2018 International Conference On Advances in Communication and Computing Technology (ICACCT)

Amrutvahini College of Engineering, Sangamner, Ahmednagar, India. Feb 8-9, 2018

Uncertainty and Disturbance Estimator Based

Control Methodology for Speed Control of PMSM
Drives
Dipti D. Bhawarkar, A. A. Godbole, A. A. Apte,
Department of Electrical Department of Electrical Department of Electrical
Engineering Engineering Engineering
All India Shri Shivaji Memorial All India Shri Shivaji Memorial All India Shri Shivaji Memorial
society society society
College of Engineering College of Engineering College of Engineering
Pune-411001 Pune-411001 Pune-411001
Email:dipti.bhawarkar@gmail.com *Email:aagodbole@aissmscoe.com Email:aapte@aissmscoe.com

Abstract—In this paper, speed control strategy for Permanent are given in literature [2],[3]. For speed control of PMSM,
magnet synchronous motor (PMSM) drive based on uncertainty linear feedback controller is developed in [2]. In [3], a simple
and disturbance estimator (UDE) is proposed. Unlike traditional adaptive disturbance observer has been proposed for the
PI controllers, the control strategy designed is robust to internal control of PMSM drive. Sliding mode control is used instead
parameters variations and external torque disturbances. The
of PI in [4] for speed control.
controller design is insensitive to different uncertainties and
tuning of controller parameters is eliminated. MATLAB In 2004, Zong and Rees proposed a robust control
simulation is presented to demonstrate the tracking and technique known as Uncertainty Disturbance and Estimator
disturbance rejection capability of UDE based speed control [5]. It is an elegant robust control strategy incorporating a first
strategy of PMSM drive. The control yields satisfactory results order filter to estimate uncertainty and disturbances. Notable
for varied types of disturbances. feature of UDE is that it is not affected by modelling
inaccuracies and does not require a priori knowledge of
Keywords—Uncertainty and disturbance estimator (UDE), disturbances.
Permanent magnet synchronous motors (PMSM), speed control. UDE has been successfully applied to diverse system
I. INTRODUCTION like nonlinear state delay system [6],robust input-output
linearization[7], robot manipulator[8],non-affine nonlinear
Research in permanent magnet synchronous motors (PMSM) system[9],voltage control of DC-DC power converter[10]. In
has attracted increasing interest in recent years for industrial this paper, a simple UDE based speed control scheme is
drive applications due to its beneficial features like high proposed.
efficiency, high power density, more precise speed control and The paper is written as follows: The mathematical
quick response [1].PMSM drive makes the best alternative for model of PMSM is presented in section II. The development
some applications compared to induction motor drive. Vector of the UDE-based control scheme is done in section III. The
control of PMSM is one of the widely used method in drive performance of UDE based controller is demonstrated through
application for control of speed. This method involves design simulation in section IV. Conclusions are presented in section
of three PI controllers. Out of the three, two PI controllers are V.
designed for electrical system (current loop) which are known
as inner loops and one is used for mechanical system (speed II. MATHEMATICAL MODEL OF PMSM
loop) which is the outer loop as explained in R. krishnan [1].
The methodology commonly used is constant delta control Vector control is best solution for speed control of inverter fed
method. In the constant delta method, reference current AC drives [1]. The main aim of vector control is to control the
trajectory id* is made ‘0’ and ݅‫ כݍ‬is obtained from the speed torque and magnetic field by controlling the ‘d’ and ‘q’
controller loop. In literature, it is refered to as ݅݀‫= כ‬0 strategy. components of the stator currents or relative fluxes.
In a practical systems, the parameters vary and also With the information of the stator currents and the
external disturbance may act on it. If PI controller is designed rotor angle, vector control technique can control the motor
in such situation, then tuning of PI controller gain is needed to torque and the flux in a very effective way. The main
get satisfactory performance. As the disturbance varies, advantages of this technique are fast response and little torque
retuning is needed. This difficulty is overcome when UDE ripple[1]. However, the knowledge of operating conditions
controller is employed. A single controller gives good results and machine parameters is required in vector control. The
for varied types of disturbances. To overcome the parameter voltage equation of PMSM in d-q reference frame is given by
tuning problem, various other approaches and control methods

978-1-5386-0926-2/18/$31.00 ©2018 IEEE 415

Authorized licensed use limited to: MAULANA AZAD NATIONAL INSTITUTE OF TECHNOLOGY. Downloaded on February 25,2021 at 07:44:33 UTC from IEEE Xplore. Restrictions apply.
 ݀݅݀ III. UDE WITH VECTOR CONTROL OF PMSM
‫ ݀ݒ‬ൌ ܴ‫ ݀݅ ݏ‬൅ ‫݀ܮ‬ െ ‫ݍ݅ ݎ߱ ݀ܮ‬
݀‫ݐ‬
݀݅‫ݍ‬
‫ ݍݒ‬ൌ ܴ‫ ݍ݅ ݏ‬൅ ‫ݍܮ‬
݀‫ݐ‬
൅ ‫ ݀݅ ݎ߱ ݍܮ‬൅ ߣ݂ܽ ߱‫( ݎ‬1) All practical systems are affected by disturbances and
Where ݅݀ ,݅‫ = ݍ‬direct and quadrature axis current ,respectively. uncertainties. Disturbances can be categorized as external
‫ ݀ܮ‬ǡ ‫ ݍܮ‬ൌ Stator direct axis and quadrature axis disturbances or internal parameter variations. UDE based
control law can effectively tackle all such uncertainties. UDE
inductances, and ‫ ݀ܮ‬ൌ ‫ ݍܮ‬ൌ ‫ܮ‬.
technique is successfully applied for robust control of electric
ܴ‫ ݏ‬ൌStator resistance. motor drive[11], control of Unmanned Aerial vehicles (UAVs)
ߣ݂ܽ ൌPermanent magnet rotor flux linkages. [12], trajectory tracking control of piezoelectric stages [13]
‫ ݍݒ‬ǡ ‫ ݀ݒ‬ൌ Voltage applied to quadrature and direct axis. and robust control of single axis gimbal platform for micro air
߱‫ ݎ‬ൌelectrical angular speed of motor. vehicles (MAV) [14].The speed loop model of PMSM is
considered which includes the external load torque disturbance
The electromagnetic torque of the machine can be expressed, term. The disturbance is estimated by UDE and speed tracking
in the ‘d-q’ reference frame, as follows: controller is designed which has disturbance estimate
݀߱݉ embedded in it. The speed loop (3) can be rewritten as
ܶ݁ ൌ ‫ܬ‬ ൅ ‫ ݉߱ܤ‬൅ ܶ‫ ܮ‬
݀‫ݐ‬ follows:

͵ܲ ݀߱ ‫ݎ‬ ͵ ܲ ‫ܤ‬ ͳ
ܶ݁ ൌ ߣ݂ܽ ݅‫ ݍ‬ሺʹሻ  ൌ ሺ ሻʹ ݅‫ ݂ܽߣ ݍ‬െ ߱‫ ݎ‬െ ܶ‫(  ܮ‬5)
ʹʹ ݀‫ݐ‬ ʹ‫ʹ ܬ‬ ‫ܬ‬ ‫ܬ‬
Where J = moment of inertia of motor,
B = frictional coefficient, Selecting state as ‫ ݔ‬ൌ ߱‫ ݎ‬, u=݅‫ݍ‬
P = number of poles,
߱݉ ൌMechanical angular speed of motor ‫ݐܭ‬ ‫ܤ‬
ܲ ‫ݔ‬ሶ ൌ ‫ ݑ‬െ ‫ ݔ‬൅ ݀(6)
‫ܬ‬ ‫ܬ‬
߱‫ ݎ‬ൌ ߱݉ 
ʹ where ݀ ൌ െ ܶ‫ ܮ‬.
All equations are communicated in electrical speed. ‫ʹכ ܬ‬

ܶ݁ represents the electromagnetic torque developed by motor

and ܶ‫ ܮ‬denotes the applied load torque. Equation (6) is written as
Equation (1) and (2) in state space form are given by
‫ݔ‬ሶ ൌ ܽͳ ‫ ݔ‬൅ ܾͲ ‫ ݑ‬൅ ݀(7)
‫ܤ‬ ‫ܭ‬
݀݅‫ݍ‬ ‫ݏܴ ݍݒ‬ ߣ݂ܽ withܽͳ ൌ െ and ܾͲ ൌ ‫ݐ‬
‫ܬ‬ ‫ܬ‬
ൌ  െ ݅‫ ݍ‬െ ߱‫ ݀݅ ݎ‬െ ߱
݀‫ݐ‬ ‫ܮ‬ ‫ܮ‬ ‫ݎ ܮ‬ The control problem is to design input ‘u’ so that motor tracks
the reference speed trajectory or regulates it at same desired
݀݅݀ ‫ݏܴ ݀ݒ‬ value.
ൌ  െ ݅݀ ൅ ߱‫ݍ݅ ݎ‬
݀‫ݐ‬ ‫ܮ‬ ‫ܮ‬
݀߱ ‫ݎ‬ ‫ݐܭ‬ ‫ܤ‬ ܲ A. Control design
 ൌ ݅‫ ݍ‬െ ߱‫ ݎ‬െ ܶ  (3)
݀‫ݐ‬ ‫ܬ‬ ‫ܬ‬ ‫ܮ ʹכ ܬ‬ PMSM model under the influence of disturbance can be
͵ ܲ
expressed as,
Where‫ ݐܭ‬ൌ ሺ ሻʹ ߣ݂ܽ
ʹ ʹ
‫ݔ‬ሶ ൌ ܽͳ ‫ ݔ‬൅ ܾͲ ‫ ݑ‬൅ ݀(8)
ܶ
Here inputs are ሾ‫ ݍݒ‬ǡ ‫ ݀ݒ‬ሿ , state of system isൣ݅‫ ݍ‬ǡ ݅݀ ǡ ߱‫ ݎ‬൧ and load Here ܽͳ and ܾͲ depend on plant dynamics and d is the external
torque is considered as disturbance therefore, disturbance.

݀ ൌ െ ܶ‫ ܮ‬.
‫ʹכ ܬ‬ As per the UDE algorithm [5], a continuous signal can be
If ߜ(torque angle) is kept atͻͲͲ , the direct axis current approximated and estimated when it is passed through a low
component ݅݀ becomes zero and quadrature axis current pass filter of suitable bandwidth. The relation between
component ݅‫ ݍ‬becomes the stator phasor current [1]. This disturbance d and its estimate݀መ is given as
current is also termed as the torque component of current.
From (1) & (2), when ݅݀ = 0, torque is given by: ݀መ ൌ ‫ ݂ܩ‬ሺ‫ݏ‬ሻ݀ (9)
͵ܲ
ܶ݁ ൌ ߣ݂ܽ ݅‫ ݍ‬ሺܰ െ ݉ሻ (4) Here ‫ ݂ܩ‬ሺ‫ݏ‬ሻ ൌ ;
ͳ
(10)
ʹʹ ͳ൅‫߬ݏ‬
where ߬ is the bandwidth of the filter.
From (4), it is observed that the electromagnetic torque From (8) and (10),
depends only on ݅‫ ݍ‬and its value is maximum when ݅݀ is
made‘0’. Therefore constantߜ method is extensively used in ݀መ ൌ ‫ ݂ܩ‬ሺ‫ݏ‬ሻሺ‫ݔ‬ሶ െ ܽͳ ‫ ݔ‬െ ܾͲ ‫ݑ‬ሻ (11)
vector control of AC drives.

416

Authorized licensed use limited to: MAULANA AZAD NATIONAL INSTITUTE OF TECHNOLOGY. Downloaded on February 25,2021 at 07:44:33 UTC from IEEE Xplore. Restrictions apply.
 The new control law that ensure speed tracking in the presence (‫) ݍܮ‬
of unknown load torque is obtained as, 5 Stator Resistance (R) 1.4 Ω
ͳ
‫ ݑ‬ൌ ൫െܽͳ ‫ ݔ‬െ ‫ܭ‬ሺ‫ ݔ‬െ ‫ כ ݔ‬ሻ ൅ ‫כ ݔ‬ሶ െ ݀መ ൯(12) 6 Pole (P) 6
ܾͲ
Where K > 0. 7 Rated Torque (ܶ‫) ܮ‬ 2.43 N-m
Substituting control law u in (7) 8 Inertia (J) 0.0012
‫ݔ‬ሶ ൌ െ‫ܭ‬ሺ‫ ݔ‬െ ‫ כ ݔ‬ሻ ൅ ‫כ ݔ‬ሶ ൅ ሺ† െ ݀መ ሻ 9 Frictional constant (B) 0.01
This gives 10 Flux Linkage (ߣ݂ܽ ) 0.1546 Wb-turns
݁ሶ ൅ ‫ ݁ܭ‬ൌ ݀ሚ (13)
Where ݁ ൌ ‫ ݔ‬െ ‫ כ ݔ‬and ݀ሚ ൌ ݀ െ ݀መ .
Equation (13) indicates that the disturbance estimation error Different types of disturbances are applied to PMSM motor in
drives the speed tracking error dynamics. Term K decides the simulations. However, in all the cases the reference value of
convergence of the state trajectory error. For constant or slow speed is maintained constant at 314 rad/sec and ݅݀‫= כ‬0.
varying disturbances ݀ሚ will be 0 and hence satisfactory speed
tracking can be obtained. The simulation results confirm this. ¾ Case I: Load torque of 2.43N-m is applied at 0.2sec.
¾ Case II: Initially torque is constant at 2.43N-m, sine
As the aim of paper is to regulate the speed of PMSM,‫ כ ݔ‬i.e. varying disturbance is added to it from 0.2sec.
߱‫ כݎ‬is kept constant at certain value and ݅‫ כݍ‬is obtained which ¾ Case III: Torque is gradually increased from 0 to
involves estimated disturbance. 2.43N-m at 0.2 sec.

The new control vector u is applied to vector control of

PMSM. Here the need of PI controller in Speed loop(outer
loop) is eliminated. Fig(1) shows the modified vector control
of PMSM with UDE. The simulation result are shown in
section IV

Fig 2: Actual and reference speed for case I

Fig1: Block diagram of proposed vector control of PMSM with UDE

IV. SIMULATION RESULTS

Fig 3: Disturbance estimation error for case I
In order to demonstrate the ability of proposed UDE based
speed control of PMSM, several simulations are performed
using SIMULINK software. The PMSM parameters used in
simulations are summarized in Table 1.

TABLE 1. MOTOR PARAMETERS

No Parameter Value
1 Inverter DC link Voltage 285 V
2 Reference speed (߱‫) ݎ‬ 314 rad/sec
3 Direct axis Inductance (‫) ݀ܮ‬ 0.0056 H
4 Quadrature axis Inductance 0.0056 H
Fig 4: Designed control law ݅‫ כݍ‬for case I

417

Authorized licensed use limited to: MAULANA AZAD NATIONAL INSTITUTE OF TECHNOLOGY. Downloaded on February 25,2021 at 07:44:33 UTC from IEEE Xplore. Restrictions apply.
 In case I, the system response is obtained when load torque of
2.43N-m is applied at 0.2sec. The actual and the reference
speed is shown in figure(2). When the disturbance is suddenly
applied at 0.2sec, the speed is restored to the set value of 314
rad/sec in nearly 0.05 sec. Figure (3) shows disturbance
estimation error (dtilda) between d and݀መ . Figure (4) shows
designed control law ݅‫ כݍ‬.

Fig 8: Actual and reference speed using PI controller for case II Kp = 1.5 and
Ki = 10.5

The results of the PMSM motor when PI controller is

employed are presented in fig (7) and fig(8). The figures
shows that tuning of PI control parameters namely Kp and Ki
is needed as the disturbance varies for getting satisfactory
Fig 5: Actual and reference speed for case II results.

The reference to the actual speed for the case of sinusoidally V. CONCLUSION
disturbance applied at 0.2 sec is shown in figure (5).
In this paper, uncertainty and disturbance estimator (UDE)
based control law has been successfully employed in the speed
control of PMSM against various types of load disturbances.
A single controller yields satisfactory results for different
types of disturbances. The control law thus derived has been
validated using simulation on MATLAB for meeting tracking
requirement. The order of the filter in UDE must be increased
if the uncertainties are fast varying.

REFERENCES

[1] R. Krishnan, Electric Motor Drives: Modeling, Analysis, and Control,

Fig 6: Actual and reference speed for case III Upper Saddle River, NJ, Unite States: Prentice-Hall, 2001.
[2] Aishwarya A Apte ,Vrunda A Joshi, Rahee A Walambe ,Ashwini A
Godbole ,“Speed Control of PMSM Using Disturbance Observer”, 4th
Figure (6) presents the speed tracking results for a slow IFAC Conference on Advances in Control and Optimization of
varying disturbance. Dynamical Systems ACODS, Volume 49 , Issue 1, 2016, pp. 308-313, 1-
5 February 2016.
[3] Mohamed, Y., “Design and implementation of arobust current-control
All the above simulations show that the UDE based control
scheme for a pmsm vector drive with a simple adaptive disturbance
law is robust to different types of disturbances. However if the observer”, IEEE Trans Ind .Electronics, 54(4), pp.1981-1988, 2007.
disturbance are fast varying then the order of the filter needs to [4] S.Li, M.Zhou, and X.Yu ,”Design and implementationof terminal
be increased for getting satisfactory results. slidingmode control method for pmsm speed regulation system”,IEEE
Trans Ind. Informatics, 9(4), pp.1879-1891, 2013.
[5] Q.-C. Zhong and D. Rees, “Control of uncertain LTI systems based on
an uncertainty and disturbance estimator,” ASME J. Dyn. Syst., Meas.,
Control, vol. 126, pp. 905–910, 2004.
[6] Alon Kuperman and Qing-Chang Zhong, “Robust UDE-Based Control
of Uncertain Nonlinear State-Delay Systems”, Dept. of Electrical
Engineering and Electronics, University of Liverpool, pp .928-
933,2009.
[7] S. Talole and S. Phadke, “Robust input-output linearisation using
uncertainty and disturbance estimation,” International Journal of
Control, vol. 82, no.10,pp. 1794–1803, 2010.
[8] Jaywant P. Kolhe, Md Shaheed, T. S. Chandar and S. E. Talole, “Robust
control of robot manipulators based on uncertainty and disturbance
estimation”, International Journal of Robust and Nonlinear Control Int.
J. Robust. Nonlinear Control, 23: pp.104–122, 2013.
[9] B. Ren, Q.-C. Zhong, and J. Chen, “Robust control for a class of non-
Fig 7: Actual and reference speed using PI controller for case I affine nonlinear systems based on the uncertainty and disturbance
Kp = 3 and Ki = 16

418

Authorized licensed use limited to: MAULANA AZAD NATIONAL INSTITUTE OF TECHNOLOGY. Downloaded on February 25,2021 at 07:44:33 UTC from IEEE Xplore. Restrictions apply.
 estimator,” IEEE Trans. Ind. Electron., vol. 62, no. 9, pp. 5881–5888, [12] Anish Ashok, Deepika Kumari, Kavya Satish, Swati Eswar and T. S.
Sept. 2015. Chandar, “Robust Control of UAVs using Uncertainty and Disturbance
[10] A. Kuperman, “UDE-based robust voltage control of DC-DC power Estimation”, International Conference on Industrial Instrumentation and
converters”, In 5th International Congress on Ultra Modern Control (ICIC) College of Engineering Pune,pp.434-438, 2015.
Telecommunications and Control Systems and Workshops (ICUMT), pp. [13] Jinhao Chen, Beibei Ren and Qing-Chang Zhong,“UDE-Based
140–145, Sept. 2013. Trajectory Tracking Control of Piezoelectric Stages”, IEEE
[11] Rohan Vimal Raj, Venkata Sreecharan Poluru, T S Chandar, “Robust Transactions On Industrial Electronics, accepted for publication,2016.
control of Electric motor drive based on UDE”, Proceedings of the 9th [14] Akshata S Kori, C. M. Ananda and T. S. Chandar, “Robust Control of
IFAC symposium on Control of Power and Energy Systems, Indian Single Axis Gimbal Platform for Micro Air Vehicles Based on
Institute of Technology, 2015. Uncertainty and Disturbance Estimation”, 7th International Conference
on Mechanical and Aerospace Engineering, pp.480-486, 2016.

419

Authorized licensed use limited to: MAULANA AZAD NATIONAL INSTITUTE OF TECHNOLOGY. Downloaded on February 25,2021 at 07:44:33 UTC from IEEE Xplore. Restrictions apply.