## Are you sure?

This action might not be possible to undo. Are you sure you want to continue?

GEOMECHANICS

Marian IVAN

University of Bucharest

Department of Geophysics

6 Traian VUIA str.

70138 BUCHAREST o.p.37

ROMANIA

e-mail : ivam@gg.unibuc.ro

ivam@math.math.unibuc.ro

Samizdat Press

2

In memoriam

Professor Marin DOROBANÞU

INTRODUCTION

These course notes present some theoretical problems related to the Mechanics of a Continuum Solid Body, of particular

importance to Applied Geomechanics, Geological Engineering and Structural Geology. In most cases, only static aspects are

discussed, but some dynamic cases are also presented.

As a rule, the modern tensorial approach is used. Some basic elements are reviewed at the beginning of the notes. More details

can be found into the excellent books of George Backus and Brian Kennett, also available at Samizdat Press. The linear

elasticity and the homogeneity of the continuum solid body are almost thoroughly assumed to be valid, but some elements of

Rheology are also presented.

In most cases, the semi-inverse method is used to solve the problems. According to it, the solution is supposed to be of a

particular form, as a consequence of the simplified hypothesis previously assumed. It is verified that solution checks both the

corresponding equations and the boundary conditions. Based on the Uniqueness Theorem of the Linear Elasticity, it follows the

assumed particular solution is just the general solution of the problem. In all the cases discussed here, the assumed simplified

hypotheses allow one to obtain simple, analytical solutions. At a first glance, the importance of such solutions is minor with

respect to the real cases, where mainly the non-homogeneity of the medium plays a great role. However, the analytical solutions

are the basis for deriving finite element algorithms, allowing one to model satisfactory the complex real cases. Such examples

are also presented.

The author will kindly appreciate any critical remarks on these notes, being very grateful to Samizdat Press for the possibility

to shear his work.

3

CONTENTS

INTRODUCTION

page

Chapter A: BASIC ELEMENTS. ………………………………………………………………………………………….... 5

A.1) The displacement vector. Lagrangean (material) and Eulerian (spatial) co-ordinates. …………………………………..5

A.2) Invariants of a tensor. Tensor deviator. ………………………………………………………………………….………..5

A.3) Strain tensor. Stress tensor. Equation of motion / equilibrium. …………………………………………………………..5

A.4) HOOKE’s law. ………………………………………………………………………………………………………...…..8

Chapter B: DEFORMATION OF A CYLINDRICAL BODY IN THE PRESENCE OF GRAVITY ……………....…9

B.1) The model… ……………………………………………………………………………………………………………….9

B.2) Equations of equilibrium. Boundary conditions. Simplifying hypothesis. ……………………..……………………..…..9

B.3) The final shape of the body. ………………………………………………………………………………………….…..12

Chapter C: LÉVY’s PROBLEM - the triangular dam……………………………………………………………………. 14

C.1) The SAINT-VENANT’s equations. …………………………………………………………………………………..…..14

C.2) The model. Simplifying hypothesis. The planar deformation state. ……………………………………………………...14

C.3) Equations of equilibrium. AIRY’s potential. ……………………………………………………………………………..15

C.4) Boundary conditions. The final shape of the dam..…………………………………………………………………..…..16

Chapter D: KIRSCH’s PROBLEM - the circular bore hole / tunnel………………………………………………..……19

D.1) The model. ……... ……………………………………………………………………………………………………..…19

D.2) The planar state of deformation in cylindrical co-ordinate system. ………………………………………………….…..20

D.3) The circle of MOHR. ………………………………………………………………………………..……………………21

D.4) AIRY’s potential in cylindrical co-ordinates. The bi-harmonic equation.…………………………………………….… 21

D.5) The divergence of a tensor in cylindrical co-ordinates. …………………………………………………………………..22

D.6) The gradient of a vector and the strain tensor in cylindrical co-ordinates. …………………………………………..….23

D.7) The bi-harmonic equation cylindrical co-ordinates. ……………………………………………………………….…….24

D.8) The stress elements. Conditions at infinity for the stress elements. ….………………………………………..…………25

D.9) Strain and displacement vector. Conditions at infinity. ………………………………………………………….………27

D.10) Boundary conditions for the stress elements on the wall of the circular cavity. ………………………………….…….28

D.11) The final shape of the wall. ………………… ………………………………………………………………….………31

Chapter E: BOUSSINESQ’s PROBLEM - concentrated load acting on an elastic semi-space………………………....33

E.1) The equations of BELTRAMI and MITCHELL. …………………………………………………………………………33

E.2) The model. …………………………………………………………………………………………………..………….…35

E.3) The equations of equilibrium and strain tensor in spherical co-ordinates. …………….…………………………………35

E.4) LAPLACE operator in spherical co-ordinates. LEGENDRE’s polynomials. …………………………………….………36

E.5) The displacement field. ………………………………………………………………………………….…..……………37

E.6) Boundary conditions for the stress elements. The final solution. …………………..…………………………….………39

Chapter F: ELEMENTS OF THIN PLATE THEORY…………………………………………………………………....41

F.1) The model of a thin elastic plane plate…………………………………………………………………………………….41

F.2) The planar state of a plate. The bending state………………………………………………………………………….….41

4

F.3) Loads acting on the plate…………………………………………………………………………………………………..42

F.4.) Odd and even functions for the planar state and for the bending state. ………………………………………………….42

F.5) Mean value of a function. Equilibrium equations for thin plates. ………………………………………………………..43

F.6) Thin plate in the bending state. ……………………………………………………………………………………….…..45

F.7) BERNOULLI’s hypothesis. …………………………………………………………………..……………………….…..45

F.8) HOOKE’s law for a thin plate. ……………….……………………………………………………………………….…..46

F.9) The infinite, 1-dimensional (1-D) plate. The flexure of the lithosphere. …………..………………………………..…..47

F.10) Exterior forces on the lateral surface of the plate. Buckling. …………………………………………………………....48

F.11) The buckling of a simply leaning thin plate. ………………………………………………………………………..…..50

F.12) The infinite extended 1-D plate. ……………………………………………………………………………….…….…..51

F.13) FOURIER transforms. Properties. ………………………………………………………………………………..….…..51

F.14) Solution of the flexure equation by using FOURIER transforms. ………..………………………………………….…..52

F.15) Finite plates. ……………………………………………………………………………….………………………..…..55

a) Significance of Σ

ij

and Μ

ij

for the bending state. …….……………………………………………………..……55

b) The rectangular plate. Boundary conditions. LÉVY’s solution. …….………………………………………..………55

F.16) Vibrations of a plate laying on a viscous substratum. …………………………………………………………..………57

a) The differential equation. …………………………………………………………………………………………...…57

b) The rectangular plate with 3 embedded sides. ……………………………………………………………………...…58

Chapter G: THE SPHERICAL SHELL……………………………………………………………………………………. 60

G.1) The model. BERNOULLI’s hypothesis. Displacement vector and strain tensor. ………………………………..………60

G.2) Quasi-mean values. Equation of motion. ……………………………………………………………………………..….61

G.3) Integrals of the stress elements. Quasi-moments. ………..…………………………………………………………...….62

G.4) Integrals of displacement vector. ……………………………………………………….……………………………..….62

G.5) Equation of motion in quasi-mean values. ……………….…………………………………………………………...….63

G.6) Quasi-mean value of the shell density. The differential equation. ……..……………………………………………..….63

G.7) The buckling of a spherical shell. ……………..……………………………………………………………………...….64

G.8) Load on the upper face. Stress on the lower surface of the shell. ……………………………………………………..….66

G.9) The differential equation of time dependent flexure. ………..………………………………………………..………….67

G.10) Spherical effects with respect to the plane plate. ………….………………………………………………………....….68

Chapter H: ELEMENTS OF RHEOLOGY………………………….…………………………………………………..…69

H.1) Introduction. …………………………………………………………………………………………………………..…..69

H.2) Linear models. …………………………………………………………………………………………………….………69

a) KELVIN-VOIGT (strong viscous) model. …………………………………………………………………………….69

b) MAXWELL (viscous-elastic) model. ………………………………………………………………………….………70

c) BURGERS (general linear) model. …………………………………………………………………………………….70

d) Remarks on the linear models. …………………………………………………………….…………………………..71

H.3) Non-linear models. …………….……………………………………………………………………………………….…73

H.4) Brittle. Creep. Empirical criteria. ………..……………………………………………….………………………………73

H.5) Empirical criteria for shear-faulting. TRESCA criterion. COULOMB-NAVIER criterion. ………………………….....74

H.6) Von MISES-HENCKY criterion for ductile flow (plasticity).. ………………………..…………………………………76

H.7) Rheological models. ……………………………………………………………………………………………..……….77

a) SAINT VENANT body (elastic-plastic material). …………………………………………………………………….77

b) BINGHAM body (visco-plastic material). …………………………………………………………………………….77

Chapter I: THE ACCRETION WEDGE………………………………………………………………………………..…. 78

I.1) The model. …………………………………………………………………………………………………………………78

I.2) Equations of equilibrium. Yield condition. Stress field. ……….……………………………………………………….…78

I.3) Boundary conditions. Final results. ……..…………………………………………………………………………………79

References…………………………………………………………………………………………..…….81

5

A) BASIC ELEMENTS.

A.1) The displacement vector. Lagrangean (material) and Eulerian (spatial) co-ordinates.

Consider an arbitrary material point inside a continuum body, subject to a deformation process. At the initial time t

0

, that

point has the position vector denoted by X

→

, with respect to the origin of a co-ordinate system (Fig.A1). At a time t t ≥

0

, the

new position vector is to be x

→

. The difference x X

→

−

→

represents the displacement vector. Taking into account that the

components ( ) x x x

1 2 3

, , of the vector x

→

are all functions of the components ( ) X X X

1 2 3

, , of X

→

, the displacement

vector can be written as

( ) U U X X X t

→

·

→

1 2 3

, , , . (a1)

This represents a Lagrangean (material) description of the deformation process. Here, ( ) X X X

1 2 3

, , are representing the

Lagrangean (material) co-ordinates. Alternately, the components ( ) X X X

1 2 3

, , of X

→

can be seen as functions of the

components ( ) x x x

1 2 3

, , of the vector x

→

. Consequently, the displacement vector can be written as

( ) u u x x x t

→

·

→

1 2 3

, , , . (a2)

This represents a Eulerian (spatial) description of the deformation process, where ( ) x x x

1 2 3

, , are the Eulerian (spatial)

co-ordinates. A basic supposition assumed thoroughly in that notes is that the deformation process is a continuous one, i.e. all

the components of u

→

or U

→

are continuous functions together their derivatives with respect to both their spatial co-ordinates or

to time.

The Lagrangean co-ordinates are usual in the Solid Mechanics, while the Eulerian co-ordinates are commonly used in Fluid

Mechanics. However, in the Linear Elasticity, the distinction between these two kinds of co-ordinates is not important, as it will

be seen in the next chapters. More details on such aspects can be found in (Aki and Richards 1980; Ranalli 1987).

Fig.A.1. The continuum deformed body and the displacement vector.

A.2) Invariants of a tensor. Tensor deviator.

A second order tensor represents mainly a 3x3 matrix. The elements of the tensor are changing according to a certain rule

with respect to a change of the co-ordinate system. Such a change with respect to a rotation will be discussed later. For

simplicity, only symmetric tensors will be considered. A symmetric tensor is equal to its transpose

T T

t

· (a3)

(or T

ij

T

ji

· ). The superscript “t” shows the transposed tensor (matrix).

Let the components of the tensor be real numbers

6

T

T T T

T T T

T T T

·

¸

¸

_

,

11 12 13

12 22 23

13 23 33

, (a4)

The scalar λ and the vector u

→

are representing the eigen-value and the eigen-vector respectively of that tensor if

T

u u u

→

·

→ →

≠

→

λ , 0 . (a5)

It’s easy to see that the eigen-values are not changing with respect to a rotation of the co-ordinates system.

Suppose now that the eigen-value λ and the components of the eigen-vector u

→

are complex numbers. By taking the complex

conjugate (denoted by an asterisk) into (a5), it follows

T

u u

* *

→

·

→

λ (a6)

Taking into account the symmetry of the tensor, the next inner product is evaluated into two different ways

<

→

→

>·<

→

→

>·

→

T

u u u u u ,

*

,

*

λ λ

2

, (a7)

and

<

→

→

>·<

→

→

>·<

→

→

>·<

→

→

>·

→

T T T

u u u

t

u u u u u u ,

*

,

*

,

*

,

* * *

λ λ

2

(a8)

From (a7) and (a8) it follows that the eigen-values (and the components of the eigen-vectors) of a symmetric tensor are real

numbers.

Eq.(a5) can be written as

T

u u −

¸

¸

_

,

→

·

→ →

≠

→

λ

1

0 0 , , (a9)

where

1

denotes the unit tensor. From (a9) it follows that the next determinant vanishes

T T T

T T T

T T T

11 12 13

12 22 23

13 23 33

0

−

−

−

·

λ

λ

λ

(a10)

Hence the eigen-values are the roots of the third degree equation

− + − + · λ λ λ

3

1

2

2 3

0 I I I , (a11)

where

( )

( )

I T T T

I T T T T T T T T T

I T T T T T T

tr

T

T

1 11 22 33

2 11 22 22 33 33 11

12

2

23

2

13

2

3 11 22 33 12 23 13

· + + ·

· + + − − −

· + + ·

,

,

... det

(a12)

With respect to a rotation of the co-ordinates system, the elements of the tensor are generally changing. Because the quantities

defined by (a12) can also be expressed as functions of the roots of (a11), it follows their values are not changing with respect to

a rotation. They represent the main invariants of the tensor. The first invariant is the trace of the tensor, while the third one

represent just its determinant.

The tensor defined by

( )

*

T T tr

T

· −

1

3

1

, (a13)

7

represents the tensor deviator, having its trace equal to zero. Elementary computations show its second invariant is

( ) ( ) ( ) I T T T T T T T T T

2

1

6

11 22

2

22 33

2

33 11

2

6

12

2

23

2

13

2 *

· − − + − + − + + +

¸

¸

_

,

¸

1

]

1

(a14)

That invariant is especially important to define constitutive equations for plasticity.

A.3) Strain tensor. Stress tensor. Equation of motion / equilibrium.

By using the spatial co-ordinates, the strain tensor is defined as

ε

·

→

+

→

¸

¸

_

,

1

2

grad u grad

t

u . (a15)

where “grad” denotes the gradient. In Cartesian co-ordinates, that symmetric tensor has the elements

( )

ij

u

i

x

j

uj

x

i

u

i, j

u

j i ε

∂

∂

∂

∂

· +

¸

¸

_

,

· +

1

2

1

2

,

. (a16)

According to the CAUCHY’s hypotheses there are two kinds of forces acting at an arbitrary point placed inside a body or on

its boundary. The first ones are represented by the mass forces, characterised by a mass density b

→

. For the problems

discussed in that book, such mass forces are ignored. Or, they are represented by the gravity, when b

→

is just the gravitational

acceleration g

→

. Suppose now a mechanical state of tension (stress) is present inside the deformed body, e.g. as a result of the

action of a pair forces t

→

T . An arbitrary cross section is considered through a certain point of the body, dividing it into a part

denoted by D

l

at left and a pert D

r

at the right respectively (Fig.A.2). A surface element dS is considered on the boundary

of D

l

, having the outer pointing normal vector denoted by n

→

. The material points of the boundary of D

r

are acting on dS

by an elementary force df

→

. It follows (e.g. Aki and Richards 1980; Ranalli 1987) that the next relation is valid

df

dS

n

→

·

→

σ

, (a.17)

where the tensor

σ

represents the CAUCHY stress tensor, spatial co-ordinates being used. According to the Principle of

the Kinetic Momentum Balance, stress is a symmetric tensor. It can be shown too that the Principle of Impulse Balance leads to

the next equation of motion /equilibrium

div b

d u

dt

σ

ρ ρ +

→

·

→

2

2

, (a18)

That equation is valid at an arbitrary point inside the body , where ρis the density and

d u

dt

2

2

→

represents the acceleration. By

projecting eq.(a18) on the co-ordinates system axes, three scalar equations are obtained.

Fig.A.2. An imaginary cross section through the deformed body.

8

A.4) HOOKE’s law.

Neglecting the initial stress (in most cases), it is further assumed a linear relation between the stress and strain tensors, i.e.

σ ε

·

H

, (a19)

or

ij ijkl kl

H σ ε

· , (a20)

where

H

is a fourth-order tensor. Eq.(a19) represents HOOKE’s law. In the usual cases discussed here, an elastic,

homogeneous, isotropic medium is considered. Then eq.(a19) takes the particular form

σ ε ε

λ µ · + tr

1

2 . (a21)

Here, tr denotes the trace of the tensor,

1

is the unit tensor(matrix) and λ µ , are the elastic coefficients of LAMÉ. Hence

( )

( )

( )

11 11 22 33

2

11

22 11 22 33

2

22

33 11 22 33

2

33

12 21

2

12 13 31

2

13 23 32

2

23

σ ε ε ε ε

σ ε ε ε ε

σ ε ε ε ε

σ σ ε σ σ ε σ σ ε

λ µ

λ µ

λ µ

µ µ µ

· + + +

· + + +

· + + +

· · · · · ·

,

,

, ,

, (a.22)

Alternately, HOOKE’s law ( a21) can be reversed to give

ε σ σ

ν ν

·

+

−

1

1

E E

tr , (a23)

where the modulus of YOUNG is

E ·

+

+

µ

λ µ

λ µ

3 2

(a24)

and the transverse contraction coefficient of POISSON is

ν

λ

λ µ

·

+ 2( )

. (a25)

By reversing (a24) and (a25), it follows

( )( ) ( )

λ

ν

ν ν

µ

ν

·

+ −

·

+ 1 1 2

1

2 1

E E , . (a26)

The parameter defined by

χ

λ µ

ν

·

+

·

−

3 2

3 3 1 2

E

( )

(a27)

represents the incompressibility or bulk modulus. For (theoretical) incompressible rocks, that modulus approaches infinity.

Other constitutive equations will be discussed in relation to the rheological bodies.

9

B) DEFORMATION OF A CYLINDRICAL BODY IN THE PRESENCE OF

GRAVITY

B.1) The model.

An elastic homogeneous isotropic body is considered (Fig.B1). Its initial shape is a right, vertical, very thin cylinder of

radius equal to r and height equal to H. The base of the body is placed on the horizontal, absolutely rigid, plane x Ox

1 2

.

The deformation of the body due to its own weight follows to be studied and the final shape of the body into the final

equilibrium state will be found. The approximations of the linear theory are assumed and the variation of the density is ignored.

The problem is solved by following the next steps:

i) - the equations of equilibrium are used, the unknowns here being the components of the stress tensor

σ

; these equations

are processed according to the simplifying hypothesis of the problem;

ii) - by using the reversed HOOKE’s law, the equations of equilibrium are processed in order to have only the components of

the strain tensor

ε

as unknowns;

iii) - by using the definition of the strain tensor, the components of the displacement vector

→

u

are obtained and the final

shape of the body is found.

Fig.B.1. (a) A vertical cylinder lying on a rigid plane; (b) The final shape of a vertical cross section (solid line) with respect to

the initial shape (dashed line). [NO SCALE]

B.2) The equations of equilibrium. Boundary conditions. Simplifying hypothesis.

A simplified approach can be derived by using cylindrical co-ordinates. However, the problem here is an introductory one. So

these co-ordinates will be used later, in relation to other problems. The equations of equilibrium in Cartesian co-ordinates are

111 12 2 13 3

0

12 1 22 2 23 3

0

131 23 2 33 3

0

, , ,

, , ,

, , ,

σ σ σ

σ σ σ

σ σ σ

ρ

+ + ·

+ + ·

+ + − · g

(b1)

Here, δ is the density and g is the gravitational acceleration. The forces acting upon the body are the reaction force of the

horizontal plane and the gravity of the cylinder. The boundary conditions are:

-on the lateral surface of the cylinder:

σ

n for x H x x

→

·

→

∈ ∈ 0

3

0

1 2

, [ , ], , Γ (b2)

-on the upper base of the cylinder

σ

n for x H x x

→

·

→

· ∈ 0

3 1 2

, , , ∆ , (b3)

Here, ∆ is the disc of radius equal to r , having the centre at the origin of the co-ordinate system and the boundary denoted by

Γ. The outer pointing normal at the lateral surface of the body is a linear combination with variable coefficients of the

horizontal unit vectors, i.e.

n C x x C x x

e e

→

·

→

+

→

1 1 2

1

2 1 2

2

( , ) ( , ) (b4)

10

For x H x x

3

0

1 2

∈ ∈ [ , ], , Γ eq.(b2) becomes

C x x C x x

e e e e e e

1 1 2

11 1 12 2 13 3

2 1 2

12 1 22 2 23 3

0 ( , )[ ] ( , )[ ]

σ σ σ σ σ σ

→

+

→

+

→

+

→

+

→

+

→

·

→

(b5)

The outer pointing normal at the upper base of the body is the unit vector

→

e

3

. For x H x x

3 1 2

· ∈ , , ∆ , eq. (b3) gives

13 1 23 2 33 3

0

σ σ σ

→

+

→

+

→

·

→

e e e

(b6)

Eq. (b5) is satisfied if the stress tensor has the form

σ

σ

¸

1

]

1

·

¸

¸

_

,

0 0 0

0 0 0

0 0

33

(b7)

on the lateral surface of the body.

Because the cylinder is a very thin one, the stress at its inner points is approximately the same one to the stress on the lateral

surface. So, it is assumed that eq.(b7) holds inside the whole volume of the body. It follows eqs.(b1a)-(b1b) are identical

verified. From eq. (b1c) it follows that

∂

∂

ρ

σ

σ

33

3

33

1 2 3

0

x

g x x x H · · · , ( , , ) (b8)

The problem represented by eq.(b8) has the next immediate solution

33

1 2 3 3 σ

ρ ( , , ) ( ) x x x g x H · − (b9)

The reversed HOOKE’s law is

( )

[ ]

( )

[ ]

( )

[ ]

11

1

1

11 11 22 33

22

1

1

22 11 22 33

33

1

1

33 11 22 33

12

1

12 23

1

23 13

1

13

ε σ σ σ σ

ε σ σ σ σ

ε σ σ σ σ

ε σ ε σ ε σ

ν ν

ν ν

ν ν

ν ν ν

· + − + +

· + − + +

· + − + +

·

+

·

+

·

+

E

E

E

E E E

( )

( )

( )

, ,

(b10)

Because

ij

i

x

j

j

x

i

u

u

ε

∂

∂

∂

∂

· +

¸

¸

_

,

1

2

, (b11)

eqs. (b10) lead to

∂

∂

νρ

∂

∂

νρ

∂

∂

ρ

∂

∂

∂

∂

∂

∂

∂

∂

∂

∂

∂

∂

1

1

3

2

2

3

3

3

3

1

2

2

1

0

2

3

3

2

0

3

1

1

3

0

u u u

u u u u u u

x

g

E

H x

x

g

E

H x

x

g

E

x H

x x x x x x

· − · − · −

+ · + · + ·

( ) , ( ) , ( )

, ,

(b12)

By integrating eq.(b12) it follows that

11

1

3 1 1 2 3

2

3 2 2 1 3

3

3

2

2

3 3 1 2

1

2

2

1

0

2

3

3

2

2

3

1

1

3

1

u

u

u

g

E

H x x f x x

g

E

H x x f x x

g

E

x

Hx f x x

f

x

f

x

f

x

f

x

g

E

x

f

x

f

x

g

E

x

· − +

· − +

· − +

+ · + · + ·

νρ

νρ

ρ

∂

∂

∂

∂

∂

∂

∂

∂

νρ ∂

∂

∂

∂

νρ

( ) ( , ) ,

( ) ( , ) ,

( ) ( , ) ,

, ,

(b13)

Hence the displacement field is found if the unknown functions f f f

1 2 3

, , are finally obtained. Differentiating (b13e) with

respect to x

1

and (b13f) with respect to x

2

and adding the results, it follows

2

2

3

1 2 3

1

2

2

1

0

∂

∂ ∂

∂

∂

∂

∂

∂

∂

f

x x x

f

x

f

x

+ +

¸

¸

_

,

· . (b14)

So, using eq.(b13d) it follows

∂

∂ ∂

2

3

1 2

0

f

x x

· (b15)

From eq.(b15) it follows that

f x x h x h x

3 1 2 1 1 2 2

( , ) ( ) ( ) · + , (b16)

where h h

1 2

, are two unknown functions, following to be found. Eqs.(b13e) and (b13f) give

∂

∂

νρ ∂

∂

νρ f x x

x

g

E

x

dh x

dx

f x x

x

g

E

x

dh x

dx

2 1 3

3

2

2 2

2

1 2 3

3

1

1 1

1

( , ) ( )

,

( , ) ( )

· − · − (b17)

The left side of eq.(b17a) is represented by a function depending on x x

1 3

, only, while the right side is a function of x

2

.

Hence both sides are equal to a constant, i.e.

∂

∂

νρ f x x

x

a

dh

dx

g

E

x a

2 1 3

3

2

2

2

2 2

( , )

, · − · + (b18)

It follows that

f x x a x g x h x

g

E

x a x b

2 1 3 2 3 2 1 2 2

2

2

2

2 2 2

( , ) ( ) , ( ) · − + · + +

νρ

(b19)

In a similar manner, eq.(b17b) gives

f x x a x g x h x

g

E

x a x b

1 2 3 1 3 1 2 1 1

2

1

2

1 1 1

( , ) ( ) , ( ) · − + · + +

νρ

(b20)

From eq. (b13d) it follows that

dg x

dx

dg x

dx

K

1 2

2

2 1

1

( ) ( )

· − · (b21)

where Kis a constant. Then

g x Kx C g x Kx C

1 2 2 1 2 1 1 2

( ) , ( ) · + · − + (b22)

For simplicity, material co-ordinates are used to obtain the final expression of the displacement field

12

→ →

·

¸

¸

_

,

·

−

−

− + +

¸

¸

_

,

+

− −

− +

+

¸

¸

_

,

+

+

¸

¸

_

,

u

u

u

u

X

g

E

H X X

H X X

X

HX X X

a X KX

a X KX

a X a X

C

C

b b

( )

( )

( )

( )

1

2

3

3 1

3 2

3

2

2

3

2

1

2

2

2

1 3 2

2 3 1

1 1 2 2

2

1

1 2

ρ

ν

ν

ν

(b23)

The first term in eq.(b23) is the true displacement, the last one is a translation while the second term is the rigid rotation

a

a

K

X

2

1

−

¸

¸

_

,

×

→

(b24)

B.3) The final shape of the body.

a) The final shape of the upper base

Consider an arbitrary point of co-ordinates equal to ( , , ) X X X H

1 2 3

· . In the initial stage, it is placed on the upper base of

the cylinder. Finally, the co-ordinates of the point are

x

x

x

X

X

H

g

E

X X

H

1

2

3

1

2

0

0

2

1

2

2

2

2

2

¸

¸

_

,

·

¸

¸

_

,

+

+ −

¸

¸

_

,

ρ

ν

( )

(b25)

Hence

x X

x X

x

g

E

X X

g

E

H H

1 1

2 2

3

2

1

2

2

2

2

2

·

·

· + − +

¹

'

¹

¹

¹

¹

¹

νρ ρ

( )

(b26)

From eq.(b26) it follows that

x x X X r

1

2

2

2

1

2

2

2 2

+ · + · . (b27)

Hence the circle representing the contour of the upper base remains a circle of the same radius. The plane of the circle is

moving downward by a quantity equal to ρ ν g H r E ( ) / ( )

2 2

2 − . The surface of the disc representing the upper base of the

body is no longer a plane one. It becomes a rotational parabolic surface having the equation

x

g

E

x x

g

E

H H

3

2

1

2

2

2

2

2

· + − +

νρ ρ

( ) (b28)

b) The final shape of the lower base

Consider now an arbitrary point initially placed on the lower base of the body. The point has the co-ordinates equal to

( , , ) X X X

1 2 3

0 · . The final co-ordinate of the point are

x

x

x

X

X

g

E

HX

HX

X X

1

2

3

1

2

0

1

2

1

2

2

2

2

¸

¸

_

,

·

¸

¸

_

,

+

+

¸

¸

_

,

νρ

( ) /

(b29)

Hence

13

x gH E X

x gH E X

x

g

E

X X

1

1

1

2

1

2

3

2

1

2

2

2

· +

· +

· +

¹

'

¹

¹

¹

¹

¹

( / )

( / )

( )

νρ

νρ

νρ

(b30)

From eq.(b30) it follows that

x x gH E X X gH E r

1

2

2

2

1

2

1

2

2

2

1

2 2

+ · + + · + ( / ) ( ) ( / ) νρ νρ , (31)

i.e. the circle representing the contour of the lower base remains a circle. The new radius is increased by a quantity equal to

νρgH E / . The initial horizontal plane of the circle is uplifted by a quantity equal to νρgr E

2

2 / ( ) . The surface of the disc

representing the lower base becomes a rotational paraboloid having the equation

x

g

E gH E

x x

3

2 1

2

1

2

2

2

·

+

+

νρ

νρ ( / )

( ) (b32)

c) The final shape of the lateral surface

Consider now a point initially placed on a generatrix line of the cylinder. Because of the cylindrical symmetry of the problem,

the point having the initial co-ordinates equal to ( , , ) X X r X

1

0

2 3

· · is considered. Finally, that point has the position

characterised by the co-ordinates

x

x r

g

E

H X r

x X

g

E

X

HX

g

E

r

1

0

2 3

3 3

3

2

2

3

2

2

·

· + −

· + −

¸

¸

_

,

+

¹

'

¹

¹

¹

¹

¹

¹

¹

¹

¹

¹

¹

νρ

ρ νρ

( ) (b33)

From (b33), it follows that the generatrix remains into the initial vertical plane. Its shape is changed from a straight line

segment to a convex parabolic segment, having the equation

( ) ( ) x

E

g

x r

E

g

x r H

g

E

H

g

E

r

3

2

2

2

1

2

2

1

2

2

2

2

· − − − + − +

ν ρ

νρ

ρ νρ

/ / (b34)

OBSERVATION. On the lower base of the cylinder it is acting the reaction force of the rigid plane, equal to the weight of the

body. When the surface of the base is decreasing, approaching the paraboloid of eq.(b32), the normal unit effort (equal to the

weight divided by the contact area) is increasing. At a certain moment, its magnitude will exceed a yielding value of the

material. Then, HOOKE’s law, valid in the elastic domain, will be no longer appropriate here.

14

C) LÉVY’s PROBLEM - THE TRIANGULAR DAM

C.1) The SAINT-VENANT ‘s equations.

Differentiating a certain element of the strain tensor

ij i, j j i

u u ε

· +

¸

¸

_

,

,

/ 2 , (c1)

it follows, for example, that

( ) ( )

( ) ( ) ( )

11 22 2211 11

22

2 2

11

1122 2 211

1 2

12

21

12

1 2 21

12

2

1212

, , ,

,

,

,

, ,

,

,

,

,

, ,

,

,

ε ε

ε

+ · + · +

· + · + ·

u u u u

u u u u

(c2)

Hence

11 22 22 11

2

12 12 , , ,

ε ε ε

+ · (c3)

Also,

22 33 33 22

2

23 23 , , ,

ε ε ε

+ · (c4)

3311 1133

2

3131 , , ,

ε ε ε

+ · (c5)

In a similar way, it follows that

( )

( )

( )

12 3 231 31 2

2

22 31

231 31 2 12 3

3

3312

31 2 12 3 231

1

11 23

, , ,

,

,

, , ,

,

,

, , ,

,

,

ε ε ε ε

ε ε ε ε

ε ε ε ε

+ − ·

+ − ·

+ − ·

(c6)-(c8)

The above equations (c3)-(c8) represent the SAINT-VENANT’s equations of compatibility.

C.2) The model . Simplifying hypothesis. The planar deformation state.

A horizontal dam of infinite length is considered. The cross-section is represented by a rectangular triangle OAB (Fig.C1).

The length of the base is AB=l and the height is OA=h. On OA catheter is acting the hydrostatic pressure of a liquid (water)

having the specific weight equal to γ. As a result, the dam is deformed. The dam is represented by an elastic homogeneous,

isotropic material. Its specific weight is equal to Γ and its elastic constants are E and ν.

Fig.C1. A vertical cross section through the dam. N.Hs. is the free surface of the water, acting on OA side by a pressure linearly

increasing with depth.

Because the shape of the dam, the displacement vector has the components like

15

1 1

1 2

2 2

1 2

3

0

u u

u u

u

x x

x x

·

·

·

¹

'

¹

¹

¹

¹

¹

( , )

( , ) (c9)

It follows the strain tensor components are like

( )

( )

( )

11 11 11

1 2

12

1

2

1 2 2 1 12

1 2

13

1

2

1 3 31

0

22 2 2 22

1 2

23

1

2

2 3 3 2

0

33 3 3

0

ε ε

ε ε

ε

ε ε

ε

ε

· ·

· + ·

· + ·

· ·

· + ·

· ·

,

( , )

, ,

( , )

, ,

,

( , )

, ,

,

u

u u

u u

u

u u

u

x x

x x

x x

(c10)

Hence the strain matrix is

( )

ε ε

ε ε

ε ε

¸

1

]

1

·

¸

1

]

1

·

¸

¸

_

,

x x

1 2

11 12

0

12 22

0

0 0 0

, , (c11)

It corresponds to a planar state of the strain (the plane here being 1-2).

The components of the stress tensor are

( )

( )

( ) ( ) ( ) ( )

11 11 22

2

11

12

2

12

13

2

13

0

22 11 22

2

22

23

2

23

0

33 11 22

2

33 11 22

2

11 22 11 22

σ ε ε ε

σ ε

σ ε

σ ε ε ε

σ ε

σ ε ε ε ε ε σ σ σ σ

λ µ

µ

µ

λ µ

µ

λ µ λ

λ

λ µ

ν

· + +

·

· ·

· + +

· ·

· + + · + ·

+

+ · +

( )

(c12)

Hence the stress matrix is

( )

( )

σ σ

σ σ

σ σ

σ σ

ν

¸

1

]

1

·

¸

1

]

1

·

+

¸

¸

_

,

x x

1 2

11 12

0

12 22

0

0 0

11 22

, (c13)

Because the component 33 of the stress has a non-zero value, eq.(c13) shows that the stress state corresponding to a planar state

of the strain is not generally a planar one too.

C.3) Equations of equilibrium. AIRY’s potential.

The only body force acting on the dam is its weight. The equations of equilibrium are

111 12 2

0

12 1 22 2

0

, ,

, ,

σ σ

σ σ

+ + ·

+ ·

¹

'

¹

¹

¹

Γ

(c14)

Because the presence of Γ, eqs.(c14) represent a non-homogeneous system. In the beginning, the homogeneous system is

solved, i.e.

16

111 12 2

0

12 1 22 2

0

, ,

, ,

Σ Σ

Σ Σ

+ ·

+ ·

¹

'

¹

¹

¹

(c15)

Using an unknown function ϕ , the first equation of (c15) is verified for

11

2

12

1

Σ Σ

· · −

∂ϕ

∂

∂ϕ

∂ x x

, (c16)

In the same way, the second equation of (c15) is verified for

12

2

12

1

Σ Σ

· · −

∂ψ

∂

∂ψ

∂ x x

, (c17)

It follows that

∂ϕ

∂

∂ψ

∂ x x

1 2

0 + · (c18)

i.e. the unknown functions are

ϕ

∂

∂

ψ

∂

∂

· · −

A A

x x

2 1

, (c19)

The unknown function

A A

x x · ( , )

1 2

represents the AIRY’s potential. It allows one to obtain the next expressions for

the components of the stress tensor when the body force are absent:

11 22 12 12

22

11

Σ Σ

Σ

· · − ·

,

,

,

,

,

A A A

(c20)

From (c13), the trace of the stress tensor can be written using LAPLACE’s operator in 1-2 co-ordinates

( )

tr

A Σ Σ Σ

∆ · + + · + ( ) ( )

*

1

11 22

1 ν ν (c21)

The components of the strain tensor are obtained using the reversed HOOKE’s law

11

1

22 22

1

11 12

1

12

ε ε ε

ν

ν∆

ν

ν∆

ν

·

+

−

¸

¸

_

,

·

+

−

¸

¸

_

,

· −

+

E E E

A A A A A

,

*

,

,

*

,

,

. (c22)

Using (c22) and (c3) it follows

,

*

,

,

*

,

, 2222

22

1111

11

2

1212

A A A A A

−

¸

¸

_

,

+ −

¸

¸

_

,

· − ν ν ∆ ∆ , (c23)

i.e.

( )

* *

1 0 − · ν ∆ ∆

A

(c24)

Because ν < 05 . , it follows that AIRY’s potential is a solution of the bi-harmonic equation

∆ ∆

* *

A

· 0 (c25)

Because the trace of a tensor is an invariant, eq.(c25) holds too in the general case of the orthogonal curvilinear co-ordinates.

However, eq.(c20) has to be modified.

C.4) Boundary conditions. The final shape of the dam.

On the side OA of the dam is acting the hydrostatic pressure. It follows that

σ

γ ( ) −

→

·

→

2

1

2

e e

x (c26)

On the side OB of the dam is acting the negligible atmospheric pressure. It follows that

σ

→

·

→

n

0 (c27)

where the outer pointing normal at the dam is

→

· −

→

+

→

n e e

sin cos α α

1 2

(c28)

On the side OA, for x h x

1

0

2

0 ∈ · [ , ], , it follows that

17

12

0

22

1

σ

σ

γ

·

· −

¹

'

¹

¹

¹

,

x

(c29)

On the side OB it follows for x h x x

1

0

2 1

∈ · [ , ], tan α that

12 11

0

22 12

0

σ σ

σ σ

α

α

− ·

− ·

¹

'

¹

¹

¹

tan ,

tan

(c30)

Eqs.(c29)-(c30) represent 4 boundary conditions, suggesting a solution of the bi-harmonic equation (c25) which depends on 4

unknown coefficients denoted by a b c d , , , , i.e.

A

x x

a

x

b

x x

c

x x

d

x ( , )

1 2

6

1

3

2

1

2

2

2

1

2

2

6

2

3

· + + + (c31)

Using (c20), the solution of the homogeneous system is

11

1 2

12

1 2

22

1 2

Σ

Σ

Σ

· +

· − +

· +

¹

'

¹

¹

¹

¹

¹

cx dx

bx cx

ax bx

( ) (c32)

A particular solution of the non-homogeneous system (c14) is

11 22

0

12

2

σ σ

σ

· ·

· −

¹

'

¹

¹

¹

Γx

(c33)

It follows the general solution of (c14) is

11

1 2

12

1 2 2

22

1 2

σ

σ

σ

· +

· − + −

· +

¹

'

¹

¹

¹

¹

¹

cx dx

bx cx x

ax bx

( ) Γ . (c34)

Replacing (c34) into (c29)-(c30) it follows that

− + − · ∈ ·

+ · − ∈ ·

− + − + − · ∈ ·

+ + + + · ∈ ·

¹

'

¹

¹

¹

¹

¹

( ) , [ , ] ,

, [ , ] ,

( ) tan ( ) , [ , ] , tan

( ) tan , [ , ] , tan

bx cx x for x h x

ax bx x for x h x

cx dx bx cx x for x h x x

bx cx x ax bx for x h x x

1 2 2

0

1

0

2

0

1 2 1 1

0

2

0

1 2 1 2 2

0

1

0

2 1

1 2 2 1 2

0

1

0

2 1

Γ

Γ

Γ

γ

α α

α α

(c35)

It follows that

a

b

c

d

· −

·

· − +

· −

¹

'

¹

¹

¹

¹

¹

γ

γ α

α γ α

0

2

2

3

Γ

Γ

/ tan

/ tan / tan

(c36)

and

11

1 2

12

2

22

1

σ

σ

σ

γ

· +

· −

· −

¹

'

¹

¹

¹

¹

¹

Ax Bx

Cx

x

(c37)

where

A h l B h l h l C h l · − · − · − γ γ γ

2 2

2

3 3 2 2

/ , / / , / Γ Γ (c38)

Hence

11 11

1 1 2 2

,

u

C x C x · · +

ε

, (c39)

where

18

C A A E C B E

1

1

2

1

2

· + − − · − ( )[ ( )] / , ( ) / ν ν γ ν (c40)

It follows that

1

1

1

2

2

2 1 2 1 2 u

C x C x x f x · + + / ( ) (c41)

where the unknown function f

1

follows to be found. In the same manner,

2

3 1 2 4

2

2

2

2 1 u

C x x C x f x · + + / ( ) (c42)

But

( )

12

1

2

1 2 2 1

1

2

2 1 1 2 3 2 2 1

1

12

1

2 ε σ

ν ν

· + · + + + ·

+

· −

+

, ,

(

’

( )

’

( )

u u

C x f x C x f x

E E

Cx (c43)

Hence

C x f x K

C x f x K

2 1 2 1

3 2 1 2

+ ·

+ · −

¹

'

¹

¹

¹

’

( )

’

( )

, (c44)

where Kis an arbitrary constant. It follows

f x C x Kx K f x C x Kx K

1 2 3

2

2

2

2 1 2 1 2

1

2

2

1 2

( ) / , ( ) / · − − + · − + + (c45)

Hence, the displacement field is

1

1

1

2

2

2 1 2 3

2 1

2

2

2

2 1

2

2

1

2

2

3 1 2 4

2

2

2

1 2

u

u

C x C x x C C E x Kx K

C x C x x C x Kx K

· + − + + − +

· − + + + +

¹

'

¹

¹

¹

/ [ ( ) / ] /

/ /

ν

(c46)

The last terms into (c46) represent a rigid roto-translation.

It should be outlined that the above boundary conditions on stress values on the sides OA and OB are not complete ones. As a

result, the unknown constants C C

3 4

, are present in (c46). Boundary conditions on stress values (or displacements) on the

side AB are required in order to obtain an unique solution of the problem

For example, consider the case when the points A and B are fixed ones. It follows

( ) ( )

( ) [ ] { }

1

1

1

2 2

2

2 1 2 3

2 1

2 2

2

2

2

2

1

2

2

3 2 1 2 2 3

2 1 2

1

u

u

C x h C x h x C E x l x

C h x C x x x h l C h C E l x h

· −

¸

¸

_

,

+ − − + + −

· −

¸

¸

_

,

+ − + − + + −

¹

'

¹

¹

¹

/ [ ( )C / ] /

/ / ( )C / / ( )

ν

ν

(c47)

An arbitrary point placed initially on the side AB has the initial co-ordinates ( ) X h X

1 2

· ; . Its final position is

x X X X h C C E X l X

x X X X X C hX l X l

u

u

1 1

1

1 2 3

2 1

2 2

2

2 2

2

1 2 2 3 2 2

· + · + + + −

· + · + −

¹

'

¹

¹

¹

( , ) [ ( ) / ] ( ) /

( , ) ( ) /

ν

(c48)

Elementary computations show that

C

E

C

E

h

l

3

2 1 1

1 2

2

2

+

+

· −

+

− − − −

¸

1

]

1

1

( )

( ) ( )

ν ν

γ ν νΓ ν γ (c49)

If

C

E

C

3

2 1

0 +

+

<

( ) ν

, (c50)

the final shape of the side AB is a concave parabolic segment. Because the possibility of the water to flow below the dam, that

situation is not recommended in real cases. Therefore, it is asked to

γ ν νΓ ν γ ( ) ( ) ( / ) 1 2

2

0 − − − − ≤ h l , (c51)

i.e.

[ ]

h l / ( ) / ( ) / ≥ − − − γ ν νΓ ν γ 1 2 (c52)

For example, assuming that γ ν · · · 1000

3

2400

3

0 25 Kgs m Kgs m / , / , . Γ it follows that h l ≥ 0 29 . .

EXERCISE. Obtain the final shape of the dam in the above hypothesis.

19

D) KIRSCH’s PROBLEM - THE CIRCULAR BORE HOLE / TUNNEL

D.1) The model.

It is assumed that the whole 3-dimensional space is represented by an elastic, homogeneous, isotropic medium, having the

elastic constants denoted by E and ν, respectively λ and µ . A co-ordinate system having the third axis positive upward will

be used. The initial state of stress is represented by the homogeneous tensor

0

σ

, corresponding to a planar state of

deformation, i.e.

0

11

0

12

0

0

12

0

22

0

0

0 0

11

0

22

0

σ

σ σ

σ σ

σ σ

ν

·

+

¸

¸

_

,

¸

¸

_

,

, (d1)

where the components

ij

0

σ

have constant values. The mass forces are ignored, hence the equilibrium equation

div

0

0

σ

·

→

(d2)

is identically satisfied.

Suppose that a circular, infinite bore hole / tunnel is performed along the third axis, its material being instantly removed. The

origin of the co-ordinate system is placed at the centre of the cavity. On the wall of the bore hole is acting now the atmospheric

pressure (or the pressure of the drilling mud), denoted by

0

p

. Consequently, a new (non-homogeneous) stress value is

obtained and the circular shape of the bore hole is changing too. It follows to obtain the new stress, denoted by

f

σ

, and the

new shape of the bore hole in the final equilibrium stage, where

div

f

σ

·

→

0 . (d3)

It is also assumed that the deformation is an elastic one, i.e. the stress perturbation

σ σ σ

· −

f 0

is related to the strain

tensor by

σ ε ε

λ µ · + tr

1

2 . (d4)

The unknown components of the displacement vector are supposed to correspond to a planar deformation state, i.e.

u u x x u u x x u

1 1 1 2 2 2 1 2 3

0 · · · ( , ) , ( , ) , . (d5)

Because the symmetry of the problem, the cylindrical co-ordinate system ( ) r z , , θ will be used, having the unit vectors

denoted by

→ → →

¸

¸

_

,

e e e

r z

, ,

θ

(see Fig.D1).

Fig.D1. The cylindrical co-ordinate system.

D.2) The planar sate of deformation in cylindrical co-ordinate system.

20

With respect to Fig.D1 it follows that

→

·

→

+

→

→

· −

→

+

→

→

·

→

¹

'

¹

¹

¹

¹

¹

¹

¹

¹

¹

e e e

e e e

e e

r

z

cos sin ,

sin cos ,

.

θ θ

θ

θ θ

1 2

1 2

3

(d6)

Hence the matrix for passing from the Cartesian co-ordinates to cylindrical co-ordinates is

Q · −

¸

1

]

1

1

1

cos sin

sin cos

θ θ

θ θ

0

0

0 0 1

(d7)

It represents a rotation of angle equal to θ in a positive (counter clockwise) sense. From (d4) and (d5) it follows that the stress

matrix in Cartesian co-ordinates is

[ ]

crt

σ

σ σ

σ σ

σ σ

ν

·

+

¸

¸

_

,

¸

¸

_

,

11 12

0

12 22

0

0 0

11 22

. (d8)

Let the stress matrix in cylindrical co-ordinates be

[ ]

cyl

rr r rz

r z

rz z zz

σ

σ σ σ

σ σ σ

σ σ σ

θ

θ θθ θ

θ

·

¸

¸

_

,

. (d9)

It follows that

[ ] [ ]

( )

cyl

Q

crt

Q

t

σ σ

σ σ

σ σ

σ σ

θ θ

θ θ

ν

θ θ

θ θ · · −

¸

¸

_

,

+

¸

¸

_

,

− ¸

¸

_

,

cos sin

sin cos

cos sin

sin cos

0

0

0 0 1

11 12

0

12 22

0

0 0

11 12

0

0

0 0 1

(d10)

By performing the computations in (d10), it follows

rr

σ

σ σ

σ

σ σ

θ θ ·

+

+ +

−

11 22

2

12

2

11 22

2

2 sin cos , (d11)

θθ

θ θ

σ

σ σ

σ

σ σ

·

+

− −

−

11 22

2

12

2

11 22

2

2 sin cos , (d12)

rθ

θ θ

σ

σ σ

σ

· −

−

+

11 22

2

2

12

2 sin cos , (d13)

( ) ( )

rz z zz rr

σ σ σ σ σ σ σ

θ

ν ν

θθ

· · · + · + 0

11 22

, . (d14)

D.3) The circle of MOHR.

21

Suppose the Cartesian co-ordinate system is selected in order its axes to be along the first two eigen vectors of the stress

tensor. In that case,

11

σ

and

22

σ

are eigenvalues of the stress tensor and

12

0

σ

· . From equations (d11)-(d13) it follows

that

rr r

σ

σ σ

σ

σ σ

θ

−

+ ¸

¸

_

,

+ ·

− ¸

¸

_

,

11 22

2

2

2 11 22

2

2

, (d15)

and an identical relation obtained by replacing

rr

σ

with

θθ

σ

. Eq.(d15) shows that

rr

σ

and

rθ

σ

are placed on a circle

of radius equal to

11 22

2

σ σ

− / . Suppose now that

rr

σ

(or

θθ

σ

) (i.e. the radial stress component, usually denoted

by

σ

) and

rθ

σ

(i.e. the tangential stress, usually denoted by

τ

) are obtained at various angles θ and the MOHR’s

circle represented by eq.(d15) is obtained. Its radius and its position of the centre allow one to obtain graphically the

eigenvalues of the stress tensor. Further discussion will be presented in relation to the empirical failure criteria of materials.

D.4) AIRY’s potential in cylindrical co-ordinates. The bi-harmonic equation.

Consider the representation of the stress components with the AIRY’s potential in Cartesian co-ordinates, i.e.

11 22 22 11 12 12

σ σ σ

· · ·

−

,

,

,

,

,

A A A

, (d16)

where the AIRY’s potential verifies the bi-harmonic equation

∆ ∆

* *

A

· 0 . (d17)

In the beginning, the derivatives in eq.(d16) will be evaluated by using the polar co-ordinates

x r

x r

r x x

a x x

1

2

1

2

2

2

2 1

·

·

¹

'

¹

· +

·

¹

'

¹

¹

¹

cos

sin

,

tan( / )

θ

θ

θ

(d18)

Than, a representation of the stress components in cylindrical co-ordinates with the help of the AIRY’s potential will be

obtained from (d11)-(d13).

But

∂

∂

θ

∂

∂

θ

∂θ

∂

θ ∂θ

∂

θ r

x

x

r

r

x

x

r x

x

r

r x

x

r

r

1

1

2

2

1

2

2

2

1

2

· · · · · − · − · · cos , sin ,

sin

,

cos

(d19)

It follows

,

cos

,

sin

, 1

1 1 1

A

A A A

A A

x r

r

x x

r

r

· · + · −

∂

∂

∂

∂

∂

∂

∂

∂θ

∂θ

∂

θ

θ

θ

(d20)

In the same way,

,

sin

,

cos

, 2

A A A

r

r

· + θ

θ

θ

. (d21)

Also,

( )

, ,

,

cos

,

sin

,

,

sin cos

,

sin

,

,

cos

cos

,

sin

,

,

,

, , sin , ,

cos

12 1

2

2

2

2

2

2

2

A A A A

A A A A

A

A A A A

r

r

r

r

r

r

r

r

rr

r

r

r

r

r

r

· · −

¸

¸

_

,

· −

¸

¸

_

,

+ −

¸

¸

_

,

· − −

¸

¸

_

,

+ −

¸

¸

_

,

θ

θ

θ

θ θ

θ

θ

θ

θ

θ

θ

θ

θθ θ θ θ

θ

(d22)

In the same way

,

cos

,

sin

, ,

sin cos

, ,

11

2 2

2

2

2

A A

A A A A

rr

r

r

r

r

r

r

· + +

¸

¸

_

,

+ − +

¸

¸

_

,

θ θ

θθ

θ θ

θ θ

, (d23)

22

,

sin

,

cos

, ,

sin cos

, ,

22

2 2

2

2

2

A A

A A A A

rr

r

r

r

r

r

r

· + +

¸

¸

_

,

− − +

¸

¸

_

,

θ θ

θθ

θ θ

θ θ

, (d24)

Hence

11 22 22 11

2

σ σ

θθ

+ · + · + +

, , ,

, ,

A A A

A A

rr

r

r

r

, (d25)

11 22 22 11

2

2 2 2

2

σ σ

θθ

θ θ

θ θ

− · − · − + +

¸

¸

_

,

− − +

¸

¸

_

,

, , ,

, ,

cos sin

, ,

A A A

A A A A

rr

r

r

r

r

r

r

, (d26)

From eqs.(d11)-(d13) it follows

rr

r

r

r

rr r

r

r

r

A A

A

A A

σ σ σ

θθ

θθ θ

θ θ

· + · · − +

, ,

,

,

,

, ,

2 2

. (d27)

D.5) The divergence of a tensor in cylindrical co-ordinates.

In the case of the cylindrical co-ordinates, the square of the elementary arc is equal to

ds dx dx dx dr r d dz

2

1

2

2

2

3

2 2 2 2 2

· + + · + + θ (d28)

Hence the differential parameters of Lamé are

h h r h

1

1

2 3

1 · · · , , , (d29)

the orthogonal curvilinear co-ordinates are equal to

c c c z

1

1

2 3

· · · , , θ , (d30)

and the unit vectors are

→

·

→ →

·

→ →

·

→

n

e

n

e

n

e

r z

1 2 3

, ,

θ

. (d31)

It follows that

∂

∂

β

δ δ

β

h

q

c

q

·

2 1

. (d32)

Substituting the above results in the formula (a2.77) (Ivan 1996) it follows the next formula for the divergence of a tensor in

cylindrical co-ordinates

divT

T

rr

r

T

r

r

T

rz

z

T

rr

T

r

r

T

r

r

T

r

T

z

z

T

r

r

T

rz

r

T

z

r

T

zz

z

T

rz

r

z

e

e e

· + + +

−

¸

¸

_

,

→

+ + + +

¸

¸

_

,

→

+ + + +

¸

¸

_

,

→

∂

∂

∂

θ

∂θ

∂

∂

θθ

∂

θ

∂

∂

θθ

∂θ

∂

θ

∂

θ

θ

∂

∂

∂

θ

∂θ

∂

∂

2

(d33)

D.6) The gradient of a vector and the strain tensor in cylindrical co-ordinates.

Substituting the above results in the formula (a2.68) (Ivan 1996) it follows the next formula for the gradient of a vector in

cylindrical co-ordinates

23

grad u

u

r

r

r r

u

r

r

r

u

r

z

r z

u

r

r

u

r

u

z

z

u

z

r

z r

u

z

r

z

u

z

z

z z

u

r

r

u

r

r

e e e e e e

e e e e e e

e e e e e e e e e e

→

·

→

⊗

→

+

→

⊗

→

+

→

⊗

→

+

→

⊗

→

+

→

⊗

→

+

→

⊗

→

+

→

⊗

→

+

→

⊗

→

+

→

⊗

→

−

→

⊗

→

+

→

⊗

→

∂

∂

∂

∂θ

θ

∂

∂

∂

θ

∂

θ

∂

θ

∂θ

θ θ

∂

θ

∂

θ

∂

∂

∂

∂θ

θ

∂

∂

θ

θ θ θ

. (d34)

The components of the strain tensor

ε

·

→

+

→

¸

¸

_

,

¸

1

]

1

1

1

1

2

grad u grad u

t

are

rr

u

r

r

r

u

r

r

u

r

u

r

rz

u

r

z

u

z

r

u

r

u

r

r

z

u

z

u

z

r

zz

u

z

z

ε ε ε

ε ε ε

∂

∂

θ

∂

∂θ

θ

∂

θ

∂

∂

∂

∂

∂

θθ

∂

θ

∂θ

θ

∂

θ

∂

∂

∂θ

∂

∂

· · − +

¸

¸

_

,

· +

¸

¸

_

,

· + · +

¸

¸

_

,

·

, ,

, ,

1

2

1

2

1

2

. (d35)

For the particular displacement field represented by (d5), the components of the vector u

→

are

u

r

u

r

r u u r u

z

· · · ( , ) , ( , ) , θ

θ θ

θ 0 . (d36)

Hence

rz z zz

ε ε ε

θ

· · · 0 . (d37)

It is the case of a planar state of deformation, i.e.

ε

ε ε

ε ε

θ

θ θθ

·

¸

¸

_

,

rr r

r

0

0

0 0 0

, (d38)

and all the strain elements are functions of r and θ .

Consequently, the stress is

( )

σ

σ σ

σ σ

σ σ

θ

θ θθ

ν

θθ

·

+

¸

¸

_

,

rr r

r

rr

0

0

0 0

, (d39)

all the components of the stress being too functions of r and θ , according to HOOKE’s reversed law. It follows from (d33)

the next equations of equilibrium are obtained in the absence of mass forces:

∂

∂

∂

θ

∂θ

θθ

σ σ σ σ

rr

r

r

r

rr

r

+ +

−

· 0 (d40)

∂

θ

∂

∂

θθ

∂θ

θ

σ σ σ

r

r r

r

r

+ + · 2 0 (d41)

D.7) The bi-harmonic equation in cylindrical co-ordinates.

By using (a2.69) and (a2.43) (Ivan 1996), it follows the LAPLACE operator in cylindrical co-ordinates is

24

∆f div grad f

r r

r

f

r

r

f f

z

· ·

¸

¸

_

,

+ + ( )

1 1

2

2

2

2

2

∂

∂

∂

∂

∂

∂θ

∂

∂

. (d42)

The AIRY’s potential is also a function of r and θ . Hence the AIRY’s potential is the solution of the bi-harmonic equation

∆ ∆

* *

A

· 0 , (d43)

where the LAPLACE operator in polar co-ordinates is

∆

*

·

¸

¸

_

,

+ · + +

1 1

2

2

2

2

2

1 1

2

2

2

r r

r

r

r r

r r

r

∂

∂

∂

∂

∂

∂θ

∂

∂

∂

∂

∂

∂θ

. (d44)

It should be noted that the singular point r · 0 is avoided in (d44) because r R ≥ > 0, where R is the radius of the bore hole.

It follows to solve (d43) by using (d44) in order to derive the AIRY’s potential. The stress components will be obtained from

(d27), imposing the boundary conditions on the wall of the bore hole. The components of the strain will be derived by using the

HOOKE’s reversed law. The displacement vector will be obtained from the definition of strain elements, allowing one to find

the final shape of the deformed bore hole wall.

Consider the FOURIER expansion of the AIRY’s potential, having the coefficients equal to functions of r

[ ] A A A B

r r

n

r n

n

r n

n

( , ) ( ) ( ) cos ( ) sin θ θ θ · + +

·

∞

∑

0

1

(d45)

It follows

∂

∂

θ θ

A

A A B

r

n

n

n

n

n

· + +

¸

¸

_

,

·

∞

∑

0

1

’ ’

cos

’

sin , (d46)

∂

∂

θ θ

2

2

0

1

A

A A B

r

n

n

n

n

n

· + +

¸

¸

_

,

·

∞

∑

’’ ’’

cos

’’

sin , (d47)

( )

∂

∂θ

θ θ

2

2

2

1

A

A B

n

n

n

n

n

n

· − +

·

∞

∑

cos sin . (d48)

Hence

∆ ∆

A A A A A B B B

r

n

r

n

r

n

r

n

n

· + + −

¸

¸

_

,

+ + −

¸

¸

_

,

¸

1

]

1

1

·

∞

∑

0

1

2

2

1

2

2

1

’’ ’

cos

’’ ’

sin θ θ (d49)

and

∆∆ ∆∆

A A A A A A A

B B B B B

r

n

r

n

r

n n

r

n

n

r

n

r

n

r

n n

r

n

· + + −

+

+

+

+

−

¸

¸

_

,

¸

·

∞

+ + −

+

+

+

+

−

¸

¸

_

,

1

]

1

1

∑

0

2 2

2

1

2

2

2

1

3

4

4

2

4

1

2 2

2

1

2

2

2

1

3

4

4

2

4

’’’’ ’’’ ’’ ’

cos

’’’’ ’’’ ’’ ’

sin

θ

θ

. (d50)

By using (d50), the bi-harmonic (d43) is verified if

∆∆

0

0

A

· , (d51)

and if the functions

A B

, are the solutions of the next differential equation

’’’’ ’’’ ’’ ’

Φ Φ Φ Φ Φ

+ −

+

+

+

+

−

·

2 2

2

1

2

2

2

1

3

4

4

2

4

0

r

n

r

n

r

n n

r

. (d52)

Because

0

A

is a function of r only, eq.(d51) is

1 1

0

0

r

d

dr

r

d

dr r

d

dr

r

d

dr

A

¸

¸

_

,

¸

1

]

1

1

¹

'

¹

¹

¹

¹

;

¹

¹

¹

· . (d53)

25

Hence

r

d

dr r

d

dr

r

d

dr

a

d

dr

r

d

dr

a r r b r

A A 1

0

0

0

0 0

¸

¸

_

,

¸

1

]

1

1

·

¸

¸

_

,

· + , ln . (d54)

But

r rdr

r

r const. ln ln · −

¸

¸

_

,

+

∫

2

2

1

2

(d55)

Hence

r

d

dr

a

r r

b

r c

A

0 0

2

2

1

2

0

2

2

0

· −

¸

¸

_

,

+ + ln (d56)

Finally, denoting again the constants, it follows

0

0

2

0

2

0 0 A

r a r r b r c r d ( ) ln ln · + + + . (d57)

In order to solve eq.(52), a solution of the form

Φ

· r

m

(d58)

is considered. Substituting (d58) in (d52), it follows that the exponent mis the solution of the algebraic equation

m m m m m m m n m m n m n n ( )( )( ) ( )( ) ( ) ( ) ( ) − − − + − − − + − + + + − · 1 2 3 2 1 2 2

2

1 1 2

2

1

4

4

2

0,(d59)

having the roots

m n m n m n m n

1 2

2

3 4

2 · − · − + · · + , , , . (d60)

Hence the AIRY’s potential is

A

r a r r b r c r d a

n

r

n

b

n

r

n

c

n

r

n

d

n

r

n

n

n

n

r

n

n

r

n

n

r

n

n

r

n

n

n

( , ) ln ln cos

sin

θ θ

α β γ δ θ

· + + + +

+

+ +

− +

+

−

¸

¸

_

,

·

∞

+

+

+ +

− +

+

−

¸

¸

_

,

·

∞

∑

∑

0

2

0

2

0 0

2 2

1

2 2

1

,(d61)

where the unknown coefficients a

i

b

i

c

i

d

i j j j j

i j , , , , , , , , , , , ... , , , ... α β γ δ · · 01 2 1 2 follows to be obtained.

D.8) The stress elements. Conditions at infinity for the stress elements.

By using (d61) and (d27), it follows

( )

θθ

θ

α β γ δ θ

σ

· · + + −

+ + + + −

−

+ − −

−

+ +

− −

¸

¸

_

,

·

∞

+ + + + −

−

+ − −

−

+ +

− −

¸

¸

_

,

·

∞

∑

∑

,

ln

( )( ) ( ) ( )( ) ( ) cos

( )( ) ( ) ( )( ) ( ) sin

rr

a r b

c

r

a

n

n n r

n

b

n

n n r

n

c

n

n n r

n

d

n

n n r

n

n

n

n

n n r

n

n

n n r

n

n

n n r

n

n

n n r

n

n

n

A 0

2 3 2

0

0

2

2 1 1

2

2 1 1

2

1

2 1 1

2

2 1 1

2

1

(d62)

At great distances from the cylindrical cavity, the elastic perturbation has to vanish, i.e.

r→∞

·

lim

θθ

σ

0 . (d63)

It follows

a b b a

n n

n

b

n n

n

0

0 2

0

2

2

2

2

0 0 1 2

0 3 4

· + + · · · ·

· · ·

, , , , , ...

, , , ...

β α

β

(d64)

Hence the AIRY’s potential is

26

( ) ( )

( )

A

r c r d b r c

n

n

n

n r

n

n

d

n

n

n

n r

n

n

( , ) ln cos sin

cos sin

θ β θ γ θ

θ δ θ

· + + + + +

− +

·

∞

+ +

−

·

∞

∑

∑

0 0 1 1

2

1

1

(d65)

and

( )

( )

θθ

θ γ θ

θ δ θ

σ

· − + − − +

−

·

∞

+ + +

− −

·

∞

∑

∑

c

r

n n c

n

n

n

n r

n

n

n n d

n

n

n

n r

n

n

0

2

2 1

1

1

2

1

( )( ) cos sin

( ) cos sin

. (d66)

From (d65), it follows that

( )

( )

,

( ) cos sin

( ) cos sin

r

c

r

b n c

n

n

n

n r

n

n

n d

n

n

n

n r

n

n

A

· + + + − + +

− +

·

∞

+ − +

− −

·

∞

∑

∑

0

1 1

2

1

1

1

1

β θ γ θ

θ δ θ

, (d67)

( ) ( )

,

sin cos sin cos

θ

θ γ θ θ δ θ

A

n c

n

n

n

n r

n

n

n d

n

n

n

n r

n

n

· − +

− +

·

∞

+ − +

−

·

∞

∑ ∑

2

1 1

, (d68)

( ) ( )

,

cos sin cos sin

θθ

θ γ θ θ δ θ

A

n c

n

n

n

n r

n

n

n d

n

n

n

n r

n

n

· − +

− +

·

∞

− +

−

·

∞

∑ ∑

2 2

1

2

1

. (d69)

Hence

( )

( )

rr

r

r

r

c

r

b

r

n n c

n

n

n

n r

n

n

n n d

n

n

n

n r

n

n

A A

σ

θθ

β

θ γ θ

θ δ θ

· + · +

+

− + −

¸

¸

_

,

+

−

·

∞

− +

¸

¸

_

,

+

− −

·

∞

∑

∑

, ,

cos sin

cos sin

2

0

2

1 1 2

2

1

2 2

1

(d70)

From (d70), it follows that

r

rr

→∞

·

limσ

0 . (d71)

Also,

( ) ( )

,

( ) sin cos sin cos

r

n n c

n

n

n

n r

n

n

n d

n

n

n

n r

n

n

A

θ

θ γ θ θ δ θ · − −

− +

·

∞

+ −

− −

·

∞

∑ ∑

2

1

1

2 1

1

. (d72)

Hence

( ) ( )

r

r

r

r

n n c

n

n

n

n r

n

n

n n d

n

n

n

n r

n

n

A A

θ

θ θ

θ γ θ θ δ θ

σ

· − +

· − − −

−

·

∞

− +

¸

¸

_

,

−

− −

·

∞

∑ ∑

, ,

( ) sin cos sin cos

2

2

1

2 2

1

(d73)

From (d73), it follows that

r

r

→∞

·

lim

θ

σ

0 . (d74)

27

D.9) Strain and displacement vector. Conditions at infinity.

From (d39) and HOOKE’s reversed law it follows that

( ) ε σ σ σ σ σ

ν ν

ν

ν

θθ

· + −

¸

1

]

1

·

+

− +

¸

1

]

1

1

1

1

1 1

E

tr

E

rr

( ) (d75)

i.e.

rr

E

rr

u

r

r

ε σ σ

ν ν

ν

θθ

∂

∂

·

−

−

−

¸

1

]

1

·

1

2

1

, (d76)

Hence

( )

( )( )

E u

r

r

c

r

b

r

n n n c

n

n

n

n r

n

n

n n d

n

n

n

n r

n

n

1

2

0

1

1

2

1 1

1 2

1

2

2

1

1

1

2

1

−

·

−

+

+

− − + +

−

−

¸

1

]

1

+

−

·

∞

−

−

+ +

− −

·

∞

∑

∑

ν

∂

∂ ν

β ν

ν

θ γ θ

ν

θ δ θ

( ) ( ) cos sin

cos sin

(d77)

Integrating (d77) it follows

( ) ( ) ( )

( )

E

u

r

c

r

b r n n c

n

n

n

n r

n

n

n d

n

n

n

n r

n

n

1

2

0

1

1

1 1

2

1

2

1

2

1

1

1

1

−

· −

−

+ + + + +

−

−

¸

1

]

1

+

− +

·

∞

+

−

+

− −

·

∞

+

∑

∑

ν

ν

β

ν

ν

θ γ θ

ν

θ δ θ ϕ θ

ln cos sin

cos sin ( )

(d78)

From

r

u

r

→∞

·

lim

0 , (d79)

it follows that

b

1 1

0 0 + · ≡ β ϕ θ , ( ) . (d80)

Hence

( ) ( )

( )

E

u

r

c

r

n n c

n

n

n

n r

n

n

n d

n

n

n

n r

n

n

1

2

0

1

1

2

1

2

1

2

1

1

1

1

−

· −

−

+ + +

−

−

¸

1

]

1

+

− +

·

∞

+

−

+

− −

·

∞

∑

∑

ν

ν

ν

ν

θ γ θ

ν

θ δ θ

cos sin

cos sin

. (d81)

Finally

( )( )

( )

u

r

E

c

r

n c

n

n

n

n r

n

n

n d

n

n

n

n r

n

n

·

+

− + + − +

− +

·

∞

¸

+ +

− −

·

∞

1

]

1

1

∑

∑

1

0

2 4

1

2

1

1

ν

ν θ γ θ

θ δ θ

cos sin

cos sin

. (d82)

In the same way,

θθ

ν

θθ

ν

ν

∂

θ

∂θ

ε σ σ

·

−

−

−

¸

1

]

1

· +

1

2

1 E

rr

u

r

u

r

r

. (d83)

It follows

E u E

u

r

r r

rr

1

2

1

2

1

−

· −

−

+ −

−

ν

∂

θ

∂θ

ν

θθ

ν

ν

σ σ

. (d84)

Substituting (d82), (d66) and (d70) into (d84) and integrating with respect to θ , it follows after some computations that

28

( )( )

( )

u

E

r n c

n

n

n

n r

n

n

n d

n

n

n

n r

n

n

θ

ν

ν ψ ν θ γ θ

θ δ θ

·

+

− + − + −

− +

·

∞

¸

+ −

− −

·

∞

1

]

1

1

∑

∑

1

1 4 4

1

2

1

1

( ) ( ) sin cos

sin cos

, (d85)

Form the condition

r

u

→∞

·

lim θ

0 , (d86)

it follows the unknown function ψ ψ · ( ) r is subject to the condition

r

r

→∞

·

lim

( ) ψ 0 . (d87)

But

r

E

r

u

r

r

u

r

u

r

θ

ν

θ

∂

∂θ

θ

∂

θ

∂

ε σ

·

+

· − +

¸

¸

_

,

1 1

2

, (d88)

or

2

1

r

r

E u

r

u r

u

r

θ

ν

∂

∂θ

θ

∂

θ

∂

σ

·

+

− +

¸

¸

_

,

. (d89)

Substituting (d73), (d82) and (d85) into (d89), it follows after some computations that

r

d

dr

ψ

ψ · , (d90)

i.e. ψ · Cr . From (d87) it follows that ψ( ) r ≡ 0 and, finally,

( )( ) ( ) u

E

n c

n

n

n

n r

n

n

n d

n

n

n

n r

n

n

θ

ν

ν θ γ θ θ δ θ ·

+

− + −

− +

·

∞

+ −

− −

·

∞

¸

1

]

1

1

∑ ∑

1

4 4

1

2

1

1

sin cos sin cos

(d91)

D.10) Boundary conditions for the stress elements on the wall of the circular cavity.

Using the previous results, the final expressions of the plane elements of the stress are equal to

( )

( )

rr

c

r

n n c

n

n

n

n r

n

n

n n d

n

n

n

n r

n

n

σ

θ γ θ

θ δ θ

· − + −

¸

¸

_

,

+

−

·

∞

− +

¸

¸

_

,

+

− −

·

∞

∑

∑

0

2

2

2

1

2 2

1

cos sin

cos sin

(d92)

( )

( )

θθ

θ γ θ

θ δ θ

σ

· − + − − +

−

·

∞

+ + +

− −

·

∞

∑

∑

c

r

n n c

n

n

n

n r

n

n

n n d

n

n

n

n r

n

n

0

2

2 1

1

1

2

1

( )( ) cos sin

( ) cos sin

(d93)

and

( ) ( )

r

n n c

n

n

n

n r

n

n

n n d

n

n

n

n r

n

n

θ

θ γ θ θ δ θ

σ

· − − −

−

·

∞

− +

¸

¸

_

,

−

− −

·

∞

∑ ∑

( ) sin cos sin cos

2

1

2 2

1

(d94)

The cavity wall has the outer normal (with respect to the rock domain) equal to −

→

e

r

and radius equal to r R · . In the case

of a bore hole, let ∆p be the difference between the mud pressure and the pressure of the fluid contained by the porous rock

29

(usually, because the atmospheric pressure is negligible, it follows in the case of a tunnel that ∆p · 0). It follows the final

stress

f

σ

satisfies the next boundary condition

f

r

p

r

for r R

e e

σ

−

→

¸

¸

_

,

·

→

· ∆ , (d92)

Hence

rr

r R

rr

p

r

r R

r

σ σ

σ σ

θ θ

·

· − −

·

· −

¹

'

¹

¹

¹

0

0

∆

. (d93)

With no loss of generality, it can be assumed that the stress at infinity is along its main axes, i.e.

12

0

0

σ

· . Hence

( ) ( )

( ) ( )

c

R

n n c

n

n

n

n R

n

n

n n d

n

n

n

n R

n

n

p

n n c

n

n

n

n R

n

n

n n d

n

n

n

n R

n

n

0

2

2

2

2

2 2

1

11

0

22

0

2

11

0

22

0

2

2

2

2

2 2

1

11

0

22

0

2

2

− + −

¸

¸

_

,

+

−

·

∞

− +

¸

¸

_

,

+

− −

·

∞

· −

+

−

−

−

−

¸

¸

_

,

− +

−

·

∞

+ +

¸

¸

_

,

− +

− −

·

∞

·

−

¹

'

¹

¹

¹

¹

¹

¹

¹

¹

¹

∑ ∑

∑ ∑

cos sin cos sin

cos

sin cos sin cos

sin

θ γ θ θ δ θ

θ

θ γ θ θ δ θ

θ

σ σ σ σ

σ σ

∆

¹

¹

¹

¹

(d94)

Hence

c

R

p

d

0

2

11

0

22

0

2

1 1

0

· −

+

−

· ·

¹

'

¹

¹

¹

¹

¹

σ σ

δ

∆

, (d95)

and

n n

c

n

R

n

n n

d

n

R

n

n n n

n

R

n

n n

n

R

n

n

n

2

2

2

2

2

2

2

2

2

11

0

22

0

2

2

+ −

¸

¸

_

,

+ +

¸

¸

_

,

+

¸

1

]

1

+ + −

¸

¸

_

,

+ +

¸

¸

_

,

+

¸

1

]

1

¹

'

¹

¹

¹

¹

;

¹

¹

¹

·

∞

·

−

∑

cos sin

cos

θ

γ δ

θ

θ

σ σ

(d96)

− −

¸

¸

_

,

+ +

¸

¸

_

,

+

¸

1

]

1

+ −

¸

¸

_

,

+ +

¸

¸

_

,

+

¸

1

]

1

¹

'

¹

¹

¹

¹

;

¹

¹

¹

·

∞

·

−

∑

n n

c

n

R

n

n n

d

n

R

n

n n n

n

R

n

n n

n

R

n

n

n

2 2

2

2 2

2

2

11

0

22

0

2

2

sin cos

sin

θ

γ δ

θ

θ

σ σ

. (d97)

It follows that

30

c

R

p

c

R

d

R

c

R

d

R

0

2

11

0

22

0

2

4

2

2

6

2

4

11

0

22

0

2

2

2

2

6

2

4

11

0

22

0

2

· −

+

−

+ ·

−

− − ·

−

¹

'

¹

¹

¹

¹

¹

¹

¹

¹

¹

¹

¹

σ σ

σ σ

σ σ

∆

. (d98

and

( ) ( ) , , , ...

( ) ( ) , , , ...

( ) ( ) , , , ...

( ) ( ) , , , ...

n n

c

n

R

n

n n

d

n

R

n

n

n n

n

R

n

n n

n

R

n

n

n n

c

n

R

n

n n

d

n

R

n

n

n n

n

R

n

n n

n

R

n

n

2

2

2

2

0 3 4

2

2

2

2

0 2 3

2 2

2

0 3 4

2 2

2

0 2 3

+ − + +

+

· ·

+ − + +

+

· ·

− + +

+

· ·

− + +

+

· ·

¹

'

¹

¹

¹

¹

¹

¹

¹

¹

¹

¹

¹

γ δ

γ δ

(d99)

Hence

c p R

c R c

n

n

d

R

d

n

n

n n

n

0

11

0

22

0

2

2

2

11

0

22

0

2

2

0 3 4

2

11

0

22

0

4

4

0 3 4

0 0 2 3

· −

+

+

¸

¸

_

,

·

−

· ·

· − −

¸

¸

_

,

· ·

· · ·

¹

'

¹

¹

¹

¹

¹

¹

¹

¹

¹

¹

¹

σ σ

σ σ

σ σ

γ δ

∆

, , , ,...

, , , , ...

, , , , ...

. (d100)

By using (d11-d13) for

12

0

0

σ

· , the expressions of the final stress are equal to

rr

f

p

R

r

R

r

R

r

σ

σ σ σ σ

σ σ

σ σ σ σ

θ

θ θ

·

+

+

−

−

+

+

¸

¸

_

,

− −

¸

¸

_

,

+ −

¸

¸

_

,

11

0

22

0

2

11

0

22

0

2

2

11

0

22

0

2

2

2

2

11

0

22

0

2

2

2

3

2

11

0

22

0

4

4

2

cos

cos cos ∆

(d101)

and

θθ

θ θ

σ

σ σ σ σ σ σ

σ σ

f

p

R

r

R

r

·

+

−

−

+

+

+

¸

¸

_

,

− −

¸

¸

_

,

11

0

22

0

2

11

0

22

0

2

2

11

0

22

0

2

2

2

3

2

11

0

22

0

4

4

2 cos cos ∆ (d102)

r

f

R

r

R

r

θ

θ θ θ

σ

σ σ

σ σ σ σ

· −

−

− −

¸

¸

_

,

+ −

¸

¸

_

,

11

0

22

0

2

2

11

0

22

0

2

2

2

3

2

11

0

22

0

4

4

2 sin sin sin (d103)

In real cases, the value of the stress at infinity are positive ones for compression.

31

D.11) The final shape of the wall.

Consider again the case when the direction of the horizontal axes of the co-ordinate system is along the corresponding eigen

vectors of the initial stress

0

σ

. In that case,

12

0

0

σ

· . Using the above results, it follows the displacement vector for the

points initially placed on the wall of the circular cavity is

u

r

r R

E

R p

u r R

E

R

( , ) ( ) cos

( , ) ( ) sin

· ·

+

+

+ + −

−

¸

1

]

1

1

1

· · −

+

−

−

¹

'

¹

¹

¹

¹

¹

¹

¹

θ

ν

ν θ

θ

θ

ν

ν θ

σ σ σ σ

σ σ

1

11

0

22

0

2

3 4

11

0

22

0

2

2

1

3 4

11

0

22

0

2

2

∆

. (d104)

Consider an arbitrary point on the wall of the bore hole. In the initial state, it has the polar co-ordinates ( ) r R, · θ . Its

position vector with respect to the centre of the circle is

→

·

→

+

→

¸

¸

_

,

X

e e

R cos sin θ θ

1 2

(d105)

Using (d6) the position vector in the final stage is

→

·

→

+

→

+

→

·

→

+

→

¸

¸

_

,

+

+

+

+ + −

−

¸

1

]

1

1

1

→

+

→

¸

¸

_

,

−

+

−

−

−

→

+

→

¸

¸

_

,

χ

σ σ σ σ

σ σ

θ

θ

θ

θ

θ θ

ν

ν θ θ θ

ν

ν θ θ θ

X

e e e e

e e

e e

u

r

R

r

u R R

E

R p

E

R

( , ) ( , ) cos sin

( ) cos cos sin

( ) sin sin cos

1 2

1

11

0

22

0

2

3 4

11

0

22

0

2

2

1 2

1

3 4

11

0

22

0

2

2

1 2

∆ . (d106)

Hence

x R

E

R p

E

R

x R

E

R p

E

R

1

1

11

0

22

0

2

3 4

11

0

22

0

2

2

1

3 4

11

0

22

0

2

2

2

1

11

0

22

0

2

3 4

11

0

22

0

2

2

1

3 4

11

0

22

0

2

2

· +

+

+

+ + −

−

¸

1

]

1

1

1

+

+

−

−

· +

+

+

+ + −

−

¸

1

]

1

1

1

−

+

−

−

¹

'

¹

¹

¹

¹

¹

¹

¹

¹

¹

¹

¹

¹

¹

¹

¹

cos ( ) cos cos

( ) sin sin

sin ( ) cos sin

( ) sin cos

θ

ν

ν θ θ

ν

ν θ θ

θ

ν

ν θ θ

ν

ν θ θ

σ σ σ σ

σ σ

σ σ σ σ

σ σ

∆

∆

. (d107)

After elementary computations, it follows

32

x

E

p

x

E

p

X

X

1

1

1

11

0

22

0

2

3 4

11

0

22

0

2

1

2

1

1

11

0

22

0

2

3 4

11

0

22

0

2

2

· +

+

+

+ + −

−

¸

¸

_

,

¸

1

]

1

1

1

· +

+

+

+ − −

−

¸

¸

_

,

¸

1

]

1

1

1

¹

'

¹

¹

¹

¹

¹

¹

¹

¹

¹

ν

ν

ν

ν

σ σ σ σ

σ σ σ σ

∆

∆

( )

( )

. (d108)

Taking into account that

1

2

2

2 2

X X

R + · , it follows the final shape of the cavity is an ellipse of equation

x

a

x

b

1

2

2

2

2

2

1 + · , (d109)

where the semi-axes of the ellipse are equal to

a R

E

p

b R

E

p

· +

+

+

+ + −

−

¸

¸

_

,

¸

1

]

1

1

1

· +

+

+

+ − −

−

¸

¸

_

,

¸

1

]

1

1

1

¹

'

¹

¹

¹

¹

¹

¹

¹

¹

¹

1

1

11

0

22

0

2

3 4

11

0

22

0

2

1

1

11

0

22

0

2

3 4

11

0

22

0

2

ν

ν

ν

ν

σ σ σ σ

σ σ σ σ

∆

∆

( )

( )

. (d110)

Fig.D2. The shape of the borehole (tunnel) in the initial state (the circle) and in the final state (the ellipse), corresponding to a

compressive stress.

In real cases, the initial stress

0

σ

is usually a compressive one, i.e. (see Fig.D2)

11

0

1

22

0

3 σ σ

· − · − S S , (d111)

where the maximum compressive stress S

1

and the minimum compressive stress S

3

have positive values 0

3 1

≤ ≤ S S . It

follows here that the major semi-axis a corresponds to the minimum stress and a b ≤ .

33

E) BOUSSINESQ’S PROBLEM - the concentrated force acting on the elastic semi-

space

E.1) The equations of BELTRAMI and MITCHELL.

It follows to obtain the partial derivative equations for the stress tensor

σ

in the particular case of an elastic, homogeneous,

isotropic media. In the beginning, the next symmetric tensor is evaluated

S

grad b grad

t

b · −

→

+

→

¸

1

]

1

1

( ) ( ) ρ ρ

, (e1)

Using the equilibrium equation, it follows that

div b

σ

ρ · −

→

(e2)

Hence

S

grad div grad

t

div ·

¸

¸

_

,

+

¸

¸

_

,

σ σ

.

(e3)

The HOOKE’s reversed law gives

ε σ

ν ν

·

+

−

1

1

E E

Θ , (e4)

so

σ ε

ν

ν

ν

·

+

+

+

E

1 1

1

Θ , (e5)

where

Θ · tr

σ

, ϑ

ν

ε

· ·

−

tr

E

1 2

Θ . (e6)

But

div f grad f ( )

1

· . (e7)

Hence

div

E

div grad

σ ε

ν

ν

ν

·

+

+

+ 1 1

Θ (e8)

Expression (e1) becomes

S

E

grad div grad

t

div grad grad grad

t

grad ·

+

+

¸

1

]

1

+

+

+

¸

1

]

1

1 1 ν

ν

ν

ε ε

( ) ( ) ( ) ( ) Θ Θ , (e9)

Because the tensor grad grad ( ) Θ is a symmetric one, it follows that

S

E

grad div grad

t

div grad grad ·

+

+

¸

1

]

1

+

+ 1

2

1 ν

ν

ν

ε ε

( ) ( ) ( ) Θ . (e10)

The tensor in the first parenthesis of (e10) has the ij - component equal to

[ ]

grad div grad

t

div

ij

div

i, j

div

j i

iq qj jq qi

i,qqj q iqj j qqi q jqi i, j j i

qq

q q

ij

ij

tr

ij ij

E

ij

ij

E

grad grad

ij

u u u u u u u

( ) ( ) ( ) ( )

,

, ,

, , , ,

,

,

,

,

,

( )

ε ε ε ε ε ε

ε ε ε ε

ν ν

+

¸

1

]

1

· + · + ·

+ + +

¸

¸

_

,

· +

¸

¸

_

,

+

¸

¸

_

,

¸

1

]

1

1

·

¸

¸

_

,

+

¸

¸

_

,

·

¸

¸

_

,

+

−

·

¸

¸

_

,

+

−

1

2

1

2

2

1 2 1 2

∆ ∆ Θ ∆ Θ

(e11)

Hence

grad div grad

t

div

E

grad grad ( ) ( ) ( )

ε ε ε

ν

+ · +

−

∆ Θ

1 2

(e12)

Using (e12) and (e10), the expression (e1) becomes

34

S

E

grad grad ·

+

+

+ 1

1

1 ν ν

ε

∆ Θ (e13)

Applying the 3-dimensional LAPLACE operator ∆ in (e4), it follows

∆ ∆ ∆Θ

ε σ

ν ν

·

+

−

1

1

E E

(e14)

Replacing (e14) into (e13) gives

S

grad grad · −

+

+

+

∆ ∆Θ Θ

σ

ν

ν ν 1

1

1

1

, (e15)

i.e.

∆ ∆Θ Θ

σ

ν

ν ν

ρ ρ −

+

+

+

· −

→

+

→

¸

1

]

1

1

1

1

1

1

grad grad grad b grad

t

b ( ) ( ) (e16)

But

tr tr

tr grad grad

tr

tr grad b grad

t

b tr grad b div b

( ) ( ) ,

( ) ,

( ) ,

( ) ( ) ( ) ( )

∆ ∆ ∆ Θ

Θ ∆ Θ

∆Θ ∆ Θ

σ σ

ρ ρ ρ ρ

· ·

·

·

→

+

→

¸

1

]

1

1

·

→

¸

1

]

1

1

·

→

1

3

2 2

(e17)

Applying the trace operator in (e16) and using eqs.(e17) gives

∆Θ ∆Θ ∆Θ +

+

−

+

· −

→

1

1

3

1

2

ν

ν

ν

ρ div b ( ) (e18)

It follows the trace of the stress tensor verifies the relation

∆Θ · −

+

−

→

1

1

ν

ν

ρ div b ( ) (e19)

Replacing (e19) into (e16) gives

∆ Θ

σ

ν

ν

ρ

ν

ρ ρ +

−

→

+

+

· −

→

+

→

¸

1

]

1

1

1

1

1

1

div b grad grad grad b grad

t

b ( ) ( ) ( ) (e20)

Eqs.(e19)-(e20) represents the equations of BELTRAMI and MITCHELL, having as unknowns only the elements of the stress

tensor . Together with appropriate conditions (in tensions) on the boundary of the elastic body, they allow one to solve the

corresponding linear static problem.

Particular cases.

a) Suppose that

div b ( ) ρ

→

· 0 , (e21)

i.e. a vector potential ψ

→

exists having the property that

ρ ψ b rot

→

·

→

, (e22)

From (e19) it follows that the traces of both stress and strain tensors are harmonic functions

∆Θ ∆θ · · 0 , (e23)

b) Suppose that

ρ ϕ b grad

→

· , where ∆ϕ · 0 , (e24)

it follows that

div b ( ) ρ

→

· · ∆ϕ 0 , (e25)

i.e. eq. (e23) is verified and eq.(20) gives

∆ Θ

σ

ν

ϕ +

+

· −

1

1

2 grad grad grad grad ( ) (e26)

35

Applying the LAPLACE operator ∆ to eq.(e26), it follows that the stress tensor is the solution of the þi-harmonic equation

∆∆

σ

· 0 . (e27)

In most real cases, the volume forces are neglected (or they are represented only by the weight of the body, satisfying eq.(24). It

follows the elastic linear problem involve solving harmonic and bi-harmonic equations.

E.2) The model.

A co-ordinate system having the third vertical axis positive downward is used. The semi-space x

3

0 ≥ is represented by an

elastic, homogeneous, isotropic medium having the elastic coefficients λ and µ (or E and νrespectively). In the origin of the

co-ordinate system is acting a vertical force having the magnitude equal to P. It follows to find the stress and the

displacements. Spherical co-ordinates will be used (Fig.E1):

x r x r x r

1 2 3

· · · sin cos , sin sin , cos ϑ λ ϑ λ ϑ (e28)

Fig.E1 The spherical co-ordinate system and a vertical force of magnitude equal to P acting at the origin.

Because symmetry, the displacement vector has the components like

r r

r r

u u u u u

· · · ( , ) , , ( , ) ϑ

λ ϑ ϑ

ϑ 0 (e29)

It follows the components of both stress and strain tensors are functions of r and ϑ only.

E.3) The equations of equilibrium and strain tensor in spherical co-ordinates.

The LAMÉ differential parameters for spherical co-ordinates are

1

1

2 3

h h h

r r · · · , , sin ϑ (e30)

The generalised curvilinear co-ordinates are

1 2 3

c c c

r · · · , , ϑ λ (e31)

The unit vectors of the axis are

1 2 3

n e n e n e

r

→

·

→

→

·

→

→

·

→

, ,

ϑ λ

(e32)

By using, for example, (IVAN, 1996), the divergence of a symmetric tensor

T

in spherical co-ordinates is

div

rr

r r

r

r

r rr

r

r

r r

r

r

r r r

r

r

r

r r

r

r r r

r

r r

T

T T T T T T T

e

T T T T T T T

e

T T T T T

e

·

+ + + + −

+ ¸

¸

_

,

→

+

+ + + + +

− ¸

¸

_

,

→

+

+ + + +

¸

¸

_

,

→

∂

∂

∂

ϑ

∂ ϑ ϑ

∂

λ

∂ λ

ϑ

ϑ

ϑϑ λλ

∂

ϑ

∂

∂

ϑϑ

∂ ϑ ϑ

∂

ϑλ

∂ λ

ϑ ϑ

ϑ

ϑϑ λλ

ϑ

ϑ

∂

λ

∂

∂

ϑλ

∂ ϑ ϑ

∂

λλ

∂ λ

λ ϑλ

ϑ

λ

1 1

2

1 1

3

1 1

3

2

sin tan

sin tan tan

sin tan

(e33)

In the same way, the gradient of a vector in spherical co-ordinates is

36

grad

r

r

r r

r

r

r

r

r

r

r

r

r

r

r

r

r r r

r

r

r r

v

v

e e

v v

e e

v v

e e

v

e e

v v

e e

v v

e e

v

e e

v

e e

v

→

·

→

⊗

→

+ −

¸

¸

_

,

→

⊗

→

+ −

¸

¸

_

,

→

⊗

→

+

→

⊗

→

+ +

¸

¸

_

,

→

⊗

→

+ −

¸

¸

_

,

→

⊗

→

+

→

⊗

→

+

→

⊗

→

+

∂

∂

∂

∂ ϑ

ϑ

ϑ

ϑ

∂

∂ λ

λ

λ

∂

ϑ

∂

ϑ

∂

ϑ

∂ ϑ

ϑ ϑ

ϑ

∂

ϑ

∂ λ

λ

ϑ

ϑ λ

∂

λ

∂

λ

∂

λ

∂ ϑ

λ ϑ

ϑ

∂

λ

∂

1 1

1 1

1 1

sin

sin tan

sin λ

ϑ

ϑ

λ λ

+ +

¸

¸

_

,

→

⊗

→

r

r r

v v

e e

tan

(e34)

The components of the strain tensor are

rr

r

r

r

r

r

r r

r

r

r

r r r

r

r

r r r r

r

r r

u u u u

u u u u u

u u u u u u

ε ε

ε ε

ε ε

∂

∂

ϑ

∂

∂ ϑ

ϑ

∂

ϑ

∂

λ

ϑ

∂

∂ λ

λ

∂

λ

∂

ϑϑ

∂

ϑ

∂ ϑ

ϑλ

ϑ

∂

ϑ

∂ λ

λ

ϑ

∂

λ

∂ ϑ

λλ

ϑ

∂

λ

∂ λ

ϑ

ϑ

· · − +

¸

¸

_

,

· − +

¸

¸

_

,

· +

· − +

¸

¸

_

,

· + +

, ,

sin

, ,

sin tan

,

sin tan

1

2

1

1

2

1 1

1

2

1 1 1

(e35)

In the particular case of eqs.(e29), it follows

rr

r

r

r

r

r

r r

r

r

r

r

r

r r

u u u u

u u u u

ε ε ε

ε ε ε

∂

∂

ϑ

∂

∂ ϑ

ϑ

∂

ϑ

∂

λ

ϑϑ

∂

ϑ

∂ ϑ

ϑλ λλ

ϑ

ϑ

· · − +

¸

¸

_

,

·

· + · · +

, ,

, ,

tan

1

2

1

0

1

0

(e36)

Hence

rλ ϑλ

σ σ

· · 0 , (e37)

By using (e33), the equilibrium equations in the absence of the volume forces are

∂

∂

∂

ϑ

∂ ϑ

ϑ

ϑ

ϑϑ λλ

∂

ϑ

∂

∂

ϑϑ

∂ ϑ

ϑ ϑ

ϑ

ϑϑ λλ

ϑ

σ σ σ σ σ σ σ

σ σ σ σ σ σ

rr

r r

r rr

r

r

r

rr

r

r

r r

r

r

r

r r

+ + + ·

+ +

+ + + +

−

·

1

3

1

3

0

tan

tan tan

, (e38)

i.e.

∂

∂

ϑ

∂

∂ ϑ

ϑ

ϑ

ϑ

∂

∂

ϑ

ϑ

∂

∂ ϑ

ϑ

ϑϑ

ϑ

λλ

σ σ

σ σ σ

r

r

rr

r

r

r

r

r

r

r r

3 2 2

3 2 2

sin sin sin ,

sin sin cos

¸

¸

_

,

+

¸

¸

_

,

·

¸

¸

_

,

+

¸

¸

_

,

·

Θ

(e39)

E.4) LAPLACE operator in spherical co-ordinates. LEGENDRE’s polynomials.

Using, for example, (IVAN, 1996) the gradient of a scalar function in spherical co-ordinates is

grad f

f

r

r

r

f

r

f

e e e

·

→

+

→

+

→

∂

∂

∂

∂ ϑ

ϑ

∂

∂ λ

λ

1 1

sin ∆

. (e40)

Also, the LAPLACE operator is

∆ f div grad f

r

r

r

f

r

f f

· ·

¸

¸

_

,

+

¸

¸

_

,

+

¸

1

]

1

1

1

2

2

1

2

2

sin

sin sin

sin

ϑ

ϑ

∂

∂

∂

∂

∂

∂ ϑ

ϑ

∂

∂ ϑ ϑ

∂

∂λ

(e41)

Consider the LAPLACE equation

37

∆ f · 0 (e42)

for the particular case when the unknown function has the form ( ) f f r · , ϑ . Equation (e42) becomes

∂

∂

∂

∂ ϑ

∂

∂ ϑ

ϑ

∂

∂ ϑ r

r

f

r

f

2

1

0

¸

¸

_

,

+

¸

¸

_

,

·

sin

sin . (e43)

By using the method of separation of the variables, a solution for eq.(43) has the form

( ) f r R r Y , ( ) ( ) ϑ ϑ · , (e44)

where R andY are two unknown functions. Eq.(e43) becomes

d

dr

r

dR

dr

R

d

d

dY

d

Y

k

2

¸

¸

_

,

· −

¸

¸

_

,

·

ϑ

ϑ

ϑ

ϑ

sin

sin

, (e45)

where k is a constant. From (e45) it follows that

r

d R

dr

r

dR

dr

kR

2

2

2

2 0 + − · (e46)

In a general case, the function R can be developed in power series. Let’s look for a particular solution having the form

R

n

r r

n

( ) · where n is a natural number. It follows from (e46) that

k n n · + ( ) 1 (e47)

Then the particular solution of eq. (e46) can be expressed with the aid of two arbitrary constants as

R

n

r A

n

r

n

B

n

r

n

( ) / · + (e48)

Because f has to approach finite values for r → ∞ , it has to take A

n

· 0.

The second relation (e45) gives

sin cos ( ) sin ϑ

ϑ

ϑ

ϑ

ϑ

d Y

d

dY

d

n n Y

2

2

1 0 + + + · (e49)

By performing the substitution z · cos ϑ eq.(e49) becomes

d

dz

z

dY

dz

n n Y 1

2

1 0 −

¸

¸

_

,

¸

1

]

1

+ + · ( ) , (e50)

The solution of (e50) is represented by the LEGENDRE polynomials denoted by P

n

z n ( ) , , , , ... · 01 2 So,

P

o

z P z z P z z ( ) , ( ) , ( ) ( ) / · · · − 1

1 2

3

2

1 2 (e51)

Because the trace Θof the stress tensor is a solution of the harmonic equation (e23), it follows that the general solution for that

trace in the case of the BOUSSINESQ problem is

Θ ·

+

·

∞

∑

a

n

r

n

n

P

n

1

0

(cos ) ϑ . (e52)

According to eq.(e6), a similar solution exists for the trace of strain tensor.

E.5) The displacement field.

We look for a displacement field having the form

r

r r

u u u

· · ·

1 1

0 ϕ ϑ

ϑ

ψ ϑ

λ

( ) , ( ) , , (e53)

where ϕ and ψ are two unknown functions following to be obtained. Substituting (e53) into (e36) it follows that

( )

( ) ( )

rr

r

r

d d r

r

d d r tg r

ε ε ε

ε ε ε

ϕ

ϑ

ϕ ϑ ψ

λ

ϑϑ

ϕ ψ ϑ

ϑλ λλ

ϕ ψ ϑ

· − · −

¸

¸

_

,

·

· + · · +

/ , / / ,

/ / , , / /

2

2 2

2

0

2

0

2

(e54)

It follows that the trace of the strain tensor is

( ) θ

ϑϑ λλ

ϕ ψ ϑ ψ ϑ

ε ε ε ε

· · + + · + + tr

rr

d d tg r / / /

2

(e55)

A comparison of (e55) to (e52) shows that

38

ϕ ψ ϑ ψ ϑ ϑ + + · d d tg a / / cos , (e56)

where a is a constant remaining to be obtained. Hence

( )

( )

( )

θ ϑ λ µ θ λ µ ϑ

λλ

θ

ϑϑ

ϑ ψ ϑ

ε ε ε

· · + · +

· − + · −

a r a r

rr

a d d r

cos / , ( ) cos / ,

cos / /

2

3 2 3 2

2

2

Θ

(e57)

By using HOOKE’s law, it follows that

( ) ( )

( )

[ ]

[ ]

rr

a r

r

d d r

a d d r

a d d r

σ σ

σ

σ

λ ϑ µϕ

ϑ

µ ϕ ϑ ψ

ϑϑ

λ ϑ µ ϕ ψ ϑ

λλ

λ µ ϑ µ ψ ϑ

· − · −

· + +

· + −

cos / , / / ,

cos / / ,

( ) cos / /

2

2

2

2

2

2

2 2

2

(e58)

Substituting (e58) into the first equilibrium equation (e39) and using (e56), it follows after elementary computations that

d

d

d

d

a

ϑ

ϑ

ϕ

ϑ

λ µ ϑ sin ( / ) sin

¸

¸

_

,

· + 2 2 . (e59)

Hence

sin ( / ) cos ( / ) ( / ) sin ,

( / ) /

sin

( / )sin

ϑ

ϕ

ϑ

λ µ ϑ λ µ λ µ ϑ

ϕ

ϑ

λ µ

ϑ

λ µ ϑ

d

d

a

b b

a

a

d

d

b a

a

· − + + · − + + +

·

− +

+ +

2

2 2

2

2 2

2

2 2

2

(e60)

where b is an integration constant. Because

dx

x

tg

x

C

sin

ln

∫

· +

2

, (e61)

where C is a new integration constant, it follows that

ϕ λ µ

ϑ

λ µ ϑ · − +

¸

1

]

1

− + + b

a

tg a C

2

2

2

2 ( / ) ln ( / ) cos . (e62)

For ϑ π → / 4 , the logarithmic term into (e62) leads to infinite radial displacements. That can be avoided by taking

b

a

− + ·

2

2 0 ( / ) λ µ . (e63)

Hence

ϕ λ µ ϑ

ϕ

ϑ

λ µ ϑ · − + + · + a C

d

d

a ( / ) cos , ( / ) sin 2 2 . (e64)

and

( )

[ ]

rr

a C r

σ

λ µ ϑ µ · + − 3 4 2

2

cos / . (e65)

Substituting eq.(e64) into eq.(e56), it follows that

( )

d

d tg

a C

d

d

a

C

a

C D

D

a a

C

ψ

ϑ

ψ

ϑ

λ µ ϑ

ϑ

ψ ϑ λ µ ϑ ϑ

ψ ϑ λ µ ϑ ϑ

λ µ λ µ ϑ ϑ

+ · + +

· + −

· − + + +

· − + + + +

( / ) cos ,

sin ( / ) sin sin ,

sin ( / ) cos cos

( / ) ( / )sin cos

3

2

3 2

4

3 2

4

3

2

3

2

(e66)

Hence

ψ

λ µ

ϑ

λ µ ϑ ϑ ·

− +

+ + +

D a a

C ctg

( / ) /

sin

( / )sin

3 4

2

3 . (e67)

Because the tangential displacements has to be finite ones for ∆ →0, it follows that

39

ψ

ϑ

ϑ

λ µ ϑ

ϑ

λ µ ϑ

ψ

ϑ

ϑ

λ µ ϑ

·

−

+ + · − + +

· − + +

C

a

C tg

a

d

d

C a

cos

sin

( / )sin ( / )sin ,

cos ( / )

( / ) cos

1

2

3

2 2

3

2

2

2

2

3

(e68)

By substituting eqs.(e64) and (e68) into eq.(e58) it follows that

r

C tg a r C tg a r

C tg a r

ϑ

µ

ϑ

ϑ

ϑϑ

µ

ϑ

ϑ

λλ

µ

ϑ

ϑ

σ σ

σ

· −

¸

¸

_

,

· −

¸

¸

_

,

−

¸

1

]

1

· +

¸

¸

_

,

−

¸

1

]

1

2

2

2

1

2

2

2

1

2

2

2

sin / , cos / ,

cos /

(e69)

Substituting eqs.(e69) into the second equilibrium equation (e39) it can be seen that the last one is identically verified.

E.6) Boundary conditions for stress elements. The final solutions.

Eqs.(e65) and (e69) contain the unknown coefficients a and C. These constants follow to be obtained taking into account

that the force P concentrated in the origin of the co-ordinate system is acting on the elastic semi-space. It can be seen that the

points of the horizontal plane x

3

0 · have the co-latitude ϑ π · / 2.

The unit vectors of the spherical co-ordinate system are related to the same vectors of the rectangular co-ordinate system by

r

t

r

e

e

e

Q

e

e

e

e

e

e

Q

e

e

e

→

→

→

¸

¸

_

,

·

→

→

→

¸

¸

_

,

→

→

→

¸

¸

_

,

·

→

→

→

¸

¸

_

,

ϑ

λ

ϑ

λ

1

2

3

1

2

3

, , (e70)

where the orthogonal matrix is

Q

· −

−

¸

¸

_

,

**sin cos sin sin cos
**

cos cos cos sin sin

sin cos

ϑ λ ϑ λ ϑ

ϑ λ ϑ λ ϑ

λ λ 0

. (e71)

The outer pointing unit vector normal to the elastic semi-space is equal to

−

→

· −

→

+

→

3

e e e

r

cos sin ϑ ϑ

ϑ

. (e72)

The resulting exterior force acting on the elastic semi-space is vanishing for all the points of the horizontal plane ϑ π · / 2,

excepting the origin, i.e.

σ

−

→

¸

¸

_

,

·

→

3

0

e

. (e73)

Substituting (e72) into (e73) for ϑ π · / 2gives

ϑϑ ϑ

σ σ

· · 0 0 ,

r

. (e74)

By using (e69), the first eq.(e74) becomes an identity and the second one leads to

a C · 2 . (e75)

So, eqs.(e65) and (e69) give

( )

[ ]

rr

C r

σ

λ µ ϑ µ · + − 2 3 4

2

cos / , (e76)

r

C tg r C tg r

ϑ

µ

ϑ

ϑϑ

µ ϑ

ϑ

σ σ

· − · −

¸

¸

_

,

2

2

2 2

2

1

2

cos / , cos / ∆ . (e77)

Consider an elastic hemisphere having the centre at the origin of the co-ordinate system. The curved surface S of the

hemisphere has the outer pointing normal equal to

→

r

e

. On that surface, the rest of the elastic body (i.e. the semi-space minus

the hemisphere) is acting on the hemisphere with a total force equal to

40

σ σ σ

ϑ ϑ r

S

dA

rr r r

S

dA

e e e

→

·

→

+

→

¸

¸

_

,

∫∫ ∫∫

, (e78)

where dA is the surface element and the unit vectors are obtained with (e70). It follows that

( )

[ ]

σ σ

λ

π π

ϑ ϑ

π λ µ ϑ µ ϑ ϑ µ

ϑ

ϑ ϑ ϑ

π

π λ µ

r

dA

S

d

r

r d

C tg d C

e e

e e

→

·

→

· + − +

¹

'

¹

¹

;

¹

→

· +

→

∫∫ ∫ ∫

∫

0

2

0

2

2

4 3 4

2

2

0

2

3

4

3

/

sin

cos cos sin cos sin

/

( )

(e79)

Because the hemisphere is into an equilibrium state, it follows that

σ

r

S

dA P

e e

→

+

→

·

→

∫∫

3

0 , (e80)

Hence

C

P

· −

+ 4π λ µ ( )

. (e81)

Finally, the non-zero components of the stress tensor are equal to

( )

[ ]

rr

P

r

σ

π λ µ

λ µ ϑ µ · −

+

+ −

2

3 4

2

( )

cos / , (e82)

r

P

tg r

∆

σ

π λ µ

µ

ϑ

ϑ ·

+ 2 2

2

( )

cos / , (e83)

ϑϑ

π λ µ

µ ϑ

ϑ

σ

·

+

−

¸

¸

_

,

P

tg r

4

1

2

2

2

( )

cos / . (e84)

The non-zero components of the displacement vector are

[ ]

r

P

r

P

tg r

u

u

·

+

+ −

·

+

− +

¸

1

]

1

4

2 2 1

4 2

3

π λ µ

λ µ ϑ

ϑ

π λ µ

ϑ

λ µ ϑ

( )

( / ) cos / ,

( )

( / )sin /

(e85)

Using eq.(e71), the components of the stress tensor into the Cartesian base can be obtained as

11 12 13

22 23

33

0

0

0 0

σ σ σ

σ σ

σ

σ σ

σ σ

σ

ϑ

ϑ ϑϑ

λλ

¸

¸

_

,

·

¸

¸

_

,

t

rr r

r

Q Q

(e86)

Also, the components of the displacement vector into the Cartesian base are

1

2

3

0

u

u

u

Q

u

u

t

r

¸

¸

_

,

·

¸

¸

_

,

ϑ

. (e87)

Of particular importance in real life are the components

33

3

2

3 5

3

1

2

2 1

1

2

3

σ

π

ν

π

ν · − ·

+

− +

¸

1

]

1

1

P

z r

E

P

r

z

r

u

/ , ( ) . (e88)

The BOUSSINESQ problem has a great importance in Geomechanics, in relation to the computation of a building foundation.

The above solution derived for a concentrated vertical force can be used in the case of arbitrary vertical forces (spread on a

certain domain) by assuming the principle of the superposition.

41

F) ELEMENTS OF THIN PLATES THEORY

F.1) The model of a thin elastic plane plate.

The thin plane plate is a cylindrical body having an arbitrary horizontal cross section in the initially non-deformed state. Its

height, denoted by H h · 2 , is much smaller than the other dimensions (usually, around 7-10 times). The material of the plate

is an elastic, homogenous, isotropic one, having the constants denoted by E and ν, respectively λ and µ . The third axis of

the co-ordinate system, denoted by Oz , is a vertical one, positive downward. Let be x x x y

1 2

· · , . In the initial, non-

deformed state, the plate is a plane horizontal one. The upper face has the equation z h · − and the down face has z h ·

respectively. The plane of equation z · 0 represents the median plane. After the plate is deformed, it becomes a median

surface.

F.2) The planar state of a plate. The bending state.

Two particular situation for a deformed plate are considered (Fig.F1).

Fig.F1. (a) The non-deformed plate; (b) The plate into a planar state; (c) The plate into a bending state. Here, the median plane

is denoted by m.p. while m.s. is the median surface.

In the first case, the horizontal components of the displacement vector are symmetrical ones with respect to the median plane,

being even functions with respect to the z -variable, while the vertical component of the displacement vector is an anti-

symmetrical one (odd function with respect to z ), i.e.

k

x y z

k

x y z k

x y z x y z

u u

u u

( , , ) ( , , ), ,

( , , ) ( , , ).

− · ·

− · − −

1 2

3 3

(f1)

From eq.(f1b) it follows that

3

0 0

u

x y ( , , ) · , (f2)

i.e. the material points placed initially into the median plane have no vertical displacement as a consequence of the deformation

of the plate (the median plane holds a horizontal plane one). This kind of deformation represents the planar state of the plate.

In the second case, the horizontal components of the displacement vector are anti-symmetrical ones (odd functions with respect

to z ), while the vertical component of the displacement vector is a symmetrical one with respect to the median plane (even

function with respect to z ), i.e.

k

x y z

k

x y z k

x y z x y z

u u

u u

( , , ) ( , , ), ,

( , , ) ( , , ).

− · ·

− · −

−

1 2

3 3

(f3)

By deformation, the material points initially placed into the median plane are displaced on the vertical direction, having no

horizontal movement. The median plane becomes a median surface. This state represents the bending state of a plate.

F.3) Loads acting on the plate.

42

For simplicity, the forces acting on the lateral surface of the plate are neglected. On the upper surface of the plate, having the

equation z h · − and the outward pointing normal vector

→

· −

→

n e

3

, it is acting the surface force

s s

x y

q q

→

·

→

( , ) . In the

same way, on the down surface, having the equation z h · and the normal vector

→

·

→

n e

3

, it is acting the surface force

j j

x y

q q

→

·

→

( , ) . Hence

σ

−

→

¸

¸

_

,

·

→

· −

e

q

s

for z h

3

, , (f4)

i.e.

13

1

23

2

33

3

σ σ σ

( , , ) ( , ) , ( , , ) ( , ) , ( , , ) ( , ) x y h

s

x y x y h

s

x y x y h

s

x y

q q q

− · − − · − − · − (f5)

Also

σ

→

·

→

·

e

q

j

for z h

3

, , (f6)

i.e.

13

1

23

2

33

3

σ σ σ

( , , ) ( , ) , ( , , ) ( , ) , ( , , ) ( , ) x y h

j

x y x y h

j

x y x y h

j

x y

q q q

· · · . (f7)

We shell see that the above presented deformation states are compatible only to certain distributions of volume or surface

forces applied to the plate.

F.4) Odd and even functions for the planar state and for the bending state.

Let f f x y z · ( , , ) be a function of three variables, supposed to be smooth enough. It can be seen that

f x y z

f x y z f x y z f x y z f x y z

( , , )

( , , ) ( , , ) ( , , ) ( , , )

·

+ −

+

− −

2 2

(f8)

Let

f x y z

f x y z f x y z

+

·

+ −

( , , )

( , , ) ( , , )

2

, (f9)

f x y z

f x y z f x y z

−

·

− −

( , , )

( , , ) ( , , )

2

(f10)

It follows that

f x y z f x y z f x y z f x y z

+

− ·

+ −

− · −

−

( , , ) ( , , ) , ( , , ) ( , , ) . (f11)

The function f

+

represents the even part of f (with respect to z -variable), while the function f

−

represents its odd part. It

follows that

f f f f f f

+

¸

¸

_

,

+

·

+ −

¸

¸

_

,

−

·

− +

¸

¸

_

,

−

·

−

¸

¸

_

,

+

· , , 0 , (f12)

i.e. the even part of the even part is equal to the even part too. A similar relation holds for the odd part. The even part of an odd

part (and the odd part of an even part) is vanishing. For k x x x y · · · 1 2

1 2

, , , , it follows that

[ ]

f

k

x

k

f x y z

x

k

f x y z

x

k

f x y z f

k

x y z

+

¸

¸

_

,

·

+

· + −

¸

1

]

1

·

+

,

( , , ) ( , , ) ( , , )

,

( , , )

∂

∂

∂

∂

∂

∂

1

2

(f13)

Hence, the partial derivative (with respect to a horizontal co-ordinate) of the even part of a certain function is equal to the even

part of that derivative. This property holds for the odd part. So,

( ) ( )

f

k

f

k

f

k

f

k

+

¸

¸

_

,

·

+

−

¸

¸

_

,

·

−

,

,

,

,

,

(f14)

Also,

43

[ ]

[ ] ( )

f

z

f x y z f x y z f x y z f x y z f

+

¸

¸

_

,

· + −

¹

'

¹

¹

;

¹

· − − ·

−

,

( , , ) ( , , )

,

( , , )

,

( , , )

,

3

1

2

1

2

3 3 3

∂

∂

(f15)

Hence the partial derivative of the even part with respect to the vertical co-ordinate is equal to the odd part of the partial

derivative of the function itself with respect to same co-ordinate. In the same way it follows that

( )

f f

−

¸

¸

_

,

·

+

,

,

3

3

. (f16)

By using the HOOKE’s law and the definition of the strain, it can be resumed that the planar state and the bending state are

characterised by the next components:

-for the planar state: - the displacement vector:

1 2 3

+ + −

u u u

, , ,

- strain tensor:

11 12 13 22 23 33

+ + − + − + +

ε ε ε ε ε ε

ϑ , , , , , , ,

- stress tensor:

11 12 13 22 23 33

+ + − + − +

σ σ σ σ σ σ

, , , , , .

-for the bending state: - the displacement vector:

1 2 3

− − +

u u u

, , ,

- strain tensor:

11 12 13 22 23 33

− − + − + − −

ε ε ε ε ε ε

ϑ , , , , , , ,

- stress tensor:

11 12 13 22 23 33

− − + − + −

σ σ σ σ σ σ

, , , , , .

F5) Mean value of a function. Equilibrium equations for thin plates.

Let f f x y z · ( , , ) an integrable function with respect to z -variable. The mean value of f computed on the thickness of the

plate is denoted by

f x y

h

f x y z

h

h

( , ) ( , , )dz ·

−

+

∫

1

2

. (f17)

For an odd function f , its mean value vanishes, i.e.

f x y

−

· ( , ) 0 , (f18)

It follows that

f f f f f f ·

+

+

−

·

+

+

−

·

+

(f19)

For a function C C x y · ( , ) depending only on the horizontal co-ordinates, eq.(f17) leads to

C C zC · · , 0 . (f20)

Differentiating eq.(f17) with respect to the horizontal co-ordinates, it follows that:

( )

f

k

f

k

k

,

,

, , · · 1 2 . (f21)

For the vertical derivative, it follows that

[ ]

f

h

f

z

dz

h

h

h

f x y h f x y h

h

f x y z h

,

( , , ) ( , , ) ( , , )

3

1

2

1

2

1

· ·

−

+

− − ·

−

·

∫

∂

∂

. (f22)

Consider the function zf x y z ( , , ) . It follows that

( ) ( ) zf zf zf zf

+

·

− −

·

+

, . (f23)

Its mean value is

( ) zf z f f zf zf zf zf ·

+

+

−

·

+

+

−

·

−

·

+

( ) . (f24)

Hence

44

zf

h

z

f

z

h

h

dz

h z

zf f

h

h

dz

h

zf

h

h

h

fdz

h

h

f x y h f x y h

f x y f x y z h f

,

( ) ( )

( , , ) ( , , )

( , ) ( , , )

3

1

2

1

2

1

2

1

2

2

·

−

+

· −

¸

1

]

1

−

+

·

−

+

¸

1

]

1

−

−

+

·

+ −

− ·

+

· −

∫ ∫ ∫

∂

∂

∂

∂

. (f25)

Neglecting the volume forces (the weight of the plate itself, for example), the equilibrium equations for the planar state are

11

1

12

2

13

3

0

12

1

22

2

23

3

0

13

1

23

2

33

3

0

+

¸

¸

_

,

+

+

¸

¸

_

,

+

−

¸

¸

_

,

·

+

¸

¸

_

,

+

+

¸

¸

_

,

+

−

¸

¸

_

,

·

−

¸

¸

_

,

+

−

¸

¸

_

,

+

+

¸

¸

_

,

·

σ σ σ

σ σ σ

σ σ σ

, , ,

, , ,

, , ,

. (f26)

Let the mean values of the stress components be denoted by

ij ij

h

ij

x y z

h

h

i j

Σ

· ·

−

+

·

∫ σ σ

1

2

1 2 3 ( , , )dz , , , , (f27)

Applying the mean value operator to eqs. (f26) gives

( ) ( ) [ ]

( ) ( ) [ ]

11

1

12

2

1

2

13 13

0

12

1

22

2

1

2

23 23

0

Σ Σ

Σ Σ

, ,

( , , ) ( , , )

, ,

( , , ) ( , , )

+ + − − ·

+ + − − ·

h

x y h x y h

h

x y h x y h

σ σ

σ σ

, (f28)

Using eqs.(f5) and (f7) it follows

( ) ( )

( ) ( )

11

1

12

2

1

2

1 1

0

12

1

22

2

1

2

2 2

0

Σ Σ

Σ Σ

, ,

( , ) ( , )

, ,

( , ) ( , )

+ + +

¸

1

]

1

·

+ + +

¸

1

]

1

·

h

j

x y

s

x y

h

j

x y

s

x y

q q

q q

. (f29)

Let

ij

z

ij

h

z

ij

x y z

h

h

i j

Μ

· ·

−

+

·

∫ σ σ

1

2

1 2 3 ( , , )dz , , , . (f30)

Multiplying eqs.(f26) by z and using again the mean value operator, it follows

( ) ( )

13

1

23

2

33 33

2

33

0

Μ Μ Σ

, ,

( , , ) ( , , )

+ +

+ −

− ·

σ σ

x y h x y h

. (f31)

Hence, using again eqs.(f5) and (f7),

( ) ( )

13

1

23

2

3 3

2

33

0

Μ Μ Σ

, ,

( , ) ( , )

+ +

−

− ·

j

x y

j

x y

q q

, (f32)

So, the equilibrium of a thin plate into the planar state leads to eqs.(f29) and (f32).

Consider the weight of the plate, the equation of equilibrium for the plate into the bending state are

11

1

12

2

13

3

0

12

1

22

2

23

3

0

13

1

23

2

33

3

0

−

¸

¸

_

,

+

−

¸

¸

_

,

+

+

¸

¸

_

,

·

−

¸

¸

_

,

+

−

¸

¸

_

,

+

+

¸

¸

_

,

·

+

¸

¸

_

,

+

+

¸

¸

_

,

+

−

¸

¸

_

,

+ ·

σ σ σ

σ σ σ

σ σ σ

ρ

, , ,

, , ,

, , ,

g

. (f33)

Here, the density of the plate is ρand the acceleration of gravity is g .

Proceeding in a similar manner to above, it follows the equations of equilibrium for the thin plate into a bending state are

45

( ) ( )

( ) ( )

( ) ( )

13

1

23

2

1

2

3 3

0

11

1

12

2

1

2

1 1

13

0

12

1

22

2

1

2

2 2

23

0

Σ Σ

Μ Μ Σ

Μ Μ Σ

, ,

( , ) ( , ) ,

, ,

( , ) ( , )

, ,

( , ) ( , )

+ + +

¸

1

]

1

+ ·

+ + −

¸

1

]

1

− ·

+ + −

¸

1

]

1

− ·

h

j

x y

s

x y g

j

x y

s

x y

j

x y

s

x y

q q

q q

q q

ρ

. (f34)

Eq.(f34b) is differentiated with respect to x and eq.(f34c) is differentiated with respect to y. The results are substituted in

eq.(f34a). Hence

( ) ( ) ( )

11

11

2

12

12

22

22

1

2

3 3

1

2

1 1

1

1

2

2 2

2

0

Μ Μ Μ

, , ,

( , ) ( , )

( , ) ( , )

,

( , ) ( , )

,

+ + + +

¸

1

]

1

+

+ −

¸

1

]

1

+ −

¸

1

]

1

·

h

j

x y

s

x y g

j

x y

s

x y

j

x y

s

x y

q q

q q q q

ρ

. (f35)

F.6) Thin plane plate in the bending state.

The component

1

u

of the displacement vector is developed in power series with respect to z -variable. It follows

1 1

0

1

0

u u

u

x y z x y z

z

x y ( , , ) ( , , ) ( , , ) ... · + +

∂

∂

(f36)

For the bending state, the first term in eq.(f36) vanishes. Because the thickness of the plate is a small one, only the second term

is kept. So the horizontal components of the displacement vector are

1

1

0

2

2

0

u

u

u

u

x y z z

z

x y x y z z

z

x y ( , , ) ( , , ) , ( , , ) ( , , ) · ·

∂

∂

∂

∂

. (f37)

The vertical displacement of the points placed into the median plane is denoted by

w w x y x y

u

· · ( , ) ( , , )

3

0 , (f38)

i.e. the mean surface has the equation z w x y · ( , ) . Here, w represents the arrow of the plate.

F7) BERNOULLI’s hypothesis.

According to BERNOULLI, it is assumed that an arbitrary material segment of the plate, initially perpendicular on the

median plane in the non-deformed state, rests perpendicular on the mean surface in the deformed state. Let A x y z ( , , ) a

certain point of the plate (not placed in the mean plane) and A x y

0

0 ( , , ) its projection on the median plane. So, the segment

A A

0

→

is perpendicular on the median plane. After deformation, the material point A is moving at the point having the co-

ordinates

( )

M x x y z y x y z z x y z

u u u

+ + +

1 2 3

( , , ), ( , , ), ( , , ) , while the point A

0

is moving at the point

( )

M x x y y x y w x y

u u 0

1

0

2

0 + + ( , , ), ( , , ), ( , ) . The horizontal displacements of the points placed in the median plane

are vanishing. So, writing the vector along the line, it follows that

( )

M M x y z x y z z x y z w x y

u u u 0

1 2 3

→

· + − ( , , ), ( , , ), ( , , ) ( , ) (f39)

Expanding in power series the even function

3

u

with respect to z -variable, it follows

3

2

u

x y z w x y z x y ( , , ) ( , ) ( , ) ... · + + β (f40)

Using eqs.(f37) and (f40), it follows that

46

M M z

z

x y z

z

x y z z x y

z

z

x y

z

x y z x y

u u

u u

0

1

0

2

0

3

1

0

2

0 1

2

→

· + +

¸

¸

_

,

· + +

¸

¸

_

,

∂

∂

∂

∂

β

∂

∂

∂

∂

β

( , , ), ( , , ), ( , ) ...

( , , ), ( , , ), ( , ) ...

(f41)

Hence the direction is

M M

M M

z

x y

z

x y z x y

z z

z x y

u u

u u

0

0

1

0

2

0 1

2

1

2

2

2

1

2

2

→

→

·

+ +

¸

¸

_

,

¸

¸

_

,

+

¸

¸

_

,

+ + +

¸

¸

_

,

| | | |

( , , ), ( , , ), ( , ) ...

( , ) ...

∂

∂

∂

∂

β

∂

∂

∂

∂

β

(f42)

The normal vector on the median surface is

→

· − −

¸

¸

_

,

¸

¸

_

,

+

¸

¸

_

,

+

n

w

x

w

y

w

x

w

y

∂

∂

∂

∂

∂

∂

∂

∂

, ,1

2 2

1 . (f43)

Comparing eq.(f42) to eq.(f43), the supplemental hypothesis of BERNOULLI is satisfied if

∂

∂

∂

∂

∂

∂

∂

∂

1

0

2

0

3

u u

u

z

x y

w

x z

x y

w

y

x y z w x y ( , , ) , ( , , ) , ( , , ) ( , ) · − · − · , (f44)

i.e., by using eq.(f37), the displacement vector has the horizontal components equal to

1 2 3

u u u

x y z z

w

x

x y z z

w

y

x y z w x y ( , , ) , ( , , ) , ( , , ) ( , ) · − · − ·

∂

∂

∂

∂

. (f45)

From eq. (f45), the components of the strain tensor are

11

2

2

12

2

22

2

2

33

0

ε ε ε ε

∂

∂

∂

∂ ∂

∂

∂

· − · − · − · z

w

x

z

w

x y

z

w

y

, , , . (f46)

The LAPLACE operator in horizontal components is denoted by

∆

*

· +

∂

∂

∂

∂

2

2

2

2

x y

. (f47)

Hence the trace of the strain tensor is

θ

ε ε ε ε

· · + + · − tr z w

11 22 33

∆

*

. (f48)

F8) HOOKE’s law for a thin plate.

By using eqs.(f46) and (f48) it follows that

11

2

11

2

2

2

σ ε ε

λ µ λ∆ µ

∂

∂

· + · − +

¸

¸

_

,

tr z w

w

x

*

. (f49)

But

z

h

z

h

h

dz

h

2

1

2

2

2

3

·

−

+

·

∫

(f50)

Hence

11 11

2

3

2

2

2

2

12

2

2

2

Μ

· · − +

¸

¸

_

,

· − +

¸

¸

_

,

z

h

w

w

x

H

w

w

x

σ

λ∆ µ

∂

∂

λ∆ µ

∂

∂

* *

. (f51)

In the same way,

47

12 12

2

2

3

2 2

6

2

Μ

· · − · − z

h w

x y

H w

x y

σ

µ

∂

∂ ∂

µ

∂

∂ ∂

, (f52)

22 22

2

3

2

2

2

2

12

2

2

2

Μ

· · − +

¸

¸

_

,

· − +

¸

¸

_

,

z

h

w

w

y

H

w

w

y

σ

λ∆ µ

∂

∂

λ∆ µ

∂

∂

* *

(f53)

Substituting eqs.(f51)-(f53) into eq.(f35) it follows the equation of Sophie GERMAIN:

D w gH

j

x y

s

x y

H j

x y

s

x y

j

x y

s

x y

q q

q q q q

∆ ∆

* *

( , ) ( , )

( , ) ( , )

,

( , ) ( , )

,

· + +

+ −

¸

1

]

1

+ −

¸

1

]

1

¹

'

¹

¹

¹

¹

;

¹

¹

¹

ρ

3 3

2

1 1

1

2 2

2

, (f54)

where

D

h E H

· + ·

−

+ −

( )

( )

( )( )

λ µ

ν

ν ν

2

2

3

3

1

1 1 2

3

12

(f55)

represents the flexural rigidity of the plate.

So, obtaining the median surface asks someone to solve a bi-harmonic equation, with certain boundary conditions on the upper/

lower faces of the plate. The equation (f55) will be solved in same cases of particular importance in real cases.

EXERCISE. Find the expressions of

13

Σ

and

23

Σ

for a thin plate into a bending state.

F.9) The infinite, 1-dimensional (1-D) plate. The flexure of the lithosphere.

Consider an infinite extended plate along y -co-ordinate. The component

2

u

of the displacement vector is equal to zero,

and the rest of the components does not depend on y -co-ordinate. In this case, the plate is assumed to be in a cylindrical

bending state. Neglecting the horizontal loads, eq.(f55) becomes

D

d w

d x

gH

j

x

s

x

q q

4

4 3 3

· + + ρ ( ) ( ) . (f56)

For the case presented in Fig.F2, on the upper face of the plate is acting the load P due to the relief and the lithostatic pressure,

i.e.

3

s

x P

r

g w x h

q

( ) ( ( ) ) · + − ρ , (f57)

where ρ

r

is the density of the filling sediments (assumed to be homogeneous ones) placed between the reference plane of

elevation equal to zero and the upper surface of the plate. On the down face of the plate, it is acting the pressure of the liquid of

density equal to ρ

m

,i.e.

3 0

j

x

m

g w x h

q p

( ) ( ( ) ) · − + ρ , (f58)

where p

o

is an unknown constant. Eq.(f59) becomes

D

d w

d x

P

m r

w x

r

gh

m

gh gH

p

4

4 0

· − − + − − + ( )g ( ) ρ ρ ρ ρ ρ . (f59)

It is assumed that in the absence of the relief (i.e. P=0), the non-deformed plate (i.e. w x ( ) ≡ 0) is in an equilibrium state under

the action of its own weight and of the pressure of the liquid, i.e.

0

0

p

r

gh

m

gh gH − − + · ρ ρ ρ (f60)

48

Fig.F2. The flexure of the lithosphere under the action of the relief (P), of the lithostatic pressure Γ

L

and of the pressure of the

liquid Γ

M

.

It results the flexure equation of the 1-D plate :

[ ]

d w

d x

D

P

m r

w x

4

4

1

· − − ( )g ( ) ρ ρ (f61)

Let

( ) ρ ρ α

m r

g D − · / / 4

4

, (f62)

where α is the flexural parameter of the plate. Eq.(f61) becomes

d w x

d x

P

D

w x

4

4

4

4

( )

( ) · −

α

(f63)

F.10) Exterior forces on the lateral surface of the plate. Buckling.

To derive Sophie GERMAIN equation (f56) for the bending state, exterior forces acting on the lateral surface of the plate

have been ignored, especially those placed into the median plane. Consider a very thin plate simply leaning like in Fig.F3. The

load P is absent and the plate is infinite developed in a direction perpendicular on the plane of the figure. An element of the

plate having the length equal to unit along that direction is considered too. Let h be the thickness of the plate. The forces per

unit length along the above direction are denoted by t

→

N , being derived from the stress σ

c

( positive for compression ) by

N

c

h · σ (f64)

Fig.F3. A plate simply leaning, subject to the action of forces placed into the median plane.

(a)A lateral view. (b) A view from above.

If the forces t

→

N are small ones, the plate will be deformed according to a plane state, attempting a final configuration similar

to Fig.1b. If the forces t

→

N are above a certain critical value, the plate loses suddenly its equilibrium state, attending a

deformation state like Fig.F4a, (or in the contrary sense, i.e. symmetrically with respect to the line of its supports. The

displacement field in this case is similar to the cylindrical bending. That phenomenon represents the buckling of the plate. It

characterises very thin plates or bars.

49

Fig.F4. (a) The equilibrium of a buckled plate acted by the forces t

→

N and by the reactions of the supports.

(b) The equilibrium of a buckled plate acted by a vertical load p and by the reactions of the supports.

(c)The equilibrium of a buckled plate acted by the forces t

→

N and by a load q having the same magnitude but a

contrary sense with respect to the load presented in Fig.F4 (b).

In order to use the previous results, it is necessary to find when the mechanical state of stress / strain corresponding to the

presence of the lateral forces is identical to the mechanical state of stress / strain corresponding to an unknown vertical load

p p x · ( ) (Fig.F4b). The equilibrium condition for the case shown in Fig.4a is

N

R

→

+

→

·

→

∑ ∑

( ) 1

0 , (f65)

and the equilibrium condition for the case shown in Fig.F4b is

p

R

→

+

→

·

→

∑ ∑

( ) 2

0 , (f66)

where

( ) 1

R

→

∑

and

( ) 2

R

→

∑

are the reactions of the supports in the above cases. Because the mechanical state of stress is

identical in both cases, particularly in the neighbourhood of the supports, the reactions will be the same, i.e.

( ) ( ) 1 2

R R

→

·

→

∑ ∑

. (f67)

It follows that

N p

→

+ −

→

·

→

∑ ∑

( ) 0 . (f68)

Hence for the corresponding state of stress / strain, the plate is in an equilibrium state if it is acted by the lateral forces and by a

vertical load q p

→

· −

→

, in the absence of the supports (Fig.F4c). Consider a plate element having the horizontal length equal

to dx and the ends denoted by A and B (Fig.F5).

Fig.F5. The equilibrium of a plate element due to the load q and to the internal tensions.

50

Because the plate is a thin one, the tangential efforts are neglected. At the point A, it is acting a force (per unit length) denoted

by t

→

, representing a normal effort, tangent to the plate. The angle between t

→

and the horizontal axis is denoted by ϕ. At the

point B, it is acting the effort t dt +

→

, making an angle equal to ϕ+dϕ with the horizontal axis. The equilibrium conditions are

( ) ( )

( ) ( )

t t dt d

t t dt d qdx

cos cos

sin sin

ϕ ϕ ϕ

ϕ ϕ ϕ

− + + ·

− + + − ·

0

0

(f69)

For small angles ϕ, it follows that

( )

( )

cos cos

sin , sin

ϕ ϕ ϕ

ϕ ϕ ϕ ϕ ϕ ϕ

ϕ ϕ

≅ + ≅

≅ + ≅ +

≅ · −

d

d d

tg

dw

dx

1

(f70)

The first equation in (f69) gives

dt

dx

· 0 (f71)

Hence t has no variation along x-axis, i.e.

t N · (f72)

The second equation in (f69) gives

( ) t t d qdx ϕ ϕ ϕ − + − · 0 (f73)

By using (f74) and (f70c) it follows that

q N

d

dx

N

d w

dx

· − ·

ϕ

2

2

. (f74)

Hence the forces on the lateral faces of the plate are mechanical equivalent to a vertical load equal to

p p x N

d w

dx

· · − ( )

2

2

. (f75)

It follows that eq. (f63) has the next general form

d w

d x

N

D

d w

d x

w

P

D

4

4

2

2

4

4

+ + ·

α

(f76)

F.11) The buckling of a simply leaning thin plate.

Consider the plate in Fig.F3. For simplicity, it is assumed that c · 0 (i.e. the plate is leaning just at its ends). The buoyancy

force and the vertical loads are neglected. Equation (f76) becomes

d w

d x

N

D

d w

d x

4

4

2

2

0 + · , (f77)

together with the next conditions:

-at the end point having x=0: w d w dx · · 0

2 2

0 , / (f78)

-at the end point having x=a: w d w dx · · 0

2 2

0 , / (f79)

The equations (f77)-(f79) has the trivial solution w ≡ 0 . It follows to find a critical buckling value N N ·

*

in order the

system (f77)-(f79) to have further non-trivial solutions. Successively, equation (f77) can be written as

d

d x

d w

d x

N

D

w

2

2

2

2

0 +

¸

¸

_

,

· , (f80)

d w

d x

N

D

w C x C

2

2

1 2

+ · + , (f81)

where C C

1 2

, are two integration constants, vanishing according to (f78)-(f79). Hence (f81) is

51

d w

d x

N

D

w

2

2

0 + · , (f82)

having the solution

w x C N Dx C N Dx ( ) sin( / ) cos( / ) · +

3 4

. (f83)

From (f78) it follows that C

4

0 · , while from (f79) it follows the critical values

N

k

D k a k

*

( / ) , , , ... · · π

2

1 2 (f84)

The lowest critical value is obtained for k · 1.

EXERCISES.

(1)Perform a study for the buckling of a 1-D plate having an embedded end point, the other being free.

(2)Perform a study for the buckling of a 1-D plate having both end points free. The plate is simply leaning at 1/3 from its length

with respect to its left end.

(3)Perform a study of the simply leaning 1-D plate in the presence of the buoyancy force.

(4)Modify the equation of Sophie GERMAIN for the 2-D plate in the presence of lateral forces.

(5)Perform a study for the buckling a 2-D rectangular plate, simply leaning at all its sides.

F.12) The infinite extended 1-D plate.

By integrating both sides of eq.(f76) it follows

D

d w

dx

dx N

d w

dx

dx

m

w x P x

4

4

2

2

−∞

+∞

+

−∞

+∞

+ −

−∞

+∞

·

−∞

+∞

∫ ∫ ∫ ∫

( )g ( )dx ( )dx ρ ρ , (f85)

Because w and its derivatives of any order are vanishing at infinite, the first two integrals in (f85) are vanishing too. It follows

that the area bounded by the median curve (the flexure) and the horizontal x-axis is proportional to the load due to the relief,

irrespective the presence of the lateral forces:

w x

m

P x ( )dx

( )g

( )dx

−∞

+∞

·

−

−∞

+∞

∫ ∫

1

ρ ρ

. (f86)

Let an approximation of the relief be a set of msteps, each one of height equal to h

j

and density equal to ρ

j

, i.e.

[ ]

P x P

j

x

j

m

P

j

x

j

gh

j

pentru x a

j

b

j

in rest

( ) ( ) , ( )

, ,

,

·

·

·

∈

¹

'

¹

¹

¹

∑

1

0

ρ

. (f87)

Equation (f86) becomes

w x

m

j

h

j

b

j

a

j

j

m

( )dx ( )

−∞

+∞

·

−

−

·

∫

∑

1

1

ρ ρ

ρ . (f88)

Hence the area bounded by the flexural curve and the horizontal axis is a linear combination of the areas approximating the

relief. In real cases, the flexural curve can be outlined along a finite interval denoted by

[ ]

−L L , , hence an upper bound for

the difference of the densities can be obtained as

ρ ρ ρ

m j

h

j

b

j

a

j

j

m

w x

L

L

− < −

·

−

+

∑

∫

( ) ( )dx

1

. (f89)

Equation (f76) will be solved by using FOURIER transforms.

F.13) FOURIER transforms. Properties.

The direct FOURIER transform ( t.F.d. ) of a function f x ( ) is the new function Φ of variable u , defined as

Φ[ ]( ) ( ) exp( )dx , f u f x iux i · −

−∞

+∞

· −

∫

1 (f90)

The inverse FOURIER transform ( t.F.i. ) of a function Φ( ) u is the function f x ( ) defined as

52

f x f u x u iux ( ) [ [ ]( )]( ) ( ) exp( )du ·

−

· +

−∞

+∞

∫

Φ Φ Φ

1

1

2π

(f91)

Differentiating both sides of eq. (f91) with respect to x, it follows that the t.F.d. of the first derivative

df

dx

can be obtained by

multiplying the the t.F.d. of f x ( ) by iu . Hence the t.F.d. of the derivative

d w

dx

4

4

can be obtained by multiplying the t.F.d.

of w x ( ) by ( ) iu u

4 4

· . Consider two functions f and g of one variable. Their convolution product is

( * )( ) ( )g( )dy f g x f y x y · −

−∞

+∞

∫

. (f92)

Permuting the integrals, it follows that the direct FOURIER transform of the convolution product is the product pf the

transforms of both factors of the product, i.e.

Φ

Φ Φ

[ * ]( ) ( )g( )dy exp( )dx

( ) exp( )dy ( ) exp( )dz [ ] [ ]

f g u f y x y iux

f y iuy g z iuz f g

· −

−∞

+∞

¸

1

]

1

1

−∞

+∞

−

· −

−∞

+∞

−∞

+∞

− ·

∫ ∫

∫ ∫

. (f93)

F.14) Solution of the flexure equation by using FOURIER transforms.

The solution of eq.(f76) is the sum of two terms, a term corresponding to the homogeneous equation and a term corresponding

to a particular solution, i.e.

w x w

h

x w

p

x ( ) ( ) ( ) · + (f94)

A particular solution w

p

w

p

x · ( ) will be obtained applying the direct FOURIER transform to eq.(f76) and by using the

above presented properties of the FOURIER transform

u w

p

u

N

D

u w

p

u w

p

u

D

P u

4 2

4

4

1

Φ Φ Φ Φ [ ]( ) [ ]( ) [ ]( ) [ ]( ) − + ·

α

, (f95)

i.e.

Φ

Φ

[ ]( )

[ ]( )

/

w

p

u

D

P u

u Ku

·

− +

1

4 2

4

4

α

. (f96)

It is assumed that the value of the positive constant K N D · / is small enough. Consider the particular case when the load

due to the relief is a load concentrated at the origin of the axes, having the magnitude equal to unit. The direct FOURIER

transform of this load is equal to unit too. The corresponding solution, denoted by w

G

, represents the elastostatic GREEN

function. It allows one to obtain the solution corresponding to an arbitrary load of magnitude equal to P . Hence

Φ[ ]( )

/ / ( / )

w

G

u

D

u K K

·

−

¸

¸

_

,

+ −

1 1

2

2

2

4

4

2

2

α

(f97)

It is assumed that the next condition is satisfied

K < 4

2

/ α (f98)

Let

A K B K · + · − 1

2

4 1

2

4 / / , / / α α (f99)

But

53

( ) ( )

( ) ( )

1

4 2

4

4

2

8

2 2 2 2

2

4

1

2 2

1

2 2

u Ku

A

u

u A B

u

u A B

u A B u A B

− +

·

+ +

−

− +

¸

1

]

1

1

+

+ +

+

− +

¸

1

]

1

1

/ α

α

α

. (f100)

The next result is valid (Rîjic and Gradstein 1955)

z z

z x

xdx z z

sgn Re( )

cos exp( ) ,

2 2

0

2

0

+

∞

· − ≠

∫

π

. (f101)

By using (f101), the next inverse FOURIER transforms are obtained

( )

( )

Φ

−

t +

¸

1

]

1

1

· −

1

1

2 2

1

2

u A B

x

B

Bx iAx ( ) exp m . (f102)

( ) ( )

( ) ( )

Φ

−

+ +

+

− +

¸

1

]

1

1

· −

1

1

2 2

1

2 2

1

u A B u A B

x

B

Bx Ax ( ) exp cos . (f103)

By using the property of the derivative, it follows

( )

[ ]

( )

Φ

1

2

2 2

B

d

dx

Bx iAx u

iu

u A B

exp ( ) − −

¹

'

¹

¹

;

¹

·

+ +

. (f104)

In the same way

( )

[ ]

( )

Φ

1

2

2 2

B

d

dx

Bx iAx u

iu

u A B

exp ( ) − +

¹

'

¹

¹

;

¹

·

− +

. (f105)

Subtracting eq.(f104) from eq.(105), it follows that

( ) ( )

[ ]

( ) ( )

Φ

1

2 2 2 2

B

d

dx

Bx Ax u

u

u A B

u

u A B

exp sin ( ) −

¹

'

¹

¹

;

¹

·

− +

−

+ +

. (f106)

Hence

( ) ( )

( ) ( )

[ ]

( ) ( ) ( )

Φ

−

− +

−

+ +

¸

1

]

1

1

· − · − −

¸

1

]

1

1

2 2 2 2

1

u

u A B

u

u A B

B

d

dx

Bx Ax Bx

A

B

Ax Ax exp sin exp cos sin

. (f107)

Using the above results, it follows after some elementary computations that

( )

( ) ( )

Φ Φ [ ]( ) exp

sin cos

w

G

u

D

Bx

Ax

A

Ax

B

· − +

¸

1

]

1

1

¹

'

¹

¹

¹

¹

;

¹

¹

¹

α

2

8

, (f108)

( )

( ) ( )

w

G

x

D

Bx

A x

A

Ax

B

( ) exp

sin cos

· − +

¸

1

]

1

1

α

2

8

, (f109)

It can be observed that w

G

→ ∞ for B→0 , corresponding to the buckling of the infinite plate in the presence of a lateral

compressive stress. From (f96) and (f97) it follows that

Φ Φ Φ [ ] [ ] [ ] w

p

P w

G

· . (f110)

Hence the solution for an arbitrary load is the convolution of the load due to the relief and the function given by (f109),

representing a general property of the GREEN function:

w

p

x P w

G

x P y w

G

x y ( ) ( * )( ) ( ) ( )dy · · −

−∞

+∞

∫

. (f111)

54

For the approximation of the relief represented by eq.(f87), it follows that

( ) ( )

[ ]

w

p

x

m

j

h

j

I b

j

x I a

j

x

j

m

( )

( )

·

−

− − −

·

∑

1

4

1

ρ ρ

ρ . (f112)

where

I z z Bz Az

K

m

g D K

Az ( ) sgn( ) exp( ) cos( )

( ) / /

sin( ) · − − +

− −

¸

1

]

1

1

1

+

¹

'

¹

¹

¹

¹

;

¹

¹

¹

2

2

4

2

ρ ρ

(f113)

The solution of the homogeneous equation can be immediately derived as

w

h

x C Ax C Ax Bx C Ax C Ax Bx ( ) [ cos( ) sin( )]exp( ) [ cos( ) sin( )]exp( ) · + − + +

1 2 3 4

(f114)

Hence the general solution is

( ) ( )

[ ]

w x

m

j

h

j

I b

j

x I a

j

x

j

m

C Ax C Ax Bx C Ax C Ax Bx

( )

( )

[ cos( ) sin( )]exp( ) [ cos( ) sin( )]exp( )

·

−

− − −

·

+ + − + +

∑

1

4

1

1 2 3 4

ρ ρ

ρ

(f115)

It follows to find the unknown coefficients C C C

1 2 3

, , and C

4

in some particular cases. For the infinite plate, the flexure

Wis subject to the next conditions:

lim ( )

x

W x

→t∞

· 0 (f116)

Hence the coefficients C C C

1 2 3

, , and C

4

are vanishing and the general solution is just the particular solution represented

by eq.(116). In the case of the semi-infinite plate the flexure Wis subject, for example, to the next conditions:

lim ( )

x

W x

→∞

· 0 (f117)

W W ( ) 0 0

0

+ · (f118)

d W

dx

W D

M

2

2

0 0

0

0

( )

’’

/ + · · − (f119)

where W W

0 0

,

’’

and

0

M

(positive when acting into a clockwise sense) are the values of the flexure, that of the second

derivative of the flexure and that of the bending moment respectively at the left end of the plate where the origin of the x-axis is

selected. It follows

( ) ( )

[ ]

( ) ( ) ( ) ( )

[ ]

C W

j

gh

j

I b

j

I a

j

j

m

C

AB ABD

j

gh

j

Ba

j

Aa

j

Bb

j

Ab

j

K

j

m

C W

1 0

1

2

1

2

1

4 2

1

1 0

· − −

·

· − − − −

·

−

¸

1

]

1

1

1

∑

∑

ρ

ρ

,

exp sin exp sin

’’

(f120)

and

C C

3 4

0 · · . (f121)

A finite plate of variable thickness can be approximated in real cases by a sum of n elements having constant thickness and

homogeneous elastic properties. To obtain the values of the unknown coefficients C C C

1 2 3

, , and C

4

for each element,

proper conditions have to be verified at the ends of each element. A finite element algorithm based on the continuity of the

values of the flexure, of its first derivative, of the bending moment and of the share force has been derived by Ivan (1997).

EXERCISE. Derive the expression of w x ( ) for a load due to a relief having the equation

P P x gh x

gh x pentru x

in rest

· · ·

∈ − +

¹

'

¹

( ) ( )

sin( / ) , [ / , / ]

,

ρ

ρ π λ λ λ

0

2 2 2

0

, (f122)

where h

0

is the amplitude of the relief and λ is its wave-length.

55

F.15) Finite plates.

In the case of the finite plates, the boundary conditions on the contour of the plate are essential ones in order to obtain the

values of the integration constants. Some particular, most common cases will analysed in detail. As a consequence, a previous

examination of the mean values Σ

ij

and Μ

ij i, j , , , . ·1 2 3

it is necessary.

F.15 a) Significance of Σ

ij

and Μ

ij

for the bending state.

According to the definition of the stress tensor elements,

ij

σ

is the projection along j -axis of the surface force acting on

the surface having the outer pointing normal equal to

→

i

n

. Because the elements

13 23

σ σ

, are even functions, the

mean values

13 23

Σ Σ

, are representing share forces, acting like in Fig.F3.

Fig.F3. The mean values Σ

13

(a) and Σ

23

(b). Both of them are share forces.

Similar considerations allow one to conclude that the mean values

11

M

and

22

M

are bending moments, while

12 21

Μ Μ

, are torsion moments (Fig.F4).

Fig.F4. The bending moments M

11

(a) and M

22

(b). The torsion moment M

12

acting on the side having the outer pointing

normal 1 is presented in Fig.4c. A similar torsion moment is acting on the side having the normal 2, but it is not presented in the

figure.

F.15 b) The rectangular plate. Boundary conditions. LÉVY ‘s solution.

Let consider a rectangular plate having the sides equal to 2a and 2b respectively (Fig.F5). Consider, for example, the side

AB, having the equation x a y b b · ∈ − , [ , ]). Among most commonly used boundary conditions are

-the embedded side: the flexure of the plate and the derivative of the flexure are both equal to zero

w x y w x ( , ) , · · 0 0 ∂ ∂ ; (f123)

-the rotating side: the flexure of the plate and the bending moment are both equal to zero:

w x y ( , ) , · · 0

11

0

Μ

, (f124)

i.e.:

56

w x y w

w

x

( , ) ,

*

· + · 0 2

2

2

0 λ µ

∂

∂

∆ . (f125)

-the free side: the share force, the bending moment and the torsion moment are all equal to zero.

Fig.F5. The rectangular plate.

Similar conditions can be derived for plates of arbitrary shape.

As an example, consider the rectangular plate with two opposite articulated sides. Neglecting the lateral forces and the

gravity, it follows to solve the simplified equation of the flexure

∆ ∆ w q x y D · ( , ) / , (f126)

with the boundary conditions (f125) written for x a · t and a 2-D LAPLACE operator. Consider a particular solution having

the form

w x y f

k

y k x a

k

( , ) ( )sin( / ) ·

·

∞

∑

π

1

, (f127)

it follows that

w a y

w

x a

k

k

f

k

k x a

( , ) ,

sin( / )

t ·

· −

·

∞

∑

0

2

2

2

2

2

1

∂

∂

π

π

. (f128)

Hence

∂ ∂ ∂ ∂

2 2 2 2

0 w x w y for x a / / , · · · t . (f129)

It follows the conditions (f125) are fulfilled for the two opposite articulated sides. Substituting (f127) into (f126), it follows that

f

k

y

k

a

f

k

y

k

a

f

k

y

k x

a

k

q x y D

""

( )

"

( ) ( ) sin ( , ) / −

¸

¸

_

,

+

¸

¸

_

,

¸

1

]

1

1

·

∞

·

∑

2

2 4

1

π π π

. (f130)

The function q x y ( , ) is expanded in FOURIER series and the coefficients are identified. Both sides of eq.(f130) are

multiplied by sin , , ,...

j x

a

j

π

· 1 2 The result is integrated on the interval [ , ] −a a , taking into account that

sin sin

,

,

k x

a

a

a

j x

a

dx

pentru j k

a pentru j k

π π

−

+

·

≠

·

¹

'

¹

∫

0

. (f131)

Hence

f

k

y

k

a

f

k

y

k

a

f

k

y

Da

q x y

k x

a

a

a

dx k

""

( )

"

( ) ( ) ( , )sin , , ,... −

¸

¸

_

,

+

¸

¸

_

,

·

−

+

·

∫

2

2 4

1

1 2

π π π

. (f132)

In the beginning, the next homogeneous equation is solved, i.e.

F

k

y

k

a

F

k

y

k

a

F

k

y k

""

( )

"

( ) ( ) , , ,... −

¸

¸

_

,

+

¸

¸

_

,

· · 2

2 4

0 1 2

π π

. (f133)

The characteristic equation is

r

k

a

r

k

a

deci r r k a r r k a

4

2

2

2

4

0

1 2 3 4

−

¸

¸

_

,

+

¸

¸

_

,

· · · − · ·

π π

π π , / , / , (f134)

hence the solution of the homogeneous equation is

57

F

k

y A

k

k y

a

B

k

k y

a

y C

k

k y

a

D

k

k y

a

( ) cosh sinh cosh sinh · + + +

¸

¸

_

,

π π π π

, (f135)

where A

k

B

k

C

k

D

k

k , , , , , , ... ·1 2 are some constants following to be obtained from the boundary conditions on the

other two (non-articulated) sides of the plate. The general solution of equation (f132) is the sum of (f135) and a particular

solution. The last one can be obtained by the usual techniques (e.g. CAUCHY method). The above approach is due to LÉVY.

F.16) Vibrations of a plate laying on a viscous substratum.

In this chapter, the flexure is considered as a function of both spatial co-ordinates and time. The corresponding differential

equation is derived, being solved in the case of the 2-dimensional plate.

a) The differential equation.

As usually, a co-ordinate system is used having the horizontal axes x and y. The z -axis is positive downward, having the unit

vector denoted by

→

e

z

. By applying the mean-value operator, the equations of motion for the bending state are

∂ ∂ ∂ ∂ ρ

xx

x

xy

y

x

d

x y t

x

u

x y t

xz

z

x

q q

u Μ Μ Σ

+ + −

¸

1

]

1

− ·

••

( , , ) ( , , ) / 2 , (f136)

∂ ∂ ∂ ∂ ρ

xy

x

yy

y

y

d

x y t

y

u

x y t

yz

z

y

q q

u Μ Μ Σ

+ + −

¸

1

]

1

− ·

••

( , , ) ( , , ) / 2 , (f137)

and

∂ ∂ ∂ ∂ ρ ρ

xz

x

yz

y

h

z

d

x y t

z

u

x y t g

z

q q

u Σ Σ

+ + +

¸

1

]

1

+ ·

••

1

2

( , , ) ( , , ) . (f138)

It will be assumed that Bernoulli’s hypothesis is valid for all time. Hence the displacement vector has the elements

x

x y z t z

w

x

y

x y z t z

w

y

z

x y z t w x y t

u u

u

( , , , ) , ( , , , ) , ( , , , ) ( , , ) · − · − ·

∂

∂

∂

∂

. (f139)

Substituting (f139) into (f136)-(f138) it follows after elementary computations that

D w gH H w H w

z

l

x y t

z

u

x y t

H

x

l

x y t

x

u

x y t x

y

l

x y t

y

u

x y t y

q q

q q q q

∆ ∆ ∆

* * *

( , , ) ( , , )

( , , ) ( , , ) ( , , ) ( , , )

· −

••

+

••

+

+ −

¸

1

]

1

+ −

¸

1

]

1

¹

'

¹

¹

;

¹

+

ρ ρ

ρ

∂ ∂ ∂ ∂

12

3

2

, (f140)

As usually, the horizontal loads for the upper face of the plate are neglected, the surface forces being assumed to be

x

u

y

u

z

u

F

gw P w g

q q q

· · · + −

••

0 0 1 , , ( / ) ρ , (f141)

where ρ

F

is the density of the filling sediments and P is the load. The last term in (f141) is an inertial one. For the material

below the plate, the next constitutive equation is assumed

σ ε ε ε

ρ λ µ η · − + + + +

•

[ ( )] p

M

g H w

M

tr

M M

0

2 2

1 1

(f142)

where p

0

is a reference pressure and ρ

M

is the density of the material below the plate. The LAMÉ elastic coefficients are

λ µ

M M

, and η

M

is the viscosity. Hence the loads on the lower face of the plate are

58

x

l

M

w

x

M

w

x

y

l

M

w

y

M

w

x

z

d

p

H

M

w

M

g H w

q q

q

· +

•

· +

•

· − + +

µ

∂

∂

η

∂

∂

µ

∂

∂

η

∂

∂

λ ρ

, ,

*

( )

0

2

∆

. (f143)

Again, a correction due to the compressive horizontal stresses σ σ

x

c

y

c

, acting along the x- and y-axes respectively at the ends

of the plate follows to be considered further. The reference pressure p

0

is selected in order the flexure w to vanish in the

absence of the load P . Finally, a generalisation of the Sophie GERMAIN equation for a time dependent flexure is obtained as

D w

H

M M

w H

x

c

w

x

y

c

w

y

M F

gw

P

M

H

w H P g w H w

∆ ∆ ∆

∆ ∆

* * *

( )

*

( / )

*

+ −

¸

¸

_

,

+ +

¸

¸

_

,

+ −

· +

•

− +

••

+

••

2

2

2

2

2

2 12

3

λ µ σ

∂

∂

σ

∂

∂

ρ ρ

η ρ

ρ

(f144)

b) The rectangular plate with 3 welded sides.

For usual materials λ µ

M M

· , hence the second term on the left side of (f144) vanishes. Because the load is mainly

represented by the relief, having the elevations much smaller then the thickness of the plate, the inertial term is usually

negligible on the right side of (15). Let w

e

w

e

x y · ( , ) be the equation of the flexure corresponding to the state of

equilibrium in the presence of the load, i.e.

D w

e

H

x

c

w

e

x

y

c

w

e

y

M F

gw

e

P ∆ ∆

* *

/ / ( ) + +

¸

¸

_

,

+ − · σ ∂ ∂ σ ∂ ∂ ρ ρ

2 2 2 2

(f145)

Consider the difference

δ δ · · − ( , , ) ( , , ) ( , ) x y t w x y t w

e

x y (f146)

It follows that

∆ ∆ ∆ ∆

* * * *

δ

ρ

σ

∂ δ

∂

σ

∂ δ

∂

ρ ρ

ρ

δ δ

η

ρ

δ δ + +

¸

¸

_

,

+

−

·

••

+

•

−

•• ¸

¸

_

,

12

2

1

2

2

2

2

2

12

3 2

1

2

6

2

12

2

a H

x

c

x

y

c

y

M F

g

H a a

M

H H

(f147)

where a denotes the velocity of P-waves through the plate.

Consider the lengths of the sides are L

x

L

y

, respectively. Suppose the plate is welded according to

δ

δ

δ

( , , ) ,

( , , ) ,

( , , ) ,

x t with x L

x

y t with y L

y

L

x

y t with y L

y

0 0 0

0 0 0

0 0

· ≤ ≤

· ≤ ≤

· ≤ ≤

, (f148)

at any time. A particular solution satisfying (f148) is

( ) δ π π τ

mn

x y t m x L

x

n y L

y n

t m n ( , , ) sin / sin[( . ) / ] ( ) , , , , ... · − · 05 1 2 (f149)

Substituting (f149) in (f147) it follows the modes are damped harmonic, i.e.

τ ζ τ

π

τ

••

+

•

+ · 2

4

2

2

0

mn

T

mn

, (f150)

For m n , , , ... · 1 2 the periods are

T

mn

H

a

A

mn

A

mn x

c

m H

x y

c

n H

y

a gH a

·

+

− + − +

2

12

2

12

2 2 2

05

2 2 2

12

2

π

π σ σ ρ ρ [ ( . ) ] / ( )

*

/

, (f151)

59

and the decay constants are

ζ

η

ρ

mn

M

H

A

mn

A

mn

·

+

3

2

12

, (f152)

where

H

x

H L

x

H

y

H L

y

M F

· · · − / , / ,

*

( ) / ρ ρ ρ ρ (f153)

and

A

mn

m H

x

n H

y

· + −

¸

1

]

1

π

2 2 2

05

2 2

( . ) . (f154)

According to (f139), the modes are both toroidal and spheroidal.

For each mode, a critical viscosity can be found from

ζ π

mn

cr

T

mn

· 2 / , (f155)

i.e.

η

πρ

mn

cr

H

T

mn

A

mn

A

mn

·

+ 2

2

3

12

. (f156)

For values of viscosity less than the critical value (f156), the motion of the plate is represented by a sum of damped oscillations.

The decay constants are obtained from (f152) and the periods are

T

mn

T

mn mn mn

cr *

/ / · −

¸

¸

_

,

1

2

ζ ζ (f157)

For values of viscosity greater than the critical value, the motion of the plate is aperiodic and the characteristic roots of (21) are

− + −

¸

¸

_

,

ζ ζ π

mn mn

T

mn

2

4

2 2

/ and − − −

¸

¸

_

,

ζ ζ π

mn mn

T

mn

2

4

2 2

/ . It follows that for large scale of times, the

solution of (f150) behaves like exp( ) −Ct , where the decay constant is

C T · − − ζ ζ π

11

11

2

4

2

11

2

/ , (f158)

all the other terms being faster attenuated. In many real applications, an approximate value of (f159) is

C

T

T

T

·

+ −

≅

4

2

11

2

11

11

2

4

2

11

2

2

2

11

11

2

π

ζ ζ π

π

ζ

/

/

(f159)

With regard to (f149), the decay constant (f158) or (f159) can be obtained as the ratio between the amplitude of the velocity

∂δ ∂ / t and the amplitude of δ . By using eq.(f152), that ratio can be used to estimate the mean viscosity of the material below

the plate

η

ρ π

M

H

C

T

A

mn

A

mn

·

+

2

3

2

2

11

2

12

(f160)

A numerical application with respect to the Moesian Platform, a micro-plate bounded by Carpathians and Balkan Mountains

and by the Black Sea is presented in detail by (Ivan 1997a,b).

60

G) THE SPHERICAL SHELL

G.1) The model. BERNOULLI’s hypothesis. Displacement vector and strain tensor.

A spherical elastic, homogeneous shell having the elastic moduli equal to λ µ , is considered and the usual spherical co-

ordinate system ( ) r, , θ ϕ is used. However, some of the derived results are also valid in the case of a more general, non-elastic

spherical shell. In the initial state, the homogeneous density is denoted by ρ

0

and the median spherical surface of the shell has

the equation

r R · , (g1)

where R is the radius of the sphere. At a certain time during the deformation, the median surface will be

r R w t · − ( , , ) θ ϕ , (g2)

where w w t · ( , , ) θ ϕ is the flexure of the shell, positive downward. Hence the unit vector normal to the median surface at a

certain point of co-ordinates ( ) r R w · − , , θ ϕ is

n

r

w

r r

w

r

w w

e e e

→

·

→

+

→

+

→

¸

¸

_

,

+

¸

¸

_

,

+

¸

¸

_

,

¸

1

]

1

1

∂

∂θ

θ

θ

∂

∂ϕ

ϕ

∂

∂θ

θ

∂

∂ϕ

1

1

1

2

2

1

2

2

sin

/

sin

. (g3)

Neglecting the quantities of the second order, it follows that

n

r

w

R R

w

e e e

→

≅

→

+

→

+

→

∂

∂θ

θ

θ

∂

∂ϕ

ϕ

1

sin

. (g4)

Consider the initial, non-deformed state of the shell and two points of co-ordinates A R

0

( , , ) θ ϕ and B R h

0

( , , ) + θ ϕ ,

where H=2h is the thickness of the shell. It follows that the segment A B

0 0

→

has the unit vector equal to

r

e

→

. At an arbitrary

time, in the deformed state of the shell, the point A

0

is displaced to a point Ahaving the position vector equal to

r

A

R

r

u R t

e

→

·

→

+

→

( , , , ) θ ϕ , (g5)

while the point B

0

is displaced to the point Bhaving the position vector equal to

r

B

R h

r

u R h t

e

→

· +

→

+

→

+ ( ) ( , , , ) θ ϕ . (g6)

Here, u

→

is the displacement vector at a point of certain spherical co-ordinates. Assuming the shell is thin, quantities of the

order h

2

are neglected and the segment AB

→

has the unit vector equal to

r

B

r

A

r

B

r

A

u

r

r

r

u

r

u

r

e e e

→

−

→

¸

¸

_

,

→

−

→

≅ +

¸

¸

_

,

→

+

→

+

→

/ 1

∂

∂

∂

θ

∂

θ

∂

ϕ

∂

ϕ

. (g7)

The partial derivatives in (g7) are computed at the point ( , , ) R θ ϕ . It is assumed that Bernoulli’s hypothesis is valid for all

time. It follows that a segment inside the shell, which is initially normal to the median spherical surface, will be always normal

to the median surface during the deformation. From (g4) and (g7), it is supposed that the displacement vector has the elements

u

r

r t w t

u r t

r R

R

w

x t

u r t

r R

R

w

y t

( , , , ) ( , , )

( , , , ) ( , , )

( , , , )

sin

( , , )

θ ϕ θ ϕ

θ

θ ϕ

∂

∂θ

θ ϕ

ϕ

θ ϕ

θ

∂

∂ϕ

θ ϕ

· −

·

−

+

·

−

+

¹

'

¹

¹

¹

¹

¹

¹

¹

1

, (g8)

where x and y are two unknown functions representing the horizontal displacements of the points initially placed on the

median sphere. A further hypothesis on x and y will be later considered. It follows the elements of the strain tensor are

61

rr

u

r

r

ε

∂

∂

· · 0 , (g9)

r

r

u

r

u

r

u

r

x

r

θ

∂

∂θ

θ

∂

θ

∂

ε

· − +

¸

¸

_

,

· −

1

2

1

2

, (g10)

r

r

u

r

u

r

u

r

y

r

ϕ

θ

∂

∂ϕ

ϕ

∂

ϕ

∂

ε

· − +

¸

¸

_

,

· −

1

2

1

2 sin

, (g11)

θθ

∂

θ

∂θ

∂

∂θ

∂

∂θ

ε

· +

¸

¸

_

,

·

−

+ −

¸

¸

_

,

1

2

2

1

r

u

u

r

r R

rR

w

r

x

w , (g12)

θϕ

θ

∂

θ

∂ϕ

ϕ

θ

∂

ϕ

∂θ

∂

∂θ θ

∂

∂ϕ θ

∂

∂ϕ

∂

∂θ θ

ε

· − +

¸

¸

_

,

·

− ¸

¸

_

,

+ + −

¸

¸

_

,

1

2

1 1 1

2

1

r

u

u u

r R

rR

w

r

x y y

sin tan sin sin tan

, (g13)

and

ϕϕ

θ

∂

ϕ

∂ϕ

θ

θ

θ

∂

∂ϕ

θ

∂

∂θ θ

∂

∂ϕ θ

ε

· + +

¸

¸

_

,

·

−

+

¸

¸

_

,

+ − +

¸

¸

_

,

1 1 1

2

2

2

1 1 1

r

u

u

r

u r R

rR

w w

r

y

w

x

sin tan

sin

tan sin tan

(g14)

Hence the trace of the strain tensor is

( ) tr

rr

R r R

r

w

r

w x

y

ε ε ε ε

θθ ϕϕ

θ

θ

∂

∂θ

θ

∂

∂ϕ

· + + ·

−

∗

+ − + +

¸

1

]

1

( )

sin

sin sin ∆

1

2 , (g15)

where the two-dimensional LAPLACE operator is

∆

∗

·

¸

¸

_

,

+

¸

1

]

1

1

w

R

w w 1

2

1

2

2

sin

sin

sin

θ

∂

∂θ

θ

∂

∂θ θ

∂

∂ϕ

. (g16)

G.2) Quasi-mean values. Equations of motion.

Consider the stress tensor in spherical co-ordinates. For each element of the tensor, the corresponding quasi-mean value is

defined, for example, by

rr rr

H

r

R

rr

dr

R h

R h

Σ

· ·

−

+

∫ σ σ

1

. (g17)

In the same way, the corresponding quasi-moment is defined by

rr

H

r

R

r R

rr

dr

R h

R h

M

· −

−

+

∫

1

( )

σ

. (g18)

The equations of motion in spherical co-ordinates are

∂

∂

∂

θ

∂θ

θ

θ θ

∂

ϕ

∂ϕ

θθ ϕϕ

ρ ρ

∂

∂

∂

θ

∂

θ

∂

θθ

∂θ θ

∂

θϕ

∂ϕ

θθ ϕϕ

θ

ρ

∂

θ

∂

∂

ϕ

∂

ϕ

∂

θϕ

∂θ

θϕ

θ θ

∂

ϕϕ

∂ϕ

ρ

∂

ϕ

∂

σ σ σ σ

σ σ σ

σ σ σ

σ σ σ

σ σ σ σ σ

rr

r

rr

r

r

r

r

r r

r

r

g

u

r

t

r

r

r

r r r r

u

t

r

r

r

r r r r

u

t

+ + + + −

+

− ·

+ + + +

−

·

+ + + + ·

2

1

2

2

3

1

2

2

3

1

2

2

2

tan sin

sin tan

tan sin

. (g19)

In the above equations, ρ denotes the density of the shell in the deformed state. Assumed to be a negligible second order

effect, that density will be replaced by the initial density ρ

0

at the right side of (g19).

62

The first equation in (g19) will be multiplied by r

2

and the quasi-mean operator defined by (g17) will be applied. The next

two equations in (g19) will be multiplied by r and the quasi-moment operator defined by (g18) will be applied. In order to do

that, some intermediary results are necessary.

G.3) Integrals of the stress elements. Quasi-moments.

Let the stress values on the upper/lower faces of the shell be denoted as

U

R h t

σ σ

θ ϕ · + ( , , , ) (g20)

and, respectively, by

L

R h t

σ σ

θ ϕ · − ( , , , ) . (g21)

Elementary computations show that

1 1

2

2 1

2

1

2

H R

r

rr

r

r

rr

R h

R h

dr

R

H

h

R

rr

U

h

R

rr

L

∂

∂

σ

σ σ σ

+

¸

¸

_

,

−

+

· +

¸

¸

_

,

− −

¸

¸

_

,

¸

1

]

1

1

∫

. (g22)

In the case of the elastic shell, the HOOKE’s law leads to

r r

x r

r

x R

θ

µ

θ

µ

θ

µ

σ ε

· · − · − 2 / , /

Σ

. (g23)

Integrating by parts, it follows that

1

2

1

2

3 2

2

1

2

1

2

H

r

R

r R

r

r

R h

R h

dr

HR

r r R

r

R h

R h

x r R dr

R h

R h

R h

R

r

U

h

R

r

L

R

r

( ) ( ) ( ) −

−

+

· −

¸

1

]

1

−

+

+ −

−

+

¹

'

¹

¹

¹

¹

;

¹

¹

¹

· +

¸

¸

_

,

+ −

¸

¸

_

,

¸

1

]

1

1

−

∫ ∫

∂

θ

∂

θ

µ

θ θ θ

σ

σ

σ σ Σ

. (g24)

Also,

1

2

3

0

H

r

R

r R

r

r

R h

R h

dr ( ) −

−

+

·

∫

θ

σ

. (g25)

Similar relations are derived for

rϕ

σ

.

Also,

θθ

λ∆

µ ∂

∂θ

λ∆

µ ∂

∂θ

Μ

·

∗

+

¸

¸

_

,

−

−

+

·

∗

+

¸

¸

_

,

∫

w

R

w

H

r R dr

R h

R h

h

w

R

w 2

2

2

2

1

2

2

3

2

2

2

2

( ) . (g26)

In the same way, it follows that

θϕ

µ ∂

∂θ θ

∂

∂ϕ

Μ

·

¸

¸

_

,

h

R

w

2

3

2

2

1

sin

, (g27)

and

ϕϕ

λ µ

µ ∂

∂θ

Μ

∆ · +

∗

−

¸

1

]

1

1

h

w

R

w

2

3

2

2

2

2

2

( ) . (g28)

G.4) Integrals of the displacement vector.

Using (g8), it follows

1

2

0

2

2

0

1

1

3

2 2

2

H

r

R

u

r

t

dr

R h

R h

R

h

R

w

t

ρ

∂

∂

ρ

∂

∂

−

+

· − +

¸

¸

_

,

¸

1

]

1

1

∫

. (g29)

Also,

63

1

2

0

2

2

0

2

2

2

2

3

2

4

5R

4

2

3

2

H

r r R

R

u

t

dr

R h

R h

R

t

h

R

h w x h

R

( ) −

−

+

· +

¸

¸

_

,

+

¸

¸

_

,

¸

1

]

1

1

∫

ρ

∂

θ

∂

ρ

∂

∂

∂

∂θ

, (g30)

and

1

2

0

2

2

0

2

2

2

2

3

2

4

5

4

1 2

3

2

H

r r R

R

u

t

dr

R h

R h

R

t

h

R

h

R

w y h

R

( )

sin

−

−

+

· +

¸

¸

_

,

+

¸

¸

_

,

¸

1

]

1

1

∫

ρ

∂

ϕ

∂

ρ

∂

∂

θ

∂

∂ϕ

. (g31)

G.5) Equations of motion in quasi-mean values.

Let the first equation in (g19) be multiplied by r

2

. By applying the quasi-mean operator defined by (g17), it follows

( )

( ) ( )

1 1

2

1

2

1

2

0

1

1

3

2 2

2

sin

sin ( )

/ /

θ

∂

∂θ

θ

θ

∂

ϕ

∂ϕ

θθ ϕϕ

ρ

ρ

∂

∂

σ σ

r

r

RH

r g r dr

R h

R h

R

H

h R

rr

U

h R

rr

L

R

h

R

w

t

Σ

Σ

Σ Σ

+

¸

1

]

1

1

− +

¸

¸

_

,

−

−

+

+ + − −

¸

1

]

1

· − +

¸

¸

_

,

¸

1

]

1

1

∫

. (g32)

The next two equations in (g19) are multiplied by r and the quasi-moment operator defined by (g18) is applied. Hence

( ) ( )

1 1 1

2

1

2

1

2

0

2

2

1

3

2

1

5

4

2

3

2

R

r

R

h R

r

U

h R

r

L

R

t

h

R

h

R

w x h

R

∂

θθ

∂θ θ

∂

θϕ

∂ϕ

θ

θθ ϕϕ

θ

θ θ

ρ

∂

∂

∂

∂θ

σ σ

Μ

Μ

Σ

Μ Μ

+

¸

¸

_

,

− +

−

+ + + −

¸

1

]

1

·

¸

¸

_

,

+

¸

¸

_

,

¸

1

]

1

1

+

¸

¸

_

,

¹

'

¹

¹

¹

¹

;

¹

¹

¹

sin tan

/ /

(g33)

and

( ) ( )

1 1

2

1

2

1

2

1

2

0

2

2

1

3

2

1

5

4

1 2

3

2

R

r

R

h R

r

U

h R

r

L

R

t

h

R

h

R

w y h

R

∂

θϕ

∂θ θ

∂

ϕϕ

∂ϕ

ϕ

θϕ

θ

ϕ ϕ

ρ

∂

∂

θ

∂

∂ϕ

σ σ

Μ Μ

Σ

Μ

+

¸

¸

_

,

− + + + + −

¸

1

]

1

·

¸

¸

_

,

+

¸

¸

_

,

¸

1

]

1

1

+

¸

¸

_

,

¹

'

¹

¹

¹

¹

;

¹

¹

¹

sin tan

/ /

sin

. (g34)

G.6) Quasi-mean value of the shell density. The differential equation.

According to (g17), the quasi-mean of a constant is equal to that constant. Applying the quasi-mean operator to the mass

balance equation in the linear approximation gives

ρ ρ

ε

· −

0

1 ( ) tr . (g35)

It will be further assumed that the quasi-mean of the density is constant and equal to its initial value. From (g35) it follows that

tr

ε

· 0 , (g36)

i.e., using (g15),

∂

∂θ

θ

∂

∂ϕ

θ ( sin ) sin x

y

w + · 2 . (g37)

For the elastic shell,

σ ε ε

λ µ · + tr

1

2 . (g38)

64

By using (g9), it follows

θθ ϕϕ

λ µ

σ σ ε

+ · + 2( )tr , (g39)

or, using (g36),

θθ ϕϕ

Σ Σ

+ · 0 . (g40)

The term ( ) h R /

4

will be neglected. Using (g40) and substituting

rθ

Σ

and

rϕ

Σ

between (g32)-(g34) it follows

( )

( ) ( ) ( ) ( )

( ) ( )

1

2

2

2

2

2

1

2

2

1

2

1

2

1

2

1

2

1

2

1

2

1

2

1

2

1

2

R

H

h R

rr

U

h R

rr

L

R

h R

r

U

h R

r

L

R

h R

r

U

h R

r

L

HR

r gdr

R h

R h

∂ θ

θθ

∂θ

θ

∂ θ

θϕ

∂θ∂ϕ θ

∂

ϕϕ

∂ϕ

∂

∂θ

θ

ϕϕ

θ ∂

∂θ

θ

θ θ

∂

∂ϕ

ϕ ϕ

ρ

σ σ σ σ

σ σ

sin

sin

sin

sin

cos

sin

/ / sin / /

/ /

Μ

Μ

Μ

Μ

+

¸

¸

_

,

+ −

¸

¸

_

,

¸

1

]

1

1

1

+ + − −

¸

1

]

1

+ + + −

¸

1

]

1

¹

'

¹

¹

;

¹

+ + + −

¸

1

]

1

−

−

+

·

∫

ρ θ

∂

∂

∂

∂

0

2

3

2

2

1

2 2

2

sin

h w

t

h

R

w

t

∆

∗

¸

¸

_

,

− −

¸

¸

_

,

¸

1

]

1

1

¹

'

¹

¹

¹

¹

;

¹

¹

¹

(g41)

A correction due to the compressive horizontal stresses

NS

σ

,

WE

σ

acting at the ends of the shell, approximately along

the θ − and ϕ − axes respectively, follows to be further considered.

G.7) The buckling of a spherical shell.

Consider the element ABCD of the mean surface of the deformed shell from Fig.G1. It is centred at the point M, where the

local unit vectors are

r

e e e

→ → →

, ,

θ ϕ

. The centres of the lateral sides are denoted by M

k

k , , , , ·1 2 3 4. On the meridian

cross section M MM

4 2

is acting a normal compressive stress

WE

σ

, having the approximate direction from West to East,

and a tangential stress

NS

τ

, having the approximate direction from North to South. On the parallel cross section M MM

1 3

is acting a normal compressive stress

NS

σ

, having the approximate direction from North to South, and a tangential stress

WE

τ

, having the approximate direction from West to East. Let Φ be the angle between the normal vector to the meridian

section M MM

4 2

and the unit vector

ϕ

e

→

. Also, let Θ be the angle between the normal vector to the parallel section

M MM

1 3

and the unit vector

θ

e

→

. The concentrated force acting at the point M

1

is

→

· −

¸

¸

_

,

−

¸

¸

_

,

→

+ −

¸

¸

_

,

→

¸

1

]

1

1

1

¹

'

¹

¹

¹

+ −

¸

¸

_

,

−

¸

¸

_

,

→

+ −

¸

¸

_

,

→

¸

1

]

1

1

¹

;

¹

¹

¹

× − +

¸

¸

_

,

Σ

Φ Φ

Θ Θ

1

2 2 2

2 2 2 2

WE

WE

d d d

r

NS

NS

d d d

r

R w

w d

Hd

e

e

e e

σ

σ

τ

τ

∂

∂ϕ

ϕ ∂Φ

∂ϕ

ϕ

ϕ

∂Φ

∂ϕ

ϕ

∂

∂ϕ

ϕ ∂Θ

∂ϕ

ϕ

θ

∂Θ

∂ϕ

ϕ ∂

∂ϕ

ϕ

θ

cos sin

cos sin

(g42)

65

Fig.G1. A rectangular element of the mean surface of the deformed shell

In the same way, the concentrated force acting at the point M

3

is equal to

→

· − +

¸

¸

_

,

+

¸

¸

_

,

→

+ +

¸

¸

_

,

→

¸

1

]

1

1

1

¹

'

¹

¹

¹

+ +

¸

¸

_

,

+

¸

¸

_

,

→

+ +

¸

¸

_

,

→

¸

1

]

1

1

¹

;

¹

¹

¹

× − −

¸

¸

_

,

Σ

Φ Φ

Θ Θ

3

2 2 2

2 2 2 2

WE

WE

d d d

r

NS

NS

d d d

r

R w

w d

Hd

e

e

e e

σ

σ

τ

τ

∂

∂ϕ

ϕ ∂Φ

∂ϕ

ϕ

ϕ

∂Φ

∂ϕ

ϕ

∂

∂ϕ

ϕ ∂Θ

∂ϕ

ϕ

θ

∂Θ

∂ϕ

ϕ ∂

∂ϕ

ϕ

θ

cos sin

cos sin

(g43)

The concentrated forces acting at the points M

2

and M

4

are respectively equal to

→

· t

¸

¸

_

,

t

¸

¸

_

,

→

+ t

¸

¸

_

,

→

¸

1

]

1

1

¹

'

¹

¹

¹

+ t

¸

¸

_

,

t

¸

¸

_

,

→

+ t

¸

¸

_

,

→

¸

1

]

1

1

1

¹

;

¹

¹

¹

× −

¸

¸

_

,

t

Σ Θ Θ

Φ Φ

2 4

2 2 2

2 2 2

2

,

cos sin

cos sin

sin( )

m

m

NS

NS

d d d

r

WE

WE

d d d

r

R w

w d

d Hd

e e

e

e

σ

σ

τ

τ

∂

∂θ

θ ∂Θ

∂θ

θ

θ

∂Θ

∂θ

θ

∂

∂θ

θ ∂Φ

∂θ

θ

ϕ

∂Φ

∂θ

θ

∂

∂θ

θ

θ θ ϕ

(g44)

But

cos , cos , sin

sin

, sin Φ Θ Φ Φ Θ Θ ≅ ≅ ≅ ≅ ≅ ≅ 1 1

1 1

R

w

R

w

θ

∂

∂ϕ

∂

∂θ

. (g45)

Let p p · ( , ) θ ϕ a surface density of forces normal to the element of the shell, having the same mechanical effect as the

presence of the compressive stress. The force due to that density is equal to

→

· −

→

Σ

p

p R w d d

r

e

( , )( ) sin θ ϕ θ θ ϕ

2

. (g46)

It follows the deformed element of the shell is into an equilibrium state due to the action of the lateral stress and to the opposite

force −

→

Σ

p

, i.e.

→

+

→

+

→

+

→

−

→

·

→

Σ Σ Σ Σ

Σ

1 2 3 4

0

p

. (g47)

Hence the next three equations of equilibrium are obtained:

66

2cos

sin

sin

sin

sin sin

( , )

( )

sin

θ

θ ∂

∂θ

∂

∂θ θ

∂

∂ϕ

θ

∂

∂θ

∂

∂θ θ

∂

∂ϕ

∂

∂ϕ θ

∂

∂ϕ

∂

∂θ

θ ϕ θ

σ

τ

σ

τ σ

τ

−

¸

¸

_

,

+

¸

¸

_

,

+ +

¸

¸

_

,

+ +

¸

¸

_

,

· −

−

R

w

NS

w

WE

w

NS

w

WE

w

WE

w

NS

w

p

R R w

H

, (g48)

NS

R

w

NS NS

σ

σ τ

θ

θ ∂

∂θ

θ

∂

∂θ

∂

∂ϕ

2 0 cos

sin

sin −

¸

¸

_

,

+ + · , (g49)

and

∂

∂ϕ

θ

θ ∂

∂θ

θ

∂

∂θ

σ

τ

τ

WE

WE

R

w

WE

+ −

¸

¸

_

,

+ · 2 0 cos

sin

sin . (g50)

For the particular case when

WE

σ

is a constant and

NS WE

τ τ

· · 0, eqs.(g48)-(g50) show that the presence of the

lateral compressive stress is equivalent to a supplemental load placed on the upper face of the shell, having the value

p

H

R

NS

w

WE

w

( , )

sin

θ ϕ

∂

∂θ θ

∂

∂ϕ

σ σ

· − +

¸

¸

_

,

2

2

2

1

2

2

2

, (g51)

a result similar to the case of the plane plate (Timoshenko and Woinowsky-Krieger 1959; Nowacki 1961). Quantities of the

second order, like ( ) ∂ ∂θ w /

2

have been neglected again.

G.8) Load on the upper face. Stress on the lower surface of the shell.

Consider now the differential equation (g41). Usually, the horizontal loads for the upper face of the shell are neglected, and it

is assumed that

rr

U F

gw P

g

w

t

r

U

r

U

σ σ σ

ρ

∂

∂

θ ϕ

· − − −

¸

¸

_

,

· · 1

1

2

2

0 0 , , , (g52)

where ρ

F

is the density of the filling sediments, P is the load and an inertial term is considered. For the material below the

shell, the next constitutive equation is assumed

M

p

M

g R h w

M

tr

M M M M

M

t

σ ε ε

ε

ρ λ µ η

∂

∂

· − − −

¸

1

]

1

+ + +

0

2 2

1 1

( ) , (g53)

where p

0

is a reference pressure, ρ

M

is the density of that material and

1

is the unit tensor. The elastic coefficients are

λ µ

M M

, and the viscosity is denoted by η

M

. The strain tensor inside the material is

M

ε

and the strain rate here is

∂ ∂

ε

M

t / .

The first boundary condition assumed on the lower face of the shell, having the equation r R h · − , is the continuity of the

displacement vector. Hence the elements of the displacement vector inside the material placed immediately below the lower

face of the shell are assumed to be equal to the same elements at the points of the shell placed on the lower face. By using eqs.

(g8), it follows

u

r

r t w t

u r t

h

R

w

x t

u r t

h

R

w

y t

( , , , ) ( , , )

( , , , ) ( , , )

( , , , )

sin

( , , )

θ ϕ θ ϕ

θ

θ ϕ

∂

∂θ

θ ϕ

ϕ

θ ϕ

θ

∂

∂ϕ

θ ϕ

· −

·

−

+

·

−

+

¹

'

¹

¹

¹

¹

¹

¹

¹

1

. (g54)

Hence the next values for the strain tensor immediately below the shell are obtained

67

rr

M

r

M

R

w x

R h

r

M

R

w y

R h

ε ε ε

θ

∂

∂θ

ϕ

θ

∂

∂ϕ

· · − −

−

· − −

−

0

1

2 2

1

2

1

2

,

( )

,

sin ( )

. (g55)

The normal vector at the lower face of the shell is

r

e

→

. From Newton’s third low, it follows a relation between the stress

S

σ

inside the shell and the stress

M

σ

inside the material below the shell, i.e.

S

r

M

r

e e

σ σ

−

→

¸

¸

_

,

· −

→

¸

¸

_

,

, (g56)

on the lower surface of the shell. Here, the next elements of the stress are obtained after some elementary computations

rr

L

p

M

g R h w

Rh

R h

M

w

σ

ρ λ · − − − −

−

∗

0

( ) ∆ , (g57)

r

L M

R

w x

R h

M

R

w

t R h

x

t

θ

µ

∂

∂θ

η

∂

∂θ

∂

∂

∂

∂

σ

· − +

−

¸

¸

_

,

−

¸

¸

_

,

+

−

¸

1

]

1

1 1 1

, (g58)

and

r

L M

R

w y

R h

M

R

w

t R h

y

t

ϕ

µ

θ

∂

∂ϕ

η

θ

∂

∂ϕ

∂

∂

∂

∂

σ

· − +

−

¸

¸

_

,

−

¸

¸

_

,

+

−

¸

1

]

1

1 1 1

sin sin

. (g59)

G.9) The differential equation of time dependent flexure.

The reference pressure p

0

in (g53) is selected in order the flexure of the shell to vanish in the absence of the load.

Substituting the loads on the upper/lower faces of the shell in eq.(s41) and taking into account the presence of the lateral

gstress, a generalisation of the Sophie GERMAIN plain plate static equation in the case of a time dependent flexure of a

spherical elastic shell is obtained as

H

w

w

R

h

R

F

gw P

g

w

t

h

R

M

gw

Rh

R h

M

w

H

R

h

R

M

R w

w

R h

H

R

h

R

M

R

w

t R h

w

t

3

12

2 2

2

1

2

1

1

2

2

1

2

2

1

2

2

2

1

2

2

( ) λ µ µ ρ

∂

∂

ρ λ µ

η

∂

∂

∂

∂

+

∗ ∗

+

∗

¸

1

]

1

1

− +

¸

¸

_

,

+ −

¸

¸

_

,

¸

1

]

1

1

+ −

¸

¸

_

,

+

−

∗

¸

¸

_

,

− −

¸

¸

_

,

∗

+

−

¸

¸

_

,

− −

¸

¸

_

,

∗

¸

¸

_

,

+

−

¸

1

]

∆ ∆

∆

∆ ∆

∆

1

+ +

¸

¸

_

,

·

∗

¸

¸

_

,

− −

¸

¸

_

,

¸

1

]

1

1

¹

'

¹

¹

¹

¹

;

¹

¹

¹

H

R

NS

w

WE

w

H

H w

t

h

R

w

t

2

2

2

1

2

2

2

0

2

12

2

2

1

2 2

2

σ σ

∂

∂θ θ

∂

∂ϕ

ρ

∂

∂

∂

∂

sin

∆

. (g60)

If quantities of the order ( / ) h R

2

are neglected with respect to unit, it follows finally that

D w

H

M M

w

M F

gw

H

NS

R

w

WE

R

w

P

H

M

t

w

w

R

H

P

g

w

t

H

t

w

∆ ∆ ∆

∆ ∆

∗ ∗

+ −

¸

¸

_

,

∗

+ − −

∗

¸

¸

_

,

+ +

∗

¸

¸

_

,

+ +

∗

¸

¸

_

,

¸

1

]

1

1

·

∗

+

∗

+

¸

¸

_

,

− +

∗

¸

¸

_

,

+

∗

2

1

2

2

2

1

2 2

2

2

2

2

2

2

2

12

3

2

2

λ µ ρ ρ ρ

∂

∂θ θ

∂

∂ϕ

η

∂

∂

ρ

∂

∂

ρ ∂

∂

σ σ σ σ

sin

( )

, (g61)

where

68

D

H

a

H

·

+

·

( ) λ µ

ρ

2

3

12

2

3

12

(g62)

is the flexural rigidity and a is the velocity of the P-wave through the shell. Also,

ρ ρ ρ

µ µ

σ

∗

· + +

¸

¸

_

,

∗

·

∗

· +

¸

¸

_

,

M F

M

gR

H

R

H

R

P P

H

R

, ,

6

2

2

1 . (g63)

G.10) Spherical effects with respect to the plane plate.

For the usual materials, λ µ

M M

· . In that case, the corresponding equation for the plane plate (Ivan 1997) is.

D w H

x

c

w

x

y

c

w

y

M F

gw

P

H

M

t

w H

P

g

w

t

H

t

w

∆ ∆

∆ ∆

∗ ∗

+ +

¸

¸

_

,

+ −

¸

¸

_

,

· +

∗

¸

¸

_

,

− +

¸

¸

_

,

+

∗

σ σ

∂

∂

∂

∂

ρ ρ

η

∂

∂

ρ

∂

∂

ρ ∂

∂

2

2

2

2

2

2

2

12

3

2

2

( )

. (g64)

With respect to (g64), a change of the density difference ρ ρ

M F

− according to (g63a) and a substitution of the real lateral

stresses

NS

σ

,

WE

σ

by their apparent values

NS WE

σ σ σ σ

+

∗

+

∗

, can be observed in eq.(s61). A supplemental

load is present according to (g63c). For usual values (e.g. Ivan 1997a) like H R / / ∝1 100,

µ µ ,

M

Pa ∝10

11

, ρ ρ

M F

kgs m , / ∝3000

3

,

NS WE

MPa

σ σ

, ∝30 , all these effects are usually negligible and

difficult to be observed in real life. To compare 2

2

w R / to ∆

∗

w in the left side of (g61), the case of a rectangular plate

having the sides equal to L

x

L

y

, is considered. Here, the flexure is proportional to a product of sins (cosines) functions, i.e.

w m x L

x

n y L

y

∝sin( / ) sin( / ) π π . It follows the LAPLACE operator of the flexure is proportional

to ∆

∗

∝ +

¸

¸

_

,

w m L

x

n L

y

m x L

x

n y L

y

π π π

2 2 2 2 2

/ / sin( / )sin( / ) . For the fundamental mode ( m n · ·1) it

follows that

2

2

1

2 2

1

2

1

2

w R

w R L

x

L

y

/

( / / ) ∆

∗

·

+ π

. (g65)

That ratio is negligible too in the usual cases.

It can be concluded that in the usual cases, the sphericity of the crustal plates can be ignored and the equation derived for the

flat plate can be used.

69

H) ELEMENTS OF RHEOLOGY

H.1) Introduction.

Especially for geological processes at a large time scale and great values of the stress, the internal friction of the material

cannot be ignored. Consequently, the HOOKE’s law has to be replaced by assuming different constitutive equations (models).

Their expressions are mainly depending on the time scale of the geophysical process to be modelled. In relation to seismic or

seismological applications, for example, short periods and short stresses are required (usually, seconds up to minutes, with a

maximum value around one hour for the fundamental mode of the free oscillations of the Earth). Here, the non-elasticity is

related to the very short period irreversible changes in the crystal defect structures of the medium (e.g. opening/closing of pre-

existing cracks) and/or to the energy lost by friction at the two sides of a crack or on the non-elastic boundary coupling grain

particles to the adjacent material (Aki and Richards, 1980; Ranalli, 1987; Wahr, 1996).

With respect to the mathematical relation between stress and strain, there are two kinds of constitutive equations (models).

H.2) Linear models.

Simplified models involves a linear relation between stress (and its derivatives of various orders with respect to time) and

strain (together with its time derivatives).

In the beginning, only the 1-D case will be discussed. More general examples follow to be presented in relation with the

dynamic aspects of the flexure of a plate (shell) and to the accretion prism. For each constitutive equation, a mechanical

analogue can be considered. The elastic part will be represented by a spring, while the inelastic (viscous) behaviour is

associated to a dashpot. Both parts are supposed to be linear ones, i.e. a linear relation

σ ε

µ · 2 is valid for the spring and

a similar linear relation holds for the dashpot

σ ε

η ·

•

2 .

Fig.H1. (a) KELVIN-VOIGT model; (b) MAXWELL model; (c) BURGERS model

a) KELVIN-VOIGT (strong viscous) model.

The mechanical analogue of the first model to be considered is represented in Fig.H1a. The total stress is the sum between the

stress in the spring and the stress in the dashpot, while the total deformation is equal to both the deformation of the spring and

the deformation of the dashpot. It follows that KELVIN-VOIGT model has the next constitutive equation

σ ε ε

µ η · +

•

2 2

K K

, (h1)

where the dot shows the (total, material) derivative with respect to time. Here, the second term is the inelastic one, η

K

being

the viscosity. Suppose now a constant stress equal to

0

σ

is applied. Elementary computations show that the differential

equation

70

0

2 2

σ ε

ε

µ η · +

K K

d

dt

(h2)

with the initial condition

( )

ε

t t · ·

0

0 (h3)

has the solution

ε

σ

µ

µ

η

· − − −

¸

¸

_

,

¸

1

]

1

0

2

1

0

K

K

K

t t exp ( ) , (h4)

for t t ≥

0

. Hence, for very great values of time, the strain approaches a limiting value equal to

∞

·

ε

σ

µ

0

2

K

. (h5)

The “flowage function”, denoted by

J t

K

K

K

t t ( ) exp ( ) · − − −

¸

¸

_

,

¸

1

]

1

1

2

1

0

µ

µ

η

, (h6)

shows that for a constant stress (equal to unit here), there is a temporal variation of the strain.

Suppose now at a certain moment t t ·

1

, the constant stress

0

σ

is removed, the corresponding strain at that moment being

equal to

1

ε

. It follows now the corresponding solution decreases towards zero as

ε ε

µ

η

· − −

¸

¸

_

,

1

1

exp ( )

K

K

t t . (h7)

b) MAXWELL (viscous-elastic) model.

Consider the mechanical analogue represented in Fig.H1b. The total stress is equal to both the stress of the spring and to the

stress of the dashpot, while the total deformation is the sum between the deformation of the spring and the deformation of the

dashpot. It follows that MAXWELL model has the next constitutive equation

•

·

•

+

ε

σ

σ

µ η 2 2

M M

, (h8)

with the initial condition represented by (h3).

Suppose again a constant stress equal to

0

σ

is applied. The spring is instantly deformed to a value equal corresponding to

the first term in the right hand of (h8), i.e.

0

0

2

ε

σ

µ

·

M

. (h9)

and the solution of (h8) (for a constant stress

0

σ

) with the initial condition (h9) is the straight line

( )

ε

σ σ

η µ

· − +

0

2

0

0

2

M

t t

M

, (h10)

having the slope related to the stress

0

σ

and to the viscosity of the dashpot. Suppose now at a certain moment t t ·

1

, the

constant stress

0

σ

is removed, the corresponding strain at that moment being equal to

1

ε

. It follows from (h8) that the

strain remains constant.

c) BURGERS (general linear) model.

The third model to be considered has the mechanical analogue represented in Fig.H1c.

EXERCISE. Show that the corresponding differential equation is

2

1

2

1

1

2

1

2

1

2

1

1

2

η µ

η

µ

η

η

µ

µ

µ

η

ε ε σ σ σ

••

+

•

·

••

+ + +

¸

¸

_

,

•

+ . (h11)

Consider now the same initial condition (h3) and suppose again a constant stress equal to

0

σ

is applied. In a similar manner

to MAXWELL model, the system is instantly deformed to a value equal to

71

0

0

2

2

ε

σ

µ

· (h12)

and the solution of (h11) (for a constant stress

0

σ

) with the initial condition (h12) is

ε

σ σ

µ η

µ

η

· + − + − − −

¸

¸

_

,

¸

1

]

1

0

2

2

0

2

2

0

1

1

1

0

( ) exp ( ) t t C t t , (h13)

where C is an unknown coefficient (because (h11) is a second order differential equation, two initial conditions are required to

obtain the complete solution). However, differentiating (h13) it follows

•

· + − −

¸

¸

_

,

ε

σ

η

µ

η

µ

η

0

2

2

1

1

1

1

0

C t t exp ( ) . (h14)

Hence, for great values of time, the solution (h13) approaches asymptotically to a straight line having the slope equal to

0

2

2 σ

η / . Suppose now at a certain moment t t ·

1

, the constant stress

0

σ

is removed, the corresponding strain at that

moment being equal to

1

ε

. It follows from (h13) that the strain decreases exponentially towards zero, i.e

ε ε

µ

η

· − −

¸

¸

_

,

1

1

1

1

exp ( ) t t . (h15)

EXERCISE. Show that two (or more) springs / dashpots connected in series (or parallel) sequence are equivalent to a single

spring / dashpot. Justify that the BURGERS model is the general linear model.

d) Remarks on the linear models.

In the most general case, the linear relation between stress and strain can be written as

P D Q D ( ) ( )

σ ε

· , (h16)

where

P D D D

n

D

n

A A A A

( ) ... · + + + +

0 1 2 2

(h17)

and

Q D D D

m

D

m

B B B B

( ) ... · + + + +

0 1 2 2

(h18)

are formal polynomials of the variable D

d

dt

· representing the derivative with respect to time, applied to stress and strain

respectively. Here,

0 1 0 1

A A A B B B

n m

, , ..., , , , ..., are fourth rank tensors. For example, with respect to the

MAXWELL body having the constitutive equation (h8), it follows that

P D

M M

D Q D D ( ) , ( ) · + ·

1

2

1

2 η µ

(h19)

A common way to solve (h16) is by using the LAPLACE transform (e.g. Sokolnikoff and Redheffer, 1958). Consider a certain

function of one real variable f t ( ) , providing that

1. f t for t ( ) , · < 0 0 ;

2. f t ( ) is piecewise continuous on every finite interval;

3. there are two constants 0 0 < ≤ M a , in order to have f t M at ( ) exp( ) ≤ , for an arbitrary t .

Under the above conditions, the LAPLACE transform of f t ( ) is a new function of the variable p , defined by

[ ]

L f p f t pt dt ( ) ( ) exp( ) · −

∞

∫

0

(h20)

EXERCISE. Show that:

(a) The LAPLACE transform of the first derivative is

[ ]

L

df

dt

p pL f f

¸

1

]

1

· − + ( ) ( ) 0 0 (h21)

(b)

[ ]

L at p

p a

a exp( ) ( ) , − ·

+

≤

1

0 (h22)

72

(c) The convolution theorem. Consider the functions f g , vanishing for negative values of their argument and a new function

(also vanishing for negative values of the argument defined) by the convolution product

( ) h t f g t f t f t

t

( ) ( ) ( )g( )d ( )g( )d · ∗ · −

−∞

+∞

· −

∫ ∫

τ τ τ τ τ τ

0

(h23)

Show that

[ ] [ ] [ ]

L f g L f L g ∗ · (h24)

As an example, consider again the MAXWELL model by applying the LAPLACE transform to both sides of (h8). It follows

that

pL

M

pL

M

L

ε ε σ σ σ

µ η

¸

1

]

1

− ·

¸

1

]

1

−

¸

¸

_

,

+

¸

1

]

1

0

1

2

0

1

2

, (h25)

where

0

0 0

0

0 0

σ σ ε ε

· + · + ( ) , ( ) are the initial stress and strain respectively.

EXERCISE. By using the above properties of the LAPLACE transform, show that

σ ε ε σ ε

µ

µ

η

µ

η

τ τ µ

µ

η

( ) ( ) ( ) exp ( ) exp t

M

t

M

M

t

M

M

t

t

d

M

M

M

t · − − −

¸

1

]

1

+ −

¸

¸

_

,

−

¸

¸

_

,

∫

2

2

2

0

0

2

0

(h26)

If the initial conditions are elastically coupled, i.e.

0

2

0

σ ε

µ ·

M

(h27)

it follows an integral representation of the stress which is independent on the initial conditions

σ ε ε

µ

µ

η

µ

η

τ τ ( ) ( ) ( ) exp ( ) t

M

t

M

M

t

M

M

t

t

d · − − −

¸

1

]

1

∫

2

2

2

0

(h28)

Hence, the actual value of the stress is related both to the actual value of the deformation and to the previous values of the

strain (i.e. the stress depends on the “history” of the deformation). If a constant strain

ε ε

( ) , t for t · ≥

0

0 is applied to

the MAXWELL body, it follows from (h28) that the stress decreases as

σ ε

µ

µ

η

( ) exp t

M

M

M

t · −

¸

¸

_

,

2

0

, (h29)

representing a “relaxation phenomenon” (stress decreases in time if a constant deformation is present). Here, the function

( ) G(t

M

t

M

) exp / · − 2µ τ (h30)

is the “relaxation” kernel and the parameter τ η µ

M M M

· / is the relaxation time.

A very similar approach (e.g. Wahr 1996) is based on the use of FOURIER transform (see Chapter F.13). Formally, the

results derived by using FOURIER transform are derived from the same results obtained with LAPLACE transform by

performing the substitution p iu · .

The above 1-D models can be generalised for the 3-D case. For example, consider again the MAXWELL body. There is a

strong experimental evidence that the MAXWELL Rheology applies only to the dissipation of the shear energy, i.e. the stress

and strain tensors in (h8) are the deviatoric tensors

σ σ σ ε ε ε

↔ − ↔ −

1

3

1

3

1 1

tr tr , (h31)

It will be further assumed that there is no dissipation of the compressional energy, i.e. a proportionality like

tr

M M

tr

σ ε

λ µ · + ( ) 3 2 , (h32)

is valid. By differentiating with respect to time

tr

M M

tr

•

· +

•

σ ε

λ µ ( ) 3 2 (h33)

Substituting (h31) and (h33) in (h8) it follows

•

−

•

·

•

−

+

• ¸

¸

_

,

+ −

+

¸

¸

_

,

ε ε σ ε σ ε

µ

λ µ

η

λ µ 1

3

1

2

3 2

3

1

2

3 2

3

1 1 1

tr

M

M M

tr

M

M M

tr (h34)

Hence

73

•

+ ·

•

+

•

+

+

σ σ ε ε ε

µ

η

µ λ

µ

η

λ µ

M

M

M M

tr

M

M

M M

tr 2

3 2

3

1 1

(h35)

By applying the FOURIER transform to both sides of (h35), it follows

Φ Φ Φ

σ ε ε

λ µ

¸

1

]

1

1

·

¸

1

]

1

1

+

¸

1

]

1

1

~ ~

tr

1

2 , (h36)

i.e. a relation similar to HOOKE’s law is valid between the FOURIER (or LAPLACE) transforms of stress and strain. Here, the

coefficients similar to LAMÉ parameters are

λ λ

µ

η

µ

λ

µ

η

µ µ

µ

η

~

,

~

·

+ +

¸

¸

_

,

+

·

+

M

iu

M

M

M

M

iu

M

M

M

iu

iu

M

M

1

2

3

(h37)

At short periods T correspond high values of the pulsation u T · 2π / . From (h37) it follows

λ λ µ µ

~

,

~

· ·

M M

(h38)

and the behaviour of the MAXWELL body is an elastic one.

At long periods T correspond low values of the pulsation and

λ

λ µ

µ

~

,

~

·

+

· ·

3 2

3

0

M M

K

M

(h39)

and the MAXWELL body is a fluid having the compressional coefficient denoted by K

M

.

H.3) Non-linear models.

For tectonic applications, large stresses and periods of thousands to millions of years are appropriate. Here, the non-elastic

behaviour is probably related to the diffusion or dislocation creep of the molecules, a major factor being the high temperatures.

There are great difficulties to consider constitutive equations with non-linear relations between stress and strain, but some

attempts have been made. A very common non-linear model is the work-hardening plasticity (e.g. Ranalli, 1994)

ij

E

n

ij

ε

σ

σ

σ

σ

·

¸

¸

_

,

−

3

2

0

1

0

*

*

, (h40)

where

ij

*

σ

is the component of the deviatoric stress,

E ij ij

* * *

σ σ σ

·

3

2

is related to the second invariant of the deviatoric

stress, and

0

σ

, n are material parameters.

H.4) Brittle. Creep. Empirical criteria.

The usual materials are reacting in an elastic manner only for small values of the (deviatoric) stress, i.e. for stress values

smaller than a limiting value representing the yield strength (or the yield stress), denoted by

Y

σ

. The yield stress is a

function of the nature of the material, of the temperature, pressure, the chemical composition of the adjacent rocks and, finally,

of the history of the deformation (i.e. the intermediate steps followed to attend the yield value). When the yield stress is

attended, there are two possibilities of behaviour of the material:

• a rupture deformation of the rock, when the continuity of the deformation is lost, usually along a fault surface; this is the

case of the brittle materials. The process is illustrated in Fig.H2a and b.

• a plastic, irreversibly flow of the material (creep), when, apparently, the continuity holds. The phenomenon is quite

similar to the usual viscous flow, but it can be observed only when the yield stress is attended. This is the case of the

ductile materials. The process is illustrated in Fig.H2c. An usual non-linear constitutive equation is the BYERLEE

power-law creep

•

· −

ε σ

A

n

H RT exp( / ) , (h41)

74

where A n , are material parameters, His the activation enthalpy, R is the gas constant and T is the absolute

temperature. It should be noted that the same material can act as a brittle or a ductile one according to the external

conditions.

Fig.H2. (a) Faulting of a brittle material; (b) Increasing of the yield stress due to the history of deformation;

(c) Creep of a ductile material.

H.5) Empirical criteria for shear-faulting. TRESCA criterion. COULOMB-NAVIER criterion.

A first criterion also (valid for plasticity) is due to TRESCA. It assumes that faulting (for brittle materials) or creep (for

ductile ones) is attended at that points of the material where the maximum value of the shear (tangential) stress is equal to a

yield value denoted by

Y

σ

. Consider now a homogeneous (constant) stress field inside the material. Such a case can be

obtained either by considering an infinitesimal volume of material or by taking into account a prismatic body with very large

(infinite) sides. Consider the eigen-values

1 2 3

σ σ σ

, , of the stress tensor. It will be assumed that they are denoted in order

to have

3 2 1

σ σ σ

< < , with the remark that ,in real life, stress is assumed to be positive for compression. Hence, with

respect to Fig.H3, let

11 1 22 3

σ σ σ σ

· − · − , and

12

0

σ

· in eqs. (d11) and (d13). It follows that

σ σ

σ σ σ σ

τ σ

σ σ

ψ

θ

ψ

· ·

+

−

−

· − ·

−

−

rr

r

1 3

2

1 3

2

2

1 3

2

2

cos

sin

, (h42)

where

σ τ

, are so-called “normal stress” and “tangential (shear) stress” respectively, acting on a plane inside the material.

The plane is at an angle ψ θ

π

· −

2

with the direction of the maximum compressive stress. The outer-pointing normal at that

plane makes an angle θ with the direction of the maximum compressive stress. With respect to a

σ τ

− reference system,

eqs.(h42) are the parametric equations of the MOHR circle (see Section D4), plotted in Fig. H3. In the most general case, stress

field is varying from point to point inside the material. Hence both the eigen-vectors of the stress tensor (i.e. the local directions

of the maximum / minimum compressive stress) and the eigen-values of that tensor (i.e. the magnitudes of the maximum /

minimum compressive stress) are also changing from point to point. For a fixed point inside the material, both normal stress

and shear (tangential) stress are varying with the angle between the plane (with respect to normal and tangential stresses are

defined) and the direction of the local maximum compressive stress. Consider a certain point inside the material and imagine

various planes passing through that point. Hence, according to TRESCA empirical criterion, failure of the material is produced

here if

ψ

τ σ max

·

Y

(h43)

Using eq.(h42b), it follows

ψ

ψ

ψ

ψ

σ σ σ σ

σ max max

sin sin

1 3

2

2

1 3

2

2

−

·

−

·

Y

(h44)

75

Consider now a homogeneous stressed material subject to progressive increasing values of the difference

1 3

σ σ

− . The

material is characterised by a yield value denoted by

Y

σ

. Eq.(h44) shows that:

• if

( )

1 3

2

σ σ σ

− < /

Y

, there is no failure inside the material;

• when the equality

1 3

2

σ σ σ

− ·

Y

(h45)

is attended, a failure is produced along the planes at angles ψ

π

· t

4

with the direction of the maximum compressive stress.

Fig.H3.(a) The MOHR circle;

(b) A plane inside the material, at an angle ψ with the direction of the maximum compressive stress;

(c) TRESCA criterion. E.d and F.D. denote the elastic domain and the failure domain respectively;

(d) COULOMB - NAVIER criterion.

Eq.(h45) represents the TRESCA criterion in terms of the eigenvalues of the stress tensor. In the case of a material subject to a

non-homogeneous stress field, eq.(h45) is a local condition. Here, the eigen-values are obtained (see Section A). For a

compressive stress, that values are expected to be negative ones. They have to be denoted by − − −

1 2 3

σ σ σ

, , , where

3 2 1

σ σ σ

< < .

A second criterion is due to COULOMB and NAVIER. It can be used to describe only the shear fracture. According to it, a

material is characterised by the cohesive strength denoted by S and by the coefficient of friction, denoted by µ φ

*

tan · .

Here, φ is the angle of internal friction ( φ · 30

o

in most rocks). According to COULOMB - NAVIER empirical criterion,

shear failure of the material is produced at its points where

ψ

µ

τ σ max

*

−

¸

¸

_

,

· S (h46)

Using eqs.(h42), it follows

76

ψ

ψ φ ψ

σ σ σ σ σ σ

max

sin tan cos

1 3

2

2

1 3

2

1 3

2

2

−

−

+

−

− ¸

¸

_

,

¸

1

]

1

1

· S (h47)

If ψ is a solution of (h47), π ψ − (or just −ψ ) is a solution too. Hence, without loss of generality, the values of the angle

ψ will be limited to the first quadrant, where eq.(h47) is

( )

[ ]

1 3

2

1

2

1 3

2

σ σ σ σ

φ

ψ

ψ φ φ

−

+ · +

+

cos

sin tan

max

S (h48)

Eq.(h48) shows that:

• if

1 3

2

1

1 3

2

σ σ σ σ

φ

φ

−

−

+

<

cos

tan S, there is no failure of the material;

• when the equality

1 3

2

1

1 3

2

σ σ σ σ

φ

φ

−

−

+

·

cos

tan S (h49)

is attended, a share fracture is produced along the planes at angles ψ

π φ

· t −

¸

¸

_

,

4 2

with the direction of the maximum

compressive stress. Eq.(h49) represents the COULOMB - NAVIER criterion in terms of the eigenvalues of the stress tensor.

Taking into account that µ φ

*

tan · , eq.(h49) can be written as

1

2

1

3

2

1 2

σ σ

µ µ µ µ

* * * *

¸

¸

_

,

+ −

¸

1

]

1

1

−

¸

¸

_

,

+ +

¸

1

]

1

1

· S , (h50)

outlining that COULOMB - NAVIER criterion is a generalisation of TRESCA criterion for a non-zero internal friction.

H.6) Von MISES-HENCKY criterion for ductile flow (plasticity).

Because the ductile (plastic) flow is independent of the co-ordinate system used, it depends only on the invariants of the stress

tensor (Section A, eq.(a22)). Hence an equation like

( ) f I I I

1 2 3

0 , , · , (h51)

will be valid. There are strong experimental evidence that the plastic flow does not depend on the hydrostatic pressure, being

also similar for compressive and tensile states of stress. It follows the function f in (h51) depends only on the second invariant

of the stress deviator (see eq.(a23)). Hence, ductile flow occurs only at those points of material where the second invariant of

the deviator stress reaches a certain value, depending on the nature of the material. Using eq.(a24), the criterion of Von MISES

- HENCKY assumes that ductile flow occurs at those points of the material where

( ) ( ) ( )

11 22

2

22 33

2

33 11

2

6

12

2

23

2

13

2

6

2

σ σ σ σ σ σ σ σ σ

− + − + − + + +

¸

¸

_

,

· k (h52)

In terms of the principal stresses, eq.(h52) is

( ) ( ) ( )

1 2

2

2 3

2

3 1

2

6

2

σ σ σ σ σ σ

− + − + − · k (h53)

Hence the criterion of Von MISES - HENCKY can be regarded too as a generalisation of TRESCA criterion, by taking into

account the presence of the intermediate stress.

77

H.7) Rheological models.

a) SAINT-VENANT body (elastic-plastic material).

The behaviour of that material is characterised by linear elasticity for stress values below the yield strength. When the yield

stress is attended, the body exhibits a pure plasticity. Its constitutive equation (using deviator tensors) has the symbolic form

* *

,

(

*

,

*

) ,

σ ε σ σ

σ σ σ

µ · <

· ·

¹

'

¹

¹

¹

¹

¹

2

2 3

S

Y

f I I

Y Y

(h54)

The above material has the mechanical analogue presented in Fig. (h4a), being referred as a SAINT-VENANT body.

b) BINGHAM body (visco-plastic material).

Similar to the SAINT-VENANT body, that material exhibits linear elasticity for stress values lower than the yield strength.,

but flows linearly above that value. The strain rate is proportional to the difference between the deviatoric stress and the yield

strength. Its constitutive equation is

* *

,

*

*

,

σ ε σ σ

σ σ

ε

σ σ

µ

η

· <

· + ≥

¹

'

¹

¹

¹

¹

¹

2

2

B

Y

Y

B

d

dt

Y

(h55)

The above material has the mechanical analogue presented in Fig. (h4b), being referred as a BINGHAM body.

Fig.H4. (a) The SAINT-VENANT body; (b) The BINGHAM body.

78

I) THE ACCRETION WEDGE.

I1) The model.

Consider a 2-D prismatic body having a triangular vertical section (Fig.I1), in the presence of gravity. The wedge rests on a

rigid basement having the slope equal to θ

0

. Both the compressional force acting on the left side of the wedge and the friction

to the basement cause thickening of the incompressible material and the development of a topographical slope equal to α. It is

assumed that the material is into a state of plastic yielding according to the VON MISES-HENCKY criterion. It follows to

obtain a condition relating the slopes of the topography and that of the basement to the geometry of the wedge, its yield strength

and the friction coefficient to the basement.

Fig.I1. The 2-D accretion wedge.

I2) Equations of equilibrium. Yield condition. Stress field.

Taking into account that the 2-D case is discussed, the stress in polar co-ordinates is

( )

σ

σ σ

σ σ

σ σ

θ

θ θθ

ν

θθ

·

+

¸

¸

_

,

rr r

r

rr

0

0

0 0

. (i1)

Because the material is assumed to be incompressible, the POISSON coefficient is ν · 1 2 / .

Using polar co-ordinates, the equilibrium equation (d40)-(d41) in the presence of gravity are

∂

∂

∂

θ

∂θ

θθ

ρ θ

σ σ σ σ

rr

r

r

r

rr

r

g + +

−

+ · sin 0 (i2)

∂

θ

∂

∂

θθ

∂θ

θ

ρ θ

σ σ σ

r

r r

r

r

g + + + · 2 0 cos (i3)

Using (i1), the yield condition (r54) is

1 3

2

σ σ

− · k (i4)

Taking into account again that the compressive stress is assumed to have positive sign, eqs.(d11)-(d13) give

rr

f r k

f r k

r

k

σ

σ

σ

θ ψ

θθ

θ ψ

θ

ψ

· − −

· − +

·

¹

'

¹

¹

¹

¹

¹

( , ) cos

( , ) cos

sin

2

2

2

, (i5)

where the trace of the stress has been denoted by f r

tr

( , ) ( ) θ

σ

·

1

2

and ψ ψ θ · ( , ) r is the angle between the radius and

the local direction of the maximum compressive stress (variable inside the wedge). However, it will be assumed that

ψ ψ θ · ( ) only. After some manipulations, substituting (i5) into (i2)-(i3) gives

∂

∂

ψ

ψ

θ

ψ ρ θ

∂

∂θ

ψ

ψ

θ

ψ ρ θ

f

r

k

r

d

d

k

r

g

f

k

d

d

k gr

· − +

· − + +

¹

'

¹

¹

¹

¹

¹

2

2

2

2

2 2 2 2

cos cos sin

sin sin cos

(i6)

79

Eqs.(i6a) and (i6b) are differentiated with respect to θ , r respectively, the results of the differentiation being equal each other.

After some elementary manipulations, it follows that

d

d

d

d θ

ψ

ψ

θ

ψ cos cos 2 2 0 −

¸

¸

_

,

· , (i7)

i.e.

d

d

C ψ

θ ψ

· + 1

2 cos

, (i8)

where C is a constant of integration. Substituting (i8) into (i6a) gives

∂

∂

ρ θ

f

r

Ck

r

g · +

2

sin , (i9)

i.e.

f Ck r gr g · + + 2 ln sin ( ) ρ θ θ , (i10)

To find the unknown function g g · ( ) θ , eq.(i10) is substituted into (i6b) to obtain

dg

d

Ck

θ

ψ · −2 2 tan . (i11)

Using (i8), it follows that

dg

d

Ck

C ψ

ψ

ψ

· −

+

2

2

2

sin

cos

, (i12)

i.e.

g Ck C A · + + ln( cos ) 2ψ , (i13)

where A is another constant of integration. Hence the final stress inside the wedge is

rr

Ck r Ck C co gr A k

Ck r Ck C co gr A k

r

k

σ

σ

σ

ψ ρ θ ψ

θθ

ψ ρ θ ψ

θ

ψ

· − − + − − −

· − − + − − +

·

¹

'

¹

¹

¹

¹

¹

2 2 2

2 2 2

2

ln ln( ) sin cos

ln ln( ) sin cos

sin

. (i14)

I3) Boundary conditions. Final results.

Consider the segment AC placed on the side OA of the wedge, having θ · 0 and OC r OA ≤ ≤ , where the point C is very

closed to the point A. The outward pointing normal vector is n e

→

· −

→

θ

. Here, is acting the lithostatic pressure due to the

topography. Hence

σ θ

ρ α

θ

−

→

¸

¸

_

,

·

→

e gr e tan , (i15)

i.e.

( )

( )

r

gr

θ

θ

θθ

θ ρ α

σ

σ

· ·

· · −

¹

'

¹

¹

¹

0 0

0 tan

. (i16)

Because the angle θ

0

has very small values, it will be assumed for all angles θ that

θθ

ρ α

σ

· − gr tan , (i17)

or

∂

θθ

∂

ρ α

σ

r

g · − tan . (i18)

However, all the next derivations are supposed to be valid at the rear of the wedge, where the topography is generated due to

the horizontal compression, i.e. the radius r is a mean value of the lengths OA and OC, the point C being closed to A. By

differentiating (i14b), eq.(i18) leads to

2Ck

r

g · ρ α tan , (i19)

where

80

r

h

≅

0

0

θ

. (i20)

Consider now the side OB of the wedge, having θ θ ·

0

and the outward pointing normal vector n e

→

·

→

θ

. Here, is acting

the friction force due to the basement, assumed to have the magnitude equal to λk , where λ is a friction coefficient. Hence

σ θ

λ e k e

r

→

·

→

, (i21)

i.e.

( )

r

k

θ

θ θ λ

σ

· ·

0

. (i22)

The next partial derivative follows to be evaluated in two ways. In the first approach, eqs.(i22), (i16a) and (i20) are used to

give

( ) ( ) ∂

θ

∂θ

θ

θ θ

θ

θ

θ

λ σ σ σ

r

r r

r r

k

h

≈

· − ·

−

·

1 0

0

0

0

0

. (i23)

The dominant stress is into the wedge is the horizontal compression. Hence ψ θ ≅ , both angles having small values. It follows

cos 2 1 ψ ≅ . Eqs.(i5c), (i8), (i19) and (i20) give

∂

θ

∂θ

ψ

ψ

θ

ψ ρ α

θ σ

r

r

k

r

d

d

k

r

C

Ck

r

k

r

g

k

h

· · + ≈ + · +

2

2

2

2

2 2 2

0

0

cos ( cos ) tan (i24)

From (i23) and (i24) it follows

ρ α θ λ gh k k

0

2

0

tan + ≅ , (i25)

showing that the friction force (resistance to sliding of the wedge onto the basement) is balanced by tqo forces. The first one is

due to the topography and the second force is related to the compressive stress and to the slope of the basement. Further details

related to the application of eq.(i25) in real cases are presented by Ranalli (1987).

80

General references

Aki, K., Richards, P.G., 1980. Quantitative Seismology, Freeman and Co., New York

Backus, G., Continuum Mechanics, Samizdat Press (http://samizdat.mines.edu)

Bayly, B., 1992, Mechanics in Structural Geology, Springer - Verlag, N.Y.

Filonenco-Borodici, M.M., 1952, Theory of Elasticity, Ed.tehnicã, Bucharest

Ivan, M., 1996, Seismology (in Romanian), v.I, Ed.Universitãtii Bucuresti

Ivan, M., 1997a. A finite element algorithm for computing the flexure of the crustal, in press, Studii si

cercetari de GEOFIZICA 35

Ivan, M., 1997b. Aperiodic motion of a crustal microplate, in press, Revue Roumaine de Géophysique 41

Jaeger, J.C., Cook, N.G.W., 1969, Fundamental of Rock Mechanics, Methuen and Co. London

Kennett, B., Continuum Mechanics, Samizdat Press (http://samizdat.mines.edu)

Mandl, G., 1988, Mechanics of tectonic faulting, Elsevier, Amsterdam

Meissner, R., 1986, The Continental Crust - A Geophysical Approach, Academic Press, Inc, N.Y.

Nowacki, W., 1961. Dynamika budowli (in Polish), Arkadi, Warszawa

Ranalli, G., 1987, Rheology of the Earth, Allen & Uniwin Inc., Boston

Rîjic, I.M., Gradstein, I.S., 1955, Tables of integrals, sums, series and products, -in Romanian, Ed.tehnicã,

Bucharest

Sokolnikoff, I.S., Redheffer, R.M., 1958, Mathematics of Physics and Modern Engineering, McGraw-Hill

Book Co., Inc., New York

Timoshenko, S., Woinowsky-Krieger, S. 1959. Theory of plates and shells, Mc-Graw-Hill, New York

Timoshenko, S.P., Gere, J.M., 1961, Theory of elastic stability, McGraw-Hill, New York

Turcotte, D.L., Schubert, G. 1982. Geodynamics, John Wiley and Sons, New York

Wahr, J., 1996, Gravity and Geodesy - Class Notes, Samizdat Press

2

In memoriam Professor Marin DOROBANÞU

INTRODUCTION

These course notes present some theoretical problems related to the Mechanics of a Continuum Solid Body, of particular importance to Applied Geomechanics, Geological Engineering and Structural Geology. In most cases, only static aspects are discussed, but some dynamic cases are also presented. As a rule, the modern tensorial approach is used. Some basic elements are reviewed at the beginning of the notes. More details can be found into the excellent books of George Backus and Brian Kennett, also available at Samizdat Press. The linear elasticity and the homogeneity of the continuum solid body are almost thoroughly assumed to be valid, but some elements of Rheology are also presented. In most cases, the semi-inverse method is used to solve the problems. According to it, the solution is supposed to be of a particular form, as a consequence of the simplified hypothesis previously assumed. It is verified that solution checks both the corresponding equations and the boundary conditions. Based on the Uniqueness Theorem of the Linear Elasticity, it follows the assumed particular solution is just the general solution of the problem. In all the cases discussed here, the assumed simplified hypotheses allow one to obtain simple, analytical solutions. At a first glance, the importance of such solutions is minor with respect to the real cases, where mainly the non-homogeneity of the medium plays a great role. However, the analytical solutions are the basis for deriving finite element algorithms, allowing one to model satisfactory the complex real cases. Such examples are also presented. The author will kindly appreciate any critical remarks on these notes, being very grateful to Samizdat Press for the possibility to shear his work.

3

CONTENTS

INTRODUCTION page Chapter A: BASIC ELEMENTS. ………………………………………………………………………………………….... 5 A.1) The displacement vector. Lagrangean (material) and Eulerian (spatial) co-ordinates. …………………………………..5 A.2) Invariants of a tensor. Tensor deviator. ………………………………………………………………………….………..5 A.3) Strain tensor. Stress tensor. Equation of motion / equilibrium. …………………………………………………………..5 A.4) HOOKE’s law. ………………………………………………………………………………………………………...…..8

Chapter B: DEFORMATION OF A CYLINDRICAL BODY IN THE PRESENCE OF GRAVITY ……………....…9 B.1) The model… ……………………………………………………………………………………………………………….9 B.2) Equations of equilibrium. Boundary conditions. Simplifying hypothesis. ……………………..……………………..…..9 B.3) The final shape of the body. ………………………………………………………………………………………….…..12

Chapter C: LÉVY’s PROBLEM - the triangular dam……………………………………………………………………. 14 C.1) The SAINT-VENANT’s equations. …………………………………………………………………………………..…..14 C.2) The model. Simplifying hypothesis. The planar deformation state. ……………………………………………………...14 C.3) Equations of equilibrium. AIRY’s potential. ……………………………………………………………………………..15 C.4) Boundary conditions. The final shape of the dam..…………………………………………………………………..…..16

Chapter D: KIRSCH’s PROBLEM - the circular bore hole / tunnel………………………………………………..……19 D.1) The model. ……... ……………………………………………………………………………………………………..…19 D.2) The planar state of deformation in cylindrical co-ordinate system. ………………………………………………….…..20 D.3) The circle of MOHR. ………………………………………………………………………………..……………………21 D.4) AIRY’s potential in cylindrical co-ordinates. The bi-harmonic equation.…………………………………………….… 21 D.5) The divergence of a tensor in cylindrical co-ordinates. …………………………………………………………………..22 D.6) The gradient of a vector and the strain tensor in cylindrical co-ordinates. …………………………………………..….23 D.7) The bi-harmonic equation cylindrical co-ordinates. ……………………………………………………………….…….24 D.8) The stress elements. Conditions at infinity for the stress elements. ….………………………………………..…………25 D.9) Strain and displacement vector. Conditions at infinity. ………………………………………………………….………27 D.10) Boundary conditions for the stress elements on the wall of the circular cavity. ………………………………….…….28 D.11) The final shape of the wall. ………………… ………………………………………………………………….………31

Chapter E: BOUSSINESQ’s PROBLEM - concentrated load acting on an elastic semi-space………………………....33 E.1) The equations of BELTRAMI and MITCHELL. …………………………………………………………………………33 E.2) The model. …………………………………………………………………………………………………..………….…35 E.3) The equations of equilibrium and strain tensor in spherical co-ordinates. …………….…………………………………35 E.4) LAPLACE operator in spherical co-ordinates. LEGENDRE’s polynomials. …………………………………….………36 E.5) The displacement field. ………………………………………………………………………………….…..……………37 E.6) Boundary conditions for the stress elements. The final solution. …………………..…………………………….………39 Chapter F: ELEMENTS OF THIN PLATE THEORY…………………………………………………………………....41 F.1) The model of a thin elastic plane plate…………………………………………………………………………………….41 F.2) The planar state of a plate. The bending state………………………………………………………………………….….41

4 F.3) Loads acting on the plate…………………………………………………………………………………………………..42 F.4.) Odd and even functions for the planar state and for the bending state. ………………………………………………….42 F.5) Mean value of a function. Equilibrium equations for thin plates. ………………………………………………………..43 F.6) Thin plate in the bending state. ……………………………………………………………………………………….…..45 F.7) BERNOULLI’s hypothesis. …………………………………………………………………..……………………….…..45 F.8) HOOKE’s law for a thin plate. ……………….……………………………………………………………………….…..46 F.9) The infinite, 1-dimensional (1-D) plate. The flexure of the lithosphere. …………..………………………………..…..47 F.10) Exterior forces on the lateral surface of the plate. Buckling. …………………………………………………………....48 F.11) The buckling of a simply leaning thin plate. ………………………………………………………………………..…..50 F.12) The infinite extended 1-D plate. ……………………………………………………………………………….…….…..51 F.13) FOURIER transforms. Properties. ………………………………………………………………………………..….…..51 F.14) Solution of the flexure equation by using FOURIER transforms. ………..………………………………………….…..52 F.15) Finite plates. ……………………………………………………………………………….………………………..…..55 a) Significance of Σ ij and Μ ij for the bending state. …….……………………………………………………..……55 b) The rectangular plate. Boundary conditions. LÉVY’s solution. …….………………………………………..………55 F.16) Vibrations of a plate laying on a viscous substratum. …………………………………………………………..………57 a) The differential equation. …………………………………………………………………………………………...…57 b) The rectangular plate with 3 embedded sides. ……………………………………………………………………...…58 Chapter G: THE SPHERICAL SHELL……………………………………………………………………………………. 60 G.1) The model. BERNOULLI’s hypothesis. Displacement vector and strain tensor. ………………………………..………60 G.2) Quasi-mean values. Equation of motion. ……………………………………………………………………………..….61 G.3) Integrals of the stress elements. Quasi-moments. ………..…………………………………………………………...….62 G.4) Integrals of displacement vector. ……………………………………………………….……………………………..….62 G.5) Equation of motion in quasi-mean values. ……………….…………………………………………………………...….63 G.6) Quasi-mean value of the shell density. The differential equation. ……..……………………………………………..….63 G.7) The buckling of a spherical shell. ……………..……………………………………………………………………...….64 G.8) Load on the upper face. Stress on the lower surface of the shell. ……………………………………………………..….66 G.9) The differential equation of time dependent flexure. ………..………………………………………………..………….67 G.10) Spherical effects with respect to the plane plate. ………….………………………………………………………....….68 Chapter H: ELEMENTS OF RHEOLOGY………………………….…………………………………………………..…69 H.1) Introduction. …………………………………………………………………………………………………………..…..69 H.2) Linear models. …………………………………………………………………………………………………….………69 a) KELVIN-VOIGT (strong viscous) model. …………………………………………………………………………….69 b) MAXWELL (viscous-elastic) model. ………………………………………………………………………….………70 c) BURGERS (general linear) model. …………………………………………………………………………………….70 d) Remarks on the linear models. …………………………………………………………….…………………………..71 H.3) Non-linear models. …………….……………………………………………………………………………………….…73 H.4) Brittle. Creep. Empirical criteria. ………..……………………………………………….………………………………73 H.5) Empirical criteria for shear-faulting. TRESCA criterion. COULOMB-NAVIER criterion. ………………………….....74 H.6) Von MISES-HENCKY criterion for ductile flow (plasticity).. ………………………..…………………………………76 H.7) Rheological models. ……………………………………………………………………………………………..……….77 a) SAINT VENANT body (elastic-plastic material). …………………………………………………………………….77 b) BINGHAM body (visco-plastic material). …………………………………………………………………………….77 Chapter I: THE ACCRETION WEDGE………………………………………………………………………………..…. 78 I.1) The model. …………………………………………………………………………………………………………………78 I.2) Equations of equilibrium. Yield condition. Stress field. ……….……………………………………………………….…78 I.3) Boundary conditions. Final results. ……..…………………………………………………………………………………79 References…………………………………………………………………………………………..…….81

Lagrangean (material) co-ordinates. x 2 . where (x1 . The elements of the tensor are changing according to a certain rule with respect to a change of the co-ordinate system. only symmetric tensors will be considered. the displacement vector can be written as t 0 . (a1) This represents a Lagrangean (material) description of the deformation process. x 2 . At a time t ≥ t 0 . A. x 3 . X 2 . A second order tensor represents mainly a 3x3 matrix. that → → U = U (X1 . Here. The superscript “t” shows the transposed tensor (matrix). Consequently. X 2 . X3 ) are representing the → x1 . as it will be seen in the next chapters. X 2 . For simplicity. X 3 .A. Fig.1. Ranalli 1987). The Lagrangean co-ordinates are usual in the Solid Mechanics. X 2 . x 2 .e. A basic supposition assumed thoroughly in that notes is that the deformation process is a continuous one.2) Invariants of a tensor. Tensor deviator. X 3 ) of X . Taking into account that the → → components (x1 .A1). the distinction between these two kinds of co-ordinates is not important. i. More details on such aspects can be found in (Aki and Richards 1980. A symmetric tensor is equal to its transpose T= T (or Tij t (a3) = Tji ). Lagrangean (material) and Eulerian (spatial) co-ordinates. (X1. t ) . all → → the components of u or U are continuous functions together their derivatives with respect to both their spatial co-ordinates or to time. X3 )of → X can be seen as functions of the (a2) This represents a Eulerian (spatial) description of the deformation process. in the Linear Elasticity. x 3 ) of the vector x are all functions of the components (X1 . The difference x − X represents the displacement vector. x 3 ) are the Eulerian (spatial) co-ordinates. A. Let the components of the tensor be real numbers . subject to a deformation process. the → → → new position vector is to be x . Such a change with respect to a rotation will be discussed later. At the initial time point has the position vector denoted by → X . while the Eulerian co-ordinates are commonly used in Fluid Mechanics.5 A) BASIC ELEMENTS. the components components (X1. the displacement vector can be written as ( → → u = u (x1 . t ) . x 2 . The continuum deformed body and the displacement vector. Consider an arbitrary material point inside a continuum body. Alternately.1) The displacement vector. with respect to the origin of a co-ordinate system (Fig. x 3 ) of the vector x . However.

(a4) (a5) It’s easy to see that the eigen-values are not changing with respect to a rotation of the co-ordinates system. (a12) 2 2 2 I 2 = T11T22 + T22 T33 + T33 T11 − T12 − T23 − T13 I 3 = T11T22 T33 + T12 T23 T13 +.u T T (a8) From (a7) and (a8) it follows that the eigen-values (and the components of the eigen-vectors) of a symmetric tensor are real numbers. the elements of the tensor are generally changing. They represent the main invariants of the tensor. By taking the complex → → * = λ u* Tu (a6) Taking into account the symmetry of the tensor. T T11 T = T12 T13 T12 T13 T22 T23 T23 T33 . (a9) 1 denotes the unit tensor.6 The scalar λ and the vector → u are representing the eigen-value and the eigen-vector respectively of that tensor if → → → → u =λ u . = det (T) With respect to a rotation of the co-ordinates system. while the third one represent just its determinant. t u * >=< u . u ≠ 0 1 T where . Suppose now that the eigen-value λ and the components of the eigen-vector conjugate (denoted by an asterisk) into (a5). (a10) Hence the eigen-values are the roots of the third degree equation (a11) where I1 = T11 + T22 + T33 = tr (T) . The first invariant is the trace of the tensor. Eq. u ≠ 0 . Because the quantities defined by (a12) can also be expressed as functions of the roots of (a11). < T u . u * >=< u .(a5) can be written as → → → → −λ u = 0 .u (a7) and → → → → → → → → →2 * >=< u .. From (a9) it follows that the next determinant vanishes T11 − λ T12 T13 T12 T22 − λ T23 = 0 T13 T23 T33 − λ − λ3 + I1λ2 − I 2 λ + I 3 = 0 . it follows their values are not changing with respect to a rotation. The tensor defined by T = T − 3 tr (T)1 * 1 . it follows → u are complex numbers. u * >= λ u .. . (a13) . the next inner product is evaluated into two different ways → → → → →2 * >=< λ u . λ* u * >= λ* u < T u .

where ρ is the density and dt 2 projecting eq. It can be shown too that the Principle of Impulse Balance leads to the next equation of motion /equilibrium div σ → d2 u represents the acceleration. → → d2 u +ρ b =ρ . that symmetric tensor has the elements 1 ∂u i ∂uj 1 = u i. (a16) According to the CAUCHY’s hypotheses there are two kinds of forces acting at an arbitrary point placed inside a body or on its boundary. spatial co-ordinates being used.g. the strain tensor is defined as ε ε ij = = → → 1 grad u + grad t u .17) dS where the tensor → b is just the gravitational → b . . they are represented by the gravity.A.7 represents the tensor deviator. having the outer pointing normal vector denoted by n . when acceleration → g . having its trace equal to zero. Aki and Richards 1980. e. three scalar equations are obtained. dividing it into a part denoted by D l at left and a pert D r at the right respectively (Fig. Or. According to the Principle of the Kinetic Momentum Balance.2). By using the spatial co-ordinates. An imaginary cross section through the deformed body. characterised by a mass density discussed in that book. Stress tensor.(a18) on the co-ordinates system axes. (a. By That equation is valid at an arbitrary point inside the body . The material points of the boundary of D r are acting on dS → by an elementary force df .2. (a14) A. An arbitrary cross section is considered through a certain point of the body. such mass forces are ignored. It follows (e. stress is a symmetric tensor. 2 (a15) where “grad” denotes the gradient. Equation of motion / equilibrium. j + u j. Suppose now a mechanical state of tension (stress) is present inside the deformed body. A surface element dS is considered on the boundary → of D l . i + 2 ∂x j ∂x i 2 ( ) . In Cartesian co-ordinates.A. The first ones are represented by the mass forces. For the problems σ represents the CAUCHY stress tensor.3) Strain tensor. Elementary computations show its second invariant is 1 2 2 2 I * = − (T11 − T22 )2 + (T22 − T33 )2 + (T33 − T11 )2 + 6 T12 + T23 + T13 2 6 That invariant is especially important to define constitutive equations for plasticity. dt 2 (a18) Fig.g. Ranalli 1987) that the next relation is valid → → df =σ n . as a result of the → action of a pair forces ± T .

8 A.e. Then eq. Neglecting the initial stress (in most cases). Hence σ11 = λ(ε11 + ε 22 + ε 33) + 2µ ε11 . an elastic. HOOKE’s law ( a21) can be reversed to give 1 . σ 23 = σ 32 = 2µ ε 23 ε = 1+ ν ν σ − E tr σ E (a. For (theoretical) incompressible rocks. (a19) σ ij = H ijkl ε kl . Other constitutive equations will be discussed in relation to the rheological bodies. µ= E . homogeneous. In the usual cases discussed here. σ or =H ε . that modulus approaches infinity. (a20) where H is a fourth-order tensor.(a19) represents HOOKE’s law. (a21) tr denotes the trace of the tensor. 2(λ + µ ) (a25) λ= The parameter defined by 1 ν E .4) HOOKE’s law. 2(1 + ν) (1 + ν)(1 − 2ν) 3λ + 2µ E = 3 3(1 − 2 ν ) (a26) χ= (a27) represents the incompressibility or bulk modulus. = λtr ε 1 + 2µ ε . σ 33 = λ(ε11 + ε 22 + ε 33) + 2µ ε 33 σ12 = σ 21 = 2µ ε12 . 1 is the unit tensor(matrix) and λ . .(a19) takes the particular form σ Here. . (a23) where the modulus of YOUNG is E=µ and the transverse contraction coefficient of POISSON is 3λ + 2µ λ+µ (a24) ν= By reversing (a24) and (a25). it is further assumed a linear relation between the stress and strain tensors. Eq. σ13 = σ 31 = 2µ ε13 . it follows λ . i. µ are the elastic coefficients of LAMÉ. isotropic medium is considered.22) Alternately. σ 22 = λ(ε11 + ε 22 + ε 33) + 2µ ε 22 .

x 2 ) e 2 (b4) . very thin cylinder of radius equal to r and height equal to H . the strain tensor ε → iii) . The forces acting upon the body are the reaction force of the horizontal plane and the gravity of the cylinder. u are obtained and the final Fig. the components of the displacement vector shape of the body is found. x 2 ∈ ∆ (b2) σ → → n = 0 .B.by using the reversed HOOKE’s law.by using the definition of the strain tensor. The boundary conditions are: -on the lateral surface of the cylinder: σ111 + σ12. The approximations of the linear theory are assumed and the variation of the density is ignored.2 + σ 33. x1 . ∆ is the disc of radius equal to r . for . x 2 ∈Γ x 3 = H. A simplified approach can be derived by using cylindrical co-ordinates. The problem is solved by following the next steps: i) .1 + σ 23.1 + σ 22. Its initial shape is a right. However. the problem here is an introductory one. (b3) Here.B1).2 + σ13.3 = 0 . The equations of equilibrium in Cartesian co-ordinates are Here. these equations are processed according to the simplifying hypothesis of the problem.2 + σ 23.3 − ρg = 0 → → n = 0 . x1 . plane x1Ox 2 . for (b1) σ -on the upper base of the cylinder x 3 ∈[0. → → → n = C1 ( x1 .2) The equations of equilibrium. the unknowns here being the components of the stress tensor σ . An elastic homogeneous isotropic body is considered (Fig. The outer pointing normal at the lateral surface of the body is a linear combination with variable coefficients of the horizontal unit vectors.9 B) DEFORMATION OF A CYLINDRICAL BODY IN THE PRESENCE OF GRAVITY B. H]. ii) . δ is the density and g is the gravitational acceleration. Boundary conditions.the equations of equilibrium are used. Simplifying hypothesis. in relation to other problems. The deformation of the body due to its own weight follows to be studied and the final shape of the body into the final equilibrium state will be found. the equations of equilibrium are processed in order to have only the components of as unknowns. vertical. absolutely rigid. The base of the body is placed on the horizontal. x 2 ) e1 + C 2 ( x1 . i.3 = 0 σ13. (a) A vertical cylinder lying on a rigid plane.e. σ12.1) The model. (b) The final shape of a vertical cross section (solid line) with respect to the initial shape (dashed line). having the centre at the origin of the co-ordinate system and the boundary denoted by Γ .1. [NO SCALE] B. So these co-ordinates will be used later.

E + ∂ u3 ∂x 2 =0 . x 2 )[σ12 e1 + σ 22 e 2 + σ 23 e3 ] = 0 (b5) → The outer pointing normal at the upper base of the body is the unit vector e . Because the cylinder is a very thin one. it is assumed that eq. For x 3 = H.(b2) becomes → → → → → → → C1 ( x1 . eq. ε 23 = E σ 23 . (b11) ∂ u1 ∂x1 = νρg (H − x3 ) . the stress at its inner points is approximately the same one to the stress on the lateral surface. x 3 = H ) = 0 (b8) The problem represented by eq. x1 . So. ∂ u3 ∂x 3 ∂ u3 ∂x1 = + ρg ( x3 − H) E ∂ u1 ∂x 3 =0 ∂ u1 ∂x 2 ∂ u2 ∂x 3 (b12) By integrating eq. (b3) gives 3 → → → → (b6) σ13 e1 + σ 23 e 2 + σ 33 e3 = 0 For Eq. E + ∂ u2 ∂x1 =0 . x 2 ∈ ∆ . ∂ u2 ∂x 2 = νρg (H − x3 ) . It follows eqs. x 2 . x 3 ) = ρg( x 3 − H ) The reversed HOOKE’s law is (b9) ε11 = E [(1 + ν) σ11 − ν(σ11 + σ 22 + σ 33)] 1 1 1 ε 22 = E [(1 + ν) σ 22 − ν(σ11 + σ 22 + σ 33)] ε 33 = E [(1 + ν) σ 33 − ν(σ11 + σ 22 + σ 33)] ε12 = Because (b10) 1+ ν 1+ ν 1+ ν σ12 . σ 33 ( x1 .(b1a)-(b1b) are identical verified. x 2 )[σ11 e1 + σ12 e 2 + σ13 e3 ] + C 2 ( x1 . From eq.(b8) has the next immediate solution σ 33 (x1. x 2 ∈ Γ eq. (b10) lead to 1 ∂ u i ∂ u j + 2 ∂x j ∂x i . H ]. x1 . ε13 = E σ13 E ε ij = eqs.(b7) holds inside the whole volume of the body.10 x 3 ∈[0.(b12) it follows that . (b1c) it follows that ∂ σ 33 ∂x 3 = ρg. (b5) is satisfied if the stress tensor has the form σ 0 = 0 0 0 0 0 σ 33 0 0 (b7) on the lateral surface of the body. x 2 .

∂f1 ∂f 2 ∂f 2 ∂f νρg + =0 .11 νρg ( H − x 3 )x1 + f1 ( x 2 . x 3 ) = −a 2 ∂x 3 It follows that .(b13d) it follows (b14) ∂ 2 f3 =0 ∂x1∂x 2 From eq. 2 ρg x 3 u 3 = E ( 2 − Hx 3 ) + f3 ( x1. x 3 ) .(b17a) is represented by a function depending on Hence both sides are equal to a constant. using eq. x 3 ) = − a1x 3 + g1 ( x 2 ) . x 2 ) = h 1 ( x1 ) + h 2 ( x 2 ) . x 3 only. x 3 ) = − a 2 x 3 + g 2 ( x1 ) . x 3 ) νρg dh ( x ) = x2 − 2 2 . material co-ordinates are used to obtain the final expression of the displacement field .(b13e) and (b13f) give ∂f 2 ( x1 . dh 2 νρg = x2 + a2 dx 2 E νρg 2 x + a 2x2 + b2 2E 2 νρg 2 x + a1x1 + b1 2E 1 (b18) f 2 ( x1 . (b13d) it follows that (b20) dg1 ( x 2 ) dg ( x ) =− 2 1 =K dx 2 dx1 where (b21) K is a constant. f 2 .e. following to be found. f 3 are finally obtained. Eqs.(b17b) gives (b19) f1 ( x 2 . g 2 ( x1 ) = − Kx1 + C 2 (b22) For simplicity. x 3 ) . h 2 ( x 2 ) = In a similar manner. h 2 are two unknown functions. x1 . eq. it follows ∂ 2 f3 ∂ ∂f1 ∂f 2 + + 2 =0 . x 3 ) νρg dh ( x ) = x1 − 1 1 ∂x 3 E dx1 (b17) The left side of eq. ∂x 3 E dx 2 ∂f1 ( x 2 . Differentiating (b13e) with x1 and (b13f) with respect to x 2 and adding the results. x 2 ) . (b16) where h1 . ∂x1∂x 2 ∂x 3 ∂x 2 ∂x1 So. ∂f3 ∂f1 νρg + = x1 E ∂x1 ∂x 3 Hence the displacement field is found if the unknown functions respect to f1 . h1 ( x1 ) = From eq. Then g1 ( x 2 ) = Kx 2 + C1 .(b15) it follows that (b15) f 3 ( x 1 . while the right side is a function of x 2 . + 3 = x2 E ∂x 2 ∂x1 ∂x 3 ∂x 2 u1 = (b13) . E νρg u 2 = E ( H − x 3 )x 2 + f 2 ( x1. ∂f 2 ( x1 . i.

it is placed on the upper base of x1 X 1 x2 = X2 x3 H Hence 0 ρg 0 + E 2 ν H 2 ( X1 + X 2 ) − 2 2 2 (b25) x1 = X1 x 2 = X 2 νρg 2 ρg 2 x 3 = ( X1 + X 2 ) − H +H 2 2E 2E From eq. The surface of the disc representing the upper base of the body is no longer a plane one.(b23) is the true displacement. the last one is a translation while the second term is the rigid rotation (b24) B. X 3 = 0) . X 2 . The final co-ordinate of the point are HX1 x1 X1 νρg HX 2 x2 = X2 + E x3 0 ( X 2 + X 2 ) / 2 1 2 Hence (b29) . The point has the co-ordinates equal to ( X1 . X 3 the cylinder. Finally. the co-ordinates of the point are = H ) . It becomes a rotational parabolic surface having the equation x3 = b) The final shape of the lower base νρg 2 ρg 2 H +H ( x1 + x 2 ) − 2 2E 2E (b28) Consider now an arbitrary point initially placed on the lower base of the body. In the initial stage. The plane of the circle is moving downward by a quantity equal to ρg( H − νr ) / ( 2 E ) .3) The final shape of the body. a) The final shape of the upper base Consider an arbitrary point of co-ordinates equal to ( X1 .(b26) it follows that (b26) 2 2 x1 + x 2 = X 1 + X 2 = r 2 2 2 2 2 . (b27) Hence the circle representing the contour of the upper base remains a circle of the same radius.12 → → u1 u ( X ) = u 2 u 3 − a1X 3 − KX 2 C 2 ν( H − X 3 )X1 ρg ν( H − X 3 )X 2 = + − a 2 X 3 + KX1 + C1 E a1X1 + a 2 X 2 b1 + b 2 2 X3 ν 2 2 ) − HX 3 + ( X1 + X 2 2 2 a2 → − a1 × X K (b23) The first term in eq. X 2 .

The surface of the disc representing the lower base becomes a rotational paraboloid having the equation x3 = c) The final shape of the lateral surface 2 ( x1 + x 2 ) 2 2E(1 + νρgH / E ) 2 νρg (b32) Consider now a point initially placed on a generatrix line of the cylinder. equal to the weight of the body. When the surface of the base is decreasing. On the lower base of the cylinder it is acting the reaction force of the rigid plane. Because of the cylindrical symmetry of the problem. approaching the paraboloid of eq. the point having the initial co-ordinates equal to ( X1 = 0. Finally. the circle representing the contour of the lower base remains a circle. its magnitude will exceed a yielding value of the material. the normal unit effort (equal to the weight divided by the contact area) is increasing. X 3 ) is considered. At a certain moment. will be no longer appropriate here. having the equation x3 = νρg 2 ρg E E r (x 2 / r − 1)2 − (x 2 / r − 1) + H − H 2 + νρg 2E 2E 2 ν 2 ρg (b34) OBSERVATION. Then. Its shape is changed from a straight line segment to a convex parabolic segment. X 2 = r. The new radius is increased by a quantity equal to νρgH / E . The initial horizontal plane of the circle is uplifted by a quantity equal to νρgr 2 / ( 2 E ) . it follows that the generatrix remains into the initial vertical plane.13 x1 = (1 + νρgH / E )X1 x 2 = (1 + νρgH / E )X 2 νρg 2 x 3 = ( X1 + X 2 ) 2 2E From eq.e.(b30) it follows that (b30) 2 2 x1 + x 2 = (1 + νρgH / E) 2 ( X1 + X 2 ) = (1 + νρgH / E) 2 r 2 2 2 . (31) i. HOOKE’s law.(b32). . valid in the elastic domain. that point has the position characterised by the co-ordinates x1 = 0 νρg ( H − X 3 )r x 2 = r + E X2 x = X + ρg 3 − HX + νρg r 2 3 3 3 E 2 2E (b33) From (b33).

it follows that (ε12. for example. (c1) ε11.Hs.2).2) The model . ε11. Because the shape of the dam. (c3) (c4) (c5) In a similar way.2 = ε 22.31 .3 − ε 23. is the free surface of the water. j + u j.12 (u 2.2 + ε12.11 = u1122 + u 2.3). A horizontal dam of infinite length is considered. the dam is deformed.22 = 2 ε 23.3 + ε 23.33 = 2 ε 31.1 − ε 31. Its specific weight is equal to Γ and its elastic constants are E and ν. that . i / 2 it follows. acting on OA side by a pressure linearly increasing with depth.12 ( + = u1.12 ε 22.14 C) LÉVY’s PROBLEM .3 = ε 33. the displacement vector has the components like .2).11 = (u11). The dam is represented by an elastic homogeneous. The planar deformation state.23 (c6)-(c8) The above equations (c3)-(c8) represent the SAINT-VENANT’s equations of compatibility.THE TRIANGULAR DAM C.22 + ε 22. N. As a result.12 = 2 ε12.33 + ε 33.2 + u 2. The cross-section is represented by a rectangular triangle OAB (Fig. Fig. Simplifying hypothesis.1 ).2 − ε12.1 + ε 31.1). A vertical cross section through the dam.1 = ε11.11 = 2 ε12. . On OA catheter is acting the hydrostatic pressure of a liquid (water) having the specific weight equal to γ.C1.22 + ε 22.31 (ε 23. The length of the base is AB=l and the height is OA=h.1).211 . isotropic material. C.12 (ε 31. Differentiating a certain element of the strain tensor ε ij = u i.2).1) The SAINT-VENANT ‘s equations.C1).22 + (u 2. = (u1.23 ε 3311 + ε11.12 (c2) Hence Also.

1) = 0 ε 22 = u 2.3) Equations of equilibrium. 1 1 ε12 = 2 (u1. C. The components of the stress tensor are σ11 = λ(ε11 + ε 22) + 2µ ε11 σ12 = 2µ ε12 σ13 = 2µ ε13 = 0 σ 22 = λ(ε11 + ε 22) + 2µ ε 22 σ 23 = 2µ ε 23 = 0 (c12) σ 33 = λ(ε11 + ε 22) + 2µ ε 33 = λ(ε11 + ε 22) = 2(λ + µ ) (σ11 + σ 22) = ν(σ11 + σ 22) λ Hence the stress matrix is = σ σ σ11 (x .2 + u 2.e.3 + u 3. σ12.15 u = u ( x1 . x ) = 1 2 ε12 0 ε12 ε 22 0 0 0 0 . In the beginning.(c13) shows that the stress state corresponding to a planar state of the strain is not generally a planar one too. The only body force acting on the dam is its weight.(c14) represent a non-homogeneous system. eq. eqs. 2 = 0 (c14) . (c11) It corresponds to a planar state of the strain (the plane here being 1-2).1) = ε12 ( x1 .2 = ε 22 ( x1 . 2 + Γ = 0 . the homogeneous system is σ111 + σ12.1 + σ 22. Γ . i. x 2 ) ε13 = 2 (u1.3 = 0 1 Hence the strain matrix is (c10) = ε ε ε11 (x .2) = 0 ε 33 = u 3. x ) = 1 2 σ12 0 σ12 σ 22 0 0 ν σ11 + σ 22 0 (c13) ( ) Because the component 33 of the stress has a non-zero value.3 + u 3. x 2 ) . AIRY’s potential. x 2 ) 1 1 u 2 = u 2 ( x1 . x 2 ) ε 23 = 2 (u 2. The equations of equilibrium are Because the presence of solved. x 2 ) u 3 = 0 It follows the strain tensor components are like (c9) ε11 = u11 = ε11(x1.

Σ12 = − ∂x1 ∂ψ (c17) ∂ϕ ∂ψ + =0 ∂x1 ∂x 2 (c18) i.16 Using an unknown function ϕ . x 2 ) represents the AIRY’s potential. Σ 22 = A . ε12 = − 1+ ν E A . (c22) Using (c22) and (c3) it follows A . It allows one to obtain the next expressions for the components of the stress tensor when the body force are absent: Σ11 = A . it follows that n → → = − sin α e1+ cos α e 2 (c28) .e. the first equation of (c15) is verified for ∂ϕ ∂ϕ Σ11 = ∂x 2 . h ]. . The final shape of the dam. + =0 Σ 12. It follows that → → n= 0 (c27) → On the side OA.22 A . eq. Σ12 = − ∂x1 Σ 111 + Σ12.e. the second equation of (c15) is verified for Σ12 = ∂x 2 It follows that ∂ψ .11 . It follows that σ σ where the outer pointing normal at the dam is → → ( − e 2) = γx1 e 2 (c26) On the side OB of the dam is acting the negligible atmospheric pressure. However.22 tr Σ = (1 + ν ) . x 2 = 0 . On the side OA of the dam is acting the hydrostatic pressure.1111 A .22 .2 (c15) (c16) In the same way. ε 22 = 1+ ν − ν∆* A E A . for x1 ∈[0.4) Boundary conditions.12 .11 (c20) From (c13).12 .11 (c23) (1 − ν)∆* ∆* A = 0 Because (c24) ν < 0.(c25) holds too in the general case of the orthogonal curvilinear co-ordinates. Σ12 = − A . C.1212 . eq. it follows that AIRY’s potential is a solution of the bi-harmonic equation ∆* ∆* =0 A (c25) Because the trace of a tensor is an invariant.(c20) has to be modified. the trace of the stress tensor can be written using LAPLACE’s operator in 1-2 co-ordinates (Σ11 + Σ 22) = (1 + ν)∆* A (c21) The components of the strain tensor are obtained using the reversed HOOKE’s law ε11 = 1+ ν − ν∆* A E A . the unknown functions are ϕ= The unknown function ∂A ∂x 2 .5 .1 Σ 22. ψ=− ∂A ∂x1 (c19) A = A ( x1 .2222 − ν ∆ A * i.2 = 0 . + − ν ∆* = −2 A .

h ] . (c34) −( bx1 + cx 2 ) − Γx 2 = 0 . B = Γh / l − 2γh 3 / l 3 Hence where . x 2 ) = 6 x1 + 2 x1 x 2 + 2 x1x 2 + 6 x 3 2 2 Using (c20). for x1 ∈[0.e. b . C = − γh 2 / l 2 (c38) (c39) u11 = ε11 = C1x1 + C 2 x 2 . 3 2 A (x1. h ] . . c .(c29)-(c30) represent 4 boundary conditions. . x 2 = x1 tan α ( bx1 + cx 2 + Γx 2 ) tan α + ax1 + bx 2 = 0 . for x1 ∈[0. d . x 2 = x1 tan α that σ12 − σ11 tan α = 0. h ] . suggesting a solution of the bi-harmonic equation (c25) which depends on 4 unknown coefficients denoted by a . for x1 ∈[0. σ 22 − σ12 tan α = 0 (c29) (c30) Eqs. x 2 = 0 ax + bx = − γx . for x ∈[0. h ] . x = 0 1 2 1 1 2 −( cx1 + dx 2 ) tan α − ( bx1 + cx 2 ) − Γx 2 = 0 . i. the solution of the homogeneous system is a b c d (c31) Σ = cx1 + dx 2 11 Σ 12 = − ( bx1 + cx 2 ) Σ 22 = ax1 + bx 2 A particular solution of the non-homogeneous system (c14) is (c32) σ11 = σ 22 = 0 σ12 = − Γx 2 (c33) It follows the general solution of (c14) is σ = cx1 + dx 2 11 σ12 = −( bx1 + cx 2 ) − Γx 2 σ 22 = ax1 + bx 2 Replacing (c34) into (c29)-(c30) it follows that . h ]. σ 22 = − γx1 On the side OB it follows for x1 ∈[0.17 σ12 = 0. x 2 = x1 tan α It follows that (c35) a = − γ b = 0 2 c = − Γ + γ / tan α 3 d = Γ / tan α − 2γ / tan α and (c36) σ = Ax1 + Bx 2 11 σ12 = − Cx 2 σ 22 = − γx1 where (c37) A = γh 2 / l 2 − Γ .

consider the case when the points A and B are fixed ones.29 l . Its final position is x1 = X1 + u 1 ( X1 . h2 l2 (c49) C3 + (c50) the final shape of the side AB is a concave parabolic segment.1) = 2 (C 2 x1 + f1( x 2 ) + C3x 2 + f 2 ( x1 ) = 1 1 1+ ν 1+ ν σ12 = − E Cx 2 E (c43) Hence C x + f ’ ( x ) = K 2 1 2 1 ’ C 3 x 2 + f1 ( x 2 ) = − K where .e. Obtain the final shape of the dam in the above hypothesis. the displacement field is = C x 2 / 2 + C x x − [C + 2(1 + ν)C / E ]x 2 / 2 − Kx + K 1 1 2 1 2 3 2 1 u 1 2 u = − C 2 x 2 / 2 + C 3 x1x 2 + C 4 x 2 / 2 + Kx1 + K 2 1 2 2 (c46) The last terms into (c46) represent a rigid roto-translation.2 + u 2. X 2 ) = h + [C 3 + 2(1 + ν)C / E]X 2 ( l − X 2 ) / 2 x 2 = X 2 + u 2 ( X1 . the unknown constants C 3 . . it is asked to γ (1 − ν) − νΓ − ( 2 − ν )γ ( h / l ) 2 ≤ 0 . assuming that γ = 1000 Kgs / m EXERCISE. f 2 ( x1 ) = − C 2 x1 / 2 + Kx1 + K 2 2 (c45) Hence. (c44) K is an arbitrary constant. that situation is not recommended in real cases. It should be outlined that the above boundary conditions on stress values on the sides OA and OB are not complete ones. i. In the same manner. As a result.18 C1 = (1 + ν )[A − ν( A − γ )] / E . u 2 = C 3 x1x 2 + C 4 x 2 / 2 + f 2 ( x1 ) But 2 (c42) ’ ’ ε12 = 2 (u1. X 2 ) = X 2 + C 3 hX 2 ( l − X 2 ) / l Elementary computations show that (c48) C3 + If 2(1 + ν) 1+ ν γ (1 − ν) − νΓ − ( 2 − ν)γ C= − E E 2(1 + ν) C< 0 E . Boundary conditions on stress values (or displacements) on the side AB are required in order to obtain an unique solution of the problem For example. C 2 = (1 − ν 2 )B / E It follows that (c40) (c41) u1 = C1x1 / 2 + C 2 x1x 2 + f1( x 2 ) where the unknown function 2 f1 follows to be found. (c51) (c52) h / l≥ 3 [γ (1 − ν) − νΓ ] / ( 2 − ν) / γ m3 . . Therefore. Because the possibility of the water to flow below the dam. ν = 0. It follows = C x 2 − h 2 / 2 + C x − h x − [C + 2(1 + ν)C / E]x l − x / 2 )2 2( 1 3 2( 2) 1 1 u1 (c47) u = C 2 h 2 − x 2 / 2 + C 3 x 2 (x1 − x 2 h / l) + {C 2 h − [C 3 + 2(1 + ν)C / E]l / 2}( x1 − h ) 1 2 An arbitrary point placed initially on the side AB has the initial co-ordinates (X1 = h.25 it follows that h ≥ 0. Γ = 2400 Kgs / For example. X 2 ) . It follows 2 f1 ( x 2 ) = − C 3 x 2 / 2 − Kx 2 + K1 . C 4 are present in (c46).

having the unit vectors → → → denoted by e . where σ f . z) will be used. its material being instantly removed. and the → div σ f = 0 .e. θ. e .THE CIRCULAR BORE HOLE / TUNNEL D. denoted by → div σ 0 = 0 (d2) p 0 . D. hence the equilibrium equation 0 is identically satisfied. corresponding to a planar state of 0 σ11 0= 0 σ σ12 0 where the components σ12 0 σ 22 0 0 . having the elastic constants denoted by E and ν . denoted by new shape of the bore hole in the final equilibrium stage. The origin of the co-ordinate system is placed at the centre of the cavity. Consequently.e. The cylindrical co-ordinate system. Suppose that a circular. infinite bore hole / tunnel is performed along the third axis. i. x 2 ) . u 3 = 0 . = σ f − σ 0 is related to the strain (d4) (d5) σ = λtr ε 1 + 2µ ε The unknown components of the displacement vector are supposed to correspond to a planar deformation state. u1 = u1 ( x1 . Because the symmetry of the problem.D1. The initial state of stress is represented by the homogeneous tensor deformation. the stress perturbation tensor by σ .1) The model.e. respectively λ and µ . i. (d3) It is also assumed that the deformation is an elastic one. isotropic medium. a new (non-homogeneous) stress value is obtained and the circular shape of the bore hole is changing too. It is assumed that the whole 3-dimensional space is represented by an elastic. r θ z Fig.2) The planar sate of deformation in cylindrical co-ordinate system. . homogeneous. u 2 = u 2 ( x1 .19 D) KIRSCH’s PROBLEM . e (see Fig. the cylindrical co-ordinate system (r. 0 0 + 0 ν σ11 σ 22 0 (d1) σ ij have constant values. x 2 ) . The mass forces are ignored. σ 0 . On the wall of the bore hole is acting now the atmospheric pressure (or the pressure of the drilling mud). A co-ordinate system having the third axis positive upward will be used. It follows to obtain the new stress.D1). i.

1 2 r → → → = − sin θ e1 + cos θ e 2 . (d12) σ rθ = − D.20 With respect to Fig.3) The circle of MOHR. 0 ν σ11 + σ 22 0 (d8) Let the stress matrix in cylindrical co-ordinates be [σ] It follows that σ rr cyl = σ rθ σ rz 0 σ11 0 σ12 1 0 σ rθ σ rz σ θθ σ θz σ θz σ zz σ12 σ 22 0 0 .D1 it follows that → → → e = cos θ e + sin θ e . σ rz = σ θz = 0 σ zz = ν(σ11 + σ 22) = ν(σ rr + σ θθ) σ12 cos 2θ (d13) . it follows σ rr = σ θθ = σ11 + σ 22 + 2 σ11 + σ 22 2 σ12 sin 2θ + σ11 − σ 22 cos 2θ 2 σ11 − σ 22 2 . . Hence the matrix for passing from the Cartesian co-ordinates to cylindrical co-ordinates is (d6) cos θ sin θ 0 Q = − sin θ cos θ 0 0 1 0 (d7) It represents a rotation of angle equal to θ in a positive (counter clockwise) sense. (d14) . σ11 − σ 22 sin 2θ + 2 . From (d4) and (d5) it follows that the stress matrix in Cartesian co-ordinates is [σ] σ11 crt = σ12 0 σ12 σ 22 0 . (d11) − σ12 sin 2θ − cos 2θ . e θ → → e z = e3 . (d9) [σ] cyl =Q [σ] cos θ crt t Q = − sin θ 0 sin θ cos θ 0 cos θ 0 sin θ 0 ν σ11 + σ12 − sin θ cos θ 0 ( ) 0 0 1 (d10) By performing the computations in (d10).

4) AIRY’s potential in cylindrical co-ordinates. usually denoted by σ ) and σ rθ (i. Consider the representation of the stress components with the AIRY’s potential in Cartesian co-ordinates. r − A . The bi-harmonic equation. (d16) where the AIRY’s potential verifies the bi-harmonic equation ∆*∆* A = 0 . From equations (d11)-(d13) it follows + σ − σ11 σ 22 rr 2 2 2 2 = σ11 − σ 22 . θ + r cos θ A .θ cos 2θ r r2 2 r r2 A . = 2 = sin θ . r − ∂x1 ∂θ ∂x1 r A . Eq.θ r = A . rr − In the same way A . σ11 = A .θθ + 2 sin θ cos θ − A . r − r A .11 r .22 .1). θ . Further discussion will be presented in relation to the empirical failure criteria of materials. rθ + A .θ r2 r r2 A.(d16) will be evaluated by using the polar co-ordinates (d17) x1 = r cos θ r = x 2 + x 2 1 2 . But x x x ∂r ∂r ∂θ sin θ = 1 = cos θ . i.(d15) is obtained. the radial stress component. x 2 = r sin θ θ = a tan( x 2 / x1 ) (d18) Than. rθ − A .e. the derivatives in eq.r . a representation of the stress components in cylindrical co-ordinates with the help of the AIRY’s potential will be obtained from (d11)-(d13). Its radius and its position of the centre allow one to obtain graphically the eigenvalues of the stress tensor.(d15) shows that σ rr and σ rθ are placed on a circle of radius equal to σ − σ 11 22 / 2 . usually denoted by τ ) are obtained at various angles θ and the MOHR’s and an identical relation obtained by replacing circle represented by eq.2 r (d22) sin θ cos θ sin θ = sin θ cos θ A . In the beginning. σ 22 = A .11 .1 = In the same way. that σ11 and σ 22 are eigenvalues of the stress tensor and σ12 = 0 .θ (d20) A . cos θ r A . D.12 .e.θ . Suppose now that σ rr (or σ θθ ) (i. (d21) A . x ∂θ cos θ = 1= ∂x 2 r 2 r (d19) A .2 = sin θ A .θθ sin 2θ + A .r = cos 2 θ A . In that case. + σ rθ 2 (d15) σ rr with σ θθ . the tangential stress. r − sin θ A . ∂A ∂x1 = ∂A ∂r ∂ A ∂θ ∂r sin θ + = cos θ A .12 = (A . (d23) .21 Suppose the Cartesian co-ordinate system is selected in order its axes to be along the first two eigen vectors of the stress tensor. =− 2 =− ∂x1 r ∂x 2 r ∂x1 r r2 It follows . θ . r + Also.e. σ12 = − A . rr + sin 2 θ + A .2 = cos θ A . r − A .

rθ + A . In the case of the cylindrical co-ordinates. Substituting the above results in the formula (a2.22 − A . (d29) c1 = 1 . θ r2 r r2 .11 = A . (d31) It follows that ∂h q = δ q 2δ β1 β ∂c (d32) Substituting the above results in the formula (a2. θθ − 2 sin θ cos θ − A . n 3 = ez . r + A . r + A . θ r r2 .5) The divergence of a tensor in cylindrical co-ordinates. c 2 = θ .θθ cos 2θ − 2 sin 2θ − A . c 3 = z and the unit vectors are . rr + σ11 − σ 22 = A . .r = sin 2 θ A . n 2 = eθ → → . r + A . rr + cos 2 θ + A . rr + From eqs.11 = − A . (d30) → → n1 = e r → → . the square of the elementary arc is equal to 2 2 ds 2 = dx1 + dx 2 + dx 3 = dr 2 + r 2 dθ 2 + dz 2 2 Hence the differential parameters of Lamé are (d28) h1 = 1 . θ r r2 r r2 . rr .22 r Hence A .77) (Ivan 1996) it follows the next formula for the divergence of a tensor in cylindrical co-ordinates → ∂T T − Tθθ ∂T ∂T divT = rr + rθ + rz + rr er ∂r r∂θ r ∂z → → Trθ ∂Trθ ∂Tθθ ∂Tθz ∂Trz ∂Tθz ∂Tzz Trz + + + +2 + + + e + e ∂r r∂θ ∂z r θ ∂r r∂θ ∂z r z (d33) D. θθ r r2 . h 2 = r . rθ + A . θθ r r2 .(d11)-(d13) it follows A . σ rθ = − A .68) (Ivan 1996) it follows the next formula for the gradient of a vector in cylindrical co-ordinates . (d26) σ rr = A . rθ + A .22 + A . h 3 = 1 the orthogonal curvilinear co-ordinates are equal to .22 A. (d24) σ11 + σ 22 = A . σ θθ = A .6) The gradient of a vector and the strain tensor in cylindrical co-ordinates. (d25) A . (d27) D.

23 → ∂u → → ∂u → → ∂u → → grad u = r e r ⊗ e r + r e r ⊗ eθ + r e r ⊗ e z r∂θ ∂r ∂z → → → → → → ∂u θ ∂u θ ∂u θ + . (d35) u r = u r ( r. according to HOOKE’s reversed law. the components of the vector u are ∂u θ u ε θθ = r∂θ + rr . By using (a2. . Consequently.7) The bi-harmonic equation in cylindrical co-ordinates. it follows the LAPLACE operator in cylindrical co-ordinates is . ε θz = θ + z r∂θ 2 ∂z . θ ) . ε rz = 2 1 ∂u r ∂u z + ∂z ∂r ∂u .43) (Ivan 1996).69) and (a2. ε zz = z ∂z → For the particular displacement field represented by (d5). (d36) ε rz = ε θz = ε zz = 0 It is the case of a planar state of deformation. ν σ rr + σ θθ 0 (d39) ( ) all the components of the stress being too functions of r and θ . (d34) e ⊗ e r + r∂θ eθ ⊗ eθ + ∂z eθ ⊗ e z ∂r θ → → → → → → → → → → uθ ur ∂u z ∂u z ∂u z + e ⊗ e r + r∂θ e z ⊗ eθ + ∂z e z ⊗ e z − r e r ⊗ eθ + r eθ ⊗ eθ ∂r z The components of the strain tensor ε ε rr = ∂u r ∂r → → t 1 = grad u + grad u are 2 u ∂u 1 ∂u . It follows from (d33) the next equations of equilibrium are obtained in the absence of mass forces: ∂ σ rr ∂ σ rθ ∂r ∂r + + ∂ σ rθ r∂θ ∂ σ θθ r∂θ + σ rr − σ θθ = 0 σ rθ = 0 r r (d40) +2 (d41) D. u θ = u θ ( r . (d38) σ σ rr = σ rθ 0 σ rθ σ θθ 0 0 . u z = 0 Hence . θ ) .e. ε rθ = r − θ + θ r 2 r∂θ ∂r ∂u 1 ∂u . (d37) ε and all the strain elements are functions of r and θ . the stress is ε rr = ε rθ 0 ε rθ ε θθ 0 0 0 0 . i.

(d47) n =1 A ∂θ 2 Hence n =1 ∑ n 2 (A n cos nθ + Bn sin nθ) n2 r2 .(d51) is 0 1 d d 1 d d A 0 r r dr = 0 . Hence the AIRY’s potential is the solution of the bi-harmonic equation ∆*∆* =0 . (d50) By using (d50). where R is the radius of the bore hole. It follows to solve (d43) by using (d44) in order to derive the AIRY’s potential. (d44) r ∂r ∂r r 2 ∂θ 2 ∂r 2 r ∂r r 2 ∂θ 2 It should be noted that the singular point r = 0 is avoided in (d44) because r ≥ R > 0 . θ ) = A 0 ( r ) + ∞ n =1 ∑ [A n ( r ) cos nθ + Bn ( r) sin nθ] .B are the solutions of the next differential equation 2 2 4 2 ’’’’ + 2 ’’’ − 2n + 1 ’’ + 2n + 1 ’ + n − 4n Φ rΦ Φ Φ r2 r3 r4 Because A is a function of r only.24 1 ∂ ∂f 1 ∂ 2 f ∂ 2 f ∆f = div( grad f ) = + . eq. (d51) A . The components of the strain will be derived by using the HOOKE’s reversed law. (d48) ∆ A = ∆ A0 + and ∞ 1 ∑ A ’’+ r A ’− n =1 A cos nθ + ’’+ 1 ’− B r B n2 r2 B sin nθ cos nθ sin nθ (d49) ∆∆ A = ∆∆ A 0 + ∞ 2 2n 2 + 1 ’’ 2n 2 + 1 ’ n 4 − 4n 2 A ’’’’ + A ’’’ − ∑ A + 3 A+ r r2 r r4 n =1 2 2n 2 + 1 ’’ 2n 2 + 1 ’ n 4 − 4n 2 + B’’’’ + B’’’ − B + 3 B+ r r2 r r4 A B . Consider the FOURIER expansion of the AIRY’s potential. allowing one to find the final shape of the deformed bore hole wall. r dr dr r dr Φ =0 . imposing the boundary conditions on the wall of the bore hole. (d43) A where the LAPLACE operator in polar co-ordinates is 1 ∂ ∂ 1 ∂2 1 ∂2 ∂2 1 ∂ ∆* = r + = + + . having the coefficients equal to functions of r A It follows ( r. the bi-harmonic (d43) is verified if ∆∆ A 0 = 0 and if the functions . The stress components will be obtained from (d27). (d52) (d53) . ∞ (d45) ∂A ∂r ∂2 A ∂r 2 ∂2 = A’ + 0 n =1 ∑ A ’n cos nθ + B’n sin nθ ∑ A ’’ cos nθ + B’’ sin nθ n n ∞ (d46) = A ’’ + 0 =− ∞ . The displacement vector will be obtained from the definition of strain elements. (d42) r + r ∂r ∂r r 2 ∂θ 2 ∂z 2 The AIRY’s potential is also a function of r and θ .

.1. a solution of the form (d57) Φ having the roots Hence the AIRY’s potential is = rm (d58) is considered. c i .. (d63) a 0 = 0 . i... Hence (d55) r Finally... b n = β n = 0 . Conditions at infinity for the stress elements. α j . rr = a 0 (2 ln r + 3) + 2b 0 − + + ∞ n =1 ∞ n =1 c0 r2 (d62) ∑ a n ( n + 2)( n + 1)r n + b n n( n − 1)r n − 2 + c n ( n − 2)( n − 1)r − n + d n n( n + 1)r − n − 2 cos nθ ∑ α n ( n + 2)( n + 1)r n + β n n( n − 1)r n − 2 + γ n ( n − 2)( n − 1)r − n + δ n n( n + 1)r − n − 2 sin nθ At great distances from the cylindrical cavity.. β j .. m 2 = − n + 2 . it follows that the exponent m is the solution of the algebraic equation m ( m − 1)( m − 2)( m − 3) + 2m ( m − 1)( m − 2) − ( 2n 2 + 1)m ( m − 1) + ( 2n 2 + 1)m + n 4 − 4n 2 = 0 .25 Hence r But d 1 d d A 0 r = a0 dr r dr dr . .2. d d A 0 r = a 0 r ln r + b 0 r dr dr .. it follows d A0 dr a 1 b = 0 r 2 ln r − + 0 r 2 + c 0 2 2 2 2 ln r + b r 2 + c ln r + d .e. 2 b 0 + 2 b 2 + 2β 2 = 0 ..(d61) ∞ + ∑ α n r n + 2 + β n r n + γ n r − n + 2 + δ n r − n sin nθ n =1 where the unknown coefficients a i . θ ) = a 0 r 2 ln r + b 0 r 2 + c 0 ln r + d 0 + ∑ a n r n + 2 + b n r n + c n r − n + 2 + d n r − n cos nθ ∞ σ θθ = A . D.8) The stress elements. δ j . By using (d61) and (d27).2. Substituting (d58) in (d52). m 4 = n + 2 .. r→∞ It follows lim σ θθ = 0 . (d54) 1 r2 ∫ r ln rdr = 2 ln r − 2 + const. m 3 = n . Hence the AIRY’s potential is (d64) . i = 0. denoting again the constants. (d60) n =1 .. γ j .(52).2. a n = α n = 0 .4. d i . it follows A ( r. n = 3. 0 0 0 (d56) A 0 (r) = a 0 r In order to solve eq. b i . j = 1. follows to be obtained.(d59) m1 = − n . the elastic perturbation has to vanish. n = 1.

A . (d71) From (d70). θ ) = c 0 ln r + d 0 + (b1 + β1 )r + + ∞ n =1 ∑ (c n cos nθ + γ n sin nθ)r − n + 2 (d65) ∞ n =1 and ∑ (d n cos nθ + δ n sin nθ)r − n ∞ c = − 0 + ∑ ( n − 2 )( n − 1)(c n cos nθ + γ n sin nθ)r − n σ θθ r 2 n =1 + n =1 ∑ n( n + 1)(d n cos nθ + δ n sin nθ)r − n − 2 ∞ . it follows that r →∞ Also. ∞ + ∑ ( − n )(d n cos nθ + δ n sin nθ)r − n −1 n =1 ∞ ∞ = ∑ n(− c n sin nθ + γ n cos nθ)r − n + 2 + ∑ n(− d n sin nθ + δ n cos nθ)r − n . r + A . ∑ n =1 n =1 ∞ (d72) σ rθ = − =− A . rθ + A .θθ n =1 n =1 (d67) (d68) (d69) Hence σ rr = A . (d74) . A .θθ = c 0 + b1 + β1 − r r2 r2 r − n =1 ∞ n =1 ∑ n 2 + n − 2 (c n cos nθ + γ n sin nθ)r − n (d70) ∞ ∑ n 2 + n (d n cos nθ + δ n sin nθ)r − n − 2 .26 A ( r. rθ = Hence ∞ n ( n − 2)(c n sin nθ − γ n cos nθ)r − n +1 + ∑ n 2 (d n sin nθ − δ n cos nθ)r − n −1 . r = r + b1 + β1 + ∑ ( − n + 2)(c n cos nθ + γ n sin nθ)r n =1 .θ n =1 n =1 ∞ ∞ = − ∑ n 2 (c n cos nθ + γ n sin nθ)r − n + 2 − ∑ n 2 (d n cos nθ + δ n sin nθ)r − n . it follows that ∞ c0 − n +1 A . it follows that r →∞ lim σ rθ = 0 .θ r r2 n =1 ∑ ( n 2 − n )(c n sin nθ − γ n cos nθ)r − n − ∑ n 2 + n (d n sin nθ − δ n cos nθ)r − n − 2 n =1 ∞ ∞ (d73) From (d73). lim σ rr = 0 A . (d66) From (d65).

∞ 1 + ∑ n(d n cos nθ + δ n sin nθ)r − n −1 1− ν n =1 E ur = ∞ 1 + ν c0 − + ∑ (n + 2 − 4 ν)(c n cos nθ + γ n sin nθ)r − n +1 E r n =2 ∞ + ∑ n(d n cos nθ + δ n sin nθ)r − n −1 n =1 . (d84) 1 − ν σ rr 1 − ν 2 ∂θ 1 − ν2 Substituting (d82). (d81) Finally . From (d39) and HOOKE’s reversed law it follows that ε i.e. (d66) and (d70) into (d84) and integrating with respect to θ . = 1 (1 + ν ) σ − νtr σ E 1 = 1+ ν −ν E σ (σ rr + σ θθ)1 .9) Strain and displacement vector. Conditions at infinity. ϕ ( θ ) ≡ 0 . (d80) ∞ c0 1 ν ur = − + ∑ n + 2 + (n − 2)(c n cos nθ + γ n sin nθ)r − n +1 1− ν r 1− ν 1 − ν2 n=2 . (d75) 1 − ν2 ν ∂u ε rr = E σ rr − 1 − ν σ θθ = ∂rr Hence (d76) ∞ ∂u r c 1 b1 + β1 ν = 0 + − ∑ ( n − 1) n + 2 + ( n − 2 )(c n cos nθ + γ n sin nθ)r − n 2 ∂r 1 − ν r 2 r 1− ν 1− ν n =2 (d77) ∞ 1 − ∑ n(n + 1)(d n cos nθ + δ n sin nθ)r − n − 2 1− ν n =1 E ∞ c0 1 ν ur = − + (b1 + β1 ) ln r + ∑ n + 2 + (n − 2)(c n cos nθ + γ n sin nθ)r − n +1 1− ν r 1− ν 1 − ν2 n=2 (d78) ∞ 1 + ∑ n(d n cos nθ + δ n sin nθ)r − n −1 + ϕ(θ) 1− ν n =1 E r →∞ Integrating (d77) it follows From lim u r = 0 . (d82) In the same way. 1 − ν2 ν ∂u θ u ε θθ = E σ θθ − 1 − ν σ rr = r∂θ + rr It follows (d83) ∂u θ E ν =− u r + r σ θθ − r .27 D. (d79) it follows that Hence b1 + β 1 = 0 . it follows after some computations that E .

dr . From (d87) it follows that ψ( r ) ≡ 0 and. let ∆p be the difference between the mud pressure and the pressure of the fluid contained by the porous rock .28 uθ = ∞ 1+ ν (1 − ν)ψ ( r ) + ∑ (n − 4 + 4 ν)(c n sin nθ − γ n cos nθ)r − n +1 E n=2 . Using the previous results. ∞ + ∑ n (d n sin nθ − δ n cos nθ)r − n −1 n =1 (d85) Form the condition it follows the unknown function r →∞ ψ = ψ ( r ) is subject to the condition lim ψ (r ) = 0 . (d86) (d87) But ε rθ = or ∂u 1+ ν 1 ∂u r u σ rθ = 2 r∂θ − rθ + ∂rθ .10) Boundary conditions for the stress elements on the wall of the circular cavity. finally. r →∞ lim u θ = 0 . (d88) 2 r σ rθ = (d89) Substituting (d73). the final expressions of the plane elements of the stress are equal to ∞ c = 0 − ∑ n 2 + n − 2 (c n cos nθ + γ n sin nθ)r − n σ rr 2 r n =1 − ∞ n =1 (d92) ∑ n 2 + n (d n cos nθ + δ n sin nθ)r − n − 2 σ θθ = − c0 + r2 n =1 ∞ + ∑ n( n + 1)(d n cos nθ + δ n sin nθ)r − n − 2 n =1 ∑ ( n − 2)( n − 1)(c n cos nθ + γ n sin nθ)r − n (d93) ∞ and ∞ 2 − n ) c sin nθ − γ cos nθ r − n − (n ) ∑ n 2 + n (d n sin nθ − δ n cos nθ)r − n − 2 (d94) n σ rθ = − ∑ ( n n =1 n =1 → The cavity wall has the outer normal (with respect to the rock domain) equal to − e and radius equal to r = R . ∞ ∞ 1+ ν ∑ (n − 4 + 4 ν)(c n sin nθ − γ n cos nθ)r − n +1 + ∑ n (d n sin nθ − δ n cos nθ)r − n −1 uθ = E n =1 n = 2 (d91) D.e. E ∂u E ∂u r − uθ + r θ 1 + ν ∂θ ∂r r dψ =ψ . (d82) and (d85) into (d89). it follows after some computations that (d90) i. ψ = Cr . In the case r ∞ of a bore hole.

It follows the final σ f satisfies the next boundary condition → → f − = ∆p er σ er . i.e. because the atmospheric pressure is negligible. it can be assumed that the stress at infinity is along its main axes. Hence 0 0 + 0 c 0 = − σ11 σ 22 − ∆p R2 2 d1 = δ1 = 0 and . it follows in the case of a tunnel that stress ∆p = 0 ).29 (usually. for r=R (d92) Hence 0 σ rr r = R = − σ rr − ∆p 0 σ rθ r = R = − σ rθ With no loss of generality. (d95) ∞ c d γ δ ∑ n 2 + n − 2 n + n 2 + n nn 2 cos nθ + n 2 + n − 2 n + n 2 + n nn 2 sin nθ n n + + R R R R n = 2 σ − σ 22 cos 2θ = 11 0 0 2 (d96) ∞ c d γ δ ∑ − n 2 − n n + n 2 + n nn 2 sin nθ + n 2 − n n + n 2 + n nn 2 cos nθ n n R R + R R + . (d93) ∞ ∞ c 0 − ∑ n 2 + n − 2 (c n cos nθ + γ n sin nθ)R − n − ∑ n 2 + n (d n cos nθ + δ n sin nθ)R − n − 2 R2 n =2 n =1 0 0 0 0 σ11 + σ 22 − σ11 − σ 22 cos 2θ − ∆p =− 2 2 ∞ ∞ n 2 − n (− c sin nθ + γ cos nθ)R − n + ∑ ∑ n 2 + n (−d n sin nθ + δ n cos nθ)R − n − 2 n n n = 2 n =1 0 0 σ11 − σ 22 sin 2θ = 2 (d94) Hence σ12 = 0 . (d97) n=2 0 0 σ11 − σ 22 sin 2θ = 2 It follows that . .

n = 3.... n = 2. R n +2 Rn γ δn + n − 2) n + ( n 2 + n ) = 0 .. n = 3.. +6 = 4 2 R4 R2 0 0 d 2 σ11 − σ 22 c2 −2 2 − 6 4 = 2 R R and (d98 2 ( n 2 ( n ( n 2 ( n 2 Hence c dn + n − 2) n + ( n 2 + n ) = 0 . n = 3... . δ n = 0 . d n = 0 .4. c 2 = n 2 4 d 2 = − 0 − 0 R σ . the value of the stress at infinity are positive ones for compression.. Rn R n +2 c dn − n) n + (n 2 + n) = 0 ...3. n = 3...4. n = 2..3. By using (d11-d13) for ..4.30 0 + 0 c 0 = − σ11 σ 22 − ∆p R2 2 0 0 c d 2 σ11 − σ 22 2 . Rn R n+2 γ δn − n) n + (n 2 + n) = 0 .. n = 2. n n+2 R R (d99) 0 0 c = − σ11 + σ 22 + ∆p R 2 0 2 0 0 σ11 − σ 22 R 2 .. (d100) σ rr f σ12 = 0 ..3.. the expressions of the final stress are equal to 0 0 0 0 σ11 + σ 22 + σ11 − σ 22 cos 2θ = 0 2 2 0 + 0 σ 22 + ∆p R 2 − 2 0 − 0 R 2 cos 2θ + 3 0 − 0 R 4 cos 2θ σ σ − 11 2 11 σ 22 2 2 σ11 σ 22 r 4 2 r r 0 0 0 0 0 (d101) and f σ θθ = σ11 + σ 22 − σ11 − σ 22 cos 2θ + σ11 + σ 22 + ∆p R 2 − 3 0 2 2 2 2 r 4 0 − 0 R cos 2θ (d102) 2 σ11 σ 22 r 4 (d103) 0 0 2 4 f = − σ11 − σ 22 sin 2θ − 0 − 0 R sin 2θ + 3 0 − 0 R sin 2θ σ σ rθ σ11 σ 22 2 2 2 11 σ 22 r 4 r In real cases..4.. 11 σ 22 4 γ n = 0 .. c = 0 .

θ ) = − E R (3 − 4 ν ) 2 . + + ∆p + (3 − 4 ν) R 1 2 E 2 2 0 0 → → 1+ ν σ11 − σ 22 sin 2θ − sin θ e + cos θ e − R (3 − 4 ν ) 1 2 E 2 Hence 0 0 0 0 x = R cos θ + 1 + ν R σ11 + σ 22 + ∆p + (3 − 4 ν ) σ11 − σ 22 cos 2θ cos θ 1 E 2 2 0 0 1+ ν σ11 − σ 22 sin 2θ sin θ + R (3 − 4 ν ) E 2 0 0 0 0 1 + ν σ11 + σ 22 σ11 − σ 22 cos 2θ sin θ x 2 = R sin θ + R + ∆p + (3 − 4ν) E 2 2 0 0 1+ ν σ11 − σ 22 sin 2θ cos θ − R( 3 − 4 ν ) E 2 After elementary computations. Its (d105) → → cos θ X = R e1 + sin θ e2 → Using (d6) the position vector in the final stage is → → → cos θ χ = X + u r ( R. it follows . Using the above results. θ) eθ = R e1 + sin θ e2 → (d106) → → 0 0 0 0 → → 1 + ν σ11 + σ 22 σ11 − σ 22 cos 2θ cos θ e + sin θ e . (d104) Consider an arbitrary point on the wall of the bore hole. (d107) .31 D. Consider again the case when the direction of the horizontal axes of the co-ordinate system is along the corresponding eigen vectors of the initial stress σ 0 . In that case. σ12 = 0 .11) The final shape of the wall. θ) e r + u θ ( R. it follows the displacement vector for the 0 points initially placed on the wall of the circular cavity is 0 0 0 0 u ( r = R. θ ) = 1 + ν R σ11 + σ 22 + ∆p + (3 − 4ν ) σ11 − σ 22 cos 2θ r E 2 2 0 − 0 1+ ν σ11 σ 22 sin 2θ u θ ( r = R . In the initial state. θ) . it has the polar co-ordinates position vector with respect to the centre of the circle is (r = R.

0 is usually a compressive one. the initial stress σ . (d109) 0 0 0 0 a = R 1 + 1 + ν σ11 + σ 22 + ∆p + (3 − 4 ν) σ11 − σ 22 E 2 2 0 − 0 0 + 0 b = R 1 + 1 + ν σ11 σ 22 + ∆p − (3 − 4 ν) σ11 σ 22 E 2 2 . i. The shape of the borehole (tunnel) in the initial state (the circle) and in the final state (the ellipse). (d110) Fig.32 0 0 0 0 x = 1 + 1 + ν σ11 + σ 22 + ∆p + (3 − 4 ν) σ11 − σ 22 X1 1 E 2 2 0 0 0 0 1 + 1 + ν σ11 + σ 22 + ∆p − (3 − 4 ν) σ11 − σ 22 x 2 = X2 2 2 E Taking into account that .e. It follows here that the major semi-axis a corresponds to the minimum stress and a ≤ b .D2.D2) 0 0 (d111) σ11 = −S1 . In real cases. (see Fig. (d108) X1 + X 2 = R 2 2 2 . corresponding to a compressive stress. it follows the final shape of the cavity is an ellipse of equation 2 x1 x 2 + 2 =1 2 b2 a where the semi-axes of the ellipse are equal to . σ 22 = −S3 where the maximum compressive stress S1 and the minimum compressive stress S 3 have positive values 0 ≤ S 3 ≤ S1 .

it follows that E 2ν t S = 1 + ν grad( div ε ) + grad ( div ε ) + 1 + ν grad(grad Θ ) . it follows that → → = − grad (ρ b ) + grad t (ρ b ) → div σ = − ρ b . (e4) σ where E ν ε + 1 + ν Θ1 1+ ν = (e5) Θ = tr σ .33 E) BOUSSINESQ’S PROBLEM .1) The equations of BELTRAMI and MITCHELL. (e10) The tensor in the first parenthesis of (e10) has the ij . (e9) 1+ ν 1+ ν Because the tensor grad( grad Θ ) is a symmetric one. It follows to obtain the partial derivative equations for the stress tensor isotropic media. homogeneous. i + 2 u q .qi = 1 1 u + u q . . . the next symmetric tensor is evaluated σ in the particular case of an elastic. In the beginning. j + u j. q = 2 . (e3) ε so = = 1+ ν ν σ − E Θ1 E .qj + ε jq . iqj + u j. S Using the equilibrium equation. j ε j.component equal to S = grad( div ) + grad t ( div ) = ( div ) + ( div ) = ε ε ij ε i. ϑ = tr ε But 1 − 2ν Θ E . ij ε ij ij E E Hence (e11) grad ( div ε ) + grad t ( div ε ) = ∆ ε + Using (e12) and (e10). qq . ij 2 i. (e1) (e2) Hence S The HOOKE’s reversed law gives = grad div σ + grad t div σ .the concentrated force acting on the elastic semispace E. ij ε ij ij ε . qqi + u q . jqi = u i. (e6) div ( f 1 ) = grad f Hence (e7) div σ Expression (e1) becomes = E ν div ε + grad Θ 1+ ν 1+ ν (e8) E ν grad ( div ε ) + grad t ( div ε ) + grad ( grad Θ ) + grad t ( grad Θ ) .qqj ∆ + tr = ∆ + 1 − 2 ν Θ = ∆ + 1 − 2 ν [grad( gradΘ )] ε . the expression (e1) becomes 1 − 2ν grad ( gradΘ ) E (e12) . i ε iq .

tr ( ∆Θ1 ) = 3∆ Θ .e.(e19)-(e20) represents the equations of BELTRAMI and MITCHELL. 1+ ν 1+ ν (e15) ∆σ − But → → ν 1 ∆Θ1 + grad gradΘ = − grad (ρ b ) + grad t (ρ b ) 1+ ν 1+ ν (e16) tr ( ∆ σ ) = ∆( tr σ ) = ∆ Θ . having as unknowns only the elements of the stress tensor . Particular cases.34 S = Applying the 3-dimensional LAPLACE operator ∆ in (e4).e. Together with appropriate conditions (in tensions) on the boundary of the elastic body. they allow one to solve the corresponding linear static problem. (e23) is verified and eq.e. (e22) (e23) From (e19) it follows that the traces of both stress and strain tensors are harmonic functions ∆Θ = ∆θ = 0 . it follows E 1 grad gradΘ ∆ε + 1+ ν 1+ ν 1+ ν ν ∆ σ − ∆Θ1 E E (e13) ∆ε = Replacing (e14) into (e13) gives (e14) S i.(e17) gives ∆Θ + It follows the trace of the stress tensor verifies the relation → 1 3ν ∆Θ − ∆Θ = − 2 div(ρ b ) 1+ ν 1+ ν ∆Θ = − → 1+ ν div(ρ b ) 1− ν (e18) (e19) Replacing (e19) into (e16) gives ∆σ + → → → ν 1 div (ρ b )1 + grad gradΘ = − grad (ρ b ) + grad t (ρ b ) 1− ν 1+ ν (e20) Eqs. (e24) it follows that (e25) i. tr ( grad grad Θ ) = ∆ Θ . → → → → tr grad (ρ b ) + grad t (ρ b ) = 2tr grad (ρ b ) = 2 div(ρ b ) (e17) Applying the trace operator in (e16) and using eqs. a) Suppose that → div (ρ b ) = 0 . where ∆ϕ = 0 . eq. → div (ρ b ) = ∆ϕ = 0 . a vector potential → ψ exists having the property that → → ρ b = rot ψ . (e21) i. b) Suppose that → ρ b = grad ϕ . = ∆σ − ν 1 ∆Θ1 + grad gradΘ .(20) gives ∆σ + 1 grad gradΘ = − 2grad ( gradϕ ) 1+ ν (e26) .

u ϑ = u ϑ ( r. In the origin of the P . It follows to find the stress and the (e28) x 3 ≥ 0 is represented by an x1 = r sin ϑ cos λ . (e27) In most real cases. h 3 = r sin ϑ (e30) c 1= r . Spherical co-ordinates will be used (Fig. x 3 = r cos ϑ Fig. n = eλ (e32) By using. u r = u r ( r. u λ = 0 . x 2 = r sin ϑ sin λ .3) The equations of equilibrium and strain tensor in spherical co-ordinates. ϑ ) . The LAMÉ differential parameters for spherical co-ordinates are h The generalised curvilinear co-ordinates are The unit vectors of the axis are 1 = 1. the gradient of a vector in spherical co-ordinates is → e λ . A co-ordinate system having the third vertical axis positive downward is used. the divergence of a symmetric tensor T in spherical co-ordinates is div T = ∂ T rr 1 ∂ T rϑ 1 ∂ T rλ 2 T rr Trϑ − + + + + ∂r r ∂ϑ r sin ϑ ∂ λ r r tan ϑ Tϑϑ + Tλλ → + e r r Tϑϑ − Tλλ r tan ϑ eϑ + → (e33) ∂ T rϑ 1 ∂ Tϑϑ 1 ∂ T ϑλ 3 T rϑ Trϑ + + + + + ∂r r ∂ϑ r sin ϑ ∂ λ r r tan ϑ ∂ T rλ 1 ∂ Tϑλ 1 ∂ Tλλ 3 T rλ 2Tϑλ + + + + ∂r r ∂ϑ r sin ϑ ∂ λ r r tan ϑ In the same way.E1): µ (or E and ν respectively). c 2=ϑ. (IVAN.E1 The spherical co-ordinate system and a vertical force of magnitude equal to P acting at the origin. homogeneous. c 3= λ (e31) → → → → → → 1= 2= 3 n e r .(e26). n eϑ . h 2=r. E.35 Applying the LAPLACE operator ∆ to eq. for example. isotropic medium having the elastic coefficients λ and co-ordinate system is acting a vertical force having the magnitude equal to displacements. ϑ ) (e29) E. Because symmetry. 1996). The semi-space elastic. the displacement vector has the components like It follows the components of both stress and strain tensors are functions of r and ϑ only. It follows the elastic linear problem involve solving harmonic and bi-harmonic equations.(24). satisfying eq.2) The model. the volume forces are neglected (or they are represented only by the weight of the body. it follows that the stress tensor is the solution of the þi-harmonic equation ∆∆ σ = 0 .

− + 2r ∂ϑ r ∂r (e35) 1 1 ∂ ur uλ ∂ uλ 1 ∂ uϑ ur . + + r ∂ϑ r r tan ϑ r + 1 ∂ σ ϑϑ 3 σ rϑ σ rϑ + + + r ∂ϑ r r tan ϑ By using (e33). 1996) the gradient of a scalar function in spherical co-ordinates is grad f = Also. ∂r ∂ 3 ∂ 2 2 r sin ϑ σ + rϑ ∂ ϑ r sin ϑ σ ϑϑ = r cos ϑ σ λλ ∂r (e39) E. ε rλ 2 r sin ϑ ∂ λ r r ∂ r ϑϑ r ∂ ϑ ε ϑλ = 1 1 ∂ uϑ 1 ∂ uλ ur uλ + 1 ∂ uλ . ε λλ = ur + uϑ r r tan ϑ (e37) σ rλ = σ ϑλ = 0 . uϑ ε = − r sin ϑ ∂ λ λλ r sin ϑ ∂ λ + r + r tan ϑ 2 r tan ϑ r ∂ ϑ In the particular case of eqs. ∂ r e r r ∂ ϑ e ϑ r sin ∆ ∂ λ e λ (e40) ∆ f = div grad f = Consider the LAPLACE equation ∂f ∂ 2 ∂f ∂ 1 ∂2f sin ϑ r + sin ϑ + 2 sin ϑ ∂ r ∂ r ∂ ϑ ∂ ϑ sin ϑ ∂λ2 r 1 (e41) . LEGENDRE’s polynomials. (IVAN.36 → grad v ∂ v r → → 1 ∂ v r vϑ → → 1 ∂ v r vλ → → = ⊗ + − ⊗ + − ⊗ + r e r e ϑ r sin ϑ ∂ λ r er eλ ∂ r er er r ∂ ϑ (e34) → ∂ vϑ → → 1 ∂ vϑ v r → 1 ∂ vϑ vλ → ⊗ → + e ⊗e + ⊗ er + + − r ∂ϑ ϑ r sin ϑ ∂ λ r ϑ r tan ϑ e ϑ e λ ∂ r eϑ → ∂ vλ → → 1 ∂ vλ → 1 ∂ vλ vr vϑ → ⊗ → + + ⊗ + ⊗ + r r tan ϑ e λ e λ ∂ r e λ e r r ∂ ϑ e λ e ϑ r sin ϑ ∂ λ The components of the strain tensor are ε rr = ∂ ur ∂r . r ε ϑλ = 0. 1 1 ∂ u r u ϑ ∂ uϑ = ε rϑ 2 r ∂ ϑ − r + ∂ r . the equilibrium equations in the absence of the volume forces are (e38) ∂ σ rϑ ∂r i. for example. it follows ε rr = ∂ ur ∂r . + ε rλ = 0 (e36) ε ϑϑ = r Hence 1 ∂ uϑ ∂ϑ ur . ε rϑ = 1 1 ∂ u r uϑ ∂ u ϑ . Using. the LAPLACE operator is ∂f → 1 ∂f → 1 ∂f → + + .(e29). σ ϑϑ − σ λλ = 0 r tan ϑ ∂ 3 ∂ 2 2 r sin ϑ σ + rr ∂ ϑ r sin ϑ σ rϑ = r sin ϑ Θ .ε = − + = + . ∂ σ rr ∂r + 1 ∂ σ rϑ 3 σ rr σ rϑ = σ rr + σ ϑϑ + σ λλ .4) LAPLACE operator in spherical co-ordinates.e.

. (e43) ( ) (e44) R and Y are two unknown functions. It follows that the trace of the strain tensor is A comparison of (e55) to (e52) shows that ε ϑλ = 0 . ϑ = R ( r ) Y( ϑ ) . ϑ ) .(e43) becomes d 2 dR d dY sin ϑ r dr dϑ dϑ dr =− = k. it follows that the general solution for that trace in the case of the BOUSSINESQ problem is Θ= n=0 E. Equation (e42) becomes (e42) ∂f ∂ 2 ∂f 1 ∂ r + sin ϑ = 0. ε λλ = (ϕ + ψ / tgϑ) / r 2 (e54) θ = tr ε = ε rr + ε ϑϑ + ε λλ = (ϕ + dψ / dϑ + ψ / tgϑ ) / r 2 (e55) . u ϑ = r ψ( ϑ ) 1 ..(43) has the form f r . It follows from (e46) that k = n ( n + 1) (e47) where Then the particular solution of eq.(e49) becomes d 2Y dY + cos ϑ + n( n + 1)Y sin ϑ = 0 2 dϑ dϑ (e49) d 2 dY (e50) 1 − z dz + n ( n + 1)Y = 0 . Eq.5) The displacement field. ε rϑ = (dϕ / dϑ − 2ψ ) / 2r 2 .37 ∆f =0 for the particular case when the unknown function has the form f = f (r. where ϕ and ψ are two unknown functions following to be obtained. (e53) ε ϑϑ = (ϕ + dψ / dϑ) / r 2 . n +1 n r (e52) According to eq. The second relation (e45) gives (e48) sin ϑ By performing the substitution z = cos ϑ eq.. P2 ( z ) = (3z 2 − 1) / 2 (e51) Because the trace Θ of the stress tensor is a solution of the harmonic equation (e23).1. Substituting (e53) into (e36) it follows that ε rr = − ϕ / r 2 . From (e45) it follows that d 2R dR r2 + 2r − kR = 0 (e46) dr dr 2 In a general case. Let’s look for a particular solution having the form R n ( r ) = r n where n is a natural number. a similar solution exists for the trace of strain tensor.(e6). it has to take A n = 0 . So. (e46) can be expressed with the aid of two arbitrary constants as R n (r) = A n r n + Bn / r n Because f has to approach finite values for r → ∞ . ε rλ = 0 u r = r ϕ( ϑ ) 1 . We look for a displacement field having the form ∑ ∞ an P (cos ϑ ) . (e45) R Y sin ϑ where k is a constant. P1 ( z ) = z . Po ( z) = 1 . uλ = 0 . ∂ r ∂ r sin ϑ ∂ ϑ ∂ ϑ By using the method of separation of the variables. the function R can be developed in power series. a solution for eq. dz The solution of (e50) is represented by the LEGENDRE polynomials denoted by Pn ( z ) . n = 0.2.

ϕ = b − ( 2 + λ / µ ) ln tg 2 2 For ϑ → π / 4 . 2 dϑ a ψ sin ϑ = − (3 + λ / µ ) cos 2ϑ + C cos ϑ + D 4 a a = D − (3 + λ / µ ) + (3 + λ / µ )sin 2 ϑ + C cos ϑ 4 2 (e65) Substituting eq. 2 ϕ = − a ( 2 + λ / µ ) cos ϑ + C . ε λλ = θ − ε rr + ε ϑϑ = (a cos ϑ − dψ / dϑ) / r 2 By using HOOKE’s law. sin ϑ 2 Because the tangential displacements has to be finite ones for ∆ → 0 . σ rϑ = µ(dϕ / dϑ − 2ψ ) / r 2 σ ϑϑ = [λ a cos ϑ + 2µ(ϕ + dψ / dϑ )] / r 2 . Θ = (3λ + 2µ )θ = (3λ + 2µ ) a cos ϑ / r 2 . σ λλ = [( λ + 2µ ) a cos ϑ − 2µdψ / dϑ] / r 2 dϕ d sin ϑ = a ( 2 + λ / µ ) sin 2ϑ . Because ∫ sin x = ln where dx tg x +C . 2 (e61) C is a new integration constant. dϑ (e62) (e63) Hence (e64) and σ rr = [(3λ + 4µ)a cos ϑ − 2µC] / r 2 . That can be avoided by taking a b − (2 + λ / µ ) = 0 . dψ ψ + = a (3 + λ / µ ) cos ϑ + C . dϑ dϑ Hence . it follows that ψ= (e67) . Hence (e56) θ = a cos ϑ / r 2 . it follows that (e66) Hence D − a (3 + λ / µ ) / 4 a + (3 + λ / µ )sin ϑ + C ctgϑ . where a is a constant remaining to be obtained.(e64) into eq. dϑ tgϑ d a (ψ sin ϑ) = (3 + λ / µ ) sin 2ϑ − C sin ϑ . it follows that ( ) (e57) σ rr = (λ a cos ϑ − 2µϕ) / r 2 .(e56). 2 2 dϑ (e60) dϕ b − a ( 2 + λ / µ ) / 2 = + a ( 2 + λ / µ )sin ϑ dϑ sin ϑ where b is an integration constant. it follows after elementary computations that (e59) sin ϑ dϕ a a = − ( 2 + λ / µ ) cos 2ϑ + b = b − ( 2 + λ / µ ) + a ( 2 + λ / µ ) sin 2 ϑ . (e58) Substituting (e58) into the first equilibrium equation (e39) and using (e56). the logarithmic term into (e62) leads to infinite radial displacements. it follows that a ϑ − a ( 2 + λ / µ ) cos ϑ + C . dϕ = a ( 2 + λ / µ ) sin ϑ .38 ϕ + dψ / dϑ + ψ / tgϑ = a cos ϑ .

. It can be seen that the points of the horizontal plane x 3 = 0 have the co-latitude ϑ = π / 2 . i. ϑ (e77) Consider an elastic hemisphere having the centre at the origin of the co-ordinate system. → → er e 1 → t → =Q e ϑ e 2 → → e 3 e λ .e. ϑ ϑ σ λλ = µ C1 + tg 2 2 − a cos ϑ / r 2 Substituting eqs. The unit vectors of the spherical co-ordinate system are related to the same vectors of the rectangular co-ordinate system by → er → = e ϑ Q → e λ where the orthogonal matrix is → e 1 → e 2 → e 3 . (e70) Q cos ϑ sin ϑ cos λ sin ϑ sin λ = cos ϑ cos λ cos ϑ sin λ − sin ϑ cos λ 0 − sin λ → → . excepting the origin. (e71) The outer pointing unit vector normal to the elastic semi-space is equal to The resulting exterior force acting on the elastic semi-space is vanishing for all the points of the horizontal plane ϑ = π / 2 .(e58) it follows that σ rϑ = µ 2C tg 2 − a sin ϑ / r 2 ϑ . e3 . σ ϑϑ = µ C1 − tg 2 2 − a cos ϑ / r 2 . (e73) σ ϑϑ = 0 σ rϑ = 0 .(e65) and (e69) give a = 2C . The curved surface S of the e r . Eqs. 2 2 2 sin ϑ dψ C a =− + (3 + λ / µ ) cos ϑ 2 ( ϑ / 2) 2 dϑ 2 cos (e68) By substituting eqs. These constants follow to be obtained taking into account that the force P concentrated in the origin of the co-ordinate system is acting on the elastic semi-space. On that surface. − e3 = − cos ϑ e r + sin ϑ e ϑ .6) Boundary conditions for stress elements. (e74) (e75) (e76) By using (e69). The final solutions. the first eq. eqs.39 ψ=C cos ϑ − 1 a ϑ a + (3 + λ / µ )sin ϑ = − C tg + (3 + λ / µ )sin ϑ .(e69) into the second equilibrium equation (e39) it can be seen that the last one is identically verified. the rest of the elastic body (i. σ ϑϑ = Cµ cos ϑ tg 2 2 − 1 / r 2 .(e65) and (e69) contain the unknown coefficients a and C .(e64) and (e68) into eq. the semi-space minus the hemisphere) is acting on the hemisphere with a total force equal to .(e74) becomes an identity and the second one leads to So. (e69) E. σ rϑ = −2Cµ tg 2 cos ∆ / r 2 → hemisphere has the outer pointing normal equal to ϑ σ rr = 2C[(3λ + 4µ ) cos ϑ − µ] / r 2 .e. → (e72) σ Substituting (e72) into (e73) for ϑ = π / 2 gives → → − = 0 .

S → (e78) dA is the surface element and the unit vectors are obtained with (e70). in relation to the computation of a building foundation. ϑ (e84) u r = 4π( λ + µ ) [2(2 + λ / µ ) cos ϑ − 1] / r Using eq. (e85) P ϑ u ϑ = 4π( λ + µ ) tg 2 − (3 + λ / µ )sin ϑ / r σ 11 σ 0 σ12 σ13 rr σ rϑ = Qt 0 Q σ 22 σ 23 σ rϑ σ ϑϑ 0 0 σ 33 σ λλ u ur 1 u = Qt u . σ ϑϑ = 4π( λ + µ ) µ cos ϑ1 − tg 2 2 / r 2 . The above solution derived for a concentrated vertical force can be used in the case of arbitrary vertical forces (spread on a certain domain) by assuming the principle of the superposition. . 2 ϑ 0 u 3 (e86) Also. the non-zero components of the stress tensor are equal to (e81) σ rr = − 2π( λ + µ ) [(3λ + 4µ) cos ϑ − µ] / r 2 . P (e82) (e83) σ r∆ = 2π( λ + µ ) µ tg 2 cos ϑ / r 2 P The non-zero components of the displacement vector are P ϑ . 3 r (e88) The BOUSSINESQ problem has a great importance in Geomechanics. It follows that 2π π / 2 → → 2 ∫∫ σ e r dA = ∫ dλ ∫ σ e r r sin ϑdϑ S = 4πC π/2 0 0 0 → → (e79) ϑ 2 ∫ [(3λ + 4µ) cos ϑ − µ]cos ϑ sin ϑ + µ tg 2 cos ϑ sin ϑ dϑ e3 = 4πC( λ + µ ) e3 Because the hemisphere is into an equilibrium state. u3 = 1+ ν P 1 2(1 − ν ) + E 2π r z2 . it follows that ∫∫ σ S Hence → → dA + P e3 = 0 . 4π( λ + µ ) → (e80) C=− Finally. the components of the stress tensor into the Cartesian base can be obtained as P .(e71).40 ∫∫ σ S where → → σ e r dA = ∫∫ rr e r + σ rϑ e ϑ dA . er P . the components of the displacement vector into the Cartesian base are (e87) Of particular importance in real life are the components σ 33 = − 3P 3 5 z /r 2π .

e.2 u 3 ( x. After the plate is deformed. the material points initially placed into the median plane are displaced on the vertical direction.2 u 3 ( x. In the second case. In the initial.− z) = − u 3 ( x. it becomes a median the co-ordinate system. is much smaller than the other dimensions (usually.e. Here. y.0) = 0 . From eq. positive downward. while the vertical component of the displacement vector is a symmetrical one with respect to the median plane (even function with respect to z ). u k ( x.− z). while m. This kind of deformation represents the planar state of the plate.p.− z) = − u k ( x. having no horizontal movement. respectively λ and µ . denoted by H = 2 h . the horizontal components of the displacement vector are symmetrical ones with respect to the median plane. is a vertical one.1) The model of a thin elastic plane plate.− z) = u k ( x. k = 1. y. (a) The non-deformed plate. This state represents the bending state of a plate.2) The planar state of a plate. y. y.41 F) ELEMENTS OF THIN PLATES THEORY F. Its height.− z) = u 3 ( x. i. The plane of equation z = 0 represents the median plane. y. having the constants denoted by E and ν . The upper face has the equation z = − h and the down face has z = h respectively. around 7-10 times). homogenous.e. u k ( x. Two particular situation for a deformed plate are considered (Fig. the horizontal components of the displacement vector are anti-symmetrical ones (odd functions with respect to z ). y. the median plane is denoted by m. Fig. (f2) i. The material of the plate is an elastic. z).F1). the material points placed initially into the median plane have no vertical displacement as a consequence of the deformation of the plate (the median plane holds a horizontal plane one). is the median surface. The thin plane plate is a cylindrical body having an arbitrary horizontal cross section in the initially non-deformed state.3) Loads acting on the plate. z). y. In the first case. (f3) By deformation. y. nondeformed state. F.F1. (b) The plate into a planar state.− z). (c) The plate into a bending state. y. being even functions with respect to the z -variable. The bending state. The median plane becomes a median surface. . i. denoted by surface. The third axis of Oz . while the vertical component of the displacement vector is an antisymmetrical one (odd function with respect to z ). the plate is a plane horizontal one. isotropic one.(f1b) it follows that (f1) u 3 ( x. Let be x1 = x . k = 1. x 2 = y .s. F.

y.42 For simplicity. y. f − ( x. On the upper surface of the plate. y. z ) = 2 f + ( x. y. y. y.2 . y. y) .− z ) = f + ( x. F. y. y. y. x1 = x . y) . → → − = s .k ( ) − .k ( ) (f14) Also. z ) = (f9) (f10) It follows that f + ( x. A similar relation holds for the odd part. f = f ( x. f + = f + . the even part of the even part is equal to the even part too.− z ) f − ( x. while the function f represents its odd part. 2 f ( x. → (f6) σ13 ( x.− z ) = f ( x. y. the forces acting on the lateral surface of the plate are neglected. having the equation z = h and the normal vector q → j → j = q ( x.− h ) = − q 3 ( x. f − = f. z ) + ∂ f ( x. h ) = q 1( x.− z ) f ( x. h ) = q 2 ( x. y. y ) . y.e. f + = f − = 0 . → e3 = q j . z) be a function of three variables.− z ) .k . So. Let j j j (f7) We shell see that the above presented deformation states are compatible only to certain distributions of volume or surface forces applied to the plate. y.4) Odd and even functions for the planar state and for the bending state. z ) = + 2 2 (f8) Let f ( x. supposed to be smooth enough. y. h ) = q 3 ( x.− z ) = − f − ( x.e. z ) . on the down surface. z ) + . z) + f ( x. y. k . In the same way. (f12) i. . z ) − f ( x. σ 33 ( x. x 2 = y . y ) . z ) = 1 ∂ f ( x. For k = 1. having the equation z = − h and the outward pointing normal vector → n = − e3 . y. z ) − f ( x.− z ) f ( x. (f11) + − The function f represents the even part of f (with respect to z -variable). The even part of an odd part (and the odd part of an even part) is vanishing. y) . for z = h .k . for z = − h . f − = f − − + . y. This property holds for the odd part.− h ) = − q 1 ( x. Hence n → = e3 . σ 23 ( x. y. y) . y. y. the partial derivative (with respect to a horizontal co-ordinate) of the even part of a certain function is equal to the even part of that derivative. y. y. It follows that f + + = f+ − . σ 33 ( x.− h ) = − q 2 ( x. it is acting the surface force σ i. y. It can be seen that f ( x.k ∂ x 2 ∂ xk ∂ xk k [ ] (f13) Hence.e. z ) + f ( x. y. y) . y) Also s s s (f5) σ i. e3 q (f4) σ13 ( x. y. it follows that f + = ∂ f + ( x. it is acting the surface force q → → → s → s = q ( x. σ 23 ( x. z) .

σ 12 . y. k . σ 13 . it follows that ( ).− z ) = f − (f15) . 1 f.3 2 ∂ z 2 [ ] ( ) Hence the partial derivative of the even part with respect to the vertical co-ordinate is equal to the odd part of the partial derivative of the function itself with respect to same co-ordinate. − − + − + − − . . ϑ . y. its mean value vanishes.strain tensor: ε 11 . y. −h (f17) f .3 .(f17) leads to f = f+ +f− = f+ +f− = f+ C = C . ε 22 .− z )] = 1 f ( x. + + − + − + + . ε 13 . (f23) zf = z( f + + f − ) = zf + + zf − = zf − = (zf )+ Hence (f24) . z ) . u 3 . i.strain tensor: ε 11 .e. (f19) (f20) (f21) Differentiating eq.the displacement vector: u1 . σ 33 . . y. σ 23 .stress tensor: σ 11 . σ 23 . y.3 ( ) (f16) By using the HOOKE’s law and the definition of the strain. The mean value of plate is denoted by 1 f ( x. − − + + + − + − + -for the bending state: . h ) − f ( x. y) = 0 . σ 12 .3 . ε 33 . u 3 .the displacement vector: u1 . + + − . Equilibrium equations for thin plates. σ 22 . σ 22 . y. y. z = h ) . ε 33 . eq. z ) + f ( x. . y ) = 2h For an odd function +h ∫ f ( x. zf − = (zf )+ . It follows that (f18) For a function C = C( x. z)dz .(f17) with respect to the horizontal co-ordinates. ε 12 . zC = 0 .− h )] = f − ( x. u 2 . z ) − f ( x.43 f + = 1 ∂ [f ( x. y. (f22) ∂z 2h h zf ( x. ε 13 . u 2 . it can be resumed that the planar state and the bending state are characterised by the next components: -for the planar state: . k = 1. σ 33 .2 .stress tensor: σ 11 . ε 23 . ε 23 . y. it follows that: f. σ 13 . y ) depending only on the horizontal co-ordinates. ε 12 . Let f = f ( x.3 . y. z ) an integrable function with respect to z -variable.3 . In the same way it follows that f − = f + . k = f For the vertical derivative. It follows that zf + = (zf )− Its mean value is . f − ( x. f computed on the thickness of the − − + − + − F5) Mean value of a function.3 = 2h Consider the function +h −h ∫ ∂f 1 1 dz = [f ( x. ϑ . ε 22 .

(f29) and (f32). h ) − σ13 ( x. z)dz i . h ) + f ( x.(f5) and (f7).2 + 21h q1j( x.3 . j = 1. −h −h −h f ( x.1 σ 23 .2 σ 23 . y) = 0 . (f31) Hence.2 σ13 . it follows the equations of equilibrium for the thin plate into a bending state are . y) + q 2 ( x.1 + (Σ12).2 + 21h [σ13( x.3 −h ( x.1 + (Μ 23). y. (f26) gives +h −h ∫ σ ij ( x.1 σ12 . y. y. (Μ13).2 + q 3 ( x.3 .3 − + − + + =0 σ 13 .3 − + − + + = 0 σ 12 .1 σ 22 .1 σ 22 . y. σ 12 .2 + σ33 − Σ 33 = 0 . y. y.1 + Σ 22 . for example). y.44 1 zf. z = h ) − f 2 ∂f 1 z dz = ∂z 2h + + + + − =0 σ 11 . Consider the weight of the plate.3 + + + + − = 0. h ) − σ 23( x. y) − q 3 ( x. y) = 0 ) .2 σ 23 .2 + 21h [σ 23( x.− h )] = 0 s (Σ11).2 σ13 . (f27) (Σ11). (f33) Here. (f32) So. y) = f + ( x.1 + (Σ12). y. it follows (Μ13). y. j = 1.1 σ 23 . using again eqs. the equilibrium equations for the planar state are (f26) Let the mean values of the stress components be denoted by 1 Σ ij = σ ij = 2h Applying the mean value operator to eqs. (f29) Μ ij = z σ ij = 1 2h +h ∫ z σ ij ( x.3 = 2h +h +h ∂ 1 + h − 1 fdz = ∫ ∫ ∂ z ( zf ) − f dz = 2h ( zf ) − h 2h ∫ . y.− h ) 2 .2. h ) + σ 33 ( x.(f26) by z and using again the mean value operator. y) 2 j j − Σ 33 = 0 .1 + (Σ 22).2 σ 33 .2.3 + + + + − + ρg = 0 σ 13 .3 +h (f25) Neglecting the volume forces (the weight of the plate itself. y) + q1 ( x.2 + 2h q 2 ( x. (f30) Multiplying eqs. the equation of equilibrium for the plate into the bending state are − + − + + = 0 σ 11 . (f28) Using eqs.2 σ 33 .− h )] = 0 (Σ12).1 + (Μ 23). the density of the plate is ρ and the acceleration of gravity is g . the equilibrium of a thin plate into the planar state leads to eqs. y. Proceeding in a similar manner to above.1 σ12 . z)dz i .− h ) − f ( x.(f5) and (f7) it follows ( ) ( Let 1 j s Σ12 .

y) + q ( x. z + u 3 ( x. y.0) . According to BERNOULLI.1 + (Σ 23). (f35) F. y) + ρg 3 1 j 1 j s s + q ( x. y.0) .0) . y) − q ( x. z) = z ∂ u1 ∂z ( x. y. z + u ( x. So.. w = w( x. y. F7) BERNOULLI’s hypothesis. y + u ( x. y) = 0 1 2 2 1 . y. y) − q ( x. z).45 j s (Σ13). y) − Σ13 = 0 2 j (Μ12).. it follows 2 u 3 ( x.0 ). z). Because the thickness of the plate is a small one. So the horizontal components of the displacement vector are u 2 ( x. So. y)+. y) + z β( x. y) = u 3 ( x. y) + q 3 ( x. w( x. y) + ρg = 0. the segment M 0 x + u1 ( x.0) + z ∂ z1 ( x. The component u1 of the displacement vector is developed in power series with respect to z -variable. y) − q s2 ( x. y. (f34) j (Μ11). the first term in eq. y. y ) . it follows that .0)+..(f36) vanishes. while the point A 0 is moving at the point 1 2 3 A 0 ( x. The results are substituted in eq. rests perpendicular on the mean surface in the deformed state.(f37) and (f40). y) − Σ 23 = 0 2 Eq. Hence . z) = w( x. z ).2 + 1 q1j( x. writing the vector along the line.2 + 1 q 2 ( x. The horizontal displacements of the points placed in the median plane are vanishing. (f37) (f38) The vertical displacement of the points placed into the median plane is denoted by i. y) + q s ( x. z) = z ∂ u2 ∂z ( x. y)) (f39) Expanding in power series the even function u 3 with respect to z -variable.2 . y.1 + (Μ 22). s (Μ11).0 ). only the second term is kept.(f34a).e. (f36) For the bending state.. (f40) Using eqs. u1( x. y. u 2 ( x. y.1 2 2 . w represents the arrow of the plate.2 + 21h q 3( x.12 + (Μ 22). Let A ( x.22 + 21h q 3( x.(f34c) is differentiated with respect to y. it is assumed that an arbitrary material segment of the plate. y. y. y.1 + (Μ12). y + u 2 ( x. z ). z ) a certain point of the plate (not placed in the mean plane) and → A 0 A is perpendicular on the median plane. y. z) − w( x. z ) . y.6) Thin plane plate in the bending state.11 + 2(Μ12).0 ) its projection on the median plane. it follows that ( ( ) ) → M 0M = (u1( x. y. initially perpendicular on the median plane in the non-deformed state. the material point A is moving at the point having the coordinates M x + u ( x. y. y. the mean surface has the equation z = w( x. After deformation.(f34b) is differentiated with respect to x and eq. It follows ∂u u1( x. z) = u1( x. Here. y. y) − q1 ( x. y ) .

0). ∂w ∂w . z) = − z ∂ x ∂w . ( x. the displacement vector has the horizontal components equal to u1( x. z) = − z ∂ y .e.0). u 3( x.. 2 2 λ∆* w + 2µ ∂ w = − H 12 ∂ x2 (f51) .0).0) = − ∂w ∂x . the components of the strain tensor are ∂2w ε11 = − z 2 ∂x ∂2w ∂2w . (f45) From eq. y. y)+. y. y. y. (f44) i. y )+.− . (f47) Hence the trace of the strain tensor is θ = tr ε = ε11 + ε 22 + ε 33 = − z∆* w .0 ). z + z 3β( x. y.1 = − ∂y ∂x ∂ w ∂x ∂ w + ∂y 2 +1 .0). (f45). ∂z ∂z Hence the direction is (f41) → M 0M = → || M 0 M|| The normal vector on the median surface is ∂ u2 ∂ u1 ( x. y) u 3( x. 2 2 n ∂ u1 ∂z ( x.(f43).(f42) to eq. the supplemental hypothesis of BERNOULLI is satisfied if ∂ u2 ∂z ( x. y)+. y.. y )+. u 2 ( x. ∂z ∂z ∂ u2 ∂ u1 = z ( x. ∂z ∂z ∂ u1 ∂z → (f42) 2 ∂ u 2 + ∂z 2 + 1 + z 2 β( x..(f46) and (f48) it follows that ∂ 2w . y. 1 + z 2 β( x. by using eq. y. y. y. ∂ x2 (f50) h2 Μ11 = z σ11 = − 3 In the same way. y) .(f37). ε12 = − z .. y. By using eqs.. (f43) Comparing eq. ( x. = λ tr ε + 2µ ε11 = − z λ∆* w + 2µ σ11 2 ∂x But (f49) 1 z2 = 2h Hence +h h2 z 2 dz = ∫ 3 −h 2 λ∆* w + 2µ ∂ w . (f46) The LAPLACE operator in horizontal components is denoted by .46 → ∂ u2 ∂ u1 M 0M = z ( x.. ε 33 = 0 .. z) = w( x. (f48) F8) HOOKE’s law for a thin plate. z) = w( x. z ( x.0) = − ∂w ∂y ∂w . .0 ).. 1 + z 2 β( x.. = −z ∂ x∂ y ε 22 ∂ y2 2 2 *= ∂ + ∂ ∆ ∂ x 2 ∂ y2 . y.

(f54) H j s s j + q ( x. Find the expressions of Σ13 and Σ 23 for a thin plate into a bending state. w ( x ) ≡ 0 ) is in an equilibrium state under the action of its own weight and of the pressure of the liquid. y) 1 2 2 1 .(f35) it follows the equation of Sophie GERMAIN: j s D∆*∆* w = ρgH + q ( x. y) − q ( x. where ρ r is the density of the filling sediments (assumed to be homogeneous ones) placed between the reference plane of elevation equal to zero and the upper surface of the plate. D (f59) 0 4 dx It is assumed that in the absence of the relief (i. The flexure of the lithosphere. p 0 − ρ r gh − ρ m gh + ρgH = 0 (f60) . with certain boundary conditions on the upper/ lower faces of the plate. i. y) + q ( x. The component u 2 of the displacement vector is equal to zero. y) + q ( x. the plate is assumed to be in a cylindrical Consider an infinite extended plate along bending state.F2.1 2 .e. EXERCISE. (f55) So. 1-dimensional (1-D) plate. In this case. on the upper face of the plate is acting the load P due to the relief and the lithostatic pressure. i.(f59) becomes d 4w = P − (ρ m − ρ r )g w( x ) + p − ρ r gh − ρ m gh + ρgH . The equation (f55) will be solved in same cases of particular importance in real cases. it is acting the pressure of the liquid of density equal to ρ m . y) 3 3 .e.9) The infinite.e.2 where 2h 3 E(1 − ν) H3 D = ( λ + 2µ ) = 3 (1 + ν)(1 − 2 ν) 12 represents the flexural rigidity of the plate. and the rest of the components does not depend on y -co-ordinate. y) − q ( x. (f58) p o is an unknown constant.e. (f52) Μ 22 = z σ 22 = − h2 3 2 2 λ∆* w + 2µ ∂ w = − H 12 ∂ y2 2 λ∆* w + 2µ ∂ w ∂ y2 (f53) Substituting eqs. On the down face of the plate.(f55) becomes D d 4w j s = ρgH + q ( x ) + q ( x ) . (f57) q 3 ( x ) = p 0 − ρ m g( w( x ) + h ) where j . Neglecting the horizontal loads. 3 3 d x4 (f56) For the case presented in Fig. q 3 ( x ) = P + ρ r g ( w( x ) − h ) s . Eq.e.(f51)-(f53) into eq.i. eq. the non-deformed plate (i.47 h 2 ∂2w H2 ∂2w Μ12 = z σ12 = −2µ 3 ∂ x∂ y = −µ 6 ∂ x∂ y . P=0). y -co-ordinate. obtaining the median surface asks someone to solve a bi-harmonic equation. F.

The load P is absent and the plate is infinite developed in a direction perpendicular on the plane of the figure. (a)A lateral view. especially those placed into the median plane. attending a If the forces deformation state like Fig. (f62) d 4 w( x ) P 4 = − w( x ) D α4 d x4 F. Consider a very thin plate simply leaning like in Fig. Eq.e.F4a.10) Exterior forces on the lateral surface of the plate. the plate loses suddenly its equilibrium state. That phenomenon represents the buckling of the plate. (b) A view from above. attempting a final configuration similar → to Fig. . (f63) To derive Sophie GERMAIN equation (f56) for the bending state. Let h be the thickness of the plate. It results the flexure equation of the 1-D plate : d 4w 1 = [P − (ρ m − ρ r )g w ( x )] d x4 D Let (f61) (ρ m − ρ r )g / D = 4 / α 4 where α is the flexural parameter of the plate. The flexure of the lithosphere under the action of the relief (P).F3.1b. The displacement field in this case is similar to the cylindrical bending. An element of the plate having the length equal to unit along that direction is considered too. the plate will be deformed according to a plane state.F3. The forces per unit length along the above direction are denoted by → ± N . A plate simply leaning. subject to the action of forces placed into the median plane.F2. Fig. If the forces ± N are above a certain critical value. of the lithostatic pressure ΓL and of the pressure of the liquid ΓM . It characterises very thin plates or bars. (or in the contrary sense. i. Buckling. being derived from the stress σ c ( positive for compression ) by N = σch (f64) → ± N are small ones. exterior forces acting on the lateral surface of the plate have been ignored.(f61) becomes .48 Fig. symmetrically with respect to the line of its supports.

in the absence of the supports (Fig.F4 (b). (c)The equilibrium of a buckled plate acted by the forces ± N and by a load q having the same magnitude but a contrary sense with respect to the load presented in Fig.F4b is . (f67) → → → N + ∑ (− p ) = 0 .F5. The equilibrium of a plate element due to the load q and to the internal tensions. the reactions will be the same. the plate is in an equilibrium state if it is acted by the lateral forces and by a vertical load to → → q = − p . (b) The equilibrium of a buckled plate acted by a vertical load p and by the reactions of the supports. ∑R It follows that → (1) = → ∑ R ( 2) .e. The equilibrium condition for the case shown in Fig. ∑ (f68) Hence for the corresponding state of stress / strain. .F5). particularly in the neighbourhood of the supports. (f65) → (1) and where ∑ R → ∑ R ( 2) are the reactions of the supports in the above cases. Because the mechanical state of stress is → → → ∑ p + ∑ R ( 2) = 0 .49 Fig. (f66) identical in both cases.F4. i. (a) The equilibrium of a buckled plate acted by the forces ± N and by the reactions of the supports.4a is → → → → → ∑ N + ∑ R (1) = 0 and the equilibrium condition for the case shown in Fig.F4b).F4c). it is necessary to find when the mechanical state of stress / strain corresponding to the presence of the lateral forces is identical to the mechanical state of stress / strain corresponding to an unknown vertical load p = p( x ) (Fig. In order to use the previous results. Consider a plate element having the horizontal length equal dx and the ends denoted by A and B (Fig. Fig.

Consider the plate in Fig. Successively. It follows to find a critical buckling value system (f77)-(f79) to have further non-trivial solutions. .→ → t . 2 d x2 D dx d 2w N + w = C1x + C 2 d x2 D where . d x4 D d x2 together with the next conditions: -at the end point having x=0: -at the end point having x=a: w = 0 .11) The buckling of a simply leaning thin plate. The angle between t and the horizontal axis is denoted by ϕ. For simplicity. C 2 are two integration constants. (f81) C1 . d 2 w / dx 2 = 0 (f78) (f79) The equations (f77)-(f79) has the trivial solution w ≡ 0 . tangent to the plate. the plate is leaning just at its ends). t=N (f72) (f73) The second equation in (f69) gives By using (f74) and (f70c) it follows that tϕ − t(ϕ + dϕ ) − qdx = 0 q = −N dϕ d 2w =N dx dx 2 . The buoyancy (f77) d 4w N d 2w + =0 . it is assumed that c force and the vertical loads are neglected. it is acting a force (per unit length) denoted t sin ϕ − (t + dt ) sin(ϕ + dϕ ) − qdx = 0 (f69) For small angles ϕ. d 2 w / dx 2 = 0 w = 0 . Equation (f76) becomes (f76) = 0 (i. At the → point B. sin (ϕ + dϕ) ≅ ϕ + dϕ ϕ ≅ tgϕ = − The first equation in (f69) gives cos ϕ ≅ cos(ϕ + dϕ) ≅ 1 dw dx dt =0 dx (f70) (f71) Hence t has no variation along x-axis. At the point A. (f63) has the next general form d2w dx 2 (f75) d 4w N d 2w 4 P + + w= D d x4 D d x2 α 4 F. representing a normal effort. vanishing according to (f78)-(f79).e. equation (f77) can be written as N = N* in order the (f80) d2 d 2w N + w = 0 . the tangential efforts are neglected. Hence (f81) is . i.e. making an angle equal to ϕ+dϕ with the horizontal axis. (f74) Hence the forces on the lateral faces of the plate are mechanical equivalent to a vertical load equal to p = p( x ) = − N It follows that eq. it follows that sin ϕ ≅ ϕ . it is acting the effort t + dt . The equilibrium conditions are t cos ϕ − (t + dt ) cos(ϕ + dϕ ) = 0 by 50 Because the plate is a thin one.F3.

(f89) ρm − ρ < Equation (f76) will be solved by using FOURIER transforms. simply leaning at all its sides.d. In real cases.F. The direct FOURIER transform ( t. ) of a function −∞ Φ( u ) is the function f ( x ) defined as . (3)Perform a study of the simply leaning 1-D plate in the presence of the buoyancy force. (5)Perform a study for the buckling a 2-D rectangular plate. (2)Perform a study for the buckling of a 1-D plate having both end points free. By integrating both sides of eq. (f86) [ ] (f87) Equation (f86) becomes +∞ ∫ w( x )dx = −∞ 1 ρm − ρ j =1 ∑ ρ jh j ( b j − a j ) m . w( x ) = C3 sin( N / Dx ) + C 4 cos( N / Dx ) . (4)Modify the equation of Sophie GERMAIN for the 2-D plate in the presence of lateral forces. b j P( x ) = ∑ P j ( x ) . (f88) Hence the area bounded by the flexural curve and the horizontal axis is a linear combination of the areas approximating the relief.12) The infinite extended 1-D plate. dx 4 dx 2 −∞ −∞ −∞ −∞ (f85) Because w and its derivatives of any order are vanishing at infinite. The plate is simply leaning at 1/3 from its length with respect to its left end. hence an upper bound for +L ∫ w( x )dx . m ρ j gh j . ) of a function f ( x ) is the new function Φ of variable u . P j ( x ) = . i = −1 (f90) The inverse FOURIER transform ( t.F. (f83) (f84) EXERCISES.2. the other being free.13) FOURIER transforms...e. 0 . It follows that the area bounded by the median curve (the flexure) and the horizontal x-axis is proportional to the load due to the relief. the flexural curve can be outlined along a finite interval denoted by the difference of the densities can be obtained as [−L .(f76) it follows +∞ 4 +∞ 2 +∞ +∞ d w d w D ∫ dx + N ∫ dx + (ρ m − ρ)g ∫ w( x )dx = ∫ P( x )dx . irrespective the presence of the lateral forces: −∞ Let an approximation of the relief be a set of m steps. defined as +∞ Φ[f ]( u ) = ∫ f ( x ) exp( − iux )dx . (1)Perform a study for the buckling of a 1-D plate having an embedded end point. pentru x ∈ a j . while from (f79) it follows the critical values N * = D( kπ / a )2 . 2 D dx having the solution From (f78) it follows that C 4 (f82) = 0 . in rest j =1 +∞ ∫ w( x )dx = −∞ 1 (ρ m − ρ)g +∞ ∫ P( x )dx .i. j =1 ∑ ρ jh j ( b j − a j ) m −L F. k = 1. L]. each one of height equal to h j and density equal to ρ j . k The lowest critical value is obtained for k = 1 .51 d 2w N + w=0 . the first two integrals in (f85) are vanishing too. i. Properties. F..

e. i. denoted by w G . of of f ( x ) by iu . (f92) −∞ Permuting the integrals.(f76) and by using the above presented properties of the FOURIER transform u 4 Φ[w p ]( u ) − i. Consider the particular case when the load Φ[w p ]( u ) = due to the relief is a load concentrated at the origin of the axes.d. Hence Φ[w G ]( u ) = It is assumed that the next condition is satisfied 1 1 2 D 2 u − K / 2 + 4 / α 4 − ( K / 2 )2 (f97) K < 4 / α2 Let (f98) A = 1 / α2 + K / 4 But .F. having the magnitude equal to unit.F.d.e.d. it follows that the direct FOURIER transform of the convolution product is the product pf the transforms of both factors of the product. The corresponding solution. it follows that the t.F. represents the elastostatic GREEN function. w( x ) = w h ( x ) + w p ( x ) p p (f94) A particular solution w = w ( x ) will be obtained applying the direct FOURIER transform to eq. The solution of eq. 4 dx w( x ) by ( iu ) 4 = u 4 . i. It allows one to obtain the solution corresponding to an arbitrary load of magnitude equal to P . (f96) D u 4 − Ku 2 + 4 / α 4 It is assumed that the value of the positive constant K = N / D is small enough. (f91) with respect to x.52 1 f ( x ) = Φ −1[Φ[f ]( u )]( x ) = 2π +∞ ∫ Φ( u ) exp( +iux )du −∞ (f91) Differentiating both sides of eq. D D α4 (f95) 1 Φ[P]( u ) . B = 1 / α2 − K / 4 (f99) .d. The direct FOURIER transform of this load is equal to unit too. N 2 4 1 u Φ[w p ]( u ) + Φ[w p ]( u ) = Φ[P]( u ) . a term corresponding to the homogeneous equation and a term corresponding to a particular solution. Consider two functions f and g of one variable. of the derivative d4w can be obtained by multiplying the t.e.14) Solution of the flexure equation by using FOURIER transforms.F.(f76) is the sum of two terms. (f93) +∞ +∞ = ∫ f ( y) exp( − iuy )dy ∫ g( z ) exp( − iuz)dz = Φ[f ] Φ[g] −∞ −∞ F. Hence the t. +∞ +∞ f ( y)g( x − y)dy exp( − iux )dx Φ[f * g]( u ) = ∫ ∫ −∞ −∞ . of the first derivative df can be obtained by dx multiplying the the t. Their convolution product is +∞ ( f * g )( x ) = ∫ f ( y)g( x − y)dy .

Φ −1 + B (u + A )2 + B2 (u − A )2 + B2 By using the property of the derivative. 1 d A exp(− B x ) sin(Ax) = exp(− B x ) cos( A x ) − sin ( A x ) = B dx B (f107) [ ] Using the above results.(105).53 1 α2 u u = − u 4 − Ku 2 + 4 / α 4 8 A (u + A )2 + B2 (u − A )2 + B2 α2 1 1 + + 4 (u + A )2 + B2 (u − A )2 + B2 The next result is valid (Rîjic and Gradstein 1955) . B 8D A sin (A x ) cos( A x ) α2 w G ( x) = exp(− B x ) + (f109) . Φ −1 2B (u ± A )2 + B2 1 1 ( x ) = 1 exp(− Bx ) cos( A x ) . representing a general property of the GREEN function: Φ[w p ] = Φ[P] Φ[w G ] . it follows that ] . −∞ (f111) . 8D B A It can be observed that w G → ∞ for B → 0 . z ≠ 0 . it follows after some elementary computations that sin( A x ) cos( A x ) α2 Φ[w G ]( u ) = Φ exp(− B x ) + (f108) .(f104) from eq. From (f96) and (f97) it follows that (f110) Hence the solution for an arbitrary load is the convolution of the load due to the relief and the function given by (f109). 2 B dx (u + A )2 + B2 In the same way [ [ ] (f104) iu 1 d Φ exp(− B x + iAx) ( u ) = 2 B dx (u − A )2 + B2 Subtracting eq. (f106) u u Φ −1 − (u − A )2 + B2 (u + A )2 + B2 . it follows (f102) (f103) iu 1 d Φ exp(− B x − iAx) ( u ) = . 2 z2 + x 2 (f101) By using (f101). (f100) ∞ ∫ 0 z sgn Re( z ) π cos xdx = exp( − z ) . the next inverse FOURIER transforms are obtained 1 ( x ) = 1 exp(− B x m iAx) . corresponding to the buckling of the infinite plate in the presence of a lateral compressive stress. (f105) u u 1 d Φ exp(− B x ) sin(Ax) ( u ) = − B dx (u − A )2 + B2 (u + A )2 + B2 Hence [ ] . w p ( x) = ( P * w +∞ G )( x ) = ∫ P( y ) w G ( x − y )dy .

m K 1 1 ’’ C2 = ∑ ρ jgh j exp − Ba j sin Aa j − exp − Bb j sin Ab j − 2 C1 − W0 2AB 4ABD j =1 [ ( ) ( ) ( ) ( )] (f120) C3 = C 4 = 0 . (f112) K I( z ) = sgn( z )exp( − B z )−2 cos( A z ) + sin( A z ) + 2 (ρ m − ρ)g / D − K 2 / 4 The solution of the homogeneous equation can be immediately derived as (f113) w h ( x ) = [C1 cos( Ax ) + C 2 sin( Ax )]exp( − Bx ) + [C 3 cos( Ax ) + C 4 sin( Ax )]exp( Bx ) Hence the general solution is (f114) w( x ) = 1 4(ρ m − ρ) j=1 + [C1 cos( Ax ) + C 2 sin( Ax )]exp( − Bx ) + [C3 cos( Ax ) + C 4 sin( Ax )]exp( Bx ) It follows to find the unknown coefficients C1 .(f87). Derive the expression of w ( x ) for a load due to a relief having the equation and ρgh 0 sin( 2π x / λ ) . EXERCISE. To obtain the values of the unknown coefficients C1 . A finite element algorithm based on the continuity of the values of the flexure.(116). (f121) A finite plate of variable thickness can be approximated in real cases by a sum of n elements having constant thickness and homogeneous elastic properties. . For the infinite plate. to the next conditions: lim W( x ) = 0 (f117) x →∞ W( 0 + 0 ) = W0 (f118) d 2W ’’ (f119) ( 0 + 0) = W0 = − M 0 / D dx 2 ’’ where W0 . In the case of the semi-infinite plate the flexure W is subject. + λ / 2] P = P( x ) = ρ gh ( x ) = . C 3 and C 4 for each element. of the bending moment and of the share force has been derived by Ivan (1997). proper conditions have to be verified at the ends of each element. pentru x ∈[− λ / 2. (f122) 0 . the flexure W is subject to the next conditions: lim W( x ) = 0 (f116) x →±∞ Hence the coefficients C1 . C 3 and C 4 are vanishing and the general solution is just the particular solution represented by eq. for example. in rest where h 0 is the amplitude of the relief and λ is its wave-length. It follows ∑ ρ jh j[I(b j − x) − I(a j − x)] m (f115) C1 = W0 − j =1 ∑ ρ jgh j[I(b j ) − I(a j )] m . that of the second 0 derivative of the flexure and that of the bending moment respectively at the left end of the plate where the origin of the x-axis is selected.54 For the approximation of the relief represented by eq. C 2 . of its first derivative. W0 and M (positive when acting into a clockwise sense) are the values of the flexure. C 2 . it follows that w p (x) = where 1 4(ρ m − ρ) j =1 ∑ ρ jh j[I(b j − x) − I(a j − x)] m . C 3 and C 4 in some particular cases. C 2 .

15 b) The rectangular plate. it is necessary. y ) = 0 . Among most commonly used boundary conditions are -the embedded side: the flexure of the plate and the derivative of the flexure are both equal to zero w ( x. Boundary conditions.F4).15 a) Significance of Σ ij and Μ ij for the bending state. Μ 21 are torsion moments (Fig.F5). ∂ w ∂ x = 0 . =0 . M11 and M 22 are bending moments. y ∈[− b. b ] ).3.e. Some particular. σ 23 are even functions. y ) = 0 . Similar considerations allow one to conclude that the mean values Μ12 . Σ 23 are representing share forces. for example. σ ij is the projection along j -axis of the surface force acting on the surface having the outer pointing normal equal to mean values Σ13 .15) Finite plates. A similar torsion moment is acting on the side having the normal 2. (f124) Μ11 i. acting like in Fig. the boundary conditions on the contour of the plate are essential ones in order to obtain the values of the integration constants. Consider. Let consider a rectangular plate having the sides equal to 2a and 2b respectively (Fig. i. while Fig. The torsion moment M12 acting on the side having the outer pointing normal 1 is presented in Fig. the Fig. The mean values Σ 13 (a) and Σ 23 (b). Both of them are share forces.F4. In the case of the finite plates. j=1. a previous examination of the mean values Σ ij and Μ ij . (f123) -the rotating side: the flexure of the plate and the bending moment are both equal to zero: w ( x. The bending moments M11 (a) and M22 (b).F3.: . n i . LÉVY ‘s solution. F. F. most common cases will analysed in detail.55 F.4c. but it is not presented in the figure. the side AB. According to the definition of the stress tensor elements. As a consequence. having the equation x = a .2.F3. Because the elements σ13 → .

2. a a (f133) In the beginning. ∂2w π2 =− ∑ k 2 f k sin( kπ x / a ) ∂ x2 a 2 k =1 (f128) Hence (f129) It follows the conditions (f125) are fulfilled for the two opposite articulated sides. y ) = 0 . consider the rectangular plate with two opposite articulated sides.F5. r 1 2 3 4 a a hence the solution of the homogeneous equation is (f134) .. (f126) with the boundary conditions (f125) written for the form x = ± a and a 2-D LAPLACE operator. i. a] ....(f130) are jπ x multiplied by sin . r = r = kπ / a . The result is integrated on the interval [− a. . ∑ a a a k =1 The function (f130) q ( x.56 2 * w + 2µ ∂ w = 0 . the bending moment and the torsion moment are all equal to zero. k = 1. y ) is expanded in FOURIER series and the coefficients are identified. it follows to solve the simplified equation of the flexure ∆ ∆ w = q ( x. (f131) ∫ a a a . ∞ (f127) w ( ± a. w ( x.. (f125) Fig.2. for x = ± a .e. Neglecting the lateral forces and the gravity. taking into account that a +a kπ x jπ x 0 .. ∞ kπ x kπ 2 " kπ 4 "" f k ( y ) − 2 f k ( y ) + f k ( y ) sin = q ( x. .. As an example. Similar conditions can be derived for plates of arbitrary shape. y ) / D . λ ∆ ∂ x2 -the free side: the share force. ∞ . deci r = r = − kπ / a . it follows that ∂ 2 w / ∂ x 2 = ∂ 2 w / ∂ y 2 = 0 . y)sin a dx . k = 1. (f132) a a Da −a 2 4 kπ " kπ "" Fk ( y) − 2 Fk ( y) + Fk ( y) = 0 . pentru j ≠ k sin sin dx = . The rectangular plate.. the next homogeneous equation is solved. Substituting (f127) into (f126). pentru j = k −a Hence +a 2 4 kπ x "" ( y ) − 2 kπ f " ( y) + kπ f ( y ) = 1 fk k k ∫ q ( x. Consider a particular solution having w( x. The characteristic equation is 2 4 4 − 2 kπ r 2 + kπ = 0 . y) = it follows that k =1 ∑ f k ( y)sin( kπ x / a ) . j = 1.. y ) = 0 .2. Both sides of eq. y ) / D .

(f141) where ρ is the density of the filling sediments and P is the load. y. Hence the displacement vector has the elements u x ( x. y. having the unit → vector denoted by •• u d ∂ Μ xx ∂x + ∂ Μ xy ∂y + q ( x. the horizontal loads for the upper face of the plate are neglected. k = 1. C k . The LAMÉ elastic coefficients are λM . t ) ∂x + ∂ q ( x. a co-ordinate system is used having the horizontal axes x and y. By applying the mean-value operator. y. t ) ∂y x y 2 x y As usually. The last one can be obtained by the usual techniques (e. y. t ) / 2 − Σ yz = ρ z u y y y and e z . the surface forces being assumed to be qx = 0 F u . y. The last term in (f141) is an inertial one. B k . t ) − q ( x. z. (f137) ∂ Σ xz ∂x + ∂ Σ yz ∂y + ∂w •• 1 d u ( x. The z -axis is positive downward. t ) + q ( x. (f140) H l u u l + ∂ q ( x. t ) = w( x. t ) . u y ( x. u z ( x. the equations of motion for the bending state are . t ) − q ( x. t ) − q ( x. y. t ) = − z ∂ x . CAUCHY method). For the material below the plate. t ) + ρg = ρ u z q z 2h z ∂w . qy = 0 u . t ) − q ( x.. z. y. (f139) Substituting (f139) into (f136)-(f138) it follows after elementary computations that •• ρ •• l u D∆* ∆* w = ρgH − ρH w + H 3 ∆* w +q ( x. The general solution of equation (f132) is the sum of (f135) and a particular solution. y. t ) + q ( x. t ) = − z ∂ y .g. the flexure is considered as a function of both spatial co-ordinates and time. being solved in the case of the 2-dimensional plate. F. The above approach is due to LÉVY.57 Fk ( y) = A k cosh kπ y kπ y kπ y kπ y + B k sinh + y C k cosh + D k sinh . a) The differential equation. y.. y. z. In this chapter. D k . t ) / 2 − Σ xz = ρ z u x x x •• u d ∂ Μ xy ∂x + ∂ Μ yy ∂y + q ( x. (f138) It will be assumed that Bernoulli’s hypothesis is valid for all time. (f136) . y.16) Vibrations of a plate laying on a viscous substratum. As usually. y. are some constants following to be obtained from the boundary conditions on the other two (non-articulated) sides of the plate. t ) z z 12 . The corresponding differential equation is derived. y. (f135) a a a a where A k . q z = ρFgw + P(1 − w / g) • + 2 µ M ε + 2 ηM ε u •• . the next constitutive equation is assumed σ where = [ p0 − ρMg( H + w )]1 + λM tr ε 1 (f142) p 0 is a reference pressure and ρM is the density of the material below the plate. y.2.. µ M and ηM is the viscosity. y. Hence the loads on the lower face of the plate are . y.

the periods are •• • 4π 2 τ + 2ζ mn τ + τ=0 .5) 2 H 2 ] / (ρa 2 ) + 12ρ*gH / a 2 A mn x x y y .5)πy / L y ] τ n ( t ) . M M w e = w e ( x.e.58 • ∂w ∂w l q x = µ M ∂x + ηM ∂x d H • ∂w ∂w l . at any time. Let equilibrium in the presence of the load. n = 1. Suppose the plate is welded according to δ( x. q = µM + ηM y ∂y ∂x . For m . σ c acting along the x. t ) = 0. 2. t ) = w( x. y.. t ) = 0. y. L y respectively. (f143) q z = p0 − 2 λM∆*w + ρMg( H + w) Again. y.. .. The reference pressure p 0 is selected in order the flexure w to vanish in the absence of the load P . y. t ) = sin(mπx / L x )sin[( n − 0. hence the second term on the left side of (f144) vanishes. (f149) Substituting (f149) in (f147) it follows the modes are damped harmonic. i..and y-axes respectively at the ends x y of the plate follows to be considered further.. (f151) 2 − 12π 2 [σ c m 2 H 2 + σ c ( n − 0. Because the load is mainly represented by the relief. i. A particular solution satisfying (f148) is with δ mn ( x. t ) − w e ( x. δ( 0. n = 1. (f148) δ( L x . y) be the equation of the flexure corresponding to the state of x D∆*∆*w e + H σc ∂2 w e / ∂x2 + σ c ∂2 w e / ∂y2 + (ρM − ρ F )gw e = P y Consider the difference (f145) δ = δ( x. 2 Tmn (f150) Tmn = 2π H a A mn + 12 . t ) = 0. m . y. with with 0 ≤ x ≤ Lx 0 ≤ y ≤ Ly 0 ≤ y ≤ Ly . a generalisation of the Sophie GERMAIN equation for a time dependent flexure is obtained as H ∂2 w ∂2 w M F + σc D∆*∆*w + λM − µ M ∆*w + H σ c y 2 + (ρ − ρ )gw x 2 2 ∂x ∂y (f144) = P + ηM •• ρ •• H *• ∆ w − (ρH + P / g) w + H 3∆* w 2 12 b) The rectangular plate with 3 welded sides. Consider the lengths of the sides are L x . Finally. For usual materials λ = µ ..2. the inertial term is usually negligible on the right side of (15). having the elevations much smaller then the thickness of the plate.e. y ) It follows that (f146) M F g 2 2 1 * •• 6ηM * • 12 •• *∆*δ + 12 1 σc ∂ δ + σ c ∂ δ + 12 ρ − ρ ∆ δ + ∆ δ= ∆ δ− δ (f147) x 2 y 2 ρ ∂x ∂y H 3a 2 a2 ρH 2 H2 ρa 2 H 2 where a denotes the velocity of P-waves through the plate.0. a correction due to the compressive horizontal stresses σ c .

H y = H / Ly . the decay constant (f158) or (f159) can be obtained as the ratio between the amplitude of the velocity ∂δ / ∂t and the amplitude of δ .b). a critical viscosity can be found from (f154) ζ cr = 2π / Tmn .e. (f156) For values of viscosity less than the critical value (f156). It follows that for large scale of times. In many real applications. that ratio can be used to estimate the mean viscosity of the material below the plate ηM = ρH 2 2 π 2 A mn + 12 3C T2 A mn 11 (f160) A numerical application with respect to the Moesian Platform. For each mode. The decay constants are obtained from (f152) and the periods are * cr 2 Tmn = Tmn / 1 − ζ mn / ζ mn (f157) For values of viscosity greater than the critical value.59 and the decay constants are 3ηM A mn ζ mn = ρH 2 A mn + 12 where . the motion of the plate is represented by a sum of damped oscillations. (f152) H x = H / Lx and . a micro-plate bounded by Carpathians and Balkan Mountains and by the Black Sea is presented in detail by (Ivan 1997a. an approximate value of (f159) is 2 4π 2 / T11 2π 2 ≅ C= 2 2 − 4π 2 / T 2 ζ11T11 ζ11 + ζ11 11 (f159) With regard to (f149). the mn mn solution of (f150) behaves like exp( − Ct ) . (f158) all the other terms being faster attenuated. where the decay constant is 2 2 C = ζ11 − ζ11 − 4π 2 / T11 . ρ* = (ρM − ρ F ) / ρ .5) 2 H 2 x y According to (f139). . By using eq. (f155) η cr = mn 2πρH 2 A mn + 12 3Tmn A mn .(f152). mn i. the modes are both toroidal and spheroidal. (f153) A mn = π 2 m 2 H 2 + ( n − 0. the motion of the plate is aperiodic and the characteristic roots of (21) are 2 2 − ζ mn + ζ 2 − 4π 2 / Tmn and − ζ mn − ζ 2 − 4π 2 / Tmn .

θ. t ) where . ϕ . Hence the unit vector normal to the median surface at a certain point of co-ordinates (r = R − w. ϕ ) is 2 2 → → ∂w → 1 ∂w → / 1 + 1 ∂w + 1 ∂w . In the initial state. (g2) w = w(θ. some of the derived results are also valid in the case of a more general. ϕ. n = e r+ + r∂θ eθ r sin θ ∂ϕ e ϕ r 2 ∂θ sin 2 θ ∂ϕ Neglecting the quantities of the second order. It follows that the segment A 0 B 0 has the unit vector equal to e . θ. A further hypothesis on x and y will be later considered. positive downward. non-elastic λ . t ) u θ ( r. (g8) where x and y are two unknown functions representing the horizontal displacements of the points initially placed on the median sphere. θ. It follows that a segment inside the shell. (g1) where R is the radius of the sphere. t ) . ϕ. ϕ . t ) = − w (θ. in the deformed state of the shell. However. t ) r − R ∂w + x(θ. t ) . Displacement vector and strain tensor. It follows the elements of the strain tensor are . ϕ) is used. Assuming the shell is thin. ϕ ) and B 0 ( R + h . the homogeneous density is denoted by ρ 0 and the median spherical surface of the shell has the equation r=R . µ is considered and the usual spherical co- r = R − w (θ . θ. t ) = R sin θ ∂ϕ + y(θ. (g6) u r ( r . homogeneous shell having the elastic moduli equal to ordinate system spherical shell. θ. the point A 0 is displaced to a point A having the position vector equal to → → → rA = R e r + u ( R . BERNOULLI’s hypothesis. ϕ ) . quantities of the → 2 order h are neglected and the segment AB has the unit vector equal to → → → → ∂u → ∂u → ∂u ϕ → rB − rA / rB − rA ≅ 1 + r e + θ e + . At an arbitrary r time. From (g4) and (g7). θ. At a certain time during the deformation. t ) . it follows that (g3) → → ∂w → 1 ∂w → n ≅ er+ + . It is assumed that Bernoulli’s hypothesis is valid for all Here. θ. ϕ ) . the median surface will be (r.60 G) THE SPHERICAL SHELL G. θ. θ. will be always normal to the median surface during the deformation. ϕ. non-deformed state of the shell and two points of co-ordinates A 0 ( R . → → where H=2h is the thickness of the shell. (g4) R∂θ eθ R sin θ ∂ϕ e ϕ Consider the initial. ϕ. ϕ. (g7) ∂r r ∂r θ ∂r e ϕ The partial derivatives in (g7) are computed at the point ( R . θ. ϕ . t ) = R ∂θ r − R 1 ∂w u ϕ ( r. ϕ.1) The model. ϕ . which is initially normal to the median spherical surface. t ) is the flexure of the shell. (g5) while the point → u is the displacement vector at a point of certain spherical co-ordinates. time. it is supposed that the displacement vector has the elements B 0 is displaced to the point B having the position vector equal to → → → rB = ( R + h ) e r + u ( R + h. A spherical elastic.

(g17) 1 M rr = H The equations of motion in spherical co-ordinates are R+h R−h ∫ r ( r − R ) σ rr dr R . the corresponding quasi-mean value is defined. ∂ 2u ϕ 1 ∂ σ ϕϕ + + =ρ r∂θ r tan θ r sin θ ∂ϕ ∂t 2 ∂ 2uθ 1 ∂ σ θϕ σ θθ − σ ϕϕ + =ρ r sin θ ∂ϕ r tan θ ∂t 2 ∂ 2u r − ρg = ρ ∂t 2 . that density will be replaced by the initial density ρ 0 at the right side of (g19).61 ∂u r =0 . ε rr = (g9) (g10) (g11) (g12) (g13) and u r − R 1 ∂2w 1 1 ∂u ϕ 1 ∂w 1 1 ∂y x + = + ur + θ = + −w+ (g14) ε ϕϕ r sin θ ∂ϕ tan θ rR sin 2 θ ∂ϕ 2 tan θ ∂θ r sin θ ∂ϕ tan θ Hence the trace of the strain tensor is tr ε = ε rr + ε θθ + ε ϕϕ = where the two-dimensional LAPLACE operator is R(r − R ) ∗ 1 ∂ ∂y ∆ w+ −2w sin θ + ∂θ (x sin θ) + ∂ϕ . by 1 Σ rr = σ rr = H In the same way. Consider the stress tensor in spherical co-ordinates. Equations of motion. (g18) ∂ σ rr ∂r ∂ σ rθ ∂r ∂ σ rϕ ∂r σ ∂ σ rθ + σ rθ + 2 rr + r r∂θ 1 ∂ σ rϕ + − r tan θ r sin θ ∂ϕ σ θθ + σ ϕϕ r σ ∂ σ θθ + + 3 rθ + r r∂θ +3 σ rϕ r + ∂ σ θϕ ρ denotes the density of the shell in the deformed state. In the above equations. ∂r u ∂u x 1 1 ∂u ε rθ = 2 r ∂θr − rθ + ∂rθ = − 2r . (g19) 2σ θϕ . the corresponding quasi-moment is defined by R+h R−h ∫ r dr R σ rr . Assumed to be a negligible second order effect. y 1 1 ∂u r u ϕ ∂u ϕ = ε rϕ 2 r sin θ ∂ϕ − r + ∂r = − 2r . ∂θ ∂u ϕ r − R ∂ 1 ∂w 1 1 ∂x ∂y uϕ 1 1 ∂u θ y ε θϕ = 2r sin θ ∂ϕ − tan θ + ∂θ = rR ∂θ sin θ ∂ϕ + 2r sin θ ∂ϕ + ∂θ − tan θ . 1 ∂u θ r − R ∂ 2 w 1 ∂x ε θθ = r ∂θ + u r = rR 2 + r ∂θ − w .2) Quasi-mean values. + 2 sin θ ∂θ ∂θ sin θ ∂ϕ 2 R 1 (g16) G. for example. For each element of the tensor. (g15) r r sin θ ∆∗ w = ∂ ∂w 1 ∂ 2w sin θ .

3) Integrals of the stress elements. (g24) R h 2 h 2 1 + σ U + 1 − σ L − R Σ = rθ rθ rθ 2 R R 1 H Also. The next two equations in (g19) will be multiplied by r and the quasi-moment operator defined by (g18) will be applied.62 The first equation in (g19) will be multiplied by r and the quasi-mean operator defined by (g17) will be applied. Let the stress values on the upper/lower faces of the shell be denoted as σ and. R+h 2 1 h 2 ∂2w ∂2u r r ρ0 dr = −ρ 0 R 1 + . R+h 2 3 σ rθ r (r − R) dr = 0 . (g20) σ 1 1 HR (g21) R+h ∂ R h 2 h 2 r 2 σ rr + 2r σ dr = 1 + σ U − 1 − σ L . 2 G. In order to do that. some intermediary results are necessary. = ( r − R ) 2 dr = Μ θθ ∫ 3 R 2 ∂θ 2 H R − h R 2 ∂θ 2 (g26) In the same way. Σ rθ = −µx / R . ∫ R 3 R ∂t 2 ∂t 2 R−h (g29) . ϕ. (g23) R+h 2 R+h R+h ∂ σ rθ r 1 2 (r − R ) dr = r ( r − R ) σ rθ + µx ∫ (3r − 2R )dr ∫ R R−h ∂r HR R−h R−h . by Elementary computations show that U= L= σ σ ( R + h . it follows that σ rθ = 2µ ε rθ = −µx / r . ∫ ∂r rr rr rr H R R R−h (g22) In the case of the elastic shell. t ) ( R − h . t ) . Using (g8). σ rϕ . Quasi-moments. it follows that h 2 2µ ∂ 1 ∂w Μ θϕ = 3 2 ∂θ sin θ ∂ϕ . θ. θ. it follows 2µ ∂ 2 w h2 ( λ + 2µ )∆∗ w − . R and (g27) Μ ϕϕ = G. ϕ . respectively. 3 R 2 ∂θ 2 (g28) 1 H Also. the HOOKE’s law leads to Integrating by parts. ∫ R r R−h (g25) 2 R+h 2 h2 ∗ λ∆∗ w + 2µ ∂ w 1 λ∆ w + 2µ ∂ w . 1 H Similar relations are derived for Also.4) Integrals of the displacement vector.

∫ 2 2 3R 2 5R 4 ∂θ R 3 R ∂t ∂t R−h (g30) 1 H R+h 2 ∂2u ϕ r (r − R) h 4 1 ∂w 2 y h 2 ∂ 2 h 2 + + ρ0 dr = ρ 0 R 2 . Hence Let the first equation in (g19) be multiplied by ( ) 1 ∂ Μ θθ 1 ∂ Μ θϕ − + + R ∂θ sin θ ∂ϕ Σ rθ = ρ0 R and Μ θθ − Μ ϕϕ R tan θ + 1 1 + h / R )2 σ U + (1 − h / R )2 σ L rθ rθ ( 2 ∂2 ∂t 2 1 h 2 1 h 4 ∂w 2 x h 2 + + 5 R ∂θ 3 R 3 R (g33) 2 Μ θϕ 1 1 ∂ Μ θϕ 1 ∂ Μ ϕϕ 2 U 2 L − + Σ rϕ + R tan θ + 2 (1 + h / R ) σ rϕ + (1 − h / R ) σ rϕ sin θ ∂ϕ R ∂θ ∂ 2 1 h 2 1 h 4 1 ∂w 2y h 2 = ρ0 R + + 5 R sin θ ∂ϕ 3 R ∂t 2 3 R . (g34) G. (g31) ∫ 2 2 3R 2 5R 4 sin θ ∂ϕ R 3 R ∂t ∂t R−h G.6) Quasi-mean value of the shell density. ∂θ ∂ϕ (g37) For the elastic shell. (g32) 1 h 2 ∂ 2w R + (1 + h / R ) 2 σ U − (1 − h / R ) 2 σ L = −ρ 0 R 1 + rr rr H 3 R ∂t 2 The next two equations in (g19) are multiplied by r and the quasi-moment operator defined by (g18) is applied. r 2 . By applying the quasi-mean operator defined by (g17). (g38) . using (g15). the quasi-mean of a constant is equal to that constant. Applying the quasi-mean operator to the mass balance equation in the linear approximation gives ρ = ρ 0 (1 − tr ε ) . σ = λtr ε 1 + 2µ ε . (g35) It will be further assumed that the quasi-mean of the density is constant and equal to its initial value. The differential equation.5) Equations of motion in quasi-mean values.63 1 H and R+h 2 ∂2uθ ∂ 2 h 2 r (r − R) h 4 ∂w 2 x h 2 ρ0 dr = ρ 0 R 2 + + . From (g35) it follows that (g36) ∂ ∂y ( x sin θ ) + = 2w sin θ . tr ε = 0 . i.e. it follows R+h ∂ Σ rϕ 1 ∂ 1 2 − Σ + Σ − sin θ Σ rθ + ϕϕ RH ∫ ρr g( r )dr θθ sin θ ∂θ ∂ϕ R−h . According to (g17)..

G1.2.64 By using (g9). Σ θθ + Σ ϕϕ = 0 . having the approximate direction from North to South. let Θ be the angle between the normal vector to the parallel section ϕ → τ σ M 1MM 3 and the unit vector eθ . Let Φ be the angle between the normal vector to the meridian → section M 4 MM 2 and the unit vector e . using (g36).7) The buckling of a spherical shell. The centres of the lateral sides are denoted by M k . Also. having the approximate direction from North to South. having the approximate direction from West to East. (g40) The term (h / R )4 will be neglected. having the approximate direction from West to East.3. approximately along A correction due to the compressive horizontal stresses the θ − and ϕ − axes respectively. Consider the element ABCD of the mean surface of the deformed shell from Fig. where the local unit vectors are cross section → → → e r . it follows σ θθ + σ ϕϕ = 2(λ + µ )tr ε . ( ) σ σ G. and a tangential stress is acting a normal compressive stress WE . eϕ . k = 1. eθ . NS . The concentrated force acting at the point M1 is → → WE dϕ WE ∂ σ cos Φ − ∂Φ dϕ e + sin Φ − ∂Φ dϕ e Σ1 = σ − ∂ϕ 2 ∂ϕ 2 ϕ ∂ϕ 2 r (g42) → → NS ∂ dϕ ∂Θ dϕ ∂Θ dϕ ∂w dϕ + τ NS − τ cos Θ − Hdθ eθ + sin Θ − er × R − w + ∂ϕ 2 ∂ϕ 2 ∂ϕ 2 ∂ϕ 2 → . follows to be further considered. (g39) or. On the parallel cross section M MM and a tangential stress τ 1 3 NS .4 . Using (g40) and substituting Σ rθ and Σ rϕ between (g32)-(g34) it follows 2 2 ∂ 2 sin θ Μ θϕ 1 ∂ sin θ Μ θθ 2 1 ∂ Μ ϕϕ ∂ cos θ Μ + + − ϕϕ ∂θ∂ϕ ∂θ sin θ sin θ ∂ϕ 2 ∂θ 2 R2 sin θ 1 ∂ 2 U 2 L + (1 + h / R )2 σ U − (1 − h / R )2 σ L + rr rr 2R ∂θ sin θ (1 + h / R ) σ rθ + (1 − h / R ) σ rθ H (g41) R+h 1 ∂ 1 2 1 + h / R )2 σ U + (1 − h / R )2 σ L − + rϕ rϕ HR ∫ ρr gdr ( 2R ∂ϕ R−h 2 2 h2 2 ∗ ∂ w − 1 − h ∂ w ∆ = ρ 0 sin θ 2 R 2 ∂t ∂t 3 NS . It is centred at the point M. On the meridian M 4 MM 2 is acting a normal compressive stress σ WE . WE acting at the ends of the shell.

sin Φ ≅ Φ ≅ 1 ∂w R sin θ ∂ϕ . cos Θ ≅ 1 . (g46) It follows the deformed element of the shell is into an equilibrium state due to the action of the lateral stress and to the opposite − Σ p .4 = m σ ± ∂θ 2 cos Θ ± ∂θ 2 eθ + sin Θ ± ∂θ 2 e r → → WE dθ ∂Φ dθ ∂Φ dθ WE ± ∂ τ cos Φ ± + τ e ϕ + sin Φ ± e r ∂θ 2 ∂θ 2 ∂θ 2 ∂w dθ × R − w m sin(θ ± dθ )Hdϕ ∂θ 2 But → (g44) cos Φ ≅ 1 .G1. ϕ )( R − w ) 2 sin θdθdϕ → er . R ∂θ (g45) Let p = p(θ. sin Θ ≅ Θ ≅ 1 ∂w . The force due to that density is equal to Σp force → = p(θ. the concentrated force acting at the point M 3 is equal to → → WE dϕ ∂Φ dϕ ∂Φ dϕ WE ∂ σ Σ 3 = − σ + ∂ϕ 2 cos Φ + ∂ϕ 2 eϕ + sin Φ + ∂ϕ 2 e r (g43) → → NS ∂ dϕ ∂Θ dϕ ∂Θ dϕ ∂w dϕ + τ NS + τ cos Θ + eθ + sin Θ + er × R − w − Hdθ ∂ϕ 2 ∂ϕ 2 ∂ϕ 2 ∂ϕ 2 → The concentrated forces acting at the points M 2 and M 4 are respectively equal to → → NS dθ ∂Θ dθ ∂Θ dθ NS ∂ σ Σ 2.65 Fig. i. → Σ1 + Σ 2 + Σ 3 + Σ 4 − Σ p = Hence the next three equations of equilibrium are obtained: → → → → → → 0 .e. (g47) . A rectangular element of the mean surface of the deformed shell In the same way. ϕ ) a surface density of forces normal to the element of the shell. having the same mechanical effect as the presence of the compressive stress.

ϕ. θ. ϕ. The strain tensor inside the material is M and the strain rate here is The first boundary condition assumed on the lower face of the shell. t ) . (g48) (g49) ∂ WE sin θ ∂w + sin θ τ =0 . u θ ( r. like (∂w / ∂θ)2 have been neglected again. Quantities of the second order. θ. it follows ∂ ε M / ∂t .66 WE sin θ ∂w NS ∂w τ ∂w + 2 cos θ − σ R ∂θ ∂θ sin θ ∂ϕ WE WE ∂ NS ∂w τ ∂w ∂ σ ∂w NS ∂w = − p(θ. is the continuity of the displacement vector. having the value ∂ σ WE p(θ. eqs. For the material below the M M = p − ρ M g( R − h − w ) + λM tr M + 2µ M M + 2η M ∂ ε . Usually. ε u r ( r. the horizontal loads for the upper face of the shell are neglected. σ rθ = 0 U . the next constitutive equation is assumed . ϕ. having the equation r = R − h . µ M and the viscosity is denoted by η M . ϕ.(g48)-(g50) show that the presence of the For the particular case when σ τ τ lateral compressive stress is equivalent to a supplemental load placed on the upper face of the shell. (g8). ϕ. t ) − h ∂w + x( θ. G. ρ is the density of that material and 1 is the unit tensor. t ) = − w(θ.8) Load on the upper face. The elastic coefficients are λM . By using eqs. Stress on the lower surface of the shell. t ) Hence the next values for the strain tensor immediately below the shell are obtained (g54) . σ rϕ = 0 U . θ. and it is assumed that 2 U = −ρ F gw − P 1 − 1 ∂ w σ rr g ∂t 2 where ρ is the density of the filling sediments. ϕ ) = − H R2 2 2 NS ∂ w + WE 1 ∂ w . t ) = R sin θ ∂ϕ + y(θ. (g53) σ 0 ε 1 ε 1 ∂t M where p 0 is a reference pressure. (g52) F P is the load and an inertial term is considered. Consider now the differential equation (g41). shell. t ) = R ∂θ − h 1 ∂w u ϕ ( r. ϕ. ϕ ) R ( R − w ) sin θ + + + sin θ + H ∂θ σ ∂θ sin θ ∂ϕ ∂ϕ sin θ ∂ϕ τ ∂θ NS ∂ NS NS 2 cos θ − sin θ ∂w + sin θ ∂ σ + τ =0 . Hence the elements of the displacement vector inside the material placed immediately below the lower face of the shell are assumed to be equal to the same elements at the points of the shell placed on the lower face. σ ∂θ ∂ϕ R ∂θ and . Nowacki 1961). (g50) + τ WE 2 cos θ − ∂θ ∂ϕ R ∂θ WE is a constant and NS = WE = 0 . σ σ ∂θ 2 sin 2 θ ∂ϕ 2 (g51) a result similar to the case of the plane plate (Timoshenko and Woinowsky-Krieger 1959.

∂θ 2( R − h ) → ε rϕ = − 2R sin θ ∂ϕ − 2( R − h ) M 1 1 ∂w y . From Newton’s third low. it follows finally that H D∆∗ ∆∗ w + λM − µ M ∆∗ w + ρ M − ρ F − ρ∗ gw 2 1 ∂ 2 w WE 1 ∂2w + H σ NS + σ ∗ + σ + σ ∗ . on the lower surface of the shell.e. 2 2 R ∂θ 2 R sin 2 θ ∂ϕ 2 ∗ 2 2 ∗ + H η M ∂ ∆∗ w + 2 w − ρH + P ∂ w + ρ H 3 ∂ ( ∆∗ w ) =P 2 ∂t g ∂t 2 12 R2 ∂t 2 where (g61) . (g60) 2 ∂w H NS ∂ 2 w H h 2 M ∗ ∂w 1 ∂2w − + + WE 1 − η R∆ + ∂t R − h ∂t R 2 σ 2 σ 2 θ ∂ϕ 2 2R R sin ∂θ 2 H2 ∗ ∂2w h ∂2w − 1 − ∆ = ρ0 H 2 R 2 ∂t 12 ∂t 2 If quantities of the order ( h / R ) are neglected with respect to unit.9) The differential equation of time dependent flexure.(s41) and taking into account the presence of the lateral gstress. the next elements of the stress are obtained after some elementary computations σ rr = p 0 − ρ M g( R − h − w ) − R − h λM ∆∗ w L Rh . + −η ∂θ R − h R ∂θ ∂t R − h ∂t y 1 ∂ ∂w 1 ∂y . it follows a relation between the stress → → S− = − M . Substituting the loads on the upper/lower faces of the shell in eq. σ rϕ = −µ M R sin θ ∂ϕ + R − h − η M R sin θ ∂ϕ ∂t + R − h ∂t L 1 ∂w (g59) G. a generalisation of the Sophie GERMAIN plain plate static equation in the case of a time dependent flexure of a spherical elastic shell is obtained as ∆∗ w 1 ∂2w H3 h2 ( λ + 2µ )∆∗ ∆∗ w + 2µ − 1 + ρ F gw + P1 − 2 12 g ∂t 2 R R 2w h2 M ∗ Rh M ∗ H h 2 +1 − ρ M gw + λ ∆ w − 1 − µ R∆ w + 2R R − h R R−h R . (g55) The normal vector at the lower face of the shell is e r . σ er σ er (g56) σ S inside the shell and the stress σ M inside the material below the shell. i.67 ε rr M=0 . The reference pressure p 0 in (g53) is selected in order the flexure of the shell to vanish in the absence of the load. Here. (g57) (g58) σ rθ = −µ M R L and x 1 ∂w M 1 ∂ ∂w + 1 ∂x . ε rθ = − 2R M 1 ∂w x − .

. It can be concluded that in the usual cases. i. (g64) H ∂ P ∂2w ρ 3 ∂2 ∗ (∆ w) = P + η M ∆∗ w − ρH + + H 2 2 ∂t g ∂t 2 12 ∂t M − ρ F according to (g63a) and a substitution of the real lateral With respect to (g64). 2 ∂2w c ∂ w + ρ M − ρ F gw D∆∗ ∆∗ w + H σ c + σy x 2 ∂x ∂y 2 . H / R ∝1 / 100 . µ. For the fundamental mode ( m = n = 1 ) it x y x y 2w / R 2 1 . µ M ∝1011 Pa .g. In that case. It follows the LAPLACE operator of the flexure is proportional to ∆ ∗ w ∝ π 2 m 2 / L2 + n 2 / L2 sin( mπx / L )sin( nπy / L ) . For the usual materials. = ∆∗ w π 2 R 2 (1 / L2 + 1 / L2 ) x y follows that (g65) That ratio is negligible too in the usual cases. A supplemental Ivan 1997a) like is present For usual values (e. the flexure is proportional to a product of sins (cosines) functions.(s61). Here.68 3 (λ + 2µ )H 3 2H D= = ρa 12 12 is the flexural rigidity and a is the velocity of the P-wave through the shell. (g62) µM H ρ∗ = ρ M + ρ F + gR R . σ NS . the case of a rectangular plate having the sides equal to L x .e. ρ M . all these effects are usually negligible and 2 ∗ difficult to be observed in real life. λM = µ M .10) Spherical effects with respect to the plane plate. σ WE by their apparent values according to (g63c). σ WE ∝ 30 MPa . (g63) G. the sphericity of the crustal plates can be ignored and the equation derived for the flat plate can be used. Also. w ∝ sin( mπx / L x ) sin( nπy / L y ) . the corresponding equation for the plane plate (Ivan 1997) is. σ NS + σ . P∗ = P1 + R .σ ∗ WE + σ ∗ can be observed in eq. To compare 2w / R to ∆ w in the left side of (g61). ρ F ∝ 3000 kgs / m 3 . L y is considered. a change of the density difference ρ stresses load σ NS . 2 ∗= µ H σ 6 2 R H .

opening/closing of preexisting cracks) and/or to the energy lost by friction at the two sides of a crack or on the non-elastic boundary coupling grain particles to the adjacent material (Aki and Richards. The mechanical analogue of the first model to be considered is represented in Fig. Elementary computations show that the differential η K being . seconds up to minutes. The elastic part will be represented by a spring.H1. Ranalli. the HOOKE’s law has to be replaced by assuming different constitutive equations (models). Especially for geological processes at a large time scale and great values of the stress. a mechanical analogue can be considered.H1a. 1987. there are two kinds of constitutive equations (models). with a maximum value around one hour for the fundamental mode of the free oscillations of the Earth). (b) MAXWELL model. a linear relation σ = 2µ ε is valid for the spring and a similar linear relation holds for the dashpot • = 2η ε . σ 0 is applied. i. Consequently. In the beginning. while the inelastic (viscous) behaviour is associated to a dashpot. For each constitutive equation. (c) BURGERS model a) KELVIN-VOIGT (strong viscous) model. Here. More general examples follow to be presented in relation with the dynamic aspects of the flexure of a plate (shell) and to the accretion prism. only the 1-D case will be discussed. Wahr. In relation to seismic or seismological applications. for example. Simplified models involves a linear relation between stress (and its derivatives of various orders with respect to time) and strain (together with its time derivatives). Suppose now a constant stress equal to equation • = 2µ K ε + 2η K ε . material) derivative with respect to time. Here. the non-elasticity is related to the very short period irreversible changes in the crystal defect structures of the medium (e. H. (a) KELVIN-VOIGT model.1) Introduction.69 H) ELEMENTS OF RHEOLOGY H.2) Linear models. It follows that KELVIN-VOIGT model has the next constitutive equation σ the viscosity. while the total deformation is equal to both the deformation of the spring and the deformation of the dashpot. short periods and short stresses are required (usually. With respect to the mathematical relation between stress and strain. the internal friction of the material cannot be ignored.e. 1996). 1980. Their expressions are mainly depending on the time scale of the geophysical process to be modelled. the second term is the inelastic one. (h1) where the dot shows the (total. The total stress is the sum between the stress in the spring and the stress in the dashpot. Both parts are supposed to be linear ones.g. σ Fig.

H1c. the strain approaches a limiting value equal to σ0 ε ∞ = 2µ K .H1b. (h6) shows that for a constant stress (equal to unit here). while the total deformation is the sum between the deformation of the spring and the deformation of the dashpot.70 σ 0 = 2µ K ε with the initial condition has the solution + 2η K dε dt (h2) ε (t = t 0 ) = 0 ε = µK ( t − t 0 ) . • • + σ 2η M . denoted by J( t ) = 1 2µ K µK ( t − t 0 ) 1 − exp − ηK . It follows from (h8) that the 0 1 c) BURGERS (general linear) model. Consider the mechanical analogue represented in Fig. 1 − exp − 2µ K ηK (h3) σ0 (h4) for t ≥ t 0 . (h8) σ 0 is applied. Suppose now at a certain moment equal to ε1 . The third model to be considered has the mechanical analogue represented in Fig. for very great values of time. It follows now the corresponding solution decreases towards zero as ε t = t1 . The total stress is equal to both the stress of the spring and to the stress of the dashpot. Suppose now at a certain moment t = t1 . EXERCISE. ηK (h7) b) MAXWELL (viscous-elastic) model. the constant stress σ is removed. there is a temporal variation of the strain. the corresponding strain at that moment being equal to ε . the constant stress σ 0 is removed. µ2 η2 η2 µ 2 (h11) Consider now the same initial condition (h3) and suppose again a constant stress equal to to MAXWELL model. Suppose again a constant stress equal to the first term in the right hand of (h8). The spring is instantly deformed to a value equal corresponding to and the solution of (h8) (for a constant stress having the slope related to the stress σ 0 and to the viscosity of the dashpot. σ 0 ) with the initial condition (h9) is the straight line σ0 σ0 ε = 2η M (t − t 0 ) + 2µ M . σ0 ε 0 = 2µ M . It follows that MAXWELL model has the next constitutive equation σ ε = 2µ M with the initial condition represented by (h3). Show that the corresponding differential equation is 2η1 •• ε • η •• η • µ µ + 2µ1 ε = 1 σ + 1 + 1 + 1 σ + 1 σ . Hence. the corresponding strain at that moment being µ = ε1 exp − K ( t − t1 ) . (h5) The “flowage function”. the system is instantly deformed to a value equal to σ 0 is applied.e. (h9) (h10) strain remains constant. i. In a similar manner .

f ( t ) = 0 . A . a ≤ 0 in order to have f ( t ) ≤ M exp( at ) . B are fourth rank tensors. differentiating (h13) it follows (h14) Hence. Sokolnikoff and Redheffer. It follows from (h13) that the strain decreases exponentially towards zero. However. Q( D ) = D 2 η M 2µ M (h19) A common way to solve (h16) is by using the LAPLACE transform (e. Under the above conditions.. A .. providing that 1... two initial conditions are required to obtain the complete solution). Suppose now at a certain moment t = t1 . µ = ε1 exp − 1 ( t − t1 ) . the solution (h13) approaches asymptotically to a straight line having the slope equal to σ 0 / 2η 2 . (h16) where P( D ) = A 0 + A1 D + A 2 D 2 +. the linear relation between stress and strain can be written as P( D ) σ = Q( D ) ε .. Show that: (a) The LAPLACE transform of the first derivative is df L ( p ) = pL[f ] − f ( 0 + 0 ) dt (b) (h21) (h22) L[exp( − at )]( p ) = 1 p+a .e 1 ε d) Remarks on the linear models. For example. Show that two (or more) springs / dashpots connected in series (or parallel) sequence are equivalent to a single spring / dashpot. 0≤a . Here.g. f ( t ) is piecewise continuous on every finite interval.. In the most general case. η1 (h15) EXERCISE. (h13) where C is an unknown coefficient (because (h11) is a second order differential equation. for t < 0 . the constant stress σ 0 is removed. the corresponding strain at that moment being equal to ε .. B . B . it follows that P( D ) = 1 1 + D . with respect to the D= MAXWELL body having the constitutive equation (h8). applied to stress and strain dt 0 1 n 0 1 m respectively... for an arbitrary t . A . Consider a certain function of one real variable f ( t ) . σ ε 0 = 2µ 0 2 (h12) . 2. there are two constants 0 < M . for great values of time. the LAPLACE transform of f ( t ) is a new function of the variable p . i.+ Bm D m are formal polynomials of the variable (h18) d representing the derivative with respect to time. 3. 1958)..71 and the solution of (h11) (for a constant stress σ 0 ) with the initial condition (h12) is µ1 σ σ ε = 2µ 0 + 2η0 ( t − t 0 ) + C1 − exp − η1 ( t − t 0 ) 2 2 µ1 µ σ ε = 2η0 + C η1 exp − η1 ( t − t 0 ) 2 1 • .+ A n D n and (h17) Q( D ) = B0 + B1 D + B2 D 2 +. defined by ∞ L[f ]( p ) = ∫ f ( t ) exp( − pt )dt (h20) 0 EXERCISE.. Justify that the BURGERS model is the general linear model..

72 (c) The convolution theorem. Consider the functions f , g vanishing for negative values of their argument and a new function (also vanishing for negative values of the argument defined) by the convolution product

+∞ t h ( t ) = (f ∗ g)( t ) = ∫ f ( τ )g( t − τ )dτ = ∫ f ( τ )g( t − τ )dτ −∞ 0 L[f ∗ g] = L[f ] L[g]

(h23)

Show that

(h24)

As an example, consider again the MAXWELL model by applying the LAPLACE transform to both sides of (h8). It follows that

where

EXERCISE. By using the above properties of the LAPLACE transform, show that

σ

0=

σ

1 1 0 pLε − ε 0 = L pL σ − σ + 2 η M σ 2µ M ( 0 + 0) , ε 0 = ε ( 0 + 0) are the initial stress and strain respectively.

,

(h25)

σ

( t ) = 2µ M

ε

(t) −

2µ 2 t M ηM

∫ε

0

µ ( t ) exp − M ( t − τ )dτ + σ 0 − 2µ M ηM

ε

**0 exp − µ M t (h26) ηM
**

(h27)

If the initial conditions are elastically coupled, i.e.

it follows an integral representation of the stress which is independent on the initial conditions

σ

0 = 2µ

M

ε

0

µ ( t ) exp − M ( t − τ )dτ ηM

σ

( t ) = 2µ M

ε

(t) −

2µ 2 t M ηM

∫ε

(h28)

0

Hence, the actual value of the stress is related both to the actual value of the deformation and to the previous values of the strain (i.e. the stress depends on the “history” of the deformation). If a constant strain the MAXWELL body, it follows from (h28) that the stress decreases as

ε

( t ) = ε 0 , for t ≥ 0 is applied to

(h29)

σ

( t ) = 2µ M

ε

0 exp − µ M t , ηM

G(t ) = 2µ M exp(− t / τ M ) is the “relaxation” kernel and the parameter τ M = η M / µ M is the relaxation time.

representing a “relaxation phenomenon” (stress decreases in time if a constant deformation is present). Here, the function (h30)

A very similar approach (e.g. Wahr 1996) is based on the use of FOURIER transform (see Chapter F.13). Formally, the results derived by using FOURIER transform are derived from the same results obtained with LAPLACE transform by performing the substitution p = iu . The above 1-D models can be generalised for the 3-D case. For example, consider again the MAXWELL body. There is a strong experimental evidence that the MAXWELL Rheology applies only to the dissipation of the shear energy, i.e. the stress and strain tensors in (h8) are the deviatoric tensors

σ

is valid. By differentiating with respect to time

1 ↔ σ − tr σ 3

1

,

ε

,

1 ↔ ε − tr ε 3

1

(h31)

It will be further assumed that there is no dissipation of the compressional energy, i.e. a proportionality like

tr σ = (3λ M + 2µ M )tr ε

• • tr σ = (3λ M + 2µ M )tr ε

(h32)

(h33)

Substituting (h31) and (h33) in (h8) it follows

1 • 1 − tr ε1 = ε 3 2µ M

Hence

•

• 3λ + 2µ • M tr + 1 − M ε1 2η M σ 3

− 3λ M + 2µ M tr ε1 σ 3

(h34)

73

σ+

•

µM ηM

σ

= 2µ M

• µ 3λ M + 2µ M tr ε 1 + λ M tr ε 1 + M ε ηM 3

~ ~ Φ σ = λ tr Φ ε 1 + 2 µ Φ ε

•

(h35)

By applying the FOURIER transform to both sides of (h35), it follows , (h36)

i.e. a relation similar to HOOKE’s law is valid between the FOURIER (or LAPLACE) transforms of stress and strain. Here, the coefficients similar to LAMÉ parameters are

**µM 2 µM 1 + ~ ~ ηM 3 λM iu , µ = µM λ = λM µ µ iu + M iu + M ηM ηM At short periods T correspond high values of the pulsation u = 2π / T . From (h37) it follows ~ ~ λ = λM , µ = µM iu +
**

and the behaviour of the MAXWELL body is an elastic one. At long periods T correspond low values of the pulsation and

(h37)

(h38)

~ 3λ + 2µ M =K λ= M M 3

~ , µ=0

KM .

(h39)

and the MAXWELL body is a fluid having the compressional coefficient denoted by

**H.3) Non-linear models.
**

For tectonic applications, large stresses and periods of thousands to millions of years are appropriate. Here, the non-elastic behaviour is probably related to the diffusion or dislocation creep of the molecules, a major factor being the high temperatures. There are great difficulties to consider constitutive equations with non-linear relations between stress and strain, but some attempts have been made. A very common non-linear model is the work-hardening plasticity (e.g. Ranalli, 1994)

n −1 * * σ ij 3 σ E , ε ij = 2 σ0 σ0

(h40)

where

**σ ij is the component of the deviatoric stress, σ E = stress, and σ , n are material parameters. 0
**

*

*

3 * * is related to the second invariant of the deviatoric 2 σ ij σ ij

**H.4) Brittle. Creep. Empirical criteria.
**

The usual materials are reacting in an elastic manner only for small values of the (deviatoric) stress, i.e. for stress values smaller than a limiting value representing the yield strength (or the yield stress), denoted by

σ Y . The yield stress is a

function of the nature of the material, of the temperature, pressure, the chemical composition of the adjacent rocks and, finally, of the history of the deformation (i.e. the intermediate steps followed to attend the yield value). When the yield stress is attended, there are two possibilities of behaviour of the material: • a rupture deformation of the rock, when the continuity of the deformation is lost, usually along a fault surface; this is the case of the brittle materials. The process is illustrated in Fig.H2a and b. • a plastic, irreversibly flow of the material (creep), when, apparently, the continuity holds. The phenomenon is quite similar to the usual viscous flow, but it can be observed only when the yield stress is attended. This is the case of the ductile materials. The process is illustrated in Fig.H2c. An usual non-linear constitutive equation is the BYERLEE power-law creep

ε = Aσ

•

n exp( − H / RT) ,

(h41)

74 where A, n are material parameters, H is the activation enthalpy, R is the gas constant and T is the absolute temperature. It should be noted that the same material can act as a brittle or a ductile one according to the external conditions.

Fig.H2. (a) Faulting of a brittle material; (b) Increasing of the yield stress due to the history of deformation; (c) Creep of a ductile material.

**H.5) Empirical criteria for shear-faulting. TRESCA criterion. COULOMB-NAVIER criterion.
**

A first criterion also (valid for plasticity) is due to TRESCA. It assumes that faulting (for brittle materials) or creep (for ductile ones) is attended at that points of the material where the maximum value of the shear (tangential) stress is equal to a yield value denoted by

σ Y . Consider now a homogeneous (constant) stress field inside the material. Such a case can be σ1 , σ 2 , σ 3 of the stress tensor. It will be assumed that they are denoted in order

obtained either by considering an infinitesimal volume of material or by taking into account a prismatic body with very large (infinite) sides. Consider the eigen-values to have

σ 3 < σ 2 < σ1 , with the remark that ,in real life, stress is assumed to be positive for compression. Hence, with respect to Fig.H3, let σ = − σ , σ 11 1 22 = − σ 3 and σ12 = 0 in eqs. (d11) and (d13). It follows that σ + σ 3 − σ1 − σ 3 cos 2ψ = −σ rr = 1 σ 2 2 , (h42) σ1 − σ 3 sin 2ψ τ = − σ rθ = 2 where σ , τ are so-called “normal stress” and “tangential (shear) stress” respectively, acting on a plane inside the material.

The plane is at an angle

π with the direction of the maximum compressive stress. The outer-pointing normal at that 2 plane makes an angle θ with the direction of the maximum compressive stress. With respect to a σ − τ reference system, ψ=θ−

eqs.(h42) are the parametric equations of the MOHR circle (see Section D4), plotted in Fig. H3. In the most general case, stress field is varying from point to point inside the material. Hence both the eigen-vectors of the stress tensor (i.e. the local directions of the maximum / minimum compressive stress) and the eigen-values of that tensor (i.e. the magnitudes of the maximum / minimum compressive stress) are also changing from point to point. For a fixed point inside the material, both normal stress and shear (tangential) stress are varying with the angle between the plane (with respect to normal and tangential stresses are defined) and the direction of the local maximum compressive stress. Consider a certain point inside the material and imagine various planes passing through that point. Hence, according to TRESCA empirical criterion, failure of the material is produced here if

max τ

ψ

= σY

(h43)

Using eq.(h42b), it follows

max

ψ

σ1 − σ 3 sin 2ψ = σ1 − σ 3

2 2

max sin 2ψ = σ Y

ψ

(h44)

75 Consider now a homogeneous stressed material subject to progressive increasing values of the difference material is characterised by a yield value denoted by

• •

if

**(σ1 − σ 3) / 2 < σ Y , there is no failure inside the material;
**

ψ=±

σ Y . Eq.(h44) shows that: σ1 − σ 3 = 2 σ Y

σ1 − σ 3 . The

when the equality

(h45)

is attended, a failure is produced along the planes at angles

π with the direction of the maximum compressive stress. 4

Fig.H3.(a) The MOHR circle; (b) A plane inside the material, at an angle ψ with the direction of the maximum compressive stress; (c) TRESCA criterion. E.d and F.D. denote the elastic domain and the failure domain respectively; (d) COULOMB - NAVIER criterion. Eq.(h45) represents the TRESCA criterion in terms of the eigenvalues of the stress tensor. In the case of a material subject to a non-homogeneous stress field, eq.(h45) is a local condition. Here, the eigen-values are obtained (see Section A). For a compressive stress, that values are expected to be negative ones. They have to be denoted by

σ 3 < σ 2 < σ1 .

material is characterised by the cohesive strength denoted by

− σ1,− σ 2 ,− σ 3 , where

A second criterion is due to COULOMB and NAVIER. It can be used to describe only the shear fracture. According to it, a

S and by the coefficient of friction, denoted by µ * = tan φ .

Here, φ is the angle of internal friction ( φ = 30 in most rocks). According to COULOMB - NAVIER empirical criterion, shear failure of the material is produced at its points where

o

max ψ

Using eqs.(h42), it follows

τ

− µ*

σ = S

(h46)

H.76 σ − σ 3 sin 2ψ − tan φ σ1 + σ 3 − σ1 − σ 3 cos 2ψ = S 1 max 2 2 2 ψ If (h47) ψ will be limited to the first quadrant. Hence an equation like f (I1 . without loss of generality. . eq. Hence. a share fracture is produced along the planes at angles ψ = ± − with the direction of the maximum 4 2 compressive stress.(a22)).(h52) is (h52) (σ1 − σ 2)2 + (σ 2 − σ3)2 + (σ3 − σ1)2 = 6k 2 (h53) Hence the criterion of Von MISES .HENCKY assumes that ductile flow occurs at those points of the material where 2 2 (σ11 − σ 22)2 + (σ 22 − σ 33)2 + (σ 33 − σ11)2 + 6 σ12 + σ 2 + σ13 = 6k 2 23 In terms of the principal stresses. Taking into account that when the equality σ1 − σ 3 µ * = tan φ . Eq. I 2 .NAVIER criterion is a generalisation of TRESCA criterion for a non-zero internal friction. eq. It follows the function f in (h51) depends only on the second invariant of the stress deviator (see eq.NAVIER criterion in terms of the eigenvalues of the stress tensor. Using eq. by taking into account the presence of the intermediate stress. I 3 ) = 0 .HENCKY can be regarded too as a generalisation of TRESCA criterion. π − ψ (or just −ψ ) is a solution too. the criterion of Von MISES .(h47) is ψ is a solution of (h47). eq. where eq.(h49) can be written as 2 2 µ * + 1 − µ * − σ µ * + 1 + µ * = 2S σ1 3 .(h49) represents the COULOMB . depending on the nature of the material.6) Von MISES-HENCKY criterion for ductile flow (plasticity).(a24).(a23)). being also similar for compressive and tensile states of stress. it depends only on the invariants of the stress tensor (Section A. the values of the angle σ1 − σ 3 2 Eq. (h50) outlining that COULOMB . (h51) will be valid. ductile flow occurs only at those points of material where the second invariant of the deviator stress reaches a certain value. there is no failure of the material.(h48) shows that: • • 1 cos φ σ σ max [sin(2ψ + φ)] = S + tan φ 1 2 3 + ψ (h48) if σ1 − σ 3 2 1 σ + σ 3 < S . − tan φ 1 2 cos φ 1 σ + σ3 = S (h49) − tan φ 1 2 2 cos φ π φ is attended. There are strong experimental evidence that the plastic flow does not depend on the hydrostatic pressure. Because the ductile (plastic) flow is independent of the co-ordinate system used. Hence.

d ε* * . Similar to the SAINT-VENANT body. (b) The BINGHAM body. * * f ( I 2 . (h4b). Its constitutive equation is σ * = 2µ B ε * . the body exhibits a pure plasticity. (a) The SAINT-VENANT body. b) BINGHAM body (visco-plastic material). When the yield stress is attended. Its constitutive equation (using deviator tensors) has the symbolic form σ * = 2µ S ε * . σ < σY σ = σY (h54) The above material has the mechanical analogue presented in Fig. Fig. σ = σ Y + 2 η B dt σ < σY σ ≥ σY (h55) The above material has the mechanical analogue presented in Fig. The strain rate is proportional to the difference between the deviatoric stress and the yield strength. being referred as a SAINT-VENANT body. The behaviour of that material is characterised by linear elasticity for stress values below the yield strength. (h4a). I 3 ) = σ Y . .. being referred as a BINGHAM body.H4. that material exhibits linear elasticity for stress values lower than the yield strength.77 H.7) Rheological models. but flows linearly above that value. a) SAINT-VENANT body (elastic-plastic material).

substituting (i5) into (i2)-(i3) gives dψ 2k ∂f 2k ∂r = r cos 2ψ dθ − r cos 2ψ + ρg sin θ ∂f = −2k sin 2ψ dψ + 2k sin 2ψ + ρgr cos θ ∂θ dθ (i6) . the stress in polar co-ordinates is 0 σ 0 ν σ rr + σ θθ Because the material is assumed to be incompressible. Taking into account that the 2-D case is discussed. However. the POISSON coefficient is ν = 1 / 2 . θ ) = 1 ( ) and ψ = ψ ( r. The wedge rests on a rigid basement having the slope equal to θ 0 . (i4) Taking into account again that the compressive stress is assumed to have positive sign. it will be assumed that ψ = ψ (θ ) only. Fig. σ rr = σ rθ 0 σ rθ σ θθ 0 . Yield condition. It follows to obtain a condition relating the slopes of the topography and that of the basement to the geometry of the wedge. (i1) ( ) Using polar co-ordinates. θ ) is the angle between the radius and 2 tr σ the local direction of the maximum compressive stress (variable inside the wedge).I1). Stress field. Both the compressional force acting on the left side of the wedge and the friction to the basement cause thickening of the incompressible material and the development of a topographical slope equal to α . Consider a 2-D prismatic body having a triangular vertical section (Fig. It is assumed that the material is into a state of plastic yielding according to the VON MISES-HENCKY criterion. θ ) − k cos 2ψ rr σ θθ = − f ( r .I1. After some manipulations. The 2-D accretion wedge. in the presence of gravity.(d11)-(d13) give (i5) where the trace of the stress has been denoted by f ( r. its yield strength and the friction coefficient to the basement. θ ) + k cos 2ψ σ rθ = k sin 2ψ . I1) The model.78 I) THE ACCRETION WEDGE. I2) Equations of equilibrium. the equilibrium equation (d40)-(d41) in the presence of gravity are ∂ σ rr ∂r ∂ σ rθ ∂r + ∂ σ rθ + r∂θ ∂ σ θθ r∂θ + σ rr − σ θθ + ρg sin θ = 0 σ rθ + ρg cos θ = 0 r r (i2) +2 (i3) Using (i1). the yield condition (r54) is σ1 − σ 3 = 2 k σ = − f ( r . eqs.

(i11) Using (i8).e.e. Final results. Hence . (i7) dψ C =1+ dθ cos 2ψ ∂f 2 Ck = + ρg sin θ r ∂r . it follows that sin 2ψ dg = −2Ck dψ cos 2ψ + C g = Ck ln( C + cos 2ψ ) + A . (i12) i. it will be assumed for all angles θ that σ θθ = −ρgr tan α . where the topography is generated due to the horizontal compression. The outward pointing normal vector is topography. eq. ∂ σ θθ ∂r = −ρg tan α . . eq. After some elementary manipulations. where the point C is very → → − e θ = ρgr tan α e θ σ i. Here. the radius r is a mean value of the lengths OA and OC. Hence the final stress inside the wedge is σ = −2 Ck ln r − Ck ln( C + co2ψ ) − ρgr sin θ − A − k cos 2ψ rr σ θθ = −2 Ck ln r − Ck ln( C + co2ψ ) − ρgr sin θ − A + k cos 2ψ = k sin 2ψ σ rθ I3) Boundary conditions. Substituting (i8) into (i6a) gives . is acting the lithostatic pressure due to the OC ≤ r ≤ OA . To find the unknown function f = 2Ck ln r + ρgr sin θ + g(θ ) g = g(θ ) .79 Eqs.e. all the next derivations are supposed to be valid at the rear of the wedge. the point C being closed to A. By differentiating (i14b). (i13) where A is another constant of integration.(i10) is substituted into (i6b) to obtain dg = −2Ck tan 2ψ dθ . Consider the segment AC placed on the side OA of the wedge. (i19) . having θ = 0 and closed to the point A. (i14) → → n = − e θ . (i9) i. r respectively. cos 2ψ dθ dθ i. the results of the differentiation being equal each other.e. it follows that dψ d − cos 2ψ = 0 . σ θθ (θ = 0) = −ρgr tan α Because the angle θ 0 has very small values.e. . i. (i10) . (i8) where C is a constant of integration.(i18) leads to 2Ck = ρg tan α r where .(i6a) and (i6b) are differentiated with respect to θ . (i16) (i17) or (i18) However. (i15) σ rθ (θ = 0) = 0 .

(i21) σ Consider now the side OB of the wedge.(i25) in real cases are presented by Ranalli (1987). eqs. (i16a) and (i20) are used to give ρgh 0 tan α + 2 kθ 0 ≅ λk . Eqs. → → θ = θ 0 and the outward pointing normal vector n = e θ .(i5c).80 i. (i20) σ rθ (θ = θ 0 ) = λk ∂ σ rθ ≈ .e. Hence ψ ≅ θ . It follows cos 2ψ ≅ 1 . both angles having small values. (i25) showing that the friction force (resistance to sliding of the wedge onto the basement) is balanced by tqo forces. (i22) 1 σ rθ (θ = θ 0 ) − σ rθ (θ = 0) λk = . The next partial derivative follows to be evaluated in two ways. (i19) and (i20) give ∂ σ rθ 2k 2kθ 0 dψ 2k 2Ck 2k = cos 2ψ = ( C + cos 2ψ ) ≈ + = ρg tan α + (i24) r∂θ r dθ r r r h0 From (i23) and (i24) it follows . Hence → → e θ = λk e r . In the first approach. assumed to have the magnitude equal to λk . is acting the friction force due to the basement. (i23) r∂θ r θ0 − 0 h0 The dominant stress is into the wedge is the horizontal compression.(i22). The first one is due to the topography and the second force is related to the compressive stress and to the slope of the basement. (i8). Further details related to the application of eq. having h r≅ 0 θ0 . Here. where λ is a friction coefficient.

R. 1959. 1996. Woinowsky-Krieger. Inc. Elsevier. Mc-Graw-Hill. 1958. J. Samizdat Press . v. Samizdat Press (http://samizdat. Gradstein. B.G. Inc. New York Wahr. K. 1987. 1992. Continuum Mechanics.mines. A finite element algorithm for computing the flexure of the crustal. Bucharest Ivan. R... I. G. 1961. New York Backus. N. 1997a.. J.. Seismology (in Romanian). Boston Rîjic. Mathematics of Physics and Modern Engineering. M... Studii si cercetari de GEOFIZICA 35 Ivan. 1986.A Geophysical Approach.. S..M. Arkadi. New York Turcotte. New York Timoshenko.. Tables of integrals.M.tehnicã.. Allen & Uniwin Inc.S.. Fundamental of Rock Mechanics.tehnicã. J. B. I. Revue Roumaine de Géophysique 41 Jaeger. S. Schubert. London Kennett.. Rheology of the Earth. 1982. 1961..Y.. N. Ed. D. Theory of elastic stability. Geodynamics. New York Timoshenko.P. John Wiley and Sons. Mechanics in Structural Geology..I.edu) Bayly. M. Bucharest Sokolnikoff. Methuen and Co. Springer .C.. McGraw-Hill Book Co.. Cook.M. G. 1980. Theory of Elasticity. 1955. -in Romanian. Gravity and Geodesy . Redheffer.. Richards. Continuum Mechanics. 1969.Y. G. S. Mechanics of tectonic faulting..Universitãtii Bucuresti Ivan.. Samizdat Press (http://samizdat. McGraw-Hill. Nowacki. Dynamika budowli (in Polish). 1997b. 1952.mines. Theory of plates and shells.. series and products. Quantitative Seismology. Gere.G.W. G. N. in press.M...Class Notes. Ed. Amsterdam Meissner.S. M. 1988.80 General references Aki. Warszawa Ranalli. sums. The Continental Crust . 1996. Academic Press. M.Verlag. Ed.L.edu) Mandl. Aperiodic motion of a crustal microplate... Filonenco-Borodici.. P. W. Freeman and Co... I. in press.

- 1.Strain26-1
- 01slides Outline
- Planilla de Tensado
- Curvature Ductility of RC Sections Based on Eurocode
- sdarticle12
- ME 7210 Syllabus
- A Two-dimensional Approach for Designing Tunnel Support
- geologia
- Solutions
- Lecture7 PressureVessel Combined Loading
- materials-06-04847-v2
- Analysis of Hyper Elastic Materials With MECHANICA
- Map3D
- Lab 1 - Compression of a Helical Spring
- Deshmukh, Mechanics of Materials
- 1-s2.0-S0022509610000712-main
- numerical-examples_2.pdf
- Structural Analysis JWFILES
- Property of Inertia
- Bonded Versus Unbonded Strip Fiber Reinforced Elastomeric Isolators Finite Element Analysis
- Shell Coupled FSI _example
- RÓTULAS PLASTICAS
- aph539
- precast foundations seizmic
- Experiment # 03
- Wang 2009
- Structural_Analysis [UandiStar.org]
- math.pdf
- bab 2 Strain.pdf

Skip carousel

- Metrology Workbench
- tmpCA93.tmp
- tmpD7EE
- Stress Induced in Two Nodal Fixed System is Directly Proportional to Nodal Displacement only
- tmpD855
- tmp4A0E
- tmp878F
- tmpDFEC
- tmp8550.tmp
- tmpA837
- tmpD900.tmp
- tmp3BB4.tmp
- Tmp 6880
- As 1391-2007 Metallic Materials - Tensile Testing at Ambient Temperature
- tmp546E.tmp
- tmpBD9C.tmp
- tmp2BB9
- tmpD49E.tmp
- tmpB917
- tmp1AE7.tmp
- tmp55E6.tmp
- Tmp 7645
- tmp6D71.tmp
- tmpB7DF
- Tmp 3831
- Tmp 8214
- tmpED9
- tmpC8B0
- tmpF7D
- tmp439

- MEHANIKATLA-Ispitna pitanja
- hronika_sreza_levackog
- Uvod u Python
- C++_how_to
- The Passive Voice
- DP-Man
- Savetnik Za Vegetarijance
- Ford Fiesta Mk5 Facelift - Restajling
- TM-Man
- Čista i Jednostavna Prirodna Kontrola Težine
- Srbija. Zemlja i Stanovnistvo Od Rimskog Doba Do Kraja XIX Veka - KANITZ Felix
- Osnove Python Jezika
- Čista i Jednostavna Prirodna Kontrola Težine
- C++_how_to
- Tcl_tutorial
- FKM Analytical Strength Assessment 5th Edition
- Nelinearna Analiza Konstrukcija
- Koral_Uputstvo
- hronika_sreza_levackog
- Deformacijska Svojstva Tla
- Arnold Ehret - Postom Do Zdravlja
- Three Dimensional Static and Dynamic Analysis of Structures
- FKM Analytical Strength Assessment 5th Edition
- User Man
- Phase2_SlopeStabilityVerification_Part1
- BM-Man
- Theory and Probems of Engineering Mechanics - Statics and Dinamics - E.W. Nelson, C.L. Best & W.G. McLean - 1st Ed -1997 - (Schaum Outline-McGraw Hill)

Sign up to vote on this title

UsefulNot usefulClose Dialog## Are you sure?

This action might not be possible to undo. Are you sure you want to continue?

Close Dialog## This title now requires a credit

Use one of your book credits to continue reading from where you left off, or restart the preview.

Loading