International J. of Engg. Research & Indu. Appls. (IJERIA).

ISSN 0974-1518, Vol.3, No. I (2010), pp 211-223

SVPWM - MRAS SPEED OBSERVER FOR SENSORLESS

CONTROL OF INDUCTION MOTOR DRIVE

MD. HASEEB KHAN AND J. AMARNATH

Abstract
This paper proposes a novel Space Vector Pulse width modulation (SVPWM) for sensorless control of
induction motor using model reference adaptive system (MRAS). The steady state ripples in the
torque are present in the conventionally used MRAS sensorless control of induction motor which
utilizes normally used voltage source inverters. Also performance of the steady state speed is not as
perfect as required having disturbances in steady state region. Hence to improve the performance of
MRAS based speed observer a novel method of SVPWM based on reference voltage vector that
utilizes the control variables as stator flux components is proposed. By using the proposed SVPWM
control of induction motor the speed disturbances which are obtained are minimized and the speed
performance is improved. Also the ripples present in the electromagnetic torque are reduced. This is
proved by the simulation results for conventional MRAS speed observer and proposed SVPWM based
MRAS speed observer.

----------------------------------------
Keywords: Sensorless Control, Model Reference Adaptive System, Reference voltage Vector,
SVPWM
212 MD. HASEEB KHAN AND J. AMARNATH

1. INTRODUCTION
Space vector pulse width modulation (SVPWM) technique is being for a long time now.
The main advantage of this technique is that the switching losses are low; the harmonic
performance is improved and produces an improved output voltage for the same dc bus
voltage [1-4], when compared to a sinusoidal PWM (SPWM). SPWM is a technique in
which PWM outputs are produced based on the principle that compares a triangular carrier
signal with a sinusoidal reference wave. [5-7]. SVPWM is an advanced PWM method and
one of the best methods among all the different types of PWM methods available in literature
for variable speed drives and hence becomes the mostly used PWM technique [8-11]. This
scheme of SVPWM may look simple theoretically but the implementation of this scheme
becomes little difficult with the use of low sampling period or high switching frequency [12].
In [13] space vector equivalent references of a carrier-based unbalanced two-phase output
space vector pulse width modulation (SVPWM) strategy is applied to a three-leg voltage
source inverter where the space vector equivalent references are obtained from conventional
balanced SVPWM method. When multilevel inverters are used the conventional SVPWM
scheme becomes more and more complicated to implement due to the increase in number of
space vectors [14-16]. To overcome this, a method of SVPWM based on reference voltage
vector which utilizes the control variable as stator flux component is proposed. By using the
proposed method not only the ripples present in the torque are reduced but also the transient
performance of the drive is improved.

2. CONVENTIONAL SVPWM
A three phase voltage source inverter is shown in fig.1 where a, b, c are the three phases and
the voltages with respect to these phases are taken as Va, Vb, Vc which are applied to the
induction motor. Six power transistors, 3 in the upper half and 3 in the lower half form a
bridge inverter with Vdc split as +Vdc/2 and Vdc/2 with 0 as the dc bus centre.
SVPWM - MRAS SPEED OBSERVER FOR SENSORLESS CONTROL OF 213

+Vdc
2
0 a b c
-Vdc
2

Figure 1. Three leg Voltage source inverter.

Every phase of three phase Voltage source inverter can be connected to either upper
transistor or lower transistor. When ever the upper transistor gets connected it is taken as 1
for a or b or c and when the lower transistor gets connected it is taken as 0 for a or b or c.
Therefore it can be seen that there are 8 (23) switching states or space vectors possible for the
above voltage source inverter which are designated as V0 to V7. The 8 switching states are
given as V0(000), V1(100), V2(110), V3(010), V4(011), V5(001), V6(101), V7(111). Out of
these 8 possible space vectors 6 vectors i.e., V1 to V6 are non-zero vectors and V0 and V7 are
zero vectors present at origin as shown in figure 2.
These non-zero vectors form an origin centered hexagon with zero vectors at origin as shown
in figure 2. The maximum boundary of the space vector forms the hexagon and the circle is
the maximum trajectory of the sinusoidal outputs in the linear modulation. The PWM output
patterns of the inverter are shown as I VI sectors.

V3 q V2

II
III I
V0,V7 d
V4 V1
IV VI
V

V5 V6
Figure 2. Voltage Space Vector of the VSI
214 MD. HASEEB KHAN AND J. AMARNATH

During the non-zero states the power gets transferred between dc bus and induction motor
and during zero states there is no power transfer between the dc bus and induction motor.
The non-zero vectors are also known as active vectors.
For the VSI considered in figure 1 the phase voltages are taken as Van, Vbn, Vcn and pole
voltages are taken as Vao, Vbo, Vco and are given as

Vao = Van + Vno

Vbo = Vbn + Vno (1)
Vco = Vcn + Vno
where Vno is common node voltage.

By using conventional three phase to two phase transformation the three phase voltages Va,
Vb, Vc can be transformed into (Vd, Vq) as the reference vector Vref which is expressed as in (2)
The decomposition of the reference vector Vref in sectorI is shown in figure 3 where the XV1
is the component along vector V1 and XV2 is along V2 vector. The values of XV1 and XV2 are
dependent on the duration (T1, T2) for which active states 1 and 2 are to be applied
respectively in a given sampling period Ts.

1 1 V
Vd 2 1 2 .V
a

= = . 2
Vref (2)
Vq 3 0 3 3 b
V
2 2 c

V2

XV2 Vref

V1
d
XV1
Figure 3. Decomposition of Vref
SVPWM - MRAS SPEED OBSERVER FOR SENSORLESS CONTROL OF 215

With the above discussion XV1 and XV2 can be given as

T1 T
X V1 = V1 and X V 2 = 2 V2 (3)
Ts Ts

T1 T2
and Vref = V1 + V2 (4)
Ts Ts

In general it is given as
Tn T
Vref = Vn + n +1 Vn +1 (5)
Ts Ts
where Tn and Tn+1 are durations for which adjacent vectors or active states Vn and Vn+1 are to
be applied in a given sampling period Ts.
The values of Tn and Tn+1 are given as

Tn =
3Ts
Vdc
{
Vd sin( n * 60 o ) Vq cos(n * 60 0 ) } (6)

Tn +1 =
3Ts
Vdc
{ [ ] [
Vd sin (n 1) * 60 0 + Vq cos (n 1) * 60 0 ]} (7)

To keep the switching frequency constant the remainder of the time is used for zero state
vectors in each sampling period and is given as
T0 = Ts Tn Tn +1 (8)

where Ts is sampling period.

3. PROPOSED SVPWM TECHNIQUE

In the proposed method, the position of the reference stator flux vector s * is derived by

the addition of slip speed and actual rotor speed. The actual synchronous speed of the stator

flux vector s is calculated from the adaptive motor model. After each sampling interval,

actual stator flux vector s is corrected by the error and it tries to attain the reference flux

space vector s * . Thus the flux error is minimized in each sampling interval. The d-axis and

q-axis components of the reference voltage vector can be obtained as follows.

216 MD. HASEEB KHAN AND J. AMARNATH

Reference values of the d-axis and q- axis stator fluxes and actual values of the d-axis and q-
axis stator fluxes are compared in the reference voltage vector calculator block and hence the
errors in the d-axis and q-axis stator flux vectors are obtained as

ds = ds * ds (9)

qs = qs * qs (10)

The knowledge of flux error and stator ohmic drop allows the determination of appropriate
reference voltage space vectors as given as

* = ds (11)
V ds R s i ds +
Ts
qs
* =
V qs R s i qs + (12)
Ts
Where, Ts is sampling period also known as the duration of sub-cycle and it is a half of
period of the switching frequency. It means that the torque and flux are controlled for each
switching half cycle or controlled twice per switching cycle. Also the reference voltage
vector d-q components are fed to the SVPWM block from which the switching times are
calculated for each inverter leg.
Vdc
Ref.
Speed e e Calculation
+ _ of Vref SVPWM VSI

Speed s s*

Cal. of Vds
Model & Vqs
Reference
Adaptive
System 3 to 2
Transformation

IM

Figure 4. Block diagram of SVPWM-MRAS speed observer for Induction Motor.

SVPWM - MRAS SPEED OBSERVER FOR SENSORLESS CONTROL OF 217

4. MODEL REFERENCE ADAPTIVE SYSTEM (MRAS)

The block diagram of conventional MRAS speed observer for sensorless control of Induction
Motor is shown in figure 5. It basically consists of two models namely reference model and
adaptive (adjustable) model. Another important content of MRAS scheme is the adaptation
mechanism which is designed in such a way so as to generate the value of estimated speed
which minimizes the error between the reference and the estimated fluxes [17, 18].

Vs Ref. Model X
Eqs. (13), (14)
is
+ ex
_
Adjustable Model X
Eqs. (15), (16)

r Adaptation
Mechanism

Figure 5. Structure of Model Reference Adaptive Systems

The reference model also known as voltage model is developed by using the following
reference rotor flux components as [18, 19]:
Lr
p dr = (Rsids + Ls pids vds ) (13)
Lm

p qr =
Lr
(Rs iqs + Ls piqs vqs ) (14)
Lm
The adjustable model also known as current model is developed by using the rotor flux
components as [18, 19]:
1
p dr = ( Lm ids dr ) r qr (15)
Tr
1
p qr = ( Lm i qs qr ) + r dr (16)
Tr
where is leakage coefficient which is given as
218 MD. HASEEB KHAN AND J. AMARNATH

2
L
= 1 m (17)
Ls Lr
The adaptation mechanism is designed in a way to reduce the error between the actual flux
and estimated flux there by reducing the error between the actual speed and the estimated
speed.
The equation for speed tuning signal is given as
ex = qr dr dr qr (18)

and the estimated speed is given as

k
r = k p + i ex (19)
p
The electromagnetic torque equation of induction motor in stationary frame is given as
3 p Lm
Te = (i qs dr i ds qr ) (20)
2 2 Lr

5. SIMULATION RESULTS AND DISCUSSIONS

Simulations of the conventional MRAS based control of induction motor and proposed
SVPWM-MRAS speed observer for sensorless control of induction motor are performed on
MATLAB / SMULINK. Induction motor having Rr=3.55, Rs=4.15, Lr=Ls=0.76mH,
Lm=0.646mH with 2 poles and moment of inertia 0.33Kg-m2 is considered for performing
the simulations. Figure 6 and 7 show the transient response of the motor when the speed is
changed from 100rpm to 150rpm at 0.5 sec with a step in load torque command where load
of 10N-m is applied at 0.2 sec and removed at 0.6 sec. In figure 8 and 9 transient responses
are shown for SVPWM based MRAS speed observer for the above conditions. The speed
obtained when SVPWM is used is better in performance as the transient disturbances
produced in the speed are less which can be observed from the waveforms. Also the ripples
present in the torque are reduced.
SVPWM - MRAS SPEED OBSERVER FOR SENSORLESS CONTROL OF 219

160

140

120 Estimated
Reference
100

80
Speed

60

40

20

-20
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Time

Figure 6. Speed response of conventional MRAS

50

40

30
Torque

20

10

-10
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Time
Figure 7. Motor Torque of conventional MRAS
160

140
Estimated
120 Reference
100

80
Speed

60

40

20

-20
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Time

Figure 8. Speed response of proposed drive

220 MD. HASEEB KHAN AND J. AMARNATH

30

25

20

15
Torque

10

-5
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Time

Figure 9. Motor Torque of proposed drive

180

160

140
Reference
120
Estimated
100
Speed

80

60

40

20

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

Time

Figure 10. Speed response of conventional MRAS

50

40

30
Torque

20

10

-10
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Time

Figure 11. Motor Torque of conventional MRAS

SVPWM - MRAS SPEED OBSERVER FOR SENSORLESS CONTROL OF 221

160

140

120 Estimated
Reference
100

80
Speed

60

40

20

-20
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Time

Figure 12. Speed response of proposed drive

30

25

20

15
Torque

10

-5
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Time

Figure 13. Motor Torque of proposed drive

The transient response of induction motor using conventional MRAS and proposed drive are
shown respectively from figure 10 to figure 13 when the speed changes from 100 to 150 rpm
at 0.2 sec and load torque of 10 N-m applied at 0.5 sec and removed at 0.8sec. From
simulation results it can be seen that the proposed SVPWM base drive give better
performance over the conventional drive. The ripples present in the torque are also reduced.

CONCLUSION
The control of the induction motor drive with conventional MRAS speed observer give rise
to ripples and these ripples becomes large with load variation as can be seen from the
waveforms. To overcome this MRAS observer using SVPWM based on reference voltage
222 MD. HASEEB KHAN AND J. AMARNATH

vector is proposed and implemented. By using the proposed method not only the transient
response is improved but also the ripples present in the torque are minimized to a large
extent even when load is varied.

REFERENCES

[1] Jabbar M. A., Khambadkone A.M., and Zhang Y., Space-vector modulation in a two-phase
induction motor drive for constant-power operation, IEEE Trans. Ind. Electron., vol. 51, no. 5,
pp. 10811088, Oct. 2004.
[2] Kumar V. S. and Kannan P. S., Harmonic studies in space vector PWM inverter drive system,
in Proc. Int. Conf. Power Syst. Technol., 2004, vol. 1, pp. 123127.
[3] Trzynadlowski M., Borisov K., Yuan L., and Ling Q., A novel random PWM technique with
low computational overhead and constant sampling frequency for high-volume, low-cost
applications, IEEE Trans. Power Electron., vol. 20, no. 1, pp. 116122, Jan. 2005.
[4] McGrath P., Holmes D. G., and Meynard T., Reduced pwm harmonic distortion for multilevel
inverters operating over a wide modulation range, IEEE Trans. Power Electron., vol. 21, no. 4,
pp. 941949, Jul. 2006.
[5] Bowes S. R. and Yen-Shin L., The relationship between space-vector modulation and regular-
sampled PWM, IEEE Trans. Ind. Electron., vol. 44, no. 5, pp. 670679, Oct. 1997.
[6] Keliang Z. and Danwei W., Relationship between space-vector modulation and three-phase
carrier-based PWM:Acomprehensive analysis, IEEE Trans. Ind. Electron., vol. 49, no. 1, pp.
186196, Feb. 2002.
[7] Bai H., Zhao Z., Meng S., Liu J., and Sun X., Comparison of three PWM strategies-SPWM,
SVPWM & one-cycle control, in Proc. 5th Int. Conf. Power Electron. Drive Syst., 2003, vol. 2,
pp. 13131316.
[8] Iqbal A., Analysis of space vector pulse width modulation for a five phase voltage source
inverterIE (I) journal-EL, Volume 89 Issue 3.September 2008 Pages 8-15.
[9] Hava Ahmet M. and Un Emre, Performance Analysis of Reduced Common-Mode Voltage
PWM Methods and Comparison with Standard PWM Methods for Three-Phase Voltage-Source
Inverters IEEE Transactions on Power Electronics, Volume. 24, Issue 1, January 2009 Pages(s):
241 252.
[10] Narayanan G., Krishnamurthy H. K., Di Z., and Ayyanar R., Advanced bus-clamping pwm
techniquesbased on space vector approach, IEEE Trans. Power Electron., vol. 21, no. 4, pp.
974984, Jul. 2006.
[11] Hyoung R. Joon, Won L. Kwang, and Sung L. Ja, A unified flux and torque control method for
dtc-based induction-motor drives, IEEE Trans. Power Electron., vol. 21, no. 1, pp. 234242,
Jan. 2006.
[12] Celanovic N. and Boroyevich D., A fast space vector modulation algorithm for multilevel
three-phase converters, in Proc. Ind. Appl. Conf., 1999, vol. 2, pp. 11731177.
SVPWM - MRAS SPEED OBSERVER FOR SENSORLESS CONTROL OF 223

[13] Kinnares Vijit and Charumit Chakrapong Modulating Functions of Space Vector PWM for
Three-Leg VSI-Fed Unbalanced Two-Phase Induction Motors IEEE Transactions on Power
Electronics, Volume. 24, Issue 4, April 2009 Pages(s): 1135 1139.
[14] Tian L., Qiang S., Wenhua L., Yuanhua C., and Jianguo L., FPGAbased universal multilevel
space vector modulator, in Proc. IECON 32nd Annu. Conf., 2005, pp. 745749.
[15] Nho N. V. and Youn M. J., Comprehensive study on spacevector- pwm and carrier-based-pwm
correlation in multilevel inverters, Proc. Inst. Elect. Eng., vol. 153, no. 1, pp. 149158, 2006.
[16] Xiaomin K., Corzine K. A., and Wielebski M. W., Overdistention operation of cascaded
multilevel inverters, IEEE Trans. Ind. Appl., vol. 42, no. 3, pp. 817824, May/Jun. 2006.
[17] Tamai S., Sugimoto H., Yano M., "Speed-sensorless vector control of induction motor with
model reference adaptive system", Conf. Record of the 1985 IEEE-IAS Annual Meeting, pp.
613-620, 1985.
[18] Shauder C., "Adaptive speed identification for vector control of induction motor without
rotational transducers", IEEE Trans. Ind. Application, Vol. 28, No. 5, pp. 1054-1061, Sept./Oct.
1992
[19] Vas P., Sensorless Vector and Direct torque control. New York: Oxford University Press, 1998

Md. Haseeb Khan

Royal Institute of Technology and Science,
Chvella, RR Dist.
Andhra Pradesh, India
Email: haseeb_md@yahoo.com,

J. Amarnath
JNTU College of Engineering,
JNT University Hyderabad,
Andhra Pradesh, India
Email: amarnathjinka@yahoo.com