Lecture 4: Pulse-Width Modulation and Current Control

ELEC-E8402 Control of Electric Drives and Power Converters

Marko Hinkkanen
Spring 2020

1 / 25
Learning Outcomes

After this lecture and exercises you will be able to:

I Explain the difference between the standard suboscillation PWM method and
the symmetrical suboscillation PWM method
I Explain the principle of three-phase synchronous-frame current control
I Understand operation of the current controller in Assignment 1

The current controller is presented here for induction motors, but it can be easily
extended for other AC machines and for grid converters (equipped with L filter).

2 / 25
3-Phase Inverter

Pulse-Width Modulation

Current Control

Anti-Windup, Sampling, PWM Update

3 / 25
3-Phase Inverter

Leg a Leg b Leg c

Udc

a b c

4 / 25
Space Vector of the Converter Output Voltages
ia a
ib b
Udc n
ic c
N

I Zero-sequence voltage does not affect the phase currents

I Reference potential of the phase voltages can be freely chosen

2 
uss = uan + ubn ej2π/3 + ucn ej4π/3 Neutral n as a reference
3
2 
= uaN + ubN ej2π/3 + ucN ej4π/3 Negative DC bus N as a reference
3

5 / 25
I Converter output voltage vector β
2 
uss = uaN + ubN ej2π/3 + ucN ej4π/3 (0, 1, 0) (1, 1, 0)
3
2 Udc
√

= qa + qb ej2π/3 + qc ej4π/3 Udc 3
3
where qabc are the switching states (0, 1, 1) (0, 0, 0) 2Udc /3 (1, 0, 0)
(either 0 or 1) (1, 1, 1) α
I Vector (1, 0, 0) as an example

2Udc
uss =
3
(0, 0, 1) (1, 0, 1)

6 / 25
Switching-Cycle Averaged Voltage

I Using PWM, any voltage vector inside the voltage hexagon can be produced
in average over the switching period
2 
uss = da + db ej2π/3 + dc ej4π/3 Udc
3
where dabc are the duty ratios (between 0. . . 1)
√
I Maximum magnitude of the voltage vector is umax = Udc / 3 in linear
modulation (the circle inside the hexagon)
I PWM can be implemented, e.g., using the carrier comparison
I Mainly switching-cycle averaged quantities will be needed in this course
(overlining will be omitted for simplicity)

7 / 25
3-Phase Inverter

Pulse-Width Modulation

Current Control

Anti-Windup, Sampling, PWM Update

8 / 25
Suboscillation Method

qa carrier
I Duty ratios for carrier comparison 1
da
1 ua,ref 1 ub,ref 1 uc,ref
da = + db = + dc = + 0
2 Udc 2 Udc 2 Udc t
qb
I Voltage vectors during Tsw in the example: 1
(0, 0, 0) → (0, 1, 0) → (1, 1, 0) → (1, 1, 1) → db
(1, 1, 1) → (1, 1, 0) → (0, 1, 0) → (0, 0, 0)
0
I Problem: only 87% of the maximum t
available voltage can be used! qc
1
I Proper zero-sequence component should be
dc
added to utilize all available voltage
0
Tsw t

9 / 25
Symmetrical Suboscillation Method
I Zero sequence

min(ua,ref , ub,ref , uc,ref ) + max(ua,ref , ub,ref , uc,ref )

u0 =
2
I Modified voltage references

u0a,ref = ua,ref − u0 u0b,ref = ub,ref − u0 u0c,ref = uc,ref − u0

I Duty ratios for carrier comparison

0 0 0
1 ua,ref 1 ub,ref 1 uc,ref
da = + db = + dc = +
2 Udc 2 Udc 2 Udc
I Whole voltage hexagon can be now used
I Following example: ua,ref = uc,ref = −Udc /4 and ub,ref = Udc /2
The symmetrical suboscillation method is identical to the continuous space-vector PWM.
10 / 25
 carrier qa carrier qa
1 1
da da
0 0
t t
qb db qb
1 1
db

0 0
t t
qc qc
1 1
dc dc
0 0
Tsw t Tsw t

Suboscillation method Symmetrical suboscillation method

11 / 25
3-Phase Inverter

Pulse-Width Modulation

Current Control

Anti-Windup, Sampling, PWM Update

12 / 25
Current Controller in Stator Coordinates
ua,ref , ub,ref , uc,ref
iss,ref s
Current us,ref αβ
controller PWM
abc

iss αβ ia , ib , ic
abc

I PI controller cannot give zero steady-state error for sinusoidal references

I Actual current does not follow its reference in the steady state
(phase shift and magnitude error)
13 / 25
Current Controller in Synchronous Coordinates

ua,ref , ub,ref , uc,ref

is,ref
Current us,ref dq
controller PWM
abc

is dq ia , ib , ic
abc

ϑs
M

I DC signals in steady state, no steady-state error

I Synchronous-frame two-degree-of-freedom PI control will be considered

14 / 25
Review: Electrical Dynamics of the DC Motor and SPM

DC motor SPM in stator coordinates

ia Ra La iss Rs Ls

ua ea uss ess

dia diss
La = ua − Ra i a − ea Ls = uss − Rs iss − ess
dt dt
Back-emf: ea = kf ωm Back-emf: ess = jωm ψf ejϑm

I Back-emf is a quasi-constant disturbance for the current controller

The DC motor and the surface-mounted permanent-magnet motor (SPM) were considered in the prerequisite course ELEC-E8405 Electric Drives.
15 / 25
Electrical Dynamics of the Induction Motor

I State-space representation in stator coordinates

I Stator current and rotor flux as state variables

diss
 
s s RR
Lσ = us − Rσ is − jωm − ψ sR Rσ Lσ
dt LM iss
| {z }
back-emf ess
dψ sR

RR
 uss ess
s
= RR is + jωm − ψ sR
dt LM

where Rσ = Rs + RR
I Rotor-flux magnitude and speed change slowly
as compared to the stator current

16 / 25
General 3-Phase Load

I Differential equation in synchronous coordinates

dis
L + jωs Lis = us − Ris − es
dt iss R L

General SPM Induction motor

uss ess
L Ls Lσ
R Rs  R σ 
es jωm ψf jωm − LRM
R
ψR
ωs ωm ωs

17 / 25
Transfer Function in Synchronous Coordinates

I Transfer function of the system to be controlled

is (s) 1
= Y (s) = es
us (s) (s + jωs )L + R
us is
I Back emf es can be seen as a load disturbance Y (s)
I Similar to the transfer function of the DC motor
I Ideal converter and PWM will be assumed
(us = us,ref )

18 / 25
Cross-Coupling Compensation and Active Resistance

dis
L = us − (R + jωs L)is − es
dt
I Term jωs Lis causes the cross-coupling between the axes

did
L = ud − Rid + ωs Liq − ed
dt
diq
L = uq − Riq − ωs Lid − eq
dt
I Cross-coupling is cancelled by adding jωs L̂is to the controller output
I Active resistance r improves disturbance-rejection capability

us = u0s + (jωs L̂ − r)is

19 / 25
Control Algorithm
Current controller Motor modelled in
synchronous coordinates
1
ki es
s

is,ref us,ref 1 is
kp
(s+jωs )L + R

jωs L̂ − r

dI
= is,ref − is
dt
us,ref = kp (is,ref − is ) + ki I + (jωs L̂ − r)is

20 / 25
Tuning

I PI controller sees the system (assuming L̂ = L)

dis is (s) 1
L = u0s − (R + r)is − es or 0
= Y 0 (s) =
dt us (s) sL + R + r
I PI controller and closed-loop transfer functions

ki is (s) K(s)Y 0 (s) αc

K(s) = kp + = 0
=
s is,ref (s) 1 + K(s)Y (s) s + αc

where αc is the desired closed-loop bandwidth

I Gain selection
r = αc L̂ − R kp = αc L̂ ki = αc2 L̂

This tuning principle is covered in more detail in connection with the DC motors in the course ELEC-E8405 Electric Drives.
21 / 25
3-Phase Inverter

Pulse-Width Modulation

Current Control

Anti-Windup, Sampling, PWM Update

22 / 25
Anti-Windup

I Maximum converter output voltage is limited: |us | < umax

I Reference |us,ref | may exceed umax for large current steps
(especially at high rotor speeds due to the large back-emf |es |)
I Realisable voltage vector us is obtained from the PWM algorithm
I Algorithm is augmented with the anti-windup

dI 1
= is,ref − is + (us − us,ref )
dt kp
us,ref = kp (is,ref − is ) + ki I + (jωs L̂ − r)is
us = PWM(us,ref , ϑs )

23 / 25
Back-Calculation Anti-Windup Method

Current controller
This signal is nonzero
1/kp only if the voltage saturates

1
ki es
s

is,ref us,ref us 1 is
kp
(s+jωs )L + R
Saturation
jωs L̂ − r

24 / 25
Discrete Implementation: Sampling and PWM Update

sample
currents
update
carrier references

da (k) qa da (k) qa
k k+1 k+2 t/Ts k k+1 k+2 k+3 k+4 t/Ts
Tsw Tsw

Single-update PWM Double-update PWM

I No switching ripple in the current samples due to synchronous sampling

I Duty ratios db and dc are updated simultaneously with da
Study the implementation in Assignment 1, where the double-update PWM scheme is used. For more advanced designs, see (Awan, Saarakkala,
and Hinkkanen, “Flux-linkage-based current control of saturated synchronous motors,” IEEE Trans. Ind. Appl., 2019) and references therein.
25 / 25