Electr Eng
DOI 10.1007/s00202-017-0592-5
ORIGINAL PAPER
Backstepping control of a separately excited DC motor
Abdelkader Harrouz1 · Houcine Becheri2 · Ilhami Colak3
· Korhan Kayisli3
Received: 10 November 2016 / Accepted: 20 June 2017
© Springer-Verlag GmbH Germany 2017
Abstract In the last decades, researches on speed control
of electric motors have been very important in industrial
systems. On the other hand, the development of power electronics and semiconductor technologies has expanded the
scope of the AC and DC machines. However, the separately
excited DC motor is widely used in the field of variable speed
applications due to its simplicity to control. In this paper,
we present the control of a separately excited DC motor
using an analogue PI controller, a backstepping method
and an adaptive backstepping control techniques. Simulations are performed in MATLAB Simulink software, and
improved results of adaptive backstepping controller have
been obtained overall system performance compared to conventional PI and backstepping.
Keywords Control · Modelling · Backstepping · PI · DC
motor · Adaptive backstepping
B
Ilhami Colak
ilhcol@gmail.com
Abdelkader Harrouz
harrouz@univ-adrar.dz
Houcine Becheri
houcine.becheri@gmail.com
Korhan Kayisli
korhankayisli@gmail.com
1
Department of Hydrocarbon and Renewable Energy,Faculty
of Science and Tech, University of Adrar, Adrar, Algeria
2
Department of Electrical Engineering, Faculty of Technology,
University of Bechar, Béchar, Algeria
3
Department of Electrical Electronics Eng., Faculty of
Engineering Architecture, Nisantasi University, Istanbul,
Turkey
List of symbols
DC
AC
PI
PM
Va
Ra
Ia
La
Lm
if
Vf
Rf
Lf
ωr
ω
R
Rh
u1
u2
u max
Kp
θ
Hc
Vc
s
P
K pω
K iω
Kθ
e1
fc
V1
V2
Direct current
Alternating current
Proportional–integral controller
Power motor
Armature voltage
Armature resistance
Armature current
Armature inductance
Mutual inductance
Field current
Field voltage
Field resistance
Field inductance
Armature (rotor) speed
Change of speed
Nominal armature resistance
Rheostat resistance
First value of the voltage
Second value of the voltage
Maximal value of the voltage
Static gain of the proportional controller
Position
Transfer function of chopper
Control voltage of chopper
Laplace variable
Poles number
Static gain of the proportional speed controller
Static gain of the integral speed controller
Static gain of position controller
Error of rotor speed
Coefficient of friction
First Lyapunov function
Second Lyapunov function
123
Electr Eng
e2
k1
k2
e3
U
L
V3
Rf
E ref
Error of current and the armature current ia
First coefficient of armature voltage
Second coefficient of armature voltage
Final signal error
Voltage
Inductance
Third Lyapunov function
Field resistance
Set point weaken the flow ωref reference speed
1 Introduction
DC motors are a type of devices that transform electrical
energy to the mechanical energy. The usage of DC motors
might be restricted since distribution systems are based on
AC current but DC motors are used in a huge number of
industrial applications. In particular, separately excited DC
motors have many fields of applications [1–4]. However, for
some applications, it may be an advantage to use DC current
motors powered by static converters that convert AC to DC.
DC motors have many benefits such as flexibility, continuous
and instantaneous speed. Therefore, they have been widely
used in many fields including electric traction, positioning a
radar and handling robots [5].
In addition, the variable speed electrical motors have very
important roles in the industrial and transportation systems.
In recent years, the development in power electronics has
expanded the scope of the DC machines [6,7]. For instance,
the separate excitation DC machine is widely used in the field
of variable speed applications with the simplicity of control.
There are several existing solutions for speed control of a
DC motor for separate excitation. The principle speed control of separately excited DC motors has developed from the
equation of electromotive force of the motor [8,9].
The speed control of a DC motor can be achieved using
the following methods:
• varying excitation flux
• varying the resistance of armature circuit
• varying the supply voltage of armature circuit
Because of its simplicity and stability, electric drives use conventional controller “proportional–integral” (PI) to control
the current, velocity or position. In a practical implementation, some physical characteristics of the motor can vary
during operation resulting in parametric variations on the
system model [10,11].
In the context of tracking, the basic idea of backstepping
is to make the loop system equivalent to stable cascade subsystems in the sense of Lyapunov, which gives it qualities
robustness and asymptotic overall stability of the tracking
123
error. For a large class of systems, this technique is a systematic and recursive method of synthesis of nonlinear control
laws. Thus, at each step of the process, a virtual command is
generated to ensure the convergence of the first-order subsystems characterizing the continuation of trajectories towards
their equilibrium states (zero tracking errors in the deterministic and non-perturbed case) [12–14]. This technique
allows the synthesis of robust control law despite some misunderstanding of the parameters of the system and some
perturbations [15,16].
The objective of this study was to apply control techniques, it is the PI, and backstepping adaptive backstepping
on a DC motor powered by a chopper. The contribution is
composed of five parts. The first is devoted to the presentation
and modelling of the DC motor and speed control methods
of DC Motor. The second part is reserved for motor control
by the PI controller. In the third part, we give the principle and the general algorithm of backstepping followed by
an application of our model. Then, the fourth we will apply
the command by adaptive backstepping. Finally, simulation
results are present and conclusions are drawn from the last
part of this article.
2 Modelling of a DC motor
Many researchers have addressed the control of electrical
machines and specifically the control of DC motors [17–
19]. Several authors have tackled the control problem of
electrical machines operating in the field-weakening region.
For example, Asghar et al. [20], Zambada [21], Krim et al.
[22], Seibel et al. [23] dealt with the control of induction
motors operating in the field-weakening region. Tan et al. [24]
explored the field-weakening control of a permanent magnet synchronous motor (PMSM). Hesari [25] studied design
and implementation of maximum solar power tracking system using photovoltaic panels, and Bakshi and Bakshi [26]
proposed a field-weakening speed control system for a selfregulated synchronous motor. Nouira et al. [27] studied a
contribution to the design and the installation of an universal
platform of a wind emulator using a DC motor.
The equations of a DC machine for continuous permanent
magnet arrangement can be fulfilled by the equations of the
field voltages and frame, as well as the equation connects
torque with speed. The equations of the motor assuming a
simple model of the load are given as follows [28–30]:
di a
+ L m i f ωr
dt
di f
Vf = Rf i f + L f
dt
Va = Ra i a + L a
(1)
(2)
Electr Eng
Cm
Cm
(R + R h1 )
(R + R h2 )
r2
r1
r0
f3
Crn
Cr
r3
f2
f1
R h1 = 0
r
rn
Fig. 1 Torque-speed curve of a DC motor at different armature resistance
Torque and rotor speed are combined with the following relationship:
Ce = L m i f i a = J
dωr
+ Fc ωr + Cr
dt
(3)
r
Fig. 2 Torque-speed curve of a DC motor at different fluxes
• the reduction in the rigidity of the mechanical characteristic
• the decrease in overload capacity
• large power losses by the Joule effect
3.2 Speed control by the flux changes
The parameters and nominal conditions of the DC machine
are chosen as given in Table 1, in Appendix.
If the drop of resistive voltage is neglected before, the voltage
(Va >> Ra Ia ) expression of the speed becomes as given in
Eq. 5. [26,28]:
3 Speed control methods of a DC motor
The motor operating point is fixed by the intersection of the
curves Cm (n) and Cr (n) as shown in Fig. 1. The speed control
is to move the characteristic Cm (n) so as to obtain a different
operating point. The adjustment process may be mechanical
or electrical. The mechanical process is a change of the speed
by means of a gear that changes the ratio. Electric processes
are for three variables dependent on speed, namely the flux
φf , the armature voltage Va and the resistance of the armature
circuit [24].
3.1 Speed control by insertion of a resistance into
armature circuit
This method is performed by using an adjustable resistor
inserted in the armature circuit. It is a very simple method
allows adjustment in the direction to decrease from its nominal speed. The speed–torque characteristic of separately
excited DC motor (or shunt) can be expressed by the following equation:
Ra
Va
−
Ia = ωro − ω
ωr =
K φf
K φf
(4)
The voltage and flux are fixed at their nominal values, and by
this way, it can reduce the speed by increasing the armature
resistance by a rheostat connected in series with the armature
circuit. It presents the following disadvantages [31,32]:
ωr =
Va
K φf
(5)
As depicted in Fig. 2, the no-load speed is inversely proportional to the flux φf (or current If ), and the tangent of the
rate equation (Eq. 5) is inversely proportional to the square
of the flux φf .
It is also known that the torque directly proportional to the
flux and the armature current Ce = K φf Ia .
3.3 Control of armature voltage
This method consists of speed control by adjusting the armature voltage and it is very effective, stable and easy to
implement. As seen in Fig. 3, the only variable controlled
in here is the armature voltage of the DC motor based on the
speed Eq. 6. [24–28]:
ωr =
Va
Ra
−
i a = ωro − ωr
K φf
K φf
(6)
When the voltage Va is reduced, the no-load speed ω ro is
reduced too. In addition, for the same value of load torque
and flux φf , the armature voltage does not affect the drop in
speed ωr . The tangent of the speed–torque characteristics
is Ra 2 that independent of the armature voltage. Therefore,
(K φf )
it is possible to obtain different values of the voltage induces
a family of the non-deformed parallel characteristics. Speed
control of a DC motor can be achieved either by varying
123
Electr Eng
u2<u1
Cm
u1 < u max
u max
Cn
Fig. 5 Speed control block
r
Fig. 3 Torque–speed curve of a DC motor at different armature voltages
a plane relative to the poles of the inner loop. According to
the mechanical equation, [29,33,34] the following definition
can be written:
P
ωr
=
Ce
fc + J · s
(7)
Block diagram of a DC motor speed control is shown in
Fig. 5:
To evaluate the performance of the speed control of a DC
motor, we performed the numerical simulations with the following conditions.
• running the motor at no-load up to 100 rad/s
• applying a torque value with nominal load torque as a
value of 29.2 Nm at t = 5 s
Fig. 4 Closed loop speed control block diagram of a DC motor
Figure 6 shows that:
the voltage across of the armature circuit or by varying the
excitation flux.
4 Control of a DC motor by analogical PI
The speed control of a separately excited DC motor fed by a
chopper works like all other converters as depicted in Fig. 4.
The difference between the two configurations is existence
in an internal loop of the current. The reason of the difference is the particular characteristics of the power stage of the
chopper. The current and speed loops will be examined, and
its characteristics are explained. There are two control structures as the speed and the position of motors with a current
fed by chopper. Figures 4 and 5 illustrate the configurations
that have a single outer loop and an inner loop in which a
controlled current value is compared with the current value
of the armature, and the error is processed through a control
circuit [8,29].
4.1 Speed control
The best outer loop is to control the speed or the position
(size having the slowest dynamics). For this reason, the poles
laid down for the outer loop is closer to the origin of roots
123
• The speed follows the reference with a slight overshoot
and a response time of 0.95 s
• The current is limited to the permissible value and
increases by increasing the load
• Good rejection of perturbation
4.2 Position control
The position controller determines the reference speed and
the reference current, thus maintaining at the same speed.
Considering that, the dynamic speed is faster than its position,
and assuming that the speed reaches its reference value, then
the open loop transfer function is written as below [8,29]:
θ
1
=
ωr
s
(8)
The controller is chosen as a proportional action. The
block diagram of the position control loop is depicted in
Fig. 7.
The closed loop transfer function can be written as follows:
kθ
θ
=
∗
θ
1 + kθ s
(9)
Electr Eng
Fig. 6 Simulation results of adjusting current and speed of a DC motor
*
e
k
*
r
1
s
Fig. 8 Simulation results of adjustment in the current position of DC
motor
5 Backstepping control technique of a separately
excited DC motor
Fig. 7 Block diagram of position control
5.1 Speed control
To evaluate the performance of the control position of a DC
motor, we performed the numerical simulations with the following conditions as shown in Fig. 8.
In this section, we propose the backstepping control technique based on the command presented in the fourth section.
This method will also be applied for the motor control centre
and speed control with current limiting [29,31,35–37].
The main objective of the control law is that the rotor
speed ωr is pursuing a reference signal ωref .
Now we go back to the steps of backstepping:
• applying a 10◦ position level
• applying a torque to the nominal load 29.2 Nm at t = 5 s.
Figure 8 highlights that the adjustment by PI controller
gives satisfactory results as follows:
• the position of the rotor reaches the reference position
with small oscillations;
• the current is limited to the permissible value;
• good rejection of perturbation.
• Step 1:
The aim is to control the rotor speed and the amplitude of the
induced current, so that the following errors can be defined:
e1 = ωr∗ − ωr and ė1 = ω̇r∗ − ω̇r
(10)
123
Electr Eng
If it is the control action, using a single Lyapunov function
V1 = 21 e12 , and its temporal derivative is given as below:
V̇1 = e1 ė1
V̇1 = e1 ω̇r∗ − (L m i f i a − f c ωr − Cr )/J
(12)
fc
cr
∗
k1 e1 + ω̇r + + ωr
=
L m if
J
J
(13)
(11)
So that the reference current i a∗ is chosen as:
i a∗
J
• Step 2:
We define another error signal between the reference current
and the armature current i a :
e2 = i a∗ − i a , and its derivative : ė2 = i̇ a∗ − i̇ a
(14)
If this error and its derivative are replaced in to the Lyapunov
function, then we obtain Eq. 15.;
V2 =
1 2 1 2
e + e , its derivative is: V̇2 = e1 ė1 + e2 ė2
2 1 2 2
(15)
V̇2 = e1 [ω̇r − (L m i f i a − f c ωr − Cr )/J ]
+e2 i̇ a∗ − (va − Ra i a − L m i f ωr ) /L a
(16)
Fig. 9 Current and speed of a DC motor obtained from the simulation
with backstepping control
The armature voltage Va can be written as follows:
⎡
k1 e1 + k2 e2 + wr∗ + i a∗ +
Va∗ = L a ⎣ f
+ Jc + LLmaif ωr + cJr
Ra
La
−
L m if
J
ia
⎤
⎦
5.2 Position control
(17)
To illustrate the influence of backstepping controller on the
system performance, the same tests are repeated with the PI
controller.
According to the simulation results, following conclusions
can be given:
• improvement in total system performance with the insertion of backstepping controller compared to conventional
PI controller
• the speed reaches its set point value along with practically
zero overshoot as depicted in Fig. 9
• total rejection of perturbation
• the current is limited to its permissible value as illustrated
in Fig. 9
123
In this section, we have proposed the continuing problem of a
DC motor using Backstepping control method. The proposed
control method does not only stabilize the motor but also
causes the speed of tracking error to converge asymptotically
to zero [24]. The model of the DC motor can be reformulated
as follows [8] :
⎧
⎪
⎨ θ̇1 = ωr
ω̇r = − Jf ωr + 1J Ce − 1J Cr
⎪
⎩ Ċ = − kφ 2 ω − L C + kφ u
e
L
r
R
e
(18)
L
This model is used to design the backstepping algorithm
in order to achieve the objective of stability and pursuit the position. For this, we followed the following
steps:
Electr Eng
• Step 1:
At this point, if the desired reference torque is chosen as:
First, we consider the speed as a control variable. Let us
define the signal position error as:
e1 = θ − θ ∗ and its derivative
ė1 = θ̇ − θ̇ ∗
= ωr − θ̇ ∗
1 2
V̇ = e1 ė1
e and its derivative 1
= e1 ωr − θ̇ ∗
2 1
• Step 3:
Now the final signal error is obtained as:
(20)
The derivative of the Lyapunov function can be rewritten as:
V̇1 = −k1 e12 + e1 k1 e1 + ωr − θ̇ ∗
(21)
e3 = Ce − Ce∗ and its derivative : ė3 = Ċe − Ċe∗
+ J (1 + k1 + k2 ) θ̈ ∗ + J θ ∗
R
fc
−
+ Ce (1 + k1 + k2 ) −
J
L
kφ
fc
+
U − Ċr
+ C r − 1 + k1 + k2 +
J
L
(22)
(29)
If Va is considered as a control variable, the third Lyapunov
function V3 is chosen as:
• Step 2:
At the second step, it is tried to converge the speed signal to
its reference. Therefore, the signal error is redefined:
e2 =
(28)
ė3 = −J (e2 − k1 e1 ) (k1 + k1 k2 )
kφ
f2
+ ωr − f c (1 + k1 + k2 ) + c −
J
L
At that point, the desired reference speed can be selected as
follows:
ωr∗ = −k1 e1 + θ̇ ∗
(27)
(19)
If ωr is the control variable, using a single Lyapunov function:
V1 =
Ce∗ = J (1 + k1 + k2 ) e2 − k12 e1 − θ̈ ∗ f c ωr + Cr
V3 =
ωr − ωr∗
= ωr + k1 e1 − θ̇ ∗
(23)
1 2
e and its derivative : V̇3 = e3 ė3
2 3
(30)
So the Lyapunov function can be rewritten as:
And Eq. (23) can be expressed as:
ė1 = e2 − k1 e1 and its derivative is given follows :
V̇3 = e3 (ė3 + k3 e3 ) − k3 e3
ė2 = k1 ė1 − θ̈ ∗ + ω̇r
At this point, the desired reference voltage can be selected:
= k1 (e2 − k1 e1 ) +
1
1
f
ωr − Ce + Cr − θ̈ ∗
J
J
J
(24)
If Ce is considered as the virtual control variable, we can use
the Lyapunov function V2 like:
1 2 1 2
e + e and its derivative:
2 1 2 2
V̇2 = e1 ė1 + e2 ė2
fc
= e1 [e2 − k1 e1 ] + e2 −k12 e1 + k2 e2 − θ̈ ∗ + ωr
J
1
1
(25)
− Ce + Cr
J
J
V2 =
The Lyapunov function can be rewritten as:
V̇2 = −k1 e12 + e1 (k1 e1 + e2 − k1 e1 ) − k2 e22
fc
+e2 k2 e2 − k12 e1 + k1 e2 − θ̈ ∗ + ωr
J
1
1
− Ce + Cr
J
J
(31)
Va∗ = −k3 e3 + J (e2 − k1 e1 ) (k1 + k1 k2 )
kφ
f c2
−
− ωr − f c (1 + k1 + k2 ) +
J
L
− J (1 + k1 + k2 ) θ̈ ∗ − J θ ∗
R
fc
−
− Ce (1 + k1 + k2 ) −
J
L
fc
+ −Ċr
− C r − 1 + k1 + k2 +
J
(32)
To evaluate the performance of position control of a DC
motor using the Backstepping technique, we performed
numerical simulations under the following conditions as
illustrated in Fig. 10:
• applying a 10◦ position level.
• applying a torque rated load 29.2 Nm at t = 5 s.
(26)
Figure 10, shows that the control performance of backstepping gives satisfactory results as below:
123
Electr Eng
ẋ = f N (x) + f (ẋ) + gf (x)u f
(34)
where:
T
T
T
x = i a i f ω , ga = L1a 0 0 , gf = 0 L1f 0 ,
⎤
⎡
⎤
⎡
1
a
− R
L a (−RaN i a − E)
L a ia
⎥
⎢
⎥
⎢
f
f N (x) = ⎣ − RLfN i f
⎦ , f (x) = ⎣ − R
L f if ⎦
f
1
− 1J Cr
J (C e − f c ω − C rN )
The objective is to develop a nonlinear adaptive controller,
which can stabilize and track the desired value (reference)
ωref and interval ωr ≥ ωrN , compensating for parametric
uncertainties resistance of the armature and the inductor, and
reject the load torque. To obtain these, the system controls
the motor speed ωr and the E defined as:
Fig. 10 Simulation results of position and armature current of a backstepped DC motor
e E = E − E ref
(35)
eω = ω − ωref
(36)
where ωref is the reference speed and E ref represents the set
point weaken the flow, which is chosen between 0.85 to 0.95
of the rated armature voltage [13,37,39].
We define the following change of coordinates:
z 1 = h 1 (x) = E
• the position reaches the desired value
• the current is limited to the permissible value
• a total rejection of perturbation
z 2 = h 2 (x) = ω
z 3 = ℓfN h(x) = ∇h · f (x)
6 Adaptive backstepping control technique of a
separately excited DC motor
In this section, a proposed adaptive backstepping control
based on the cascade setting discussed in the first section
of this chapter and the work presented in [33,35] will be
applied for the speed control of the DC motor with an estimate of the armature resistance, inductor resistance and load
torque Cc observer. This method is inspired by the research
presented by Jianguo Zhou and Zuo Z. Liu [33,35]:
where the notations ℓf h(x) = ∇h · f (x) is ℓif h(x) =
ℓf (ℓi−1
f h) are used for the Lie derivation of the function
h(x) along a vector field f (x) and its iterative form, respectively.
This change of coordinates is one-to-one in s =
x ∈ R 3 : i f = 0 and ω = 0}.
Using the coordinate transformation above, the motor
model (22) becomes as below:
ż 1 = ℓ fN h 1 (x) + ℓ f h 1 (x)
+ ℓga h 1 (x)Va + ℓgf h 1 (x)Vf
ż 2 = z 3 + ℓ f h 2 (x)
ż 3 =
ℓ2fN h 1 (x) + ℓ f ℓfN h 2 (x)
+ℓga ℓfN h 2 (x)Va + ℓgf ℓfN h 2 (x)V f
Ra = Ra − RaN Rf = Rf − RfN Cr = Cr − CrN
(33)
where:
K
RfN
E + i f (Ce − f c ω − CrN ),
Lf
J
T
Rf
r
ωi f i f ,
ℓ f h 1 = K − L f − C
J
ℓfN h 1 = −
Unknown deviations from the parameters of its nominal values are RaN , RfN , and CrN . The values of Ra and Rf change
due to ohmic heating (increase in temperature during operation), while the value of Cr is practically unknown.
Whereas the two control variables Va and Vf , the motor
model can be reformulated as:
123
(37)
1
Cr
(Ce − f c ω − CrN ), ℓ f h 2 = −
,
J
J
K
RfN
ℓ2fN h 2 =
i f (−RaN i a − E) −
Ce
J La
J Lf
ℓfN h 2 =
(38)
Electr Eng
fc
(Ce − f c ω − CrN )
J2
1
Rf
BCr
a
−K R
ℓf ℓfN h 2 =
La + Lf
J
J
K
ℓga h 1 = 0, ℓgf h 1 =
ω, ℓga ℓf h 2 =
Ls
K
ℓgf ℓf h 2 =
ia .
J Lf
θ3 =
−
ia if 1
K
if ,
J La
T
1
a
−K R
La +
J
Rf
Lf
BCr
J
, ζ2 (x) = 1
After the backstepping design [19], we define the new variables
,
ē1 = e1
ē2 = e2
ē3 = e3 − α
V̂a = ℓga h 1 (x)Va + ℓgf h 1 (x)Vf
V̂f = ℓga ℓ fN h 2 (x)Va + ℓgf ℓ fN h 2 (x)Vf
(39)
(44)
where α is for stabilizing control e2 , and
α = −k2 ē2 − θ̂2 ζ2 (x)
Then, Eq. (38) becomes:
= −k2 ē2 − θ̂2
⎡
⎤
⎡
⎤ ⎡
ℓfN h 1 (x) + ℓf h 1 (x)
ż 1
1
⎣ ż 2 ⎦ = ⎣ z 3 + ℓf h 2 (x)
⎦ + ⎣0
ℓ2fN h 2 (x) + ℓf ℓfN h 2 (x)
ż 3
0
⎤
0
V̂a
0⎦
V̂f
1
(40)
Taking the derivative of (32), we have
ē˙1 = ℓfN h 1 + V̄a + θ1 ζ1 (x)
= ℓfN h 1 + V̄a + θ̂1 ζ1 (x) − θ̃1 ζ1 (x)
˙ē2 = ē3 + α + ℓ f h 2
Define a linear reference model, as:
ż m = Am z m + Bm u
(45)
(41)
= ē3 − k2 ē2 − θ̃2
˙ē3 = ℓ2f h 2 + V̄f + θ̂3 ζ3 (x) − θ̃3 ζ3 (3) − α̇
N
= ℓ2f N h 2 + V̄ f + θ̂3 ζ3 (x) − θ̃3 ζ3 (3) − k2 ē˙2 − θ̂˙2
with,
T
T
z m = z 1m z 2m z 3m , etu = E ref 0 ωref ,
⎡
⎡
⎤
⎤
−a1m 0
a1m 0
0
⎦ , Bm = ⎣ 0 0 ⎦
Am = ⎣ 0
0
1
0
−a2m −a3m
0 a2m
(46)
where :
∂α ˙
∂α ˙
θ̂2
ē2 +
∂ ē2
∂ θ̂2
= −k2 ē˙2 − θ̂˙
α̇ =
where a1m , a2m and a3m are constant design parameters that
can specify the responses of the dynamic system. The tracking error e is defined as:
(47)
and the parameter errors :
θ̃1 = θ̂1 − θ1
⎤
⎤ ⎡
z 1 − z 1m
e1
e = ⎣ e2 ⎦ = ⎣ z 2 − z 2m ⎦
z 3 − z 3m
e3
⎡
θ̃2 = θ̂2 − θ2
(42)
So, its dynamic error is:
ė = A(x) + A(x) + B V̄
where
V̄a
V̂a + a1m z 1m − a1m E ref
=
V̄f
V̂f + a2m z 2m + a3m z 3m − a2m ωref
⎡
⎤
ℓf h 1 (x)
⎦,
A(x) = ⎣ e3
2
ℓf h 2 (x)
⎤
⎡
⎡
⎤ ⎡
⎤
ℓf h 1 (x)
θ1 ζ1 (x)
10
A(x) = ⎣ ℓf h 2 (x) ⎦ = ⎣ θ2 ζ2 (x) ⎦ , B = ⎣ 0 0 ⎦
01
ℓf lfN h 2 (x)
θ3 ζ3 (x)
T
C
r
Cr
f
θ1 = K − R
, θ2 = −
, ζ1 (x) = ω · i f i f ,
Lf − J
J
V̄ =
(43)
θ̃3 = θ̂3 − θ3
(48)
Then, we define the following Lyapunov function:
1 2
1 2
1 2
1 2
2
2
ē + ē2 + ē3 + θ̃1 + θ̃2 + θ̃3
V =
2 1
γ1
γ2
γ3
(49)
Taking the derivative of this function with respect to time,
and replacing ē˙1 , ē˙2 , and ē˙3 with Eq. (38), we obtain:
V̇ = ē1 (ℓ f N h 1 + V̄a + θ̂1 ζ1 − θ̃1 ζ1 ) + ē2 (ē3 − k2 ē2 − θ̃2 )
+ ē (ℓ2 h + V̄ + θ̂ ζ − θ̃ ζ − k ē˙ − θ̂˙ )
3
fN 2
f
3 3
3 3
1
1
1
+ θ̃1 θ̃˙1 + θ̃2 θ̃˙2 + θ̃3 θ̃˙3
γ1
γ2
γ3
2 2
2
(50)
To do,
V̇ ≤ −k1 ē12 − k2 ē22 − k3 ē32 ≤ 0
(51)
123
Electr Eng
• application of a nominal load torque 29.2 Nm at time
t = 5 s and t = 10 s and t = 15 s.
According to the simulation results, we note following
conclusions:
• improved overall system performance with the inclusion
of adaptive backstepping controller compared to conventional PI and backstepping
• when starting, the speed reaches its set value with virtually no overshoot
• total rejection of disturbance
• the current is limited with its allowable value
7 Conclusions
In this paper, we have modelled a separately excited DC
motor. The mathematical model was developed for the DC
motor with separate excitation in the form of linear differential equations. These equations describing the dynamic
characteristics of the motor were solved using block diagram
and then simulated.
Thereafter, the DC motor was run by a PI speed control.
After that, simulation for speed adjustment by the backstepping and adaptive backstepping techniques was developed.
The simulation results under the MATLAB Simulink environment proved the efficiency and improvement of the
system performance of this PI controller.
Fig. 11 Simulation results of position by adaptive backstepping control
Appendix
We design the following control law:
Table 1 Motor driver system
parameters
Parameter
Per unit values
V̄a = −ℓfN h 1 − θ̂1 ζ1 − k1 ē1
Ra
0.6 Ω
V̄f = −ℓfN h 2 − ē2 − k3 ē3 − θ̂3 ζ3 − k2 (ē3 − k2 ē2 ) − θ̂˙2
Rf
240 Ω
La
0.012 H
Lf
120 H
Lm
1.8 H
J
1 Kg m 2
Fc
5 × 10−5 Nms
(52)
With adaptive laws:
θ̂˙1 = γ1 ē1 ζ1 θ̂˙2 = γ2 (ē2 + k2 ē3 )θ̂˙3 = γ3 ē3 ζ3
(53)
To illustrate the influence of the adaptive backstepping
based controller on system performance, we performed
numerical simulations under the following conditions as
depicted in Fig. 11.
• application of a speed step of 100 rad / s
• increase in reference speed
123
References
1. Zribi M , Al-Zamel A (2007) Field-weakening nonlinear control
of a separately excited DC motor, Hindawi Publishing Corporation
mathematical problems in engineering, Vol 1
2. Koreboina VB, Narasimharaju BL, Kumar DMV (2016) Performance evaluation of switched reluctance motor PWM control in
Electr Eng
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
PV-fed water pump system. Int J Renew Energy Res (IJRER)
6(3):941–950
Rinderknecht F, Hellstern H, Schier M (2011) Electric wheel hub
motor for aircraft application. Int J Renew Energy Res (IJRER)
1(4):298–305
Oguz Y, Harun O, Yabanova I, Oguz E, Kırkbaş A (2012) Efficiency
analysis of isolated wind-photovoltaic hybrid power system with
battery storage for laboratory general illumination for education
purposes. Int J Renew Energy Res (IJRER). 2(3):440–445
Talavaru P, Nagaraj N, Kishore Kumar Reddy V (2014) Microcontroller based closed loop speed and position control of DC motor.
Int J Eng Adv Technol (IJEAT) ISSN 3(5):2249–8958
Das HS, Dey A, Tan CW, Yatim A (2016) Feasibility analysis of standalone PV/wind/battery hybrid energy system for rural
Bangladesh. Int J Renew Energy Res (IJRER). 6(2):402–412
Alphonse I, Thilagar H, Singh FB (2012) Design of Solar Powered
BLDC Motor Driven Electric Vehicle. Int J Renew Energy Res
(IJRER) 2(3):456–462
Kadwane S (2006) Converter based DC motor speed control using
TMS320LF2407A DSK 1st IEEE conference on industrial electronics and applications, pp 1–5, Print ISBN: 0-7803-9513-1, 26
May 2006
Lekhchine S, Bahi T, Soufi Y (2013) Direct torque control for
double star induction motor. Int J Renew Energy Res (IJRER)
3(1):121–125
Scott J, Round WH (2009) Speed control with low armature loss
for very small sensor less brushed DC motors. IEEE Trans Ind
Electron 56(4):1223–1229
Mirosevic M, Maljković Z, Pavlinović D (2011) The dynamics of
diesel-generator unit in isolated electrical network. Int J Renew
Energy Res (IJRER) 1(3):126–133
Gruber P, Balemi S (2010) Overview of non-linear control methods,
28 May 2010
Hamida MA, Ezzat M, Glumineau A, De Leon et J, Boisliveau R
(2012) “ Commande par Backstepping avec action intégrale pour la
MSAP : Tests expérimentaux ” , Conférence Internationale Francophone d’Automatique (CIFA2012), Grenoble, France. 4–6 Juillet,
pp 973–978
Emna M, Kheder A (2016) An adaptive backstepping flux observer
for two nonlinear control strategies applied to WGS based on
PMSG. Int J Renew Energy Res (IJRER) 6(3):914–929
Ye XD (2012) Nonlinear adaptive control by switching linear controllers. Syst Control Lett 61(4):617621
Wang JH (2008) Nonlinear control of voltage PWM source rectifier,
1st edn. China Machine Press, Beijing, pp 10–12
Bodson M, Chiasson J (1998) Differential-geometric methods
for control of electric motors. Int J Robust Nonlinear Control
8(11):923–954
Elmaguiri O, Giri F (2007) Digital backstepping control of induction motors. In: IEEE international symposium on industrial,
electronics, pp 221–226
Wahsh SA, Shazly JH, Hassan AY (2017) Effect of rotor speed on
the thermal model of AFIR permanent magnet synchronous motor.
Int J Renew Energy Res (IJRER) 7(1):21–25
Asghar MSJ (2003) Speed control of wound rotor induction motors
by AC regulator based optimum voltage control. In: Power electronics and drive systems, the fifth international conference, pp
1037–1040
Zambada J (2007) Field-oriented control for motors, Motor
Design.com
22. Krim S, Gdaim S, Mtibaa A, Mimouni MF (2016) FPGA contribution in photovoltaic pumping systems: models of MPPT and
DTC-SVM algorithms. Int J Renew Energy Res (IJRER) 6(3):866–
879
23. Seibel BJ, Rowan TM, Kerkman RJ (1997) Field-oriented control
of an induction machine in the field-weakening region with DC-link
and load disturbance rejection. IEEE Trans Ind Appl 33(6):1578–
1584
24. Tan Y, Chang J, Tan H (2003) Adaptive backstepping control and
friction compensation for AC servo with inertia and load uncertainties. IEEE Trans Ind Electron 50(5):944–952
25. Hesari S (2016) Design and implementation of maximum solar
power tracking system using photovoltaic panels. Int J Renew
Energy Res (IJRER) 6(4):1221–1226
26. Bakshi V, Bakshi U (2009) Electrical machines. Technical Publications, Pune, India
27. Nouira I, Khedher A, Bouallegue A (2012) A contribution to the
design and the installation of an universal platform of a wind
emulator using a DC motor. Int J Renew Energy Res (IJRER)
2(4):797–804
28. Krause PC, O’Connell TS, Wasynczuk O, Hasan M (2017) Introduction to electric power and drive systems. In: IEEE press series
on power engineering, ISBN: 1119214254/ 9781119214250, Wiley
29. Hughes A (2009) Electrical motors and drives: fundamentals, types
and applications, 3rd edn. Newnes Publications, Elsevier, Oxford,
UK
30. Lekhchine S, Bahi T, Soufi Y (2012) Comparative performances
study of a variable speed electrical drive. Int J Renew Energy Res
(IJRER) 2(4):591–595
31. Chaouch S, Nait-Said MS (2006) Backstepping control design for
position and speed tracking of DC motors. Asian J Inf Technol
5(12):1367–1372
32. Farshbaf RA, Azizian MR, Shazdeh S, Mokhberdoran A (2016)
Modelling of a new configuration for DFIGs using T-type converters and a predictive control strategy in wind energy conversion
systems. Int J Renew Energy Res (IJRER) 6(3):975–986
33. Zhou J , Wang Y, Zhou R (2000) Adaptive backstepping control of
separately excited DC motor with uncertainties PowerCon 2000.
In: 2000 International conference on power system technology.
Proceedings (Cat. No.00EX409), Perth, vol 1, pp. 91–96
34. Tsotoulidis SN, Safacas AN (2011) Analysis of a drive system in a
fuel cell and battery powered electric vehicle. Int J Renew Energy
Res (IJRER) 1(3):140–151
35. Liu ZZ, Luo FL, Rashid MH (2003) Adaptive MIMO backstepping
controller for high-performance DC motor field weakening. Electr
Power Compon Syst 31:913–924
36. Mishra SK, Purwar S, Kishor N (2016) An optimal and non-linear
speed control of OWC wave energy plant with wells turbine and
DFIG. Int J Renew Energy Res (IJRER) 6(3):995–1006
37. Nehrir MH, Fatehi F (1996) Tracking control of DC motors via
input-output linearization. Electr Mach Power Syst 24(3):237–247
38. Khedher A, Khemiri N, Mimouni MF (2012) Wind energy conversion system using DFIG controlled by backstepping and sliding
mode strategies. Int J Renew Energy Res (IJRER) 2(3):421–430
39. Elgammal A, Sharaf AM (2012) Dynamic self adjusting FACTSswitched filter compensation schemes for wind-smart grid interface
systems. Int J Renew Energy Res (IJRER) 2(1):103–111
123