A Novel Technique for the Design of Controller of a VectorControlled Permanent Magnet Synchronous Motor Drive

Shirazul Islam , 2Farhad Ilahi Bakhsh, 3Mohammad Khursheed, 4Shamsuddin Ahmad and 5Atif Iqbal
1 ,2
Student Member, IEEE, 4 ,5Senior Member, IEEE
1

2, 4

Department of Electrical Engineering, Teerthankar Mahavir University, Moradabad, India.

Department of Electrical Engineering, Z. H. C. E. T., Aligarh Muslim University, Aligarh, India.
3
Department of Electrical Engineering, Integral University, Lucknow, India.
5
Department of Electrical and Computer Engineering, Texas A&M University, Qatar.

Abstract The design of the speed controller is important from

the point of view of imparting desired transient and steady-state
characteristic to the speedcontrolled permanent magnet
synchronous motor (PMSM) drive system. The design and
analysis of a vector-controlled closed-loop permanent magnet
synchronous motor (PMSM) drive is presented in this paper
work. The effect of different factors like the time constant of
proportional speed controller and the motor time constant on
drive performance is analyzed. The D-partition technique is used
to investigate the interaction of proportional speed time constant,
amplifier gain and armature time constant on the system
stability. In this paper work, a procedure has been suggested to
find the optimum value of the amplifier gain to give minimum
settling time, which is the most important requirement of high
performance drives or high fidelity drives. Using the D-partition
technique, a comprehensive performance analysis of system
design is carried out in this paper work in MATLAB
environment.
Keywords PMSM Drive, Degree of stability, Settling time, High
performance drive, D-partition technique, Neimark Hatching.

I. INTRODUCTION
Introduction to permanent magnets to replace the
electromagnets which have windings and require an external
electric energy source resulted in compact dc machines. The
synchronous machine with its conventional field excitation in
the rotor is replaced by the PM excitation: the slip rings and
brushes assembly are dispensed with. With the development
in the field of power semiconductor controlled switches, the
mechanical commutator has been replaced with an inverter,
which led to the development of the PMSM and brushless dc
machines. Therefore the excitation field has been transferred
to the rotor with the help of PM poles and the armature circuit
on the stator [1-4].
The permanent magnet synchronous machine (PMSM) is
primarily associated with high performance applications and
is normally fed by a voltage source inverter (VSI). The
machine is of the synchronous type and the rotor field is
created by permanent magnets attached to the rotor. The
material of the permanent magnets can differ but the best
materials are of rare earth type, such as Samarium-Cobalt
(Sm-Co) or Neodymium-Iron- Boron (NeFeB). The NeFeB
magnets combine a high flux density with a large coercive
force [5-6].
The poly-phase PMSM control is rendered equivalent to
that of the dc machine by a decoupling control known as
vector control. The vector control separates the torque and
flux channels in the machine through its stator excitation

inputs. A dynamic model is required to derive the vector

control algorithm to decouple the air gap flux and torque
channels in the derive system [7].
In recent years, Permanent Magnet Synchronous Motors
(PMSMs) are increasing applied in several areas such as
traction, automobiles, robotics and aerospace technology [8].
Accurate digital simulation tools are necessary to evaluate
their field performance particularly when they are driven with
solid-state drives connected to larger electrical networks. One
of the areas of interest is the design of controllers for these
motor drives. In many applications the physical controls have
to be designed and tuned for best performance [9].
The design of the speed controller is important from the
point of view of imparting desired transient and steady state
characteristics to the speed-controlled PMSM drive system. A
proportional-plus-integral controller is sufficient for many
industrial applications. For high fidelity drive the optimum
setting of the gain and time constant of PI controller is
necessary. There are many conventional techniques available
for the design of the speed controller [10-11], but in this paper
work, the D-partition technique has been used for its synthesis.
Based on analytically characterizing the D-partition
boundaries of the controller parameter space, necessary
conditions of the maximum degree of stability are derived.
With these derived conditions, the problem of maximizing the
degree of stability is converted to a set of parametric
optimization problems, whose solutions can be obtained by an
existing method [12]. Also the effect of variation of the other
parameters of the drive can also be studied with this technique.
To the best of the information available to the author,
performance and stability analysis of vector-Controlled
PMSM drive, using D-partition technique, has not been
carried so far. The parameter plane technique is proposed to
study the stability of such drive and design carried out for
fastest transient response. A comprehensive analysis of system
design is presented, which shows that the value of PI
controller time constant influence the value of PI controller
gain for minimum settling time.
The work related to the performance and stability analysis
of indirect vector-controlled close-loop drive using the same
technique has been explored by the same author [13].
II. DESCRIPTION OF THE DRIVE
From the previous discussion, the vector controlled PM
synchronous motor drive schematic is obtained and is shown
in figure 1.

for a balanced three phase operation by substituting for

from equation (4) and (5)

and

(7)
These statorphase current commands are amplified by the
inverter and its logic and fed to the PMSM.
The rotor position is obtained with a position encoder or
synchronous resolver. The velocity signal
is extracted
from the rotor position by using processors.
III. BLOCK DIAGRAM DERIVATION AND EVALUATION OF
TRANSFER FUNCTION
The Block diagram of vector-controlled PM synchronous
motor is derived in this section. This is made possible by
deriving the transfer functions of the various operating
systems such as the induction machine, inverter, speed
controller and feedback transfer function.
A. Derivation Of Transfer Function:
The motor q-axis voltage equation with the d-axis current
being zero become:

V r qs = ( Rs + Lq p)i r qs + r af

(8)
And the equation for electromechanical dynamics is given by

Fig.1 Vector-controlled PM synchronous motor drive

The torque reference is a function of the speed error and the

speed controller is usually of PI type. For fast response of the
speed, a PID controller is appropriate. The product of torque
*
reference and the air gap flux-linkage
in p. u. generates
*
the torque producing component iT of the stator current. The
reason for this block is to adjust the torque producing
component of the stator current both in the constant-torque
and in the constant torque and in the constant power regions
of operation. The function generator on the speed has the
following characteristics:
(1)

d r
P
+ B1r = (Te Tl )
dt
2

(9)

Where the value of electromagnetic torque is given by

Te =

3P
af iqsr
22

(10)

And the value of frictional load is given by

Tl = Blm

(11)
When these values are substituted in equation (9), the
dynamics
equation
for
load
becomes

d r
3 P
+ B1 r = ( ) 2 af iqsr Bl r = K t .iqsr Bl r
dt
2 2
(12)

(2)
By using these relations the transfer function between the
and
(3) speed and the torque producing component is given by
sets the reference for the
The function generator
Km
IT ( s)
=
resultant mutual flux linkages, involving a constant Kf. The
r ( s ) 1 + sTm
(13)
flux linkages reference
determines the field current
P
1
J
reference required to counter the flux linkage vector due to
where, Bt = Bl + B1, Km = ;Tm =
rotor magnet.
2
Bt
Bt
The stator-current phasor magnitude
and the torque
and Kt = 3 ( P )2 af
angle
command are evaluated by
2 2
Transfer
Function
of the Inverter:
=
(4)
The inverter is modelled as a gain with a time lag by
=
(5)
K in
=1

The instantaneous position of the stator current is phasor

command is given by
(6)
From which the stator-phase current commands are obtained

Gin ( s ) =

1 + sTin

where, K in = 0.65

(14)

Vdc
1
, Tin =
2 fc
Vcm

Speed Controller:
The transfer function of speed controller is given by

Gs ( s ) =

K s (1 + sTs )
sTs

(15)

Current Feedback Transfer Function:

The current feedback signal function has only gain block
whose value is given by

Gc (s) = H c

(16)

Speed feedback Transfer Function:

The speed feedback signal is processed through a first order
filter whose transfer function is given by

Gw ( s ) =

rm ( s )
Kw
=
r ( s ) 1 + sTw

(17)
By incorporating above equations, the complete block
diagram of the vector-controlled PMSM close-loop drive with
constant rotor flux linkages is shown in figure 2.
Block Diagram of Speed-Controlled PMSM Drive:

The closed loop transfer function is given by

Kt K s Km Ka Kin (1 + a1s + a2 s 2 )
G( s)
=
1 + G(s) H (s) A1s5 + A2 s 4 + A3 s3 + A4 s 2 + K (1 + b1s) (19)
From the equation (19), the characteristic equation of the
closed loop system can be written as
5
4
3
2
1+ G(s)H(s) = As
1 + A2s + A3s + A4s + K(1+ bs
1 ) = 0 (20)

where,

A1 =TinTT
a mTT
w s
A2 =Ts[TinTT
a m +Tw(TinTa +TT
a m +TinTm)]
A3 =Ts[TinTa +TT
a m +TinTm +Tw(Ta +Tm +Tin) +KaKT
b inTw +HcKaKinTmTw]
A4 =Ts[Ta +Tm +Tin +Tw +KaKb(Tin +Tw)+HcKaKin(Tm +Tw)+KaKb +HcKaKin +1]
K = KKK
t s mKaKinHc
a1 =Ts +Tw
a2 =Ts.Tw
b1 =Ts
L
1
Ka = ,Ta = q , Kb = KK
t mf
Rs
Rs
(21)
B. Drive Specifications:
To illustrate the analysis, the drive represented by figure 2,
with the following specifications is considered:
Three phase, 5 HP, 60Hz, 6-pole, 200V, 14.38A, Y connected
Mutual flux linkages due to rotor magnets af = 0.1546 Wbturn.
Switching frequency of the inverter fc = 2 KHz.
Maximum control voltage Vcm =10V.
DC link input voltage to the inverter = 285V
Stator resistances Rs

= 1.4,

Stator direct axis and quadrature axis inductances:

Ld = 0.0056 , Lq = 0.009 H
Moment of inertia of motor and load: J = 0.006 Kg.
Viscous friction coefficient: Bt = 0.01 N m s/rad
Steady state rotor speed: ro = 370 rad /sec

Fig. 2 Block diagram of speed-controlled PMSM drive.

Gain of the tachometer Hw = 0.05 V/V,

Time constant of the tachometer Tw = 0.002 sec,
Gain of the inverter Kin = 18.525 V/V,
Time constant of the inverter Tin = 0.00025 sec,
Gain of the PI controller Ks = 8.6638,
Time constant of the PI controller Ts = 0.0155 sec,
Gain of the current feedback loop Hc = 0.8 V/A.

C. Use of Parameter Plane Technique for Stability

Analysis and Determination of Amplifier Gain for
minimum Settling Time
The
parameter plane technique is used to study the effect of
KKKKK
t s m a in(1+sTs)
Gs
( )=
system
parameters on the drive stability and determine the
3
2
sTs[TTTs
in a m +(TT
in a +TT
a m +TT
in m)s +(Ta +Tm +Tin)s+KK
a b(1+sTin)+HKK
c a in(1+sTm)+1] value of amplifier gain for minimum settling time. Now we
(18) have to select the parameters of interest [14, 15]. Depending

Solving the block diagram of the system shown in figure 2,

the open loop transfer function G(s) is given by

upon the occurrence of parameters of interest in the

characteristic equation following two cases arises;
1. Linear case
2. Non-linear case
Effect of Variation of System Parameters on Dynamic
Stability:
The parameter plane technique also known as the D-partition
technique can be used for studying effect of variation of
system parameters, considering two of them at a time. In the
case under study, there are three important parameters of the
interest Ta, Ts and gain K. The dynamic stability has been
studied for variations in these three parameters in the
following two combinations:
(a) Ta and K varied; Ts kept constant.
(b) Ts and K varied; Ta kept constant.
Variation of Armature Time Constant Ta and Gain K:
In this case, the parameters Ta and K occur independently, this
give rise to linear case. From equation (20), the characteristic
equation is given by
Ta ( K1 s 5 + K 2 s 4 + K 3 s 3 + b1 s 2 ) + K (1 + b1 s ) + L1 s 4 + L2 s 3 + L3 s 2 + L4 s = 0

(22)
where,

K1 = TinTmTwTs
K2 = TsTinTm + TwTS (Tin + Tm )
K3 = TinTs + TsTm + TwTs
K4 = Ts
L1 = TinTmTwTs
L2 = Ts [TinTm + Tw (Tm + Tin ) + Ka KbTinTw + Hc Ka KinTmTw ]

Fig. 3 D-partition curve for the linear case in Ta-K plane

As the region R4, corresponds to the negative values of K

and, it is not a feasible region. The most stable region is
therefore is the region R1.This reveals that any combination of
positive values of Ta and K from region R1, will lead to the
stable operation.
Variation of Controller Time Constant Ts and Gain K:
In this case, the parameters of interest selected are time
constant of PI controller Ts and overall gain K. From the
equation (20) the value of characteristic equation in terms of
these
parameters
is
given
by
5
4
3
2
1 + G(s) H (s) = (M1s + M 2 s + M 3s + M 4 s + M 5 s + K + KTs s)
(24)

L3 = Ts [Tm + Tin + Tw + Ka Kb (Tin + Tw ) + Hc Ka Kin (Tm + Tw )] where,

M1 = TinTaTmTw
L4 = Ts [Ka Kb + Hc Ka Kin +1]
Substituting s = j in equation (22) and solving for Ta and K,
M2 = TinTaTm +Tw (TinTa +TaTm +TinTm )
Ta = (S1.Q2 - S2.Q1) / 1 , K = (P1.S2- P2.S1)/ 1
(23) M3 = TinTa + TaTm +TinTm +Tw (Ta +Tm +Tin ) + Ka KbTinTw + Hc Ka KinTmTw
where,

P1 = 2.0933 *10 5 * 4 0.0155 * 2 ;

P2 = 4.65 *10 9 * 5 0.0093 * 3 ;
Q1 = 1;
Q 2 = 0.0155 * ;
S1 = 4.65 *10 9 * 4 + 0.1089 * 2 ;
S 2 = 2.1796 *10 4 * 3 0.5367 *
1 = P1 .Q 2 P2 .Q1 ;
From equation (23), the D-partition boundary in the (Ta-K)
plane can be plotted as shown in figure 3. Following the
Neimark hatching rule [7], the region of stability is
determined. The entire (Ta-K) plane is divided into regions
marked R1, R2, R3R5, and the number within the
parenthesis denotes the relative number of roots present in the
left half of S-plane. The most stable regions are R1 and R4,
which corresponds to maximum number of roots (n).

M4 = Ta +Tm +Tin +Tw + Ka Kb (Tin +Tw) + Hc Ka Kin (Tm +Tw )

M5 = Ka Kb + Hc Ka Kin +1
It may be noted here that the parameters of interest Ts and K,
in this case do not occur independently, in the characteristic
equation (24), therefore this give rise to non-linear case.
Substituting s = j in equation (24) and separating the real
and imaginary parts, the corresponding system of two
simultaneous equations can also be written as
H1 * Ts * K + G1 * Ts + E1 * K + D1 = 0,

H 2 * Ts * K + G2 * Ts + E2 * K + D2 = 0,

(25)
These two equations provide two independent sets of values
of controller time constant and gain obtained as (Ts, K) and
(Ts, K) as

Ts = {e + (e 2 4* a * c)1/ 2 } / 2* a;
K = (G1 * Tin + D1 ) / ( H1 * Ts + E1 );

Ts ' = {e (e 2 4 * a * c )1/2 } / 2 * a;
K ' = (G1 * Tin '+ D1 ) / ( H1 * Ts '+ E1 );
where,

a = G 2 * H 1 G1 * H 2 ;
b = E 2 * H 1 E1 * H 2 ;
c = E1 * D 2 E 2 * D1 ;
e = G 2 * E1 G1 * E 2 + H 1 * D 2 H 2 * D1 ;
J = a * Tin + b * K + G1 * E 2 G 2 * E1 ;
H1 = 0;
H 2 = ;
G1 = (8.9432 *106 * 4 7.0315* 2 );
G2 = (1.92*109 * 5 0.0179* 3 + 34.6462* );
E1 = 1;
E2 = 0;
D1 = 0;
D2 = 0;

D. Design of the Drive System for Maximum Stability

with Minimum Settling Time
From the most stable region of figure 4, a number of values
(26)
of gain K can be selected for a given value of controller time
constant Ts. Out of these values of K, a particular value that
gives the minimum settling time and hence the fastest
transient response may be determined. This value of K has
been determined as explained below.
Substituting s = (- + j ) in the characteristic equation
(24), a set of different D-partition boundaries can be obtained
for different increasing positive values of . In each case, the
most stable region can be determined as before. The larger the
value of , narrower is the most stable region, and there exist
a maximum value of beyond which the most stable region
disappears. This critical value of is called c. The Dpartition boundary for = c can be plotted and the mast
stable region marked. A value of gain K corresponding to a
given value of Ts selected from this region will give the
minimum settling time. It can therefore be inferred that the
value of gain K is directly influenced by time constant Ts if
the system is designed for minimum settling time.
For the case under study, the value of c is obtained as 30.5,
for which the D-partition boundary is plotted as shown in
figure 5. The most stable region is R1 containing a maximum
number of roots in the left half of s-plane. From this region,
the value of gain K is obtained corresponding to a given
values of Ts. From the most stable region R1, it can be noted
that for the given value of Ts = 0.0155 sec, the corresponding
value of the gain K, selected from the most stable region is
253. Knowing the value of K, the value of the controller gain
for the given value of PI controller time constant Ts, can be
calculated as Ks = 1.8328 from the following relation:
Ks = K / (Kin. Km. Ka. Hw )
From this observation we can infer that if we have to design
the PI controller for maximum stability and minimum settling
time, the value of the gain Ks of the PI controller is 1.8328. As
the D-partition technique indicates the relative stability of the
system, the absolute stability is varied using Routh criteria.

Fig. 4 D-partition curve for the nonlinear case in Ts-K plane

From equation (26), the D-partition boundary in the Ts-K

plane can again be plotted. It is shown in the figure 4.
Following the Neimark hatching rule, the region of stability is
determined. The entire Ts-K plane is divided into regions
marked R1, R2 upto R6, and the number within parenthesis
denotes the relative number of roots present in the left half of
s-plane. The most stable region is R1 and R5, which
correspond to maximum number of roots (n). As the region
R5, corresponds to the negative values of K and, it is not a
feasible region. The most stable region is therefore the region
R1. It is again observed that any combination of the positive
values of Ts and K can be selected to provide the most stable
operation.
Fig. 5 D-partition curve for the nonlinear case in Ts-K plane

IV. CONCLUSIONS
The work presented in this paper work deals with the
stability analysis of a vector-controlled PMSM closed-loop
drive. The mathematical model of the above drive system has
been developed. By using the same mathematical model, the
transfer function as well as the characteristic equation of the
drive system has been derived.
The Parameter Plane Technique is used to investigate the
effect of system parameters, like the gain and time constant of
the PI controller, on the stability of the drive system. The
analysis has been carried out for two different cases
depending upon the parameters of interest selected i.e. for
linear case and non-linear case. For the latter case, the drive
system is designed for minimum settling time. In the first case
the parameters of interest selected are armature time constant
Ta and overall gain K of the system while in the last two cases,
the parameters of interest selected are time constant of the PI
controller Ts and overall gain K of the system. From the value
of the overall gain K, the value of the controller gain Ks can be
obtained. The range of the possible gain and time constant for
fastest transient response can be computed from the Dpartition boundary plot as discussed is the paper. Thus the
proposed technique provides a handy tool for obtaining the
correct system parameter for obtaining most stable operation
of the drive system and a guide line for the design of high
fidelity drives.

[6]

[7]

[8]

[9]

[10]
[11]
[12]

REFERENCES:

[1] B. K. Bimal, Modern Power Electronics and AC

Drives, Prentice Hall, New Delhi, India, 2005.
[2] B. Cui, J. Zhou, and Z. Ren, "Modeling and simulation
of permanent magnet synchronous motor drives," Fifth
International Conference on Electrical Machines and
Systems, Shenyang , China, pp. 905-908, 18 - 20 Aug.,
2001.
[3] R. Krishnan, Electric Motor Drives Modeling, Analysis,
and Control, Pearso Education, 2001.
[4] B. K. Bose, Power Electronics and Variable Frequency
Drives, Wiley, John & Sons, 1996.
[5] T. Sebastian, G. Slemon, and M. Rahman, "Modelling of
permanent magnet synchronous motors," IEEE

[13]

[14]
[15]

Transactions on Magnetics, vol. 22, pp. 1069-1071, Sep.

1986.
T. M. Jahns, G. B. Kliman, and T. W. Neumann,
"Interior Permanent-Magnet Synchronous Motors for
Adjustable-Speed Drives," IEEE Transactions on
Industrial Applications, vol. IA-22, no. 4, pp. 738-746,
July 1986.
P. Pillay, and R. Krishnan, "Modeling, simulation, and
analysis of permanent-magnet motor drives," IEEE
Transactions on Industry Applications, vol. 25, no. 2, pp.
265-273, Mar/Apr. 1989.
L. Samaramayake, and Y. K. Chin, "Speed Synchronized
Control of Permanent Magnet Synchronous Motors with
field-weakening," International Conference on Power
and Energy Systems, Euro PES 2003, vol. 3, pp. 547-552,
Sep. 2003.
A. B. Dehkordi, A. M. Gole, and T. L. Maguire,
Permanent Magnet Synchronous Machine Model for
Real- Time Simulation, International Conference on
Power Systems Transients (IPST05), Montreal, Canada,
Paper No. IPST 05 159, June 19-23, 2005.
A. O Dwyer, Handbook of PI and PID Controller
Tuning Rules, Imperial College Press, London, 2003.
C. C. Yu, Autotuning of PID Controllers: Relay
Feedback Approach, Springer, Berlin, 2001.
C. Hwang, S-H. Hwang, and J-F. Leu, Tuning PID
controllers for time-delay processes with maximizing the
degree of stability, IEEE 5th Asian Control Conference,
Melbourne, Australia, pp. 466-471, 20-23 July, 2004.
S. Islam, M. Khursheed, F.I. Bakhsh, and A. Iqbal,
Design and Performance Analysis of an Indirect
Vector-Controlled Closed-loop Drive System using Dpartition Technique, IEEE International Conference on
Process Automation, Control and Computing,
Coimbatore, 20-22 July, 2011.
M. V. Meerov, "Introduction to the Dynamics of
Automatic Regulating of Electrical Machines,"
(Butterworths, 1971), translated from Russian.
D. D. Siljak, Non-Linear System-The Parameter
Analysis and Design, John Wiley and Sons, New York,
1979.