IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS, VOL. IE-31, NO.

2, MAY 1984 141

Digital Simulation of Field-Oriented Control

of Induction Motor
S. SATHIKUMAR, STUDENT MEMBER, IEEE, AND JOSEPH VITHAYATHIL, MEMBER, IEEE

Abstract-This paper describes the simulation of a control scheme using

the principle of field orientation for the control of a voltage source in-
verter-fed induction motor. The control principle is explained, followed by
an algorithm to simulate various components of the system in the digital
computer. The dynamic response of the system for the load disturbance
and set-point variations have been studied. Also, the results of the simula-
tion showing the behavior of field coordinates for such disturbances are
given.
Fig. 1. Components of stator current vector and the rotor flux linkage in the
arbitrarily selected rotating reference frame.
I. INTRODUCTION
THE DYNAMIC PROPERTIES of an induction motor as and
a control plant can be described by a set of nonlinear
differential equations. These equations relate the electrical Te =PL
Lr
(Omrl * s2-mr2 *isi) (2)
quantities currents and voltages of the motor with the me-
chanical quantities torque and speed. The control of such a where Tn = (Ln/Rn) and w5 is the slip angular frequency.
plant becomes complicated because of intricate coupling be- Here, the components Om ri and Om r2 of the flux vector Om r
tween all the control inputs. This problem can be overcome by and the components of is, and i42 of the stator current vec-
the control of field-oriented quantities which reduces the con- tor is are all unidirectional quantities (Fig. 1).
trol of an ac induction motor to that of a separately excited dc These equations show that both stator current compo-
motor [1]. nents influence both variables, namely the rotor flux and the
This method enables the control of field and torque of the torque. A decoupling can be accomplished by properly aligning
machine independently by manipulating the corresponding the coordinate system with the rotor flux vector. Fig. 2 shows
field-oriented quantities. This system adapts to any load such a system in which the components isd of is parallel to
disturbances and/or set-point variation as fast as a dc motor Omr and the component iSq of is quadrature to Omr are
can operate. This is essentially because of the fast response shown. These field-oriented coordinates isd and isq of the
of the field-oriented quantity responsible for torque. stator current are important for the torque and field. The
Various control schemes have been proposed for the coordinate isd parallel to the field causes the build up of the
voltage source inverter-fed drive system [31, [4] in recent field, while the coordinate isq perpendicular to the field
years. In the scheme proposed here for the simulation, a together with the field Om r produces the torque.
flux position controller has been incorporated which deter- By defining a space phasor of the current as
mines the desired position of the flux to establish the re-
quired field components. i5(t) = is,(t) + i52(t)ef2'f13 +i±3(t)e-f27f3 (3)
II. PRINCIPLE OF FIELD-ORIENTED CONTROL it can be shown that for sinusoidal balanced three-phase
The vector diagram of Fig. 1 shows the vector of the stator quantities is, I5 2, and 43
current is and the vector of the rotor magnetic flux linkage
kmr. These quantities as seen by the stationary reference i() = 3/2
is(t iP~
isep(t) (4)
axis (cw, ) are alternating quantities. By selecting a rotating 2

reference frame (1, 2), it can be shown that [2] the electrical and
torque (Te) developed and the flux build up are expressed
by the following equations: d
w(t) = d p(t) (5)
d mrl -Ws -llTr- ¢mr Lm Fisl
dt smr2 -lTr ws (1) is the electrical angular frequency of the supply currents.
iisj~O ~r2 Tr I~i
Lkrr2 ±
The stator terminal voltage is defined accordingly as

Manuscript received August 16, 1982; revised November 21, 1983. Vs(t0 Vs 1(t) + VS2(t)e12 7r3 + Vs3 (t)e-'27r3 (6)
The authors are with Department of Electrical Engineering, Indian Institute
of Science, Bangalore-56 012, India. The machine equations for a two-pole induction motor with
0278-0046/84/0500-0141$01.00 © 1984 IEEE
 142 IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS, VOL. IE-31, NO. 2, MAY 1984

Fig. 2. Components (field coordinates) of stator current in the rotating r Fig. 3. Angular relationship of current and flux linkage vector.
ence frame aligned with rotor flux linkage vector.

a short-circuited rotor electrical circuit, in a stator refere

frame using space phasor rotation are [3]
d d
- (ire)) i-y
Rsis +Ls dt- (is) +Lm dt (7) Fig. 4. Induction motor model in rotating reference frame.

d d By transforming the machine model expressed by (7)-(8)

Rrir+±Lr-dt (ir)r +Lm dt
- (ise 1 )= O (8) to the rotating reference frame with definitions in (12)-
(15), we get
dw d
J-=
dt
(Te - T1) =2Lm tis(iree*]-T
3E* T (9) as Ts -isd +isd
dt
where
- Vsd ±WTsq_()Ts dmr (16)
Ls = (1 + Us)Lm Rs ~~~~~dt

isq ± 's - Vsq

d
s5Ts s iusd5 - (1 - Y)WTs inmr
with
1 (17)
Ur= 1- (10) d
(1 + Us)(1 + ar) -imr + imr = isd
Tr
dt (18)
and
dp isq
de -(Wd + Wr (19)
Wr =
dt
(I l) dt Tr im r
dw 2 Lm
the rotor angular frequency. J
dt'
-Te-Tklkimrisq-Ti (20)
3 (1 + ar)
This model can be transformed into the rotating reference
frame td get the relation between field-oriented quantities. where Ts = Ls/Rs, and J is the constant of inertia.
As discussed earlier, a decoupling effect is achieved if the
coordinates are properly aligned with the rotor flux vector. III. CONTROL STRATEGY

For this purpose, a magnetizing current vector imr(t) repre- The machine model in the rotating reference frame reduces
senting the rotor flux reference can be defined as essentially to two flow paths, one for the flux governed by
(18) with 4sd as the control input, and the other for the
imr(t) = i5(t) + (1 + Or)ir(t)e -JE = imr ef' (12) torque governed by (20) with control input isq together
with q,,,, the rotor flux. This model is shown in Fig. 4.
and The flux path in this model is similar to the field circuit, and
d4r the torque path is similar to the armature circuit of a separately
W
= -. (13) excited dc motor. As it is normally done in a dc motor, the flux
dt
level in the machine can be maintained at its maximum level,
In this frame of reference, the component ids of i5(t) parallel depending on the ac voltage available and the operating flux
to the rotor flux and the component isq perpendicular to it density of the motor, up to the base speed, above which it
are related to is as follows (Fig. 3): may be weakened.
The speed control can be imposed on isq and the flux
)eise -iS(t).
(isd +±/iq (14) control on isd- When these manipulated quantities are given,
the field-oriented quantities are established only after a tran-
The voltage vector Vs(t) is related to its components as sient process, which is caused by a negative feed back of the
field angle T. For this purpose, a flux position controller
(Vsd +iVSq)el, V(t). was incorporated, which gives the desired position of the
flux vector.
 SATHIKUMAR AND VITHAYATHIL: FIELD-ORIENTED CONTROL OF INDUCTION MOTOR 143'

Fig. 5. Block schematic of the control scheme.

IV. CONTROL SCHEME V. DIGITAL SIMULATION OF THE SYSTEM

The block schematic of the control scheme used for simula- INDUCTION MOTOR
tion is shown in Fig. 5. The flux controller, which operates To study the performance of the induction motor, the two
on the difference between the rotor flux reference and the axes (a, 3) stationary reference frame model is chosen. In this
magnitude of the actual rotor flux vector, sets the reference frame of reference, a two-pole induction motor, with a short-
for iSd(isd ref)- Similarly, the speed error wer serves as the circuited rotor circuit, is characterized by the following
input to the speed controller, which sets the reference for equations:
the torque. This reference value when divided by the magni-
tude of the actual flux vector gives the reference for isq(isq ref). d
In a current-controlled inverter drive, these current references dt- sci= Vs, -Rsisc
are converted to three phase quantities to control the current
in each phase of the motor. To achieve a fast response drive d
system, the corresponding voltage has to be impressed on the
-
dt
Orc = Rrlrce Wrfrp
motor. The voltage reference values Vsdm and Vsqm can be
obtained from (16) and (17), which define the relation between d
the stator current and the terminal voltage in the field coordi- dt rg = -Rrir, + Wrtra-
nates. But these equations which have the motor parameters
will introduce an error because of the variation in the motor d
parameter due to the temperature effect and the accuracy dt sp
s -V5-Rsis5
with which they are measured. Hence, a closed-loop control
is required for iSd and isq which is achieved by the id con- d
-
Wr = (Te-TI)/J (21)
troller and iq controller as shown in Fig. 5. dt
Once the manipulated variables Vsdm and Vsqm are given
they are to be properly positioned in the stationary reference where
frame. The desired position of the voltage vector is given by
a flux position controller. Equation (19) defines the desired
slip, and the actual slip is calculated as the difference between Fs(i Lr4sce-LmOra]
Ls5'rj- Lmsoz
w (the angular velocity of the rotor flux) and wr (that of the (22)
rotor). The slip error, when properly enumerated, serves as irj Lsrkp -Lm ks5g
the input to the flux position controller. The output of the
controller gives the incremental value At of I, which, when isJ
s-LmOr,J
[Lrs
added to the actual flux position I, gives the desired position
To of the flux vector. and
Knowing the desired position of the flux vector, the termi-
nal voltages Vs and Vs, of the stationary reference frame can Te = 1 - 5Lm (ispir -isa irp) (23)
be calculated, which are then converted to three phase quanti-
ties serving the reference voltage for the inverter. For the purpose of convenience, the per unit model of the
 144 IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS, VOL. IE-31, NO. 2, MAY 1984

motor is used for simulation. Converting the above equa- by the following difference equation (see Appendix II):
tions to represent the per unit model, we get, with WB as the
base angular frequency Yn =Y(n-1) + G C-G(n - l) (30)
C
d s = (Vsa -Rs'isce)wB where
di
T
C=-
-d ra =(-Rr'iroz' -Wr r3)WB T+DT
cit
T = time constant of the controller,

cit
'ro
O = (Rr irg' Wr qra J)WB G = gain of the controller,
DT = sampling time.
d
dt Ose -Vsp-p-RKsispg)WB All the feed-back quantities are compared with the correspond-
ing reference quantities, and the error quantity E forms the
input to the controller.
dt
VII. CURRENT TO VOLTAGE TRANSFORMATION IN
where the form of the defined variable equation (22) remains FIELD COORDINATES
unchanged, and The manipulated variables isdm, isqm are converted to the
manipulated variables Vsdm, Vsqm using the relations (16)
Te' = Lm '(isp'ira -1sor
i'r )WB (25) and (17). Rearranging the equations and converting them
into per unit quantities, we get
(The dashed quantities represent the per unit quantities.)
The required feed-back variables are the currents iso ', V+ d
is', 4ra', Orf,' and Wr'. In a practical condition, the currents Vsdm=-Rs'isdm +TB aTS'Rs' isdm
are directly accessible, whereas the rotor flux quantities
are not. The rotor flux can be estimated from the air gap
flux [4] as a function of the coupling coefficient of the + (1-U )R'TS d im, -WuGT'RS'ivqm (31)
rotor and the stator current. The magnitude of the flux
cit __ssq
vector Om5r and angle T are found from q5roa' and 5rp' as
follows: Vsqm'=RS 'isqm + TB [TsRtsqm-i
+ w 's dS+ (1-)cit q m

Om r= (/(roa + ¢rf3) (26) ± UsRisdm ±(1 -O I)WT sRsimr (32)

Or I

TI=A sin (Y '\ (27) where

\Omr /
isdm
'mr- and TB = /IWB-
The stator currents is' and isg'
are transformed to field-
oriented unidirectional quantities isd and iSq with the fol-
¢Imr
lowing coordinate transformation: Equations (31) and (32) are written in difference equation
form suitable to implement in a digital computer. They essen-
is [cos T sin 'I' Su
(21 across add the air gap voltage drop of the machine, and drop
tially
isq -sinm' cos 4 sQ' (28) the stator winding resistance, and leakage reactance.
The quantities VSd and Vsq are unidirectional quantities
The angular velocity of the rotor wr' i passed through a first- and are responsible for establishing the flux component of
order delay filter, which is characterized by the following dif- the current iSd and the torque component of the current isq
ference equation (see Appendix I): independently. The magnitude of the motor terminal voltage
when varied, varies not only the flux magnitude but also the
WFILn =A[wn + W(n-1)] -BWFIL(n-1) (29) torque magnitude. This interaction that arises due to non-
linearity in the machine is decoupled in the field-oriented
where A and B are filter constants, which depend on the gain quantities VSd and V5q.
and time constant of the filter and the sampling time.
VIII. FLUX POSITION CONTROLLER
VI. PI CONTROLLER The positioning of the voltage components in the stationary
The controllers used in this scheme are the proportional reference frame to establish the demanded field components
and integral (PI) type. This type of controller is characterized are quite important. The acceleration or deceleration of the
 SATHIKUMAR AND VITHAYATHIL: FIELD-ORIENTED CONTROL OF INDUCTION MOTOR 145

flux vector depends on the slip reference (demanded slip)

and the slip of the rotor from the flux vector. The slip
reference (SLR) is calculated from (19), repeated here
for convenience

w-w r =_ isqm SLR

Trimr
A slip limit of 3 percent is imposed on this reference based
on the pull-out torque criterion. The actual rotor speed wr
and the speed of the flux vector w are used to calculate the
actual slip. The slip error (SLE) was formed as per the algo-
rithm given in Fig. 6. This error quantity is the error in the Fig. 6. Algorithm for enumerating slip error (SLE).
angular velocity at any instant required for the flux vector
to attain the desired position. The integral of the error quan- 2. Step change in the load.
tity then gives the required displacement for the flux vector. 3. Speed reversal.
This is done by the flux position controller, which is basically
a PI controller with small gain. The output of the controller The behavior of the field coordinates and control variables
gives At, the incremental value of I. The desired position are plotted in Figs. 7-10. Fig. 7 shows the starting of the
I0 of the flux vector is then given by motor with set speed 0.5 PU. At t - 2 s, the speed reference
is taken to 1.2 PU, i.e., above base speed. The action of the
field weakening mechanism is shown by the flux waveform.
Fig. 8 shows the change of speed from 0.5 PU to 0.01 PU, i.e.,
IX. COORDINATE TRANSFORMATION near zero-speed operation. The flux waveform adapts to the
Then, using the following inverse coordinate transfor- new speed reference by changing the frequency of rotation.
mation, the voltage components are positioned: Fig. 9 shows the starting and running of the motor at 0.5
PU. A step load is applied at t = 2 s. The motor regains the
LvtJ cos To -six' j [Vsdmj (34) original speed after a small dip for a short duration. Following
the application of load, the field coordinate isq increases,
pVs, sin 4f cos To vsqJ showing the increased electrical torque generated to cater
Thfiethree phase quantities are obtained from V., and for the increased load. But the magnitude of the flux remains
by 2-axis to 3-axis transformation: constant. Fig. 10 shows the speed reversal operation of the
motor from 0.5 pu to -0.5 PU. The reversal operation with
va I 0 0.1 PU of load is completed within 0.48 s. Because of the slow
response of the flux there is a dip in magnitude during reversal
Vb = -2 Vr3/2 IV4o (35) which is regained subsequently.
_ 2 -V\[3/2 Li
X. CONCLUSION
These three phase quantities serve as the reference voltage
to the inverter. The inverter was modeled as a small delay Field-oriented control is found to be very efficient and
element, which can be ignored considering the whole dynamic fast for continuous-speed control of an induction motor from
response of the system. Hence, the reference voltage serves high frequency (above base speed) down to the stand-still
as the input to the motor. operation in all four quadrants of the torque-speed plane.
For the 2-axis model of the motor, the three phase quan- The control scheme involves a good amount of arithmetic
tities are converted to two phase quantities through the fol- and trignometric calculations. This makes it impractical to
lowing transformation: implement the scheme with discrete hardware. But the ad-
vancement of microprocessor technology will progressively
establish it as an efficient and economical means for imple-
menting such control schemes. The authors are presently
UYs] L° 1/4 -1/4 0VJ (36) working on the implementation on this control scheme.
The results of the system implemented will be published
in due course.
The program was written in Fortran language and run in
a DEC 1090 computer. The timesharing interactive terminal APPENDIX I
facilitated the tuning of the controllers. The responses of the The transfer function of the first-order delay filter using
system for the following disturbance were studied. the Laplace operator can be written as
1. Step change in the reference speed:
K
(a) above base-speed operation by field weakening, Ha(s)-
(b) near zero-speed operation. 1 +Ts
 146 IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS, VOL. IE-31, NO. 2, MAY 1984
I

I i,

1.0 I .. I.00 4.."

'. V -.U
.*
1. TI Mr .0" 1.00 2.00 3." 4.."
1 TI"E
_ n-

L --- -,.l
ik

8 -

S .wl
0.00 1.00 2.00 1.00 4.00
I Ni 0.00 1.00 1.00 .00 4.00
T ME
.1

K 1.00 1 .00 0.00 4.00

C1
I
110,

i
TiNE -.00 2.00 3.00 4.4so
-
T INm
.II
214
iL

0.0 3.00 4.00

! .-0
1.0- .00 I-
.4 44
tlmI 0 00 1.00 1.00 3." 4."

/0.L
TI Ing
IME
-:

I I3P,
I
T 4I0 I

*-I" 1.00 2.00

Tint
I." 4.09 a

' 0.00
T

1.00 2.00
T1 41
I
3.00
I
4.00

II
0I.0 a.00
1.00 3.00 p-

11~~
I FM
4.00
TIN p0.00 1.00 T1 1.00 4.00

I-a
1V..
a

1
V .~

I S 1.00 I ." a." D." s.0o

-a

I
TINE F 9 2 .0 0.00 4.00
TI N
li'i I
-
'3

1.00 *.00 3.00 4.00

------T-
0.00 1.00 2.00 3.00
TimE 4.00
TIME
Fig. 7. Step response of the speed reference (above base-speed operation).
Fig. 8. Step response of the speed reference (Near zero-speed operation).
where K is the gain of the filter, and T is the time constant Taking the inverse z transform
of the filter.
Applying the bilinear transformation s - az 1/ z + 1 - Yn =A(,n + n-l)-Byn-l
K K(z + 1)
H(z)- = APPENDIX II
z-1
z(1 + Ta) + (1-Ta) The finite difference equation that approximates the
1 + Ta ac-
z +1 tion of a continuous PI controller is written as
Y(z) z+ 1 K 1-Ta
= H(z) = A where A= -' B=
H(s
X(z) +b 1 + Ta
=
) KC (1+-)-KC
z 1 + Ta )Ts- (1+ 1S- 1)
and
where 7 is the time constant Kc is the gain.
Y(z) =
AX(z) + X(z)z - -B Y(z)z -1 Using the simple rectangular approximation for s- 1 with Ts
SATHIKUMAR AND VITHAYATHIL: FIELD-ORIENTED CONTROL OF INDUCTION MOTOR 147

a a

- ----------------

/
a 4
0 I T --I 0
0.0 1.00 2.00 3.00 4.00 '1.0a 1.00 2.00 3.00 4.00
TIME TIME
8

8 .L
"M-,
- ,.

Ia
0.00 1.00 3.00 3.00 4.00 1.00 3.00 3.00 4.00
TIME
TIME
IS."
a 1
r-
C~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
13

I
~~~~~~~~
0.00 ,.oo *.00 3.00 4.oo 0.00 I.0 . ." 3.00 4.00
TI PK

!1
TIHE

x
:.l
w

ti
6.G ."00 1 ."0 2.00 3.00 4.00 1.00 3.40 3.00 4.00
TI HE T Ing
a

II
I'
0
O." I1.00 2.00 1.00
!
w
0.00
/ I-.* .40
I
1.00 4,"0
TIME TIME
a

p0l
a
0
LK
1 ." 1.00
9
2.00 3.00 4.00 0.4a0 3.00 4.00
TI Mt T I ME

U- o, L, ,L..
, .A.l.i,A,lL.l,,.,.,.al., t.|j~~IiI,t.iAf.,^.l
1111s L%1I.
,i 8 .01.0 00S-o
I I I I
0.00 1.60 2.I0 3.00 4.00 S.4
TIRE tINN
1
q
-ilwlwwwwwwwjwm 0 911whilloldl Alb UWIWiddhai-
Mpilmmmnmmm ml mmmimmoii VI
I;
...............
.. ..... ..

0." 1.00 t." 3."0 4.00

5.0. S I -a B." .a." 4.00
I(ME TlI
Fig. 9. Step response of the load disturbance. Fig. IO. Speed reversal operation.
 148 IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS, VOL. IE-31, NO. 2, MAY 1984
as the sampling time, we get Motor Constants in PU
11-z-1 + ,ls R = 0.039 LrI = 1.532
M(z) 1 T"s \
D(z) =E
E(z) r 1 +Z-l Kc I1-z-1 RrI= 0.0143 Lm ' = 1.48
Ls = 1.532 =
= 114.23

- 11 REFERENCES
/ | Kc Z -j7T TS) [1] F. Blaschke and K. Bohm, "Verfahnen der Felderfassung bei der
/ _( + Ts) z/-1 Regalung stromrichtergespeister asynchronmaschinen," in Proc. 1st
IFAC Symp. Control in Power Electronics and Electrical Drives,
1974, pp. 653-649.
Kc z-a 7 [2] L. J. Garces, "Parameter adaption for the speed controlled static AC
D(z) = c , where a = drive with squirrel-case induction motor," in IAS Conf. Proc., 1979,
ae z -1 + Ts pp. 843-850.
[31 R. Gabriel, W. Leonhard, and C. Nordby, "Field oriented control of a
standard AC-motor using microprocessors," in IAS Conf. Proc.,
K 1979, pp. 910-916.
M(z)(z-1) = - (z -ct)E(z). [41 W. Flotter and H. Ripperger, "Field oriented closed-loop control of an
a induction machine with the new transvector control system," Siemens
Rev., vol. 39, pp. 248-252, 1972.
Taking the inverse z transform [51 R. Krishnan, W. A. Maslowski, and V. R. Stefanovic, "Control
principles in current source induction motor drives," in IAS Conf.
Proc., 1980, pp. 605-617.
K [6] F. Blaschke, "The principle of field orientation as applied to the new
Mn=Mn- +- En-KcEn-- transvector closed loop control systems for rotating field machines,"
Siemens Rev., vol. 39, no. 5, pp. 217-220, 1972.
[7] M. Kaimoto, M. Hashii, T. Yanase, and T. Wakano, "Performance
APPENDIX III improvement of current source inverter fed induction motor drives,"
IEEE Trans. Ind. Appl., vol. IA-18, no. 6, pp. 703-709, Dec. 1982.
Motor Details [8] M. Akamatsu, K. Ikeda, H. Tomei, and S. Yano, "High performance
IM drive by coordinate control using a controlled current inverter,"
2.2 kW, 400 V, 4.5 A, 2870 rpm, 3 ph, 50 Hz. IEEE Trans. Ind. Appl., vol. IA-18, no. 4, pp. 382-392, Aug. 1982.