Geared systems

Important examples of equivalent systems are given by devices

called mechanical transformers. Such devices as levers, gears,
cams, and chains transform the motion at the input into a related
motion at the output. Their analysis is simplified by describing
the effects of their mass, elasticity, and damping relative to a
single location, such as the input location. The result is a
lumped-parameter model whose coordinate system is centered at
that location.
We can characterie a transformer by a single parameter b such
that in terms of the general displacement variable Q,
Q
1
= bQ
2
!".#-$%&
where the subscripts $ and " refer to the input and output states,
respectively. Since the rate variable r is related to Q by dQ/dt=r,
we also have
r
1
=br
2
!".#-$'&
if b is constant. (n ideal transformer neither stores nor
dissipates energy. Therefore, the output power equals the input
power, or e
1
r
1
= e
2
r
2
. This shows the effort variables to be
related as
" $
$
e
b
e
!".#-$)&
( specific example is the gear pair shown in Table ".#, with
Q
,
T e
, and
r
. The parameter b is the gear ratio N. If
N>1, the pair is a speed reducer whose output torque T
2
is
greater than the input torque T
1
.
Suppose an element possessing inertia, elasticity, or damping
properties is connected to the output shaft of a gear pair, as
shown in Table ".#. *ow can we represent the geared system by
an equivalent gearless system+ If we wish to reference all
motion to the input shaft, we can proceed as follows. We
assume that the inertia, twisting, and damping of the input shaft
are negligible. If not, they can be included in a manner similar to
the following. We derive an equivalent system by requiring its
,inetic energy, potential energy, and power dissipation terms to
be identical to those of the original system. The expressions for
these quantities in the equivalent system are
83
Table ".# -ear pair relations
84
,inetic energy .
"
$
"
$

e
I
potential energy .
"
$
"
$

e
k
power dissipation .
"
$ $ $ $ $
& !
e e
c c T
If we neglect the gear inertias, gear tooth elasticity, and gear
friction, the corresponding expressions for the original system
are
,inetic energy .
"
"
"
$
I
potential energy .
"
"
"
$
k
power dissipation .
"
" " "
c T
/quating these expressions and using the gear ratio formulas, we
obtain the parameters for the equivalent system
I
N
I I
e
"
"
$
"
$
& !

!".#-$0&
k
N
k k
e
"
"
$
"
$
& !

!".#-"1&
c
N
c c
e
"
"
$
"
$
& !

!".#-"$&
The next two examples illustrate the lumping process.
System with three inertias
2or systems with more than one inertia on a single shaft, it is
frequently possible to add the inertias to produce an equivalent
system. ( pump is presented in 2igure "."3. The pump impeller
I
2
is driven by the shaft I
3
connected to an electric motor
producing a torque T
1
on its rotor I
1
. The viscous friction of the
fluid acting on the impeller is represented by the damper. (
model is desired for the behavior of the angular displacement "

due to the applied torque T
1
. 4iscuss how a lumped-parameter
model can be developed !2igure "."3b&.
85
2igure "."3 System with three inertias. !a& 5riginal system. !b&
6umped-parameter equivalent system.
( good initial assumption might be to neglect the twist " $


between the inertias I
1
and I
2
. This corresponds to neglecting the
potential energy in the shaft and implies that " $

and " $

.
/quating the ,inetic energy of the original and equivalent
systems gives
"
$ #
"
$ "
"
$ $
"
"
$
"
$
"
$
"
$
I I I I + +
With a coordinate reference on the motor shaft, $

, and thus,
# " $
I I I I + +
!".#-""&
The equivalent system model is given by
dt
d
c T
dt
d
I


$
"
"
or, since
dt
d
$
$


$ $
$

c T
dt
d
I
!".#-"#&
If the damping force is large or if the torque T
$
is applied
suddenly, the shaft twist will become important, and a model
incorporating shaft elasticity will be needed.
Geared pump-motor system
86
Suppose that the pump motor drives the impeller through a gear
pair !2igure "."7&.
2igure "."7 -eared system.
(ssume also that the gear and shaft inertias are significant.
8eglect the elasticity of the gear teeth and shafts, and assume
the friction and bac,lash in the gears are small.
4evelop a lumped-inertia model of the impeller velocity as a
function of the motor torque.
The angular displacements of the gears are related through the
gear ratio 8 such that
C B
N
!".#-"3&
B
C
D
D
N
8ote that a positive displacement B

produces a positive
displacement C

. Thus a negative sign is not needed in !".#-"3&

even through the directions of rotation are opposite. 2rom this,
we also see that C B
N
.
With negligible shaft elasticity, the twist is ero, and thus,
D C
D C
B A
B A

and
87
D A
N
(lso, we can use the results from !".#-""& to obtain I
1
, the
lumped inertia on shaft AB, as follows9
B AB A
I I I I + +
$
Similarly, the lumped inertia on shaft :4 is
D CD C
I I I I + +
"
2or a single-inertia model, we must reference all quantities to
one shaft. 6et us choose the motor shaft AB as the reference.
Then, the inertia equivalent to I
2
referenced to shaft AB is
"
"
"
$
I
N
I
e

The total equivalent inertia I is the sum of all the inertias
referenced to shaft AB, or
" $ e
I I I +
!".#-"7&
The viscous damping acting on shaft :4 has an equivalent value
on shaft (; of
c
N
c
e
"
$

!".#-"%&
Therefore, the desired model is
A e A
A
c T
dt
d
I

!".#-"'&
With I and c
e
given earlier.
Suppose that we had chosen shaft :4 as the reference shaft. (
good exercise is to use energy equivalence to derive the
following model.
D A
D
c NT
dt
d
I N I

+ & !
"
"
$
!".#-")&
5f course, !".#-"'& and !".#-")& are equivalent, as can be
verified substituting D A
N
into !".#-"'&.
This model can be used to compute the motor torque required to
maintain the load at some desired constant speed. 2or constant
speed, !".#-"'& shows the required torque to be
D e A e A
N c c T
!".#-"0&
Rotational-translational systems
There are many mechanisms which involve the conversion of
rotational motion to translational motion or vice versa. 2or
88
example, there are rac,-and-pinion, shafts with lead screws,
pulley and cable systems, etc.
To illustrate how such systems can analyed, consider a rac,-
and-pinion system shown in 2igure "."%.
2igure "."% <ac,-and-pinion system
The rotational motion of the pinion is transformed into
translational motion of the rac,. :onsider first the pinion
element. The net torque acting on it is (T
in
- T
out
). Thus,
considering the moment of inertia element, and assuming
negligible damping,
dt
d
I T T
out in


!".#-#1&
where I is the moment of inertia of the pinion and

its angular
velocity. The rotation of the pinion will result in a translational
velocity v of the rac,. If the pinion has a radius r, then
r v
.
*ence we can write
dt
dv
r
I
T T
out in

!".#-#$&
8ow consider the rac, element. There will be a force of T/r
acting on it due to the movement of the pinion. If there is a
frictional force of cv then the net force is
dt
dv
cv
r
T
out

!".#-#"&
89
/liminating T
out
from equations !".#-#1& and !".#-#"& gives
dt
dv
r
r
I
rcv T
in
& ! +
and so
& &! !
"
rcv T
r I
r
dt
dv
in

+

The result is a first-order differential equation describing how

the output is related to the input.
Electromechanical systems
/lectromechanical devices, such as potentiometers, motors and
generators, transform electrical signals to rotational motion or
vice versa. ( potentiometer has an input of a rotation and an
output of a potential difference. (n electric motor has an input
of a potential difference and an output of rotation of a shaft. (
generator has an input of rotation of a shaft and an output of a
potential difference.
!otentioeter
The rotary potentiometer, 2igure "."', is a potential divider and
thus9
max

"
"
o
where " is the potential difference across the full length of the
potentiometer trac, and max

is the total angle swept out by the

slider in being rotated from one end of the trac, to the other.
The output is "
out
for the input

.
2igure "."' <otary potentiometer.
90
DC otor
The 4: motor is used to convert an electrical input signal into a
mechanical output signal, a current through the armature coil of
the motor resulting in a shaft being rotated and hence the load
rotated !2igure ".")&.
2igure ".") 4: motor is driving a load.
The motor basically consists of a coil, the armature coil, which
is free to rotate. This coil is located in the magnetic field
provided by a current through field coils or permanent magnet.
When a current #
i
flows through the armature coil then, because
it is in a magnetic field, forces act on the coil and cause it to
rotate !2igure "."0&
2igure "."0 5ne wire of armature coil.
91
The force $ acting on a wire carrying a current i
#
and of length
% in a magnetic field of flux density B at right angles to the wire
is given by $=Bi
#
% and with N wires is $= Nbi
#
%. The forces on
the armature coil wires result in a torque T, where T = $b, with
b being the breadth of the coil. Thus
T=Nbi
#
%b
The resulting torque is thus proportional to Bi
#
, the other factors
all being constants. *ence we can write
T=k
1
Bi
#
Since the armature is a coil rotating in a magnetic field, a
voltage will be induced in it as a consequence of
electromagnetic induction. This voltage will be in such a
direction as to oppose the change producing it and is called the
bac, e.m.f. This bac, e.m.f. v
b
is proportional to the rate or
rotation of the armature and flux lin,ed by the coil, hence the
flux density B. Thus
B k v
b "

, where

is the shaft angular velocity and k

2
a
constant.
:onsider a 4: motor which has the armature and field coils
separately excited. With a so-called armature-controlled motor
the field current i
&
is held constant and the motor controlled by
ad=usting the armature voltage "
#
. ( constant field current
means a constant magnetic flux density B for the armature coil.
Thus

# "
k B k v
b

, where k
3
is a constant. The armature circuit can
be considered to be a resistance '
#
in series with an inductance
%
#
!2igure ".#1&.
92
2igure ".#1 4: motor circuits.
If v
#
is the voltage applied to the armature circuit then, since
there is a bac, e.m.f., we have
# #
#
# b #
i '
dt
di
% v v +
We can thin, of this equation in terms of the bloc, diagram
shown in 2igure ".#$a.
2igure ".#$ 4: motors9 !a& (rmature-controlled, !b& 2ield-
controlled.
The input to the motor part of the system is v
#
and this is
summed with the feedbac, signal of the bac, e.m.f. v
b
to give an
error signal which is the input to the armature circuit. The above
equation thus describes the relationship between the input of the
error signal to the armature coil and the output of the armature
current i
#
. Substituting for v
b
9
93
# #
#
# #
i '
dt
di
% k v +
#
The current i
#
in the armature generates a torque T. Since, for the
armature-controlled motor, B is constant we have
# #
i k Bi k T
3 $

where k
(
is a constant. This torque then becomes the input to the
load system. The net torque acting on the load will be
8et torque . T - damping torque
The damping torque is
c
, where c is a constant. *ence, if any
effects due to the torsional springiness of the shaft are neglected,
8et torque .
c i k
#

3
This will cause an angular acceleration of
dt
d
, hence

c i k
dt
d
I
#

3
We thus have two equations that describe the conditions
occurring for an armature-controlled motor, namely
# #
#
# #
i '
dt
di
% k v +
#
and

c i k
dt
d
I
#

3
We can thus obtain the equation relating the output

with the
input v
#
to the system by eliminating i
#
.
With a so-called field-controlled motor the armature current is
held constant and the motor controlled by varying the field
voltage. 2or the field circuit !2igure ".#1& there is essentially
=ust inductance %
&
in series with a resistance '
&
. Thus for that
circuit
dt
di
% i ' v
&
& & & &
+
We can thin, of the field-controlled motor in terms of the bloc,
diagram shown in 2igure ".#$b. The input to the system is v
&
.
The field circuit converts this into a current i
&
, the relationship
between v
&
and if being the above equation. This current leads to
the production of a magnetic field and hence a torque on the
armature coil, as given by #
Bi k T
$

. ;ut the flux density B is

proportional to the field current i
&
and i
#
is constant, hence
& #
i k Bi k T
7 $

, where k
)
is a constant. This torque output is then
converted by the load system into an angular velocity

. (s
earlier, the net torque acting on the load will be
94
8et torque . T - damping torque
The damping torque is
c
, where c is a constant. *ence, if any
effects due to the torsional springiness of the shaft are neglected,
8et torque .
c i k
&

7
This will cause an angular acceleration of
dt
d
, hence

c i k
dt
d
I
&

7
The conditions occurring for a field-controlled motor are thus
described by the equations
dt
di
% i ' v
&
& & & &
+
and

c i k
dt
d
I
&

7
We can thus obtain the equation relating the output

with the
input v
&
to the system by eliminating i
&
.
95
2.3. Hydraulic and liquid level systems
We can model fluid behavior as incompressible, which means
that the fluid>s density remains constant despite changes in the
fluid pressure. If the density changes with pressure, the fluid is
compressible. ;ecause all real fluids are compressible to some
extent, incompressibility is an approximation. ;ut this
approximation is usually sufficiently accurate for most liquids
under typical conditions, and it results in a simpler model of the
system. *ydraulics is the study of incompressible liquids, and
hydraulic devices use an incompressible liquid, such as oil, for
their wor,ing medium. 6iquid level systems consisting of
storage tan,s and connecting pipes are a class of hydraulic
systems whose driving force is due to relative differences in the
liquid heights in the tan,s.
*ere, we will avoid complex system models by describing only
the gross system behavior instead of the details of the fluid
motion patterns. To do this, we need only one basic physical
law9 conservation of mass. 2or incompressible fluids,
conservation of mass is equivalent to conservation of volume,
since the fluid density is constant. If we ,now the density and
the volume flow rate, we can compute the mass flow rate. That
is,
* *

, where *

and * are the mass and volume flow rates,

and

is the fluid density.

Basic principles
The primary variables for hydraulic systems are pressure, mass,
and flow rate.
2luid resistance is the constitutive relation between pressure and
mass flow rate.
2luid capacitance is the constitutive relation between pressure
and mass.
Integral causality relates mass m to mass flow rate, because

dt *
!".#-$&
96
:onservation of mass can be stated as follows. 2or a container
holding a mass of fluid , the time rate of change of mass d/dt
in the container must equal the total mass inflow rate minus the
total mass outflow rate. That is,
o i
* *
dt
d

!".#-"&
where *
i
is the inflow rate and *
o
is the outflow rate. The fluid
mass m is related to the container volume " by
"
!".#-#&
2or incompressible fluid,

is constant, and
dt
d"
dt
d

. 6et *
1

and *
2
be the total volume inflow and outflow rates. Thus,
$
* *
i

and "
* *
o

. Substituting these relationships into !".#-"&
gives
" $
* *
dt
d"

or
" $
* *
dt
d"

!".#-3&
This is a statement of conservation of volume for the fluid, and
it is equivalent to conservation of mass, equation !".#-"&.
2luid capacitance and resistance
The tan, shown in 2igure ".#" illustrates these concepts. 6et A
be the surface area of the tan,>s bottom. If the tan,>s sides are
vertical, the liquid height + is related to m by
A+
!".#-7&
In analogy with the electrical capacitance relation, v=Q/C,
pressure plays the role of voltage, and mass m plays the role of
charge. Thus, the fluid capacitance relation is
C

!".#-%&
97
where : is the fluid capacitance of the system. The hydrostatic
pressure due to the height + is
,+ -
!".#-'&
2igure ".#" ( liquid-level system.
where , is the acceleration due to gravity.
:omparing !".#-7&-!".#-'&, we see that
A
,
-
!".#-)&
Thus, the fluid capacitance of the tan, in 2igure ".#" is C=A/,.
When the container does not have vertical sides, the relations
between m and h, and between p and m are not linear. In this
case, we define the capacitance to be the slope of the m versus p
curve, and thus there is no single value for the container>s
capacitance.
With reference to 2igure ".#", if atmospheric pressure pa exists
at both the liquid>s surface and the pipe outlet, the pressure
difference across the pipe is !-? -
#
& --
#
.-. The outflow rate *
o

should obviously depend on - somehow. The greater the
pressure -, the greater the outflow rate. 2or now, let us assume
that the relation between *
o
and - is linear. In analogy with the
98
electrical resistance relation, i=v/r, the linear fluid resistance
relation is written as
'
-
*
o

!".#-0&
where are is the fluid resistance.
Liquid heiht model
(ssume that we are given *
i
as a function of time and that we
want a model for the behavior of the height + for the system in
2igure ".#".
@sing !".#-"&, !".#-7&, !".#-'&, and !".#-0&, we obtain
+
'
,
*
dt
d+
A
i


!".#-$1&
If we write this equation in terms of the volume inflow rate $
*
,
the density

can be divided out of the equation to obtain

+
'
,
*
dt
d+
A
$ !".#-$$&
!nterconnected storae elements
When a hydraulic system contains more than one storage
element, apply the conservation of mass equation !".#-"& to each
element. Then use the appropriate resistance relations to couple
the resulting equations. It is necessary to assume that some
pressures 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.
99
./#-0e 2.1
4evelop a model for the heights +
1
and +
2
in the liquid level
system shown in 2igure ".##a. The input volume rate * is given.
(ssume that laminar flow exists in the pipes. The laminar
resistances are '
1
and '
2
, and the bottom areas of the tan,s are
A
1
and A
2
. (lso draw the system bloc, diagram.
Solution
(ssume that +
1
>+
2
so that the mass flow rate *
1
is positive if
going from tan, $ to tan, ". :onservation of mass applied to
each tan, gives
2igure ".## ( system of two coupled tan, elements and its
bloc, diagram.
2or tan, $
& !
" $
$
$
$
$
$
+ +
'
,
*
* *
dt
d+
A

100
2or tan, ",
"
"
$
"
"
+
'
,
*
* *
dt
d+
A
o
o

Substituting for *
1
and *
o
and dividing by

gives the desired

model.
& !
" $
$
$
$
+ +
'
,
*
dt
d+
A
!".#-$"&
"
"
" $
$
"
"
& ! +
'
,
+ +
'
,
dt
d+
A
!".#-$#&
The density

does not appear in the final model because of the

incompressibility assumption. The bloc, diagram shown in
2igure ".##b is easily constructed from the preceding equations.
101
2." Hydraulic devices
*ydraulic devices use an incompressible liquid, such as oil, for
the wor,ing medium. They are widely used with high pressures
to obtain large forces in control systems, allowing large loads to
be moved or high accelerations to be obtained. *ere, we analye
three devices commonly found in hydraulic systems9 !$&
dampers limit the velocity of mechanical elementsA !"& hydraulic
servomotors produce motion from a pressure sourceA !#&
accumulators reduce fluctuations in pressure or flow rates.
# $luid damper
( damper or dashpot is a mechanical device that exerts a force
as a result of a velocity difference across it. 5ne common design
relying on fluid friction is shown in figure ".#3, and it is similar
to that used in automotive shoc, absorbers.
2igure ".#3 *ydraulic damper
( piston of diameter 4 and thic,ness 6 has a cylindrical hole of
diameter d drilled through it. The housing is assumed to be
stationary here, but the results can easily be generalied to
include a movable housing. The piston rod extends out of the
housing, which is sealed and filled with a viscous
incompressible fluid, such as an oil. ( force f acting on the
piston rod creates a pressure difference !p$-p"& across the piston
such that if the acceleration is small,
102
! A - - A & & !
" $ !".3-$&
The net cross-sectional piston area is
" "
&
"
! &
"
!
d D
A
2or a very viscous fluid at relatively small velocities, laminar
flow can be assumed, and the *agen-Boiseuille law can be
applied to give the volume flow rate
- C -
%
d
*
$
3
$")

!".3-"&
The volume flow rate * is also given by
dt
d/
A *
!".3-#&
because the fluid is incompressible. Substituting !".3-$& and
!".3-"& into !".3-#& shows that
dt
d/
c &
!".3-3&
where the damping coefficient is !from Tables&
"
"
$
"
$ )
1
1
]
1

,
_

d
D
%
C
A
c
!".3-7&
2rom the principle of action and reaction, & can also be
considered to be the force exerted by the damper when moving
at a velocity
dt
d/
. If the housing is also in motion,
dt
d/
is ta,en to
be the relative velocity between housing and piston.
If the relative velocity is large enough to produce turbulent flow
in the passage, the nonlinear resistance formulas should be used.
8ote that the fluid inertia, or ,inetic-energy storage
compartment, has been assumed to be negligible and that
potential energy resulting from pressure difference acts against
the fluid resistance to create the system>s dynamics. ( similar
analysis can be performed for air dampers in which the wor,ing
fluid is air. The common storm door damper is such a device.
#ccumulators
103
The compliance of fluid elements can have detrimental or
beneficial effects. In hydraulic control systems operating with
high force levels or at high speeds, the accuracy of the controller
can be affected by the tendency of the nominally incompressible
fluid medium to compress slightly. This compliance effect is
greatly increased by the presence of bubbles in the hydraulic
fluid, which ma,es the control action less positive and causes
oscillations.
5n other hand, compliance can be useful in damping
fluctuations or surges in pressure or flow rate, such as those
resulting from reciprocating pumps and the addition or removal
of other loads on common pressure source. In such cases, a
compliance element called an accumulator is added to the
system. This general term includes such different elements as a
surge tan, for liquid level system, a fluid column acting against
an elastic element for hydraulic systems, and a bellows for gas
systems. They all have the property of compliance, which
allows them to store fluid during pressure pea,s and release
fluid during low pressure. The result is more efficient and safe
operation by damping out pressure pulses the way a flywheel
smoothes the acceleration in mechanical drive or an electrical
capacitor smoothes or CfiltersC voltage ripples.
Two typical arrangements for hydraulic systems are shown in
2igure ".#7. 2igure ".#7a, a plate and spring oppose the pressure
of the liquid, and in 2igure ".#7b, the compliance is provided by
a gas-filled bag that compresses as the liquid pressure increases.
The simplest arrangement is a closed tube with air at the top.
The air>s compressibility acts against the liquid. This device is
the common cure for the water-hammer effect in household
plumbing.
(ll of the preceding devices use an equivalent elasticity to
oppose the pressure, and thus they can be modeled similarly. If
we ta,e the displacement / of the interface to be proportional to
the force & due to the liquid pressure, then
k/ &
, where k is the
elastic constant of the device.
6et A be the area of the interface, - the liquid pressure, and *
1
,
*
2
the volume flow rates into and out of the section.
104
2igure ".#7 *ydraulic accumulators. !a& Dechanical spring
principle. !b& :ompressible gas principle.
8eglect the pressure drop between the flow =unction and the
plate. 2rom mass !volume& conservation,
" $
& ! * *
dt
d/
A A/
dt
d

8eglecting the inertia of the interface gives the force balance
k/ -A
The model for dynamic behavior of the pressure as a function of
the flow rates is obtained by combining the last two expressions.
" $
"
* *
dt
d-
k
A

!".3-'&
When the resistances upstream and downstream are defined, the
flow rates *
1
and *
2
can be expressed as functions of the
upstream and downstream pressures.
(ssume that the laminar resistances in the left-and right-hand
pipe are '
1
and '
2
. Then,
& !
$
$
$
$
- -
'
*
!".3-)&
& !
$
$
"
"
- -
'
*
!".3-0&
105
and !".3-'& becomes
& !
$
& !
$
"
"
$
$
"
- -
'
- -
' dt
d-
k
A

!".3-$1&
This equation can be solved for - as a function of time if we are
given -
1
and -
2
as time functions.
%he hydraulic servomotor
( hydraulic servomotor is shown in 2igure ".#%.
2igure ".#% *ydraulic servomotor.
The pilot valve controls the flow rate of the wor,ing medium to
the receiving unit !a piston&. The pilot valve shown is ,nown as
a spool-type valve because of its shape. 2luid under pressure is
available at the supply port. 2or the pilot valve position shown
!the Cline-on-lineC position&, both cylinder ports are bloc,ed and
no motion occurs. When the pilot valve is moved to the right,
the fluid inters the right-hand piston chamber and pushes the
piston to the left. The fluid displaced by this motion exits
through the left-hand drain port. The action is reversed for a
valve displacement to the left. ;oth drain ports are connected to
a tan, !a sump& from which the pump draws fluid to deliver to
the supply port. The filters and accumulators necessary to clean
106
the recirculated fluid and dampen pressure fluctuations are not
shown.
6et 1 denote the displacement of the pilot valve from its line-on-
line position and / the displacement of the load and piston from
their last position before the start of the motion. 8ote that a
positive value of / !to the left& results from a positive value of 1
!to the right&. The flow through the cylinder port uncovered by
the pilot valve can be treated as flow through an orifice, and the
orifice flow relation can be applied !from tables&. 6et - be the
pressure drop across the orifice. Thus,
E "- A C *
o d

!".3-$$&
where * is the volume flow rate through the cylinder port, A
o
is
the uncovered area of the port, C
d
is the discharge coefficient
!which usually lies between 1.%-1.) for this application&, and

is the mass density of the fluid. The area A
o
is equal to 21, where
2 is the port width !into the page&. If C
d
,

, -, and 2 are ta,en

to be constant, !".3-$$& can be written as
*=C1 !".3-$"&
:onservation of mass requires the flow rate into the cylinder to
equal the flow rate out. Therefore,
dt
d/
A *
:ombining the last two equations gives the model for the
servomotor.
1
A
C
dt
d/

!".3-$#&
2or high accelerations, this model must be modified, because it
neglects the inertia of the load and piston. The piston force was
assumed sufficient to move the load and is generated by the
pressure difference across the piston.
(lthough !".3-$#& is accurate enough for many applications, we
will now develop a more detailed model to account for the
effects of load. 2igure ".#' gives the necessary details. Dachine
tools for cutting metal are one application for such a system.
107
2igure ".#' *ydraulic servomotor with load.
The applied force & can be supplied by a hydraulic servomotor.
The represents the mass of a cutting tool and the power
piston, while k represents the combined effects of the elasticity
naturally present in the structure and that introduced by the
designer to achieve proper performance. ( similar statement
applies to the tool through its prescribed motion.
The spool valve shown in 2igure ".#' has two lands. If the
width of the land is greater than the port width, the valve is said
to be overlapped !2igure ".#)a&. In this case, a dead one exists
in which a slight change in the displacement produces no
power piston motion. Such dead ones create control difficulties
and are avoided by designing the valve to be underlapped !the
land width is less than the port width-2igure ".#)b&. 2or such
valves there will be a small flow opening even when the valve is
in the neutral position !1=3&. This gives it a higher sensitivity
than the overlapped valve.
6et -
4
and -
o
denote the supply and outlet pressures,
respectively. Then the pressure drop from inlet to outlet is -
4
--
o
,
and this must equal the sum of the drops across both valve
openings and the power piston. That is,
- - - -
v o 4
+ "
108
where the drop across the valve openings is
o 4 v
- - - - -
$ "
and the drop across the piston is
$ "
- - -
Therefore,
& !
"
$
- - - -
o 4 v

2igure ".#) Spool valve construction. !a& 5verlapped.
!b& @nderlapped.
If we ta,e -
4
and -
o
to be constant, we see that v
-
varies only if
-
changes. The derivation of !".3-$#& treated v
-
as a constant.
Thus the variables 1 and
-
determine the volume flow rate, as
& , ! - 1 & *
109
This relation is nonlinear. 2or the reference equilibrium
condition !1=3,
-
.3, *=3&, a lineariation gives
- C 1 C *
" $ !".3-$3&
where the variations from equilibrium are simply 1,
-
, and *.
The lineariation constants are available from theoretical and
experimental results. The constant C
1
is identical to C in
!".3-$#&, and both C
1
and C
2
are positive. The effect of the
underlapping shows up in the nonero flow predicted from
!".3-$3& when 1=3. The pressure drop
-
is caused by the
reaction forces of the load mass, elasticity, and damping and by
the supply pressure difference !-
4
--
o
&.
./#-0e 2.2
!a& 4erive a model of the system shown in 2igure ".#'.
Ta,e into account the inertia effects of the load and
assume the valve underlapped.
!b& Show that the model reduces to !".3-$#& when
c=k=3 and either of the following approximations is valid9
$. /A

1 !large piston force compared to the load inertia&

". C
2

1 !q independent of
-
&
!a& (ssume an incompressible fluid. Then conservation of mass
and !".3-$3& give
- C 1 C *
dt
d/
A
" $
The force generated by the piston is
- A
, and from 8ewton>s
law,
- A k/
dt
d/
C
dt
/ d
+
"
"
where / is measured from the equilibrium point of the mass .
Substitute
-
from the second equation to obtain
& !
"
"
" $
/
A
k
dt
d/
A
c
dt
/ d
A

C 1 C
dt
d/
A + +
or
1 C /
A
k C
dt
d/
A
A
cC
dt
/ d
A
C
$
" "
"
"
"
& ! + + +
!".3-$7&
110
This is the desired model with 1 as the input and / as the output.
!c& 2or c=k=3, !".3-$7& becomes
!d&
1 C
dt
d/
A
dt
/ d
A
C
$
"
"
"
+
!".3-$%&
When either /A

1 or C
2

1, we obtain
1 C
dt
d/
A
$

!".3-$'&
This is equivalent to !".3-$#&, because C
1
=C.
111
2.& 'neumatic elements
while hydraulic devices use an incompressible liquid, wor,ing
medium in a pneumatic device is a compressible liquid, such as
air. Industrial control systems frequently have pneumatic
components to provide forces greater than those available from
electrical devices. (lso, they afford more safety from fire
haards. The availability of air is an advantage for these devices,
because it may be exhausted to the atmosphere at the end of the
device>s wor, cycle, thus eliminating the need for return lines.
5n the other hand, the response of pneumatic systems is slower
than that of hydraulic systems because of the compressibility of
the wor,ing fluid.
The analysis of gas flow is more complicated than for liquid
flow. 2or compressible fluids, the mass and volume flow rates
are not readily interchangeable. Since mass is conserved,
pneumatic systems analysis uses mass flow rate, here denoted
*

to distinguish from volume flow rate *. ;ecause of the

relatively low viscosity and density of gases, their flow is more
li,ely to be turbulent. 2inally, the possibility of supersonic
!greater than the speed of sound& flow is more li,ely in gas
systems.
Bressure is normally the only effort variable used for such
systems. The ,inetic energy of a gas is usually negligible, so the
inductance relation is not employed.
%hermodynamic properties o$ ases
Temperature, pressure, volume, and mass are functionally
related for a gas. The model most often used to describe this
relationship is the perfect gas law, which describes all gases if
the pressure is low enough and the temperature high enough.
The law states that
T ' -"
,

!".7-$&
where - is the absolute pressure of the gas with volume (, is
the mass, T its absolute temperature, and '
,
the gas constant that
depends on the particular type of gas.
112
The prefect gas law allows us to solve for one of the variables -,
", , or T if the other three are given. We frequently do not
,now the values of three of the variables and additional
information. 2or a perfect gas, this information is usually
available in the form of a pressure-volume or CprocessC relation.
The following process models are commonly used9 constant-
pressure, constant-volume, constant-temperature !isothermal&,
and reversible adiabatic !isentropic&. In the latter process, no
heat is transferred between the gas and its surroundings. These
four processes only approximate reality, but they allow some
modeling simplifications to be made. ( real process can be
more accurately described by deriving or measuring an
appropriate value for the exponent n in the polytropic process
equation
. & ! con4t

"
-
n

!".7-"&
If the mass is constant, the polytropic process describes the four
previous processes if n is chosen as 1,

, $, and

,
respectively, where

is the ratio c
-
/c
v
of the specific heats of
the gas.
'neumatic resistance
*ere we develop methods for determining pneumatic resistances
when the flow is compressible but subsonic.
The theory of gas flow through an orifice provides the basis for
modeling other pneumatic components. We assume that perfect
gas law applies and the effects of viscosity !friction& and heat
transfer are negligible, hence, an isentropic process is assumed.
:onsider an orifice with a cross-sectional area A. The upstream
absolute pressure is -
1
, and the absolute bac, pressure -
b
is that
pressure eventually achieved downstream from the orifice. (s -
b
is lowered below -
1
, the flow rate through the orifice increases.
If the difference between -
1
and -
b
becomes large enough, the
gas flows through the orifice at the speed of sound. The value of
-
b
at which this occurs is the critical pressure -
c
. 2or air,
3 . $

and the critical pressure is
113
$
7") . 1 - -
c

!".7-#&
The results of chief interest to us are the expressions for the
mass flow rate as functions of the pressures -
1
and -
b
. 8ewton>s
laws, conservation of mass, and the isentropic process formula
can be combined to show that the mass flow rate *

in the
subsonic case !-
b
>-
c
& is
& !
"
$
$
b b
,
d
- - -
T '
A C *
!".7-3&
where C
d
is an experimentally determined discharge coefficient
that accounts for viscosity effects and T
1
is the absolute
temperature upstream from the orifice. 2or the sonic case, *

is
independent of -
b
as long as -
b
5-
c
.
( typical condition in pneumatic systems is such that a small
pressure drop exists across the component. *ence, the flow is
frequently subsonic. (lso, the pressure changes often consist of
small fluctuations about an average or steady-state constant
pressure value. @nder these circumstances, the compressible
flow resistance of pneumatic components can be modeled in the
form of turbulent resistance relation, written here as
- * '
-

"
!".7-7&
where
-
is the pressure drop across the component and '
-
is
the pneumatic resistance.
To show that !".7-3& reduces to !".7-7&, we assume that the input
pressure -
1
fluctuates about a constant pressure -
4
by a small
amount -
i
. Thus, -
1
=-
4
6-
i
. Similarly, the bac, pressure can be
written as -
b
=-
4
6-
o
, and we assume that -
o
is also small relative
to -
4
. 8ote that -
1
--
b
=-
i
--
o
and consider the following term from
!".7-3&
o i 4 o i
4
o
4 o i o 4 o i o 4 b b
- - - - -
-
-
- - - - - - - - - - - - + + + & $ ! & &! ! & !
$
because $ ? p
o
Ep
s

$ from our assumptions. Thus, !".7-3&

becomes
114
o i
,
4
d
- -
T '
-
A C *
$
"
!".7-%&
This has the same form as !".7-7&, where b o i
- - - - -
$ , and
the pneumatic resistance is
"
"
$
" A C -
T '
'
d 4
,
-

!".7-'&
Solve !".7-7& for *

assuming that
1 > -
-

'
-
*

!".7-)&
where we ta,e the positive square root, because *

>3 if
1 > -
.
2or !".7-)&, the slope of *

is infinite at
1 -
. This physically
unrealistic result is due to certain assumptions made to derive
!".7-3& that are not true as $
- -
b

. 2or -
b
very close to -
1
, the
orifice flow relation resembles that of laminar flow, and will
model it as
-
'
*
%

$
!".7-0&
where '
%
is the equivalent resistance. If we match the solution of
!".7-)& and !".7-0& at some intermediate value of
-
, say,
-
.B, then '
%
is related to '
-
by
- %
B' '
!".7-$1&
where the value of B must be determined by experiment. 2or
this value of '
%
, the slopes of !".7-)& and !".7-0& also match at
-
.B.
Thus, we have three models for subsonic orifice flow9 !".7-3&
for
-
largeA !".7-)& for
-
small but greater than BA and !".7-
0& for
-
very small.
'neumatic capacitance
In pneumatic systems, mass is the quantity variable and pressure
the effort variable. Thus, pneumatic capacitance is the relation
115
between stored mass and pressure. Specifically, we ta,e
pneumatic capacitance to be the system>s capacitance C !or
compliance&, defined as the ratio of the change in stored mass to
the change in pressure, or
d-
d
C
!".7-$$&
2or a container of constant volume " with a gas density

, this
expression may be written as
d-
d
"
d-
" d
C


& !
!".7-$"&
If the gas undergoes a polytropic process,
con4t
-

"
-
n
n

& !
!".7-$#&
and
n-"

n- d-
d


2or a perfect gas, this shows the capacitance of the container to
be
T n'
"
n-"
"
C
,

!".7-$3&
8ote that the same container can have a different capacitance for
different expansion processes, temperatures, and gases, since C
depends on n, T, and '
,
.
116
)odelin pneumatic systems
;ecause fluid inertia is usually neglected in pneumatic analysis,
the simplest model of such systems is a resistance-capacitance
model for each mass storage element.
./#-0e 2.3
(ir passes through a valve !modeled as an orifice& into a rigid
container, as shown in 2igure ".#0. 4evelop a dynamic model of
the pressure change - in the container as a function of the inlet
pressure change -
i
. (ssume an isothermal process and that - and
-
i
are small variations from the steady-state pressure -
4
.
Solution9
2rom conservation of mass, the rate of mass increase in the
container equals the mass flow rate through the valve.
2igure ".#0 -as flow into a rigid container.
Thus, if -
i
-
-
>3
dt
d-
C
dt
d-
d-
d
dt
d
*
dt
d

-
i

'
- -
*

!".7-$7&
where C and '
-
are given by !".7-$3& and !".7-'&, with T = T
1

and n=1. The desired model is
117
-
i
'
- -
dt
d-
C

!".7-$%&
If -
i
--53, the flow *

is reversed, and the model is

-
i
'
- -
dt
d-
C


!".7-$'&
If -
i
-- is very close to ero, the linear resistance relation !".7-0&
can be used, and !".7-$7& is replaced by
& !
$
- -i
'
*
%

!".7-$)&
In this case, the system model is
& !
$
- -
' dt
d-
C
i
%

!".7-$0&
and the flow reversal is accounted for automatically if -
i
--
changes sign.
./#-0e 2.(
( pneumatic bellows, shown in 2igure ".31, is an expandable
chamber usually made from corrugated copper because of this
metal>s good heat conduction and elastic properties. We model
the elasticity of the bellows as a spring. The spring constant k
for the bellows can be determined from vendor>s data or
standard formulas. This device, when employed with a variable
resistance '
-
is useful in pneumatic controllers as a feedbac,
element to sense and control the pressure -
i
. The displacement /
is transmitted by a lin,age or beam balance to a pneumatic valve
regulating the air supply. 4evelop a model for the dynamic
behavior of / with -
i
a given input.
118
2igure ".31 Bneumatic bellows.
We assume that the bellows expands or contracts slowly. Thus,
the product of its mass and acceleration is negligible, and a force
balance gives
k/ -A
!".7-"1&
where the force exerted by the internal pressure is -A, A is the
area of the bellows, and / is the displacement of the right-hand
side !the left-hand side is fixed&. The volume is "=A/, and thus
"
k/ A/
A
k/
-"
The assumption of a slow process suggests an isothermal
process. Thus, the time derivative of the perfect gas law gives
dt
d
T ' -"
dt
d
,
& !
2rom the law of resistance !".7-0&, assuming that -
i
-- is very
small,
& !
$
- -
'
*
dt
d
i
%

Since
dt
d/
k/
dt
k/ d
"
& !
"

these expressions give the following model9

& ! "
A
k/
-
'
T '
dt
d/
k/
i
%
,

!".7-"$&
This is nonlinear because of the cross-product between / and its
derivative.
119
2.* %hermal systems
In thermal systems, energy is stored and transferred as heat. 2or
this reason, many energy conversion devices are examples of
thermal systems, as are chemical processes where heat must be
added or removed in order to maintain an optimal reaction
temperature. 5ther examples occur in food processing and
environmental control in buildings.
The effort variable is temperature, or more precisely,
temperature difference
T
. This can cause a heat transfer rate
+
*
by one or more of the following resistance modes9
conduction, convection, or radiation. The transfer of heat causes
a change in the system>s temperature. Thermal capacitance
relates the system temperature to the amount of heat energy
stored. It is the product of the system mass and its specific heat.
8o physical elements are ,nown to follow a thermal inductance
causal path.
In order to have a single lumped-parameter model, we must be
able to assign a single temperature that is representative of the
system. Thermal systems analysis is sometimes complicated by
the fact that it is difficult to assign such a representative
temperature to a body or fluid because of its complex shape or
motion and resulting complex distribution of temperature
throughout the system. *ere, we develop some guidelines for
when such an assignment can be made. When it cannot, several
coupled lumped-parameter models or even a distributed-
parameter model will be required.
Heat trans$er
*eat can be transferred in three ways9 by conduction !diffusion
through a substance&, convection !fluid transport&, and radiation
!mostly infrared waves&. The effort variable causing a heat flow
is a temperature difference. The constitutive relation ta,es a
different form for each of the three heat transfer modes. The
linear model for heat flow rate as a function of temperature
difference is given by 8ewton>s law of cooling, which is valid
for both convection and conduction.
120
T '*
+

!".%-$&
The thermal resistance for convection is
+A
'
$

!".%-"&
where + is the film coefficient of the fluid-solid interface. 2or
conduction, we will see that
kA
d
'
!".%-#&
where k is the thermal conductivity of the material, A is the
surface area, and d the material thic,ness.
:onvective heat transfer is divided into two categories9 forced
convection-due, for example, to fluid pumped past a surface-and
free or natural convection, due to motion produced by density
difference in the fluid. The heat transfer coefficient for
convection is a complicated function of the fluid flow
characteristics especially.
Significant heat transfer can occur by radiation, the most notable
example is solar energy. Thermal radiation produces heat when
it stri,es a surface capable of absorbing it. It can also be
reflected or refracted, and all three mechanisms can occur at a
single surface. When two bodies are in visual contact, a mutual
exchange of energy occurs by emission and absorption. The net
transfer of heat occurs from the warmer to the colder body. This
rate depends on material properties, geometric factors affecting
the portion of radiation emitted by one body and stri,ing the
other, and the amount of surface area involved. The net heat
transfer rate can be shown to depend on the body temperatures
raised to the fourth power !a consequence of the Stefan-
;oltmann law&.
& !
3
"
3
$
T T *
+

!".%-3&
121
The absolute body temperatures are T
1
and T
2
, and

is a factor
incorporating the other effects. This factor is usually very small.
Thus, the effect of radiation heat transfer is usually negligible
compared to conduction and convection effects, unless the
temperature of one body is much greater than that of the other.
Some simpli$yin appro+imations
The thermal resistance for conduction through a plane wall
given by !".%-#& is an approximation that is valid only under
certain conditions. 6et us ta,e a closer loo, at the problem. If
we consider the wall to extend to infinity in both directions, then
the heat flow is one dimensional. If the wall material is
homogeneous, the temperature gradient through the wall is
constant under steady-state conditions !2igure ".3$a&.
2igure ".3$ 5ne-dimensional heat transfer through a wall. !a&
Steady-state temperature gradient. !b& 6umped approximation.
122
2ourier>s law of heat conduction states that the heat transfer rate
per unit area within a homogeneous substance is directly
proportional to the negative temperature gradient. The
proportionality constant is the thermal conductivity ,. 2or the
case shown in 2igure ".3$a, the gradient is (T
2
-T
1
)/d, and the
heat transfer rate is thus
d
T T kA
*
+
& !
$ "


!".%-0&
where ( is the area in question. :omparing this with !".%-$&
shows that the thermal resistance is given by !".%-#&, with
" $
T T T
.
In transient conditions where temperatures change with time, the
temperature distribution is no longer a straight line. If the
internal temperature gradients in the body are small, as would be
the case for a large value of k, it is possible to treat the body as
being at one uniform, average temperature and thus obtain a
lumped-instead of distributed-parameter model. 2or solid bodies
immersed in a fluid, a useful criterion for determining the
validity of the uniform-temperature assumption is based on the
;iot number, defined as
k
+%
N
B

!".%-%&
where % is the ratio of the volume to surface area of the body
and h is the film coefficient. If the shape of the body resembles a
plate, or sphere, it may be considered to have a uniform
temperature if N
B
is small. If N
B
53.1, the temperature is usually
ta,en to be uniform, and the accuracy of this approximation
improves if the inputs vary slowly.
123
./#-0e 2.)
:onsider a copper sphere $in in diameter with k."$" ;tuEhr-ft-
o
2 at 7'1
o
2 and immersed in a fluid such that +.7 ;tuEhr-ft
"
-
o
2.
Show that its temperature can be considered uniform, and
develop a model of the sphere>s temperature as a function of the
temperature T
o
of the surrounding fluid.
Solution9
The surface area and volume of the sphere are
" "
$1 $) . " &
"3
$
! 3

/ A
3 #
$1 1# . # &
"3
$
! &
#
3
!

/ "
Thus, % = "/A . 1.1$#0, and N
B
. #.")x$1
-3
, which is much less
than 1.$.
(ccording to the ;iot criterion, we may treat the sphere as a
lumped-parameter system with a single uniform temperature T.
The amount of heat energy in the sphere is CT, where C is its
thermal capacitance. If the outside temperature T
o
is greater than
T, heat is transferred into the sphere at the rate *
+
given by
!".%-$&, where
T T T
o

. Thus, from the conservation of
energy,
+
* CT
dt
d
& !
@sing !".%-$& and noting that C is constant, we obtain
& !
$
T T
' dt
dT
C
o

!".%-'&
2or the copper sphere, !".%-"& gives
C =c
-
.

"c
-
. $'.#!#.1#x$1
-3
&!".0#& .1.1$73
<.
$' . 0
& $1 $) . " ! 7
$
"

/
124
# wall model
:onductive and convective resistances to heat transfer
frequently occur when one or more layers of solid materials are
surrounded on both sides by fluid, such as shown in 2igure
".3$b. We will use the figure to show how thermal resistance is
computed when both conduction and convection are present.
(ssume that the wall>s material is homogeneous, and that it is to
be treated with a lumped approximation. The wall temperature is
considered to be constant throughout and is denoted by T

. The
temperatures T
1
and T
2
are the temperatures in the surrounding
fluid =ust outside the wall surface. There is a temperature
gradient on each side of the wall across a thin film of fluid
adhering to the wall. This film is ,nown as the thermal boundary
layer. The film coefficient h is a measure of the conductivity of
this layer. The temperatures outside of these layers are denoted
by T
i
and T
o
.
The thermal capacitance C is the product of the wall mass times
its specific heat. The capacitance of the boundary layer is
generally negligible because of its small fluid mass. Thus, since
no heat is stored in the layer, the heat flow rate through the layer
must equal the heat flow rate from the surface to the surface to
the mass at temperature T

considered to be located at the center

of the wall. The length of this latter path is d/2.
2or the left-hand side, this gives
& !
" E
& !
$ $ $ $ i +
T T
d
kA
T T A + *
!".%-)&
This allows us to solve for T
1
as a function of T
i
and T

.
$
$
$
"
"
d+ k
T d+ kT
T
i
+
+

!".%-0&
Similar equations can be developed for the right-hand flow *
+2

and temperature T
2
.
(n energy balance for the wall mass gives
125
" $ + +

* *
dt
dT
C
!".%-$1&
(fter T
1
and T
2
have been eliminated, this gives an equation for
T

with T
i
and T
o
as inputs. 4efine
" , $ ,
"
"

+
7
d+ k
A k+
#
7
7
7 !".%-$$&
and !".%-$1& becomes
& ! & !
" $ o i

T T # T T #
dt
dT
C
!".%-$"&
In !".0-$"&, #
1
is the reciprocal of the resistance for the path
from T
i
to T

. We can write this resistance as

kA
d
A + A k+
d+ k
#
'
"
$
"
" $
$ $
$
$
$
+
+

!".0-$#&
:omparing this with !".0-"& and !".0-#& shows that '
1
is the sum
of the thermal boundary layer resistance and the wall>s
conductive resistance along the path of length d/2.
In many applications, the wall>s thermal capacitance is small
compared to the capacitance of the mass of fluid on either side
of the wall. In this case, we model the wall as a pure resistance,
and the total wall resistance is the sum of the path resistances.
/quation !".%-$"& can be used to show this. If either the rate of
change of T

or the wall capacitance C is very small, the right-

hand side of !".%-$"& can be ta,en to be ero and thus *
+1
equals
*
+2
. This relation can be manipulated to show that the wall
temperature is
" $
" $
# #
T # T #
T
o i

+
+

!".%-$3&
and that the heat flow rate is
& !
" $
" $
o i +
T T
# #
# #
*
+

!".%-$7&
126
Thus, the total resistance between T
i
and To is !#
1
6#
2
&E#
1
#
2
.
With !".%-$$& and some algebra, the total resistance can be
shown to be
A + k
d
k
d
+
'
$
&
$
" "
$
!
" $
+ + +
!".%-$%&
When it is not possible to identify one representative
temperature for a system, several temperatures can be chosen,
one for each distinct mass. The resistance paths between the
masses are then identified, and conservation of heat energy is
applied to each mass. This results in a model whose order equals
the number of representative temperatures.
./#-0e 2.8
( simplified representation of temperature dynamics of a room
is shown in 2igure ".3".
2igure ".3" ( model of room temperature.
The room is perfectly insulated on all sides except one, which
has a wall with a thermal capacitance C
2
. The inner wall
resistance is '
1
, as given by !".%-$#&. The outer resistance is '
2

and is found by similar expression. The representative
temperature of the room>s air is designated T
i
, that of the wall is
T

, and the outside air temperature is T

o
. The room air
capacitance is C
1
. 4evelop a model of the behavior of T
i
.
127
Solution
(ssume that T
o
>T

>T
i
. Then the heat flow is from T
o
to T

to T
i
.
:onservation of energy applied to the capacitances C
1
and C
2

gives
& !
$
$
$ i
i
T T
' dt
dT
C
!".%-$'&
& !
$
& !
$
$ "
" i o

T T
'
T T
' dt
dT
C
!".%-$)&
This is the desired model. If the heat is not in the assumed
direction, the mathematics ta,es care of the reversal
automatically.
If the wall capacitance is negligible, then C
2
=3, and we can
substitute T

from !".%-$)& into !".%-$'&. 5r, equivalently, we

can use the principle of !".%-$%& to find the total wall resistance
'. This is
' = '
1
6 '
2
!".%-$0&
The heat flow rate through the wall is (T
o
-T
i
)/', and the model
becomes
& !
$
$ i o
i
T T
' dt
dT
C
!".%-"1&
128
./#-0e 2.9
;uild up a model for the thermal system shown in 2igure ".3#.
2igure ".3# ( thermometer-liquid thermal system.
:onsider a thermometer at temperature T which has =ust been
inserted into a liquid at temperature T
%
. If the thermal resistance
to heat flow from the liquid to the thermometer is ', then,
'
T T
*
%

where * is the net rate of heat flow from liquid to thermometer.

The thermal capacitance C of the thermometer is given by the
equation
dt
dT
C * *
" $
Since there is only a net flow of heat from the liquid to the
thermometer, *
1
=* and *
2
=3. Thus
dt
dT
C *
Substituting this value of q in the earlier equation gives
'
T T
dt
dT
C
%

<earranging this equation gives

129
%
T T
dt
dT
'C +
This equation, a first-order differential equation, describes how
the temperature indicated by the thermometer T will vary with
time when the thermometer is inserted into a hot liquid.
130
./#-0e 2.:
2igure ".33 shows a thermal system consisting of an electric fire
in a room. This fire emits heat at the rate *
1
and the room loses
heat at the rate *
2
. (ssuming that the air in the room is at a
uniform temperature T and that there is no heat storage in the
walls of the room, derive an equation describing how the room
temperature will change with time.
2igure ".33 /lectric fire-room thermal system.
If the air in the room has a thermal capacity C, then
dt
dT
C * *
" $
If the temperature inside the room is T and that outside the room
T
o
, then
'
T T
*
o

"
where ' is the resistivity of the walls. Substituting for *
2
gives
dt
dT
C
'
T T
*
o

$
*ence
o
T '* T
dt
dT
'C + +
$
131
Broblems
".$ 4erive the mathematical model for the following system.
The input is the force $, and the output is the displacement /.
"." 4erive an equation relating the input angular displacement

i
with the output angular displacement

o
".# 4erive the relationship between the output, the potential
difference across the resistor ' of v
'
, and the input v for the
series %C' circuit shown below.
132
".3 4erive the relationship between the height +
2
and time for
the hydraulic system shown below. 8eglect inertance.
".7 The following figure shows a thermal system involving
two compartments, with one containing a heater. If the
temperature of the compartment containing the heater is T
1
, the
temperature of the other compartment T
2
and the temperature
surrounding the compartments T
3
, develop equations describing
how the temperature T
1
and T
2
will vary with time. (ll the walls
of the containers have the same resistance and negligible
capacity. The two containers have the same capacity C.
133
".% 4erive the differential equation for a motor driving a load
through a gear system, as shown in the following figure, which
relates the angular displacement of the load with time.
".' 4erive the mathematical model for a 4: generator. The
generator may be assumed to have a constant magnetic field.
The armature circuit has the armature coil, having both
resistance and inductance, in series with the load. (ssume that
the load has both resistance and inductance.
".) 4erive the mathematical model for a permanent magnet
4: motor.
134
".0 4evelop the mathematical model for the field-controlled
dc motor shown in the following figure with e
&
as input and

as output.
".$1 The following system might represent a machine tool slide
driven through a rac, and pinion by a motor, or an
electrohydraulic actuator in which the mass
1
represents the
spool valve.
The gear shaft is supported by frictionless bearings and thus can
only rotate. (ssume that the only significant masses in the
system are the motor with inertia I and the slide with mass
1
.
5btain the system>s differential equations with the motor torque
T as the input.
135
".$$ ( load inertia I
2
is driven through gears by a motor with
inertia I
1
. The shaft inertias are I
#
and I
3
A the gear inertias are I
)

and I
8
. The gear ratio is 79$ !the motor shaft has the greater
speed&. The motor torque is T
1
, and the viscous damping
coefficient is c=1.2 0b-&t-4ec/r#d. 8eglect elasticity in the
system, and use the following inertia values !in 4ec
2
-&t-0b/r#d&9
I
1
=3.31 I
2
=3.)
I
3
=3.331 I
(
=3.33)
I
)
=3.32 I
8
=3.33
!a& 4erive the model for the motor shaft speed

1
with T
1
as
input.
!b& 4erive the model for the load shaft speed

2
with T
1
as
input.
136