15.10)
: i
^k«. mm<5ka i
kll

These tensors are identical to the inertia tensor encountered in rigid-

body dynamics.

Upon linearization and using (4.14), (15.7), and (15.11), we get

(15 11) J
- KL ' Jk*6Kk6L*

Global balance equations for mass and mlcroinertla are obtained

by integrating (15.4) and (15.7) over the volume of the body. Thus

(15.12) jo dV - /pdv
V 0 ü (,

(15 13)
- /VKL^O " /pikAkXudv

where V is the undeformed volume and 1/ is the deformed material volume.

11. Principle of Balance of Momentum. The time rate of change of momentum

is equal to the sum of all forces acting on £ body.

 -76-

The mechanical momentum of a microelement Av is the product of

its mass with the velocity, namely, p v Av . The total momentum of

a macroelement is the vector sum of the micromomenta of its microelements.

For a micropolar body we have

Ap - ^(a)y(a)Av(a) - Ip(a)(y+ OAv (a)

a a

J (a)A (a) r (a)rA (a)

« YIQ 'Av + v x £p ^Av
a a

The last term vanishes and in the limit we write

dp ■ pvdv

The total momentum of the body is therefore given by

(15.14) p ■ /pvdv
1/ '

The principle of balance of momentum is expressed by

(15.15) - pdv (n)

• da + /pfdv
1/ "

Here t/ .. is the surface traction per unit area acting on the surface of

the body S with an outward directed normal n so that the surface integral

is the vector sum of all forces acting on S. The vector sum of the body

forces is given by the volume integral on the right. Equation (15.15) is

none other than that given in classical continuum mechanics.

III. Balance of Moment of Momentum. The time rate of change of moment

of momentum about a^ point is equal to the sum of all couples and the moment

of all forces about that point.

 -77-

The mechanical moment of momentum of a microelement is defined as

the moment of its momentum, namely,

x(a) « p(a)v(a)Av(a)

The total moment of momentum of a macroelement is calculated by

r (a) (a)
O (a)
(a)Av(a)
dM - lfy x pv yv yv 'Av
a

- !(«♦$) x p<a)(v+ ÖAv(a)

a
Upon carrying out the multiplication, we get

Am- x * v^(a)Av(a) ♦ [« x p(a)Uv(Ql)

a a

x C (a)' (a) r (a)rA (a)

+ » 2,P §Av - y " 2,P CAv
a a

The last two summations vanish since 5 is measured from the center of mass

of the deformed macroelement. Upon carrying the expression ^ from (11.20),

in the limit we may write

(15.16) dm - px * vdv + podv

where

(15.17) poAv i lo(a\ * (v x OAv(a)

a

is called the intrinsic spin. In component form this reads

(15.18) vJkJlvk

where we used (15.8) and (15.11). after expanding the triple product.

The total moment of momentum of a macroelement, therefore, is the

 -78-

vector sum of its angular momentum and the intrinsic spin. The total moment

of momentum of a micropolar body is now calculated by

(15.19) M - /(x x pv + po)dv

1/

The principle of moment of momentum is expressed by

— /(x x py + pg)dv - ^(x x t, ,. + m, ,)da

(15.20)

+ /p(l + x x f)dv
1/

The right-hand side gives the sum of all moments about the origin as in (14.6)
 -79-

16. STRESS AND COUPLE STRESS

The state of Internal loads and their connections to surface loads

may be found by applying the principles of global balance of momenta to

small regions fully and partially contained in the body, Fig. 16.1. To this

end, we first consider a small macrovolume, v + s, fully contained in the

body. At a point x of s, the effect of the remainder of the body is equi-

pollent to a surface force per unit area, t. ,, called the stress vector,

and a couple per unit area, m, ,, called the couple stress vector These

loads depend on the position x, time t, and the orientation of the surface

s at x which is described by the exterior normal n to s at x. This latter

dependence can be found explicitly by applying the mechanical principles of

momenta to a region v + s adjacent to the surface of the body This approach

does in fact also provide the connection of surface loads to the internal

loads. Consider a small tetrahedron with three faces taken as the coordinate

surfaces and the fourth face being a part of the surface of the body. Fig.

16.2. We denote the stress vectors on any coordinate surface x, ■ const, by

-t. and on s by t, ,. The equation of balance of momentum (15.15) can be

-k -(n)
applied on this tetrahedron. Using the mean-value theorem to estimate the

volume and surface integrals, we write

d * * * * *
— (pv Av) ■ t- >. Aa - t. Aa, + pf Av
at - • (n) ~ K k.

where the quantities marked with asterisks are the values of those without

asterisks at some points of v + s The volume is denoted by Av and the sur-

face areas by Aa, and Aa. The mass is conserved so that

(16.1) — (pdv) - 0
 -80-

Upon dividing both sides of the foregoing equation by Av and letting Aa

and Av approach zero, we see that Av/Aa ♦ 0 and we obtain

(16.2) t, ,da - t, da.

(n) ~k k

The four surfaces of the tetrahedron form a closed surface, therefore, the

limit of the sum of area vectors da, must add up to da- Hence

(16.3) da - nda - da^

From this we get

(16.4) da, - n. da

Substituting this into (16.2) we get

(16 5)
- Un) ' *k\

where t, is independent
r of n. Thus we found that the stress vector t . , is
-k (n)
a linear function of n. At two sides of a surface, n changes sign. From

(16.5) we therefore see that

(16.6)
^-n) - "?(«)

which proves that the stress vectors on oppos ite sides of the same surface at

,a given point are equal in magnitude and opposite in sign.

The application of the above method, with the use of the equation of

balance of moment of momentum (15.20) leads to

(16.7) ■
(n) ■ Vk

(16.8) m, , - -m, ,
~ (.-n) (n)
 -81-

The concepts of stress tensor t, p and couple stress tensor m. now follow

from the decompositions

(16 9)
- h - \zh

(16.10) ^ - ^1^

Thus t is the I component of the stress vector t, which acts on the

K. It "■ K

surface x, = const, and m, is the I component of the couple stress vector

which acts on the same surface. The positive directions of t and those

of m. t are shown on Figs. 16,3 and 16.4 respectively. We use double-headed

arrows for UL .

From (16.5), (16.7), (16.9), and (16.10) it follows that

(16 11)
- l(.) - \i\h

(16.12) m(9) - m^n^

It is thus clear that the moment vectors for the couple stress have

the identical sign convention to those of the stress vectors. The plane of

each couple is of course perpendicular to the couple vector, and the direc-

tion is as described bv the right-hand screw rule.

The expanded form of the components of t, . and m, . in rectangular

coordinates are

t, .-tn+tn+t n
(n)x xx x yx y zx z

(16.13) t, t«tft+tn+ta
(n)y xy x yy y zy i

t,, "t n +t n +t n
(n)z xz x yz y zz z
 -82-

m,, -mn+mn+mn
(n;x xx x yx y zx z

(16.14) m,, "mn+mn+mn

(n)y xy x yy y zy z

m
(n) »mn+m n+mn
z xz x yz y zz z
 -83-

17. LOCAL BALANCE LAWS

Local balance laws are obtained by postulating that the global

balance laws are valid for every part of the body. For the conservation

of mass, we convert the volume integral over \l to V. T>us

(17.1) /(p - pJ)dv - 0

V

where

(17.2) J E det
^y
is the jacobian of the transformation. Postulating that (17.1) is valid

for every part of the body, we obtain the equation of local mass conser-

vation.

(17.3) po/p - J

Another form often used in hydrodynamics is obtained from this by taking

the material derivative of (17.3). Thus

pj + pj - 0

and we can show that (Eringen [1962, Art. 19])

J Jv 0
" k,k '

Consequently

p ♦ Pvkjk = 0

or since

p
n+ P'kVk
 -84-

thls reads

(17.4) lt+(pV'lc

This is the celebrated equation of continuity. One may, of course, equally

employ (17.3) in place of (17.5). Alternative forms which follows from (15.4)

are

(17.5) p0dV0 - pdv . ~ 0

An expanded form of (17.4) in rectangular coordinates is

3(pv ) 3(pv ) 3(pv )

(17.6) i£. + _JL.;
3t 3x 9y 3z

where (v , v , v ) are the rectangular components of the velocity field.

Equations of local balance for the microinertia are already given by

(15.7) and (15.9). For the linear theory using (15.11) we get

Dj
ia
(17.7) -~ - 0

The local balance of momenta follows from (15.15) and (15.20). Upon

carrying out the indicated differentiation and using (17.5)., and writing

a = v, v ■ x,we see that

(17.8) /padv - it, .da + /pfdv

V~ S(n) V

(17.9) /(x x pa + po)dv - ^(x * t, , + m. .)da + /p(t + x >< f)dv

V - ^ - (n) ~(n) '

These are other forms of the global balance of momenta. We now take V + S
from
to be a small internal portion v + s of the body. Substituting^ (16.5) and

(16.7) we write
 -85-

(17.1Ü) /padv - ^t n, da + /pfdv

v s v

x
(17.11) /p(x x a f ö)dv - ^(x t, + m,)!^da + /pU + x < f)dv
v s v

In rectangular coordinates, the Green-Gauss theorem is expressed as

(i7.i2) h^\da • hk k
dv
s v '

If we now apply this theorem to (17.10) and (17.11) to convert the surface

integrals to volume integrals, we obtain

(17.13) f[t. k + o(f - a)]dv - Ü

v '

(17.14) /[m, , + i, x t, + p(£ - ö)]dv

V

+ /x x [tk k + p(f - a)]dv ■ Ü

For these equations to be valid for any arbitrary volume v in the body,

the necessary and sufficient condition is the vanishing of the integrands,

Hence

+ f ( v) = 0
(17.15) ^k k - "

+
(17.16) 'J'k k ^ * tjj "•" PU - C) " 0

Note that the second integrand in (17.14) vanishes by virtue of (1,7.15).

These equations are the expressions of the local balance of momenta. They

are identical to those given in Bringen (1962, eqs. (32.7) and (32.8)] with

the exception of the spin inertia term po. This term arises from the postulate
 -86-

of an indenendcit microrotntlnn. In fact, without such an internal degree

of freedom, the existence of m, and I is-questionable. Upon substituting

(lb.9) and (16.10) into (17.15) and (17.1b), we obtain the component form

of these equations, namely.

(17 17)
- ^k.i + ^k-V •0

(17.18) m0, „ + c. t + p(l. - o, ) = U

lktl kmn mn k k

These are the fi rst and second laws of motion of Cauchy which express the

local balance of momenta for micropolar bodies . When the body is nonpolar,

that is, when o ■ ■. " i ■ 0, then (17.18) gives the classical result

which expresses the symmetry of the stress tensor. For micropolar bodies,

we see that the stress is in general nonsymmetrical and the new set of differ-

ential equations (17.18) must be employed replacing (17.19).

In rectangular coordinates, the expanded expressions of (17.17) and

(17.18) are recorded below

3t at 3t
—^- f —^ + —^ + p(f - v ) = 0
3x 3y z xx

it 3t 9t
(17.20) -r—t- + —XZ + —tt + P(f - v > - 0
dx dy 3z y

3t It JI
xz yz z^
(f - v ) = ii
3x y JZ z Z

If we disregard the relation of o. to given by (15.18), these balance

laws are valid for the nonlinear theory, and (17.17) and (17. .8) are exact
expressions, cf. Eringen and Suhuoi [lybAa & b].
 -87-

dm dm 9m
^* + Ji + _**+ t t + p(«. - 0 )
dx oy dz yz zy xx

dm dm dm
(17 21) —SL+ —n.+ _2X + t t + P(£y - Oy)
U/.^i; 3x 3y 3z zx xz

3m 3m 3m
t + p(£ - o )
3x 3y 3z xy yx z z
 -SS-

IS. CONSERVATION OF ENERGY

An Important axiom of thermomechanics is the principle of conservation

of energy which may be stated as follows:

IV. Principle of Conservation of Energy. The time rate of change of the

sum of the kinetic energy and the internal energy is equal to the sum of the

mechanical energy, heat energy, and other energies. Here we exclude chemical

and electrical energies so that we may express this law mathematically as

•
(18.1) K + E - W T (2

Here K, E, W and Q are, respectively, the kinetic energy, the internal energy,

the work of applied loads per unit time, and the heat energy. For a micro-

polar continuum, these quantities may be expressed as:

K +
(is.2) "7^Vk WkVdv

(IS.3) E ■ /pedv
1/

(IS.4) W - |(t^vk ♦ .lkvk)d.| r.

Jp{£
+ kVk + Vk)dv

(IS. 5) £ - ^q, da + /phdv

S V

The physical meaning of some of the terms occurring in these equations is

known to us in the classical continuum. For example, the first term in the

integrand of (IS..'.) is the kinetic energy of the macromotion. The second

term is, however, new and it is the kinetic energy of the microrotation. In

(IS.3), c is the internal energy density per unit mass. In (18.4), the surface

integral is the work of surface tractions and surface couples per unit time,

while the volume integral is the work of the body force and body couple, per

unit time.
 ■89

Finally, in (18.5) the surface integral gives the heat input, and the volume

integral the heat source.

The equation of local energy balance is obtained by postulating that

(18.1) is valid for any arbitrary volume contained in the body. To this

end we first carry out the indicated differentiation with respect to time

for K and £. Thus

(18.6) K ' /o(akvk + Vk)dv

(18.7) £ - /pcdv
V

where, in anticipation of the local laws, we employed the equations of

conservations of local mass and inertia (17.5)-, (17.7), namely,

(18.8) -Ü- - 0 , -rr1 - 0

Ddv * Dt

Next we convert the surface integrals of (18.4) and (18.5) into volume

integrals by use of the Green-Gauss theorem. Hence

W
■ /(tUcvM +
■ikvk.t)dv +
^(ttM + pf )v
k k + (
-ik,l + Dr )v ]dv
:. k

0. ■ !<%K Kk ♦ Ph)dv
1/ »

Lpon carrying (18.6), (18.7), and the above equations into (18.1), and using

the equations of local balance of momenta (17.17) and (17.18), we obtain

/(pe - t., v, „ + e. t_v. - m.. v „ - q, , - Dh)dv - 0

i, Ik k,i kmn mn k ilk k,£. ^k,k

This is assumed to be valid for every part of the body. Thus we muat have
 -90-

(18.9) pe - t£kvk(£ - e^t^ + m£kvk(ll ♦ qk>k ♦ ph

This is the differential equation of the local balance ai. energy of a micro-

polar body . In expanded form it reads

3v 3v 3v
pc - t xx -5»+t
3x
-*+ t zx -^
yx 3y dz

3v 3v 3v
+1 —^- + t —2- + t —^■
xy 3x yy 3y zy 3z

3v 3v 3v
z z z
+1 -r^ + t — + t —7
xz 3x yz 3y ZZ c»Z

-(t - t )v - (t - t _)V - (t - t )v
yz zy x zx xz y xy yx z
(18.lü) ^ 3v
- x . x
+m —r + m -n~ + m ——
xx 3x yx 3v zx dz

3v 3v 3v
+m —r^- + m —^- + m —■
xy 3x yy 3y zy 3z

3v 3v 3v
■ z z
+m —r- + m —r- + m —:—
xz 3x yz 3y zz dz

3q aq aq
+-^ + —-^ + —2- + ph
3x dy 3z

We also note that

3e 3e 3e or.
,(18.11)
LO,x L
'- -' £;""Tr
3t
+ —
"^3x v
x -t"—
3y v y +T—
3z v z

Again the energy balance equation (18.9) is exact in this form and is
valid for the nonlinear theory-
 -91-

19. PRINCIPLE OF ENTROPY

For certain classes of physical phenomena within a range of expected

changes, a material body is characterized by certain constitutive equations.

These equations define an ideal material approximating the real material

under consideration. For any thermomechanical change, the constitutive

equations assumed must not violate the second law of thermodynamics In

continuum mechanics, this law may be stated as follows:

V. Principle of Entropy (Claufiius-Uuhem Inequality) . The time rate of

change of the total entropy H is^ never less than the entropy influx through

the surface S of_ the body and the volume entropy supply B ^in the body. T.us

is postulated to be true for all parts of the body and for all independent

processes (Eringen [1966c]).

Accordingly we write

(19.1) f r -TT- - B - is • da > 0

dt I "

where r so defined is the total entropy production. For simple thermomechani-

cal processes we have

(19.2) H = /pndv
1/

h
(19.3) b s t A
dv
/ e

q
(19.4) s ■ e

where n, h, q and 6 are, respectively, the entropy density, heat source, heat

vector, and the absolute temperature. Substituting (19.2) to (19.A) into (19.1)

gives
 -92-

(19.5) r E -{L /pndv _ ;k dv _ ^ Ji da > 0

Using the Green-Gauss theorem to convert the surface integral to a

volume integral and carrying out the differentiation with respect to time,

we get

(19.6) Hpn ■■ (-^).k - ^]dv > 0

Since this is to be valid for all parts of the body, we must have

(W.7) pn - (^).k-|i0

This is the well-known Glausius-Duhem inequality of classical continuum

mechanics. In micropolar bodies, this is considered to be unchanged.

Upon substituting h solved from (18.9), we can rearrange (19.7)

into the following form

(19.8) PY s P(n - i)
6
♦ 46 t.k£.v.£.,k. - 46 C-kmn t mn v.k

+
em£k\,£ +
~2 q e
k 'k-0
Ö

Still another form, convenient for some cases, is found by introducing the

Helmholtz free energy

(19.9) V - e - on

Hence

(i9.io) PV -r - f-(^
6
+ re
) +
i6 t.k£0v0£,k - -6 kmn
, t mn .k

+
e m£k\,£ +~7
8
V'k-0
 -93-

The Clauslus-Duhem inequality (19.8) or (19.10) is postulated to be valid

for all independent thermomechanical changes. This implies that we must

an
know the independent variables which affect Xi n. tj.« t nVo» d q. at the

outset. This in turn requires writing constitutive equations for these

variables. An example of this is to be found in the following article.

We also note that the _ntropy inequality (19.5) can be shown to

lead to an inequality restricting the normal component of q/9 on the surface

of the body (Eringen [l()66c]) ,i .e. ,

q
n 0 on S
(19.11) [T] ' '

where a boldface bracket indicates the difference on the quantity enclosed

calculated from two different sides of S, on S.

 -94-

20. THEORY OF MICROPOLAR ELASTICITY

A micropolar elastic solid is distinguished from an elastic solid

by the fact that it can support body and surface couples. These solids

can undergo local deformations and microrotations. Such materials may

be imagined as bodies which are made of rigid sliorL rylinders or

dumbbell type molecules.

From a continuum mechanical point of view, micropolar elastic solids

may be characterized by a set of constitutive equations which define the

elastic properties of such materials. A linear theory as a special case of

the nonlinear theory of microelastic solids was first constructed by Eringen

and Suhubi [196Aa,b]. Later, Eringen [1965], [1966] reorganized and extended

this theory. Here we give a self-contained account of this theory.

In linear micropolar elasticity, the strain measures are (cf. equations

(4.17) and (4.18))

(20.1) e, „ - e, . + e, „ (r - $ ) - u. . + £„. *
kl ki kim m m lvk ikm m

(20.2) v. . - e,. $
kim k£n n,m

Since only the nine components i>, 0 of v are independent and non-vanishing

(cf. equation (4.37)), we may instead of > use the axial tensor i>, for
K. x in K j *

simplicity. Upon arbitrary rotations and reflections of the spatial coordinates

represented by Q, „, i.e.,

(20.3) ^ - 2k|,|

(20.4) Q. Q - 0 0 - 6 . det 0 - '1

' \i^mi nkSa km ' "k«

and the tensors c, and $, transform according to

 -95-

i
t * Q c 0
ki SmwKTta
(20.5)

where in the last equation the plus sign is for det Q ■ +1 and the minus

sign for det Q. = -1. This is because *. is an axial vector. Equations

(20.5) express the fact that both c. . and 0. are ob lectlve tensors and

are appropriate for use as independent constitutive variables To this

list of variables we also include the temperature 6 so that the material

properties of these materials may depend on the temperature as well. The

constitutive dependent variables are

v and n
\i' \i> V

We now propose a set of constitutive equations of the form

S
^1 VfrS' *r.s' Ü)

"kl * VCrs' *r.s' e)

(20.6) qk - Gk(crs . ^^ 6)

ii ' y(f , * , e)
rt r.s

n - H(i , (> ,6)

rs r, s

The above equations are legitimate for linear homogeneous materials whether

Isotropie or not. For nonlinear Isotropie mattri.^ t they are acceptable in

form. However, since we are employing the infinitesimal strain measures, a

nonlinear constitutive theoiy in terms of linear strain measures would be

 -96-

inconsistent. For the nonlinear theory, the reader is referred to Eringen

and Suhubi [1964a (, h].

The constitutive equation (20.6) must be consistent with the second

law of thermodynamics as expressed by (19.10). Thus, upon substituting

(20.6) into (19.10), we have

' * (^ ^ + ^ '^ + "^ + r,6) + i '"'"

(20.7) . .
+ +
?\£^,k 7V'k^0

Consistent with the linear theory we write

— (A ) « 4

(20.8)
e V e
k£ ^ k,e ~ kimVra

The inequality (20.7) is postulated to be valid for all independent processes.

• • •
Here e. » *k «.♦ e an
^ e,
W can
^e var e
^ ^ independently. Since this inequality

is linear in all these variables, we must set the coefficients of these varia-

bles equal to zero. Hence

at

(20 9)
- V-pT^—
i,k

qk=0

at
 -9 7-

We therefore see that for ji micropolar elastic solid stress, the couple stress

and entropy denslty are derivable from a^ potential and the heat vector vanishes.

Since we did not consider the temperature gradient, we have no heat conduction.

Nevertheless, the free energy I and consequently the material moduli will

depend on the temperature 6. Since all terms in (20.7) vanish, we have the

entropy production density y also vanibhing.Thus.the micropolar elastic solid

is in thermal equilibrium.

Here we are concerned with the linear theory. We therefore consider a

polynomial for # which is second degree in the strain measures t, and $, , i.e.,

Ao + + A
P* " \i\i 2 kÄmn£mn+ \i\,i+ 2 \lmn\ti* m,n
(20.10)
+ C f (t
k8.mn'klc m,n

where A, A, , A, . , B.
, . . . are functions of 6 only. Since <(). is an axial
fourth and the
vector, upon a reflection of the spatial axes the^last termswill change sign

while Lhe other terms do not. For the function r' to be invariant B = 0,

C, . "0. We further note the following symmetry conditions which are clear
K».mn
from various summations in (20.10)

TcHmn mnk£ k^mn mnk£

which shows that for the most general mir/oj3olaj-_ anisotropic elastic solid,

the number of A. „ and B , „ is 45 each■ In addition, we have nine A, ,

which give rise to an initial stress in the undeformed state of the body.

Upon substituting (20.10) into (20.9). and (20 9)2, we obtain

ki Kl ikmn mn

(20.13) m, - B„, *
kl ikmn ni,n
 ■98-

These are the linear forms of the stress and couple stress constitutive

equations for the anlsotroplc micropolar elastic solids. When the Initial

stress Is zero, then we must also have A. . ■ 0. Thus, for the micropolar

solid free of Initial stress and couple stress, we have

k£ Tcimn mn

(20.15) m, - B.. $
Teil Ikmn m,n

Various material symmetry conditions place further restrictions on the con-

stitutive coefficients A,, and B„, . These restrictions are found in

ikmn ikmn
the same manner as In classical elasticity. Here we obtain the case of

Isotropie solids. If the body la Isotropie with respect to both the stress

and couple stress, we call the solid mlcroisotropic■ In this case, the

constitutive coefficients must be Isotropie tensors. For the second and

fourth-order Isotropie tensors, we have the most general forms

A. . - A6. , A. . =A16, „6 + A06, 6. + A-6, 6.

\i ki' lamn 1 kH mn 2 km in 3 kn £m
(20.16)
B, , = B-6. .6 + B06, 6, + B-6. 6.
kimn 1 kü mn 2 km Hn 3 kn im

where A, A., A?, A , B1, B , and B. are functions of 8 only In this case

then, (20.12) and (20.13) take the special fcrms

(20.17) tk£-A\^A1rrr6k£ + Vla + A3e4k

+
(20.18) V-^W Vr.rWf,k Vk,*

For vanishing initial stress A = Ü. Introducing

 -99-

A. ; A , A« = u + < , A. = vi

(20.19)
B1 a , B2 £ Y , B3 = ß

the above equations can be written as

(20.20) tki - XcrAl ♦ (u + <)eki ♦ uhk

(20 21)
- \tm »♦r.At4 ß*k.£ + Y
*£.k

For the free energy In this case we find

p tAe
'- 2 kkeU+ (lJ+ <)e
k£ek£+ "«ki'lk1
(20.22)
.1
(a
2 *k.kV£+ ß*k.£^.k+ ^k.A.^

An alternative form to (20.20) to (20.22) Is

(20.23) t, . - Ae 6 + (2u + <)e, . + <e. . (r - ♦)

ki rr kl ki klm m m

(20.24) m, . - aO (5. „+ ß*. . + Y#

■M - a$r,rÖk£+ B\.£ + Y^.k

(20.25) pf - i [Aekkeu ♦ (2u + Oe,^] ♦ .(rk - ^ (rk - ^

+ (a
2 *k.k^.£+ ^k,£*£.k+Y\.A.£)

We note the difference between Isotropie micropolar elasticity and classical

elasticity by the presence of four extra elastic moduli, namely, <t a, 6 and

Y. When these are set equal to zero, the above equations (20.23) to (20.25)

revert to Hooke's law of the linear Isotropie elastic solid.

 -100-

21. RESTRICTIONS ON MICROPOLAR l.LASTIC MODULI

The stability of materials requires that the stored elastic energy

be nonnegative. This condition is also essential for the uniqueness of

the soJutlons. This requirement places certain restrictions on the micro-

polar elastic moduli. The following theorem, Eringen [1966a].provides these

conditions for * independent of 6.

Theorem. The necessary and sufficient conditions for the internal

energy to be nonnegative are

0^3X + 2ii+t<, 0j_v , 0<_<

(21.1)
0 < 3a + 2> , -Y •■ ß < ^, 0 •

The sufficiency of (21.1) is proven by observing that when these ineaualities

hold,each one of the following energies constituting the internal energy

density is nonnegative

e + £ + e
(21.2) ^ E R M

where

Pe £ tXe
E 2 kkeU+ (2lJ + <)e
klie£k]

(21.3) 0eRE *(rk-*kHrk- V

PF ( + ß
'M ' 2 ^k,k^,£ \,£*£,k+ ^l.k^l.^

The fact that pe., is nonnegative under the conditions (21.1), and (21.1)„
t. 12
is well known for the classical elasticity. It is simple to observe that

pf is nonnegative for r, f (l>. whenever < j_ 0. To see the same for pe ,

we write this expression as

 -101-

^M-T(3a+ß +
^k.kV+ (Y
" ß)
*Ik>il*(k.t]
(21.4)
+ ( + b)
^ ^(k,£) -T*r.r\£][*(k.£) -T*8.36k£]

where

E U +
'(k.O 2 M •l.^
(21.5)
ä u
'[k.l] 2 k,£ " *e,k)

From (21.4) it is clear that when

(21.6) 3a+ß + Y^0 , >-ti_>.0 , Y + ßj_0

we have otxt > 0 so that (21.1) are sufficient for PEW > 0.
M — M —

Conditions (21.1) are also necessary for the nonnegativeness of pe.

To prove this, we recall that e , r, - 4). and $. can be varied independ-

ently of each other. Since the above three energies are uncoupled with

respect to these variables, each one of these energies must be nonnegative

independent of each other. The fact that (21.1). and (21.1). are necessary

for oe > 0 is known to us from classical elasticity. Excluding the case of

r, ■ 0. (indeterminate couple s tress theory), we see that pe is nonnegative

if and only if < ^_ 0. For the case of r, ■ $,, by replacing 2- + * by a

new modulus 2u we shall have < disappear from the constitutive equations.

Thus, it remains to prove the necessity of (21.1) for the nonnegativeness

of PF-M. To this we write it as a quadratic form in a nine-dimensional space,

i.e..

a = a
(21.7) P^-^.y^j . ij ji • (i.J = 1,2, ,9)

where
 -102-

y y E y 5
l " n.l ' 2 ^2.2 • 3 ^3,3

y E y E
4 - ^1.2 • ^5 *2,1 " 6 *2.3

'8 " ♦*
; 5
(21.8) y7
'7 *..
3,2-)
lf •* fa *3,1i •' yc
'9 " 1,3

a11 - a22 - a33 - a + 6 + y. a44 - a..,. - a66 - a77 - a88 = a99 - y

a12 - a^ - a23 - ex . a^ - a67 - a89

all other a, = Ü

The characteristic values a of a are obtained by solving the equation

(21.9) det (a.. - a6 .) = 0

The nine roots a, of this equation are

a1 = a2 = a3 - Y - ß > a^ - a. « a6 - a7 = a8 « Y + ß

a9 - 3a + ß + Y

In order for peM ^_ 0 to be satisfied for all y , it is necessary (and suffi-

cient that.) (21.7) be an ellipsoid in nine-dimensional space, i.e.,

Y-ß^0 , y+ß^O , 3a + ß+Y^0

This set of conditions is the same as the last three conditions of (21.1).

Hence the proof of the theorem.

The nonnegative character of the internal energy density has important

implications in regard to uniqueness theorems in both static and dynamic micro-

polar elasticity. For these and other important results see Eringen [1966i,hi-
 -103-

22. FIELD EQUATIONS, BOUNDARY AND INITIAL CONDITIONS

The field equations of linear micropolar elasticity are obtained by

substituting (20.23) and (20.24) into (17.17) and (17.18). Hence

(22.1) (X + M)«Mk ♦ (u + <)ukM + «^.^ * P(fk - uk) - 0

(22.2) (a ♦ BUitlk * Y*k>u + <^mumf, - 2<*k + p^ - j^) - 0

where we have taken j. . ■ J^i,» ^or the microisotropic solid. In the linear

theory, p and J are considered constants and the accelerations ii and ij.

are calculated by their approximate expressions

2 2
(22.3) B. - 1 , ♦. • 1

The vectorial forms of these equations are found to be convenient for the

treatment of problems in curvilinear coordinates. These are readily obtained

from the above equations by simply multiplying them by i, and observing that

(22 4)
- •Wk"-'» ' WM4k"YM
u
k,uik - vv . u - v ,< v * u

where V is the gradient operator so that

7(> 5 grad $ , V • u ^ div u , V x u 5 curl u

Hence

(22.5) (A + 2u + <)VV • u - (u + <)V x 7 x u + icV x ^ -f p(f - U) ■ 0

(22.6) (a + 3 + Y)VV • $ - yV * V* $ + <V * u - 2<<p + p(t - jlf) - 0

 -104-

For an initial value problem, the initial conditions have the form

(22.7) u(x,0) - uo(x) , u(x,0) - v0(x)

in 1/
(22.8) $(x,0) - $0(x) , ♦(x.O) - vo(x)

anc v are
where UQ, VQ, ^O» * 0 prescribed in V at time t - 0.

Many different types of boundary conditions are suggested in applica-

tions. For example, we may prescribe

(22.9) u(x',t) = u'

x1 on S
(22.10) ♦<?,,t) - f'

on the boundary surface S of the body. An equally permissible set of boundary

conditions requires the prescription of the tractions and couples, i.e.,

(22 n) C
- ttni " t
(n)k
on
(22.12) m£lA - m(n)k

where t. and m are the stress and the couple stress tensors given by (20.23)

and (20.24) and t, .. and m, ., are the prescribed tractions and couples on
(n)k (n)k
S whose exterior normal is n.

In some other problems, a mixture of the above two types of conditions

occurs, e.g., on some part S of S one may have (22.9) and (22.10) and un the

remainder S, 5 S - S. the conditions (22.11) and (22.12). Still other types

of mixed conditions involving some components of one set and the remaining

components of the other set are possible. All admissible sets of boundary con-

ditions allowing unique solutions must satisfy (Eringen fl966a])

 -105-

(22 13)
- Wk+ «(IOÄ -0 onS
' t
-0
where u, 4. t, ., and m, v aie, respectively,
v the difference of u, 4, t, ,
- '' • (n)* -(n) ' ' »i» .(n)
and m, , from their respective
v values on S.
~(n)
The field equations (22.1) and (22.2) are valid only for micropolar

Isotropie solids. Note that for vanishing <, a, ß, i, and j, equation (22.2)

reduces to 0 - 0, and (22.1) gives the celebrated equations of Navler of

classical elasticity.

For the anlsotroplc micropolar elastic solid, the field equations

replacing (22.1) and (22.2) are obtained by substituting (20.14) and (20.15)

Into (17.17) and (17.18)

(22.14) A„1.__(u_ _„ + e__^r ^ + D(fk - tt^ • 0

Hkmn n,in£ nmr

(22.15) B. . <) „ + e. A (u + e (|) ) + p(£, - JJ.JL) ■ 0

kUmn m.nJ, kmn mnpq q,p qpr r k ilk «,

For the expressions of e.. and o, we used (20.1) and (15.18).

Finally, we record below the expanded forms of the field equations

(22.1) and (22.2) In rectangular coordinates.

2 2 2
3u 3u 3u 3 u 3 u ä u
(x + u
) i£ (-^+ 1^ ^ + (
^ >(—!
3x
+
—r
3y
+
—f)
3z
2
3*30 3 u
^(-37-1^+p(fx---f
at
)=0

2 2 2
s 3u 3u -u 3 u 3 u 3 u
^+ ^ ^7' (-^ + ^' + ^ + (u +
^(-i
3x
+
—^
3y
+
^
3z
2
(22.16) 1% 34) 3*u
+ ^(-T3z1 - -A
3x + P(f y 2 ■ 0
1)
ot
2 2 2
3u 3u 3u 3 u 3 u 3 u

' 3x 3y 3z
2
3<)) 3(J> 3 u

dt
 •106-

2 2 2
j| 3(t 941 3$ 341 ^^ du HI

7 3zy
3x 3y dz '

32^
- 2K* + pil - j 1) - 0
at

2 2 2
di|) 3i|> 34) 3 $ 3 $ 3 ^ 3u 3u
(a + ß) ~ s
3x (-53x + "i3y1 + -Ä+YC
3z —I
^ 2 ♦ —I"
„2 ♦ —h
^2 ♦ ^(-r
3z
5
3x
' 3x 3y 3z

(22.17)
324)
- 2<* + pU - j —fe - 0
y y st2

2 2 2
3i(i 3^ 3$ 3 41 3 (|) 3 ()) 3u 3u

dx 3y 3z 3y

320)
- 2K* + p(£ - j h ' 0
Z 2
»t^
 -107-

23. INDETERMINATE COUPLE STRESS THEORY

In Art. 22 we indicated that classical elasticity is a special case

of the micropolar theory when <»a"B m

y m
i m
0 and i ■ 0. The same

is also true for motions in which the microrotation vanishes. There exist

other classes of constrained motions which have attracted the attention of

research workers. The most popular among them is the indetermine couple

stress theory which is contained in the work of the Cosserats [1909]. Recently,

Truesdell and Toupin [1960], Grioli [1960], Aero and Kuvshinskii [1960], Mindlin

and Tiersten [1962], Toupin [1962], and Eringen [1962] independently presented

new derivations and supplied various missing parts «^f the theory. This theory

can be obtained as a special case of the micropolar theory if the constraints

k k 2 kirn m,l

are imposed. In this case, the stress constitutive equations (20.23) reduce

to

(23 2) Ae + (2y +
- ^ki) " rA£ ^k^

where a parenthesis enclosing indices as usual indicates the symmetric part

of the stress tensor. We also use a bracket to denote the antisymmetric part

of tensors, e.g.,

a S (a + a
(k£) 2 k£ ik) ' a
[k)l] E
T (a
U " a
lk)

Thus, when (23.1) is valid, the antisymmetric part of the stress disappears

firom the constitutive equations. We can, however, remedy this situation by

anothei artifice. The equations of moment of momentum (17.18) can be used

to solve for the antisymmetric part of the stress tensor. Multiplying (17.18)
 •108-

by1 e, , we solve for

krs

<23-3) '[k*] " 4erk*-nr,« * i pcrkt(lr " V

where we used the idennity

(23.4) c , 0e
rk£ rmn - <5,km 60Hn - «5,kn <50Um

Upon carrying (23.1) Into (20.24) we have

(23.5) ^„--le,
m - e u „+^e u
k£ 2 krsUs,r£ + 2 en£rsUs,rk

Carrying (23.5) and the expression of

ü
(23.6) ».r " j#.
J
r " t
2 J«
J
rmn n,m

into (23.3), we have

(23 7)
- 'CMl •i^CMJ -I0(Erk££r +
JÜ[k.£])

2
where V is the laplacian operator in rectangular coordinates

(23 8) V E U
- \ £.kk

If we not substitute (23.2) and (23.7) into

(23 9) +
- \l ' '(ki) '(kl]

we get the total stress tensor

(23 i0) Xu
- ^ - r.r5k£ +
^+7)(\.£ + U
£.k)

+ 2 v2u
[k.i] " ^ p(e
rktlr * ^(M^
 -109-

The presence of the body couple I and the acceleration U. in this equation

is certainly disturbing since the constitutive equations, in general, should

not contain such terms. When (23.9) is carried into the equation of balance

of momentum (17.17), we find

(23.11) U+U+f+fv^u^+^+f-^V2^

£k

By use of identities (22.4), we may also obtain the vector form of

these equations:

(23.12) (X + 2u + <)VV • u - (y + j - J V2)V x y x y

+ p(f + -j V x £) - p(]. + ^ v x Vx)U - 0

where for the lapalcian operator we have used

(23.13) V2^ - VV • A - V x v x A

Equation (23.12) takes the form (3.27) obtained by Mindlin and Tiersten [1962]

in an entirely different way if we write y for y + K/2 and n for y/U and j ■ 0,

Thus, these author^ as with others, have neglected the micropolar rotatory

inertia. Equations (23.11) or (23.12) are the field equations of the theory

known as the (indeterminate) couple stress theory. It is to be observed that

in this theory the skew-symmetric part of stress and consequently the stress

are dependent on the applied loads and inertia, and they are not determined

solely as a result of the constitutive character of the medium. This violates

the axiom objectivity since the applied loads and inertia terms involved are

not objective quantities. A second relevant point is that while according to

 -110-

(23.5) one can determine both the symmetric and the skew-symmetric parts

of the couple stress, this in fact is not the case of the couple stress

theory in which constitutive equations are derived from a free energy

function, as is done in Art. 20. It is clear from (20.9)-, for example,

that all components of ^ , can no longer be used as independent variables.

In fact, if one uses (23.1) in the argument of free energy f, one finds

that all nine components of m, are not independent. Moreover, the skew-

syirantric parts of the stress and couple stress remain indeterminate .

This is the reason for the use of the terminology "indeterminate".

This situation has certain similarity to the isochoric motions of

compressible bodies as compared to the motions of incompressible solids.

As is well-known, in the latter case the pressure is not determined through

the constitutive equations. Finally, in the indeterminate couple stress

theory, the number of boundary conditions on the surface tractions and

couples must be reduced from six to five. A consistent set of boundary

conditions must not violate the uniqueness theorem. Mindlin and Tiersten

[1962] have obtained a uniqueness theorem for the following set of boundary

conditions.

Let (x., x„, x_) be a set of orthogonal curvilinear coordinates

taken in such a way that x- ■ x_ locally coincides with the surface S of

the body. The boundary conditions consist of specifying at x_ ■ x. one

factor in each of the five products

(23.14) p^ , p2ü2. E(33)G3 , m31r1 , S^

In this regard, see the discussion given in Eringen [1962, Art. 40]. See
also Mindlin and Tiersten [1962], Toupin [1962], and Eringen [1964b].
 -111-

where

p1 i t(31) -- (mi2;1 - m33.2)

(23.15)
P : + (m n,
2 '(32) 2 il;i ' 33;l)

Here an index pJaced after a semicolon denotes directional differentiation

along the corresponding curvilinear coordinate, and a superposed bar the

boundary values of the quantities involved. Curvilinear components of the

displacement vector on S are denoted by u, , the couple stress by m, , and

the stress tensor by t, .

If an edge is an intersection of two orthogonal surfaces x - xx

and x ■ x , then we must also specify

(23.16) [m,.] - [m ] or u
33 11 2
vO
X vO
X
3 l

The reduction of the number of boundary conditions from six to the above

five, (23.15), is similar to the one encountered in the Bernoulli-Euler

theory of thin plates. Conditions (23.16) are the analogs to the corner

conditions .

Under suitable regularity assumptions, the above five boundary condi-

tions, together with the assignment of the body force field pf, curl (pH),

and the initial values of u and u (with j ^ 0), are sufficient for the unique

determination of t.,t r, , t,. sv and m,. ,. The displacement field u and

the rotation r are unique too within an arbitrary rigid body displacement

field.

The indeterminate couple stress theory described above has many

apparent limitations. Whether or not bodies can undergo such constrained

 -112-

motlons is not known. The appearance of nonobjective quantities in the

constitutive equations, the limitations on the spin inertia and the body

couple field, and physically unnatural boundary conditions leaves much

to be desired. Experimental comparison for certain practical applications

exists (cf. Schijve [1966]). A discussion of the weakness of this theory

indicating the disagreement of the theoretical results with the experiments

has been given by Kaloni and Ariman [1967]. Nevertheless, recent litera-

ture contains a large number of solutions in the field.

 -113-

1U. PROPAGATION OF WAVES IK AN INFINITE

MICROPOLAR ELASTIC SOLID1

Here and in the following several articles we investigate solutions

of certain dynamic and static problems in linear Isotropie micropolar elas-

ticity. Essential to these problems are the field equations (22.5) and

(22.6) and boundary conditions of the type listed in (22.11) to (22.12).

The propagation of linear Isotropie micropolar elastic waves with

vanishing body loads is governed by (22.5) and (22.6) or

(24.1) (cj + c^)W • u - (C2 + 0^)7 x V * u + c^V « $ - Ü

(24.2) (cf4 + c,)W

5 • A - c.V * V * A + u^V
4~.t 0 * u - 2UK<1>
1
o = *

i nere

2 \+2u 2 2
C C c
l P ' 2 P ' 3 P
(24.3)
2
c
2 = JL 2 a+ß 2 3 <
C
' Pj 5 PJ* j Pj

We decompose the vectors u and % into scalar and vector potentials as

follows:

u-Vu+7xu | V'U-0
(24.4)
^»V^+Vx* t v.$-o

Substituting these into (24.1) and (24.2), we see that these equations are

satisfied if

The present section is based on the work of Parfitt and Eringen [1966]
 ■114-

(24.5) (cj ♦ c^)V2u - U

2 2 2 2
(24.6) (cT + cJV 4, - 2ur 1
•4 5' O

(24.7) {c22 + c23)V2U + c^V x * - Ü

(24.8) cfv2* - 2a)2* + w V x U - *

It may be observed that (24.5) and (24.6) are uncoupled for the scalar

potentials u and (J, while equations (24.7) and (24.8) for the vector

potentials are coupled.

Plane waves advancing in the positive direction of the unit vector

n may be expressed as

(24.9) (u, $, U, •} - {a, b, A, B} exp [ik (n • r - vt)]

where (a,b) are complex constants, (A, B) are complex constant vectors, k

is the wave number, and r is the position vector. Thus

(24.10) It s if , r - «^

in which t is the wave length and i are the unit rectangular base vectors,

Substituting (24.9) into (24,5) gives

(24.11) v2 = c2 + c2 - (X + 2u + O/p

which shows that a plane wave with the displacement vector

(24.12) u = ik an exp [ik (n • r - v t)]

may exist in the direction of propation n, Fig. 24.1. These waves are the

counterpart of the classical irrotational waves and reduce to them when < ■ 0,
 •115-

We designate these waves, longitudinal displacement waves.

A second scalar plane wave is the solution of (24.6) in the form

(24.9). The wave speed in this case is

2 2 2 2-2
(24.13) v^ - cT + c^ + 2(A
2 4 5 0

If we introduce the angular frequency u

(24.14) u - 27Tf - Ztv./i - kv

the wave speed may be expressed as

2
2 '1
(24.15) v2 (c2 + c2)(1. V . ^Y
245
^ pJ(i--^7)
Pju2

ITie speed of these waves depends on the frequency. Hence they are disper-

sive. Since

(24.16) o + ß + Y 1 0

we see that such waves can exist whenever

(24.17) Cü2 > S2mQ

These waves will be called longitudinal microrotation waves. Fig. 24.1,

The microrotation vector is given by

(24.18) J • ?♦ - ik2bn exp [Ik-Cn ' r - v2t)]

For the case of u. ■ >^2w^ ^ w , we have v_ ■ ^ and the wave does not exist
2 0 c 2
When w„ < ^2^^, v„ becomes purely imaginary, i.e..
 -116-

(24.19) v2 - *i|v2! , i . /T

It may be seen that a standing wave of the form

W
2
(24.20) ()i - b exp (- i i n • r) exp (-iw t)

is possible.

For such waves, propagation is possible if w„ > '•tt.' Hence,

i/2w„ = w is a cut-off n
frequency for these waves.
0 c '
The vector wave solutions are obtained by substituting (24.9) into

(24.7) and (24.8). This results in two simultaneous vector equations for

the unknowns A and B.

a.A + iaDn * B = 0
A- B
(24.21)

ißAn * A + ß B » 0
A~ B~

which for nonvanishing aA , a,,^., and ß„ are subject

J to
ABA B

(24.22) n'A-0 , n-B=0

resulting from (24.4) and (24.4) . Here

-,2,2 2 2. - . 2
aA = k (v - c2 - c3) , aB -. kc3

(24.23)
2 2 2 2 2-2
ß E kW ß £ k (v c 2w k
A 0 • B " ; - 0 >

Equations (24,22) show that the vectors A and B lie in a common plane whose

unit normal is n. Solving from (24.21) for B we have

(24.24) B - -i ~ n x A
6
B "
 -117-

Hence, the three vectors n, A, and B are mutually perpendicular. Moreover,

vanishing A implies vanishing B. These two types of waves are, therefore,

coupled and cannot exist without each other. From (24.4). and (24.4). we

see that u and $ corresponding to U and $ are normal to each other and to

the direction of propagation n. Hence these waves are transverse waves.

We call the waves that are associated with U. transverse displacement waves.

and those that are associated with $, transverse microrotation waves. Fig.

24.2. The transverse displacement waves have their classical analogues in

equivoluminal waves and in_the limit they reduce to these waves.

The velocities of propagation of these waves are determined by

carrying (24.24) into (24.21) and using (24.22) . This gives

(24.25) av4 + bv2 + c = 0

where

o 2 ü) "2
a = 1i - 2Iü
0

(24.26) b 5 -[c2 + Cjd - 2u>2y2) + c^l - a^aT2) ]

2/ 2 J 2,
c = c4(c2 + c3)

The positive real roots of (24.25) are

1/2
i n—
v. « fe*- (-b + /b -4ac )]
3 2a
(24.27) i ...
1/2
1 /2
V4 ' l
2l ("b " ^ -4ac^

By studying the discriminant

 -118-

-
jfi
rb -4ac ;r 2 - c.2 - c 2 +J 2(c
, ■ {[c. 2, 2u -l}
_ 2 + —1 C.)u ]

1/2^
.222 -2,}1/
+ *c3c4«0«

unde r the conditions «r ^_0, Y 2. 0 compatible with (21.1), we find that v_

is real when w > w and v, is real for all values of w. The frequency
c a
u ^ "^Wn ^8 a a n
8 i a
cut-off frequency for waves propagating with velocity
c

V
3*
In summary, we found that in an infinite micropolar elastic solid

six different types of plane waves propagating with four distinct speeds

of propagation can exist. These are:

(a) A longitudinal displacement wave propagating with speed v .

(b) A longitudinal microrotation wave with speed v propagating

in the longitudinal direction whenever the frequency of these waves is

above the cut-off frequency u . These two types of waves are uncoupled.

(c) Two sets of coupled transverse displacement waves and transverse

microtation waves at speeds v and v . Of these, the waves having velocity

v can exist when u » w otherwise they degenerate into distance decaying sinu-

soidal vibrations.

A detailed analyses of the wave speeds v?, v , and v is sketched in

Figs. (24.3) and (24.4).

By use of (21.1), Parfitt and Eringen [1966] have shown that a con-

sistent solution for v. and v. requires the existence of

3 4 ^

2 2 J 2
4-2 3

or the additional inequality

(24.28) \ L * + *
 -119-

In addition, the wave speeds as a function of . must satisfy

(24.29) v.2(0) < v2(a) ) < vj(»)

A — 4 c — 4

(24.30) v^-) > VjC-) > vj(»)

Also they found that

(24.31) v^ > vj

The study of the relative magnitudes of v , v , and v, requires

sum of the
knowledge of the relative magnitude of the/fconstitutive coefficients

ci + d with respect to jic. Thus

2 2 2 1
v0 > v. > v. for UJ<U) , a+ß> — Jjk
2 J 4 c 2

2 2 2 * 1
(24.32) v. > v0 > v. for u <ü)<ü) , a+ß^—jk
3—2—4 c— 2J

2 2 2 c *
v„ > v_ > v. for on < w
2 — 3—4

* 2 * 2 * *
where w is a solution of v (w ) ■ V.(üJ ) in the range u < u < <x. For

other details tne reader is referred to Parfitt and Eringen [1966].

 -120-

25, REFLECTIÜiN OF A LONGITUDINAL DISPLACEMENT WAVE

In this section we study the reflection of a plane longitudinal

displacement wave at a stress-free plane surface of a micropolar half

space. If x ■ 0 is the plane of the incident displacement vector, then

the reflected waves from the boundary surface z * 0 can be shown to

remain in this plane so that the problem is a two-dimensional one. On

the boundary plane z = 0 being free from traction we must have

t, , - m, , "0. Since we also have u « *„ - 4>„ ■ 0 through (22.11) and

~(n) ~(n) 2 3
(22.12) for the tractions on z = 0, we will have t, . = t , t, , = t
(z)z zz(z)y zy
and m, x = m , or usina0 the constitutive equations and (2A.4) we have
(z)x zx

t = A(u + u ) + (2u + <)(u - U ) = 0

zz »yy »zz »zz x,yz

(25.1) t = u(u - U ) + (u + <)(u + U ) + <(* - $ ) = 0

zy ,yz x.yy ,yz x,zz z,y y,z

m ■ y(4 - ♦ ) = 0
zx z»yz y»zz

The nonvanishing components of the solution vectors (24.9) appropriate to

this problem are given by

u ■ a exp [i(k.n • r - u),t)]

a a la 1

(25.2) Uß - lAßx exp [i(kgng • r - ^t)]

%' (V + B
dzk) eXp [U
hnc ' ?" "g^1

where u. - k v (a » l,2;ß ■ 3,4) and the repeated indices are not summed.

The coefficients A and B are related to each other by

iu) A
0 3x
(25.3) B^ s S-iS- r (n. j - n_ k)
3 , / 2 2, o 2, -2 3z^ 3y
k3(v3 - 2«Jc3 - cp
 -121-

wlth a similar equation for B,.

The potentials listed in (25.2) satisfy the boundary conditions (25.1)

at z - 0 if

13 4

(25.4) k.n. - k.n. - k.n. ■ k.n.

1 ly 1 2y 3 3y 4 4y

k.n. - k.n„ - k„n„ - k.n.

1 Ix 1 2x 3 3x 4 4x

and

(25.5) [XkJ ♦ (2|i + <)kjn^ la. + [ XkJ + (2u + <)kjnj ]a9

1 ilzl 1 12z2

2 2
- (2u + <)k n_ n^ A_ - (2y + <)k n. n A ■ 0
3 3y 3z 3x 4 4y 4z 4x

Since the incident wave is in the x » Ü plane,

r n, "0 and (25.4)- yields
Ix 3

n « n. - n. - 0
2x 3x 4x

which is the proof of our statement that all reflected waves are on the

x = U plane. Waiting

n, = cos 8, , n„ = cos 6,, , n. ? cos e_ , n. S cos e.

ly 1 2y 2 3y 3 4y 4

(see Fig. 25.1) and m. ■ k v. from (25.4) we get

V V
3 4
(25.6) cos e„ » cos 9, , cos 9« ■ — cos 0. , cosO . - — cos Ö.
2 1 ' 3 v 1 4 v1 1

where v are the speeds of the various waves found in Art. 24. From (25.6)

it is clear that Ö - ö».

 -122-

Substituting (25.2) into the remaining two equations of (25.1), we

obtain two other equations. These two equations and (25.5) are adequate

to determine the amplitude ratios a^/a,, A- /a. and A. /a.. The amplitude
2 1 3x 1 4x 1
B is determined from equations (25.3) and similar ones for B. .

If we set n = 0 in (25.5), we find that a = -a.. This shows

that for normal incidence the reflected longitudinal displacement wave

is also normal to the boundary. For n1 ^ 0, the following are the solu-

tions for the ratios of the wave amplitudes

(25.7) — = {[A + (A+2u+<) tan2 Q,][(\*<) tan2 6. - y - (u+<)g? tan 9, tan 6. - Q2

a1 1 4 13 4 2

tan
„ 2 Ö, „ 2
+ (UQ1+Q3) tan g ] + UvHcy tan ei tan ^(Q*-!))

2 2 2
x{-[X+ (A+2u+<) tan 0 ][(u+<) tan o - p - (U+K)^, can H tan 0, - Q»

2 tan 6
/ 1
+ (WQJ-Hjj) tan g ] + (2y+<) (Q^-l) tan Bj tan 6^}

A
(25.8) —-= -2(2u+<)[A + (A+2u+<) tan G. ] tan 8,
a1 1 1

"{-[A + (A+2U+K) tan ei][(u+<) tan 6. - u - (u+<)Q2 tan 6. tan 6.

1 4 13 4

- Q2 + (UQ1+Q3) tan /] + (2P+K)2(Q2-1) tan ei tan e^}"1

2 2
(25.9) Qj £ [v*(l - —|) - c;](v2(l - —A-) - c2)

02 !iV4 r 2n
2
V 2 -1
Q2 " ~~2" (V (1
4 2) " C ,
4
U) V. u
4y
2 2 _ 2
2 V3 2 1 2
oQ3 ^
- -f-f- .[v4n ^ - cj
(l --7) ."
3y
 ■123-

These results were studies in detail by Parfitt and Eringen [1966] for

various special cases. A sununary of their findings is given below:

A longitudinal displacement wave at a plane stress free boundary

in general produces three reflected waves (as compared with the two waves

of the classical theory); one longitudinal displacement wave reflected

at an angle which is the same as the angle of incidence, and two sets

of coupled transverse waves, one travelling with speed v_ and the other

set with speed v, (Fig. 25.1). Their angles of reflection are calculated

by (25.6). For normal incidence (0 = 90°) the waves at speeds v. and v,

vanish and the reflected wave is a longitudinal displacement wave normal

to the boundary. The amplitude ratios as functions of the angle of inci-

dence 6. are given by (25.7) to (25.9) and similar equations for B..

The general solution prevails as Ö. decreases from 90° to a critical

angle 6. which makes 0.« 0 and A. =0. At this angle of incidence we

1 3 4x
have a surface motion travelling at speed v and a reflected longitudinal
* *
wave at speed v and angle 6„ ■ 9.. ■ 6. . As 9. decreases from 9 to zero,

the angle 9 becomes complex. The interpretation in this case is that a

longitudinal wave is reflected into the medium at an angle 9 and a coupled

transverse wave (decaying with depth and travelling at a speed c, in the

range v < c j_ v„, is reflected. For 9 ♦ 0, a limit solution is possible

Parfitt and Eringen have also studied the reflection of coupled

transverse shear and microrotational waves and the reflection of a longi-

tudinal microrotational wave. For these and other interesting results, the

reader is referred to the foregoing reference.

 ■124-

26. MICROPOLAR SURFACE WAVES1

In this section we investigate the propagation of surface waves

in a micropolar half space. In Section 25 we have shown that the inci-

dent and reflected waves propagate in the same plane. We select this

plane to be the x - 0 plane. Thus we take u. = <t>2 = $- * 0, im i- m

0,

and u2 s v(y,z,t) u ^ w(y,z,t) and #. i#(y(s(t) as functions of y, z,

and t only. In this case (22.1) and (22.2) reduce to

0C

„2. ,dv 3v, „ t n

0
YV 41 + <(T— 77) - z** - oi —t "
3y •■ jt2

2
where V is the two-dimensional laplacian operator, e.g.,

2 2
2 2
'>• 9z

We consider waves which are propagating in the plane x " 0 with an amplitude

decay in the z direction.

v = A exp (-£,z) exp [iq(y - et) j

(26.2) w = B exp (-^z) exp [iq(y - ct) )

$ - C exp (-:z) exp [iq(y - ct)J

1
Suhubi and Eringen (1964bJ
 -125-

Substltutlng (26.2) Into (26.1) gives three equations

[-(X + 2u + <)q2 -f (u + <)C2 + pq2c2]A - (X + uUqCB - <& - 0

(26.3) - (X + u)iqCÄ + [(X + 2w + <) ;2 - (u + <)q2 + pq2c2]B - lq<C - 0

<a + lq<B + [Y(C2 - q2) - 2< + pjq2c2]C - 0

A nonvanlshlng solution for A, B, and C may exist If the determinant of

the coefficients Is zero. This gives

2 2 —
2 „
C C C fc
2 2 2
(26.4) [(I + -i-K -(14 -^ )q ] {[j(H - j(e - ^j)q2 - 2e] «
C C C
2 2 2 2
2 2 2 2 2
[e + l)c - (e + 1 - ^q ] + e (c - q )} - 0
C
2
where

(26.5) e I c2/c2 , 0 = c2/c2

and c. to cs are given by (24.3). A set of approximate roots of (26.4) Is

2
obtained by neglecting the terms containing E . Hence

9 2
(1
4 ' " TTT*
l 3 C + C

9 2
(26.6) Cj « (1 - (1 - t) ^jjq^
C
2

S-2 F+ a-^
C
2
4

in order for the waves to be surface waves, we must consider only the posi-

tive values of the roots 4,, C„, and c«. The displacement and mlcrorotation

field can now be expressed as

 -126-

3
- ^ Aj^ exp (-;kz) exp Uq(y - ct) ]
k-1

3
ex
(26.7) »• I W P ("^k2^ exp [iq(y
" Ct
^
k-1

; - u.A- exp (-;3z) exp [iq(y - ct)]

where
A A
(26.8) l q ' 2 ;2 '

2 2
iq j ,,, w2e c 2. c 2,

C C
4 2

On stress-free boundary surface we must have

t .xiv + (x + 2 j iw _ 0
zz 9y 9z
at z - 0

(26.9) t - y|2-+ (u + <) ir + «^ - 0

zy dy dz

in • Y T^ - 0
ZX ;Z

Substituting (26.7) into (26.9), we obtain a set of three homogeneous equa-

tions for A1, A_, and A_. The determinant of the coefficients must vanish,

Hence

(26.10) '.D
3
- 8w2 + 8(3 - k)aj - 16(1 - k) - 16e(l - kw) - 0

(26.11) £;3U3 - 0

where

(26.12) M I C2/C2 , k I c^/cj

For c ■ 0, (26.10) reduces to the classical expression of the Rayleigh surface

waves. Denoting the values of N for this case, with w to a first-order

K
 •127-

tpyi jüimbtion in r.,we obtain

16(l-kco )
(2C U) u) - u)D +
R
3^ - 16ui + 8(3-2k)
R R

For k ^ l/J, which correspond to Poissons ratio 1/A, and for the incom-

pressihle solids (k - 0) we have, respectively,

c - 0.919 (1 + 0.932e)c2 , (k - 1/3)

(26.1A)
c - 0.955 (1 + 0.783e)c2 , (k - 0)

Tie tc-is Lcntaining c are the first-order corrections to the Rayleigh

wave velocity in each case.

Equation (26.11) gives the speed of propagation of a new type

surl.i t wave not encountered in classical elasticity. This is given by

Ui.lS) — - [2e/j(l + I - Q)]1/2q,~1 ■ (2e/j)1/V1

2

This v.ev wave speed depends on /e and it is dispersive.

 -128-

27. STRESS CONCFNTRATION AROUND A CIRCULAR HOLE

For the determination of the stresses In a plate with a circular

hole, we need to express the basic equations of micropolar elasticity

in plane polar coordinates (r, 6), Fig. 27.1. For the equations of

balance of momenta, parallel to classical elasticity, we have

3t 3t
rr 1 er + irr_Llee+ f *\

ät
i27A)
rf1 n re+
+ 1 ^96,+ tjJ^JM , pff . p \
^r 7 TT —;— e ~7
at

dm , jm m -it
rz 1 Qz rz z
K J
3r r 30 r re Or z ^ .^2
dt

For the nonzero components of the strain tensor e. given by (4.17) through

the methods presented in Eringen [1962, Appendix] we obtain

3u
r
Srr " 3r

e
e9"7 (—+ V
(27.2)
v
re 3r z

3u
1
eer ■ 7 ^? - ( u ) +
e *z

The constitutive equations for the stress read

t - (X + 2u + <)e + X£„n
rr rr 69

tee - Aerr H- (A + 2u + 0«M

(27.3)

^9 " ('u + 0e
r9 +
^t.

+ (M + <U
^r " ^re er
 -129-

.rn these equations

n (t , t „,...), (e , e „,...) and u and u., are,
rr rO* ' rr r6 r 8
respectively, the physical components of the stresses, strains, and

displacements.

The equations of compatibility (8.9) In plane polar coordinates

take the form

6r Or r9 1, rr z _ _
3r ' r 7 80 " 3r "

3e e £ 3e H
ee , 99 " rr 1 r9 1 z ,
(27.4)
3r r ' r 39 " r 39 "

3mr,_ m„ , 3m
92 9z 1, rz .
3r r ' r 39

For static problems and vanishing body loads upon Introducing the

stress functions F(r, 9) and G(r, 9) by

i fr . \ »rf 1 32G 1 3G
rr " r 3r 2 ..2 " r 3r39 2 39
r 39 r

32F 1 32G _LiG

99 " ^ 2 r 3r39 " 2 39
3r r

1 32F 1 3F 1ZQ 1 32G

(27.5)
r9 " " 7 3r39 2 39 " r 3r " 2 .Q2
r r 39

i a2? + _Li£ + ii.

9r " " r 3r39 2 39 .2
r 3r

3G 1 3G
rz 3r ' 9z r 39

we see that equations (27.1) are satisfied identically. From (27.3), we

solve for the strains in terms of the stresses and then substitute (27.5)
 -130-

for the stress components, ii we now use the expressions (27.4), we get

^(G-cVG)--2(l-v)b2±^(V2F)
(27.6)

^(G-c2V2G).2(l-v)b2^(V2F)

where

c2 , T(M t itf. ba
K(2\i + K) » • • 2(2u + K)

(27.7)
vV I 1 v2 = _£2 + 1 J.
r är
+ 1_
2
_£2
2A + 2u + < ' 3r r ae

Solutions of the following equations are also solutions of (2 7.6)

(27.8) V2F - 0 , V2(G - CVG) - 0

For the problem of a circular hole In a plate (or cylindrical cavity In an

Infinite solid) subject to a field of simple tension at Infinity, an appro-

priate solution of (27.8) is

F - 7 r2(l - cos 29) + A. log r + (-7 + Aj cos 26

4 12 3
r
(27.9)
A
4
G - [—+ A5K2(r/c)] cos 29
r

where K„ is the modified Bessel function of the second kind and second order,

This set satisfies (27.6) if

(27.10) A. - 8(1 - v)c2A.5

A 3

The remaining four constants A , A?, A , and A are to be determined from the

boundary conditions
 -131-

t ■tQ"m "0 for r ■ a

rr r9 rz

t - 7 (1 + cos 2 6)
rr 2
for r - "
(27.11)
t - - ^7 sin 26
r6 2

m - 0
rz

Here T is the constant tension field at a plane x ■ const at infinity, Fig.

27.1. By use of (27.5) and (27.9), we calculate the conponents of the stress

and couple stress censors

_ A 6A 4A 6A
t - f (1 + cos 26) + -i - (—| + —I J-) cos 26
r r r r

2A 2
*■ « ^? K0(r/c) + (1 + ^K^r/c)] cos 26
r

_ A 6A„ 6A,
teo - | (l - cos 26) - -i + (-^ - -J) cos 26
r r r

2A 2
- -£ [^ K0(r/c) + (1 + ^^(r/c)] cos 26
r
6A 2A 6A
(27.12) t^ - _(I + -2 + _| - ^) 8in 26
r r r

+ — [— Ko(r/c) + (1 + ^-)K1 (r/c) ] sin 26

cr r * 1
r

6A 2A 6A
K - 9in20
ir -t*-t*"T--r
r r r

A . 2 3
♦ -j [d +^2)Ko(r/c) + (tt^i2|^L(r/e)] sin 26
c r r

mrz y sin 29 - -*(«» Ko(r/c) + (1 + ^^(r/c) ] sin 26

r r
2A 2A
4 5 2c
möz - <—J + 'T2' (Ko(r/c) + ^^(r/c)]} cos 26
r
 -132-

Uslng the boundary conditions (27.11), we find that

Ta4(1
T 2 - V
1 • - T2 a-
A, , "A 2
A„ A(1 + F x

2 2 2
Ta 4(1 - v)a b T
(27.13) A, - 0,, ^ v , A
3 2(1 + F) 4 1 + f1

TacF.
Ac - -
(1 + F^K^a/c)

where

(27.14) b2 r,
Fi,8(1_v)^[4 +
a2
^ +
2aK0(a/c)'
^-^
c c 1

Substituting (27.13) into (27.12), we obtain the stress and couple stress

fields. The value of tQQ at the perifery of the circular hole is of great
99

interest. For this we obtain

(27.X5) ^.„X^IJMIJI,

The maximum value of this t„„ occurs at 9 ■ i

v/2.
d9max
3 + F
,(27.16)
ni K-N
t*a
00 max . _ 1 i
^^»sg^.p^

The quantity S so defined is the stress concentration factor. From (27.14)

It is clear that S depends on v, a, b, and c.

The above result, (27.16), was given by Kaloni and Ariman [1967] who

adopted the solution of the same problem for the indeterminate couple stress
2 2
theory given by Mindlln and Tiersten [1962]. If we set b /c «1, we obtain

for F the one given by Mindlin [1963], namely.

 -133-

2 Ko(a/£) •*
(27.17) f|.ba.v)C4 4^^^I_jrl
it X

2 2
where £ is a physical length for a given material. The case of b /c - 1

gives K ■ -2\i which is not acceptable on grounds of uniqueness and physical

reasonableness since K cannot be as great as twice the shear modulus. Other

discussion of these results and comparison of the stress concentration factors

of the indeterminate couple stress theory and micropolar theory are to be

found in Kaloni and Ariman [1967]. Below we reproduce several of their curves
 -134-

28. GALERKIN AND PAPKOVITCH REPRESENTATIONS

A useful mathematical representation for the displacement and micro-

rotation fields can be constructed as follows: Let

E
Xi -^ . T H -^ , Q = xj + X2 + X^ , Q1 ^ (A + 2u -t- <)Q - pT2

(28.1)
Q2 Mu + <)Q - PT2, Q3 MoH-&fY)Q-PjT2-2<, Q^ Y Q - PJT2 - 2<

and let the matrix L« C^-iJ . (i.J " 1,2,...,6) be given by

(28.2)

Q2 + (A+M)XJ

(A+w)X1X2
(X+M )X1X2

Q2 + (X+u)X2
(kflOLI.

(X-(-u)X2X3
0

<X.
-<X. KX,

-KX,
1
(A+^X^ (A+u)X2X3 Q2 + (A+u)X3 -<X, <x.

-<X3 <X2 QA + (ori-ß)X1 (Qtfß)X1X2 (a+ß)X1X3

KX, -KX, (a+ß)X1X2 QA + (cH-ß)X2 (a+ß)X2X3

-ICX, KX, (a+ß)X1X3 (Qri-ß)X2X3 Q4+(a+ß)X3

The equations (22.1) and (22.2) can be expressed in the matrix form

Sandru [1966]
 -135-

u f
l l '
U f
2 2
U f
3 3
(28.3) - -P

*1 s
l
*2 2
L
*3 . 3 ,

For the inverse L of the matrix L. we have, formally,

(28.4)
-1 Jl +
'lj Q1Q3(Q2Q4 rQ

where

N Q {Q Q [(X+ll) <2lX } (i i 2 3)
ii " 3 1 A ' Q4 " i » " * '

Nii " Q {Q Q
1 2 3 " [(a+ß)
Q2 ' <2]X
i-3}' (i
" 4 5 6)
' '

Nij " Nji " QS^-^+^QA +

< i\^i » » ^ Ji U»J " MtD

N N B
lj " ji Q1l-(cri-ß>Q2 + < 1X
i-3Xj-3' i
^ ^ (i j
" A 5 6)
' '
(28.5)

N14 25 36 W
41 ^52 W
63

N
15 " -N51 " N
42 " -N24 - *W3

N
16 - -N61 " N
43 " -N34 - -*W2

N
26 " -N62 " N
53 " -N35 " ^l^h

Consider now
 -136-

u. F
' l "
U- F
2
F
3
(28.6) " N *
ft. F
i
|L F
a
F
*3 . 3 .

If * - Q F and 4» - Q-F through (28.5) and (28.6), we find that

u(x,t) - Q^ □4t1 - [(x + u)D4 - <2]YY • f| -< D3Y " *2

(28.7)

$(x,t) - D9 □,*„ - [(a + |) U - <2]VV • f. 1-V x I'1

-KD
2 3-2

where

D 5 (X + 2u + <)V2 - P -^2 . O (p + <)VZ - p -~

3t at'
(28.8)

D9 H (a + ß + Y)V2 - pj -^ - 2K, D, YV2 - pj —, -2<

3t '

From (28.3), (28.4), and (28.7), it follows that ♦. and *. satisfy the follow-

ing uncoupled equations

□1<D2D44«V)t1 - -pf
(28.9)
D3(D2DA + <2v2)*2 - -pi

If we take < - 0 in (28.9)^ we obtain a representation known in the classical

theory of elasticity.

In the static case, we set T - 0 in (28.7) and (28.8) and obtain the

Galerkin representation for the micropolar elasticity, namely.

 u(x) - (X + 2M + <)V2(YV2 - 2K)* - [Y(A ♦ ii)V2 - <{2\ + 2w + <)]

2
7V • * - <((a + ß + Y)^ " 2icl7 x *2

(28.10)
$(x) - (u + ic)V2[(a + ß + Y)V2 - 2<]*2 - [(u + <)(a + ß)V2 - K2]VV • *2

- <(X + 2y + K)V2(7 x *1)

where :. and !,, satisfy the equations

4 2
(X + 2u + <)V [(y + IC)YV - <(2u + <)]*, - -pf

(28.11)
[(o + ß + Y)V2 - 2<][(u + <)YV4 - <{2u + K)]*2 - -pi

For K - 0. (28.10)1 gives Galerkln's representation of classical elasticity.

We decompose the body force and body couple Into Irrotatlonal and

solenoldal parts as follows:

pf - vrio + v x n
(28.12)
* *
pi - Yn0 + v x n

and note that

( D2 □4 + <2V2)*1 - VA0 , D^ - 7 x A
(28.13)

(D2 □4 + <2v2)*2 - VAJ , D3*2 - Y x A*

From (28.7) and (28.9) It now follows that

* *
(28.14) D1Ao - -Ho , n3A o - -no

and we obtain
 -138-

u - VA0 + 7 x ( D,A) - <V x (v x A )

(28.15)

| . -KV x (V x A) + V x (O^A )
2

Thus, if we determine AQ, A and A for a given problem, the displacement

and microrotation fields can be calculated from (28.15).

Another useful decomposition can be deduced from (28.11) with

f « J. ■ Q by selecting *, and *? as the sum of three special vector

functions. This has the form

2X+ 2
u - A, + A„ - y A0 +
1 "2 4(X+2u+
U+<) ^ h '' K(2P+IC) ~- ^1 *2;

(28.16) + 27 x B, - T-4- V x B
-1 2u+< - -J

i • W • ■ + ?B - W <^ VV • B + 77 • B + — V x A + ^ S 7x4
I YY ?2 3 <(2u+<) •- ?3 — -1 2 - n K * ~2

where AQ, A1, A«, ^, B., and B satisfy

72A0.O. 7^.0, (1 - ^2; fe) V2)A2 - 0

(28.17)

These results when if ->■ « reduce to the correspo-ding decomposition for the

indeterminate couple stress theory obtained by Mindlin and Tiersten [1962].

 -139-

29. MICROPOLAR INFINITE SOLID SUBJECT TO

CONCENTRATED FORCE AND CONCENTRATED COUPLE

The problem of an Infinite solid subject to a concentrated force F

and a concentrated couple C at the origin of coordinates x ■ 0 is of fund-

amental interest. In classical theory, this problem is known as Kelvin's

problem. This problem for the static case was treated by Sandru [1966],

(i) Concentrated Force. Let F be a concentrated force acting at

the origin of the coordinates x - 0. We write

(29.1) pf - F5(x)

where 6(x) is the Dirac delta function.

The solution of (28.12) is given by

no(x) - - 4- /pf(Ov(7)dv(0
«47i ^ - ~ r
(29.2)

n(x) - - ^ MU) * ?(7)dv(p

where

1/2
(29.3) r | [(x1 - e,^2 + (x2 - C2)2 + (X3 - CJ2]

Using (29.1) we find

F
ä • ?4>
(29.4)
1 - - ^ r - v(±)

where

2 1/2
(29.5) R ; (xj2 + x^2 + xp
 -140-

From (28.13). „ for the static case we get

(A + 2y + <)72Ao m
^ I ' ify
(29.6)

[Y(U + <)V4 - <(2u + <)]A - ^7 ? x

y(|)

These equations possess the solutions

1 F • x
AO 8TT(A+2U+<) R
(29.7)
A x
^2 F
- I AsSi I (FR) +
rsfesi v * tr (i - e"RM) ]

where

(29.8) I1 1 Tfe t;)

(2u + <)<

With A - 0, (28.15) gives for the static case

5X-h6y+3< X 2X+2\i+K \ ' j{

y
" 8TT(2u+<)(X+2m-<) R 8Tt(2u+^)(A+2u+<) D3 •
K

(29.9) +_I_rV. {V . [i(e-R/%]l

For K - Y « 0, (29.9)1 reduces to the well-known solution of Kelvin's problem,

Love, [1944, p. 185]. The solution of this problem for the Indeterminate

couple stress theory Is obtained by letting «: - <*. This gives the result

obtained by Mindlln and Tiersten [1962].

(11) Concentrated Couple. Let M be a concentrated couple acting at

x » 0. We write

pi - M6(x)
 -141-

In this case, we express the solution of (27.12) in the same way as in (28.2).

Using6 (28.13)„2,3- for the static case we have

2
[(a + ß + Y)V - 2<]AJ - -If

(29.10)
[yiu ♦ <)v4 - <(2u + K)V2]A • -n*

where

(29.11) nj - - -^ M • v(|) . n* - - -^ M x V(i)

Equations (29.10) have the solutions

(29.12)
* 1 £2 M
-R/L
87T(2U+<)< - - 47r(2u+<)< ~ R

where

2 , a + B + Y
(29.13) 1
' 2.

Substituting these results into (28.15) for the static case, and with AQ = 0

and A > 0, we obtain

1 M
-R/Ä,
R/fc
V x [^ (1 - e )]
" 4Tr(2u+<)
(29.14)
M

These results may be used in (20.23) and (20.24) to obtain the stress and

couple stress fields.

It is worth noting here that the concentrated couple is a fundamental

problem which is not deduced as the limiting solution of two equal parallel
 -142-

forces directed In opposite sense. We note that In micropolar elasticity,

the concept of body couple Is totally Independent of the force and It can

exist even when the body force Is absent. For micropolar elasticity, there-

fore, force and moment singularities will have totally different natures.

This Is then expected to affect uniqueness theorems for Infinite and for

the finite regions considerably.

 -143-

SUMMARY

The theory of micropolar elasticity presented here should provide

an adequate background for pursuing analytical work and starting a badly

needed experimental program. We believe the theory Is well-posed with field

equations, boundary and Initial conditions. Certain wide classes of unique-

ness theorems have been proved though not presented here (cf. ErIngen [1966a]).

The Important implications of the theory are brought to the surface especially

In connection with the problems In the field of wave propagation. Existence

of additional waves over those existing In classical elasticity should be

attractive to workers In the field of experimental wave propagation. The

linear theory Is simple enough to lend Itself to the solution of non-trivial

boundary and initial value problems. With the additional internal degrees

of freedom provided by the microrotatlon and spin Inertia, it incorporates

the problems Involving concentrated body and surface couples into the funda-

mental singularities of the field.

The field is rather new and not even partly explored. Experimental

works are badly needed. Nevertheless, it seems to us that the logical founda-

tion of the theory is solid and promising for the understanding of the mechanics

of micropolar solids.
 -144-

POSSIBLE DIRECTIONS FOR FUTURE RESEARCH

The theory of micropolar elasticity presented here is one of the

simplest extensions of the classical theory of elasticity for the treat-

ment of materials with microstructure. Extensions of this theory to

micropolar fluids and viscoelasticity exist (cf. Eringen [1964], (1966b],

[196/]). A nonlinear theory is also contained in the more general theory

of microelasticl'v given by Eringen and Suhubi [1964a,b]. The theory of

micromorphic materials, of which the theory of microelasticity is a repre-

sentative field for the oriented solids, possesses promise for entering

the granular and molecular world of materials from the continuum side. It

should not be surprising if such a theory would be in wide use for the full

description of material properties of composites, granular and fiberous

solids. The connection of these theories to continuum dislocation theory

has already been exhibited. The intimate ties between the plasticity theory

and continuum dislocations has been recognized for some time by seme research

workers in this field (cf., Kondo [1962], [1963], Kröner [1963], Bilby [1960],

Bilby, Gardner and Stroh [1967]). Existing theories of miccmechani-s, multi-

polar theories, and the continuum theory of dislocations have not been

amalgamated into a unified structure as yet, although some attempts exist

in this direction. Presently, serious efforts are being made to bring some

order to the world of the mlcrocontinuum. The field cf micromorphic materials

needs and deserves attention from both theoretical and experimental workers.

Especially, the need for rational experiments is felt bad2y. The theory of

micropolar elasticity is certainly ready for such -. test The theory of

micromorphic materials with Its logical structure nd wide possible applica-

tions opens new and promised rich lands fcr fut .""e workers.
 -US-

ACKNOWLEDGEMENT

This work was supported by the Office of Naval Research. The

author is indebted to Dr. Harold Liebowltz for his encouragement and

enthusiasm. Acknowledgement is due also to my students Messrs. Charles

B. Kafadar and William D. Claus Jr. for their care in proofreading and

to my secretary Mrs. Susan Lynch for the excellent typing.

 -146-

REFERENCES

Aero, E. L., and E. V. Kuvshinskii [I960]: "Fundamental equations of the

theory of elastic media with rotationally interacting particles."
Fiziko Tverdogo Tela 2. 1399-1409; Translation, Soviet Physics Solid
State 2, 1272-1281 (1961).

Bilby, B. A. [I960]: "Continuous Distributions of Dislocations " Progress

in Solid Mechanics I, edited by J. N. Sneddon and R Hill, Chap. VII,
331-398. North Holland Publishing Company, Amsterdam.

Bilby, B. A., L. R. T. Gardner, and A. N. Stroh [1937]: "Continuous Distri-

butions of Dislocations and the Theory of Plasticity " IXe Congres
International de Mechanique Appliquee. Extrait dee Acts, Tome VIII,
35-42.

Cosserat, E., and F. [1909]: "Theorie des Corps Deformables " A Hermann
et Fils, Paris.

Eringen, A. C. [1962]: "Nonlinear Theory of Continuous Media " McGraw-Hi.U

Book Co., New York.

Eringen, A. C. [1964]: "Simple Microfluids." Iru. ±. Engng. Sei . 2, 205-217

Eringen, A. C. [1964]: "Mechanics of Micromorphic Materials." Prcieedings of

the Eleventh International Congress of Applied Mechanics Munich,
Germany. Ed. by Henry Görtier, Springer-Verlag, 1966, pp. 131-138.

Eringen, A. C. [1966a]: "Linear Theory of Micropolar Elasticity " J_ Math

Meeh.. 15, 909-924.

Eringen, A. C. [1966b]: "Theory of Micropolar Fluids " J. Math Mech , 16,

1-18.

Eringen, A. C. [1966c): "A Unified Theory el TkftU -nechanical Materials" Int

J. Engng. Sei.. 4, 179-202,

Eringen, A. C. [1967]: "Theory of Micropolar Plates " 2- App] Math Fh^t

(ZAMP) , 18, 12-30.

Eringen, A. C. , and E. S. Suhubi [1964a]: "Nonlinear Theory of Simple Micro-

elastic Solids-I." Int.. J_. Engng. Sei. . 2, 189-203
 ■147-

Eringen, A. C. , and E. S. Suhubi [196Ab]: "Nonlinear Theory of Microelasnic

Solids-II." Int. J. Engng. Sei.. 2, 389-404.

Green, A. E., P. M. Naghdt, and R. S. Rivlin [1965]: "Directors and Multi-

polar Displacements in Continuum Mechanics." Int. J^. Engng. Sei.. 2,
611-870.

Green, A. E. and R. S. Rivlin [1964]: Multipolar Continuum Mechanics." Arch.

Rat. Mech. Anal.. 17, 113-147.

Grioli, G. [I960]: "Elasticita Asimmetrica." Ann. di Mat- Pura ed Appl..

Ser. IV, 50, 389-417.

Günther, W. [1958]: "Zur Statik und Kinematik des Cosseratschen Kontinuums."

Abhandlungen der Braunschweigischen Wissenschaftlichen Gesellschaft.
10, 195-213.

Kaloni, P. N., and T. Ariman [I960]: "Stress Concentration Effects in Micro-

polar Elasticity." J. Appl. Math. Phys■ (ZAMP), 18, 136-141.

Kondo, K. [1962]: "Non-Holonomlc Structures of Plastic Manifolds." RAAG

Memoirs III, 91-133.

Kondo, K. [1963]: "Non-Riemannian and Flnslerian approaches to the theory

of yielding." Int.. J. Engng. Sei. . 1, 71-88.

Kroner, E. [1963a]: "On the physical reality of torque stresses in continuum

mechanics." Int. J_- Engnft. Sei.. 2, 261-278.

Kroner, E. [1963b]: "Dislocation: A New Concept in the Continuum Theory of

Plasticity." J. Math Phys. 42, 1, 27-37.

Love, A. E. H. [1944]: A Treatise on the Mathematical Theory of Elasticity.

Fourth ed.. New York: Dover.

MaeCullagh, J. [1839]: "An Essay towards a Dynamical Theory of Crystalline

Reflection and Refraction." Trans. Roy. Irish Acad. Sei., 21, 17-50.

Mindlin, R. D. [1963]: "Influence of Couples Stresses on Stress Concentra-

tions." Experimental Mechanics. 10, 1-7.
 -148-

Mindlin, R. D. [1964]: "Micro-structure in Linear Elasticity." Arch. Rat.

Mech. Anal.. 16, 51-78.

Mindlin, R. D., and H. F. Tiersten [1962]: "Effects of Couple-Stresses in

Linear Elasticity." Arch. Rat. Mech. Anal., 11, 415-448,

Palmov, V. A. [1964]: "fundamental Equations of the Theory of Asymmetric

Elasticity." PMM. 28, 401-408.

Parfitt, V. R., and A. C. Eringen [1966]: "Reflection of Plane Waves from

the Flat Boundary of a Micropolar Elastic Halfspace." General
Technology Corp. Report No. 8-3.

Sandru, N. [1966]: "On some problems of the linear theory of asymmetric

elasticity." Int. J.. Engng. Sei.. 4, 81-96.

Schäefer, H. [1962]: "Versuch einer Elastizitatstheone des zweidimensionalen

ebenen Cosserat-continuums." Miszellaneen der angewandten Mechanik.
Akademie-Verlag, 277-292.

Schijve, J. [1966]: "Note on Couple Stresses." J. Mech. Phys. Sol. . 14, 113-120

Sokolnikoff, 1. S. [1956]: Mathematical Theory of Elasticity. Second ed..

New York: McGraw-Hill Book Co.

Spencer, A. J. M. [1966]: "Treatise on Continuum Physics I." Theory of

Invariants, edited by A. C. Eringen, Chap. 3. To be published by
Academic Press .

Thomson, W. (Lord Kelvin), and P. G. Tait [1867]: Treatise on Natural Philosophy.

First ed., Oxford University Press.

Toupin, R. A. [1962): "Elastic Materials with Couple-Stresses." Arch. Rat.

Mech. Anal.. 11, 385-414.

Toupin, R. A. [1964]: "Theories of Elasticity with Couple-Stress." Arch

Rat. Mech. Anal.. 17, 85-112.

Truesdell, C, and R. A. Toupin [I960]: "The Classical Field Theories."

Handbuch der Physik ni/1. Berlin: Springer-Verlag
 -149-

Voigt, W. [1887]: Theoretische Studien über die Elasticitätsverhältnisse

der Krystalle. Abh. Ges. Wiss. Göttingen 34. [1894]: Über Medien
ohne innere Kräfte und eine durch sie gelieferte mechanische Deutung
der Maxwell-Hertzschen Gleichungen. Gott. Abh. 72-79.
 -150-

LIST OF SYMBOLS

a Acceleration vector
k
a, , Third-order microdeformation rate tensor
kfcm
b, „ Second-order microde formation rate tensor
k£
c Wave speed

c Micropolar elastic wave speeds

CL , C Spatial and material deformation tensors

d Differential operator as in dx

da Area element

dv, dV Spatial and material volume elements

d. „ Deformation rate tensor

— i ( ) Material derivative operator

e. . , E,,T Spatial and material strain tensors

kÄ. KL
f. Body force per unit mass

h Heat source per unit mass

i, ., I Spatial and material microinertia tensors

j, J Jacobians

I, Body couple per unit mass

m. Couple stress tensor

m, ,, m, ,, Surface couple
K
(n)' (n)k
n. , n Exterior normal vector to a surface

q. , q Heat flux vector directed outward of the surface

r, , RK Spatial and material macrorotarion vectors

r.., K Spatial and material macrc : .'. . ticn tensors

M, S Spatial and material surfaces

l Time
 -151-

Stress tensor
hül
Surface traction
UnV* (n)k
U Spatial and material displacement vectors
\' K
v, V Spatial and material volumes

v v Velocity vector
' k
w
k
Vorticity vector

W
k£
Vorticity tensor

?. \ Spatial rectangular coordinates

X. X,. Material rectangular coordinates

a Micropolar elastic constant

a, a. Microacceleration vector
•• k
ß Micropolar elastic constant

Y Micropolar elastic constant

ViK o■m » ^t w
KLM
Third-order microstrain tensors
6 , 6 Kronecker deltas (»1 when indices take the same number,
zero othervise)

E Internal energy density per unit mass

e. ., E Second-order microstrain tensors

Sam Permutation symbol (e^ - t^ - t^ ■ -t^ - -e^ - -e^-1

and zero otherwise)

n Entropy density per unit mass

6 Absolute temperature

•c Micropolar elastic constant

A, u Elastic constants

v, v Gyration vector

v Gyration tensor

£,, i. Spatial relative position vector

E, 5 Material relative position vector

 -152-

p Mass density

0, o, Intrinsic spin vector

$, $ Spatial microrotation vector

(> , * Microdlsplacement tensors

*, <t Material microrotation vector

K
\» ^vi. Spatial microdisplacement vectors

Xi/» Xuv Material microdisplacement vectors

i|/ Helmholtz free energy

iji. „, f— Microdeformation tensors

iC Ar KL

1, II, III Invariants of tensors

7 Gradient operator
2
V Laplacian operator

Q Wave operator

Cartesian tensor notation is used where indices take values 1,2,3,

Repeated indices indicate summation over the range 1,2,3 unless other-

wise stated. Indices following a comma indicate partial differentiation,

e.g., x. = 3x, /3X^.

A superposed dot indicates time rate with material
K
3x
point fixed, e.g., * ^ TT I x ■

Indices enclosed within parenthesis and brackets indicate symmetric

and antisymmetric parts, e.g.,

t
iki)z 2 U
ki + C
tk) ' E[k£r I (e
k)l ~ Cik)
To convert into expanded engineering notation use

x1 - x , x2 - y , x3 - z

u. - u , u_ « u , u, ■ u
I x' 2 y 3 z

II xx x 12 xy xy
where o , x ,... are the conventional stress components sometimes used

in engineering literature.
 FIGURE CAPTIONS

Flg. 1.1 Mass density versus volume

Fig. 2.1 Material and spatial coordinates

Fig. 2.2 Deformation of a mlcrovolume

Fig. 3.1 Displacement vectors

Fig. 4.1 Microrotation

Flg. 5.1 Deformation of a rectangular parallelepiped

Fig. 5.2 Average rotation

Fig. 5.3 Macrodeformatlon and mlcrodeformatlon

Fig. 5.4 Mlcrodeformatlon with 5 ■ fixed

Fig. 5.5 Mlcrodeformatlon with X ■ fixed (minl-deformatlon)

Fig. 9.1 Uniform microdelatatlon

Fig. 9.2 Uniaxlal strain

Fig. 14.1 Macrovolume element with forces acting on microelements

Fig. 14.2 Macroelement with equipollent force and couple

Fig. 14.3 Extrinsic surface loads on a macrosurface Aa

Fig. 14.4 Equipollent force and couple on Aa

Fig. 14.5 Surface and body loads

Fig. 15.1 Micromass elements containing ralcromass elements

Fig. 16.1 Surface loads

Fig. 16.2 A tetrahedron with surface loads

Fig. 16.3 Stress tensor

Fig. 16.4 Couple stress tensor

Fig. 24.1 Longitudinal displacement and microrotatlon waves

Fig. 24.2 Coupled transverse vector waves propagating with speeds v. and v,
2
Fig. 24.3 Sketch of v„ versus M
 2 2
Flg. 24.4 Sketch of v» and v, versus H

Flg. 25.1 Reflection of a longitudinal displacement wave

Fig. 27.1 Cylindrical hole In tension field

Fig. 27.2 Stress concentration factors for — ■ 0.20, v « 0

Fig. 27.3 Stress concentration factors for — - 0.20, u ■ 0.5

c
Fig. 27.4 Stress concentration factors for — ■ 0.10,v ■ 0

Fig. 27.5 Stress concentration factors for — - 0.10, v - 0.5

Molecular Classical
Micro-
&
Atomic
continuum
^r continuum

AV AV

FIG. 1.1. MASS DENSITY VERSUS VOLUME

FIG. 2.1. MATERIAL AND SPATIAL COORDINATES
FIG. 2.2. DEFORMATION OF A MICROVOLUME
FIG. i. I . DISPLACI ' . .<! \.< i "
FIG. 4.1. MICROROTATION
 c2dx2

FIG. 5.1. DEFORMATION OF A RECTANGULAR PARALLELEPIPED

FIG. 5.2. AVERAGE ROTATION
 dJ((C,) < dX+t

^dEK - dZ

(C^^d^ ^ dx + dy

FIG. 5.3. MACRODEFORMATION AND MICRODEFüRMATION

dx3i3

2+C2)dX2

FIG. 5.4 MICRO DEFORMATION W'TH = HELD CONSTANT

FIG. 5.5. MICRODEFORMATION WITH X - HF.LD 'JONSTANT (MIN'I-DEFOICIATION)
 DdX I
3 -3

DdX2l2

FIG. 9.1. UNIFORM MICRODELATATION

X,.x
^ A

FIG. 9.2. UN1AXIAL STRAIN

FIG. 14.1. MACROVOLUME ELEMENT WITH
FORCES ACTING ON MICROELEMENTS
FIG. 14.2. MACROELEMENT WITH EQUIPOLLENT
FORCE AND COUPLE
FIG. 14,3. EXTRINSIC SURFACE LOADS ON
A MACROSURFACE Aa
FIG. 14.A. EQUIPOLLENT FORCE AND COUPLE ON Aa
m, ,
(n)

FIG. 14.5. SURFACE AND BODY LOADS

FIG. 15.1. MACRÜMASS ELEMENTS CONTAINING M1CRÜMASS ELEMENTS
FIG. 16.1. SURFACE LOADS
FIG. 16.2. A TETRAHEDRON WITH SURFACE LOADS
 > t
22

FIG. 16 3. STRESS TENSOR

FIG. 16.A. COUPLE STRESS TENSOR
 Direction of propagation

Plane longitudinal wave of

speed v.

Plane microrotation wave of speed v.

FIG. 24.1. LONGITUDINAL DISPLACEMENT AND

MICROROTATION WAVES
 Direction of propagation

Microrotation wave
Displacement wave

FIG. 24.2. COUPLED TRANSVERSE VECTOR WAVES

PROPAGATING WITH SPEEDS v0 AND v,
3 4
 ♦ Cr

FIG. 24.3. SKETCH OF v^ VERSUS u

FIG. 24.4. SKETCH OF v^ AND v^ VERSUS u
3 4
 zz

'* rt

FIG. 25.1. REFLECTION OF A LONGITUDINAL DISPLACEMENT WAVE

 X^
—T

<
7* '

FIG. 27.1. CIKCULAK HOLE IN 1ENS10N FIELD

 Classical theory
3.0 _^:

2.0 ..
Couple stress theory

1.0 .

If) 12 a,

FIC. 27.2, STRESS CONCENTRATION FACTORS FOR - 0.20, v - Ü

c
 Classical theory
3.0 - Si

Couple stress theory

2.0

1.0 ..

-t—

2 10 12 a/c

FIG. 27.3. STRESS CONCENTRATION FACTORS FOR - = 0,20, - 0.5

c
». *

Classical theory
3.0

2.0 .
Couple stress theory

1.0

—■»—

10 12 a/c

FIG. 27.4. STRESS CONCENTRATION FACTORS FOR - 0.10, = 0

c
S
c ♦

Classical theory
3.0

Couple stress theory

2.0

1.0 .,

4 6 10 12 a/c

FIG. 27.5. STRESS CONCENTRATION FACTORS FOR - - 0.10. u = 0.5

c
 UNCLASSIFIED
Security Classification
DOCUMENT CONTROL DATA • R&D
(Security claamiticmlion oi itlle body of mbntrmcl *nd indeKtng mnnotmtion muat be entered when the overmll report te cleeailied)
I ORiniNATIN G ACT|V/|TV (Corpormle mulhor) 2« HCPO^T JECU«IT¥ C L ASSIFICA TION
Unclassi fied
2b GROUP
Princeton University

3 REPORT TITLE

THEORY OF MICROPOLAR ELASTICITY

4 DESCRIPTIVE NO'ES (Type ol report end mclutiye detee)
Research Report
S A{JTHOR(S) (Lmet nmme llntnmme inlltel)

Eringen, A. C.

6 REPORT DATE 7m TOTAL f OF PACE» 7b NO OF RtFS

June 196 7 194 39
8a CONTRACT OR GRANT NO 9« ORIGINATOR'» RTPORT NUMBER'S)

N00ni4-6 7-A-0151-0004
b PROJEC T NO No. 1
064-410
9b OTHER REPORT NO(S) (Any other numbere thml may be metltned
thie report)

10 AVAILABILITY L IMIT A TION NOTICES

Qualified requesters may obtain copies of this report from DDC.

11 SUPPL EMENTARY NOT« 12 SPONSORING MILITARY ACTIVITY

Office of Naval Research
For inclusi "i in Treatise on Fracture.
Washington, D. C.
to be published by Academic Press, ed.
by H. Liebowitz.
13 ABSTRACT

This article presents a self-contained account of the recent theory of

micropolar elasticity. Micropolar elastic materials possess extra independent
degrees of freedom for the local rotations different from the rotations of the
classical elasticity. These materials respond to spin inertia and body and
surface couples, and as a consequence they exhibit certain new static and dyna-
mic effects, e.g., new types of waves and couple stresses.
Extensive discussions are presented on deformatio.., strain, microstrain,
rotations, microrotations, kinematics, and balance laws. The thermodynamics
of micropolar solids is formulated and the consequences of entropy and inequality
are discussed. Field equations, boundary and initial conditions are obtained.
The indeterminate couple stress theory is shown to result as a special case of
the theory. Several static and dynamic problems are solved on the subjects of
reflection of various types- of micropolar waves, surface waves, stress concen-
tration around a circular hole, and force and moment singularities In Infinite
solids.

DD FORM
1473 UNCLASSIFIED
Security Classifirahon
 UNCLASSIFIED
Securitv Classificati.
LINK A LINK B LINK C
KEY WORDS
ROLt

Micropolar elasticity
Cosserat continuum
Multipolar theory
Oriented media
Continuum with directors

INSTRUCTIONS
1. ORIOINATINQ ACTIVITY: Enter the name and address imposed by security classification, using standard statements
nl the < ^mractor, suhconlrai tor. grantee, Department of De- such as:
fea.se artiviiv or other organi button f rorpnrato author) issuing
(1) "Qualified requesters may obtain copies of this
thr i ep' r».
report from DDC. "
Ja KFPOR P SFCUMTY I LASSIFICATION: Enter the over-
(2) "Foreign announcement and dissemination of this
all security classification if Ikt report. Indu ate whether
report by DDC is not authorized."
"Rjitncte't Datn" is included. Marking is to be in accord-
ance with appropriate security regulations. (3) "U. S. Government agencies may obtain copies of
this report directly from DDC. Other qualified DDC
2h i .HOUP: Automatic downgrading is spec fied in DoD Di-
users shall request through
rective S200. 10 and Armed Forces Industrial M.inu..! F.nter
th»- group numb"! Alsn, when applicable, show thai optional
n,H king*: have been used for flrrtup 3 *nd Group 4 as author-
i •»■ i
(4) "V. S. military agencies may obtain copies of this
report directly from DDC Other qualified users
KPTOPI riTI.F, Fnter the complete report title in all shall request through
• i 'i.i in«, i Titles in all cases should be unclassified.
■ i i i' I iiile c annot be selected without classifica-
*i ■> .hov iitl* I'lnssifu atmn in all capitals in parenthesis (5) "All distribution of this report is controlled. Qual-
■ .»■ f illowin,. the title. ified DDC users shall request through
i. ÖVSTRIHTIVE NOTES If appropriate, enter the type of
u-port, »■.■.*,,. intcitm, progress, summary, annual, or final. If the report has been furnish-d to the Office of Technical
(itvt t-w inclusive 'l-it*-'. when a specific reporting period is Seivices, Department of Commerce, for sale to the public, indi-
red* cate this fact and enter the price, if known.
** AUT llOI.'iSl Enter Ihe name(s) of authorfs) as shown on 11. SUPPLEMENTARY NOTES: Use for additional explana-
•• i" Ihe n-, ■ rt. F.n'ei last name, first name, middle initial. tory notes.
*t n lit.».», show Link fni branch of servic»*. The name of
the ;.iin. ipat «'thnr is an af.solute minimum requirement. 12. SPONSORING MILITARY ACTIVITY: Enter the name of
the departmental protect office or laboratory sponsoring (pmy
'. REPORT DATÜ Enter the date of the report as day, infi tot) the research and development. Include address.
m ifth. vejr, or ncinth, year. If more than one date appears
on 'he nvort, us« dati of publk Hlion. 13 ABSTRACT: Enter an abstract giving a brief and factual
summary of the do« ument indicative of the report, even though
.. TOI AL NUMBER OF PACKS The total page count
it may also appear elsewhere in the body of the technical re-
sh ulil f'.IIow nori,,il pagination procedures, i.e., enter the
port If additional space is required, a continuation sheet shall
oumbt-r <.( pages containing information. be attached
7h rJLMllEk OF REFERENCES! F.nter the total number of It is highly desirable that the abstract of classified reports
i. fotcncei < iled in the report. be unclassified Each paragraph of the abstract shall end with
Ho CONTRACT OR ORANT NUIBER If appropriate enter an indication of the military security classification of the in-
't.e apcl i.-abli' number of the contract or grant under whu h formation in the paragraph, represented as (TS). (S), (C). or (U)
Ih • nt oil was AiimrL There is no limitation on the length of the abstract. How
14 II \. *<t PROJEC T NUMBER: Fnter the appropriate ever, the suggested length is from 150 to 225 words.
u.ilil.irr department identifi< ation, su< h as protect number,
sulp-oiect number system numbers, task number, etc. 14 KEY WORDS: Key words are technically meaningful terms
or short phrases that characterize a report and may be used as
'*«. ORIGINATOR'S REPORT NUMBER(S) Enter the offi- index entries for cataloging the report Key words must be
( >.<! r -port number t>v which the document will be ir.entified selected so that no security classification is required Identi-
and i ■•p'lMIe i li\ the iiginating activity. This number must fiers, such as equipment model designation, trade name, military
be unirfi«- In (his report. protect code name, geographic location, may be used as key
Oh OTHKR REPOR1 N"MHI-.R(Si ll the report has been words but will be followed by an indication of technical con-
ipm ■( tny other r.-port numbers feithet h\ tbf nnfimatnr text. The assignment of links, roles, and weights is optional.
'if f>y th* «pofis'-rl, also enter this number(s).
10 AVAILABILITY/LIMITATION NOTICES Enter any lim-
it" ' r. ,ri further dissemination of the report, olh-r than those

roRM
0D r< 1473 (BACK) I'NiCLASSIFIED
Security Classification