C

7
A

Fluid and Thermal Systems

CHAPTER OUTLINE

Part I. Fluid Systems 397

7.1 Conservation of Mass 397
7.2 Fluid Capacitance 402
7.3 Fluid Resistance 407
7.4 Dynamic Models of Hydraulic Systems 411
7.5 Pneumatic Systems 427
Part II. Thermal Systems 430
7.6 Thermal Capacitance 431
7.7 Thermal Resistance 432
7.8 Dynamic Models of Thermal Systems 441
Part III. MATLAB and Simulink Applications 449
7.9 MATLAB Applications 449
7.10 Simulink Applications 453
7.11 Chapter Review 458
Reference 458
Problems 458

CHAPTER OBJECTIVES

When you have finished this chapter, you should be able to

1. Apply the conservation of mass principle to model
simple hydraulic and pneumatic systems.
2. Determine the appropriate resistance relation to use
for laminar, turbulent, and orifice flow.
3. Develop a dynamic model of hydraulic and
pneumatic systems containing one or more fluid
containers.
4. Determine the appropriate thermal resistance
relation to use for conduction, convection, and
radiation heat transfer.
5. Develop a model of a thermal process having one or
more thermal storage compartments.
6. Apply MATLAB and Simulink to solve fluid and
thermal system models.

fluid system uses one or more fluids to achieve its purpose. The dampers, shock
absorbers, and door closer we saw in Chapter 4 are examples of fluid systems
because they depend on the viscous nature of a fluid to provide damping. A
fluid might be either a liquid or a gas. Part I of this chapter concerns the study of
fluid systems, which can be divided into hydraulics and pneumatics. Hydraulics is
the study of systems in which the fluid is incompressible, that is, its density stays
approximately constant over a range of pressures. Pneumatics is the study of systems in
which the fluid is compressible. Hydraulics and pneumatics share a common modeling

396

7.1

Conservation of Mass

principle: conservation of mass. It will form the basis of all our models of such systems.
Modeling pneumatic systems also requires application of thermodynamics, because the
temperature and density of a gas can change when its pressure changes.
Thus pneumatics provides a bridge to the treatment of thermal systems, which is
the subject of Part II of the chapter. Thermal systems are systems that operate due to
temperature differences. They thus involve the flow and storage of thermal energy, or
heat, and conservation of heat energy forms the basis of our thermal models.
Part III illustrates applications of MATLAB and Simulink to fluid and thermal
systems.
Fluid and thermal systems are more complicated than most electrical and mechanical systems. While, for example, there are formulas available to compute the spring
constant of typical elastic elements, few formulas are available for the coefficients that
will appear in our fluid and thermal models, and the coefficients values often must be
determined experimentally. For this reason, the methods for developing models from
data, covered in Chapter 8 and Appendix C, are most important for modeling fluid and
thermal systems.

PART I. FLUID SYSTEMS

In addition to providing damping, other applications of fluid systems include actuators and processes that involve mixing, heating, and cooling of fluids. Active vehicle
suspensions use hydraulic and pneumatic actuators to provide forces that supplement
the passive spring and damping elements. Water supply, waste treatment, and other
chemical processing applications are examples of a general category of fluid systems
called liquid-level systems, because they involve regulating the volumes, and therefore
the levels of liquids in containers such as tanks.
Because all real fluids are compressible to some extent, incompressibility is an
approximation. But this approximation is usually sufficiently accurate for most liquids
under typical conditions, and it results in a simpler model of the system. For this reason
we will begin our study of fluid systems with hydraulics.

7.1 CONSERVATION OF MASS

We will avoid complex system models by describing only the gross system behavior
instead of the details of the fluid motion patterns. The study of such motion belongs to
the specialized subject of fluid mechanics and will not be treated here.
For incompressible fluids, conservation of mass is equivalent to conservation of
volume, because the fluid density is constant. If we know the mass density and the
volume flow rate, we can compute the mass flow rate. That is, qm = qv , where qm
and qv are the mass and volume flow rates. The FPS and SI units for mass flow rate
are slug/sec and kg/s, respectively. The units for volume rates are ft3 /sec and m3 /s,
respectively. Other common units for volume are the U.S. gallon, which is 0.13368 ft3 ,
and the liter, which is 0.001 m3 .

7.1.1 DENSITY AND PRESSURE

The units for mass density are slug/ft3 and kg/m3 . Sometimes one encounters weight
density, whose common symbol is . Its units are lb/ft3 or N/m3 , and it is related to
the mass density as = g, where g is the acceleration due to gravity. The mass
density of fresh water near room temperature is 1.94 slug/ft3 , or 1000 kg/m3 . The mass

397

398

CHAPTER 7

Fluid and Thermal Systems

density of air at sea level and near room temperature is approximately 0.0023 slug/ft3 or
1.185 kg/m3 .
Pressure is the force per unit area that is exerted by the fluid. The FPS and SI units
of pressure are lb/ft2 and the Pascal (1 Pa = 1 N/m2 ), respectively. Another common
unit is psi (lb/in.2 ). At sea level near room temperature, atmospheric pressure, usually
abbreviated pa , is 14.7 psi (2117 lb/ft2 ) or 1.0133105 Pa. Gage pressure is the pressure
difference between the absolute pressure and atmospheric pressure, and is often abbreviated as psig. For example, 3 psig is 17.7 psi absolute (which is abbreviated as psia).
Hydrostatic pressure is the pressure that exists in a fluid at rest. It is caused by the
weight of the fluid. For example, the hydrostatic pressure at the bottom of a column of
fluid of height h is gh. If the atmospheric pressure above the column of liquid is pa ,
then the total pressure at the bottom of the column is gh + pa .
E X A M P L E 7.1.1

A Hydraulic Brake System

Problem

Figure 7.1.1 is a representation of a hydraulic brake system. The piston in the master cylinder
moves in response to the foot pedal. The resulting motion of the piston in the slave cylinder
causes the brake pad to be pressed against the brake drum with a force f 3 . Obtain the expression
for the force f 3 with the force f 1 as the input. The force f 1 depends on the force f 4 applied by the
drivers foot. The precise relation between f 1 and f 4 depends on the geometry of the pedal arm.
Solution

The forces are related to the pressures and the piston areas as follows: f 1 = p1 A1 and f 2 = p2 A2 .
Assuming the system is in static equilibrium after the brake pedal has been pushed, we see that
p1 = p2 + gh, where h is the height between points 1 and 2. Thus, if h is small, that is, if
the pressure gh is negligible compared to p2 , then p1 = f 1 /A1 = p2 = f 2 /A2 . The forces are
therefore related as f 2 = f 1 A2 /A1 , and if the area A2 of the slave piston is greater than the area
A1 of the master piston, the force f 2 will be greater than the force f 1 . So we see that this system
serves to amplify the pedal force.
The force f 3 can be obtained from the lever relation f 3 = f 2 L 1 /L 2 , assuming static
equilibrium or negligible lever inertia.
x1

Figure 7.1.1 A hydraulic

brake system.

Pivot
f1

Master cylinder p1
A1

f4

A2
f2

p2

L1

x2
Pivot

Slave cylinder

L2
Drum

f3
Brake
pad

Brake
pedal

7.1

Conservation of Mass

399

The tradeoff for force amplification is that the master piston must move a distance greater
than that of the slave piston. We may see this effect by equating the fluid volume displaced by
each piston.
volume = A1 x1 = A2 x2
Thus x2 = x1 A1 /A2 , and so x2 < x1 if A1 < A2 .

Conservation of mass can be stated as follows. For a container holding a mass of

fluid m, the time rate of change m of mass in the container must equal the total mass
inflow rate minus the total mass outflow rate. That is,
m = qmi qmo
(7.1.1)
where qmi is the mass inflow rate and qmo is the mass outflow rate.
The fluid mass m is related to the container volume V by m = V . For an incom Let qvi and qvo be the total volume inflow
pressible fluid, is constant, and thus m = V.
and outflow rates. Thus, qmi = qvi , and qmo = qvo . Substituting these relationships
into (7.1.1) gives
V = qvi qvo
Cancel to obtain
(7.1.2)
V = qvi qvo
This is a statement of conservation of volume for the fluid, and it is equivalent to
conservation of mass, equation (7.1.1), when the fluid is incompressible.

A Water Supply Tank

E X A M P L E 7.1.2

Problem

Water is pumped as needed at the mass flow rate qmo (t) from the tank shown in Figure 7.1.2a.
Replacement water is pumped from a well at the mass flow rate qmi (t). Determine the water
height h(t), assuming that the tank is cylindrical with a cross section A.
Figure 7.1.2 A water supply tank.

qmi
qmo
h

Distribution
qmi

Well

(a)

qmo

(b)

400

CHAPTER 7

Fluid and Thermal Systems

Solution

We model the system as shown in part (b) of the figure. The volume of water in the tank is Ah,
and therefore the mass of water in the tank is Ah, where is the mass density of water. From
conservation of mass, we have
d
( Ah) = qmi (t) qmo (t)
dt
Since and A are constant, we have
A

dh
= qmi (t) qmo (t)
dt

which can be integrated as follows:

h(t) = h(0) +

1
A

[qmi (u) qmo (u)] du

0

Once we know the flow rates, we can evaluate the integral.

A common hydraulic actuator is the piston-and-cylinder actuator used on many

types of heavy equipment, such as the backhoe shown in Figure 7.1.3. When the operator
moves a handle, hydraulic fluid under high pressure is sent through the line to the
cylinder. The fluid acts on the piston within the cylinder and produces a force that
is equal to the pressure times the piston area. This large force moves the linkage.
Example 7.1.3 develops a simple model of such a device.
Figure 7.1.3 A backhoe.

Piston
rod
Cylinder
Operator
controls

Hydraulic
lines

E X A M P L E 7.1.3

A Hydraulic Cylinder
Problem

Figure 7.1.4a shows a cylinder and piston connected to a load mass m, which slides on a
frictionless surface. Part (b) of the figure shows the piston rod connected to a rack-and-pinion
gear. The pressures p1 and p2 are applied to each side of the piston by two pumps. Assume the
piston rod diameter is small compared to the piston area, so the effective piston area A is the
same on both sides of the piston. Assume also that the piston and rod mass have been lumped
into m and that any friction is negligible. (a) Develop a model of the motion of the displacement

7.1

Pinion

x 
m

A
p1

p2

p1

401

Figure 7.1.4 A hydraulic

cylinder for (a) translating a
mass and for (b) rotating
a pinion gear.

5 x
R
R

Conservation of Mass

Rack

p2

(a)

(b)

x of the mass in part (a) of the figure, assuming that p1 and p2 are given functions of time.
Also, obtain the expression for the mass flow rate that must be delivered or absorbed by the two
pumps. (b) Develop a model of the displacement x in part (b) of the figure. The inertia of the
pinion and the load connected to the pinion is I .
Solution

a.

Assuming that p1 > p2 , the net force acting on the piston and mass m is ( p1 p2 )A, and
thus from Newtons law,
m x = ( p1 p2 )A
Because p1 and p2 are given functions of time, we can integrate this equation once to
obtain the velocity:
x(t)
= x(0)

b.

A
m

[ p1 (u) p2 (u)] du
0

The rate at which fluid volume is swept out by the piston is A x,

and thus if x > 0, the
pump providing pressure p1 must supply fluid at the mass rate A x,
and the pump
providing pressure p2 must absorb fluid at the same mass rate.
Because we want an expression for the displacement x, we obtain an expression for the
equivalent mass of the rack, pinion, and load. The kinetic energy of the system is
KE =

1 2 1 2
1
m x + I =
2
2
2

m+

I
R2

x 2

because R = x.

Thus the equivalent mass is

I
R2
The required model can now be obtained by replacing m with m e in the model developed
in part (a).
me = m +

A Mixing Process
Problem

A mixing tank is shown in Figure 7.1.5. Pure water flows into the tank of volume V = 600 m3
at the constant volume rate of 5 m3 /s. A solution with a salt concentration of si kg/m3 flows
into the tank at a constant volume rate of 2 m3 /s. Assume that the solution in the tank is well
mixed so that the salt concentration in the tank is uniform. Assume also that the salt dissolves
completely so that the volume of the mixture remains the same. The salt concentration so kg/m3
in the outflow is the same as the concentration in the tank. The input is the concentration si (t),

E X A M P L E 7.1.4

402

CHAPTER 7

Figure 7.1.5 A mixing

process.

2 m3/s
Solution si

Fluid and Thermal Systems

5 m3/s
Water

V 5 600 m3
so

qvo

Mixer

whose value may change during the process, thus changing the value of so . Obtain a dynamic
model of the concentration so .
Solution

Two mass species are conserved here: water mass and salt mass. The tank is always full, so the
mass of water m w in the tank is constant, and thus conservation of water mass gives
dm w
= 5w + 2w w qvo = 0
dt
where w is the mass density of fresh water, and qvo is the volume outflow rate of the mixed
solution. This equation gives qvo = 5 + 2 = 7 m3 /s.
The salt mass in the tank is so V , and conservation of salt mass gives
d
(so V ) = 0(5) + 2si so qvo = 2si 7so
dt
or, with V = 600,
dso
(1)
= 2si 7so
dt
This is the model. The time constant for the mixing process is 600/7 = 85.7 s. Thus, if si is
initially zero and then becomes a nonzero constant value S, the salt concentration in the outflow
will eventually become constant at the value 2S/7 after approximately 4(85.7) = 343 s.
600

7.2 FLUID CAPACITANCE

Sometimes it is very useful to think of fluid systems in terms of electrical circuits.
Table 7.2.1 gives the fluid quantity, its common nomenclature, its linear relation, and
its analogous electrical property. Fluid resistance is the relation between pressure and
mass flow rate. Fluid capacitance is the relation between pressure and stored mass.
We will concentrate on fluid resistance and capacitance. Fluid resistance relates to
energy dissipation while fluid capacitance relates to potential energy. Fluid inertance
relates to fluid acceleration and kinetic energy.
Fluid systems obey two laws that are analogous to Kirchhoffs current and voltage
laws; these laws are the continuity and the compatibility laws. The continuity law is
simply a statement of conservation of fluid mass. This says that the total mass flow into

7.2

Fluid Capacitance

403

Table 7.2.1 Analogous fluid and electrical quantities.

Fluid quantity
Fluid mass, m
Mass flow rate, qm
Pressure, p
Fluid linear resistance, R
R = p/qm
Fluid capacitance, C
C = m/ p
Fluid inertance, I
I = p/(dqm /dt)

Electrical quantity
Charge, Q
Current, i
Voltage, v
Electrical resistance, R
R = v/i
Electrical capacitance, C
C = Q/v
Electrical inductance, L
L = v/(di/dt)

a junction must equal the total flow out of the junction. This is analogous to Kirchhoffs
current law. Flow through two rigid pipes joined together to make one pipe is an example
where this applies. If, however, the flow is through flexible tubes that can expand and
contract under pressure, then the outflow rate is not the sum of the inflow rates. This
is an example where fluid mass can accumulate within the system and is analogous to
having a capacitor in an electrical circuit.
The compatibility law is analogous to Kirchhoffs voltage law, which states that the
sum of signed voltage differences around a closed loop must be zero. It is an expression
of conservation of energy. The compatibility law states that the sum of signed pressure
differences around a closed loop must be zero.

7.2.1 FLUID SYMBOLS AND SOURCES

Figure 7.2.1 shows the commonly used symbols for fluid system elements. The
resistance symbol is used to represent fixed resistances, for example, due to pipe flow,
orifice flow, or a restriction. A valve that can be manually adjusted, such as a faucet,
is a variable resistance and has a slightly different symbol. An actuated valve, driven,
for example, by an electric motor or a pneumatic device, has a different symbol. Such
valves are usually operated under computer control.
Just as there are ideal voltage and current sources in electrical systems, so we use
ideal pressure and flow sources in our fluid system models. An ideal pressure source
is capable of supplying the specified pressure at any flow rate. An ideal flow source is
capable of supplying the specified flow.
These ideal sources are approximations to real devices such as pumps. For example,
Figure 7.2.2 shows the steady-state flow-pressure relation for a centrifugal pump, where
qm is the mass flow rate produced by the pump when the pressure difference across the
pump is p. When the outlet pressure is greater than the inlet pressure, p > 0. Such

Resistance

p1

ps

Manually
adjusted valve

Actuated
valve

p2

ps 5 p2 2 p1

qs

Ideal pressure
source

Ideal flow
source

Pump

Figure 7.2.1 Fluid system

symbols.

404

CHAPTER 7

Figure 7.2.2 Steady-state

flow-pressure relation for a
centrifugal pump.

Fluid and Thermal Systems

s3

s2
s1
qm

Dp

curves depend on the pump speed, labeled s1 , s2 , and so on in the figure. To determine
the operating condition of the pump for a given speed, we need another relation between
qm and p. This relation depends on the load connected to the pump outlet. We will
see how such a relation is obtained in Example 7.4.8 in Section 7.4.

7.2.2 CAPACITANCE RELATIONS

Fluid capacitance is the relation between stored fluid mass and the resulting pressure
caused by the stored mass. Figure 7.2.3 illustrates this relation, which holds for both
pneumatic and hydraulic systems. At a particular reference point ( pr , m r ) the slope
is C, where

dm 
(7.2.1)
C=
d p  p= pr
Thus, fluid capacitance C is the ratio of the change in stored mass to the change in
pressure.

Figure 7.2.3 General fluid

capacitance relation and its
linear approximation.

m
C
1

mr

pr

7.2

Fluid Capacitance

405

Capacitance of a Storage Tank

E X A M P L E 7.2.1

Problem

Consider the tank shown in Figure 7.2.4. Assume that the sides are vertical so that the crosssectional area A is constant. This is the case, for example, with a cylindrical tank whose horizontal
cross section is circular, or with a tank having vertical sides and a rectangular horizontal cross
section. Derive the expression for the tanks capacitance.
Solution

Because the tanks sides are vertical, the liquid height h is related to m, the liquid mass in the
tank, by m = Ah. The total pressure at the bottom of the tank is gh + pa , but the pressure due
only to the stored fluid mass is p = gh. We can therefore express the pressure as a function of
the mass m stored in the tank as p = mg/A.
Thus,
m=

pA
g

and the capacitance of the tank is given by

C=

dm
A
=
dp
g
Figure 7.2.4 Capacitance of a
storage tank.

pa
h

(a)

(b)

When the container does not have vertical sides, such as the one shown in Figure 7.2.5, the cross-sectional area A is a function of the liquid height h, and the relations
between m and h and between p and m are nonlinear. In such cases, there is no single
value for the containers capacitance. The fluid mass stored in the container is
 h
A(x)d x,
m = V =
0

which gives
dm
= A
dh
Figure 7.2.5 A storage tank
of arbitrary shape.

h
qmi

qmo

406

CHAPTER 7

Fluid and Thermal Systems

For such a container, conservation of mass gives

dm
= qmi qmo
dt

(7.2.2)

but
dm d p
dp
dm
=
=C
dt
d p dt
dt
Thus
C

dm
dp
=
= qmi qmo
dt
dt

(7.2.3)

Also
dm dh
dh
dm
=
= A
dt
dh dt
dt
So
A

dh
= qmi qmo
dt

(7.2.4)

Equations (7.2.2), (7.2.3), and (7.2.4) are alternative, but equivalent, hydraulic
models of a container of fluid. They suggest that either pressure p, mass m, or height h
can be chosen as the models variable. These variables are all indicators of the systems
potential energy, and as such any one can be chosen as a state variable. If the containers
cross-sectional area is constant, then V = Ah and thus the liquid volume V can also
be used as the model variable.
E X A M P L E 7.2.2

Capacitance of a V-Shaped Trough

Problem

(a) Derive the capacitance of the V-shaped trough shown in Figure 7.2.6a. (b) Use the capacitance
to derive the dynamic models for the bottom pressure p and the height h. The mass inflow rate
is qmi (t), and there is no outflow.
Solution

a.

From part (b) of the figure, D = 2h tan , and the vertical cross-sectional area of the liquid
is h D/2. Thus the fluid mass is given by

m = V =

Figure 7.2.6 A V-shaped

trough.

A
h

D
h

2

2

L
(a)

(b)

1
hD
2

L = ( L tan )h 2

7.3

But p = gh and thus,


m = ( L tan )

p
g

From the definition of capacitance,

C=
b.

dm
=
dp

From (7.2.3) with qmo = 0, C p = qmi , or

2L tan
g 2

2


=

L tan
g 2

2L tan
g 2


p

Fluid Resistance

407


p2


p

dp
= qmi
dt

which is a nonlinear equation because of the product p p . We can obtain the model for the
height by substituting h = p/g. The result is
(2 L tan )h

dh
= qmi
dt

7.3 FLUID RESISTANCE

Fluid meets resistance when flowing through a conduit such as a pipe, through a component such as a valve, or even through a simple opening or orifice, such as a hole. We
now consider appropriate models for each type of resistance.
The mass flow rate q m through a resistance is related to the pressure difference p
across the resistance. This relation, p = f (q m ), is illustrated in general by Figure 7.3.1.
We define the fluid resistance Rr as the slope of f (q m ) evaluated at a reference equilibrium condition ( pr , qmr ). That is,



d p
d p 

Rr =
=
(7.3.1)
d q m q m =q mr
d q m r
If we need to obtain an approximate linear model of the pressure-flow rate relation,
we can use a Taylor series expansion to linearize the expression p r = f (q m ) near
a reference operating point ( pr , qmr ) as follows (after dropping the second order and
higher terms in the expansion):


d p
p = pr +
(qm qmr ) = pr + Rr (qm qmr )
(7.3.2)
d q m r
where Rr is the linearized resistance defined by (7.3.1).
p^

Figure 7.3.1 General fluid

resistance relation and its
linear approximations.

Rr
1

p
pr

qm

qmr

q^m

408

CHAPTER 7

Fluid and Thermal Systems

Referring to Figure 7.3.1, we define a new set of variables p and qm , called deviation
variables, that represent small but finite changes in p and q m from their reference values
pr and qmr . From Figure 7.3.1, we see that
p = p p r

(7.3.3)

qm = q m q mr

(7.3.4)

In terms of these deviation variables, we can rewrite (7.3.2) as

p p r = Rr (q m q mr )
or p = Rr qm , or
p
(7.3.5)
Rr
which is a linear relation. Thus the resistance Rr is called the linearized resistance.
Thus we may think of the linearized relation (7.3.5) as
small pressure change
= resulting small change in mass flow rate
Rr
The values of pr and qmr depend on the particular application. So the resistance Rr
depends on these values, as well as the functional form of f (q m ), which depends on
the application.
In a limited number of cases, such as pipe flow under certain conditions, the relation
of p versus q m is linear so that p = R q m , or
p
(7.3.6)
q m =
R
where R is the linear resistance.
In some other applications the relation is a square-root relation.

p
q m =
(7.3.7)
B
where B is a constant that often must be determined empirically.
The relation (7.3.7) gives p = B q m2 and thus


d p
Rr =
= 2Bqmr
d q m r
and
p
qm =
(7.3.8)
2Bqmr
This is the linearized model corresponding to the relation (7.3.7).
This relation can be

expressed instead in terms of pr as follows. Because qmr = pr /B, we obtain



Rr = 2B pr /B = 2 Bpr
qm =

and thus


1

p
(7.3.9)
qm =
2 Bpr
When only the curve of p versus q m is available, we can obtain a linearized model by
graphically computing the slope S of the tangent line that passes through the reference
point ( p r , q mr ). The equivalent, linearized resistance Rr is the slope S.
The resistance symbol shown in Figure 7.3.2 represents all types of fluid resistance,
whether linear or not. Although the symbol looks like a valve, it can represent fluid
resistance due to other causes, such as pipe wall friction and orifices.

7.3

qm
p1

R1

R2
p2

qm

qm

p3

p1

Fluid Resistance

Figure 7.3.2 Combination of

(a) series resistances and
(b) parallel resistances.

p3

R 5 R1 1 R2
(a)
R1 qm1
qm

qm

R
qm
1 51 11
R
R1 R2

R2 qm2
qm 5 qm1 1 qm 2
(b)

As with electrical resistances, linear fluid resistance elements obey the series and
parallel combination rules. These are illustrated in Figure 7.3.2. Series fluid resistances
carry the same flow rate; parallel fluid resistances have the same pressure difference
across them.

7.3.1 LAMINAR PIPE RESISTANCE

Fluid motion is generally divided into two types: laminar flow and turbulent flow.
Laminar flow can be described as smooth in the sense that the average fluid particle
velocity is the same as the actual particle velocity. If the flow is rough, the average
particle velocity will be less than the actual particle velocity, because the fluid particles
meander while moving downstream. This is turbulent flow. You can see the difference
between laminar and turbulent flow by slightly opening a faucet; the flow will be smooth.
As you open the faucet more, eventually the flow becomes rough.
If the pipe flow is laminar, the following linear relation applies.
q m =

p
R

(7.3.10)

or equivalently
p
(7.3.11)
R
The laminar resistance for a level pipe of diameter D and length L is given by the
Hagen-Poiseuille formula
qm =

R=

128L
D 4

(7.3.12)

where is the fluid viscosity. The viscosity is a measure of the stickiness of the fluid.
Thus molasses has a higher value of than that of water.
Not all pipe flow is laminar. A useful criterion for predicting the existence of
laminar flow is the Reynolds number Ne , the ratio of the fluids inertial forces to the
viscosity forces. For a circular pipe,
Ne =

v D

409

(7.3.13)

410

CHAPTER 7

Fluid and Thermal Systems

where v = qv /( D 2 /4), the average fluid velocity. For Ne > 2300 the flow is often
turbulent, while for Ne < 2300 laminar flow usually exists. The precise value of Ne
above which the flow becomes turbulent depends on, for example, the flow conditions
at the pipe inlet. However, the criterion is useful as a rule of thumb.
The resistance formula (7.3.12) applies only if the so-called entrance length L e ,
which is the distance from the pipe entrance beyond which the velocity profile no
longer changes with increasing distance, is much less than 0.06D Ne . Because laminar
flow can be expected only if Ne < 2300, L e might be as long as 138 pipe diameters.
Of course, for small Reynolds numbers, L e is shorter. The smaller L e is relative to the
pipe length, the more reliable will be our resistance calculations.

7.3.2 TORRICELLI'S PRINCIPLE

An orifice can simply be a hole in the side of a tank or it can be a passage in a valve. We
saw an example of orifice flow in Example 1.5.2 in Chapter 1, in which we analyzed
the flow rate of water through a small hole in the side of a plastic milk bottle. We found
that the fitted function is f = 9.4h 0.558 , where f is the outflow rate in ml/s and the water
height h is in centimeters. It turns out that the empirically determined exponent 0.558
is close to its theoretical value of 0.5, as we will now demonstrate.
Around 1640 Torricelli discovered that the flow rate through an orifice is proportional to the square root of the pressure difference. This observation can be simply
derived by considering a mass m of fluid a height h above the orifice (see Figure 7.3.3).
The potential energy of the mass is mgh. As the mass falls toward the orifice its
potential energy is converted to kinetic energy mv 2 /2. If all the potential energy is
converted to kinetic energy at the orifice, then mgh =mv 2 /2, and the maximum speed
the fluid mass can attain through the orifice is v = 2gh. Because the pressure
drop
across the orifice is p = gh, we can express the maximum speed as v = 2 p/.
Thus the massflow rate qm through the orifice of area Ao can be no greater than
Ao v = Ao 2 p/ = Ao 2 p. The actual flow rate will be less than this value
because of friction effects. To account for these frictional effects, we introduce a factor
Cd in the flow rate equation as follows:

q m = Cd Ao 2 p
(7.3.14)
The factor Cd is the discharge coefficient, which must lie in the range 0 < Cd 1. A
typical value for water is 0.6.
Because p = gh, (7.3.14) can be expressed in terms of the volume flow rate q v
and the height h as follows:


q v = Cd Ao 2g h 0.5
Figure 7.3.3 Derivation of
Torricelli's principle.

PE 5 mgh, KE 5 0

h
2
PE 5 0, KE 5 m v
2

Ao

7.4

Dynamic Models of Hydraulic Systems

411

Thus the theoretical value of the exponent (0.5) is close to the value obtained in the
bottle experiment.
Equation (7.3.14) depends on the orifice area being small enough so that the pressure variation over the orifice area is negligible compared to the average pressure at
the orifice. For a liquid-level system with a circular orifice, this implies that the liquid
height above the orifice must be large compared to the orifice diameter.
The orifice relation (7.3.14) can be rearranged in the form of (7.3.7).

 
p
(7.3.15)
q m = Cd Ao 2 p =
Ro
where the orifice resistance is defined as
Ro =

1
2Cd2 A2o

(7.3.16)

7.3.3 TURBULENT AND COMPONENT RESISTANCE

For us, the practical importance of the difference between laminar and turbulent flow
lies in the fact that laminar flow can be described by the linear relation (7.3.5), while
turbulent flow is described by the nonlinear relation (7.3.7). Components, such as
valves, elbow bends, couplings, porous plugs, and changes in flow area resist flow
and usually induce turbulent flow at typical pressures, and (7.3.7) is often used to
model them. Experimentally determined values of B are available for common types
of components.

7.4 DYNAMIC MODELS OF HYDRAULIC SYSTEMS

In this section we consider a number of hydraulic system examples dealing with liquidlevel systems, dampers, actuators, pumps, and nonlinear systems.

7.4.1 LIQUID LEVEL SYSTEMS

In liquid-level systems energy is stored in two ways: as potential energy in the mass
of liquid in the tank, and as kinetic energy in the mass of liquid flowing in the pipe.
In many systems, the mass of the liquid in the pipes is small compared to the liquid
mass in the tanks. If the mass of liquid in a pipe is small enough or is flowing at a small
enough velocity, the kinetic energy contained in it will be negligible compared to the
potential energy stored in the liquid in the tank. If the kinetic energy of the liquid is
significant, more advanced fluid-flow theory is required. This is usually not the case
for the scope of applications considered here.

Liquid-Level System with an Orifice

Problem

The cylindrical tank shown in Figure 7.4.1 has a circular bottom area A. The volume inflow rate
from the flow source is q vi (t), a given function of time. The orifice in the side wall has an area
Ao and discharges to atmospheric pressure pa . Develop a model of h, the deviation of the liquid
height from a reference equilibrium value h r , assuming that h 1 > L.

E X A M P L E 7.4.1

412

CHAPTER 7

Fluid and Thermal Systems

q^vi

Figure 7.4.1 A liquid-level

system with an orifice.

pa
h
h1

hr

h^

Solution

If the inflow rate q vi is held constant at the rate qvir , the liquid level eventually becomes constant
at the height h r . Using the orifice flow relation (7.3.14), this height can be found from
qvir = Cd Ao

2(gh r )

or
1
hr =
2g

qvir
C d Ao

2

So if we are given qvir , we can determine h r , or vice versa.

Noting that h 1 = h + h r + L, the rate of change of liquid mass in the tank is
d( Ah 1 )
d(h + h r + L)
dh
= A
= A
dt
dt
dt
Conservation of mass implies that
A


dh

= q vi Cd Ao 2 p
dt

(1)

where the pressure drop across the orifice is

p = pa + g(h + h r ) pa = g(h + h r )
Therefore, equation (1) becomes
A


dh
= q vi Cd Ao 2g 2 (h + h r )
dt

Canceling gives the desired model.

A


dh
= q vi Cd Ao 2g(h + h r )
dt

or, because h = h + h r ,
A

d h
= q vi Cd Ao
dt


2g h

(2)

Note that the height L does not appear in the model because the liquid below the orifice does
not affect the pressure at the orifice.

7.4

Dynamic Models of Hydraulic Systems

413

Linearizing a Model

E X A M P L E 7.4.2

Problem

Consider the liquid-level system with an orifice, treated in Example 7.4.1. The model is given
by equation (2) of that example.

d h
= q vi q vo = q vi Cd Ao 2g h
dt

Consider the case where A = 2 ft2 and Cd Ao 2g = 6. Estimate the systems time constant for
two cases: (i) the inflow rate is held constant at q vi = 12 ft3 /sec and (ii) the inflow rate is held
constant at q vi = 24 ft3 /sec.
A

Solution

Substituting the given values, we obtain

2


d h
= q vi q vo = q vi 6 h
dt

(1)

When the inflow rate is held constant at the value q vir , the liquid height h reaches an equilibrium

value h r that can be found from the preceding equation by setting h = h r and d h/dt
equal to
2
zero. This gives 36h r = qvir .
The two cases of interest to us are (i) h r = (12)2
/36 = 4 ft and (ii) h r = (24)2 /36 = 16 ft.
Figure 7.4.2 is a plot of the outflow flow rate q vo = 6 h through the orifice as a function of the
The two points corresponding to h = 4 and h = 16 are indicated on the plot.
height h.
In Figure 7.4.2 two straight lines are shown, each passing through one of the points of
interest (h = 4 and h = 16) and having a slope equal to the slope of the curve at that point. The
general equation for these lines is
q vo






d q vo

= 6 h = 6 hr +
(h h r ) = 6 h r + 3h r1/2 (h h r )
d h r

30

Figure 7.4.2 Linearized

approximations of the
resistance relation.

3
6 h^ < 24 1 4 (h^ 2 16)

25

20

3
6 h^ < 12 1 2 (h^ 2 4)

6 h^ 15
10

0
0

12
h^

16

20

24

414

CHAPTER 7

Fluid and Thermal Systems

and is the same as a Taylor series expansion truncated after the first order term. Noting that
h r = h r , this equation becomes

q vo = 6

h r + 3h r1/2 (h h r )

(2)

For Case (i), this equation becomes

3
q vo = 12 + (h 4)
2
and for Case (ii),
3
q vo = 24 + (h 16)
4
These are the equations of the straight lines shown in the figure.
Noting that h h r = h, equation (2) can be expressed in the simpler form

q vo = 6

h r + (3h r1/2 )h

Substitute this into equation (1), and note that d h/dt

= dh/dt and q vi 6 h r = qvi , to obtain
dh
(3)
= qvi + (3h r1/2 )h
dt
This is the linearized model that is a good approximation of the nonlinear model (1) near the
reference height h r .

The time constant of the linearized model (3) is 2 h r /3, and is 4/3 sec. for h r = 4 and
8/3 sec. for h r = 16. Thus, for Case (i), if the input flow rate is changed slightly from its
equilibrium value of 12, the liquid height will take about 4(4/3), or 16/3, sec to reach its new
height. For Case (ii), if the input flow rate is changed slightly from its value of 24, the liquid
height will take about 4(8/3), or 32/3, seconds to reach its new height.
Note that the models time constant depends on the particular equilibrium solution
chosen

for the linearization. Because the straight line is an approximation to the q vo = 6 h curve, we
cannot use the linearized models to make predictions about the systems behavior far from the
equilibrium point. However, despite this limitation, a linearized model is useful for designing
a flow control system to keep the height near some desired value. If the control system works
properly, the height will stay near the equilibrium value, and the linearized model will be accurate.
2

E X A M P L E 7.4.3

Liquid-Level System with a Flow Source

Problem

The cylindrical tank shown in Figure 7.4.3 has a bottom area A. The total mass inflow rate from
the flow source is q mi (t), a given function of time. The total mass outflow rate q mo is not given
and must be determined. The outlet resistance R is the linearized resistance about the reference
condition (h r , qmir ). Develop a model of h, the deviation of the liquid height from the constant
reference height h r , where h = h r + h.
Figure 7.4.3 A liquid-level
system with a flow source.

q^mi

pa
h^

pa

q^mo

7.4

Dynamic Models of Hydraulic Systems

415

Solution

The total mass in the tank is m = Ah = A(h + h r ) and from conservation of mass
d[ A(h + h r )]
dh
dm
=
= A
= q mi q mo
dt
dt
dt
because , h r , and A are constants. Expressing q mi and q mo in terms of the deviation variables
qmi and qmo , we have
A

dh
= (qmi + qmir ) (qmo + qmor ) = (qmir qmor ) + (qmi qmo )
dt

Because the reference height h r is a constant, the outflow rate at equilibrium must equal the
inflow rate. Thus qmir qmor = 0, and the model becomes
A

dh
= qmi qmo
dt

(1)

Because R is a linearized resistance, then for small changes h in the height,

qmo =

1
1
[(gh + pa ) pa ] = gh
R
R

Substituting this into equation (1) gives the desired model:

A

dh
1
= qmi gh
dt
R

which can be rearranged as

R A dh
R
+h =
qmi
g dt
g
So the time constant is = R A/g. If qmi = 0 (so the inflow rate remains constant at qmir ),
then h will be essentially zero for t > 4 , which means that the height will return to nearly the
equilibrium height h r at that time.

Some engineers are helped by thinking of a fluid system in terms of an analogous

electric circuit, in which pressure difference plays the role of voltage difference, and
mass flow rate is analogous to current. A fluid resistance resists flow just as an electrical
resistor resists current. A fluid capacitance stores fluid mass just as an electrical capacitor
stores charge. Figure 7.4.4 shows an electric circuit that is analogous to the tank system
of Figure 7.4.3. The circuit model is
C

1
dv
= is v
dt
R

The input current i s is analogous to the inflow rate qmi , the voltage v across the capacitor
is analogous to the fluid pressure gh, and the electrical capacitance C is analogous to
the fluid capacitance A/g. It is a matter of personal opinion as to whether such analogies
help to understand the dynamics of fluid systems, and you should decide for yourself.
Always keep in mind, however, that we should not get too dependent on analogies for
developing models, because they might not always properly represent the underlying
physics of the original system.

Figure 7.4.4 Electric circuit

analogous to the hydraulic
system shown in Figure 7.4.3.

is

416

CHAPTER 7

Fluid and Thermal Systems

E X A M P L E 7.4.4

Liquid-Level System with a Pressure Source

Problem

The tank shown in cross section in Figure 7.4.5 has a bottom area A. A pressure source
p s = ps (t) + psr is connected through a resistance to the bottom of the tank, where ps (t)
is a given function of time. The resistances R1 and R2 are linearized resistances about the reference condition ( psr , h r ). Develop a model of h, the deviation of the liquid height from the
constant reference height h r , where h = h r + h.
pa

Figure 7.4.5 A liquid-level

system with a pressure source.

h^
pa

p^s

A
R1

pa
R2

Solution

The total mass in the tank is m = Ah = A(h + h r ), and from conservation of mass
d[ A(h + h r )]
dh
dm
=
= A
= q mi q mo
dt
dt
dt
or
dh
A
= (qmi + qmir ) (qmo + qmor ) = (qmi qmo ) + (qmir qmor )
dt
Because at the reference equilibrium, the outflow rate equals the inflow rate, qmir qmor = 0,
and we have
dh
A
= qmi qmo
(1)
dt
This is a linearized model that is valid for small changes around the equilibrium state.
Because the outlet resistance has been linearized,
1
gh
qmo =
[(gh + pa ) pa ] =
R2
R2
Similarly for the mass inflow rate, we have
1
1
qmi =
[( ps + pa ) (gh + pa )] =
( ps gh)
R1
R1
Substituting into equation (1) gives
gh
1
R1 + R2
1
dh
( ps gh)
=
ps g
h
=
A
dt
R1
R2
R1
R1 R2
This can be rearranged as
R1 R2 A dh
R2
+h =
ps
g(R1 + R2 ) dt
g(R1 + R2 )
The coefficient of dh/dt gives the time constant, which is = R1 R2 A/g(R1 + R2 ).

When a fluid system contains more than one capacitance, you should apply the conservation of mass principle to each capacitance, and then use the appropriate resistance
relations to couple the resulting equations. To do this you must assume that some pressures or liquid heights are greater than others and assign the positive-flow directions
accordingly. If you are consistent, the mathematics will handle the reversals of flow
direction automatically.

7.4

Dynamic Models of Hydraulic Systems

417

Two Connected Tanks

E X A M P L E 7.4.5

Problem

The cylindrical tanks shown in Figure 7.4.6a have bottom areas A1 and A2 . The total mass
inflow rate from the flow source is q mi (t), a given function of time. The resistances are linearized
resistances about the reference condition h 1r , h 2r , qmir . (a) Develop a model of the liquid heights
h 1 and h 2 . (b) Suppose the resistances are equal: R1 = R2 = R, and the areas are A1 = A and
A2 = 3A. Obtain the transfer function H1 (s)/Q mi (s). (c) Use the transfer function to solve for
the steady-state response for h 1 if the inflow rate qmi is a unit-step function, and estimate how
long it will take to reach steady state. Is it possible for liquid heights to oscillate in the step
response?
Solution

a.

Using deviation variables as usual, we note that

h 1 = h 1r + h 1
h 2 = h 2r + h 2
q mi = qmir + qmi
For convenience, assume that h 1 > h 2 . This is equivalent to assuming that the flow rate
from tank 1 to tank 2 increases. From conservation of mass applied to tank 1, we obtain
d[ A1 (h 1 + h 1r )]
d( A1 h 1 )
dh 1
=
= A1
= q 1m = (q1m + q1mr )
dt
dt
dt
From physical reasoning we can see that the two heights must be equal at equilibrium, and
thus q1mr = 0. Therefore,
A1

dh 1
= q1m
dt

Because R1 is a linearized resistance,

q1m =

g
(h 1 h 2 )
R1

q^mi
h^1

h^2

A1

A2

R1

R2
(a)

v1

R1

i1

v2
i2

C1

R2

C2

(b)

is

Figure 7.4.6 (a) Two

connected tanks.
(b) Analogous electric circuit.

418

CHAPTER 7

Fluid and Thermal Systems

So, after canceling on both sides, the model for tank 1 is

A1

g
dh 1
= (h 1 h 2 )
dt
R1

(1)

Similarly for tank 2,

d[ A2 (h 2 + h 2r )]
dh 2
d( A2 h 2 )
=
= A2
dt
dt
dt
Conservation of mass gives
dh 2
= q mi + q 1m q mo = (qmi + qmir ) + (q1m + q1mr ) (qmo + qmor )
dt
Recalling that q1mr = 0, we note that this implies that qmir = qmor , and thus
A2

dh 2
= qmi + q1m qmo
dt
Because the resistances are linearized, we have
A2

A2

b.

dh 2
g
g
= qmi +
(h 1 h 2 )
h2
dt
R1
R2

(2)

The desired model consists of equations (1) and (2).

Substituting R1 = R2 = R, A1 = A, and A2 = 3A into the differential equations and
dividing by A, and letting B = g/R A we obtain
h 1 = B(h 1 h 2 )
3h 2 =

qmi
qmi
+ B(h 1 h 2 ) Bh 2 =
+ Bh 1 2Bh 2
A
A

Apply the Laplace transform of each equation, assuming zero initial conditions, and
collect terms to obtain
(s + B)H1 (s) B H2 (s) = 0
B H1 (s) + (3s + 2B)H2 (s) =

1
Q mi (s)
A

(3)
(4)

Solve equation (3) for H2 (s), substitute the expression into equation (4), and solve for
H1 (s) to obtain
R B 2 /g
H1 (s)
(5)
= 2
Q mi (s)
3s + 5Bs + B 2
c.

The characteristic equation is 3s 2 + 5Bs + B 2 = 0 and has the two real roots

5 13
B = 1.43B, 0.232B
s=
6
Thus the system is stable, and there will be a constant steady-state response to a step input.
The step response cannot oscillate because both roots are real. The steady-state height can
be obtained by applying the final value theorem to equation (5) with Q mi (s) = 1/s.
h 1ss = lim s H1 (s) = lim s
s0

s0

3s 2

R B 2 /g
R
1
=
+ 5Bs + B 2 s
g

The time constants are

1 =

1
0.699
=
1.43B
B

2 =

1
4.32
=
0.232B
B