International Journal of Engineering Research and Applications (IJERA) ISSN: 2248-9622

National Conference On Advances in Energy and Power Control Engineering (AEPCE-2K12)

DC Link approach to Variable-Speed, Sensorless,

Induction Motor Drive
Ch.U.Phanendra.Kumar 1
Asst.Professor ,HOD,LIMAT
[email protected]

SK.Mohiddin2
Asst.Professor, LIMAT
[email protected]

Abstract This paper presents a new control strategy for

three-phase induction motor which includes independent speed &
torque control loops and hence
current regulation
thereby overcoming the limitation (i.e. sluggish response) of
volts per hertz controlled method. For close-loop control,
the feedback signals including the rotor speed, flux and torque
are not measured directly but are estimated by means of
an algorithm. The inputs to this algorithm are the
reconstructed waveforms of stator currents and voltages
obtained from the dc link indirectly on stator side. The proposed
drive thus requires only one sensor in the dc link to
implement the close-loop speed and torque control of a threephase induction motor. The simulation results on a 2.2 kW
induction motor drive in Matlab/Simulink software show fast
dynamic response and good agreement between the actual
values and the estimated values of torque and speed.
Replacement of the open-loop control strategy of existing v/f
drive by the proposed close-loop strategy appears to be possible
without requiring any additional power components and sensors.
Index Terms Speed-sensorless, estimation, dc link, band-pass
filter, reconstruction, three-phase induction motor, space-vector.

NOMENCLATURE
Rs Rr '
Lm

Lr '

Stator and rotor resistances ( )

Magnetizing and rotor inductances (Henry)
I.

INTRODUCTION

The widespread industrial use of induction motor (IM) has

been stimulated over the years by their relative cheapness, low
maintenance and high reliability. The control of IM variable
speed drives [1] often requires control of machine currents,
which is normally achieved by using a voltage source inverter.
A large number of control strategies have been registered so
far [2]-[4]. The volts per hertz (v/f) IM drives with inverters
are widely used in a number of industrial applications
promising not only energy saving, but also improvement in
productivity and quality. The low cost applications usually
adopt v/f scalar control when no particular performance is
required. Variable-speed pumps, fans are the examples.
For those applications which require higher dynamic
performance than v/f control, the dc motor like control of IM
that is called, the field oriented control (FOC) is preferred.
During the last few years, a particular interest has been noted

Vignans Lara Institute of Technology and Science

A.Hanumaiah 3
Professor,VLITS
[email protected]

on applying speed sensorless FOC to high performance

applications that is based on estimation of rotor speed by using
the machine parameters, instantaneous stator currents and
voltages [1]-[6]. The benefits of speed sensorless control are
the increased reliability of overall system with the removal of
mechanical sensors, thereby reducing sensor noise and drift
effects as well as cost and size. However to exploit the
benefits of sensorless control, the speed estimation methods
must achieve robustness against model and parameter
uncertainties over a wide speed range. To address this issue, a
variety of approaches have been proposed.
The adaptive observers (AO) like Luenberger observer or the
extended Kalman filter [1], [5] gets accurate estimates under
detuned operating conditions but these solutions are
computationally intensive, require more memory space and are
difficult to tune because the initial values of three covariance
matrices have to be assumed and selected after much trial and
error. So their application in low cost drives is limited. The
model reference adaptive system is also an AO technique [7],
where the same quantity is calculated by two different ways.
One of them is independent of variable to be estimated while
the other one is dependent on it. The two computed quantities
are used to formulate the error signal. The error signal is then
fed to an adaptation mechanism which in most cases is a PI
controller. The output of the adaptation mechanism is the
estimated quantity.
While all the speed sensorless techniques eliminate the use
of mechanical speed sensor, they require the stator current and
stator voltage signals as input. This requires at-least two
current sensors and two voltage sensors on the stator side. It is
difficult to get current sensors with equal gains over the wide
range of frequencies, voltages and currents used in a practical
inverter. The problem is exacerbated if the motor windings are
not perfectly balanced or if the current sensors have some dc
offset. Over last few years, techniques of stator current
reconstruction from the dc link current have been suggested in
literature [8]-[9].
In this paper, a new speed sensorless control strategy for IM
is proposed that includes the speed control, torque control and
current regulation. Unlike conventional close loop estimators,
it involves less computation and is less dependent on machine
parameters. The stator currents and stator voltages are

Page 60

International Journal of Engineering Research and Applications (IJERA) ISSN: 2248-9622

National Conference On Advances in Energy and Power Control Engineering (AEPCE-2K12)
reconstructed from dc link quantities and the inverter
switching signals. For faithful reconstruction of currents, use
of adaptable gain band-pass filter is proposed in the scheme.
The simulation results of proposed scheme shows fast
performance as compared to v/f control and therefore can be
regarded as an improvement. For the close loop speed control,
a single current sensor in the dc link is sufficient. Thus it is
suitable for low-cost, moderate performance, sensorless IM
drive applications. The proposed drive is modeled in
Matlab/Simulink software for a 2.2 kW IM. The simulation
results are presented to verify the workability of proposed
strategy.
II. PROPOSED SCHEME
Fig. 1 shows the block diagram of the proposed scheme. It
consists of a speed (frequency loop), a torque loop, and a
current regulator. The output of speed/frequency regulator
represents the torque reference for the torque loop. The torque
Regulator generates the q-axis current command iqe.The di qs is directly generated from the
axis current command i*
reference rotor flux r * as given by (1) [1]. This
eliminates an additional PI controller and reduces the
computational burden. These dc commands expressed in
synchronously rotating reference after transformation to the
three phase current commands are than compared with the
actual three- phase currents (reconstructed waveforms) to
generate the switching signals for the inverter. In the proposed
scheme, all the feedback signals including the stator currents
and stator voltages are estimated/reconstructed from the
dc link quantities.
Idc*
III.

(1)

RECONSTRUCTION OF STATOR VOLTAGES & CURRENTS

FROM DC LINK

As indicated in [1], [6] ,the stator flux,torque and speed can

be derived from the stator voltatges and currents expressed
in d-q reference frame.The phase currents and voltages
are related to the dc link current and voltage by inverter
switching states.A coltage source inverter IM drive is shown
in Fig 2.where Vdc is the dc link voltage ,Idc is the
instantaneous dc link current and and ia, ib, ic are the
instantaneous three-phase winding currents.

Generally, IGBTs associated with snubber protection and

feedback diode are used as switch in inverters. When a
switch is being turned-on and the conducting diode at the same
leg is being blocked off by this turn-on, because of the reverse
recovery effect of the diode,this leg is shorted through . To
establish the basic relationship between dc link current,
winding currents and inverter switching pattern, the switches
shown in Fig. 2 are considered as ideal; the diode recovery
effect and the snubber action are not considered.

. Figure 2.

Voltage source inverter fed induction motor drive

A. Space-Vectors
During normal state, there are eight switching states of inverter
which can be expressed as space voltage vector (SA,SB,SC) such
as (0,0,0), (0,0,1), (0,1,0), (0,1,1), (1,0,0), (1,0,1), (1,1,0) and
(1,1,1). SA =1 means upper switch of leg A is on while
the lower one is off, and vice versa. The same logic is
applicable to SB and SC also. Amongst above eight voltage
vectors, (0, 0,0) and (1,1,1) are termed as zero vectors while
the other six as active vectors. The switching vectors
describe the inverter output voltages.
B. Basic Principle of Phase Voltage & Line Current
Reconstruction
For different voltage vectors, the phase voltage that will
appear across stator winding can be determined by circuit
the stator winding is star connected. From this table,the
reconstructed expressions of three phase voltages are:
Va

(2)

Vb

(3)

Vc

(4)

The stator voltages are expressed in stationary d-q frame are:

Vqs=Va

(5)

Vds=

(6)

Figure 1. Block diagram of the proposed scheme.

.
DC LINK CURRENT & PHASE VOLTAGES

Vignans Lara Institute of Technology and Science

Page 61

International Journal of Engineering Research and Applications (IJERA) ISSN: 2248-9622

National Conference On Advances in Energy and Power Control Engineering (AEPCE-2K12)
500
va(V)

IV.SIMULATION STUDIES
In order to predict the behavior of the drive during steady-state
and transient conditions, detailed simulation studies of the
scheme shown in Fig.1 are carried out on a 2.2kW IM by
using Simulink software. Fig. 3. shows the internal structure of
the controller that consists of the speed loop, torque loop and
the current regulation loop in synchronously rotating frame of
reference. The switching signals for inverter are generated by
comparing the command ac currents with reconstructed ac
currents. For the reconstruction of stator voltages and ac line
currents, the dc link quantities with Vdc = 600V are sampled
with a sampling time of 2e-6 seconds and than segregated into
the three-phase voltages and three ac currents as per (2)-(4)
and (7)-(9) respectively. The simulation was carried out for
five different operating conditions as is presented ahead. A
variable- step ode23tb(stiff/TR-BDF2) solver was used. The
waveforms of reconstructed phase voltages and the three ac
line currents as reflected in the dc current, are presented in
Fig.4. From these waveforms, it is clear that the samples of
phase currents available in the dc link current are not evenly
spread and being discontinuous, the set of resulting points do
not constitute an acceptable reconstruction. Therefore a zeroorder hold is employed followed by a band-pass filter. The

0
-500
0.05

0.06

0.0 7

0.0 8

0.09

0.1

0.06

0.0 7

0.0 8

0.09

0.1

0.06

0.0 7
0.0 8
Ti m e (s )

0.09

0.1

500
0
-500
0.05

vc(V)

500
0
-500
0.05

iAindclink(A
)

(a)
5
0
-5

iBindclink(A
)

1 .78

1 .8 2

1 .84

1 .8

1 .8 2

1 .84

1 .8 2

1 .84

0
-5
1 .78

iCindclink(A
)

1 .8

5
0
-5
1 .78

1 .8

Time (A)

(b)
Figure 4. Reconstructed waveforms of (a) three phase voltages and (b) three
line currents separated from the dc link current.

currents'a,b,c',(A)

Figure 3. Simulink model of control strategy

Reconstructedline

values of time constants T for the band-pass filter are selected

by trial and error. The simulation output of band-pass filter
which represents the reconstructed ac line currents is shown in
Fig.5(a). For the sake of comparison, the actual ac line
currents are illustrated in Fig.5(b). The reconstructed and
actual waveforms of ac line currents during 100% speed
reversal at no-load are presented in Fig.6(a) & (b). The
response of speed sensorless drive during different
5
0
-5
1.78

1 .8

1.82

1.84

Tim e (s )

A
ctual line

currents'a,b,c' (A
)

(a)
5

-5
1 .78

1 .8

1 .82

1 .84

Time (s )

(b)

currents(A)

Reconstructed

Figure 5. Stator currents at rated load (a) reconstructed (b) actual waveforms
10
0
-10
1.1

1.2

1.3

Time (s)

(a)
Vignans Lara Institute of Technology and Science

Page 62

Figure 6. Stator currents during reversal at no-load (a) reconstructed

waveform (b) actual waveform

dynamic conditions was studied in detail. To check the

accuracy of estimated variables, these variables were obtained
by two different methods. In the first one, the machine
variables which include the flux, torque, synchronous-speed,
slip-speed and rotor-speed are estimated by using (15)-(24)
and in the second method, these variables are calculated with
the help of dynamic model of IM [1] by using the stator
currents and voltages measured directly. The simulation
results of the first method were treated as estimated values
while those of the latter method as actual values.
Case 1: Free acceleration characteristics:
The machine was allowed to accelerate from zero speed to
rated speed at no-load. The steady-state was reached at 0.3
seconds. The waveform of estimated speed show faster
response (less damped) as compared to its actual counterpart.
This is shown in Fig. 7(a) & (b).
Case 2: Step change in speed reference:
Step change in speed reference was applied two times. At 0.5
sec., from +100% to +60% and vice-versa at 1 sec. was
applied. The response is shown in Fig.8. The torque becomes
negative during the first change to decelerate the motor. Upon
reaching steady state, the torque becomes equal to the load
torque. The response time of the drive for this step change is
100ms. The estimated values of torque and speed vary in
accordance with their corresponding actual values.
Case 3: Speed reversal:
A step change in speed reference from +100% to -100% is
applied at 1.5 seconds. This step change is equivalent to 100%
speed change. The response is shown in Fig. 9. The phase
sequence reverses to rotate the motor in reverse direction. The
drive reaches steady state after the change in reference speed
in 700 ms. this proves that the speed estimation is stable even
at very low speeds.
Case 4: Step change in load:
A step change in load is applied at 0.5 seconds. The response
of the drive is shown in Fig.10. The electromagnetic torque
increases to correct the speed error. Upon reaching the steady
state, the torque becomes equal to the load torque. The rotor
speed, after an initial droop attains back its earlier speed . The
motor reaches the steady state in 300ms.
Case 5: Low speed operation:
The response of the drive at 40% and 20% of rated speed is
shown in Fig.11. For the machine under consideration, 20%
corresponds to 3.14 rad/sec angular mechanical speed. The
speed estimation is very stable even at this low speed range.
VII.

CONCLUSION

In this paper, a new control strategy for induction motor drive

is proposed. The drive is operated under torque control with an
Vignans Lara Institute of Technology and Science

Estim
atedtorque(p.u) Estim
atedspeed(p.u.)

(b)

A
ctual speed(p.u.)

1. 3

Actual torque(p.u.)

1.2
Time (A)

1
0.5
0
0

0.1

0.2

0.3

0.4

0.5

0.1

0.2
0.3
Tim e (s )

0.4

0.5

15
10
5
0
-5

(a)
1
0.5
0
0

0.1

0.2

0.3

0.4

0.5

0.1

0.2
0.3
Ti m e (s)

0.4

0.5

15
10
5
0
-5

(b)
Figure 7. Free-acceleration characteristics (a) estimated & (b) actual values
u
E
stim
a
te
dto
rq
e(p
.u
.) E
e
stim
a
te
dsp
e
d(p
.u
.

1. 1

1
0.5
0
0.4

0.6

0.4

0.6

0.8

1.2

1.2

10
0
-10

0.8
Time (s )

(a)
Actual speed(p.u.)

0
-10

Actual torque(p.u.)

10

outer speed loop and is very similar to open-loop v/f drive in

terms of power components and sensors required. Due to the
inclusion of torque control loop, the drive response is fast and
stable. Simulation results confirm the effectiveness of the
proposed scheme. The technique uses only dc link voltage and
dc link current measurements to generate the estimates of

1
0.5
0
0.4

0.6

0.4

0.6

0.8

1.2

1.2

10
0
-10
0.8
Tim e (s)

(b)
Figure 8. Variation in rotor speed and electromagnetic torque for step
changes in reference speed (a) estimated values, (b) actual values
Estim
atedtorque(p.u.)Estim
atedspeed(p.u.

Actual currents(A)

International Journal of Engineering Research and Applications (IJERA) ISSN: 2248-9622

National Conference On Advances in Energy and Power Control Engineering (AEPCE-2K12)

1
0
-1
1.4

1 .6

1.4

1 .6

1 .8

2 .2

1 .8

2 .2

5
0
-5
-1 0

Time (s)

(a)

Page 63

Actual speed(p.u.)

International Journal of Engineering Research and Applications (IJERA) ISSN: 2248-9622

National Conference On Advances in Energy and Power Control Engineering (AEPCE-2K12)
phase voltages, line currents, flux, torque and rotor speed. If
the dc link voltage is assumed as constant, only one current
sensor in the dc link is sufficient to give the estimates of all
required feedback variables. Moreover, the same current
sensor that is already available in the dc link of an open-loop
v/f drive for protection purpose can be used. Thus the openloop control strategy in an existing v/f drive can be replaced
by the proposed close-loop control strategy without requiring
any additional power components or the physical sensors. The
proposed strategy appears to be a good compromise between

1
1.4

1.6

1.4

1.6

1.8

2.2

1.8

2.2

Actual torque(p.u.)

-1

5
0

Time (s )

(b)

-5
-10

Estimatedtorque(p.u.)Estimatedspeed(p.u.)

Figure 9. Variation in rotor speed and electromagnetic torque during reversal

(a) estimated values, (b) actual values

1
0.5
0
0

0.2

0.4

0.6

0.8

0.2

0.4

0.6

0.8

15
10
5
0
-5

Time (s)

Actual torque(p.u.)

Actual speed(p.u.)

(a)

the high-cost, high-performance field-oriented drives and the

low-cost, low-performance v/f drives.
Practical implementation of the proposed scheme on a 16
bits floating point arithmetic Texas Instrument TMS320C31
processor are the subject of future follow-up research work.
APPENDIX
MACHINE PARAMETERS
Rs = 11.1; Rr = 2.2605
Ls = 0.7329H; Lr = 0.7329H
Lm = 0.71469H; P = 4
REFERENCES
[1]

[2]

0.5
0

0.2

0.4

0.6

0.8

15

[3]

10
5
0
-5

0.2

0.4
0.6
Time (sec )

0.8

[4]

(b)

Actual torque(p.u.)

Actual speed(p.u.)

E
stim
atedtorque(p.u.) Estim
atedspeed(p.u.)

Figure 10. Variation in rotor speed and electromagnetic torque with step rise
in load(a) estimated values, (b) actual values

[5]

1
0 .5

[6]

0
0

0 .5

[7]

-20
0 .5

1 .5

Time (s)

(a)

[8]

0 .5

0
0

0 .5

0 .5

1 .5

1 .5

20

[9]

-2 0

Time (s)

(b)
Figure 11. Variation in rotor speed and electromagnetic torque in low-speed
region (a) estimated values, (b) actual values

Vignans Lara Institute of Technology and Science

Electron., vol. 49, no. 1, pp. 87-95, Feb. 2002.

L. Harnefors, M. Jansson, R. Ottersten, and K. Pietilainen,
Unified sensorless vector control of synchronous and
induction motors, IEEE pp. 153-160, Feb. 2003.
M. Comanescu and L. Xu, An improved flux observer based
on PLL frequency estimator for sensorless vector control of
induction motors, IEEE Trans. Ind. Electron., vol. 53, no. 1,
pp. 50-56, Feb. 2006.
Radu Bojoi, Paolo Guglielmi and Gian-Mario Pellegrino,
Sensorless direct field-oriented control of three-phase
induction motor drives for low-cost applications, IEEE
Trans. Ind. Appl., vol. 44, no. 2, pp. 475-481, Mar. 2008.
I. Boldea and S. A. Nasar: Taylor& Francis, 2006.

1 .5

20

B. K. Bose, Power Electronics and Motor Drives, Delhi,

India, Pearson Education, Inc., 2003.
M. Rodic and K. Jezernik, Speed-sensorless sliding-mode
torque control of induction motor, IEEE Trans. Ind.

[10]
[11]

S. Maiti, C. Chakraborty, Y. Hori, and Minh. C. Ta, Model

reference adaptive controller-based rotor resistance and speed
estimation techniques for vector controlled induction motor
drive utilizing reactive power, IEEE. Trans. Ind. Electron.
vol. 55, no. 2, pp. 594-601, Feb. 2008.
B. Saritha and P. A. Janakiraman, Sinusoidal three-phase
current reconstruction and control using a dc-link current
sensor and a curve-fitting observer, IEEE Trans. Ind.
Electron., vol. 54, no. 5, pp. 2657-2662, Oct. 2007.
H. Kim and T. M. Jahns, Current control for AC motor
drives using a single dc-link current sensor and measurement
voltage vectors, IEEE Trans. Ind. Appl., vol. 42, no. 6, pp.
1539-1546, Nov./Dec. 2006.
P. Vas, Sensorless Vector and Direct Torque Control,
Oxford, U.K. Oxford Science, 1998.
J. Zhao, B. K. Bose, Neural-network-based waveform
processing and delayless AC drives, IEEE Trans. Ind.
Electron., vol. 51, no. 5, pp.981-991, Oct. 2004.

Page 64