2nd-Year Fluid Mechanics, Faculty of Science & Engineering, Curtin University

ENGR2000: FLUID MECHANICS

For Second-year Chemical, Petroleum, Civil & Mechanical Engineering

FLUID MECHANICS LECTURE NOTES

CHAPTER 3: MASS CONSERVATION, MOMENTUM AND ENERGY

3.1 Introduction

In order to introduce the ideas of conservation, it is often necessary to isolate

a part of the system and consider what is going on within that part of the
flow field. This is achieved by using a control volume (CV). We enclose a
part of the whole system by a notional boundary and perform our analysis
using the CV – what is happening within it and what is happening at its
boundaries. Examples of CV’s are seen in Fig. 3.1.

FIGURE 3.1: Examples of control volumes

Chapter 3 − Page 1
2nd-Year Fluid Mechanics, Faculty of Science & Engineering, Curtin University

Note that a CV can include regions where there is no fluid. The CV surface
- called the Control Surface (CS) - should not cross solid boundaries nor,
strictly, include solid components; however, we can ‘isolate’ such elements by
putting a separate CS around them. Often (see lower example in Fig. 3.1)
we do not bother to draw this isolating boundary in since it does not affect
the analysis.

The CV’s illustrated above are stationary; we can use moving CV’s – would
these be of greater use? In mathematical terminology should we use an
Eulerian (fixed frame) or Lagrangian (moving frame) approach to under-
standing flow behaviour? Consider the following analogy to fluid flow: Traffic
is travelling along a road and we wish to characterise it. The Eulerian would
set up a check-point at a particular location and measure the number of cars
per hour passing that point. In contrast, the Lagrangian would follow partic-
ular vehicles and determine their destinations. Each is valid form of analysis
meeting different interests.

In Fluid Mechanics, we are not generally interested in the fate of a particular

‘lump’ of fluid and so we adopt a Eulerian approach. In Solid Mechanics the
Lagrangian approach is used because we are concerned with the motion (eg.
locii and orbits) of particular parts of the structure.

3.2 Mass conservation

The principle of mass conservation is based upon the physical recognition that
mass is neither created nor destroyed (certainly in the Newtonian world!).
In Fluid Mechanics, this is interpreted in terms of mass flow. For a control
volume in which the mean flow field is time independent, we can say that
the rate of mass flow into a CV must be equal to the rate at which it leaves.
To illustrate this idea, consider Fig. 3.X1.

Chapter 3 − Page 2
2nd-Year Fluid Mechanics, Faculty of Science & Engineering, Curtin University

FIGURE 3.X1: Fluid mass flow through a simple pipe system

Denoting a mass flow-rate as ṁ (the over-dot means ’rate with time’, i.e. the
time derivative) with units kg/s, then it is obvious that

ṁ1 = ṁ2 + ṁ3 (3.X1)

to show that this is a conservation principle, re-write it as

0= ṁ1 − ṁ2 − ṁ3 (3.X2)

| {z }
Net mass flow-rate into CV
Thus the net mass flow rate into (or out of) the CV is zero so no fluid mass
is ‘being created’ inside the CV.

With the above ideas in mind, we now undertake a general analysis of mass
conservation that can be applied to any flow field.

For simplicity we shall consider a two-dimensional flow. We set up a CV in

the flow as shown in Fig. 3.2. To make life easy we choose a square CV with
sides in the x- and y-directions. The choice of CV can be very important in
facilitating the analysis.

Chapter 3 − Page 3
2nd-Year Fluid Mechanics, Faculty of Science & Engineering, Curtin University

FIGURE 3.2: Fluid mass entering and leaving a CV

ṁAB (etc.) is the rate at which fluid mass crosses the boundary AB. This
type of quantity is called a mass flux. Mass conservation tells us that:

Rate of increase of mass in CV = net flow of mass into CV across CS

Z Z 
∂
ρ dxdy = ṁAB − ṁCD + ṁDA − ṁBC (3.1)
∂t ABCD

The integral notation is used on the LHS because the fluid density, ρ, could
vary within the CV. If the mean density of fluid within the CV is invariant
with time and the CV is of fixed size, then the LHS will be zero. (This is
often the case but be wary of chemical reactions etc. going on inside a CV).
We need to evaluate the mass fluxes in Eqn. (3.1):

Chapter 3 − Page 4
2nd-Year Fluid Mechanics, Faculty of Science & Engineering, Curtin University

FIGURE 3.3: (a) Element of mass flux across AB (b) Surface notation

Only the x-component of the velocity, u, allows mass flux across the face
AB. The amount of mass, δm0 , crossing δy in time δt is:
dm0
δm0 = ρ(uδt)δy ⇒ = ṁ0 = ρuδy (3.2a, b)
dt
Summing all the increments of mass flux, ṁ0 , across the face AB gives that:
Z
ṁAB = ρudy (3.3)
AB

Note that if the density and velocity of the flow across the face AB is uniform
- for example a unidirectional jet of water from left to right - then ρ and u
are constants that come out of the integral. Equation 3.3 then reduces to
Z
ṁAB = ρu 1 dy = ρuAAB
AB

where AAB is the area of the face AB. Thus, if it were a uniform jet of speed
U in the x-direction and cross-sectional area A, its mass flow-rate would be
ρU A.

Returning to the general case, in a similar fashion to Eqn. 3.3, then:

Z Z Z
ṁCD = ρudy ṁDA = ρvdx ṁBC = ρvdx
CD DA BC

Instead of writing out the ṁ terms separately, we can be more succinct by

using vector notation. Referring to Fig 3.3b an outward normal is defined on
the CV surface with an element of surface being δs. Equation 3.2a becomes:

δm0 = −ρ(v.n δt)δs (3.4)

Chapter 3 − Page 5
2nd-Year Fluid Mechanics, Faculty of Science & Engineering, Curtin University

which now holds for any of the four faces. (The minus sign comes in because
we are using an outward normal.) Conservation of mass (Eqn. 3.1) can now
be written: Z Z  I
∂
ρ dxdy = − ρv.nds (3.5)
∂t V S

This generalised form can be used for a (two-dimensional) CV of any shape.

Finally we note that Equation 3.5 is easily extended for three-dimensional

flows: Z Z Z  Z Z
∂
ρ dxdydz = − ρv.ndA (3.6)
∂t CV CS

where dA is an area element of the two-dimensional control surface (CS)

which surrounds the CV.

A simple example: consider an inviscid (so, no shear or velocity profile)

incompressible flow in the circular converging pipe seen in Fig. 3.5. Set up
a suitable CV as shown.

FIGURE 3.5

(i) Incompressible and steady flow, so:

Z Z Z 
d
ρ dxdydy = 0
dt CV

(ii) v.n = 0 on the side walls so the boundary integration only needs to be
considered over the two end faces of the CV.

Chapter 3 − Page 6
2nd-Year Fluid Mechanics, Faculty of Science & Engineering, Curtin University

(iii) Inviscid so velocity is constant over each end face – this makes the
integration trivial. Face 1: v.n = −U1 . Face 2: v.n = U2 .

Thus, Eqn. 3.6 gives that:

0 = ρU1 A1 − ρU2 A2

or
U1 A1 = U2 A2
A complicated way to arrive at a simple answer! We could have found this
by just saying “flow rate in = flow rate out”. Mass conservation is a simple
concept: the formalisation above will assist us in the following sections.

In many applications we find that that there is no spatial variation of flow

velocity over the face of a control volume (as occurs above). In other words
there is uniform velocity across the face. When this is the case we can use
the simpler
ṁ = ρU A
formula to evaluate the mass flux across each face and then use the mass
conservation equation in words as

Rate of increase of mass in CV = net flow of mass into CV across CS

but be careful... U must be the component of the flow velocity that is per-
pendicular to the control surface area A.

Chapter 3 − Page 7
2nd-Year Fluid Mechanics, Faculty of Science & Engineering, Curtin University

3.3 The momentum equation

We can now use a similar approach to establish a momentum balance for a

CV. Since there could be forces acting inside the CV, we must account for
these by using Newton’s second law: Force = rate of change of momentum.
It is also important to note that momentum (like force) is a vector quantity.
In vector form and where there may be more than one force (hence the sum
of forces), Newton’s second law is written:
X d
F= (mv) (3.7a)
dt
Consider the fluid element in Fig. 3.6.

FIGURE 3.6: Momentum flux into/out of a CV

Equation 3.7a can be written as:

Sum of applied forces = Rate of increase of momentum inside CV
– net rate at which momentum enters CV across CS (control surface)

(3.7b)

The momentum of a fluid volume element, δxδy, is:

δ(M omentum) = δ(M) = ρδxδyv

Chapter 3 − Page 8
2nd-Year Fluid Mechanics, Faculty of Science & Engineering, Curtin University

So, the first term on the RHS of Eqn 3.7b is:

Z Z 
∂
Rate of inc. of M in CV = ρvdxdy (3.8)
∂t V

What about the rate at which momentum crosses the CV? Consider face AB
(in Fig. 3.6). The amount of momentum, δM0 , crossing δy in time δt is:

dM0
δM0 = ρ(uδt)vδy ⇒ = Ṁ0 = ρuvδy (3.9a, b)
dt

Summing all the increments of momemtum flux, Ṁ0 , across the face AB gives
that: Z
ṀAB = ρuvdy (3.10)
AB
In a similar fashion, then:
Z Z Z
ṀCD = ρuvdy ṀDA = ρvvdx ṀBC = ρvvdx
CD DA BC

Again, it is more concise to use vector form. So, Eqn. 3.9b becomes:

dM0
= Ṁ0 = −ρ(v.n)vδs (3.11)
dt
and the momentum flux across the whole CS is given by:
I
net rate at which M enters CV. = − ρv(v.n)ds (3.12)
S

We can now assemble the (two-dimensional) momentum equation, using

Eqns.3.7b, 3.8 and 3.12:
Z Z  I
X ∂
F= ρvdxdy + ρv(v.n)ds (3.13)
∂t V S

which is applicable to any shape of CV. Note: This is a vector equation....

use it appropriately.

This form is readily extended to a three-dimensional CV of arbitrary shape:

Z Z Z  ZZ
X ∂
F= ρvdxdydz + ρv(v.n)dA (3.14)
∂t CV CS

Chapter 3 − Page 9
2nd-Year Fluid Mechanics, Faculty of Science & Engineering, Curtin University

where dA is an area element of the control surface area.

The vector momentum equation can be broken down into two (in 2-D) scalar
equations in the x- and y-directions, since we have that:
F = Fx i + Fy j and v = ui + vj
For the rectangular CV seen in Fig. 3.6, the x- and y-momentum equations
are:

Sum of forces in x-direction rate of inc. of x-momentum in CV

z Z Z }| {
∂
zX}| {
Fx = ρudxdy
Z Z Z ∂t V Z 
2 2
− ρu dy − ρu dy + ρuvdx − ρuvdx (3.15a)
AB CD DA BC
| {z }
[rate of x-momentum IN - rate of x-momentum OUT]

Sum of forces in y-direction rate of inc. of y-momentum in CV

z Z Z }| {
∂
zX }| {
Fy = ρvdxdy
 Z Z Z ∂t V Z 
2 2
− − ρvudy + ρvudy − ρv dx + ρv dx (3.15b)
AB CD DA BC
| {z }
[rate of y-momentum IN - rate of y-momentum OUT]

When using these scalar equations, be careful with the signs of the flow
speeds.

Finally, when we have uniform free jets or ‘contained jets’ (such as inviscid
flow in a pipe), then we can use the following simplified forms of momentum
flux (rate of momentum) across a control surface.
Ṁx = ṁu Ṁy = ṁv
for momentum flux in, respectively, the x- and y-directions across a CS
boundary where the mass flux (mass flow rate) is ṁ across the boundary
and u and v are the components of the velocity at the boundary.

In many of the examples/applications that we will consider, this is an easier

way to proceed in tandem with the written form of Eqns. 3.15a and 3.15b
above. (However, the integral forms remain important for the general case
of fluid momentum and the forces that fluids exert.)

Chapter 3 − Page 10
2nd-Year Fluid Mechanics, Faculty of Science & Engineering, Curtin University

EXAMPLE: A jet of fluid impinges on the fixed plate as shown in Fig. 3.7.
Assuming inviscid, incompressible flow and neglecting gravitational effects,
what is the force experienced by the plate?

FIGURE 3.7: A fluid jet impinging on a flat plate showing a suitable CV

(i) Note that on the CS-faces of interest:

At face AB v = U1 i + 0j Unit normal: − i

At face BC v = 0i + U2 j Unit normal: + j
At face AD v = 0i − U2 j Unit normal: − j

(ii) The flow is steady, (∂/∂t = 0), therefore zero rate of increase of momen-
tum inside the CV. (First term on RHS of Eqn. 3.13 is zero)

(iii) Use Eqn. 3.13 with only one force:

Z Z Z
F = 0+ ρU1 i(U1 i.(−i))dx+ ρU2 j(U2 j.(j))dx+ ρ(−U2 j)(−U2 j.(−j))dx
AB BC AB

The second and third integrals cancel each other. The first integral is trivial
since the flow velocity is constant across area A1 .

Chapter 3 − Page 11
2nd-Year Fluid Mechanics, Faculty of Science & Engineering, Curtin University

The force on the fluid is:

F = −ρA1 U12 i + 0j

ie. to the left. It follows that the force on the plate must be of the same
magnitude but to the right.

Again, this might seem a very complicated way to arrive at a simple answer.
Common sense would have told us that (i) symmetry gives that Fy = 0 (ii)
Fx must be the rate at which all the x-momentum is arriving across AB is
being ‘cancelled’ (since no x-momentum leaves across BC or AD). This rate
is
ṁU1 = ρ(U1 A1 )U1 = ρA1 U12
and then figure out the direction of the force on the plate.... obvious. How-
ever, the above does illustrate how to use the formal Eqn. 3.13. We will
consider more elaborate examples/problems later.

Chapter 3 − Page 12
2nd-Year Fluid Mechanics, Faculty of Science & Engineering, Curtin University

3.4 The Reynolds Transport Theorem

The derivations seen above for mass and momentum conservation have much
in common - look at the final (two-dimensional) equations again:

d ∂
RR
H
0 = (m) = ∂t CV
ρ dV + CS ρv.ndA (3.14a)
dt
X d ∂
RR
H
F = (mv) = ∂t CV
ρvdV + CS ρv(v.n)dA (3.14b)
dt

where dV = dxdydz is an element of volume.

We will now generalise this form of transport equation to accommodate any

flow quantity (ie. mass, momentum...). Call the flow quantity B and the
amount of B per unit mass β (β is the intensive value of B). The amount of
B inside a control volume is:
ZZ
B= ρβdV (3.16)
CV

Consider Fig. 3.8 which shows a CV occupied by a ‘lump’ (or system) of

fluid (containing quantity, B) at time t and then the boundary of the same
system at time t + δt when it contains quantity B 0 .

FIGURE 3.8

Chapter 3 − Page 13
2nd-Year Fluid Mechanics, Faculty of Science & Engineering, Curtin University

We are interested in how the property B 0 changes within the moving

system and to relate this change to values of the fixed CV.
∂ 1 1
B = B(t + δt) − B(t)
∂t δt δt
1 1
= [B 0 (t + δt) − (ρβδV )out + (ρβδV )in ] − B 0 (t) (3.17)
δt δt
having used the fact that B = B 0 at time t. Substituting δV = AU δt and
rearranging, we find:

dB 0 1 ∂
≈ [B 0 (t + δt) − B 0 (t)] = B + (ρβAU )out − (ρβAU )in (3.18)
dt δt ∂t
|{z} | {z } | {z }
(i) (ii) (iii)

The approximation becomes exact in the limit δt → 0.

This is the (one-dimensional) Reynolds transport theorem for a fixed CV. It

relates the variation of a quantity, B, in a moving fluid to values of B in a
fixed CV. The terms on the RHS of Eqn. 3.18 are:

(i) The rate of change of B within the CV.

(ii) The flux of B passing out of the CV.

(iii) The flux of B passing into the CV.

Equation 3.18 is applied to a general control volume in vector form for two
dimensions to give:
Z Z  I
d ∂
(B) = ρβdV + ρβ(v.n)dA (3.19)
dt ∂t CV CS

Notice that we have not specified what the property B is..... If B = mass
then substitute β = 1 and Eqn. 3.14a is found. If B = momentum = mv
then substitute β = v and Eqn. 3.14b is found.

Chapter 3 − Page 14
2nd-Year Fluid Mechanics, Faculty of Science & Engineering, Curtin University

3.5 The Energy equation

The first law of thermodynamics states that:

dQ dW dE
dQ − dW = dE or − = (3.21a, b)
dt dt dt
where: dE ≡ system energy; dQ ≡ heat added to the system; dW ≡ work
done by the system. (Note: positive Q is heat added to the system, positive
W is work done by the system.)

We will use the Reynolds transport theorem, Eqn. 3.19, to generate the
energy equation. Thus, use B = E and therefore β = E/m = e, the energy
per unit mass. The resulting expression is:
Z Z  I
d ∂
Q̇ − Ẇ = (E) = ρedV + ρe(v.n)dA (3.22)
dt ∂t CV CS

The dot notation indicates the time derivative. (Hence Ẇ ≡ power.)

Fluid energy (per unit mass), e: This can encompass different types of
energy, e.g.:
e = einternal + ekinetic + epotential + eother
Neglecting eother (chemical reactions, magnetic field effects etc.), the above
expression becomes:
1
e = û + |v|2 + gz (3.23)
2
where z is an ‘up’ coordinate.

Rate of Work (per unit mass), Ẇ : This can be divided up as follows.

Ẇ = Ẇshaf t + Ẇpressure + Ẇviscous stress (3.24)
Ẇshaf t is the work done to a pump, fan blade, piston etc. which protrudes
into the CV. Generally the work done by viscous forces, Ẇviscous stress , is
negligible. The work done by the pressure, Ẇpressure needs to be evaluated:
since pressure acts in all directions, the sum of the work done inside the CV
is zero. What remains is the work done at the CS. Since:
δ Ẇp = −p(−v.n)δA
then summing over the CS of the CV, we have:
I
Ẇp = p(v.n)dA (3.25)
CS

Chapter 3 − Page 15
2nd-Year Fluid Mechanics, Faculty of Science & Engineering, Curtin University

When introduced into Eqn. 3.22, the above term can be taken over to the
RHS and included with the transport terms (area integral).

Combining Eqns. 3.22, 3.23, 3.24 and 3.25, we have:

Z Z  I
∂ 1 2 p 1
Q̇−Ẇs = ρ(û + |v| + gz)dV + ρ(û+ + |v|2 +gz)(v.n)dA
∂t CV 2 CS ρ 2
(3.26)
This is the (two-dimensional) energy equation. Often the enthalpy, ĥ = û+ ρp ,
is used in the second term of the RHS above.

3.5.1 The steady-flow energy equation

For steady (so, ∂/∂t = 0) one-dimensional flow with one inlet and one outlet
as shown in Fig. 3.9,

FIGURE 3.9: One-dimensional flow in a duct.

where the flow speed across area cross-sections, A1 and A2 , is constant, then
the surface integrations become trivial. Equation 3.26 collapses to:
1 1
ṁ2 (ĥ2 + U22 + gz2 ) = ṁ1 (ĥ1 + U12 + gz1 ) + Q̇ − Ẇs (3.27)
2 2
From mass continuity, ṁ1 = ṁ2 ; dividing by this quantity gives:
1 1
ĥ2 + U22 + gz2 = ĥ1 + U12 + gz1 + q − ws (3.28a)
2 2

Chapter 3 − Page 16
2nd-Year Fluid Mechanics, Faculty of Science & Engineering, Curtin University

where: q = Q̇/ṁ = Q/m (the heat transferred per unit mass) and ws =
Ẇs /ṁ = Ws /m (the work done per unit mass of fluid). Thus the steady-
flow energy equation states that the downstream stagnation enthalpy,
H2 = ĥ2 + 21 U22 + gz2 differs from the upstream value, H1 , only if there is
heat transfer or shaft work done as the fluid passes from station 1 to station
2.

Note that Eqn. 3.28a has units of energy/unit mass, (J/kg). Often, the
equation is written in terms of head (i.e. length) and seen as:

p2 û2 U22 p1 û1 U12

+ + + z2 = + + + z1 + hq − hs (3.28b)
ρg g 2g ρg g 2g

where hq = q/g and hs = ws /g are the head changes due to heat transfer
and (shaft) work done.

You might note some similarities between the above equation and the Bernoulli
equation (for those of you who have seen it). In the next chapter we study
the Bernoulli equation and then highlight its relationship with the energy
equation.

Chapter 3 − Page 17