# Strain, Strain Rate, Stress-2

date post

23-Feb-2018Category

## Documents

view

223download

0

Embed Size (px)

### Transcript of Strain, Strain Rate, Stress-2

7/24/2019 Strain, Strain Rate, Stress-2

1/26

RHEOLOGICAL MODELING SHORT COURSEApril 10-11 2007, Napoli

Strains, strain rate, stressesG. MarrucciUniversity of Naples, Italy

Broad definitions

The strain measures the change in shape of a small material element. Ratio of lengths,e.g. before and after the deformation, is used. Hence strains are nondimensionalquantities.

The strain rate is the strain achieved per unit time, and is particularly important in fluidmaterials. Dimension is inverse time, typically s1. The strain rate is related to the velocitygradient of the flow.

The stress is related to the force per unit area that a small material element exerts bycontact on its surroundings. A simple example is the pressure in a stagnant liquid, wheresuch a force is the same in all directions (isotropy). More complex is the situation arisingwhen the material is deformed, and the stress becomes anisotropic. Typical unit is Pa(Pascal = Newton/m2).

7/24/2019 Strain, Strain Rate, Stress-2

2/26

7/24/2019 Strain, Strain Rate, Stress-2

3/26

3/17RHEOLOGICAL MODELING SHORT COURSEApril 10-11 2007, Napoli

Deformation tensor

If we deform in an arbitrary way a small volume of material, we may considerthe relationship between vector ro connecting any two material points P and Q

before, and vector r linking the same material points after, the deformation.

P P

Q Qr

o r

Because of the smallness of the material element, the relationship between ro andr is linear. A general linear relationship between vectors is described by a tensor.We write:

r = E ro (1)

where E is the deformation tensor, fully describing how the undeformed element(to the left in the figure) changes its shape (to the right). It is important to notethat linearity arises from the smallness of the element, not of the deformation.

Arbitrarily large deformations can be considered, as are often encountered inrubber and rubber-like materials.

7/24/2019 Strain, Strain Rate, Stress-2

4/26

4/17RHEOLOGICAL MODELING SHORT COURSEApril 10-11 2007, Napoli

Matrix of deformation tensor

Let us use a Cartesian coordinate system. If xo, yo, zo are coordinates of amaterial point of the body before deformation, and x, y, z those of thesame point after deformation, the three following functions:

x(xo,yo,zo), y(xo,yo,zo), z(xo,yo,zo)

fully describe body deformation.

At any point in the body, we may consider the 9 partial derivatives ofthose functions, that we organize in the following matrix 3 x 3:

This is the matrix of tensor E defined in (1), also called deformation

gradient. The determinant of the matrix in (2) gives the volume ratio due to

deformation. Hence, if volume preserving deformations are considered,the determinant must be unity.

000

000

000

zzyzxz

zyyyxy

zxyxxx

E

= (2)

7/24/2019 Strain, Strain Rate, Stress-2

5/26

5/17RHEOLOGICAL MODELING SHORT COURSEApril 10-11 2007, Napoli

Example 1: shear deformation

This deformation is described by the following (linear) functions:

generating the matrix:

Notice that the determinant is unity.

0

0

00

zz

yy

yxx

=

=

+=

y

x

100

010

01

=E

(3)

(4)

7/24/2019 Strain, Strain Rate, Stress-2

6/26

6/17RHEOLOGICAL MODELING SHORT COURSEApril 10-11 2007, Napoli

Example 2: Uniaxial elongation

Linear functions for this deformation are:

where the stretch ratio is larger than unity. The deformation matrix is:

Also here the determinant is unity.

0

0

0

zzyy

xx

==

=

x

y

100

010

00

=E (5)

7/24/2019 Strain, Strain Rate, Stress-2

7/26

7/17RHEOLOGICAL MODELING SHORT COURSEApril 10-11 2007, Napoli

Elongational deformations in general

All elongational deformations are described by a symmetric tensor E. Then, by asuitable choice of Cartesian coordinates, the matrix of E has the canonical form:

where 123 = 1 for volume preservation. Hence, some of the s will be larger, andothers smaller, than unity.

If two of the s are equal, the deformation is called uniaxial, like the uniaxial stretch ofthe previous example. Similarly, one can have a uniaxial compression. In both cases,there is an axis of symmetry.

If one is unity, the deformation is called planar, since in one direction there is no

deformation.

In the general case, the deformation is called biaxial, because one has to assign thevalues of in two directions, the third being determined by volume conservation.

3

2

1

00

00

00

=E(6)

7/24/2019 Strain, Strain Rate, Stress-2

8/26

8/17RHEOLOGICAL MODELING SHORT COURSEApril 10-11 2007, Napoli

More on Eqs. (1) and (2)

In terms of Cartesian components, Eq. (1) defining tensor E can be interpreted asfollows. Imagine that the 3 components of vector ro are written as a column matrix:

Multiplication row by column of the matrix in Eq. (2) times the matrix of vector rothen gives the matrix of vector r.

If vector ro is taken from the origin, and hence its components coincide with thecoordinates xo, yo, zo of the vector tip, one can readily verify that multiplication rowby column of, for example, the shear deformation matrix of Eq. (4) times the matrixof ro indeed gives the components of vector r as reported in Eq. (3).

For future reference, it is worth noting that linear operators on vectors (i.e. tensors)

can be combined to operate in series, i.e. one after the other. For example, twoconsecutive deformations E1 and E2 combine in the single deformation E given by

E = E2 E1

where the matrix multiplication of E2 to E1 is rows by columns.

z

y

x

o

o

o

o

r

r

r

r =

7/24/2019 Strain, Strain Rate, Stress-2

9/26

9/17RHEOLOGICAL MODELING SHORT COURSEApril 10-11 2007, Napoli

Rotations

Eq. (1) defining tensor E also includes particular deformations which are infact rigid rotations of the element.

However, we can easily recognize a tensor representing a rigid rotation.

Indeed, suppose tensor Ris a rigid rotation. We then consider its transposeRT. (The transpose of a tensor has a matrix where rows and columns havebeen interchanged.)

Next we calculate the product R RT. If, and only if, Ris orthogonal, thensuch a product generates the unit tensor. Tensors representing rigidrotations are orthogonal.

For example, let us consider tensor R

representing a rigid rotation by an angle

around the z axis. Its matrix is:

The matrix of the transpose is:

One can readily verify that multiplying these two matrices rows by columnsgives the unit matrix.

1000cossin

0sincos

=

T

R

100

0cossin

0sincos

=R

7/24/2019 Strain, Strain Rate, Stress-2

10/26

10/17RHEOLOGICAL MODELING SHORT COURSEApril 10-11 2007, Napoli

Why considering rigid rotations

Generally deformations include both deformations proper, called puredeformations, and rigid rotations.

Tensors representing pure deformations are symmetric, i.e. E = ET forthem.

Elongational deformations previously encountered are pure deformations,and it is immediately verified that the matrix of E is indeed symmetric.

Conversely, for a shear deformation E is not symmetric, which implies thatthe deformation is not pure, and also incorporates a rotation.

The fact that, generally, deformations also include rotations is unfortunatebecause we expect that the stress arising in materials is due to puredeformations, not to rigid rotations. In other words, we cannot expect tolink the stress to tensor E as such, because E generally contains a rotationas well.

Fortunately, there exists a theorem stating that any tensor E can be splitin a single way into the product of a symmetric positive-definite tensor

(called U) to an orthogonal one R, i.e.:

E = U R (7)

7/24/2019 Strain, Strain Rate, Stress-2

11/26

11/17RHEOLOGICAL MODELING SHORT COURSEApril 10-11 2007, Napoli

E is split into rotation times pure deformation

The meaning of Eq. (7) is illustrated, for example, by a shear deformation E

which can be split into a rigid rotation Rfollowed by a pure deformation U:

As said previously, the pure deformation U is a symmetric positive-definitetensor, positive-definite meaning that all 3 principal values (or eigen-values) of the tensor are positive numbers (here, the 3 principal stretchratios).

E

R U

7/24/2019 Strain, Strain Rate, Stress-2

12/26

12/17RHEOLOGICAL MODELING SHORT COURSEApril 10-11 2007, Napoli

A simple way to eliminate rotation

As said before, we want to link stress to pure deformation. Hence we needto eliminate the rigid rotation Rfrom tensor E.

A simple way to obtain such a result is to operate the product:

B = E

E

T (8)

Indeed, in view of Eq. (7), and since the transpose of a product is madeby transposing the factors in reverse order, we obtain:

B = (U R) (U R)T = U R RT UT = U UT = U U = U2

In this calculation we have used the property that an orthogonal tensor Rtimes its transpose gives the unit tensor (disappearing from the product).We have also used the symmetry of the pure deformation tensor U,

whereby the product with the transpose is equivalent t

Recommended

*View more*