Performance of Fuzzy-Logic Based Indirect

Vector Control for Induction Motor Drive

Rabinarayana Parida, M.I.E.E.E, BPUT,Orissa,[email protected]
Prof. K.B. Mohanty, M.I.E.E.E,NIT,Rourkela,[email protected]
Abstract-This paper presents a novel speed control scheme of an compared to the conventional PI, PID, and their adaptive
induction motor (IM) using fuzzy-logic control. The fuzzy-logic versions, the FLC has some advantages such as: 1) it does
controller (FLC) is based on the indirect vector control. The not need any exact system mathematical model; 2) it can
fuzzy-logic speed controller is employed in the outer loop. The handle nonlinearity of arbitrary complexity; and 3) it is based
complete vector control scheme of the IM drive incorporating
on the linguistic rules with an IF-THEN general structure,
the FLC is experimentally implemented using a digital signal
processor board DS-1104 for the laboratory 1-hp squirrel-cage which is the basis of human logic. However, the application
IM. The performances of the proposed FLC-based IM drive are of FLC has faced some disadvantages during hardware and
investigated and compared to those obtained from the software implementation due to its high computational
conventional proportional-integral (PI) controller-based drive burden [7]. The earlier reported works for fuzzy-logic
both theoretically and experimentally at different dynamic applications in motor drives [8]–[11] are mainly theoretical
operating conditions such as sudden change in command speed, and based on either simulation or experimental results at low-
step change in load etc. The comparative experimental results speed operating conditions. This paper investigates the
show that the FLC is more robust and, hence, found to be a successful application of the FLC for normal speed control of
suitable replacement of the conventional PI controller for the
IM drives. The complete vector control scheme of IM
high-performance industrial drive applications.
incorporating the FLC has been successfully implemented in
Index Terms - Digital signal processor, fuzzy-logic real time using digital-signal-processor (DSP) controller-
controller, induction motor, PI controller, real-time board DS1104. The performances of the proposed drive have
implementation, speed control. also been compared with those obtained from the
conventional PI controller both theoretically and
1. INTRODUCTION experimentally. It is found that the proposed FLC is
AC MOTOR drives are used in a multitude of industrial and insensitive to temperature changes, inertia variations, and
process applications requiring high performances. In high- load torque disturbances. This novel FLC could be a suitable
performance drive systems, the motor speed should closely replacement for the conventional PI controller for high-
follow a specified reference trajectory regardless of any load performance drive systems.
disturbances, parameter variations, and model uncertainties. II. DESIGN OF FLC FOR IM
In order to achieve high performance, field-oriented control For the proposed FLC, the speed error and rate of change of
of induction motor (IM) drive is employed [1]. However, the the speed error are considered as the input linguistic variables
controller design of such a system plays a crucial role in and the torque-producing current component is considered as
system performance. The decoupling characteristics of the output linguistic variable.
vector-controlled IM are adversely affected by the parameter Thus, the functional relation of the FLC can be expressed as
changes in the motor. The motor-control issues are [14].
traditionally handled by fixed-gain proportional-integral (PI)
and proportional-integral-derivative (PID) controllers. i q ( n ) = ∫ discrete ∆ i q ( n ) = f ( ∆ e ( n ), ∆ ω r ( n )) (9)
However, the fixed-gain controllers are very sensitive to where ∆ e ( n ) = ∆ ω r ( n ) − ∆ ω r ( n − 1 ) is the change of
parameter variations, load disturbances, etc. Thus, the
speed error, is the present sample of speed error,
controller parameters have to be continually adapted. The
∆ ω r ( n ) = w r* ( n ) − ω r ( n ) is the sample of speed error,
problem can be solved by several adaptive control techniques
such as model reference adaptive control (MRAC) [2], ∆ωr (n − 1) past sample of speed error, is the present sample
sliding-mode control (SMC) [3], variable structure control *
of actual, speed, ωr (n) is the present sample of command
(VSC) [4], and self-tuning PI controllers [5], etc. The design
of all of the above controllers depends on the exact system speed, and f denotes the nonlinear function. The main goal of
mathematical model. However, it is often difficult to develop the control system is to track the command speed by
an accurate system mathematical model due to unknown load providing the appropriate torque-producing current
variation, unknown and unavoidable parameter variations component iq depending upon the operating conditions. In
due to saturation, temperature variations, and system real time, the motor position information and output of the
disturbances. In order to overcome to above problems, FLC, which is considered as the command q-axis current, as
recently, the fuzzy-logic controller (FLC) is being used for well as the command d-axis current, are used to get the
motor control purpose [7]–[12]. The mathematical tool for command phase current and using (5). The electrical position
the FLC is the fuzzy set theory introduced by Zadeb [6]. As of the motor can be expressed as
θ e = θ r + θ sl The trapezoidal functions are used as membership functions
for all the fuzzy sets except the fuzzy set ZE (zero) of the
input vectors. The triangular membership functions are used
for the fuzzy set ZE of the input vectors and all the fuzzy sets
of the output vector. The trapezoidal and triangular functions
are used to reduce the computation for online
implementation. Mathematically, the trapezoidal membership
function can be defined as
where θ e is the rotating field position, θr is the rotor
position due to slip speed, and θ sl is the slip position due to Trapezoidal: f (x; a, b, c, d) =
slip speed. In the next step, the scaling factors Kω, Ke and Ki
are chosen for fuzzification, as well as for obtaining the
actual output of the command current. These scaling factors
play a vital role for the FLC. The factors Kw and Ke are The triangular membership function can be obtained from the
trapezoidal function by setting b = c. The rules used for the
chosen to normalize the speed error ∆ wn , and the change of proposed IM specific FLC algorithm are shown in Table I.
speed error ∆ en , respectively, so that these remain within the Based on the above rules, the fuzzy-rule-based matrix is
shown in Table II. For this study, Mamdani-type fuzzy
limit of ± 1 . Factor Ki is so chosen that one can get the rated inference is used [14]. The values of the constants,
current for rated conditions. Here, the constants are taken as membership functions, fuzzy sets for the input/output
Kw = w r* (command speed), Ke =10, and Ki = 10 in order variables, and the rules used in this study are selected by trial
and error to obtain the optimum drive performance. In this
to get the optimum drive performances. After selecting the study, the center of gravity defuzzification is used [15]. The
scaling factors, the next step is to choose the membership output function is given as
function of ∆ w rn , ∆ en and i *qn , which perform the N
important task of the FLC. The membership functions used
for the input and output fuzzy sets are shown in Fig.2.
∑ iµ
k =1
c (k ) (i)
output i = N

∑µ
k =1
c (k ) (i)

where N is the total number of rules and µ c (k ) (i) denotes the

output membership grade for the kth rule with the output
subset C.

III. EXPERIMENTAL IMPLEMENTATION

The proposed FLC-based vector control of IM is
experimentally implemented using DSP-board DS1104
through both hardware and software [16]. The DSP board is
installed in a personal computer (PC) with uninterrupted
communication capabilities through dual-port memory. The
hardware schematic for real-time implementation of the
proposed FLC-based IM drive is shown in Fig.3. The
DS1104 board is based on a Texas Instrument (TI)
Incorporated TMS320C31 32-bit floating-point DSP. The
DSP has been supplemented by a set of on-board peripherals
used in digital control systems, such as A/D, D/A converters, output signals in a digital storage oscilloscope. The complete
and incremental encoder interfaces. The DS 1104 is also IM drive is implemented through software by developing a
equipped with a TI TMS320P14 16-bit micro controller DSP program in high-level American National Standards Institute
that acts as a slave processor and provides the necessary (ANSI) “C” programming language. The program is
digital input/output (I/O) ports and powerful timer functions compiled by the TI “C” code generator. Finally, the program
such as input capture, output capture, and pulse width is downloaded to the DSP controller board using loader
modulation (PWM) waveform generation. In this study, the program LD31 [16]. The sampling frequency for
slave processor is used for digital I/O configuration. The experimental implementation of the proposed FLC-based IM
actual motor currents are measured by the Hall-effect motor drive system is 5 kHz.
sensors, which have good frequency response and are fed to
IV. RESULTS AND DISCUSSIONS
the DSP board through the A/D converter. As the motor
neutral is isolated, only two-phase currents are fed back and Several tests were performed to evaluate the performance of
the third phase current is calculated from them. The rotor the proposed FLC-based vector control of the IM drive
position is measured by an optical incremental encoder, system both theoretically and experimentally. The speed-
which is mounted at the rotor-shaft end. It is then fed to the control loop of the drive was also designed, simulated, and
DSP board through an encoder interface. The encoder experimentally implemented with the PI controller in order to
generates 4096 pulses per revolution. By using a fourfold compare the performances to those obtained from the
pulse multiplication, the number of pulses is increased to 4 × respective FLC-based drive system. The speed responses are
4096 in order to get better resolution. A 24-bit position observed under different operating conditions such as a
counter is used to count the encoder pulses and is read by a sudden change in command speed, step change in load, etc.
calling function in the software. Some sample results are presented in the following section.
The motor speed is calculated from the rotor position The PI controller is tuned at rated conditions in order
by backward difference interpolation. A digital moving to make a fair comparison. Figs. 4 and 5 show the simulated
average filter is used to remove the noise from the speed starting performance of the drive with PI-and FLC-based
signal. drive systems, respectively. Although the PI controller is
tuned to give an optimum response at this rated condition, the
fuzzy controller yielded better performances in terms of
faster response time and lower starting current. Fig.6 (a) and
(b) shows the speed responses of the drive system using the
PI and FLC, respectively, with a step change in the reference
speed. It is evident from Fig.6 (a) and (b) that the proposed
FLC-based IM drive system can follow the command speed
without any overshoot and steady-state error. Thus, the FLC-
based drive system is not affected by the sudden change of
the command speed. Thus, a good tracking has been achieved
for the FLC, whereas the PI-controller-based drive system is
affected with the sudden change in command speed. Fig.7(a)
and (b) shows the speed responses for step change in the load
torque using the PI and fuzzy controller, respectively. It is to
be noted that, in Fig.7 (a), the vertical scale for current iq is
to be divided by five and, in Fig. 7(b), it is to be divided by
four. The motor starts from standstill without load and, at t
=0.8s, a sudden full load is applied. The motor speed follows
its reference with zero steady-state error and a fast response
using a fuzzy controller. On the other hand, the PI controller
shows steady-state error with a high starting iq current. It is
to be noted that the speed response is affected by the load
The input vectors of the FLC are generated from the present conditions. This is the drawback of a PI controller with
and the delayed samples of the speed error. The command varying operating conditions.
currents are generated from the FLC. The hysteresis current
controller compares the command currents with the
corresponding actual motor currents and generates the logic
signals, which act as firing pulses for the inverter switches.
Thus, these six PWM logic signals are the output of the DSP
board and fed to the base drive circuit of the inverter power
module. The D/A channels are used to capture the necessary
 The performance has been investigated at different dynamic
operating conditions both theoretically and experimentally. It
is concluded that the proposed FLC has shown superior
performances over the PI controller.
APPENDIX
MOTOR PARAMETERS
1hp, 3φ, 208 V, 50 Hz, 3.4 A, P =4, Rs= 4.0Ω, Rr =
1.143Ω, Ls = 0.3676 H, Lr = 0.3676 H, Lm = 0.3489, Jm =
0.03Kg.m2, Bm = 0.00098 (N.m)/rad/s.
REFERENCES
[1] F. Blaschke, “The principle of filed orientation as applied to the new
transvector closed-loop control system for rotating-field machines,”
Siemens Rev., vol.34, no.3, pp.217-220, May 1972.
[2] H. Sugimoto and S. Tamai,”Secondary resistance identification of an
induction motor applied model reference adaptive system and its
characteristics,” IEEE Trans. Ind. Applicant., vol.IA-23,pp.296-
303, Mar./Apr. 1987.
[3] C.Y. Won and B.K. Bose, “An induction motor servo system with
improved sliding mode control,” in Proc. IEEE IECON’92, pp. 60–
66.
[4] T.L. Chern and Y.C. Wu, “Design of integral variable structure controller
and application to electro hydraulic velocity servo systems,” Proc.
Inst. Elect. Eng., vol.138, no.5, pp.439-444, Sept.1991.
[5] J.C. Hung, “Practical industrial control techniques,” in Proc. IEEE
IECON’94, pp.7–14.
[6] L.A. Zadeh, “Fuzzy sets,” Inform. Control, vol.8, pp.338-353, 1965.
The simulated results are verified by the experimental results. [7] S.Bolognani and M. Zigliotto, “Hardware and software effective
Fig.8 (a) and (b) shows the experimental speed responses of configurations for multi-input fuzzy logic controllers,” IEEE Trans.
the drive system using the PI and FLC, respectively. It is to Fuzzy Syst. vol.6, pp. 173-179, Feb.1998.
be noted that the fuzzy controller gives better responses in [8] I. Miki, N. Nagai, S. Nishiyama, and T. Yamada, “Vector control of
induction motor with fuzzy PI controller,” in IEEE IAS Annu. Rec.,
terms of overshoot, steady-state error, and fast response. The 1992, pp.464-471.
experimental speed responses with step increase in command [9] Y. Tang and L. Xu, “Fuzzy logic application for intelligent control of a
speed are shown in Fig. 9(a) and (b) for the conventional PI variable speed drive,” IEEE Trans. Energy Conversion, vol. 9,
and the proposed FLC-based IM drive system, respectively. pp.679-685, Dec.1994.
[10] E.Cerruto, A.Consoli, A.Raciti, and A.Testa, “Fuzzy adaptive vector
These figures also show that the FLC-based drive system can control of induction motor drives,” IEEE Trans. Power Electron,
handle the sudden increase in command speed quickly vol.12, pp.1028–1039, Nov.1997.
without overshoot, undershoot, and steady-state error, [11] C.H. Won, S.Kim, and B.K. Bose, “Robust position control of
whereas the PI-controller-based drive system has steady-state induction motor using fuzzy logic control,” in IEEE IAS Conf.
Rec., Houston, TX, 1992, pp.472-481.
error and the response is not as fast as compared to the FLC. [12] A. Rubaai, D.Ricketts, and D. Kankam, “Experimental verification
Thus, the proposed FLC-based drive has been found superior of a hybrid fuzzy control strategy for a high performance brush less
to the conventional PI-controller-based system. DC drive,” IEEE Trans. Ind. Applicant, vol.37, no.-32, pp.503-512,
Mar./Apr.2001.
VI. CONCLUSION [13] B.K. Bose, “Adjustable speed ac drives– A technology status review,”
Proc. IEEE, vol.70, pp.116-135, Feb.1982.
A novel FLC- based indirect vector control of an IM has [14] M.N. Uddin and M.A. Rahman, “Fuzzy logic based speed control of
been presented in this paper. The FLC has been designed for an IPM synchronous motor drive,” J. Advanced Comput. Intell,
a speed-control loop. The simulation has been carried out vol.4, no.3, pp.-212–219, 2000.
[15] H.T. Nguyen, M.Sugeno, R.Tong, and R.R. Yager, Theoretical Aspects
using the Simulink Fuzzy Logic Toolbox Manual Guide [17]. of Fuzzy Control. New York: Wiley, 1995.
The complete IM drive incorporating the FLC has been [16] “Digital signal processing and control engineering,” in Manual Guide.
successfully implemented in real time using a DSP controller Paderborn, GmbH, Germany: dSPACE, 1996.
board DS 1104 for the prototype 1-hp motor. In order to [17] Fuzzy Logic Toolbox User Guide. Natick, MK : Math Works,
1997.
minimize the real-time computational burden, simple
membership functions and rules have been used. Since exact
system parameters are not required in the implementation of
the proposed controller, the performance of the drive system
is robust, stable, and insensitive to parameters and operating
condition variations. In order to prove the superiority of the
FLC, a conventional PI-controller-based IM drive system has
also been simulated and experimentally implemented.