PI-CONTROLLERS DETERMINATION FOR VECTOR

CONTROL MOTION
Minárech Peter 1), Makyš Pavol 2) and Vittek Ján 3)

Abstract
This paper presents determination of PI-controller parameters for vector control of
the drive with permanent magnet synchronous motor. Vector control is widely used
and popular control algorithm of the electric drives. Proper calculation of PI-
controller parameters ensures correct operation of vector controller and high
precision of the drive. Inputs to the control algorithms for calculation of
proportional gain (P component) and integral time constant (I component) are motor
parameters together with settling times of current, speed and position control loops,
which can be chosen by the designer. Settling time of each control loop can be
changed to achieve the best results of the motion control (speed or position control).
Compensation of an external load torque acting on the shaft of the motor is included.

1 VECTOR CONTROL SYSTEM DEVELOPMENT

A) Current PI-Controllers Design
For the design of PI-controllers it is necessary to know the closed loop transfer function. Block
diagrams of permanent magnet synchronous motor (PMSM) in d and q-axis are shown in Fig.1.

Fig.1 - a) Block diagram of PMSM in d-axis

Fig.1 - b) Block diagram of PMSM in q-axis

To describe appropriate model of PMSM d_q coordinate system fixed to machine rotor is used.
Model is divided into two orthogonal axis enabling independent control of flux and torque component.
If Mason´s formula is used the transfer function between flux component, i d and input voltage, ud in d-
axis is:
 Similar way the transfer function between torque component, iq and input voltage, uq in q-axis is:

Both transfer functions in d_q coordinate system can be replaced by the first order transfer
functions with gain Ka and with electrical time constants Ta-d and Ta-q.

For the design of PI-controller of magnetic flux control in d-axis following block diagram is used:

Fig.2 – Current loop of PMSM in d-axis

To achieve best results of the motion control and full flux of the machine condition id_dem = 0
must be secured. If current PI-controller in d-axis secures such condition, then the only magnetic flux
created will be flux of permanent magnets. Closed loop transfer function with PI-controller can be
written in following way:

For the design of PI-controller by ˝pole placement method˝, the demanded transfer function (5)
of corresponding order is compared with denominator of closed loop transfer function:

Where n is order of the system and ω0 is minimal natural frequency obtained from Dodd´s
formula (6), which respects prescribed settling time Tu of corresponding loop.
 The comparison of the ideal first order transfer function with denominator of the closed-loop
transfer function of current control loop results in values of proportional gain KRid and time integral
constant TRid of PI-controller in d-axis while respecting prescribed settling time:

Similar way, constants of current PI-controller in q-axis can be calculated. For the design of PI-
controller of current loop in q-axis for control of electromagnetic PMSM torque following block
diagram is used:

Fig.3 – Current loop of PMSM in q-axis

Values of current PI-controller in q-axis are calculated similar way as it was described for d-axis
and the results are as follows:

where Tuq is settling time of current loop in q-axis, KRiq is proportional gain and TRiq is time
integral constant of PI-controller in q-axis.

B) Speed PI-Controller Design

For the design of speed PI-controller block diagram shown as Fig.4 is used. Inner current
control loop is replaced by its first order transfer function respecting settling time of current control
loop Tuq, where Tp is time constant of current loop. Following relation between these two time
constants is valid:

Fig.4 – Speed loop of PMSM in q-axis

 The closed loop diagram for speed control of PMSM includes filtering block on the input of
demanded speed (the first order transfer function), which compensates overshooting of speed. Time
constant of compensator Tcom is then equal:

Closed loop transfer function of speed control loop with PI-controller is as follows:

For designing speed PI-controller by pole placement method, denominator of closed-loop

speed transfer function is compared with ideal transfer function of the same order, which for
order n = 3 is as follows:

For calculation of proportional gain KRω and integral time constant TIω of speed PI-controller,
ideal transfer function (13) is designed exploiting Dodd´s formula (6) which respects settling time of
speed control loop. Then the ideal transfer function respecting settling time of speed control loop is as
follows:

Comparing (12) and (14) speed PI-controller parameters are obtained:

For the proper function of speed control loop with chosen settling time Tuω, settling time of
current loop in q-axis should be adjusted to satisfy following condition (where equation (10) is used):

If condition (17) is respected, speed loop for chosen settling time will work with high precision.
On the other hand, if settling time of current control loop in q-axis will not fit condition (17) dynamics
of speed control loop will differ from the demanded.
 C) Position P-Controller Design

P-controller of position control loop can be calculated similar way as it was for current or speed
control loop. For the design of position P-controller block diagram shown in Fig.5 is used. For proper
function of position controller the settling time of current control loop in q-axis is respected and
parameters of PI speed controller and must be calculated as describe equations (22a and 22b).

Fig.5 – Position loop of PMSM in q-axis

Closed loop transfer function of position control loop with P-controller is as follows:

To design P-controller for position control and speed PI-controller by pole placement method,
denominator of closed speed loop transfer function is compared with ideal transfer function of the
same order. Demanded transfer function now of order n = 4 is as follows:

Using Dodd´s formula (6) the ideal transfer function (19) is adjusted to fit settling time for
position control loop as:

Comparing (18) and (20) position P-controller and speed PI-controller parameters are obtained:

To secure proper function of position control loop with chosen settling time Tup, the settling
time of current control loop in q-axis must be adjusted to satisfy following condition:
 D) GUI Application

The „Graphical User Interface‟ application shown in Fig.6 was created using „switched board
programming‟ method. This application enables fast determination of PI-controllers for position and
speed vector control of PMSM. At first user must define its own parameters of PMSM, due to fact
that some parameters are redefined as initial values. Then settling times of current control loops in d
and q axis, settling time of speed control loop and settling time of position control loop can be
changed by typing new value to the corresponding edit box or slider cursor can be used for changing
these settling times. If settling time of one of loop is set then PI-controller parameters of chosen loop
are calculated.

Fig.6 – Basic windows of GUI application

Bode and Nyquist diagrams can be also plotted to consider control quality. Application
calculates transfer function of chosen closed-loop control system together with step response to verify
correct controller design. User can convince himself that the designed controller fits chosen settling
time by displaying step response function. If settling times of control loops are changed independently
a new calculation of PI-controller parameters follows and a new motion behavior can be displayed.
But for proper and precise function of motion vector controller (speed or position) conditions for
settling times given by equations (17) and (23) must be satisfied.

Fig.7 – Bode, Nyquist and step response diagram of GUI application

References

[1] Skalický, J.: „Teorie řízení“, Vysoké učení technické v Brňe, Fakulta elektrotechniky a
komunikačních technologií, Ústav výkonové elektrotechniky a elektroniky, Technická 8, 616
00 Brno, 2002
[2] Vittek, J., Dodds, S.J.: „Riadenie elektrických pohonov s vnútenou dynamikou –
Forced Dynamics Control of Electric Drives“, EDIS – Publishing Centre of Ţilina University,
June 2003, ISBN 80-8070-087-7, http://www.kves.uniza.sk/?page=Elearn

[3] Makyš, P., Štulrajter, M.: „Záverečná správa úlohy:30-129-01EVPU“, Centrum

výskumu mechatronických systémov, Ţilinská univerzita v Ţiline, December 2008

[4] I. Boldea, A. S. Nasar, “Vector Control of AC Drives,” CRC Press, London 1992.

1) PhD. student, Department of Power Electrical Systems, Faculty of Electrical Engineering,

University of Ţilina, Veľký Diel 010 26 Ţilina Slovakia, phone: 041 513 2270, email:
Peter.Minarech@kves.uniza.sk
2) PhD. researcher, Department of Power Electrical Systems, Faculty of Electrical Engineering,
University of Ţilina, Veľký Diel 010 26 Ţilina Slovakia, phone: 041 513 2154, email:
Pavol.Makys@kves.uniza.sk
3) Professor, Department of Power Electrical Systems, Faculty of Electrical Engineering,
University of Ţilina, Veľký Diel 010 26 Ţilina Slovakia, phone: 041 513 2155, email:
Jan.Vittek@fel.utc.sk