Chapter II

Modelling and control of DC motor.

 II-1) Introduction:
A model is simply a representation of the construction and working of some systems of
interest. A model is like but simpler than the system it represents. The process of producing a
model is known as modelling. Basically, a good model is a judicious trade-off between
realism and simplicity. One of the important issues in modelling is model validity, which is
the process of simulating the model under known input conditions and comparing model
output with system output [2].

II-2) DC motor model:

The DC motor equations can be written using the equivalent circuit diagram. The excitation
of the armature is the terminal voltage V a , also called the armature voltage and the armature
current in the rotor winding is denoted by I a. The electric block represents the armature
circuit, which can be modeled as a simple serial-connected RL circuit, and the voltage E
induced by the rotation of the rotor acts against the armature voltage so that it is connected in
series with the RL circuit in the model, thus reducing the voltage on the RL-elements. The
mechanical block of the model describes the rotation of the rotor and the mechanical
characteristics of the rotational motion [13].
For illustration, let us consider the separately excited dc machine shown schematically in Fig.
For ease of analysis, the following assumptions are made:
i. The axis of armature MMF is fixed in space, along the q-axis.
ii. The demagnetizing effect of armature reaction is neglected.
iii. Magnetic circuit is assumed linear (no hysteresis and saturation). As a result, all
inductances (which came into play in dynamic analysis) are regarded as constant.
The two inductance parameters appearing in Fig are defined below:
La: armature self-inductance caused by armature flux; this is quite small and may be
neglected without causing serious error in dynamic analysis.
Lf : self-inductance of field winding; it is quite large for shunt field and must be accounted for
Mutual inductance (between field and armature) = 0; because the two are in space quadrature.
Further for dynamic analysis it is convenient to use speed in rad/s rather than rpm [14].

II-2-1) The electric equations:

The rotor is assumed to be a single coil characterized by inductance La and resistance Ra, but
it must be considered the back EMF of the motor in equation (I  13) [7].
Applying Kirchhoff’s law to the armature circuit:
d
V a (t )=E(t)+ R a i a(t)+ La i (t). (II  1)
dt a
 '
E=K ϕ ϕΩ ( t )=K Ω ( t ).

Where: K '=K ϕ ϕ .

Similarly for the field circuit:

d
V f (t)=R f i f (t)+ Lf i (t). (II  2)
dt f

Where:
V a is the applied armature voltage.

V f is the field voltage.

E is the back emf.

Ra is the resistance of the armature.

R f is the resistance of the field winding.

i a is the armature current.

i f is the field current.

Φ is constant [14].

II-2-2) The mechanical equations:

The motor is driving a mechanical load which must be represented by model that relates the
speed to torque. For simplicity, we assume that the load has a moment of inertia J ( kg . m2 / s2 ),
a viscous friction B(N . m/rad /s) and a net torque T L. Therefore, the motion equation can be
obtained as below [7]:
dΩ
T =J + BΩ+T L. (II  3)
dt
T =K ϕ ϕ I =K ' I .

Where:
T is the torque developed by the motor.

J is the equivalent moment of inertia of motor shaft and load referred to the motor.
B is the equivalent coefficient of motor and load referred to the motor shaft.
T L The resistant torque.

Energy storage is associated with the magnetic fields produced by i f and i a and with the
kinetic energy of the rotating parts. The above equations are a set of nonlinear (because of
products i f (t)Ω(t) and i f (t)i a (t )) state equations with state variables i f , i a and Ω . The
solution must be obtained numerically.
II-2-3) Transfer Functions and Block Diagrams:
In the simple linear case of motor response to changes in armature voltage, it is assumed that
the field voltage is constant and steady- state is existing on the field circuit, i.e. I f = constant.
Equations (II−1), (II−2) and (II−3) now become linear as given below:
d
V a (t )=K ' Ω(t)+ R a i a (t )+ L a i (t). (II  4)
dt a
d
T (t)=K ' i a (t )=J Ω( t)+BΩ (t)+T L (t) . (II  5)
dt
Laplace transforming eqs (II  4) and (II  5):
V a (s)=K ' Ω(s)+(Ra + s La )I a (s). (II  6)

T (t)=K ' I a ( s)=(B+ sJ )Ω(s)+T L (s). (II  7)

These equations can be reorganized as:

V a (s)−K ' Ω(s)
I a ( s)= .
(Ra + s La)

V a (s)−K ' Ω(s)

I a ( s)= (II 
R a (s τ a+1)
8)
Where:
La
τ a= : is known as the time constant of motor armature circuit.
Ra

K ' I a (s )−T L (s)

Ω(s)= .
(B+ sJ )

K ' I a (s )−T L (s)

Ω(s)= (II 
B (s τ m +1)
9)
Where:
J
τ m= : is known as the mechanical time constant of the motor.
B
T L ( s)
V a( s) 1 I a( s) T( s) 1 Ω(s)
K
R a (s B(s τ +1)
’ a
E g( s) τ a+ 1¿

K
’

Figure (II  1): DC motor block diagram

II-3) DC motor control :
The relationship of the speed of a DC motor (shunt or separate excitation) is given by:
V a−Ra I a
Ω= . (II 
Kϕϕ
10)
Exploring this relationship, it becomes clear that there are three possibilities for speed
adjustment:

 Action on Ra (dynamic adjustment).

 Action on ϕ (flow adjustment).
 Action on V a (adjustment by tension).

II-3-1) Dynamic Adjustment:

Since the voltage and flux are fixed at their nominal value, the speed can be reduced by
increasing the resistance of the armature by means of a rheostat ( Rh ) connected in series with
the armature. We have the following parametric relations:
¿ (II  11)
Va
 Pour T =0; I a=0  Ω= : This speed does not depend on Rh , so the corresponding
K'
point is immutable.
Va Va
 Pour Ω=0; I a=  T =K ' : When you increase Rh , the torque T decreases.
Ra + Rh R a+ R h
We obtain a bundle of concurrent lines.
This setting is bad both technically and economically. Indeed, from a technical point of view,
the characteristics being concurrent, they become more "series", i.e. with a resistor inserted,
the speed drop increases with the load. Technically, a good adjustment should result in the
features being moved parallel to the original feature.
In addition, this setting is bad from an economic point of view because the energy
consumption in the rheostat is even more important the higher the speed drop required. Thus,
at half speed, as much energy is consumed in the rheostat as in the engine. In practice, this
adjustment process is only used for starting or braking.
T
Rh2 Rh1 Rh =0 Rh2 >Rh1

TL

Ω
Ω2 Ω1 Ωn

Figure (II  2): Dynamic Adjustment: Adjustment characteristics.

II-3-2) Flow Adjustment:

It is important to note that, by virtue of the relationship of the torque ( T =K ϕ ϕ I a), it is always
in the interest of applying the maximum flow at start-up to allow the engine to bring to the
nominal speed all the masses that are initially at rest. In addition, due to the saturation of the
machine's iron, this flux value cannot be increased further. It follows therefore and in
accordance with equation (II  10) giving the speed of the motor; that the adjustment is
obtained by increasing the speed in relation to the nominal speed by reducing the value of the
flux (excitation current). This is achieved by introducing a field rheostat into the excitation
circuit. We have the following relationships:

{
T =K ϕ ϕ I a .
V −R a I a (II 
Ω= a .
Kϕ ϕ
12)
Va
Pour T =0 ; I a=0  Ω= : When the value of ϕ is reduced, the speed increases.
Kϕϕ

Va Va
Pour Ω=0; I a=  T =K ϕ ϕ : When the value of ϕ is reduced, the torque decreases.
Ra Ra

This setting is bad from a technical point of view; the characteristics being concurrent. It is
good from an economic point of view because the power dissipated in the inductor is very
low compared to the power absorbed; The engine efficiency will not be affected. However,
the following remarks should be made :
With this process, the motor speed can only be increased compared to its nominal speed. If
the load torque is constant T =K ϕ ϕ I a=cst , the current will increase when the flow is reduced
and the motor may heat up.
Under these conditions, the engine must be sized accordingly. It should be noted that this
disadvantage does not occur if the drive is at constant power because by virtue of the
relationship P=V a I a=cst , as the voltage U a is constant, the current I a will remain constant.

T
ϕn
ϕ1
ϕ2
ϕn > ϕ1 > ϕ 2

TL
Ω
Ωn Ω1 Ω2

Figure (II  3): Flow adjustment feature.

II-3-3) Voltage adjustment :

Since the excitation flux is set at its nominal value, the adjustment is achieved by reducing
the speed from the nominal speed by reducing the value of the supply voltage. By virtue of
the relationship:

{
T =K ' I a
V a−R a I a (II 
Ω= .
K'
13)
Va ∆Va
 Pour T =0 ; I a=0  Ω=  ∆ Ω= .
K' K'

When the voltage V a is reduced, the speed decreases.

Va Va ∆Va
 Pour Ω=0; I a=  T =K '  ∆ T =K ' .
Ra Ra Ra

∆T
When the voltage value is reduced, the torque decreases. Let's calculate the slope:
∆Ω
2
∆T K'
= =cst .
∆ Ω Ra

Features move parallel to the original feature.

 T

U2 U1 Un U1 > U2

TL

Ω
Ω2 Ω1 Ωn

Figure (II  4): Adjustment Feature by Armature Voltage.

II-4) Realization of a variable armature voltage:

To achieve a variable armature voltage, the WARD-LEONNARD assembly was used for a
long time. With the advent of semiconductors, it was found interesting to use a power
electronics converter (Figure (II 5)).

Circuit de
puissance
Ia

Source Va M
Rsh Tachymeter

Imes Ωmes
Command Regulator
(I)
-+ Regulator
(Ω)
-+ Ωref

Figure (II  5): Closed-loop variable DC voltage.

The electronic converter supplies the armature of the motor, and its speed is captured by a
tachometer dynamo. The measured value (Ωmes) is compared to the reference speed (Ωref).
The control gap is the cruise control input that generates the signal at the control unit input.
To this velocity loop, a current loop is added; A shunt R s h gives a voltage proportional to
applied I c to the input of the current regulator. The limiters placed at the output of the
regulators are used to maintain the current and armature voltage between the maximum
allowable values.
The converters used are classified as follows:
 AC-DC converters (Rectifiers): A distinction is made between non-reversible and
reversible assemblies.
 Continuous-to-continuous converters (chopper): There are also two types of assemblies:
non-reversible and reversible.
II-5) DC motor-Chopper association:
The main problem of modern electric control is the economical and progressive speed control
within wide limits and with high operational reliability. Manufacturing efficiency, production
quality, productivity and cost are all always linked to manufacturing automation. The
assurance of the rational operation of technological process depends very much on the speed
of machine tools. The speed may vary during the variation of the parameters of the electric
motors (resistors, inductors) or the parameters of the power source (voltage, frequency).

II-5-1) Chopper drives:

When the equipment is supplied with direct current, as is the case with a battery of
accumulators or the direct current catenary in electric traction, the variable direct voltage
applied to the armature V a is obtained by means of a chopper.
The value of V a can be influenced either by an analogue assembly or by means of a digital
control. The chopper allows the adjustment of the power transfer between a DC voltage
source and a current source. Several cases may arise:

II-5-2) Series chopper (buck):

The series chopper is used when the motor is to work only in quadrant 1:

exterior
command
Duty cycle  circuit
ID
IT K Ia Ie
VT
DC
source
V DL Va M Re Ve

Figure (II  6) : Association hacheur série avec un DC motor.

T is considered the operating period and ton the duration of the IGBT conduction intervals.
If K is conductor and D is blocked for 0 ≤ t ≤ ton:

V a =V → I T =I a →V T =0 → V D =−V → I D=0.

If K is blocked and D is conductor for ton ≤ t ≤ T:

V a =0 → I T =0 →V T =V →V D=0 → I D =I a .
 Figure (II  7): Motor currents and voltages, IGBT and diode.
According to the curves shown in Figure (II  7), the average value of the voltage across the
motor is:
t on
1 t
V a = ∫ V dt= on V =αV . (II  14)
T 0 T
where: α ∈ [0,1] α: is called a duty cycle.
The rotational velocity expression is related proportional to α [1]:

V a−Ra I a α V −Ra I a
Ω= = . (II  15)
K' K'

II-6) Closed-loop Control of DC Drives:

The speed of dc motors changes with the load torque. To maintain a constant speed, the
armature (and or field) voltage should be varied continuously by varying the duty cycle of
dc–dc converters. In practical drive systems it is required to operate the drive at a constant
torque or constant power; in addition, controlled acceleration and deceleration are required.
Most industrial drives operate as closed-loop feedback systems. A closed-loop control system
has the advantages of improved accuracy, fast dynamic response, and reduced effects of load
disturbances and system nonlinearities.
The block diagram of a closed-loop converter-fed separately excited dc drive is shown in
Figure (II  8). If the speed of the motor decreases due to the application of additional load
torque, the speed error V e increases. The speed controller responses with an increased control
signal V c change the duty cycle of the converter and increase the armature voltage of the
motor. An increased armature voltage develops more torque to restore the motor speed to the
original value. The drive normally passes through a transient period until the developed
torque is equal to the load torque.
 Power
supply
TL
Vr Ve Speed Vc Va Dc
Converter
controller motor

Speed
sensing

Figure (II  8): Block diagram of a closed-loop converter-fed dc motor drive.

The circuit arrangement of a converter-fed separately excited dc motor drive with open-loop
control is shown in Figure (II  9). The motor speed is adjusted by setting reference (or
control) voltage V r . Assuming a linear power converter of gain K 2, the armature voltage of
the motor is:
V a =K 2 V r . (II  16)
ia If
La Lf
+ Converter + Ra Vf
vr of va Rf
- gain, K 2 -
M
TL
T
B

Figure (II  9): Converter-fed separately excited DC motor drive.

Equations (II  16), (II  8), and (II  9) can be used to draw the open-loop block diagram as
shown in Figure (II  10). Two possible disturbances are control voltage V r and load torque
T L. The steady-state responses can be determined by combining the individual response due
to V r and T L.
T L ( s)
V r( s) V a( s) 1 I a( s) T d ( s) 1 Ω(s)
K2 K
R a (s B(s τ +1)
’ a
E g( s)

K
’

Figure (II  10): Open-loop block diagram of separately excited DC motor drive.

The response due to a step change in the reference voltage is obtained by setting T L to zero.
From Figure (II  10), we can obtain the speed response due to reference voltage as:
Ω(s) K 2 K ' /(R a B)
=GΩ−V = . (II
V r (s) 2 2
(τ a τ m)s +(τ a +τ m)s +1+ K ' /(R a B)
 17)
The response due to a change in load torque T L can be obtained by setting V r to zero. The
block diagram for a step change in load torque disturbance is shown in Figure (II  11):
 T L ( s) 1 Ω(s)
B(s

T (s)
τ m+ 1¿
I a( s) 1 E g( s)
K -1 K
R a (s
’ ’
τ a+ 1¿
Figure (II  11): Open-loop block diagram for torque disturbance input.

Ω(s) (1/B)(s τ a+1)

=GΩ−T = . (II  18)
T ( s) 2 2
(τ a τ m)s +(τ a +τ m) s +1+ K ' /( R a B)
The speed response due to simultaneous applications of disturbances in both input reference
voltage V r and the load torque T L can be found by summing their individual responses. Eqs
(II  17) and (II  18) give the speed response as:
Ω(s)=GΩ−V V r +GΩ −T T (II  19)

II-6-1) Converter control models:

We can notice from Eq. (II  14) that the average output voltage of a dc–dc converter is
directly proportional to the duty cycle α which is a direct function of the control volt age. The
gain of a dc–dc converter can be expressed as:

Vc
K r =α = . (II  20)
V cm

 V c : is the control signal voltage.

 V cm: is the maximum value of the control signal voltage.

Therefore, a converter can be modelled with a transfer function Gc (s) of a certain gain and
phase delay as described by:
τr
Gc ( s )=K r e .

which can be approximated as a first-order lag function given by:

Kr
G r ( s )= . (II  21)
1+ s τ r

τ r: is the time delay of the sampling interval.

Once a switching device is turned on, its gating signal cannot be changed. There is a delay
until the next device is gated and a corrective action is implemented. The delay time is
generally half of the interval between two switching devices.

Therefore, the delay time for a frequency of f s can be determined from:

 1
τ r= (for dc-dc converter).
2fs

II-6-2) Closed-loop transfer function:

Once the models for motors are known, feedback paths can be added to obtain the desired
output response. To change the open-loop arrangement in Figure (II  9) into a closed-loop
system, a speed sensor is connected to the output shaft. The output of the sensor, which is
proportional to the speed, is amplified by a factor of K 1 and is compared with the reference
voltage V r to form the error voltage V e . The complete block diagram is shown in Figure (II 
12)

T L ( s)

V r( s) Ve
K2
Va 1 Ia
K’
T 1 Ω(
R a (s B (s

Vb E g( s) τ a+ 1¿ τ m+ 1¿
K’

K1

Figure (II  12): Block diagram for closed-loop control of separately excited dc motor.

The closed-loop step response due to a change in reference voltage can be found from Figure
(II  12) with T L =0. The transfer function becomes:

K2K
Ω(s) Ra B
= 2
. (II 
V r (s) K +K 1 K2 K
(τ a τ m)s 2 +(τ a +τ m)s +1+
Ra B
22)

V r : is the reference voltage.

The response due to a change in the load torque T L can also be obtained from Figure (II  12)
by setting V r to zero. The transfer function becomes:

Ω(s) 1 s τ a+1
= .
T (s) B K +K1 K 2 K
2
(II 
(τ a τ m )s 2+(τ a+ τ m )s+1+
Ra B
23)

II-6-3) Closed-loop current control:

The block diagram of a DC machine contains an inner loop due to the induced emf E , as
shown in Figure (II  13). B and BL are the viscous frictions for the motor and the load,
respectively. H c is the gain of the current feedback; a low-pass filter with a time constant of
less than 1 ms may be needed in some applications. The inner current loop will cross this
back emf loop. The interactions of these loops can be decoupled by moving the K b block to
the I a feedback signal, as shown in Figure (II  14). This will allow splitting the transfer
function between the speed ω and the input voltage V a into two cascaded transfer functions,
(a) between the speed and the armature current, and (b) then between the armature current
and the input voltage. That is:

Ω(s) Ω(s ) I a ( s)
= × .
V a (s) I a (s) V a ( s)

T1

ia* Vc va 1 ia T 1 Ω
 R a + sL a
Kb
B + sJ

Current controlled ve
iam controller converter
Kb
Hc

Figure (II  13): Inner current loop.

V a(s)
1 Ia(s) Kb Ω(s
R a + sL a B + B l + sJ

E (s)

K b2
B + B l + sJ

Figure (II  14): Manipulation of moving the block K b for I a feedback.

The blocks in Figure (II  14) can be simplified to obtain the following transfer function
relationships:

Ω(s) Kb
= .
I a (s) B (1+s τ m )

I a (s ) 1+s τ m
=K m .
V a (s) (1+ s τ 1)(1+ s τ 2)

Where:

B
Km= 2 .
K + Ra B
b

[( ) √( ) ( )]
2 2
−1 1 1 Ra B Ra B Ra B+ K b
s1 ,2 = ,− = − + ± + −4 . (II  24)
τ1 τ2 2 La J La J J La

Two block representation of the open-loop motor transfer function between the output speed
and the input voltage is shown in Figure (II  15), where total viscous friction Bt =B+ B L.
 1+ s m I s( s) K b /B t
K1
(1 + s 1)( 1+ s 2) 1+ s m

Figure (II  15): Open-loop motor transfer function.

The overall closed-loop system with a current feedback loop is shown in Figure (II  16),
where Bt is the total viscous friction of the motor and the load. The converter can be
represented by Eq (II  21). Using proportional-integral type of control, the transfer functions
of the current controller Gc (s) and the speed controller Gs (s) can be represented as:

K c (1+s τ c )
Gc (s)= . (II  25)
s τc

K s (1+s τ s )
Gs (s)= . (II  26)
s τc

Where:

 K c and K s are the gains of the current and speed controllers, respectively.
 τ c and τ s are the time constants of the current and speed controllers, respectively.

Ω
¿
Ωr G s( s)
ia*
G c( s)
vc
G r( s )
va
K1
1+ s m ia K b /B t
(1 + s 1)( 1+ s 2) 1+ s m
Speed Current Converter
Limiter controller
controller
Ωmr Hc

GΩ (s)

Figure (II  16): Motor drive with current-control loop.

A dc tachogenerator along with a low-pass fitter of a time constant less than 10 ms is often
required. The transfer function of the speed feedback filter can be expressed as:

KΩ
GΩ (s)= .
1+ s τ Ω

Where:

 K Ω and τ Ω are the gain and time constants of the speed feedback loop.

Figure (II  17) shows the block diagrams with the current-control loop. i am is the feedback
armature- motor current. The simplified diagram is shown in Figure (II  18).
 K c (1 + s c) Vc Kr Va 1+ s m
i a* K1 ia
s c 1+ s r (1 + s 1) (1 + s 2)

iam
Hc

Figure (II  17): Current-control loop.

Kc K 1 m
i a* Kr ia
Tc 1+ s 3

Hc

Figure (II  18): Simplified current-control loop.

II-6-4) Design of Current Controller:

The loop gain function of the motor drive in Figure (II  17) is given by:

G(s) H (s )=
( Km Kc K r H c
τc ) (1+s τ c )(1+ s τ m)
s (1+ s τ 1 )(1+s τ 2 )(1+ s τ r )
.

To reduce the system to a second order, the following assumptions can be made for practical
motor drives:

1+ s τ m ≈ s τ m; τ 1 > τ 2> τ r and τ 2=τ c .

With these assumptions, Eq. (14.104) can be simplified to:

K
G(s) H (s )= . (II 
(1+ s τ 1 )(1+s τ r )
27)

Km Kc K r H cτ m
Where: K= .
τc

The characteristic equation of the loop gain in Eq. (14.105) is given by:

1+G(s)H (s )=(1+ s τ 1)(1+ s τ r )+ K =0.

This gives the natural frequency ω n and damping factor ξ of the current-control loop as:

ω n=
√ 1+ K
τ1 τr
.
 τ 1 +τ r
τ1τ r .
ξ=
2 ωn

Setting the damping ratio ξ=0.707 for critically damped, assuming K>>1 and τ 1 > τ r , the gain
of the current controller can be expressed as:

1 τ1 τc 1
K c= .
2 τr Km K r Hc τ m

II-6-5) Design of speed Controller:

The design of the speed controller can be simplified by replacing the second-order model of
the current loop with an approximate first-order model. The current loop is approximated by
adding the time delay τ r of the converter to the time delay τ 1 of the motor, as shown in Figure
(II  18). The transfer function of the current-control loop in Figure (II  18) can be expressed
as:

I a (s) Ki
¿ = . (II 
I a (s) (1+ s τ i)
28)

Where:

τ3
τi= .
1+ K 1

τ 3 =τ 1 +τ r.

K i=
1
(K1
H c 1+ K 1
. )
Km Kc Kr H c τm
K 1= .
τc

Replacing the speed controller with its transfer function in Eq (II  26) and approximating
current-control loop in Eq (II  28), the block diagram of the outer speed- control loop is
shown in Figure (II  19). The loop gain function is expressed as:

G(s) H (s )=
( K s Ki Kb HΩ
B τs ) (1+s τ s )
s (1+ s τ i)(1+ s τ m)(1+ s τ Ω )
.
 Ωr
¿ K s (1 + s s) ia* Ki ia K b /B t Ω
s s 1+ s i 1+ s m

Speed controller Current loop

Ωmr HΩ
1+ sΩ

Figure (II  19): Outer speed-control loop.

Assuming 1+τ m ≈ τ m and combining the delay time τ s of the speed controller with the delay
time τ Ω of the speed feedback filter to an equivalent delay time τ e=τ s + τ Ω:

G(s) H (s )=K Ω ( )
K s (1+ s τ s )
τ s s2 (1+ s τ e )
. (II  29)

Where:

τ e=τ s + τ Ω.

K i Kb HΩ
KΩ= .
B τm

The closed-loop transfer function of the speed in response to speed reference signal is given
by:

[ ]
KΩKs
(1+ s τ s)
Ω(s) 1 τs
¿ = . (II 
Ω (s) K Ω 3 2 KΩ Ks
s τe + s + s KΩ Ks+
τs
30)

Equation (II  30) can be optimized by making the magnitude of its denominator such that the
coefficients of ω2 and ω4 in the frequency domain are zero. This should widen the bandwidth
operating over a wide frequency range. It can be shown that the gain K s and the time constant
τ s of the speed controller under the optimum conditions are given by:

1
Ks= .
2 KΩ τe

τ s=4 τ e .

Substituting K s and τ s into Eq. (II  30) gives the optimized closed-loop transfer function of
the speed in response to speed input signal as:
Ω(s)
¿ =
1
[
(1+ 4 s τ e )
Ω (s) K Ω 1+ 4 s τ e +8 s 2 τ 2e + 8 s 3 τ 3e
.
]
It can be shown that the corner frequencies for the open-loop gain function H(s)G(s) are
1/(4 τ e ¿ )¿ and 1/τ e. The gain crossover occurs at 1/(2 τ e ¿ )¿ at a slope of the magnitude
response—20 dB/decade. As a result, the transient response should exhibit the most desirable
characteristic for good dynamic behaviour. The transient response is given by:

[ ( )]
−t −t

1 + e 2 τ −2 e 4 τ cos √
1 3t
Ω(t)= e e
. (II  31)
KΩ 4 τe

This gives a rise time of 3.1τ e, a maximum overshoot of 43.4%, and a settling time of 16.5 τ e.
The high overshoot can be reduced by adding a compensating network with a pole-zero in the
speed feedback path, as shown in Figure (II  19). The compensated closed-loop transfer
function of the speed in response to speed input signal is:

Ω(s)
¿ =
1
[ 1
Ω (s) K Ω 1+ 4 s τ e +8 s 2 τ 2e + 8 s 3 τ 3e
.
]
¿
Ωr 1 1 1+ 4 e s
1+ 4 e s HΩ 1+ 4 e s +8  e 2 s2 +8 e 3s3

Figure (II  19): Adding a pole-zero compensating network.

The corresponding compensated transient response is given by:

[ ( )]
−t −t

1 + e 4 τ − e 4 τ sin √
1 2 3t
Ω(t)= e
.e
(II  32)
KΩ √3 4τe

This gives a rise time of 7.6 τ e, a maximum overshoot of 8.1%, and a settling time of 13.3 τ e.
It should be noted that the overshoot has been reduced to approximately 20% of the value,
and the settling time has come down by 19%. But the rise time has increased, and the
designer must make a compromise on the rise time and the overshoot.

II-7) PID controller:

PID controllers have a rich history that dates to the early 20th century. James Clerk Maxwell
developed the foundational concepts in 1868, analysed governors for controlling steam
engines, and introduced the idea of proportional control. Later, in 1922, Nicolas Minorsky
formalized the PID framework while studying automatic ship steering systems. His research
emphasized the benefits of combining proportional, integral, and derivative actions to
enhance system stability and eliminate steady-state errors.
PID controller model Due to its simplicity and efficacy, the proportional-integral derivative
(PID) controller is a regularly used feedback control mechanism for managing the speed of
DC motors. It operates by continuously calculating the error between the desired speed and
the motor's actual speed. The controller then applies a correction based on three components:
proportional, integral, and derivative. The output of the controller adjusts the motor's speed
accordingly.
Popular tuning techniques include manual tuning, Ziegler Nichols, and Cohen-Coon. It
provides an empirical technique that uses the system's ultimate gain and oscillation period to
recommend starting parameter settings. Upon establishing these values, precise modifications
are implemented to equilibrate trade-offs across performance indicators, including rise time,
overshoot, and settling time. Figure (II  20) shows the PID configuration model.

Figure (II  20): the configuration model of PID controller

According to the process of PID control, mathematical representation of PID controller is
given by [15]:
t
de (t)
U (t )=K p e (t)+ K i∫ e(t)dt+ K d . (II  33)
0 dt

Where:
e (t): is the error between the desired and actual speed.
K p : is the proportional gain.

K i: is the integral gain.

K d : is the derivative gain.

The following will outline the main types of regulator’s actions:

 Proportional control:
According to the expression, the proportional control is proportional to the current control
error:
U (t )=K p e (t).
If the proportional value is high, it results in a larger output for the same error. However, an
excessively high proportional gain can lead to system instability. Conversely, a low
proportional gain yields a smaller output for the same error, making the controller not too
sensitive. Yet, if the value of proportional is too small, the support signal is likely to be
insufficiently large to correct for interference effects.
 Integral control:
Integral control is positively dependent on the integral of the set of control uncertainties. The
expression is shown below:
t
U (t )=K i∫ e(t)dt
0

Integral control reduces the time to converge to setpoint and eliminates some of the steady
state error. The higher the integrated charge gain, the shorter the time to converge to setpoint,
but since the integral control cumulates whole previous errors, it may result in an overshoot
of the return value.
 Derivative control:
The role of derivative control is to forecast future values based on the control error. The
expression is as follows:
de (t)
U (t )=K d .
dt
Differential control improves tuning time and system stability [16].

II-7-3) Ziegler–Nichols Rules for Tuning PID Controllers:

Ziegler and Nichols proposed rules for determining values of the proportional gain K P,
integral time T i, and derivative time T d based on the transient response characteristics of a
given plant. Such determination of the parameters of PID controllers or tuning of PID
controllers can be made by engineers on-site by experiments on the plant. (Numerous tuning
rules for PID controllers have been proposed since the Ziegler–Nichols’s proposal. They are
available in the literature and from the manufacturers of such controllers).
PID controller is defined as:

(
Gc ( s )=K p 1+
1
Tis )
+s T d . (II  34)

Where:
Kp
T i= : is the integral time.
Ki

Kd
T d= : is the derivative time.
Kp

There are two methods called Ziegler–Nichols tuning rules: the first method and the second
method. We shall give a brief presentation of these two methods.
First Method:
In the first method, we obtain experimentally the response of the plant to a unit-step input, as
shown in Figure (II  21). If the plant involves neither integrator(s) nor dominant complex-
conjugate poles, then such a unit-step response curve may look S-shaped, as shown in Figure
(II  22). This method applies if the response to a step input exhibits an S-shaped curve. Such
step-response curves may be generated experimentally or from a dynamic simulation of the
plant. The S-shaped curve may be characterized by two constants, delay time L and time
constant T. The delay time and time constant are determined by drawing a tangent line at the
inflection point of the S-shaped curve and determining the intersections of the tangent line
with the time axis and line c(t)=K, as shown in Figure (II  22).

Figure (II  21): Unit-step response of a plant.

Figure (II  22): S-shaped response curve.

C (s )
The transfer function may then be approximated by a first-order system with a transport
U (s )
lag as follows:

C ( s ) K e− Ls
= . (II 
U ( s ) T s +1
35)

Ziegler and Nichols suggested to set the values of K p , T i, and T d according to the formula
shown in Table (II1):

Type of Controller Kp Ti Td
T
P ∞ 0
L
T L
PI 0.9 0
L 0.3
T
PID 1.2 2L 0.5 L
L
Table (II1): Ziegler–Nichols Tuning Rule Based on Step Response (First Method).
Notice that the PID controller tuned by the first method of Ziegler–Nichols rules give:
 2
1
(
Gc ( s )=K p 1+
1
Tis ) T
(
+T d s =1.2 1+
L
1
2 Ls
+ 0.5 Ls . )
Gc ( s )=0.6 T
(s+ )
s
L
.
(II  36)
Thus, the PID controller has a pole at the origin and double zeros at −1/ L.
Second method:

In the second method, we first set T i=¿ and T d=0. Using the proportional control action only
(see Figure (II  23)), increase K p from 0 to a critical value K cr at which the output first
exhibits sustained oscillations. (If the output does not exhibit sustained oscillations for
whatever value K p may take, then this method does not apply.) Thus, the critical gain K cr and
the corresponding period Pcr are experimentally determined (see Figure (II  24)). Ziegler and
Nichols suggested that we set the values of the parameters K p , T i and T d according to the
formula shown in Table (II2):

Type of Controller Kp Ti Td

P 0.5 K cr ∞ 0
1
PI 0.45 K cr P 0
1.2 cr
PID 0.6 K cr 0.5 Pcr 0.125 Pcr
Table (II2): Ziegler–Nichols Tuning Rule Based on Critical Gain K cr and Critical Period Pcr
(second Method)

Figure (II  23): Closed-loop system with a proportional controller.

Figure (II  24): Sustained oscillation with period Pcr ( Pcr is measured in sec.)

Notice that the PID controller tuned by the second method of Ziegler–Nichols rules give:
 2
4
(
Gc ( s )=K p 1+
1
Tis ) (
+T d s ¿ 0.6 K cr 1+
1
0.5 Pcr s
+0.125 Pcr s )
Gc ( s )=0.075 K cr P cr
(s +
P cr
s
)
.

(II  37)

Thus, the PID controller has a pole at the origin and double zeros at s=−4 /P cr .