Chapter 7

CONSTITUTIVE EQUATIONS

Outline

§ 7-1 Count equations and unknowns

§ 7-2 Boundary conditions

1. Generalized Newtonian models: used to describe steady-state shear flows and have been widely used by
engineers for designing flow systems.
2. Linear viscoelastic models: used to describe unsteady-state flows in systems with very small
displacement gradients and have been used by chemists interested in
understanding polymer structure.
3. Non-linear viscoelastic models: used to describe all types of flow.

7-1

7-1 Constitutive equations-for non-Newtonian fluids

7-1-1 Count Equations and Unknowns

Equations Unknowns (body force is known)

Compressible Incompressible
1 Continuity Equation 1 Density 0
3 Equation of Motion

3 Components of v 3
1 Pressure 1
6 Components of  6
4 Total Equations 11 Unknowns 10

The additional equations, which are needed, are constitutive equations.


Rheological Constitutive Equation Stress Tensor = f(v)
Thermodynamic Constitutive Equation P=P(,T) for compressible fluid

7-2
7-2 Boundary conditions

7-2-1 Reynold’s Transport Theorem with Phase Interfaces

Consider V(t) to be a material volume (contains the same particles):

S+ Phase interface
+  (Singular surface)
n*
V+

 S*
U* V- S-

-
where V  t   V  V
S  t   S  S 
U*  velocity of this singular surface.

n *  unit normal vector of S* point from - phase to  phase.
7-3

The phase interface is a singular surface- a surface that is discontinuous w.r.t

quantities like density or velocity.
Since mass can be transferred across a phase interface, V+ and V- are not
material volume. Hence we can not talk about
D
  dV
Dt V  t 
(X)

in the same way before, now, we consider

D
  dV
Dt V t 
However, we can not evaluate this directly since all variables of the problem may
not continuous in V(t).

Although the phase interface moves through the material, a single particle of the
material does not necessary move. e.g., melting of an ice cube.

water ice t=0

water ice t=t1

7-4
 The transport theorem for a region containing a singular surface can be written as
(see p.20-21, 24-25 of Slattery’s book).

    * *
D
Dt V t 
  t  dV   t
dV    v  n dS    

  
U 
 n dS
V t  S t  S*


where n  outward unit normal vector to S(t)
  
U* n *  normal component of U*  speed of displacement of the surface.

If  is continuous, i.e., +=-, we recover the usual form of transport theorem.

Recall that if  is continuous, then

D       
   t  dV      v   dV   dV   v n dS
Dt V  t  V t  
t  V t 
t S t 
7-5

7-2-2 Jump Mass Balance and the Jump Linear Momentum Equation

S+ Phase interface
V+  * (Singular surface)
n
+ phase

 S*
- phase
U*
S-
V-

The time rate of change of the mass in the volume is

D
Dt V t 
 dV  0

Apply the generalized transport theorem, the above equation becomes

    

  t
 
dV   vn dS      U*  n * dS  0 (A)
V V S   S S*

The continuity equation is valid everywhere in the V+ phase and everywhere in the
V- phase, and we can integrate it over each region:

7-6
   
  t  v  dV  0 (B)
V 
t  
   (c)
  t   v  dV  0
V t  

Use Green’s theorem, we have

S+
V+ 
n*
+ phase     
   v  dV   vn dS    v    n*  dS
U* S* V  t  S  t 
 S*
n*

 n*
    
U*
S* S-   v  dV   v n dS    v   n *  dS
V-
V  t  S  t  S*
+ phase

7-7

Hence, equation (B) and (C) become

    
 t
dV   v  n dS    v   n * dS  0 (D)
V t S  t
S *

    
 t
dV   v  n dS    v   n * dS  0 (E)
V t S t 
S *

where + and - are the values of the density on the + and – sides of the surface,
and v  and v - are the velocity vectors of the material on the + and – sides of the
interface. These vectors are in general different from the velocity of the singular
surface S* since the singular surface is not necessarily material.

If we subtract equation (D) and (E) from equation (A), we have

 
   
   v  n *    U * n *    v  n *    U *  n *  dS  0
*  
S

Since this is true for any size surface element of the phase interface, hence the
integrated must be zero. That is

7-8
Jump Mass Balance
   
 
 
 

 v  n *  U* n *   v  n *  U*  n * 
 mass gain in  phase    mass loss in  phase 

Jump Momentum Balance

           
   
 v  v  n *  U* n *  n *     v  v  n *  U* n *  n *  

Remarks: Here, interface effects (such as surface tension) are considered.

7-9

Special cases

(A)No Mass Transferred Across the Interface (e.g., moving boundary problem in an
injection-molding die)
   
 
  
 
 v  n *  U* n *   v  n *  U* n *  0 
Hence
     
v  n *  v  n *  U* n *

Also, from the jump momentum balance equation, we have

 
n *    n *   
n*
Since 
  + phase
t   n *      force area  that  phase exerts on  surface 
 * 
t  n     force area  that  phase exerts on  surface - phase

Clearly,
 force area  that the   force area  that the 
     
t  t i.e.,   phase exerts on      phase exerts on 
 the  phase   the  phase 
    7-10
In RCC system

v i n *i  vi n *i  U*i n *i
n *j ji  n *j ji

(B) No Mass transferred Across Interface and Surface has No Normal Velocity
Component
 
In this case, U*  n *  0 , we have
v i n *i  vi n *i  0
n *j ji  n *j ji

Additional Hypothesis at Phase Boundary-No Slip Boundary Condition:

The tangential components of velocity are continuous at the phase interface.

 
Tangential components of v +  Tangential components of v 

This equation is an assumption (no proof for it), but it can rationalized from
experimental data and also from the face that this assumption excludes infinite
velocity gradients.
7-11

Formulation of Stress and Velocity Boundary Conditions

In this section, we consider the case that there is no mass transfer across the
interface only.

(A) Velocity Boundary Conditions

Since we have
vi n *i  vi n *i

In addition, we also postulate the following that tangential component in the

two phases are equal,
vi  vi

(B) Stress Boundary Conditions

We have
 
n*   n* 

7-12
[Example]
+ phase
In RCC system  z
n*

y
- phase x
Velocity boundary conditions:
v x  v x The application of the equations
v y  v y
vi  vi
v v

z

z
 
Stress boundary conditions n *    n* 
n *j ji  n *j ji 

to RCC system is straightforward, but the
 zx  zx utilization of these results to a phase of interface

 zy   zy which is the surface of a curvilinear coordinate

 zz   zz system presents some difficulty because some
aspect of curvilinear tensors must be introduced.

7-13

We now consider the following three types of phase interfaces.

(1) Solid-Fluid Interfaces

 
v  fluid   v  solid 
Since solids are usually considered not deformable, the stress B.C.’s are not
used to solve the mechanics problems, but can be applied after the problem is
solved, to calculate the force of the fluid exerts on the solids.

(2) Liquid-Gas Interface

In most ChE problems, we try to avoid solving the velocity field in the gas phase
so that the equation
 
v  gas   v  liquid 
is not of much use. We apply the stress condition to such interface.

7-14
 + phase (Gas)
 z
n*

y
x
- Phase (Liquid)
 zx  zx
 zy   zy
 zz   zz

Assume the liquid to be an INF, then

 v v j 
ij  Pij    ij  i 
 x x 
Assume the gas to be an ideal or inviscid fluid (   g   l )
ij  Pij (in Gas phase)

Thus we have

7-15

 v z v x  
 
      
   0 v z v x
 0
 x z 
zx zx
 x z
  v y   
   v z  v z v y
 zy   zy       0    0
 y  z  y  z
  
   v z

    v z  
 P  2  P
 zz   zz  P  2 P
z  z

Note that the pressure is not necessarily continuous at the gas-liquid interface.

(3) Liquid-Liquid Interface

We need the velocity field in both liquid phases so we must use both velocity
and stress boundary conditions

+ phase (Liquid)
 z
n*

y
x
- Phase (Liquid)
7-16
Assume that both liquids be INF

Velocity conditions:
v x  v x
v y  v y
v z  v z
Stress Conditions:

  v z v x 
   v z

v x  
             zx
 x z   x z 
zx


 v z v y 

  v y
 v z
 
             zy
zy  y  z  y z
   
 
   v z   v z 
 zz  P  2  P  2   zz
z z

Note again that the pressure is nor necessarily continuous across a liquid-liquid
interface.

7-17