/conversion/tmp/scratch/383337191.

doc

First Order Concerns

"Scalar properties"
Definitions
Weight: wi
Volume: vi
Volume fractions: Vf Vm Vvoid ... or Vi
Weight fraction: Wi
Density of composite: r c
Burnout
n
r c =�r iVi
i =1

1
rc = n

�W / r i i
i =1

wi =r i / r cVi
rc
Vi = wi
ri

Some "General" Equations

Note: There are lots of ways to express these relationships. (For example, two forms
are given for r c .
/conversion/tmp/scratch/383337191.doc

Micromechanics and Macromechanics

Continuous fiber composites

Goal: To determine the composite

behavior from the properties of the
constituents. (i.e. to determine the
"effective" or "homogenized"
properties) Typical requirement:
periodicity of the material structure.
(unit cell) Examples

How are effective properties equivalent to the original heterogeneous configuration?

Consider the following comparisons
 Simulated experiment of composite vs. equivalent material
 Volume averaged stress vs. volume averaged strain of composite vs. that for
equivalent material
 Strain energy of composite vs. equivalent material

What are the constituents?

Micromechanics: constituents = fibers + matrix

Result = effective engineering properties

Macromechanics: constituents = lamina of different types and orientations

Result varies: effective engineering properties, ABD matrices
/conversion/tmp/scratch/383337191.doc
/conversion/tmp/scratch/383337191.doc

Approximations: geometric and otherwise

What is the source of error

for the fiber-direction
effective modulus?

What are the sources of

error for the effective
modulus perpendicular to
the fiber direction?

What stresses are assumed to exist?

What variations of stresses are assumed?
…Ditto for strains

Approximate analyses for determining effective properties violate reality in one or more
ways:
 Distribution of constituents
 Shape of constituents
 Compatibility of displacements
 Equilibrium
 Dimensionality (e.g. 1D or 2D analysis)
/conversion/tmp/scratch/383337191.doc

Strength of Materials Micromechanics

 Assume x1 = “fiber direction”

Property Assumptions Idealized model

E11 * Matrix and fibers loaded in parallel

* Same axial strain
* Determine average stress

E 22 * Matrix and fiber loaded in series

* Same axial stress
* Determine average strain

v12 * Matrix and fiber loaded in parallel

* Same axial strain
* Determine v12 from net contraction
 Because the constitutive matrix is symmetric,

v12 v E
 21  v 21  22 v12
E11 E 22 E11

G12 * Matrix and fiber loaded in series

* Same shear stress
* Determine average strain

* In each case we assume only one  and  exist.

* In each case we assume either constant  or . Also, within a single constituent both
the stress and strain are constant.

Notation for solutions

 superscript “c”  composite value … either same for all constituents or is the
“effective” value
 subscript f  fiber value
 subscript m  matrix value
 Vf, Vm = volume fractions of fiber and matrix, respectively
/conversion/tmp/scratch/383337191.doc

C
Effective Longitudinal Modulus, E11

c
 f   m  

 f   m
b
P
E f af
Em am

P  f da bi  ba bg
f m m

 f  Ef c

 m  E m c
P
c 
da b  a bi
f m

Use these in definition

c c
E11 
c
c
Result: E  V f E f  Vm E m

General: E   Vi E i
c
/conversion/tmp/scratch/383337191.doc

C
Effective Transverse Modulus, E 22

 f  m

 f m c
E f a f
E m a m
b

d i
 c a f  am   f a f   mam

c c
Also,  f  , m 
Ef Em

Combine these with definition

c
Ec 
c
Result
1
Ec 
1 1
Vf  Vm
Ef Em

General
1
Ec 
V
 Ei
i
/conversion/tmp/scratch/383337191.doc

C
Effective Shear Modulus, G12

x2

P b g b g
12 f 12 m

a G
x1 b g b g 
12 f 12 m
C

f f
a m G m
b
x3
L

C
 12 d
a f  a m   12 i b ga  b gaf f 12 m m

Since for this 1-D analysis

 12 
u1 F
Gu IJ
Hx K
2
0
x 2 1

b g G
12 f
C

b g G
12 m
C

Combine these with the definition

C
GC  C
 12
Result:
1
GC 
1 1
Vf  Vm
Gf Gm

 Same form as transverse modulus

 not considered very accurate
/conversion/tmp/scratch/383337191.doc

C
v12

X2

f af X1
m am

Load in x1 direction
 22c
v12C  -
 11c
11 f  11m  11
 22 f  - v12f 11
 22 m  - v12m 11
Du2  22 f a f   22 m am
 22
C
 
a f  am a f  am
  22 f V f   22 mVm
 22
C
 - v12f 11V f - v12m 11Vm

- C22
 C
v12   v12f V f  v12
m
Vm  rule of mixtures
 11
C
C E 22
and v 21  C
v12  consequence of symmetry of constitutive matrix
E11
C
Challenge: Calculate v21 directly using SOM.
/conversion/tmp/scratch/383337191.doc

Some Simple Formulas for Transport Properties

Transport Properties
p.105
 heat conduction
 electrical conduction
 moisture diffusion
 transport of electrical and magnetic fields

Estimate
Longitudinal: k L  V f k f  Vm k m (i.e. rule of mixtures)
Transverse: Use Halpin-Tsai equations
kf
-1
k T 1   V f km
 ;n
km 1 - V f kf

km
a
and log   3 log
b
 natural log?

Moisture Absorption (mass transport … diffusion)

p.106
 Fickian … like heat conduction
 Non-Fickian
 Text gives solution for 1-D Fickian (I think).
 Some “useful” equations for estimating moisture content of plate vs. time (how long
for saturation)
 We will return to this if time permits.
/conversion/tmp/scratch/383337191.doc

Thermal Expansion

p. 101 Schapery

L 
1
EL
d i
 f E f V f   m E mVm (3.67)

d i
 T  1  f  f Vf b
1  g
 V m m m (3.68)

- L LT

For V f  25%

b g
 T   f V f  1   m  mVm

We may return to derivation later.

How would you calculate thermal expansion coefficients using finite elements?
/conversion/tmp/scratch/383337191.doc

Moisture Induced Expansion

p. 104
Usually assume  L  0 (since fibers prevent expansion)

T 
C
m
b g
1m m


densities
Compare with thermal expansion equation
b g
 T   f V f  1   m  mVm

assume  f  0

b g
  T  1   m  mVm