Professional Documents
Culture Documents
Lecture 5
Lecture 5
Lecture 5
Marko Hinkkanen
Autumn 2020
1 / 26
Learning Outcomes
2 / 26
Introduction
3 / 26
Example: Connection of a DC Voltage Source to the Terminals
ia Ra La
Udc ua ea = kf ωM
4 / 26
Outline
Simulation Examples
Time-Scale Separation
5 / 26
DC Motor Model
I Voltage equation
dia
La = ua − Ra ia − ea
dt
ia Ra La
where ea = kf ωM is the back emf
I Motion equation
ua ea
dωM
J = TM − TL
dt
where TM = kf ia is the electromagnetic torque
I For simplicity, the flux factor kf is assumed to be
constant in the following
6 / 26
Electrical and Mechanical Dynamics Are Coupled
Physical model
TM = kf ia Block diagram
ia Ra La ua ia
TL
ua ea = kf ωM ⇒ TL
DC motor
ωM
ωM
7 / 26
Electrical Dynamics in the Time Domain
In some textbooks, the symbol p = d/dt is used for the differential operator in the time domain.
8 / 26
Electrical Dynamics in the Laplace Domain
1 ea
ia (s) = [ua (s) − ea (s)] Ra
sLa + Ra
I Transfer function (admittance)
1 1/Ra
Ya (s) = = ua 1 ia
sLa + Ra 1 + τa s
sLa + Ra
where τa = La /Ra ea
In the Laplace domain, s = σ + jω is a complex variable. However, the differential operator and the Laplace variable can be used
interchangeably in many cases.
9 / 26
Useful Block Diagram Algebra
d
y(s) G1 (s)G2 (s) r y
= G1 (s) G2 (s)
r(s) 1 + G1 (s)G2 (s)H(s)
y(s) G2 (s)
=
d(s) 1 + G1 (s)G2 (s)H(s) H(s)
10 / 26
Block Diagram of the DC Motor
dia
La = −Ra ia − kf ωM + ua
dt
dωM
J = kf ia − TL
dt
TL
ua 1 1 ia TM 1 1 ωM
kf
La s J s
Ra
ea
kf
ea
kf
I Armature current depends on the armature voltage and the load torque
I Could you derive the transfer functions based on the block diagram?
12 / 26
Transfer Function From ua (s) to ωM (s)
kf
JLa Kω02
Gωu (s) = =
Ra k2 s2 + 2ζω0 s + ω02
s2 + s+ f
La JLa
I Last form is a typical generic form of 2nd-order systems
I Undamped angular frequency, damping ratio, and DC gain
r
kf Ra J 1
ω0 = √ ζ= K=
JLa 2kf La kf
You don’t need to remember these more complex transfer functions, but practise deriving them based on the block diagram instead. However, you
should remember the generic form used above.
13 / 26
2nd-Order System in the Time Domain: Step Response
y/K
ζ =0
I 2nd-order system 2
y(s) ζ = 0.2
G(s) = 1.5
u(s)
Kω02 ζ = 0.7
= 1
s2 + 2ζω0 s + ω02
ζ =1
I Response y(t) to the ζ =2
0.5
step input u(t) is shown
I No overshoot if ζ ≥ 1
0
0 π 2π 3π 4π
ω0 t (rad)
Step responses can be easily plotted using numerical simulation tools. If needed, an analytical solution could be obtained using the inverse
Laplace transformation.
14 / 26
2nd-Order System in the Frequency Domain
I 2nd-order system
Kω02
G(s) = |G(jω)|
s2 + 2ζω0 s + ω02
K ζ =0
I Consider a sinusoidal input 2.5 ζ = 0.2
2
u(t) = U sin(ωt)
1.5
I For ζ > 0, the output in ζ = 0.7
steady state is 1
ζ =1
0.5 ζ =2
y(t) = AU sin(ωt + φ)
0
where 0.01 0.1 1 10 100 ω/ω0
A = |G(jω)| φ = G(jω)
15 / 26
Transfer Function From ua (s) to ia (s)
s/La
Giu (s) =
Ra k2
s2 + s+ f
La JLa
I Characteristic polynomial remains the same
(holds also for other transfer functions of the system)
I Zero at s = 0 in this transfer function
I If J → ∞ (i.e. ωM is constant)
1
Giu (s) = = Ya (s)
sLa + Ra
16 / 26
State-Variable Form
I State-variable model consists of coupled 1st-order differential equations
I Derivatives dx/dt depend on the states x and the system input u
dx
= Ax + Bu
dt
y = Cx
I States x depend on the history, but not on the present values of the inputs
I Output y depends only on the states (in physical systems)
I State variables are typically associated with the energy storage
I Current i of an inductor (or its flux linkage ψ = Li)
I Voltage u of a capacitor (or its charge q = Cu)
I Speed v of a mass (or its momentum p = mv)
I Choice of state variables is not unique (as shown in the parenthesis above)
17 / 26
State-Variable Form of the DC Motor
" Ra #
− La − Lkfa 1
d ia ia L 0
= kf + a ua + 1 TL
dt ωM 0 ω M 0 − J
| {z } | J {z } | {z } | {z }
x A Bu BT
ia = 1 0 x ωM = 0 1 x
| {z } | {z }
Ci Cω
I Transfer functions of the system are unique, i.e. the state-variable form leads
to the previous transfer functions
I Poles of the transfer function are eigenvalues of the system matrix A
18 / 26
Outline
Simulation Examples
Time-Scale Separation
19 / 26
Time-Domain Simulation Examples
20 / 26
Voltage-Step Response
ia (A)
200
undesirable 100
J = 0.05 kgm2
50
0
0 0.02 0.04 0.06 0.08 0.1 t (s)
21 / 26
Load-Torque-Step Response
ia (A)
20
15
10
I Armature voltage is constant
5
(rated)
0
I Initially no-load condition 0 0.02 0.04 0.06 0.08 0.1 t (s)
I Rated load torque is applied at
ωM (rad/s)
t = 0.01 s
135
J = 0.05 kgm2
130
125
J = 0.005 kgm2
Simulation Examples
Time-Scale Separation
23 / 26
Time-Scale Separation
I When considering the slow mechanical dynamics, the quickly converging
electrical dynamics may be approximated with the DC gain
TL
ua 1 ia TM 1 ωM
kf
Ra sJ
ea
kf
I When considering the fast electrical dynamics, the slowly varying rotor speed
may be assumed to be constant
ua 1 ia
sLa + Ra
ea = constant
24 / 26
Reduced-Order Model for Slow Mechanical Dynamics
ia (A) First-order approximate model
200
Full-order model
150
I Response to the rated
100
voltage step
50
I Electrical dynamics are
0
approximated with the 0 0.02 0.04 0.06 0.08 0.1 t (s)
steady-state gain
ωM (rad/s)
I Response of the
150
reduced-order model is
close to the full-order model 100
50
0
0 0.02 0.04 0.06 0.08 0.1 t (s)
25 / 26
Reduced-Order Model for Fast Electrical Dynamics
ia (A) First-order approximate model
I Response to the rated 200
Full-order model
voltage step 150
26 / 26