This action might not be possible to undo. Are you sure you want to continue?
1, 2011
* Electromechanical Engineering Department ,University of Technology/Baghdad
141
Modeling, Analysis and Speed Control Design Methods of
a DC Motor
Dr. Jamal A. Mohammed*
Received on: 23/12/2009
Accepted on:5 /1/2011
Abstract
Modern manufacturing systems are automated machines that perform the
required tasks. The electric motors are perhaps the most widely used energy
converters in the modern machinetools and robots. These motors require
automatic control of their main parameters (position, speed, acceleration,
currents).
With the help of an example, a DC motor system, the use of
MATLAB/Simulink for comprehensive study of modeling, analysis and speed
control design methods has been demonstrated.
Keywords: DC motor, open loop, closed loop, system, speed control.
ميمصت قرطو ليلحتو ةجذمن ا ةرطيسل رايتلا كرحم ةعرس ىلع ملا رمتس
ةصخلا :
ةمظنأ ةثيدحلا عينصتلا يھ ةبولطملا تامھملا زجنت ةيلآ نئاكم . امبر ةيئابرھكلا تاكرحملا
يلا ناسناو ةثيدحلا نئاكملا تادعم يف
ً
امادختسا رثكا ةقاطلا توحم لثمت .
ل ةيلآ ةرطيس تاكرحملا هذھ بلطتت اھتماعم ةيساسا ) لاك عقوم و يجعتلاو ةعرسلا تارايتلاو ل ( .
ضارعتسا
ّ
مت يلاحلا ثحبلا يف ةلماش ةسارد ةجذمنل ىلع ةرطيسلا ميمصت قرطو ليلحتو ةعرسلا
ل ـلا جمانرب مادختساب رمتسملا رايتلا وذ كرحملا ةموظنم MATLAB/Simulink .
1Introduction:
he mechatronic systems,
robots and low to medium
power machinetools often use
DC motors to drive their work loads.
These motors are commonly used to
provide rotary (or linear) motion to a
variety of electromechanical devices
and servo systems.
There are several well known
methods to control DC motors such
as: ProportionalIntegral PI,
ProportionalIntegralDerivative PID
or bipositional [1]. Despite a lot of
researches and the huge number of
different solutions proposed, most
industrial control systems are based
on conventional PID controllers.
The purpose of developing a
control system is to enable stable and
reliable control. Once the control
system has been specified and the
type of control has been decided, then
the design and analysis are done.
There are three major objectives of
system analysis and design:
producing the desired transient
response, reducing steadystate error,
and achieving stability [2].
A simple technique for designing
a DC motor speed controller with
continuous time based on system
theory concepts was presented in [1].
In [3], Rusu, et al successfully
designed and implemented an open
loop control system to control the
speed of a DC motor. In [4], Ayasun
T
Eng. & Tech. Journal, Vol. 29, No. 1, 2011 Modeling, Analysis and Speed Control
Design Methods of a DC Motor
142
and Nwankpa described the
MATLAB/Simulink realization of the
DC motor speed control methods
(field resistance, armature voltage and
resistance) and feedback control
system for DC motor drives. In [5],
Othman successfully designed a
speed controller for closed loop
operation of the BLDC motor. In [6],
Allaoua, et al presented a new design
method to determine optimal PID
controller parameters using the
Particle Swarm Optimization method.
The current paper intends
presenting comprehensive study for
all the techniques of modeling,
analyzing, and designing a DC motor
speed controller based on system
theory concepts.
The speed control design methods
used are studied under MATLAB and
Simulink in the following sections.
2 MATLAB Control Design
Method for Speed Modeling:
i System Equations:
To perform simulations of a
system, an appropriate model needs
to be established. This paper was
presented the system contains a DC
motor based on the motor
specifications needs to be obtained.
This is achieved by developing the
open loop transfer function of a DC
motor with the use of the system
equations of the motor as given by
Rashid [7].
The design method uses the concepts
of the system theory, such as signals
and systems, transfer functions, direct
and inverse Laplace transforms. This
requires building the appropriate
Laplace model for each component of
the whole control system.
In order to build the DC motor’s
transfer function, its simplified
mathematical model has been used.
This model consists of differential
equations for the electrical part,
mechanical part and the
interconnection between them. The
electric circuit of the armature and the
free body diagram of the rotor are
shown in the Fig. 1. All the values for
the physical parameters are listed in
Table A in the Appendix.
The motor torque, T
m
, is related to
the armature current, i, by a constant
factor K
t
. The back emf, e
m
, is related
to the rotational speed, θ
&
by the
following equations [7]:
i K T
t m
= ….. (1)
θ
&
e m
K e = ….. (2)
Assuming that, K
t
(torque constant) =
K
e
(electromotive force constant) =
K
m
(motor constant).
From Fig. 1 and Table A, the
following equations can be written
based on Newton’s law combined
with Kirchhoff’s law:
L m
T i K b J − = + θ θ
& & &
….(3)
θ
&
m m m
K V i R
dt
di
L − = + ….. (4)
a Transfer Function
Using Laplace Transforms, the
above equations can be expressed in
terms of sdomain.
( ) ) ( ) ( ) ( s T
L m
s I K s b s J s − = + θ
&
..(5)
( ) ) ( ) ( ) ( s s K s V s I s
m m m
R L θ
&
− = +
….(6)
By eliminating I(s), the following
openloop transfer function can be
obtained, where the rotational speed
θ
&
is the output and the voltage V is
the input.
When the motor is used as a
component in a system, it is desired
to describe it by the appropriate
transfer function between the motor
voltage and its speed. For this
purpose assuming (load torque) T
L
=0
and (friction torque) T
f
=0, since
neither affects the transfer function.
Therefore
Eng. & Tech. Journal, Vol. 29, No. 1, 2011 Modeling, Analysis and Speed Control
Design Methods of a DC Motor
143
( )( )
2
0 ) (
) (
) (
m m m
m
s T
K s b s J
s V
s
R L
K
L
+ + +
=
=
θ
&
[1,8,9] (7)
b. StateSpace Representation:
In the statespace form, Eqs. 3
and 4 can be expressed by choosing
θ
&
and i as the state variable and V as
an input. The output is chosen to
beθ
&
.
V
L
i
L
R
L
K
J
K
J
b
i dt
d
m
m
m
m
m
m
(
(
¸
(
¸
+
(
¸
(
¸
(
(
(
(
¸
(
¸
− −
−
=
(
¸
(
¸
1
0
θ θ
& &
[ ]
(
¸
(
¸
=
i
θ
θ
&
&
0 1 ….(8)
ii Design Requirements:
Since the most basic requirements
of a motor are that it should rotate at
the desired speed, the steadystate
error e
ss
of the motor speed should be
less than 1%. The other performance
requirement is that the motor must
accelerate to its steadystate speed
ss
θ
&
as soon as it turns on. In this case, it
is desirable to have a settling time T
s
in less than 2sec. Since speed faster
than the reference may damage the
equipment, a peak overshoot Mp of
less than 5% is wanted.
The definition of these parameters
can be demonstrated as in [10].
iii MATLAB Representation and
Openloop Response:
The transfer function in Eq. 7 can
be represented into MATLAB by
defining the numerator (num) and
denominator (den) matrices as
follows: num=K
m
and
den=(Js+b)(L
m
s+R
m
)+K
m
2
,respectivel
y, and the corresponding plot is
shown as in Fig. 2.
The motor is assumed to be
exposed at steadystate speed to step
change in speed with 0.1 p.u. (10% of
its steadystate speed).
From the plot, it can be seen that when a
step input voltage is applied to the system
at steadystate speed, the motor can only
achieve a maximum speed of 0.11 p.u.;
ten times smaller than our desired speed
(1 p.u.). The motor parameters were
determined on the basis of its step
response about a steadystate operating
point, namely input voltage V= 6 volts
and speed θ
&
= 600 r/s. Also, it makes the
motor takes 3sec to reach its steadystate
speed; this does not satisfy the 2sec
settling time criterion.
The system can also be
represented using the statespace Eqs.
8 to get the same output plot obtained
in Fig. 2.
3 Continuous PID Control Design
Method:
PID controllers are commonly
used to regulate the timedomain
behavior of many different types of
dynamic plants [8].
Considering the unity feedback
system architecture as shown in Fig.
3, where it can be assumed that the
plant is a DC motor whose speed
must be controlled and the Controller
provides the excitation for the plant;
designed to control the overall system
behavior.
Different characteristics of the
motor response (steadystate error,
peak overshoot, rise time, etc.) are
controlled by selection of the three
gains that modify the PID controller
dynamics.
First, taking a look at how the PID
controller works in the closedloop
system shown in Fig. 3. The variable
e represents the tracking error; the
difference between the desired input
value R and the actual output Y. This
error signal will be sent to the PID
controller, and the controller
computes both the derivative and the
integral of this error signal.
Eng. & Tech. Journal, Vol. 29, No. 1, 2011 Modeling, Analysis and Speed Control
Design Methods of a DC Motor
144
Therefore, the PID controller is
defined by the relationship between
the controller input e and the
controller output u that is applied to
the motor armature:
dt
de
K dt e K e K u
D I P
+
∫
+ = (9)
, where K
P
= Proportional gain, K
I
=
Integral gain, and K
D
=Derivative
gain.
The signal u will be sent to the plant,
and the new output Y will be obtained
and sent back to the sensor again to
find the new error signal e. The
controller takes e and computes its
derivative and it’s integral again. This
process goes on and on. By adjusting
the weighting constants K
P
, K
I
, and
K
D
, the PID controller can be set to
give the desired performance. Effects
of each of these parameters on a
closedloop system are summarized
as in [8].
Taking the Laplace transform of Eq. 9
gives the following transfer function:
s K
s
K
K
s E
s U
s K
D
I
P
+ + = =
) (
) (
) (
s
K s K s K
I P D
+ +
=
2
[8]
….(10)
This transfer function clearly
illustrates the proportional, integral,
and derivative gains that make up the
PID compensation.
With adding 0.1 p.u. step input to
the steadystate speed and the design
criteria previously mentioned, the
PID controller can be designed and
added into the system, recalling the
transfer function of a PID controller
in Eq. 10.
i. Design Steps of PID Controller:
To designing a PID controller for
a given system, the following steps
must be followed to obtain a desired
response.
1. Obtaining an openloop
response and determining what needs
to be improved.
2. Adding a K
P
to improve the rise
time.
3. Adding a K
D
to improve the
overshoot (increase damping).
4. Adding a K
I
to eliminate the
steadystate error.
5. Adjusting each of K
P
, K
I
, and
K
D
until obtaining a desired response.
The MATLAB is used to obtain a
closedloop transfer function directly
from the openloop transfer function.
Using K
P
with a gain of 100, the step
response can be shown as in Fig. 4.
From the plot in Fig. 4, it can be
seen that, both the steadystate error
and the overshoot are too large.
Adding an integral term will
eliminate the steadystate error and a
derivative term will reduce the
overshoot. Trying a PID controller
with small K
I
and K
D
, the resulting
response will be plotted as in Fig. 5.
ii. Tuning the Gains:
By tuning the gains of the PID
controller and producing the optimum
response using trial and error method
(it provides simplest way to achieve a
good compensator [11]), the
simulation start with the best initial
gains.
The settling time shown in Fig. 5
is too long. So it is necessary to
increase K
I
to reduce the settling time.
Changing K
I
to 200 makes the
response looks as in Fig. 6.
The response in the Fig. 6 seems
much faster than before, but the large
K
I
has worsened the transient
response (big overshoot). Therefore,
K
D
must increase to reduce the
overshoot.
Changing K
D
to 10 can get the plot
seen in Fig. 7.
So, it must be known that using (after
several trial and error runs) a PID
Eng. & Tech. Journal, Vol. 29, No. 1, 2011 Modeling, Analysis and Speed Control
Design Methods of a DC Motor
145
controller with K
P
=100, K
I
=200, and
K
D
=10, all of the design requirements
will be satisfied, providing the
desired response.
4 Root Locus Control Design
Method:
For designing a controller using
the root locus method, it should
follow the following steps:
i. Openloop Root Locus:
The main idea of root locus
design is to find the closedloop
response from the openloop root
locus plot. Then by adding zeros
and/or poles to the original plant, the
closedloop response can be
modified. So, first, it is necessary
viewing in the root locus for the
plant.
Three arguments must be founded
are the damping ratio (zeta) term (ξ >
0.8 corresponds to an overshoot Mp <
5%), the natural frequency (ω
n
) term
(= 0 corresponds to no rise time
criterion), and the sigma term (4.6/2 =
2.3 sec). Using these criteria results a
root locus plot shown in Fig. 8.
The following three equations are
used in continuous system designs:
ξ ω
n
≥ 4.6 / T
s
(Sigma term) (11)
ω
n
≥ 1.8 / T
r
….(12)
ξ ≥[((lnMp /π)
2
/(1+(lnMp /π)
2
)]
0.5
….(13)
, where ξ: Damping ratio, ω
n
: Natural
frequency, T
s
: Settling time, T
r
: Rise
time and Mp: Peak overshoot [11].
ii. Finding the Gain:
Recalling that, the settling time
and the overshoot must be as small as
possible. Large damping corresponds
to points on the root locus near the
real axis. A fast response corresponds
to points on the root locus far to the
left of the imaginary axis. Matlab can
find the gain corresponding to a point
on the root locus. The gain can be
found and the step response plotted
using this gain all at once.
When a point is selected on the root
locus plot in Fig. 8, halfway between
the real axis and the damping
requirement, say at 6+2.05i, the
MATLAB should return the output
similar to the following: Selected
point=6.0204+2.0401i, gain;
K=10.0708, poles=6.0000±2.0400i,
Mp %=0.0101 and the ξ = 0.947.
Noticing that, the values returned
in MATLAB results may not be
exactly the same, but should at least
have the same order of magnitude.
With the above results shown in Fig.
9, it can be seen that, there is no
overshoot and the settling time is
about 1sec, so the overshoot and
settling time requirements are
satisfied. The only problem can be
seen from this plot is that; the steady
state error is about 50%. If the gain is
increased to reduce the steadystate
error, the overshoot becomes too
large. So it is necessary adding a Lag
Controller (in transfer function form)
to reduce the steadystate error.
iii. Adding a Lag Controller:
The plot in Fig. 8 shows a very
simple root locus. The damping and
settling time criteria were met with
the proportional controller. The
steadystate error is the only criterion
not met with the proportional
controller. A Lag Compensator [11]
can reduce the steadystate error. By
doing this, it might however increase
the settling time. By using the
following lag controller first:
(s + 1) / (s + 0.01)
….(14)
, the root locus will be shown as in
Fig. 10, which looks very similar to
the original one.
iv. Closedloop Response:
By closing the loop and seeing the
closedloop step response, the
Eng. & Tech. Journal, Vol. 29, No. 1, 2011 Modeling, Analysis and Speed Control
Design Methods of a DC Motor
146
following Matlab results will be
obtained: selected point = 5.6122+
4.4444i, gain; K=19.9563, poles=
5.6138±4.4445i, and 0.7824.
When prompted to select a point, pick
one that is near the damping
requirement (diagonal dot line in Fig.
10). The resultant plot will be shown
as in Fig. 11.
The gain should be about 20. It can be
seen from the figure that the response
is not quite satisfactory. It may also
note that even though the gain was
selected to correlate with a position
close to the damping criterion, the
overshoot is not even close to 5%.
This is due to the effect of the lag
controller; its pole is much slower.
What this means is that it can go
beyond the dotted lines that represent
the limit, and getting the higher gains
without worrying about the
overshoot.
Keep trying until getting a
satisfactory response. It should look
similar to the following (using a gain
of around 50): Selected point= 
5.5714+8.8889i, gain; K = 49.8876,
poles = 5.5498 ± 8.8887i and 
0.9104.
Fig. 12 shows the corresponding
response with gain =50.
It can be seen from the figure that the
steadystate error is smaller than 1%,
and the settling time and overshoot
requirements have been met. As
result, it concluded that, the design
process for root locus is very much a
trial and error process. That is why it
is nice to plot the root locus, picking
the gain, and plotting the response all
in one step. If it has not been able to
get a satisfactory response by
choosing the gains, it can have tried a
different lag controller, or even added
a Lead Controller.
5 Frequency Control Design
Method:
i. Original Bode Plot:
The main idea of frequencybased
design is to use the Bode plot of the
openloop transfer function to
estimate the closedloop response.
Adding a controller to the system
changes the openloop Bode plot,
therefore changing the closedloop
response. First, the Bode plot for the
original openloop transfer function is
drawn as shown in Fig. 13.
ii. Adding Proportional Gain:
From the bode plot in Fig. 13, it
can be seen that the phase margin Ф
m
[11] can be greater than about (180
0

120
0
=60
0
) if ω is less than 10 r/s.
Adding gain to the system so the
bandwidth frequency [11] is 10 r/s,
which will give Ф
m
of about 60
0
. The
gain at 10 r/s can be found by reading
off the Bode plot in Fig. 13 (it looks
to be slightly more than 40dB, or
0.01 in magnitude) and the exact
magnitude: mag = 0.0139. To have an
openloop gain of 1 (0dB) at 10 r/s, it
may be needed to add a control gain
of 1/0.0139≈72 (≈37dB) to get the
Bode plot shown as in the Fig. 14.
From this plot: (mag=0.0139, phase =
123.6835, and ω=10 r/s).
iii. Closedloop Response:
From the plot in Fig. 14, it can be
seen that the phase margin is now
quite large and the corresponding
closedloop response looks like as in
Fig. 15. The figure shows that the
settling time is fast enough, but the
overshoot and the steadystate error
are too high.
iv. Adding a Lag Controller:
Reducing the gain to 50, and
trying the lag controller in Eq. 14,
which should reduce the steadystate
error by a factor of 1/0.01=100 (but
could increase the settling time) gives
the bode plot shown in Fig. 16 with
Eng. & Tech. Journal, Vol. 29, No. 1, 2011 Modeling, Analysis and Speed Control
Design Methods of a DC Motor
147
(K=50≈34dB) with good phase
margin (≈70
0
).
Closing the loop gives the step
response shown in Fig. 17, which
meets the design requirements.
It can see that adding a lag controller
can reduce the steadystate error and
at the same time, the overshoot can be
reduced by reducing the gain.
6 A StateSpace Control Design
Method:
For the main problem, the
dynamic equations in statespace
form are given as in the Eqs. 8.
i. Designing the Fullstate Feedback
Controller:
The schematic for a fullstate
feedback system is shown as in the
Fig. 18.
The characteristic polynomial for this
closedloop system is the determinant
of (sI(ABK)) where s is the Laplace
variable. Since the matrices A and
BK are both 2x2 matrices, there
should be 2 poles for the system. By
designing a fullstate feedback
controller, these two poles can move
anywhere wanted them. The first try
is to place them at:
5±i ;( poles= ξω
n
± jω
n
(1ξ
2
)
0.5
)
[11] corresponds to ξ = 0.98 which
gives 0.1% overshoot and a
sigma=ξω
n
=5 which leads to≈1sec
settling time by using Eqs. (1113).
Once coming up with the wanted
poles, MATLAB will find the
controller matrix, K.
After adding the K matrix into the
system in Fig. 18, the statespace
equations become:
( ) Bu X BK A X + − =
&
CX Y = [11]
….(15)
, and the plot of the closedloop
response can be shown as in Fig. 19.
ii. Adding a Reference Input:
From the plot in Fig. 19, it can be
seen that the steadystate error is too
large. Computing the input can be
done in one step by adding a constant
gain Nbar after the reference R. The
Matlab can find the scale factor N
which will eliminate the steadystate
error to a step reference for a
continuoustime, singleinput system
with fullstate feedback using the
schematic shown in Fig. 20. The
corresponding step respond plot is
shown in the Fig. 21 and the
MATLAB results are: Nbar=10.0048
and poles = 4.3600±1.0000i.
Now, the steadystate error is much
less than 1% and all the other design
criteria have been met as well.
7 Discrete PID Control Design
Method:
A digital DC motor model can be
obtained from conversion of the
analog model. The controller for this
example will be designed by a PID
method.
To discretise the controller, the
integral and derivative terms of the
controller have to be approximated.
The integral becomes a sum and the
derivative becomes a difference. The
continuous time signal e is sampled at
a sample period T [8].
i. Continuous to Discrete
Conversion:
The first step in designing a
discrete control system is to convert
the continuous transfer function in
Eq. 7 to a discrete one [12] as
follows:
2369 . 0 0877 . 1
0057 . 0 0092 . 0
) (
) (
2
+ −
+
=
z z
z
z V
z θ
&
….(16)
The numz of Eq. 16 shown above, has
one extra zero in the front, so, must
rid of it before closing the loop.
Eng. & Tech. Journal, Vol. 29, No. 1, 2011 Modeling, Analysis and Speed Control
Design Methods of a DC Motor
148
By generating the vector of discrete
output signals and connecting them,
the closedloop response without any
control will be shown as in Fig. 22.
From the design requirement, the
sampling time T = 0.1*τ = 0.1*0.6T
s
=
0.12sec, (which is 1/10 the time
constant τ of the system with a
settling time T
s
of 2sec) [8].
ii. PID Controller:
For mapping the continuoustime
transfer function for a PID controller
in Eq. 10 from the splane to zplane,
the bilinear transformation [10] will
be used. According to the PID design
method for the DC motor in Sec. 4,
K
P
=100, K
I
=200 and K
D
=10 are
satisfied the design requirement. All
of these gains will be used in this
example. Finally, the closeloop stair
step response will be plotted as
shown in Fig. 23.
It can be seen from this figure that,
the closedloop response of the
system is unstable. Therefore, there
must be something wrong with
compensated system. So, it should
take a look at root locus of the
compensated system shown in Fig.
24.
From this rootlocus plot, it can be
seen that the PID controller has a pole
at 1 in the zplane, outside the unit
circle and the system will be unstable.
The pole location can be changed by
changing the compensator design. It
is chosen to cancel the zero at 0.62.
This will make the system stable for
at least some gains.
Furthermore, an appropriate gain can
be chosen from the root locus plot to
satisfy the design requirements.
The new zplane will have a pole
at 0.625 instead of 1, which almost
cancels the zero of uncompensated
system as shown on the rootlocus
plot in Fig. 25. The MATALB results
are: ξ =1 and Mp=0.
Then MATALB will return the
appropriate gain and the
corresponding compensated poles,
and plot the closedloop compensated
response as shown in Fig. 26, where:
selected point=0.62650.0i,
K=68.3287, poles=0.6265,
0.7344±0.1294i, and 0.2275. The plot
shows that the settling time is less
than 2 sec and the overshoot is around
3%. In addition, the steadystate error
is zero. Therefore this response
satisfies all of the design
requirements.
8 Simulink Design Method for
Speed Modeling:
A basic feedback control system
shown in Fig. 3 represents a very
common block diagram form.
For the proposed system, the Plant
will be a DC motor.
i. Building the Model:
The motor system shown in Fig.
1 will be modeled by summing the
torques acting on the rotor inertia and
integrating the acceleration to give
the velocity. Also, Kirchoff's laws
will be applied to the armature circuit.
First, the integrals of the rotational
acceleration and of the rate of change
of armature current will be modeled:
dt
d
dt
d θ θ
=
∫ 2
2
, and i
dt
di
=
∫
…(17)
Next, both Newton's law and
Kirchoff's law will be started to
model. These laws applied to the
motor system to give the following
equations:
dt
d
b T
dt
d
J
m
θ θ
− =
2
2

¹

\

− =
dt
d
b i K
J dt
d
t
θ θ 1
2
2
….(18)
m m m
e V i R
dt
di
L − + − =
Eng. & Tech. Journal, Vol. 29, No. 1, 2011 Modeling, Analysis and Speed Control
Design Methods of a DC Motor
149

¹

\

− + − =
dt
d
K V i R
L dt
di
m m
m
θ 1
.(19)
ii. Openloop Response:
To simulate the motor system,
first, an appropriate simulation time
must be set. By selecting parameters
from the Simulation menu and
entering “3” in the Stop Time field,
the openloop response can be
viewed. The physical parameters
must be set as J=0.01; b=0.1;
K
m
=0.01; R
m
=1; and L
m
=0.5. Running
the simulation model shown in the
Fig. 27 will give the output on the
Scope shown in the Fig. 28.
iii. Extracting a Linear Model into
MATLAB:
A linear model of the system (in
state space or transfer function form)
can be extracted from a Simulink
model into MATLAB.
To verify the model extraction, an
openloop step response of the
extracted transfer function will be
generating in MATLAB. The
corresponding plot will be exactly
similar to that shown in Fig. 2 which
is equivalent to the Scope's output.
iv. Implementing Lag Compensator
Control:
In the motor speed control root
locus example in Sec. 5, a Lag
Compensator was designed with the
following transfer function:
50(s+1)/(s+0.01).
To implement this in Simulink, the
openloop system model in Fig. 27
will be contained in a Subsystem
block, and inserting a Lag
Compensator into a closedloop
around the plant model by feeding
back the plant output as shown in Fig.
29 . The output of the Sum block will
provide the error signal, which will be
feed into a Lag Compensator. Finally,
a step input is applying for viewing
the output on the scope.
v. Closedloop Response:
To simulate the proposed system,
first, an appropriate simulation time
must be set, then selecting parameters
from the Simulation menu and
entering “3” in the Stop Time field.
3sec is suitable for simulation, since it
is more than the desired settling time.
The physical parameters must be set.
As a result, the Scope output will be
shown as in Fig. 30.
9Conclusions:
There are many motor control
system design methods that may be
more or less appropriate to a specific
type of application. The designer
engineer must choose the best one for
his work. Therefore, the current work
makes comprehensive study of the
analysis, modeling and speed control
design techniques of a DC motor.
An important advantage of the
rootlocus method is that the roots of
the characteristic equation of the
system can be obtained directly,
which results in a complete and
accurate solution of the transient and
steadystate response of the
controlled variable.
The frequencyresponse approach
yields enough information to indicate
whether the system needs to be
adjusted or compensated and how the
system should be compensated.
PID controller enables the motor
to reach the speed smoothly and
within an acceptable period of time. It
found that using (after several trial
and error runs) a PID controller with
K
P
=100, K
I
=200, and K
D
=10, all of
the design requirements will be
satisfied, providing the desired
response
It has observed that both the
transfer function and the PID control
could have a large influence upon the
response of the system.
Eng. & Tech. Journal, Vol. 29, No. 1, 2011 Modeling, Analysis and Speed Control
Design Methods of a DC Motor
150
It is found the design process for
root locus is very much a trial and
error process.
Future suggestion is to replace the
conventional speed controller by
adaptive self tuning of PID controller.
10 References:
[1] M. S. RUSU, and L. Grama, The
Design of a DC Motor Speed
Controller, Fascicle of Management
and Tech. Eng., Vol. VII (XVII),
2008, pp. 10551060.
[2] J. J. Jordan, A DC Motor Drive
for a DynoMicrocontroller and
Power Electronics, University of
Queensland, Australia, BSc. thesis,
Oct. 2001.
[3] J. Santana, J. L. Naredo, F.
Sandoval, I. Grout, and O. J. Argueta,
“Simulation and Construction of a
Speed Control for a DC Series
Motor,” Mechatronics, Vol. 12, issues
910, Nov.Dec. 2002, pp. 1145
1156.
[4] S. Ayasun and C. O. Nwankpa,
DC Motor Speed Control Methods
Using MATLAB/Simulink and Their
Integration into Undergraduate
Electric Machinery Courses, IEEE
Trans Educ, March, (2007), 347354.
[5] A. S. Othman, “Proportional
Integral and Derivative Control of
Brushless DC Motor,” European
Journal of Scientific Research, Vol.
35 No. 2 (2009), pp. 198203.
[6] B. Allaoua, B. Gasbaoui and B.
Mebarki, “Setting Up PID DC Motor
Speed Control Alteration Parameters
Using Particle Swarm Optimization
Strategy”, Leonardo Electronic
Journal of Practices and
Technologies, Issue 14, Jan.June
2009, pp. 1932.
[7] M. H. Rashid, Power Electronics,
Circuits Devices and Applications,
2
nd
Edition, Prentice Hall
International, Inc., New Jersey, 1993.
[8] B. Shah, Field Oriented Control of
Step Motors, MSc. Thesis, SVMIT
Bharuch, India, Dec. 2004.
[9] W. P. Aung,'' Analysis on
Modeling and Simulink of DC Motor
and its Driving System Used for
Wheeled Mobile Robot'', World
Academy of Science, Engineering
and Technology 32, 2007, pp.299
306.
[10] B. Kuo, Automatic Control
Systems, PrenticeHall, Englewood,
Cliffs. NJ, 1995.
[11] N. S. Nise, Control Systems
Engineering (3
rd
Edition), John Wiley
and Sons Inc., New York, 2000.
[12] A. Klee, Development of a
Motor Speed Control System Using
MATLAB and Simulink,
Implemented with a Digital Signal
Processor, MSc. Thesis, Orlando,
Florida, 2005.
Eng. & Tech. Journal, Vol. 29, No. 1, 2011 Modeling, Analysis and Speed Control
Design Methods of a DC Motor
151
R Y
u
+

Plant Controller
e
Fig. 3: Feed back system schematic Fig. 4: Closedloop, step response;
K
P
= 100
Fig. 5: Step response using a PID Fig. 6: Step response using a PID
controller with small K
I
and K
D
controller with large K
I
+

V
R
m
e
m
=K
m
θ
.
+

bθ
.
θ
T
m
J
L
m
Fig. 1: The electric circuit of the Fig. 2: The Openloop step response of the
armature and the free body diagram motor system using T. F. or statespace;
of the rotor for a DC motor with initial speed=0.1 p.u. at steady state
Eng. & Tech. Journal, Vol. 29, No. 1, 2011 Modeling, Analysis and Speed Control
Design Methods of a DC Motor
152
ξ = 0.8
Fig. 9: Openloop system, step Fig. 10: Root locus with
 response with gain; K=10 a lag controller
ξ = 0.8
Fig. 7: Step response using a PID Fig. 8: Root locus of the
controller: K
P
=100, K
I
=200, K
D
=10 openloop system
Fig. 11: Closedloop stepresponse Fig. 12: Closedloop, stepresponse
with a lag controller and gain=20 with a lag controller and gain=50
Eng. & Tech. Journal, Vol. 29, No. 1, 2011 Modeling, Analysis and Speed Control
Design Methods of a DC Motor
153
K
x
R
y
u

x = Ax + By
y = Cx
.
Fig. 17: Step response with a lag Fig. 18: The schematic for a full
controller and gain=50  state feedback system
Fig. 13: Bode plot of the original open Fig. 14: Bode plot of the plant
 loop transfer function with proportional gain=72
Ф
m
Ф
m
Ф
m
Fig. 15: Step response of the open Fig. 16: Bode plot of the plant
 loop system with gain=72 with a lag controller and gain=50
K
x
y
u
+

x = Ax + By
y = Cx
.
N
R
Fig. 19: Closedloop step response Fig. 20: Schematic for a fullstate feed
with a K controller back system with adding Nbar gain=10
Eng. & Tech. Journal, Vol. 29, No. 1, 2011 Modeling, Analysis and Speed Control
Design Methods of a DC Motor
154
Fig. 25: Root locus of the stable system Fig. 26: Stairstep stable response
with PID controller
p =  0.625
p =  1
z =  0.62
Fig. 23: Closeloop stairstep response Fig. 24: Root locus of the compensated
system with discrete PID controller
Fig. 21: Closedloop, stepresponse Fig. 22: Closedloop, stairstep
with adding Nbar =10 response with no control
Fig. 27: Simulation of the openloop Fig. 28: Openloop step response;
motor system model scope output
Eng. & Tech. Journal, Vol. 29, No. 1, 2011 Modeling, Analysis and Speed Control
Design Methods of a DC Motor
155
Appendix
Table A: Physical parameters of
the DC motor
Fig. A: Transient response (closed loop)
of a DC motor
Fig. 29: Closedloop system using Fig. 30: Closedloop step response;
lag compensator control scope output
J
Moment of Inertia of
the Rotor (kg.m
2
)
0.01
b
Damping ratio of the
Mechanical System
(Nms)
0.1
K
m
Motor Constant
(Nm/A)
0.01
R
m
Motor Electric
Resistance (Ω)
1
L
m
Motor Electric
Inductance (H)
0.5
V Input Voltage (Volt) 6
θ
&
Rotating Speed (r/s) 600
Eng.(5) (L s + R ) I ( s) = V ( s) − K s θ&( s) Lm m m m …. The design method uses the concepts of the system theory. This paper was presented the system contains a DC motor based on the motor specifications needs to be obtained. by a constant factor Kt.MATLAB Control Design Method for Speed Modeling: i. the above equations can be expressed in terms of sdomain. 2.(3) di & + Rm i = V − K mθ …. & Tech.System Equations: To perform simulations of a system. (2) m e Assuming that. et al presented a new design method to determine optimal PID controller parameters using the Particle Swarm Optimization method. The motor torque. Othman successfully designed a speed controller for closed loop operation of the BLDC motor.. θ by the following equations [7]: Tm = K t i ….(6) By eliminating I(s). For this purpose assuming (load torque) TL=0 and (friction torque) Tf=0. since neither affects the transfer function. is related to the armature current. its simplified mathematical model has been used. an appropriate model needs to be established. This model consists of differential equations for the electrical part. the following equations can be written based on Newton’s law combined with Kirchhoff’s law: & & J θ& + b θ = K m i − T L …. and designing a DC motor speed controller based on system theory concepts. Therefore 142 . The current paper intends presenting comprehensive study for all the techniques of modeling. direct and inverse Laplace transforms. From Fig. armature voltage and resistance) and feedback control system for DC motor drives. Vol. analyzing. This requires building the appropriate Laplace model for each component of the whole control system. such as signals and systems. When the motor is used as a component in a system.Transfer Function Using Laplace Transforms. In [6]. (1) & e =K θ …. the following openloop transfer function can be obtained. No. All the values for the physical parameters are listed in Table A in the Appendix. where the rotational speed & θ is the output and the voltage V is the input. Allaoua. 1 and Table A. In [5]. it is desired to describe it by the appropriate transfer function between the motor voltage and its speed. & s ( J s + b ) θ ( s ) = K m I ( s ) − TL ( s ) . This is achieved by developing the open loop transfer function of a DC motor with the use of the system equations of the motor as given by Rashid [7]. Kt (torque constant) = Ke (electromotive force constant) = Km (motor constant). Journal. (4) dt a. The back emf. 29. mechanical part and the interconnection between them. 1. i. em. Analysis and Speed Control Design Methods of a DC Motor and Nwankpa described the MATLAB/Simulink realization of the DC motor speed control methods (field resistance.. The electric circuit of the armature and the free body diagram of the rotor are shown in the Fig. Tm. transfer functions. 1. In order to build the DC motor’s transfer function.. is related & to the rotational speed.. 2011 Modeling. The speed control design methods used are studied under MATLAB and Simulink in the following sections.
where it can be assumed that the plant is a DC motor whose speed must be controlled and the Controller provides the excitation for the plant. (10% of its steadystate speed). the motor can only achieve a maximum speed of 0. From the plot. 3. Km b − & & 0 d θ J J θ + 1 V = K dt i − m − Rm i L m Lm Lm θ& & θ = [1 0] i …. it makes the motor takes 3sec to reach its steadystate speed.1 p. this does not satisfy the 2sec settling time criterion. No. Also. the difference between the desired input value R and the actual output Y. peak overshoot.8. designed to control the overall system behavior. Analysis and Speed Control Design Methods of a DC Motor & θ ( s) V ( s) T L = ( s) = 0 (J s + b)(Lm s + Rm ) + K m Km 2 change in speed with 0. StateSpace Representation: In the statespace form.. 3. Different characteristics of the motor response (steadystate error.respectivel y.9] (7) b. 1. The definition of these parameters can be demonstrated as in [10]. Considering the unity feedback system architecture as shown in Fig. 3 and 4 can be expressed by choosing & θ and i as the state variable and V as an input. a peak overshoot Mp of less than 5% is wanted. The motor is assumed to be exposed at steadystate speed to step The system can also be represented using the statespace Eqs. & Tech. [1.u. 3.(8) ii.11 p. Eqs. The motor parameters were determined on the basis of its step response about a steadystate operating point. taking a look at how the PID controller works in the closedloop system shown in Fig.). The other performance requirement is that the motor must accelerate to its steadystate speed θ&ss as soon as it turns on.u. it can be seen that when a step input voltage is applied to the system at steadystate speed. 29. ten times smaller than our desired speed (1 p. The variable e represents the tracking error. 143 . The output is chosen to be θ& . Journal. rise time. the steadystate error ess of the motor speed should be less than 1%.) are controlled by selection of the three gains that modify the PID controller dynamics. it is desirable to have a settling time Ts in less than 2sec.Continuous PID Control Design Method: PID controllers are commonly used to regulate the timedomain behavior of many different types of dynamic plants [8]. 2. 2011 Modeling. Since speed faster than the reference may damage the equipment. 8 to get the same output plot obtained in Fig. 7 can be represented into MATLAB by defining the numerator (num) and denominator (den) matrices as follows: num=Km and den=(Js+b)(Lms+Rm)+Km2. First. iii. and the controller computes both the derivative and the integral of this error signal. This error signal will be sent to the PID controller. Vol. In this case. 2. etc.MATLAB Representation and Openloop Response: The transfer function in Eq.Design Requirements: Since the most basic requirements of a motor are that it should rotate at the desired speed.Eng.u. and the corresponding plot is shown as in Fig. namely input voltage V= 6 volts & and speed θ = 600 r/s.
Taking the Laplace transform of Eq. Obtaining an openloop response and determining what needs to be improved. KI. Analysis and Speed Control Design Methods of a DC Motor Therefore. The controller takes e and computes its derivative and it’s integral again. 4. 10. Changing KI to 200 makes the response looks as in Fig. ii. KD must increase to reduce the overshoot. So. and KD=Derivative gain. the PID controller can be set to give the desired performance. 6. it can be seen that. where KP = Proportional gain. the PID controller can be designed and added into the system.1 p. Using KP with a gain of 100. 3. and KD until obtaining a desired response. the resulting response will be plotted as in Fig. 5. the following steps must be followed to obtain a desired response. = 1. 6 seems much faster than before. step input to the steadystate speed and the design criteria previously mentioned. Adjusting each of KP. KI. The settling time shown in Fig. The signal u will be sent to the plant. Adding a KP to improve the rise time. 7.Eng. Vol. 4. 2011 Modeling. recalling the transfer function of a PID controller in Eq. By adjusting the weighting constants KP. Effects of each of these parameters on a closedloop system are summarized as in [8]. This process goes on and on. Changing KD to 10 can get the plot seen in Fig. but the large KI has worsened the transient response (big overshoot). With adding 0. The response in the Fig. the step response can be shown as in Fig. it must be known that using (after several trial and error runs) a PID 144 . 5 is too long. 29. 5. Adding a KI to eliminate the 4. No. and derivative gains that make up the PID compensation. & Tech. and the new output Y will be obtained and sent back to the sensor again to find the new error signal e. the simulation start with the best initial gains.(10) This transfer function clearly illustrates the proportional. Therefore. Trying a PID controller with small KI and KD. 2. steadystate error.u. The MATLAB is used to obtain a closedloop transfer function directly from the openloop transfer function. Adding an integral term will eliminate the steadystate error and a derivative term will reduce the overshoot. Adding a KD to improve the overshoot (increase damping). KI = Integral gain. So it is necessary to increase KI to reduce the settling time. Design Steps of PID Controller: To designing a PID controller for a given system. 1. integral. Journal. the PID controller is defined by the relationship between the controller input e and the controller output u that is applied to the motor armature: de u = K P e + K I ∫ e dt + K D (9) dt . Tuning the Gains: By tuning the gains of the PID controller and producing the optimum response using trial and error method (it provides simplest way to achieve a good compensator [11]). 9 gives the following transfer function: K U (s) K (s) = = KP + I + KD s E (s) s KD s2 + KP s + KI [8] s …. both the steadystate error and the overshoot are too large. From the plot in Fig. and KD. i.
5 …. halfway between the real axis and the damping requirement.0401i. but should at least have the same order of magnitude. 8 shows a very simple root locus. Using these criteria results a root locus plot shown in Fig. it might however increase the settling time. first. 2011 Modeling. Finding the Gain: Recalling that. providing the desired response. Journal. 29. When a point is selected on the root locus plot in Fig.(13) . If the gain is increased to reduce the steadystate error. Three arguments must be founded are the damping ratio (zeta) term (ξ > 0. The damping and settling time criteria were met with the proportional controller. and the sigma term (4. Mp %=0. the overshoot becomes too large. say at 6+2. iv. Large damping corresponds to points on the root locus near the real axis.(14) .Eng. 1.Root Locus Control Design Method: For designing a controller using the root locus method. So it is necessary adding a Lag Controller (in transfer function form) to reduce the steadystate error. iii.8 corresponds to an overshoot Mp < 5%). the 145 . the natural frequency (ωn) term (= 0 corresponds to no rise time criterion). it should follow the following steps: i. The following three equations are used in continuous system designs: ξ ωn ≥ 4. By doing this. So.3 sec). Adding a Lag Controller: The plot in Fig.0708. poles=6. the settling time and the overshoot must be as small as possible. there is no overshoot and the settling time is about 1sec. the MATLAB should return the output similar to the following: Selected point=6. Openloop Root Locus: The main idea of root locus design is to find the closedloop response from the openloop root locus plot. it is necessary viewing in the root locus for the plant. Tr: Rise time and Mp: Peak overshoot [11].0000±2. With the above results shown in Fig. Then by adding zeros and/or poles to the original plant. Vol. A Lag Compensator [11] can reduce the steadystate error. A fast response corresponds to points on the root locus far to the left of the imaginary axis. Noticing that. 4. which looks very similar to the original one.8 / Tr ….0204+2. Closedloop Response: By closing the loop and seeing the closedloop step response.01) …. Analysis and Speed Control Design Methods of a DC Motor controller with KP=100.6 / Ts (Sigma term) (11) ωn ≥ 1. and KD=10. KI=200.05i. The gain can be found and the step response plotted using this gain all at once. the steadystate error is about 50%. K=10. gain. ii. The only problem can be seen from this plot is that. 8. No.6/2 = 2. By using the following lag controller first: (s + 1) / (s + 0. Matlab can find the gain corresponding to a point on the root locus. where ξ: Damping ratio. The steadystate error is the only criterion not met with the proportional controller. all of the design requirements will be satisfied.0101 and the ξ = 0. the values returned in MATLAB results may not be exactly the same. so the overshoot and settling time requirements are satisfied. the closedloop response can be modified.947.(12) ξ ≥[((lnMp /π)2/(1+(lnMp /π)2)]0. ωn: Natural frequency. 8.0400i. 9. the root locus will be shown as in Fig. Ts: Settling time. 10. it can be seen that. & Tech.
the design process for root locus is very much a trial and error process. gain. but the overshoot and the steadystate error are too high. 13 (it looks to be slightly more than 40dB.9104. K=19. 29. 13.Frequency Control Design Method: i. ii.0139≈72 (≈37dB) to get the Bode plot shown as in the Fig. 14. It may also note that even though the gain was selected to correlate with a position close to the damping criterion. When prompted to select a point. 12 shows the corresponding response with gain =50. 16 with 146 . 14. & Tech. and getting the higher gains without worrying about the overshoot. It can be seen from the figure that the steadystate error is smaller than 1%. it concluded that. and 0. Fig.0139. gain. Vol. it can be seen that the phase margin Фm [11] can be greater than about (18001200=600) if ω is less than 10 r/s. The gain at 10 r/s can be found by reading off the Bode plot in Fig. Adding Proportional Gain: From the bode plot in Fig. and plotting the response all in one step. First. Adding a Lag Controller: Reducing the gain to 50.5714+8.01 in magnitude) and the exact magnitude: mag = 0.Eng. 15. It should look similar to the following (using a gain of around 50): Selected point= 5.6122+ 4. If it has not been able to get a satisfactory response by choosing the gains. 5. K = 49. its pole is much slower. or 0. 2011 Modeling. therefore changing the closedloop response. iv. From this plot: (mag=0. poles = 5. 13.4445i. This is due to the effect of the lag controller. The figure shows that the settling time is fast enough. Journal. 11. The resultant plot will be shown as in Fig. and trying the lag controller in Eq. Adding gain to the system so the bandwidth frequency [11] is 10 r/s. the Bode plot for the original openloop transfer function is drawn as shown in Fig. As result. it may be needed to add a control gain of 1/0. Keep trying until getting a satisfactory response. 14. To have an openloop gain of 1 (0dB) at 10 r/s.0139. Closedloop Response: From the plot in Fig.9563.7824. That is why it is nice to plot the root locus. it can have tried a different lag controller.8876. pick one that is near the damping requirement (diagonal dot line in Fig. Adding a controller to the system changes the openloop Bode plot. phase = 123. What this means is that it can go beyond the dotted lines that represent the limit. No.6835.01=100 (but could increase the settling time) gives the bode plot shown in Fig. picking the gain. the overshoot is not even close to 5%.5498 ± 8. and the settling time and overshoot requirements have been met.8887i and 0. 1. It can be seen from the figure that the response is not quite satisfactory. poles=5. 10). Analysis and Speed Control Design Methods of a DC Motor following Matlab results will be obtained: selected point = 5.6138±4. The gain should be about 20. which will give Фm of about 600. it can be seen that the phase margin is now quite large and the corresponding closedloop response looks like as in Fig. or even added a Lead Controller.8889i. which should reduce the steadystate error by a factor of 1/0. and ω=10 r/s). Original Bode Plot: The main idea of frequencybased design is to use the Bode plot of the openloop transfer function to estimate the closedloop response. iii.4444i.
Journal.5) [11] corresponds to ξ = 0. No. (1113).3600±1. singleinput system with fullstate feedback using the schematic shown in Fig. Analysis and Speed Control Design Methods of a DC Motor (K=50≈34dB) with good phase margin (≈700). Once coming up with the wanted poles. Now. Continuous to Discrete Conversion: The first step in designing a discrete control system is to convert the continuous transfer function in Eq. the statespace equations become: & X = ( A − BK )X + Bu Y = CX [11] ….Discrete PID Control Design Method: A digital DC motor model can be obtained from conversion of the analog model. ii. 1. must rid of it before closing the loop. It can see that adding a lag controller can reduce the steadystate error and at the same time. By designing a fullstate feedback controller.0048 and poles = 4. 19. 6. Computing the input can be done in one step by adding a constant gain Nbar after the reference R. The characteristic polynomial for this closedloop system is the determinant of (sI(ABK)) where s is the Laplace variable. Closing the loop gives the step response shown in Fig. 17. K. Designing the Fullstate Feedback Controller: The schematic for a fullstate feedback system is shown as in the Fig. 147 .A StateSpace Control Design Method: For the main problem. and the plot of the closedloop response can be shown as in Fig. 16 shown above. 8. 18. the steadystate error is much less than 1% and all the other design criteria have been met as well. the integral and derivative terms of the controller have to be approximated. Since the matrices A and BK are both 2x2 matrices. 19.0000i. 2011 Modeling. it can be seen that the steadystate error is too large. The controller for this example will be designed by a PID method. & Tech. The Matlab can find the scale factor N which will eliminate the steadystate error to a step reference for a continuoustime. 20. 7. The continuous time signal e is sampled at a sample period T [8]. 7 to a discrete one [12] as follows: & θ ( z) 0. i.0877 z + 0.1% overshoot and a sigma=ξωn=5 which leads to≈1sec settling time by using Eqs. the dynamic equations in statespace form are given as in the Eqs. The integral becomes a sum and the derivative becomes a difference. The first try is to place them at: 5±i . Vol. so. which meets the design requirements.2369 …. 18. has one extra zero in the front.(16) The numz of Eq. 29.(15) .0057 = 2 V ( z ) z − 1.98 which gives 0. MATLAB will find the controller matrix.( poles= ξωn ± jωn (1ξ2)0.Eng. The corresponding step respond plot is shown in the Fig. 21 and the MATLAB results are: Nbar=10. After adding the K matrix into the system in Fig. i. these two poles can move anywhere wanted them.0092 z + 0. the overshoot can be reduced by reducing the gain. To discretise the controller. Adding a Reference Input: From the plot in Fig. there should be 2 poles for the system.
The plot shows that the settling time is less than 2 sec and the overshoot is around 3%. In addition. All of these gains will be used in this example. Building the Model: The motor system shown in Fig. the closeloop stair step response will be plotted as shown in Fig. From the design requirement. According to the PID design method for the DC motor in Sec.6Ts= 0.6265.3287. 10 from the splane to zplane.62. it should take a look at root locus of the compensated system shown in Fig. ii. 2011 Modeling. poles=0. Kirchoff's laws will be applied to the armature circuit. & Tech.Simulink Design Method for Speed Modeling: A basic feedback control system shown in Fig. 22.625 instead of 1. where: selected point=0. Therefore this response satisfies all of the design requirements. 29. It is chosen to cancel the zero at 0. The MATALB results are: ξ =1 and Mp=0.2275.1294i. the closedloop response without any control will be shown as in Fig. It can be seen from this figure that. 26. For the proposed system. Therefore. This will make the system stable for at least some gains.Eng. These laws applied to the motor system to give the following equations: ∫ ∫ J d 2θ dθ = Tm − b 2 dt dt 2 d θ 1 dθ = Kt i − b 2 dt J dt di = − Rm i + V − em dt …. the closedloop response of the system is unstable. and plot the closedloop compensated response as shown in Fig. 4. Journal.7344±0. 25. which almost cancels the zero of uncompensated system as shown on the rootlocus plot in Fig. 8. K=68. KP=100. the sampling time T = 0. Furthermore.(18) Lm 148 . and 0. the Plant will be a DC motor. it can be seen that the PID controller has a pole at 1 in the zplane. 0. KI=200 and KD=10 are satisfied the design requirement.12sec. the bilinear transformation [10] will be used. outside the unit circle and the system will be unstable. and 2 dt dt dt Next. The pole location can be changed by changing the compensator design. The new zplane will have a pole at 0.1*0. Vol. 3 represents a very common block diagram form. (which is 1/10 the time constant τ of the system with a settling time Ts of 2sec) [8]. the steadystate error is zero. Finally. So. From this rootlocus plot. the integrals of the rotational acceleration and of the rate of change of armature current will be modeled: d 2θ dθ di = i …(17) = . No. there must be something wrong with compensated system. both Newton's law and Kirchoff's law will be started to model. 1 will be modeled by summing the torques acting on the rotor inertia and integrating the acceleration to give the velocity. Also. 1. 24. Then MATALB will return the appropriate gain and the corresponding compensated poles.62650. First. i.0i. an appropriate gain can be chosen from the root locus plot to satisfy the design requirements. PID Controller: For mapping the continuoustime transfer function for a PID controller in Eq. Analysis and Speed Control Design Methods of a DC Motor By generating the vector of discrete output signals and connecting them. 23.1*τ = 0.
01).Eng. providing the desired response It has observed that both the transfer function and the PID control could have a large influence upon the response of the system. which will be feed into a Lag Compensator. an appropriate simulation time must be set. which results in a complete and accurate solution of the transient and steadystate response of the controlled variable. An important advantage of the rootlocus method is that the roots of the characteristic equation of the system can be obtained directly. the current work makes comprehensive study of the analysis.(19) dt Lm dt ii. As a result. Openloop Response: To simulate the motor system. iii. the openloop system model in Fig. 27 will be contained in a Subsystem block. since it is more than the desired settling time. 27 will give the output on the Scope shown in the Fig. 2 which is equivalent to the Scope's output. Finally. Running the simulation model shown in the Fig. a Lag Compensator was designed with the following transfer function: 50(s+1)/(s+0. 29. The corresponding plot will be exactly similar to that shown in Fig. To verify the model extraction. iv. the Scope output will be shown as in Fig. The output of the Sum block will provide the error signal. 30. Km=0. 29 . The designer engineer must choose the best one for his work.1. 9Conclusions: There are many motor control system design methods that may be more or less appropriate to a specific type of application. b=0. Implementing Lag Compensator Control: In the motor speed control root locus example in Sec. By selecting parameters from the Simulation menu and entering “3” in the Stop Time field. all of the design requirements will be satisfied. The physical parameters must be set as J=0. then selecting parameters from the Simulation menu and entering “3” in the Stop Time field. PID controller enables the motor to reach the speed smoothly and within an acceptable period of time. No. The physical parameters must be set.01. 149 . 3sec is suitable for simulation. Therefore. an openloop step response of the extracted transfer function will be generating in MATLAB. first. Rm=1. Journal. To implement this in Simulink. Vol. first. Extracting a Linear Model into MATLAB: A linear model of the system (in state space or transfer function form) can be extracted from a Simulink model into MATLAB. The frequencyresponse approach yields enough information to indicate whether the system needs to be adjusted or compensated and how the system should be compensated. an appropriate simulation time must be set.01. modeling and speed control design techniques of a DC motor. 28. the openloop response can be viewed. 1. Analysis and Speed Control Design Methods of a DC Motor 1 di dθ = − Rm i + V − K m .5. and KD=10. and inserting a Lag Compensator into a closedloop around the plant model by feeding back the plant output as shown in Fig. 5. It found that using (after several trial and error runs) a PID controller with KP=100. v. a step input is applying for viewing the output on the scope. 2011 Modeling. & Tech. Closedloop Response: To simulate the proposed system. and Lm=0. KI=200.
J. 2000.. MSc. Oct. [2] J. Jordan. P. “Simulation and Construction of a Speed Control for a DC Series Motor. Vol. Rashid. [5] A. thesis. Kuo. Othman. Nise. Naredo. Dec. Field Oriented Control of Step Motors. SVMITBharuch. 10.. 2004. 347354. Mebarki. 1. 2 (2009). pp. H. RUSU. PrenticeHall 2nd International. Englewood. PrenticeHall. 10551060. India. NJ. 1993. [9] W. Orlando. 12. Circuits Devices and Applications. Nwankpa. Santana. Engineering and Technology 32. [12] A. New Jersey. University of Queensland.” Mechatronics. Control Systems Engineering (3rd Edition). DC Motor Speed Control Methods Using MATLAB/Simulink and Their Integration into Undergraduate Electric Machinery Courses. March. 2005. “Setting Up PID DC Motor Speed Control Alteration Parameters Using Particle Swarm Optimization Strategy”. 29. Vol. [7] M. No. L.References: [1] M. Inc. Aung. pp. 2008. J. I. John Wiley and Sons Inc. 2011 Modeling. pp. Jan. Analysis and Speed Control Design Methods of a DC Motor It is found the design process for root locus is very much a trial and error process. Grama. Allaoua. 11451156. 1995. S. S. Fascicle of Management and Tech. J. Nov. Power Electronics. issues 910. 35 No. A DC Motor Drive for a DynoMicrocontroller and Power Electronics. 2001. New York. Klee. MSc. 2002. [8] B. Vol. 150 . Cliffs. Development of a Motor Speed Control System Using MATLAB and Simulink. Argueta.June 2009. Implemented with a Digital Signal Processor. Journal. IEEE Trans Educ. Australia. [6] B. Shah. B. (2007).” European Journal of Scientific Research. Issue 14. Florida. Grout. Thesis. [4] S. Gasbaoui and B. Future suggestion is to replace the conventional speed controller by adaptive self tuning of PID controller. VII (XVII). BSc.'' Analysis on Modeling and Simulink of DC Motor and its Driving System Used for Wheeled Mobile Robot''. Thesis. S. “Proportional Integral and Derivative Control of Brushless DC Motor. pp. Sandoval. Leonardo Electronic Journal of Practices and Technologies. 1932. The Design of a DC Motor Speed Controller. Edition.Dec. O. and L.. 198203. Vol. Eng. Ayasun and C. 2007. [10] B. and O. [3] J. World Academy of Science. pp. Automatic Control Systems.Eng. F.299306. & Tech. [11] N.
with initial speed=0.1 p. F. 6: Step response using a PID controller with large KI 151 . 1: The electric circuit of the armature and the free body diagram of the rotor for a DC motor Fig. No. em=Km θ bθ . 3: Feed back system schematic Fig. or statespace. KP = 100 Fig. step response. 1. 5: Step response using a PID controller with small KI and KD Fig.Eng. J Fig. Analysis and Speed Control Design Methods of a DC Motor Rm Lm Tm + V + θ  .u. 4: Closedloop. Vol. Journal. 2011 Modeling. at steady state R e u Controller Y Plant +  Fig. 29. & Tech. 2: The Openloop step response of the motor system using T.
KD=10 Fig. 7: Step response using a PID controller: KP=100. 9: Openloop system.response with gain. K=10 Fig. Journal. 11: Closedloop stepresponse with a lag controller and gain=20 Fig. Vol. KI=200. 12: Closedloop. No.8 Fig. 29.8 Fig.Eng. Analysis and Speed Control Design Methods of a DC Motor ξ = 0. 8: Root locus of the openloop system ξ = 0. step . 1. 2011 Modeling. & Tech. 10: Root locus with a lag controller Fig. stepresponse with a lag controller and gain=50 152 .
Eng. 2011 Modeling. Journal. Vol. Analysis and Speed Control Design Methods of a DC Motor Фm Фm Fig. 15: Step response of the open . 14: Bode plot of the plant with proportional gain=72 Фm Fig. y K x Fig. y K Fig. 19: Closedloop step response Fig. 20: Schematic for a fullstate feed with a K controller back system with adding Nbar gain=10 153 . 18: The schematic for a full . 17: Step response with a lag controller and gain=50 Fig.loop system with gain=72 Fig.loop transfer function Fig. 13: Bode plot of the original open . & Tech. 29.state feedback system R + N u  x = Ax + By y = Cx . 1. 16: Bode plot of the plant with a lag controller and gain=50 R  u x = Ax + By y = Cx x . No.
1.0.625 Fig. 25: Root locus of the stable system Fig. 22: Closedloop. 2011 Modeling.62 Fig. Vol. 23: Closeloop stairstep response Fig. stepresponse with adding Nbar =10 Fig. & Tech.Eng.0. 21: Closedloop. Journal. 26: Stairstep stable response with PID controller Fig. stairstep response with no control p=1 z = . scope output 154 . 27: Simulation of the openloop motor system model Fig. 28: Openloop step response. Analysis and Speed Control Design Methods of a DC Motor Fig. No. 24: Root locus of the compensated system with discrete PID controller p = . 29.
1. A: Transient response (closed loop) of a DC motor 155 . 30: Closedloop step response.1 0. Vol.01 0. & Tech. scope output Appendix Table A: Physical parameters of the DC motor J b Km Rm Lm V θ& Moment of Inertia of the Rotor (kg.01 1 0.5 6 600 Fig. No. Analysis and Speed Control Design Methods of a DC Motor Fig.Eng.m2) Damping ratio of the Mechanical System (Nms) Motor Constant (Nm/A) Motor Electric Resistance ( ) Motor Electric Inductance (H) Input Voltage (Volt) Rotating Speed (r/s) 0. Journal. 29: Closedloop system using lag compensator control Fig. 29. 2011 Modeling.
This action might not be possible to undo. Are you sure you want to continue?
We've moved you to where you read on your other device.
Get the full title to continue listening from where you left off, or restart the preview.