 Open Access
 Total Downloads : 500
 Authors : Prashant Bhopale, Swati Dhobaley, D. R. Mehta
 Paper ID : IJERTV3IS041946
 Volume & Issue : Volume 03, Issue 04 (April 2014)
 Published (First Online): 29042014
 ISSN (Online) : 22780181
 Publisher Name : IJERT
 License: This work is licensed under a Creative Commons Attribution 4.0 International License
Unscented Kalman Filter Based Observer for Vector Controlled Induction Motor
Prashant Bhopale1, Swati Dhobaley2, D. R. Mehta3
1,2,3 Dept. of Electrical Engg. Veermata Jijabai Technological Institute, Mumbai19., India.
AbstractThe extended kalman filter(EKF) suffers from 1st order approximation error which causes deviation in the mean and covariance of estimation while Unscented kalman Filter(UKF) uses the nonlinearity in the system without linearization to overcome this errors. The nonlinear approach of unscented transform has been used for rotor flux estimation in the presence of white Gaussian noise and results inferred.
KeywordsExtended Kalman Filter; Unscented Kalman Filter; nonlinear;white gaussian noise.

INTRODUCTION
The induction motor drives are widely used in industrial application due to the advantages in terms of robustness and prize, hence the motor control industry has become a strong and aggressive sector. During the last few decades the field of controlled electrical drives has undergone rapid expansion hence better estimation of the control parameter and states of control vectors are necessary. The Vector control induction motor consists of controlling the stator currents represented by a vector. The different methods of speed and flux estimation needs model which is sensitive and based on machine parameters and also require apriori knowledge of electrical and sometimes mechanical characteristics.
Some type of speed estimation is essential for speed control of induction motor. Magnetic saliencies depending speed estimation like rotor slotting [1], variation of leakage reactance [2] or rotor asymmetries [3] which are considered as true measurement and independent of machine parameters. Due to reduced range of frequency they are not widely used. Although machine model dependent estimations which uses open loop speed calculators[4,5],Model reference adaptive system(MRAS)[6,7,8,9] and Extended Kalman Filter[10]. Also deriving machine flux used methods are from machine models are integration of back emf [4,5], flux observer
[22]Julier showing nonlinear estimation gives better performance than linearized strategies like EKF.This paper is organized as follows. The fundamental model of vector control induction machine is described in Section II together with the theoretical basis for FOC strategy. The UKF algorithm is presented in Section III. In section IV MATLAB/Simulink simulation results obtained with the proposed algorithm applied on an ideal model of the machine is provided. In Section VI conclusions are drawn.

SYSTEM DESCRIPTION.

Vector Controlled Induction Motor Drive
First, The block diagram of the proposed system is shown in Fig.1 The continuoustime mathematical model of the induction machine can be described in statespace form, with rotor speed treated as a timevarying parameter,
= , + + () (1)
= 1,2, 4
x is state variable and the input voltage vector u are given as x = [ids iqs dr qr ] where s and r stands for stator and rotor and d and q are vectors of respective fixed stator frame and ()is process noise.

Mathematical Modelling of Vector Controlled Induction Motor Drive
The four first order differential equations resulting from the expansion of eqn. 1 are:
[6,7,8,9, 11], Extended Luenberger Observer [12], Monitoring local saturation effect [13] and extended kalman filterR
s
L
x 1 =
m
Rr L2
m 1
+ x +
r
L2 L
Rr Lm
r
L2 L
x3 +
wr Lm
r
L2 L
x4 +
uds (2)
1
L
[13,14,15,16,17]. But the above mentioned methods which aremodel dependent have considered the linearized model of
x = Rs + Rr L2 x
r

L L2 L 2

wr Lm x
r
L2 L 3
+ Rr Lm x
r
L2 L 4
+ 1 u
L qs
(3)
induction motor, although where 1st order approximation
using EKF can introduce mean and covariance errors.
x = Rr Lm x Rr x w x


Lr 1 Lr 3 r 4
(4)
In practice, the rotor currents or fluxes are not easily available and measurable. Therefore an observer is required to estimate the unknown states and can also provide better estimates of known current states that are contaminated by noise in some circumstances,. In this paper we propose the UKF based observer for flux estimation of vector controlled induction motor which takes into account the nonlinearity of the model using nonlinear approach of unscented transform
x = Rr Lm x

Lr 2

+ wr x3

Rr x
Lr 4
(5)
= = 1
= 1,2,3, 2 (11)
Where,
2( +)
= 2 + (12)
Fig.1 Indirect Vector Control Drive
Where L is linkage inductance i.e.
L
2
L = Ls m (6)
Lr
The states x1 and x2 are the stator dq axis currents, which are usually measurable. The statesx3 and x4 are the dq axis rotor fluxes, which are not easily measurable and require estimation. Sometimes, the measured stator currents are contaminated by noise, and require to be estimated as well. In some circumstances the estimates can be more reliable.
An output or observation equation is required by the state space model and for the practical case in which stator and rotor flux are chosen as measurements:
= (()) (7)
= [ dr qr ] (8)


ALGORITHM OF UNSCENTED KALMAN FILTER
determines the spread of sigma points around and usually set to a small positive value. is a secondary scaling parameter which is usually set to 0 and is a parameter used to incorporate any priory knowledge about the distribution of
. + is the th row or column of the matrix
square root of + and is the weight which is associated with the th point. These sigma points are propagated through the function.
B. Implementation of UKF
For L dimension state vector, 2L+1 sigma points are related as follows,
Step1: Creating Sigma Points
1 = 1 1 + 1 1 1
(13)
Where, = + and = 2 +
Step 2: Propagating Sigma Points or Prediction
, = ,1 , = 1,2, . . ,2 + 1
(14)
Step 3: Calculating Mean & Covariance
=0
Unscented Kalman Filter (UKF) also known as Sigma Point Kalman Filter[22], priory needed to created sigma points which are selected by priory knowledge of previous state and
1 = 2
, (15)
covariance matrices. 2L+1 Sigma points are more useful when
= 2
+
creating sigma points for L dimensional state vector. Following are the steps for UKF algorithm.
A. Sigma Points:
=0
(16)
,
1
,
1
1
The dimensional random variable with mean and covariance is approximated by 2 + 1 weighted points given by,
, = ,
(17)
=0
= 2 ,
(18)
0 =
= Â± + = 1,2,3, 2 (9)
= 2
[_(, ) _]
= , = + 2 + (10)
=0
(19)
,
1
0 +
0 +
= 2
+
(20)
=0
,
,
1
Step 4: Correction

CONCLUSIONS
In this paper, we have implemented nonlinear model of
=
(21)
1
indirect vector control induction motor without linearization, and simulation results are presented. The UKF based flux observer is created and performance of observer is presented
= 1 +
(22)
in the presence of noise. Result shows that the Stator and Rotor flux of Vector Control Induction Motor can be
=
(23)
successfully estimated by observer designed using UKF.

ACKNOWLEDGMENT
Process noise and Measurement noise are
characterized by
= 0 & = 0
The authors would sincerely like to acknowledge the support of Mr. Pratik K. Bajaria, VJTI, Matunga, Mumbai, India, for his valued contributions to this paper.
(24)
= 0 &
APPENDIX
Rating and Parameters of Induction Motor used for
=
(25)
0
simulation
4 Poles,60 Hz,220 V,20 Hp
Rotor Resistance( ) = 1.772 Stator Resistance( )=1.282


SIMULATION AND ESTIMATION
The rotor flux estimation is obtained by applying UKF algorithm. The UKF used the nonlinear model (2)(5) The simulation results for UKF is shown in Fig.2 where 9 sigma points are created by considering 4 state including rotor and stator current along with rotor and stator flux as augmented states which are estimated by observer. Scaling parameters and process and measurement noise of UKF are selected as [23]. Sampling time taken 0.01sec and simulation carried out for 10 sec when motor accelerates from rest to 255 rpm.
Fig.2 Results obtained by UKF observer
Stator Inductance( )=0.282mH Rotor Inductance( )=0.101mH Mutual Inductance( )=0.161mH
VII. REFERENCES

Atkinson,DJ; Acarnley,PP;Finch,JW,Observer for InductionMotor State and ParameterEstimation,IEEE Transaction on Industry Application,1991,Vol.27,No.6,pp.1191127

Jansen, P.L.; Lorenz, R.D., "Transducerless position and velocity estimation in induction and salient AC machines," Industry Applications, IEEE Transactions on , vol.31, no.2, pp.240,247, Mar/Apr 1995
doi: 10.1109/28.370269

Holloday,D,,Fletcher,J.E. and William,B.W., noninvasive rotor position and and speed sensing of Asynchronous Motor,Proceeding of EPE 95,Sevilla,1995 Vol.1,pp. 1.3331.337

Xingyi Xu; Novotny, D.W., "Implementation of direct stator flux orientation control on a versatile DSP based system," Industry Applications, IEEE Transactions on , vol.27, no.4, pp.694,700, Jul/Aug 1991
doi: 10.1109/28.85484

Habetler, T.G.; Profumo, F.; Pastorelli, M.; Tolbert, L.M., "Direct torque control of induction machines using space vector modulation," Industry Applications, IEEE Transactions on , vol.28, no.5, pp.1045,1053, Sep/Oct 1992

Jansen, P.L.; Lorenz, R.D., "Accuracy limitations of velocity and flux estimation in direct field oriented induction machines," Power Electronics and Applications, 1993., Fifth European Conference on , vol., no., pp.312,318 vol.4, 1316 Sep 1993

Kubota, K.; Matsuse, K.; Nakano, T., "DSPbased speed adaptive flux observer of induction motor," Industry Applications, IEEE Transactions on , vol.29, no.2, pp.344,348, Mar/Apr 1993

Tajima, H.; Hori, Y., "Speed sensorless fieldorientation control of the induction machine," Industry Applications, IEEE Transactions on , vol.29, no.1, pp.175,180, Jan/Feb 1993

Yang, G.; Chin, T.H., "Adaptivespeed identification scheme for a vectorcontrolled speed sensorless inverterinduction motor drive," Industry Applications, IEEE Transactions on , vol.29, no.4, pp.820,825, Jul/Aug 1993

Sc Hroedl,M.,Sensorless Control of Induction Motor at Low Speed and Standstill, Proceeding ICEM92,1992,pp.863867

KubotMAtsuse,K.,Speed sensorless fieldorientation control of the induction Motor with Rotor Resistance Adaption,IEEE Industry Application Society Annual Meeting,1993,Vol.1,pp.414418

Du, T.; Brdys, M.A., "Implementation of extended Luenberger observers for joint state and parameter estimation of PWM induction motor drive," Power Electronics and Applications, 1993., Fifth European Conference on , vol., no., pp.439,444 vol.4, 1316 Sep 1993

Atkinson, D.J.; Acarnley, P.P.; Finch, J.W., "Observers for induction motor state and parameter estimation," Industry Applications, IEEE Transactions on , vol.27, no.6, pp.1119,1127, Nov/Dec 1991

Henneberger,G.,Brunscach,B.J. and Klepsch,T.,FieldOriented Control of Synchronous and Asynchronous Drive Without Mechanical Sensors using Kalman Filter, Proceeding of the EPC conference,Firenze,1991,Vol.3,pp.664671

Brunsbach,BJ;Henneberger,G,FieldOriented Control of an Induction Motor Without Mechanical Sensors using a KalmanFilter,Archiv Fur Elektrotechnil 1990, Vol.73, No.5,pp.325335

Kim, Y.R.; SeungKi Sul; Park, M.H., "Speed sensorless vector control of induction motor using extended Kalman filter," Industry Applications, IEEE Transactions on , vol.30, no.5, pp.1225,1233,
Sep/Oct 1994

Loron,L. And Laliberte,G.,Application of the Extended Kalman Filter to Parameter Estimation of Induction Motor,Proceeding of EPE conference,Brighton,1993,Vol.5,pp.8590

B.K.BOSE,Power electronics and AC drives,(PrenticeHall,1986).

A K Chattopadhyay, Advances in vector control of AC motor drivesa review

Finch, J.W.; Atkinson, D.J.; Acarnley, P.P., "Scalar to vector: general principles of modern induction motor control," Power Electronics and VariableSpeed Drives, 1991., Fourth International Conference on , vol., no., pp.364,369, 1719 Jul 1990

Kataoka, T.; Toda, S.; Sato, Y., "Online estimation of induction motor parameters by extended Kalman filter," Power Electronics and pplications, 1993., Fifth European Conference on , vol., no., pp.325,329 vol.4, 1316 Sep 1993

Julier, S.J.; Uhlmann, J.K.; DurrantWhyte, H.F., "A new approach for filtering nonlinear systems," American Control Conference, Proceedings of the 1995 , vol.3, no., pp.1628,1632 vol.3, 2123 Jun 1995

Rigatos, G.; Siano, P., "Sensorless nonlinear control of induction motors using Unscented Kalman Filtering," IECON 2012 – 38th Annual Conference on IEEE Industrial Electronics Society , vol., no., pp.4654,4659, 2528 Oct. 2012

Jafarzadeh, S.; Lascu, C.; Fadali, M.S., "Square Root Unscented Kalman Filters for State Estimation of Induction Motor Drives," Industry Applications, IEEE Transactions on , vol.49, no.1, pp.92,99, Jan.Feb. 2013