Electric Drives

Q An electric drive is a system that converts

electrical energy to mechanical energy

• Parts:
– electric motor (or several)
– control system (including software)

• Constant-speed drives
– only a start/stop and protection system
in addition to the electric motor

• Variable-speed drives (VSDs)

– include an electronic power converter

1
Electric Drive and the Surrounding System

Energy

Power
supply
Electric drive

Display and Drive

Converter Process
control panel control

Effective Process
Measurements work control

Motor Gear Load

Field bus

2
 Acceleration of Inertial Mass
• Torque needed for
accelerating the
moment of inertia J:

dΩ m
TJ = J
dt

• Moment of inertia of
a thin-walled cylinder
J = mr 2

• Moment of inertia of
a solid cylinder
1 2 π
J= mr J= ρ lr 4
2 2

3
 Mechanical Transmissions
• Gear ratio
Number of teeth Z1

n1 Ω m1
i= =
n2 Ω m2

Z2 r2
i= =
Z1 r1 Number of teeth Z2

• Torque reduced • Moment of inertia reduced

to the motor shaft to the motor shaft
(power is preserved): (kinetic energy is preserved):
1 1
Ω m1T1 = Ω m2T2 J1Ω m2 1 = J 2 Ω m2 2
2 2
Ω T Ω m2 2 J
T1 = m 2 T2 = 2 J 1 = 2 J 2 = 22
Ω m1 i Ω m1 i
4
 Reduction of Linear Motion
Ω m1T1 = v2F2

v2
T1 = F2
Ω m1

1 1
J 1Ω m2 1 = m2 v22
2 2

v22
J1 = m2
Ω m2 1

5
 Equation of Motion
d Ωm d Ω m Te − TL
Te = J + TL =
dt dt J

Acceleration of a Acceleration of a constant-torque/

constant-torque drive constant-power drive

6
 Field-Weakening (Constant Power)
• The nominal speed is obtained at nominal armature voltage.
• Higher speeds can be achieved by decreasing the magnetic flux.
• The torque must be restricted (constant power operation).

ua ≈ kφ Ω m Te = kφ ia
kφ = K aΦ Pm = Ω mTe
7
 Four-Quadrant Operation
• Operation in all four quadrants of the Wm-T plane can be achieved:
motor and generator (braking) operation in both rotational directions.
• The direction of the armature current is changed for reversing the
torque direction.
T

Generator Motor

Speed Operating
reversal point

Ωm

Motor Generator

8
 DC Motor Drives

Electromechanics and Electric Drives

9
 Dynamic Model of a DC Motor
• Armature circuit
dia
u a = Ra i a + La + kφ Ω m kφ = K aΦ = f(i f )
dt
• Equation of motion and torque

dΩ m f( i f )
Te = J + TL
dt

Te = kφ ia
if
• Field circuit

dψ f
uf = Rf if + ψ f = f(i f )
dt

10
 Dynamic Model of a DC Motor

dia dΩ m
u a = Ra i a + La + kφ Ω m Te = kφ ia Te = J + TL
dt dt

After Laplace transform:

ua ( s) = Ra ia ( s) + sLa ia ( s) + kφ Ω m( s) Te ( s) = kφ ia ( s) Te ( s) = sJ Ω m( s) + TL ( s)

11
 Block Diagram of a DC Motor
1/ Ra Armature
ia ( s ) = ⎡ u (s ) − kφ Ω m (s )⎤⎦
La ⎣ a time constant
1+s L
Ra τa = a
Ra
Te (s ) = kφ ia (s ) Electromechanical
time constant
Te (s ) − TL (s ) JRa
Ω m (s) = τm =
sJ kφ2

Block Diagram

12
 Derivation of Transfer Functions

TL
ua + 1/ Ra ia Te − 1 Ωm
kφ Ω m ( s)
− 1 + sτ a + sJ =?
uai ua ( s)
kφ

D( s)
R( s) + + Y ( s) Y ( s) G1( s)G2 ( s)
G1( s) G2 ( s ) =
− + R( s) 1 + G1( s)G2 ( s)H( s)

Y ( s) G2 ( s)
H( s) =
D( s) 1 + G1( s)G2 ( s)H( s)

13
Transfer Functions of a DC Motor

Ω m ( s) 1 1
=
ua ( s) kφ 1 + sτ m + s2τ mτ a

ia ( s) 1 sτ m
=
ua ( s) Ra 1 + sτ m + s2τ mτ a

Ω m ( s) Ra 1 + sτ a
=−
TL ( s) kφ2 1 + sτ m + s2τ mτ a

14
 Open-Loop Speed Control

• The angular speed is adjusted by means of the armature voltage ua.

• The speed will change if the load or the supply voltage is varying.
• The method can be used only for coarse speed adjustment in low-
power drives.

15
 Closed-Loop Speed Control

• The method is rarely used (only in very small servo motors).

16
 Cascade Speed Control

• Voltage disturbances are corrected by the inner control loop.

• Load changes are corrected by the outer control loop.
• It is easy to limit the current by ia,ref.
17
 Feedback Control Objectives
Controller Plant
+ e(s )
xref (s ) Gc (s ) Gp ( s ) x(s )
−

Q Feedback control makes the system

insensitive to disturbances and parameter
variation.
Q Control objectives:
• zero steady-state error
• good dynamic response (fast response,
small overshoot and short settling time)

18
 Definitions
Controller Plant
+ e(s )
xref (s ) Gc (s ) Gp ( s ) x(s )
−

• Open-loop transfer function GOL ( s) = Gc ( s)Gp ( s)

x( s) GOL ( s)
• Closed-loop transfer function GCL ( s) = =
xref ( s) 1 + GOL ( s)

• 0 dB crossover frequency fc ,ωc GOL ( jωc ) = 0 dB

19
 Closed-loop Step Response
1 1
x(s ) =
s 1+ s
1
kOL
0.9 xref (t )

0.8
x(t )
0.7
τ
0.632
0.6 x(t ) = 1 − e −t /τ
0.5
1
0.4 τ= = 0, 5 ms
kOL
0.3
0.2
0.1
0
0 0.5 1 1.5 2 2.5 3
t (ms)

20
Steps in Designing the Controller
1. Assume that the system is linear about the
steady-state operating point and design the
controller using linear control theory.
2. Simulate the design under large-signal
conditions and "tweak" the controller as
necessary.

• For small-signal analysis, it is assumed that

the steady-state operating point = 0.
• If the highest bandwidth is at least an order of
magnitude lower than the switching frequency,
the switching-frequency components can be
ignored.

21
 Cascade Control

reference reference
speed torque torque speed

reference + Position + Speed + Torque Electrical Mech. 1 position

position controller controller controller system system s
− − −
torque (current)

speed
position

• Torque loop: fastest

• Speed loop: slower
• Position loop: slowest

22
 Modeling of DC Machines

TL ( s)

ia (s ) Te (s ) − 1 Ω m (s )
+ 1 / Ra
ua ( s ) kφ
1 + sτ a + sJ
−

uai ( s)
kφ

23
 Controller Design
Q Procedure
• Design the torque loop (fastest) first.
• Design the speed loop assuming the torque loop
to be ideal.
• Design the position loop (slowest) assuming the
speed loop to be ideal.

24
 Torque (Current) Loop
• Simplifying assumption: The moment of inertia J is assumed to be high
enough, and the feedback from the angular speed can be ignored.

TL ( s)

ia ,ref (s) + ua ( s ) ia (s ) Te (s ) − 1 Ω m (s )
1 / Ra
PI kT kφ
+ 1 + sτ a + sJ
− −

uai ( s)
kφ
ia (s )

PI
ia ,ref (s) + kiI ⎛ k pI ⎞ ua ( s ) 1/ Ra i a (s )
⎜ 1 + s ⎟ kT
− s ⎝ kiI ⎠ 1 + sτ a

i a (s )

25
 Design of the Torque (Current) Loop
PI
ia ,ref (s) + kiI ⎛ k pI ⎞ ua ( s ) 1/ Ra i a (s )
⎜ 1 + s ⎟ kT
− s ⎝ kiI ⎠ 1 + sτ a

i a (s )

k ⎛ kPI ⎞ 1/ Ra
GI ,OL (s ) = iI ⎜ 1 + s k
⎟ T
s ⎝ kiI ⎠ (1 + sτ a )

• The zero of the PI controller is selected to cancel the motor pole:

k ⎛ k ⎞ 1/ Ra k pI
GI ,OL ( s) = iI ⎜⎜ 1 + s PI ⎟⎟ kT = τa
s ⎝ kiI ⎠ ( 1 + sτ a ) kiI

• kiI is chosen to achieve the desired 0 dB crossover frequency ωcI:

kiI kT RaωcI
GI ,OL (j ωcI ) = =1 kiI =
ωcI Ra kT
26
 Speed Loop

Ω m ,ref (s) + ia ,ref (s) ia (s ) Te (s ) 1 Ω m (s )

PI 1 kφ
− sJ

Ω m (s )

• The current loop is assumed to be ideal (represented by unity).

• The open-loop transfer function is

k iΩ ⎛ k pΩ ⎞ kφ
GΩ ,OL ( s) = ⎜ 1 + s ⎟ ⋅ 1 ⋅
s ⎝ k iΩ ⎠ sJ

or
k pΩ
1+ s
kiΩ kφ k iΩ
GΩ ,OL ( s) =
J s2

27
 Design of the Speed Loop
k pΩ
1+ s
kiΩ kφ k iΩ
GΩ ,OL ( s) =
J s2

• The 0 dB crossover frequency ωcΩ is chosen an order of magnitude

lower than ωcI with a reasonable phase margin φpm,Ω (e.g. 60°):

GΩ ,OL (s )
s = jωcΩ
=1 ∠ ⎡⎣GΩ ,OL (s)⎤⎦ s= jω = −180° + φ pm ,Ω
cΩ

2
k pΩ ⎛ k pΩ ⎞
1 + jω c Ω 1 + ⎜ ω cΩ ⎟
kiΩ kφ k iΩ kiΩ kφ ⎝ ki Ω ⎠
= =1
J −ωc2Ω J ωc2Ω

⎛ k pΩ ⎞
⎜ k k 1 + jωcΩ ⎟ k pΩ
∠ ⎜ iΩ φ
kiΩ ⎟ = arctan ⎜⎛ ω ⎞
⎟ − 180° = −180° + φ pm ,Ω
cΩ
⎜ J −ωc2Ω ⎟ ⎝ k iΩ ⎠
⎜ ⎟
⎝ ⎠
28
 Design of the Position Loop
θm ,ref (s ) + Ω m ,ref (s) Ω m (s ) 1 θm (s)
kθ 1
− s

θm (s)

kθ
Gθ ,OL ( s) =
s

• The speed loop is assumed to be ideal.

(corresponds to unity).
• Proportional gain kθ alone is adequate due to the presence
of a pure integrator.
• The 0 dB crossover frequency ωcθ is chosen

kθ
Gθ ,OL (j ωcθ ) = =1 kθ = ωcθ
ωcθ

29
 Further Issues
Q Feedforward: to improve dynamic response

30
 Further Issues
Q Effect of limits: nonlinearity, further delay

Q Antiwindup integration
• Integration is suspended
when the output
saturates.

31
Effects of Measurements and Converter
Unidealities

32
 Rectifier-Fed Drives

• Transfer function of a rectifier: gain KT and delay τT

1
GT ( s ) = KT e− sτ T e − sτ
=
( sτ )2
1 + sτ + +…
2

• Approximate model: • DC-DC chopper with a

high switching frequency:
KT
GT ( s ) ≈ GT ( s ) ≈ KT
1 + sτ T
33