International Journal on Recent and Innovation Trends in Computing and Communication ISSN: 2321-8169

Volume: 2 Issue: 6 1554 – 1559

_______________________________________________________________________________________________
Model Reference Adaptive System (MRAS) Based Speed Sensorless Vector
Control of Induction Motor Drive
P N H Phanindra Kumar1, D M Deshpande2, Manisha Dubey3
1
M.Tech Scholar in Electrical Department, MANIT, Bhopal, India
2,3
Professor in Electrical Department, MANIT, Bhopal, India
1
hemanthrvr@gmail.com
2
dinesh.deshpande1949@gmail.com
3
md_manit@yahoo.com

Abstract - This paper presents a Model Reference Adaptive System (MRAS) based speed sensorless estimation of vector controlled Induction
Motor Drive. MRAS based techniques are one of the best methods to estimate the rotor speed due to its performance and straightforward
stability approach. Depending on the type of tuning signal driving the adaptation mechanism, MRAS estimators are classified into rotor flux
based MRAS, back e.m.f based MRAS, reactive power based MRAS and artificial neural network based MRAS. In this paper, the performance
of the rotor flux based MRAS for estimating the rotor speed was studied. Overview on the IM mathematical model is briefly summarized to
establish a physical basis for the sensorless scheme used. Further, the theoretical basis of indirect field oriented vector control is explained in
detail and it is implemented in MATLAB/SIMULINK.

Keywords - Model Reference Adaptive System (MRAS), Sensorless Control, Vector Control, Indirect Field Oriented Control.
__________________________________________________*****_________________________________________________

I. INTRODUCTION The rotor flux FOC scheme is based on the frame

Induction motors are rugged and inexpensive machines, transformation of all quantities to a rotating frame fixed to the
therefore much attention is taken while implementing the drive rotor flux. In this rotating rotor flux frame, all quantities
system for various applications with different control rotating at synchronous speed will appear as DC quantities
requirements [1]. An induction machine has many advantages, [12]. If the flux is aligned to the d axis of this reference frame,
especially cage rotor induction machine, when compared with then the d component of the stator current represent the flux
DC machine [2]. However, an induction machine requires and q component of the stator current represent torque
more complex control schemes than DC motors because of its component. This means that utilizing FOC, the control of IM
highly nonlinear and coupled dynamic structure [3]. is transformed to a simple control scheme similar to the DC
Conventional open-loop control of variable frequency motor control where the torque and flux components are
induction motor drives may provide a satisfactory solution decoupled.
under limited conditions. However, these methods are The way the rotor flux position is obtained determines the
unsatisfactory when high performance dynamic operation is type of FOC as either direct FOC or indirect FOC. In indirect
required. [4]. Therefore, highly developed control schemes are FOC, the flux position is obtained by adding the slip position
needed to improve the performance of the induction motor to the measured rotor position, where as in direct FOC it is
drive comparable with DC motors [5]. Recent evolutions in calculated (or can also be measured) based on the terminal
the area of control systems, power electronics, powerful and variables and rotor speed [13].
cheap microcontrollers, DC motors are replaced by an Another emerging area of research involves the sensorless
induction motor in the industry. control of drive system which is different from conventional
Variable speed IM drives use mainly PWM techniques to methods because it doesn‘t require speed or position sensors.
generate a polyphase supply of a given frequency [6, 7]. Most Removing these sensors gives a number of advantages such as
of these induction motor drives are based on keeping a increased reliability, lower production costs, reduced size and
constant voltage/frequency (V/f) ratio in order to maintain a removal of excess cabling. Sensorless drives require less
constant flux in the machine. Although the control of V/f maintenance and are also more suitable for harsh inaccessible
drives is relatively simple, the torque and flux dynamic environments [14].
performance is extremely poor [7]. As a consequence, a great The study of speed sensorless control of the IM has
quantity of industrial applications that require good torque, undergone through maturing years when new techniques came
speed or position control still use DC machines [8, 9]. into introduction to improve the previous techniques. The
Over the past few decades a great deal of work has been motivation is to find one method that can cater the entire
done into techniques such as Field Oriented Control, Direct problem related to speed sensorless IM. Among them, MRAS
Torque Control and Space Vector Pulse Width Modulation based techniques have been proven to be one of the best
[10]. Field oriented control (FOC) or vector control (VC) was methods being proposed by the researchers due to its good
introduced by Hasse and Blaschke from Germany, in 1969 and high performance ability and straight-forward stability
1971 respectively [6]. On the contrary to the scalar control, the approach. The method was first proposed in [15] followed by
development of FOC control scheme is based on dynamic [16] which consists of a reference model (RM), an adjustable
model of the IM where the voltages, currents and fluxes are model (AM) and an adaptation mechanism. RM is independent
expressed in space vector forms [11]. of the rotor speed whereas AM requires the rotor speed
The representation of the motor's quantities using space information. Through Landau‘s idea of comparing the outputs
vectors valid under both steady state and transient conditions of RM and of AM, the error between the two models can be
hence with FOC, excellent transient response can be achieved. minimized using the adaptation mechanism [17].

1554
IJRITCC | June 2014, Available @ http://www.ijritcc.org
_______________________________________________________________________________________
International Journal on Recent and Innovation Trends in Computing and Communication ISSN: 2321-8169
Volume: 2 Issue: 6 1554 – 1559
_______________________________________________________________________________________________
The paper has been organized as follows: section II briefly III. INDIRECT FIELD ORIENTED CONTROL
explains the mathematical modeling of induction motor,
A. Principle of Vector Control
section III demonstrate the indirect Field Oriented Control of
To explain principle of vector control, an assumption is
the induction motor drive, section IV describes the Rotor Flux
made that the position of rotor flux linkages phasor, Ψr is
based MRAS Speed Observer, and the results of implemented
known. Ψr is at θf from a stationary reference, θf is referred to
Matlab models are shown in section V.
as field angle hereafter, and the three stator currents can be
II. MATHEMATICAL MODELLING transformed into q and d axes currents in the synchronous
reference frame by using the park‘s transformation given
The dynamic behavior of an induction motor is complex
below [7].
due to the coupling effect between the stator and rotor phases. � �
Fig. 1 shows the dynamic d-q equivalent circuits of an � � − � +
induction machine. = � � � � −
�
� � +
� (5)
The dynamic model of induction motor represented in . . .
terms of voltages and currents can be given in matrix form as Stator current phasor, Is can be derived as
[6]: 2
� � + Þ � Þ � � = + 2 (6)
� −� � + Þ −� Þ �
= .
� Þ � −� � + Þ � −� � And stator phase angle is � = tan−1 (7)
� − � −� Þ − � −� � + Þ �
e e
These equations are expressed in general reference frame where I qs and I ds are the q and d axes currents in the
denoted by the superscript ‗c‘ and ‗Þ‘ represents the derivative synchronous reference frames that are obtained by projecting
operator, d/dt. The dynamic model of the induction motor can the stator current phasor on q and d axes, respectively. That
also be rearranged with the stator and rotor flux linkages as the the current phasor magnitude remains same regardless of the

The current phasor Is produces the rotor flux r and torque

state variables [6]. reference frame chosen to view it is evident from Fig. 2.
−
−�
�′ �′
� �

to be in phase with r. Therefore resolving stator current

�
− Te. The component of current producing rotor flux phasor has
� �′ �′ � �

phasor along r reveals the component if

= − �
+
�
(2)
� − � −�
�′ �′
� −
� is the field
� −�
�′ �′ producing component and it is the torque producing
Where �′ = � � , �′ = � � , = , = , � = − component perpendicular to it. By writing rotor flux linkages
The speed ωr in the above equations is related to the torque by and torque in terms of these components
the following mechanical dynamic equation, � ∝� (8)
� ∝�� ∝� � (9)
� −� = + �� (3)
� For vector control operation of the induction motor, the
� = � � −� � (4) arbitrary reference frame must be aligned along the rotor flux

that the position of the rotor flux linkage space phasor f, be
iqs iqr linkage space phasor at every instant. It is therefore essential
L1s L1r
+ - - +

instantaneous rotor flux phasor position, f can be written as

Rs Rr
accurately known at every instant. From the Fig. 2, the
ωcλds (ωc-ωr)λdr
� =� +�
Where r is the rotor position and sl is the slip angle. In
(10)
λqs Lm λqr
Vqs Vqr
terms of the speeds and time, the field angle is written as
� = � +� = � (11)
This knowledge of rotor flux linkage space phasor
(a) q-axis circuit position can be acquired either by measuring the flux directly
ids idr
or by estimating the flux from terminal variables i.e. by
L1s L1r indirect means. This leads to two possible control techniques
+ - - +
of induction motor namely: Direct field oriented control and
Rs Rr Indirect field oriented control.
ωcλqs (ωc-ωr)λqr
Vs
λds Lm λdr
Vqse Is
Vds Vdr
Iqse=iT
Ф θ
T
θf θs Vdse θsl Idse=if
θr
(b) d-axis circuit
Fig. 1 Dynamic d-q equivalent circuits of an IM Rotor Reference
Frame
Stator Reference
Frame

Fig. 2 Phase diagram of vector control

1555
IJRITCC | June 2014, Available @ http://www.ijritcc.org
_______________________________________________________________________________________
International Journal on Recent and Innovation Trends in Computing and Communication ISSN: 2321-8169
Volume: 2 Issue: 6 1554 – 1559
_______________________________________________________________________________________________
B. Indirect Field Oriented Control �
+� = � (16)
�
In an Indirect Field Oriented Control (IFOC) a flux � = (17)
1+� þ
estimator is used to estimate the required flux linkage space defining an equivalent rotor magnetizing current, i mr as
phasor magnitude and angular position θf as shown in Fig. 3. �
The shaft position is usually needed for estimating flux linkage � = (18)
space phasor position. This gives a more adaptable drive � = � + � þ� (19)
system, but this method would generally result in a more From the above equation, the equivalent rotor
complex control system [18]. Since it is generally desirable to magnetizing current imr is obtained by passing the direct axis

pass filter having time constant r .The position of the rotor

have a scheme which is applicable for all induction motors, the component of the stator current ids through a first order low

flux linkage space phasor  is obtained by integrating ωe

indirect field oriented has emerged as the more popular
method. In the indirect field orientated control method the flux
linkage space phasor is estimated from the motor model and which is given by the sum of the electrical rotor speed ω r and
which is sensitive to variations in machine parameter like the the slip speed ωsl.
stator time constant or rotor time constant. In the rotor flux If rotor flux � = constant, which is usually the case, then from

the rotor time constant r, of the motor. In stator flux oriented
oriented control, the indirect rotor flux estimator is sensitive to eqn. 16
� = � (20)
control, the indirect stator flux estimator is sensitive to the In other words, the rotor flux is directly proportional to current
stator time constant of the motor. In the air gap flux oriented ids in steady state.
control, the indirect air gap flux estimator is sensitive to both �
Where ωsl is given by � = (21)
the stator and the rotor time constants. Therefore, if the value � �
�
of the motor parameter varies, the desired decoupled of the From eqn. 21, � = (22)
� �
flux and the torque components of the stator current space
phasor is not achieved and this leads to reduce the
performance of the dynamic behavior of the drive system. In iqse
this paper, only rotor field orientation control is considered. 1/τr
Torque Vqso
ωsl ωe θf
Iqs,ref
Controller
PI
Vqs Vqs,ref
de-qe
÷ 1/þ
PWM idse
Ids,ref
Iqs
Vds,ref VSI
IM
1/(1-þτr) ωr
Vds
PI
abc
Ids Flux
Controller Vdso Fig. 4 Rotor Flux Estimator
IV. MODEL REFERENCE ADAPTIVE SYSTEMS (MRAS)
e e
d -q MRAS is one of the most popular adaptive control
θf Flux method used in motor control applications for tracking and
Estimator observing system parameters and states. There are different
abc model reference adaptive control methods such as series
model, parallel model, direct model and indirect model etc. are
Fig. 3 Block diagram for Indirect Field Oriented Control available [14] [15].
C. Rotor Flux Linkage Estimator

For indirect field oriented control, it is essential to

estimate the flux linkage space phasor position. Therefore it is
necessary to model the rotor flux linkages for rotor field
orientation technique.
From the induction motor modeling Fig. 1, by eliminating
the rotor currents
�
+ � − � −� � =0 (12)
�
+ � − � −� � =0 (13)
where ωsl =ωe-ωr
For decoupling control, it is desirable that Fig. 5 Generalized model reference adaptive system
MRAS estimators consist of reference model and
� =0 (14)
� adjustable model as shown in Fig. 5.The speed-adaptation
that is, =0 (15) laws adjusts the estimate speed based on the outputs of
So that the total rotor flux is � directed on the de axis. reference and adjustable models. MRAS used in this model
Substituting the above conditions in the eqns. 12 & 13, we get compares both the outputs of a reference and adaptive models,
1556
IJRITCC | June 2014, Available @ http://www.ijritcc.org
_______________________________________________________________________________________
International Journal on Recent and Innovation Trends in Computing and Communication ISSN: 2321-8169
Volume: 2 Issue: 6 1554 – 1559
_______________________________________________________________________________________________
and processes the error between these two based on the
appropriate adaptive laws that do not disturb the stability of
the applied system [16]. In the MRAS technique, the desired
process response to a command signal is specified by means of
a parametrically defined reference model. An adaptation
mechanism keeps track of the process output and the model
output and calculates a suitable parameter setting such that
difference between these outputs tends to zero [17].
An important issue in MRAS is the design of adaptive
laws. The first examples of adaptive law designs made use of
sensitivity models, and later the stability theory of Lyapunov,
and Popov`s hyper stability theory, served as standard design
methods, yielding a guaranteed stable adaptive system. In this
paper, the detail of the adaptation mechanism design is not
elaborated since it has been clearly discussed in [16] and [17]. Fig. 6 MRAS based on rotor flux estimation
A. Rotor Flux MRAS Estimator In designing the adaptation algorithm for the MRAS, it is
important to take account of the overall stability of the system
In this MRAS scheme the rotor flux linkage (� ) is used
and to ensure that the estimated speed will converge to the
as speed tuning signal. The motor voltages and currents are
desired value with satisfactory dynamic characteristics. Using
measured in a stationary frame of reference. It is also
Popov‘s criteria for hyper stability for a globally
convenient to express these equations in that stationary frame.
asymptotically stable system, we can derive the following
The speed can be calculated by the model referencing adaptive
relation for speed estimation:
system (MRAS), where the output of the reference model is �
compared with the output of an adjustable model until errors � =� + (27)
between the two models vanish to zero. A block diagram for where � = − = � � − � � (28)
speed estimation by this MRAS technique is shown in the Fig. In steady state ξ = 0, Balancing the fluxes; in other
6. Consider the voltage model‘s stator side equations (23) and words, � = � and � = � . The MRAS in the Fig. 6
(24) which are defined as a reference model. can be interpreted as a vector PLL in which the output flux
The model receives the machine stator voltage and current vector from the reference model is the reference vector and the
signals and calculates the rotor flux vector signals, as adjustable model is a vector phase shifter controlled by � .
indicated. From the stator voltage equations in the stationary
frame, the Reference model equations can be obtained as: V. SIMULATION RESULTS
Reference model equations The following induction motor parameters are chosen for
� = � − ( +� )� (23) the simulation studies:
� = � − ( +� )� (24) = 0.855Ω, = 1.15Ω, = 0.10432 , = 0.10432
= 0.1004 , = 50 �, = 0.06 − 2 , � = 6.
Where Ψ is flux linkage, , , are inductances, is Fig. 7 shows the complete Simulink model of induction
2
resistance and � = 1 − is motor leakage coefficient. The motor, Fig. 8 shows the Simulink Model of Rotor Flux based
MRAS Speed estimator and Fig. 9 shows the complete
subscripts r and s denotes the rotor and stator values,
Simulink Model of sensorless indirect field oriented control of
respectively, referred to the stator and subscripts d and q
induction motor with Rotor Flux MRAS speed Observer.
denote d-axis and q-axis components in the stationary
reference frame
The current model flux equations (25) and (26) are defined as
an adaptive model in the Fig. 6. This model can calculate
fluxes form the input stator currents only if the speed signal �
is known. With the correct speed signal, ideally, the fluxes
calculated from the reference model and those calculated from
the adaptive model will match, that is, � = � and
� = � , where � and � are the adaptive model outputs.
An adaptation algorithm with P-I control, as indicated, can be
used to tune the speed � so that the error ξ =0.
Adaptive model equations
1
� = � −� � − � (25)
1
� = � +� � − � (26)
Fig. 7 Simulink Model of Induction Motor
where � is rotor electrical speed and = is the rotor time
constant.

1557
IJRITCC | June 2014, Available @ http://www.ijritcc.org
_______________________________________________________________________________________
International Journal on Recent and Innovation Trends in Computing and Communication ISSN: 2321-8169
Volume: 2 Issue: 6 1554 – 1559
_______________________________________________________________________________________________

Fig. 11 speed error of the motor for changes in speed and load torque

Fig. 8 Simulink Model of Rotor Flux Based MRAS Speed Observer

Fig. 12 Generated torque of the motor with step changes in load

torque

Fig. 9 Simulink Model of Indirect Field Oriented Control of IM Drive

with MRAS Speed Observer
Sensorless vector control simulation of the induction
motor is presented below. The response of the induction motor
is shown in two different cases
1. with step changes in speed reference and load torque
2. with speed reversal (Four quadrant operation) Fig. 13 Phase voltage vao and 3-ph currents of the motor with step
changes in load torque

Fig. 14 Four quadrant Estimated speed and reference speed

Fig. 10 Estimated and reference speed of the motor with step changes
 With the application of the load only torque component i qs
From the above results we can conclude that
in load torque
changes but not flux producing component ids.

1558
IJRITCC | June 2014, Available @ http://www.ijritcc.org
_______________________________________________________________________________________
 International Journal on Recent and Innovation Trends in Computing and Communication ISSN: 2321-8169
Volume: 2 Issue: 6 1554 – 1559
_______________________________________________________________________________________________
 Actual speed of the motor traces the Reference speed [9] R. Saidur, ―A review on electrical motors energy use and
irrespective of the load torque within permissible limits. energy savings,‖ Renewable and Sustainable Energy
There can be a slight dip in speed at the instant of Reviews 2010; 14:877–98.
application of load torque but it should settle to the [10] Y. Oguz and M Dede, ―Speed estimation of vector

 Generated torque (Te) and load torque (T L) waveforms

reference speed after some time. controlled squirrel cage asynchronous motor with
artificial neural networks,‖ Energy Conversion and
follow the same trace. Management 2011; 52:675–86.
[11] M. Hajian, G. R. Arab Markadeh, J Soltani and S.
VI. CONCLUSIONS Hoseinnia, ―Energy optimized sliding- mode control of
In this paper, Indirect Field Oriented Control and sensorless induction motor drives,‖ Energy Conversion
sensorless vector control with MRAS observer technique for and Management 2009; 50:2296–306.
the control of induction machine are presented. First, [12] H. C. Stanley, ―An analysis of the induction motors,‖
generalized dynamic mathematical model of the induction AIEE Trans., vol. 57, pp. 751-755, 1938.
motor is studied. Next, mathematical model of induction [13] J. W. Finch and D. Giaouris, ―Controlled AC electrical
motor developed in synchronous reference frame is simulated drives,‖ IEEE Transactions on Industrial Electronics
and investigated. By using this motor dynamic model an 2008; 55:481–91.
indirect field orientation control is simulated. An adaptive [14] J. Holtz, ―Sensorless control of induction machines: with
state observer, MRAS is tested to observe rotor speed. The or without signal injection,‖ IEEE Transactions on
high performance of this scheme is shown in simulation Industrial Electrons 2006; 53:7–30.
results. Using this observer, Sensorless vector control is [15] S. Tamai, H. Sugimoto and M. Yano, "Speed-sensorless
simulated and dq-axis rotor-stator fluxes, rotor speed were vector control of induction motor with model reference
estimated and found. In Sensorless vector control also proper adaptive system", Conf. Record of the 1985 IEEE-IAS
field orientation is achieved because the value of q axis flux is Annual Meeting, pp. 613-620, 1985.
zero and there is no change in d axis current due to the [16] C. Shauder, "Adaptive speed identification for vector
application of load torque. The MRAS provides the estimation control of induction motor without rotational transducers",
of only one state or one parameter instantaneously. Therefore IEEE Trans. Ind. Application, Vol. 28, No. 5, pp. 1054-
this criterion depends on the requirements of FOC algorithms. 1061, Sept./Oct. 1992.
One may use MRAS not as a state observer but an online [17] Y.P. Landau, "Adaptive Control: The model reference
parameter tuning tool that tunes different state observers. approach", Marcel Dekker, New York, 1979.
[18] O. Burak a d L. M. Tol ert, “i uli k
VII. REFERENCES
implementation of Induction Machine Model-A
[1] R. Saidur, S. Mekhilef, M. B. Ali, A. Safari and H. A. odular approa h , IEEE Tra s Power Ele tro i s,
Mohammed, ―Applications of variable speed drive (VSD)
in electrical motors energy savings,‖ Renewable and
Vol. 39, No. 3, pp 728-734, May 2003.
Sustainable Energy Reviews 2012; 16:543–50.
[2] C. Saravanan, J. Sathiswar and S. Raja, ―Performance of
three phase induction motor using modified stator
winding,‖ International Journal of Computer Applications
2012; 46:1–4.
[3] S. Amjad, S. Neelakrishnan and R. Rudramoorthy,
‖Review of design considerations and technological
challenges for successful development and deployment of
plug-in hybrid electric vehicles,‖ Renewable and
Sustainable Energy Reviews 2010; 14:1104–10.
[4] C. A. Martins and A. S. Carvalho, ―Technological trends
in induction motor electrical drives,‖ In Proceedings of
IEEE Porto Power Tech, vol. 2; 2001.
[5] G. S. Buja and M. P. Kazmier kowski, ―Direct torque
control of PWM inverter-fed AC motors—a survey,‖
IEEE Transactions on Industrial Electronics 2004;
51:744–57.
[6] B. K. Bose, ―Power Electronics and AC Drives‖, Prentice
Hall, 1986
[7] D. W. Novotny and T. A. Lipo, ―Vector Control and
Dynamics of AC Drives‖, Oxford University Press Inc.,
Oxford, New York, 1997.
[8] E. A. Abdelaziz, R. Saidur and S. Mekhilef, ―A review on
energy saving strategies in industrial sector,‖ Renewable
and Sustainable Energy Reviews 2011; 15: 150–68.

1559
IJRITCC | June 2014, Available @ http://www.ijritcc.org
_______________________________________________________________________________________