See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/4254607

Linearizing Control of an Induction Motor

Conference Paper · January 2007

DOI: 10.1109/ICIT.2006.372323 · Source: IEEE Xplore

CITATIONS READS
29 989

2 authors, including:

Kanungo Mohanty
National Institute of Technology Rourkela
179 PUBLICATIONS 2,430 CITATIONS

SEE PROFILE

All content following this page was uploaded by Kanungo Mohanty on 20 June 2016.

The user has requested enhancement of the downloaded file.

 Linearizing Control of an Induction Motor
K. B. Mohanty and N. K. De
Dept. of Electrical Engineering
* National Institute of Technology, Rourkela
E-mail: kbmohantygnitrkl.ac.in
* Indian Institute of Technology, Kharagpur
ABSTRACT: An attempt has been made in the present using the amplitude and frequency of the supply voltage as
work to improve the dynamic and steady state inputs. A voltage command input-output linearization
performances of the vector controlled induction motor controller is developed in [7], and a current command input-
drive. Input-output linearization technique is used to output linearization controller is reported in [2]. Feedback
linearize the model of drive system and decouple the flux linearizing control technique is used in [8] to design a
and torque loop. State feedback controllers are designed controller for switched reluctance motor. A new control
for the linearized system using pole placement technique to scheme is presented in this paper based on [2-1 1] for
obtain desired dynamic and steady state responses. Test linearization and decoupling control of induction motor, and
results show that the performance of the drive system with also to improve the dynamic torque and speed responses for
the proposed control technique and designed controller is high performance motion control applications.
an improvement over the vector controlled drive and the In section II, the induction motor model in synchronously
proposed drive system is more flexible. rotating reference frame is reviewed. In section III, the
induction motor model is linearized by input-output
I. INTRODUCTION
linearization technique. Linear state feedback controllers are
synthesized to obtain good dynamic and steady state response.
Induction motors are widely used in industry, for their Test results are discussed in section IV.
simple and robust structure, higher torque-to-weight ratio,
higher reliability and ability to operate in hazardous II. STATE VARIABLE MODEL
environment. However, unlike DC motor, their dynamic
response is sluggish and control is a challenging task, because From the voltage equations of the induction motor in the
of the inherent coupling between rotor current and airgap flux, synchronously rotating d-q axes reference frame, the state
responsible for the torque production. The control of IM in space model with stator current and rotor flux components as
field coordinates using vector control (also known as field state variables is:
oriented control) [1], leads to decoupling between the flux and d F is - A1l A12] is 1 FB1,
torque, thus, resulting in improved dynamic torque and speed L =
dt Ir LA21
S +
A22 i r i 0 i
Vs (1)
responses. Significant advances have been made in vector r
control of induction motors since its inception. A disadvantage The stator current which is measurable is taken as the output,
of the conventional field-oriented controller is, the method which is expressed as
assumes that [1], the magnitude of the rotor flux is regulated to i
a constant value. Though good dynamic current (or torque) is =[1 o] (2)
and speed responses are obtained with vector control, the
torque is only asymptotically decoupled from the flux, i.e.,
decoupling is obtained only in steady state, when the flux
where,
rT
is = y =[ids iqs], I r = [dr vqr] , and

amplitude is constant. Coupling is still present, when flux is VS = LVds Vqs I

weakened in order to operate the motor at higher speed within
the input voltage saturation limits, or when flux is adjusted in
All = -a1 I -Oe J A12 = a2 I-P a3 Or J A21 = a5 I
order to maximize power efficiency [2]. This has led to further A22 = -a4 I - ( w) - PWr ) J, B1 = c I.
research on application of differential geometry [2-7], to Ft 0 F 0 -1
develop the control techniques for linearization and 1=011 and J 1I 0°
decoupling control. After the theory was proposed [3, 4], it has L
drawn attention of many researchers for further development c = Lr/( Ls Lr - Lm )'
and implementation. These techniques have resulted in a = cR + cRR L2 L2 cR L / 2
solutions to several problems, including feedback Lm /L rar9
linearization, input-output linearization and decoupling a3 = c Lm / Lr, a4 = Rr / Lr, a5 = Rr Lm / Lr
control. Reference [5] achieved decoupling of torque and flux Rs, Rr, Ls ^ Lr , Lm Motor parameters (given in appendix),
by a static multivariable state-feedback controller. Decoupling P: Number of pole pairs,
is also obtained in [6] by a static state-feedback controller c0r : Mechanical rotor angular velocity,

1-4244-0726-5/06/$20.OO '2006 IEEE 45

 Coe Synchronous electrical angular velocity, induction motor system is now transformed into two linear and
Vds Vqs d-q axes stator phase voltages, decoupled subsystems: electrical and mechanical.
ids, iqs d-q axes stator phase cuvoents, Electrical subsystem is represented by the state equation:
vIdr , vIqr: d-qaxesrotor fluxes. Ids -alids +a2Vdr +ul (10)
The torque developed by the motor is: *dr = -a41dr + a5ids (11)
Te = Kt ( VIdr iqs - vIqr ids ) (3) Mechanical subsystem is represented by the state equation:
where, Kt = 3PLm /2Lr * Te = - (a, +a4 ) Te+u2 (12)
Cr = (Te - TL - P (Or )/ J (13)
111. CONTROLLER DESIGN The stator input voltage components vdS and vqs in terms of
ul and u2 are:
A. Linearizing Control vdS =( Ce iqs + ul) / c (14)
The conditions required for vector control [1] are:
Iqr = 0 and Tqr = ° Vqs =-P (o)r( ids + a3 ]dr)+ K2 (15)
From (1), t]'r
Vqr = a5 lqs - (ce - Pvdr a4 ()qr
or The transformed model given above is valid only for vdr . 0.
Since the induction motor system described by (10)-(13) is
Indirect vector control is obtained, when linear and decoupled, the developed torque (or the speed) and
a5 iqs = ( Ce - P (or ) Vdr the rotor flux are independently controlled. State Feedback
co e = Pco r + a5 i qs/Vdrr
or, ^ (4) Controllers (SFC) are designed to obtain desired transient and
steady state performance.
When (4) is satisfied, the dynamic behaviour of the induction
motor is: B. Design of State Feedback Controller (SFC)
ids = - al ids + a21dr + Whe 1qs + C vdS (5) A general representation of the electrical subsystem (10),
( 1), and the mechanical subsystem (12), (13) is given by
lqs = -
e ids - al iqs -P a3 (OrIdr +c vqs (6) x=Ax+Bu+Ed (16)
Idr = -a4 Vdr + a5 ids (7) y=Cx (17)
Te = Kt Vdr iqs (8) where, for the electrical subsystem: u ul ,y = ids, d = 0,
iIds
The concept behind field oriented control is that rotor flux
can be controlled according to (7), with ids acting as the
L Idr
A =-a
F
L a5 -a4
=j,B
a2
0
]c =It o]
control input. The q-axis current component iqs serves as an
input in order to control the torque (8) as a product of Nldr and
For the mechanical subsystem: U U2, Y r, d TL,
iqs. Even the field oriented induction motor model described -(al +a4) 0 l 0
by (5-8) has nonlinearity and interaction. The speed emf term A= 1 B= C
((Or I dr ) appearing in (6) makes the current dynamics L JJ
nonlinear and speed dependent. Equations (5) and (6) show Since the state feedback control is basically a proportional
that interaction between current components exists, in the control, the steady-state error may exist. To remove this error,
rotating reference frame. The transition from field oriented the deviation term between the output y and the reference
voltage components, vds and vqs to current components as in value yr is introduced as a new state.
(5) and (6) involves leakage time constants and interactions. Y = Y -Yr (18)
During the flux transient period (7), coupling of flux and Assuming y. and d to be constant, differentiating (16) and (18)
torque is apparent from (5) to (8). The interaction between and arranging, a standard form of state equation is obtained as:
current components and nonlinearity in the overall system is A I+Bv
I

eliminated by using the linearization control approach given in

[2-7]. This approach consists of change of coordinates and use where a new augmented state vector, z and a new control
of nonlinear inputs to linearize the system equations. vector, v are defined as
The developed torque, T, is considered as a state variable, Fx] d FA 0] BFB1
replacing iqs in the induction motor model. Differentiating (8) z Li vu and A LC oj BL
and simplifying with substitution of(4), (6), (7): A linear state feedback control law ofthe form:
Te (a1 + a4) Te + Kttdr [ C Vqs - P (or (ids +a3V dr) ](9) v Kz (20)
The nonlinearities in (5) and (9) are put together and then can be designed for the augmented system of (19) by the pole
replaced by nonlinear functions of the form u1 and u2 placement technique [9], where K is the feedback gain matrix.
respectively. With these linearizing inputs u1 and u2, (5) and For the closed loop system (19), with the state feedback
(9) are then modified to (10) and (12) respectively. The

46
control (20) to have the desired eigenvalues, controllability stop from -800 rpm in 0.5 sec. Dynamic response to the
conditions should be satisfied as follows [10]: benchmark test is shown in Fig. 4. After starting of motor, the
1. the pair (A, B) is controllable rated rotor flux linkage of 0.45 V.s is established in 0.5 sec
FA and remains at the rated value after that, irrespective of speed
2. the matrix Bl has full rank. and load torque changes. The quadrature component of the
LC 0] rotor flux remains zero, through-out, indicating decoupling of
Since these conditions are satisfied for the given system, the flux and torque. The rotor speed tracks the reference speed
control law can be derived by substituting z into (20), and then with a little delay, and reaches 800 rpm at t=0.7 sec without
integrating. The linear state feedback control law for both any overshoot. When the rated load torque is applied, there is
electrical and mechanical subsystems is a temporary dip in speed of 145 rpm. After the load is
t released, again there is a temporary speed overshoot of 140
u=Kp x+Kif(y-y,)dt (21) rpm. When the reference speed is reduced to zero linearly,
0 rotor speed follows with a delay, and becomes zero at t=2.14
The above control law comprises of the feedback of the sates sec. For reverse motoring (second half cycle), the dynamic
(first term) as well as the integral of the output errors (IOE) response is similar to that of forward motoring.
(second term) and does not require the knowledge of the In the second case, the reference speed is changed in steps.
disturbance vector. The IOE feedback makes the controller First the speed is increased from 1000 rpm to 1500 rpm at
fairly robust by making it insensitive to modeling rated rotor flux linkage of 0.45 V.s and then further increased
imperfections and step like disturbances. to 1800 rpm with flux weakening, and then decreased to 1500
For the electrical subsystem, the control law is rpm with flux strengthened to rated value. Load torque is kept
ul=
u=Kpl i ds +pK2 Pdr +Kil 0 (ids -idj)dt
t
*) dt

(22)
(22)
constant at 1 N.m. The dynamic response is shown in Fig. 5.
During flux weakening control above the base speed (1500
rpm to 1800 rpm), the rotor flux linkage reduces from 0.45 V.s
The control law for the mechanical subsystem is to 0.375 V.s in 0.5 sec. The flux linkage also increases from
= Kp3 Te + Kp4 0)r + Ki2 J (()r - ) dt (23) 0.375 V.s to 0.45 V.s in 0.5 sec, when speed is reduced from
0 1800 rpm to 1500 rpm. During constant torque mode of
In the above two control laws, the proportional gains Kpl ,Kp2, operation, the rotor flux linkage is constant at 0.45 V.s. The
Kp3, Kp4 and the integral gains K,1 , K12 are determined using quadrature component of the flux is zero, through-out,
the pole placement technique. The eigenvalues of the indicating decoupling of flux and torque. Speed response does
augmented system matrix of the electrical subsystem are not have any overshoot.
-288.55, -10.22 and 0.0. To place the closed loop poles of the The fact that, speed response does not have any overshoot,
electrical subsystem at -288.55, -20 and -20, the gains of the is a great advantage of this controller. Speed response also
state feedback controller are, Kpl = -29.78 , Kp2=-2130.2 and tracks the command value very fast. The responses are better
K., = -28,922. The eigenvalues of the augmented matrix of the than vector controlled IM drive and this is achieved when the
mechanical subsystem are -298.77 -0.34 and 0.0. To place motor is fed from a Voltage Source Inverter (VSI). Fast torque
the closed loop poles ofthe mechanical subsystem ato 298.77, response is another advantage of this system. However, when
-10 and -8, the gains of the state feedback controller are, the model of induction motor iS not known exactly, or the
Kp3 = -17.66, Kp4 = -47.1 and K2 = -210.33. The block parameters used in the model changes due to change In
diagram of the electrical subsystem with SFC is shown in Fig. operating conditions, the linearization does not hold good. The
1. The block diagram of the mechanical subsystem with SFC decoupling of torque and flux is not obtained, and so also the
is shown in Fig. 2. The test results of the closed loop system linearization of the motor model. Because the nonlinearity
with these controllers are presented in the next section. cancellation is not perfect, under such conditions torque and
flux still remains coupled during the transient period, and
hence fast transient response can not be obtained. The drive
IV. TEST RESULTS AND DISCUSSIONS system with the proposed control is now being implemented in
The proposed controller has been simulated using the laboratory.
MATLAB on an induction motor drive system, whose data are TABLE - I RATING AND PARAMETERS OF THE MOTOR
listed in Table- I. Two cases of simulation tests are presented
here. First the drive is subjected to a benchmark test, as Threephase, 50lz, 0.75 kW, 220V,3A, 1440rpm
shown in Fig. 3. The unloaded motor is required to accelerate R0= 6.37
Stator and rotor resistances: QV, R = 4.3
from standstill condition to 800 rpm in 0.5 sec. Then reference Stator and rotor self inductances:L= Ls = 0.26 H
speed is kept constant at 800 rpm from time, t=0.5 sec to 1.5 Mtal inductancebs an r Lm 0.2 H4
Motu of Tiinrf;a ofemo
Mome%nt-4 ofkiertia
tor and :
in2
sec. Rated load torque of 5 N.m is applied from t=0.8 sec to o-f moto-r%r anA load:
lo-ad: T = 08
0.08KgRK m"r
1.2 sec. Then reference speed is changed from 800 rpm at
t=1.5 sec to -800 rpm at t=2.5 sec in no load condition, kept |Viscous friction coefficient: D 0.003 N m s/radl
constant at -800 rpm until t=3.5 sec, load torque of-S N.m is
applied from t=2.8 sec to 3.2 sec. After that motor is made to

47
 V. CONCLUSION [4] T. J. Tarn, A. K. Bejczy, A. Isidori and Y. Chen,
A linearization and decoupling control technique for
"Nonlinear feedback in robot arm control," Proceedings of
23 Conference on Decision and Control, December 1984,
induction motor drive is developed. State feedback controllers pp.736-751.
are designed using pole placement technique to control the [5] Z. Krzeminski, "Nonlinear control of induction motor,"
dynamic and steady state responses of the drive. The test IFAC 10t World Congress on Automatic Control, vol. 3,
results show that the dynamics are controlled in about 0.5 Munich, 1987, pp. 349-354.
second, fast for such systems, and also the voltage and current [6] A. De Luca, and G. Ulivi, "Design of exact nonlinear
variations are within limits. There is no overshoot in speed at controller for induction motors," IEEE Trans. on
all. This is an improvement of conventional vector controlled Automatic Control, 34 (12), Dec. 1989, pp 1304-1307.
IM drives. The implementation is underway in the laboratory. [7] R. Marino, S. Peresada, and P. Valigi, "Adaptive input-
ACKNOWLEDGMENT output linearizing control of induction motors," IEEE
Trans. on Automatic Control, vol. 38, no. 2, 1993, pp.
Dr. K.B. Mohanty is highly grateful to the Ministry of Human 208-221.
Resource Development, Government of India for providing [8] M. Illic'-Spong, R. Marino, S. M. Peresada and D. G.
necessary financial assistance, vide Sanction Order No. F.26- Taylor, "Feedback linearizing control of switched
14/2003. TS.V. dated 14.1.2004, to carry out the research reluctance motors," IEEE Trans. on Automatic Control,
work presented in this paper. vol. 32, no. 5, May 1987, pp. 371- 379.
[9] W. H Wonham, "On pole assignment in multi input
REFERNCES controllable linear systems," IEEE Trans. on Automatic
Control, vol. 12, December 1967, pp. 660-665.
[1] W. Leonhard, Control of Electrical Drives, Springer-
Verlag: Berlin, 190
Verag Beli, 1990.
[10]H.lW Smit and
[IO] regulators",
H. W. Smith B.cJmbDavison,
and E. J. "Desig of'id
119, no."Design
Proc. IEE, vol.Davison,
of industrial
8, August 1972, pp.
H
[2] D. I. Kim, I. J. Ha and M. S. Ko, "Control of induction 1210-1216.
motors via feedback linearization with input-output 1K. B. Mohanty7
[-ll- K. .
and N. K. De, ''Nonlinear controller for
decoupling," Int. Journal of Control, 51 (4), 1990, pp. induction motor drive," Proc. of IEEE International
863-883. Conference on Industrial Technology (ICIT), 2000, Goa,
[3] A. Isidori, A. J. Krener, C. Gori-Giorgi, and S. Monaco, India, pp. 382-387.
"Nonlinear decoupling via feedback: A differential-
geometric approach," IEEE Trans. on Automatic Control,
vol. 26, 1981, pp. 331-345.

KIl E4ids

Ki1
| ± Ut 1 a5 51 -dr

ids S
-
±s+aj I 'I s+a4±
+a

Kp2 ~~~~a2

Fig. 1. Block diagram ofthe electrical subsystem with SFC

3
TL
(Or
_T_I_

± J+j+4P\~
5~~ s+f

Fig. 2. Block diagram ofthe mechanical subsystem with SFC

48
 800 150

600-
00

) 40
0

0F 40

0) 200| C

-600~ -4000-
~ ~ ~ ~ ~ ~ ~ ~~~~~~~~~~~~~~~~~4

-200 -III

8400-

1700
-600-

10 5
800 1500
0 0. 1 . . .

00~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~10
0)0 0 0.5 1 1. 2 25 3 .5 40 0.5 1.5 2 2.5 3 3.5 4
~~~~~~~~~~~ 00 ~ ~ ~ ~ ~ ~ Time
~ ~ (see)
~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~~~~10 Time (see)
-600-0
04
0.5 1 1.5 2 2.5 3 3.5 40
Fig.~~~~~~~~~~~~~~~~~~~~~~Fg
0) 4 (a)
-800~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~10

400 -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~9

00 150Tie
se) im (ee
0 4(a-~~ig 1400

Time (sec) 10
111dr'~~~~~~~~~~~~~~~c ~~~~~~~~~~~~~~Time
(sec)
Fig.~~~~~~~~~~~~~~~~~~~~~~Fg 4 (b)
Fig. BenhmarktestFig. Bechmak tes reslts () Sped rspons, (b Tirqe an flu

0)~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~10

0 0.5 1.5 2 2.5 3 3.5 4

Time (see) 2 25 3 35 4

Time (see)
Fig. 4 (b)
Fig. 5 (b)

49
 200.

150

100

50

'A -50

-100
ct~ ~ ~~~~~~Fg 5 (c
-150

-200
0 0.5 1~ 2 2.5 3 3.5 4
Time (see)

Fig. 5 (c)
Fig. 5. Step change in speed command with flux weakening (a) Speed
response, (b) Torque and flux linkage, (c) a- phase stator voltage

View publication stats