You are on page 1of 40

Lecture 2

- what de nes a dynamic model? -

1
fi

Increasing sophistication / value
Model development over time

Time

2 Source:PricewaterhouseCoopers Analysis
What this course is about...
• Introduction to dynamic systems theory
• Model + analyze + design of feedback systems
• Basic principles of feedback
• Goal: understand how to use these principles to develop
innovative technologies

Analysis + design of systems

Black box methodologies

Model-based methodologies










Model-based methodologies
• Use mathematical methods of addressing problems
• Analysis + design based on models
• A prediction of how the system will behave
• Feedback can lead to counter-intuitive behavior
• Help sort out what is going on

Today
• What are models?
• De ne concepts of state, dynamics, inputs and outputs
• Overview dynamic modeling techniques:
- di erential equations
- di erence equations

6
ff
ff
fi

Visualizing world

7
The world map

8
Research expenditure

9
Research employees

10
Research papers

11
Research growth

12
Population 1960

13

Population in 2050?

14
Predictions by the U.N.

15
What are models?

16
A simpli ed, quanti ed representation of a
system or process used to answer questions

17
fi
fi
How?
via mathematical analysis and simulation

What for?
to assist calculations and predictions

18



Models serve as a means of understanding the
mechanism of a process, predicting
relationships and outcomes, and inferring the
existence and role of [information in a system]
Je G. Bohn, Thinking Systematically About Policy,
IEEE Technology and Society Magazine. winter 2000/2001
ff

What are dynamic models?

20
Dynamical v. statistical modeling
• Statistical modeling focuses on how certain variable
correlate with other variables
➡ What in uences what?
• Dynamic modeling focuses on the structure, not
statistical techniques
• Tries to answer the “why” question by describing the
structure of the system
• “Causation across time” occurs because a variable's
derivative has been a ected instantaneously

21 courtesy from cortneybrown.com


fl

ff

Dynamical v. statistical modeling


• Statistical modeling focuses on how certain variable
correlate with other variables
➡ shows in uence
• Dynamic modeling focuses on the structure

independent
variables

dynamic modeling
statistical modeling

dependent
variables
22 courtesy from cortneybrown.com
fl


What is dynamic modeling?


• Think dynamically, not just what in uences what
• Not a statistical technique
• Tries to answer the “why” question by describing the
structure of the system
• “Causation across time” occurs because a variable's
derivative has been a ected instantaneously

23
ff

fl

Democritus (“Father of modern science”)

“I would rather discover one causal


relation than be the King of Persia”

Work
Ethic
..
Mathematic
..
Literatur
...
.

Albert Einstein

“Development of Western Science is based on two


great achievements, the invention of the formal
logical system (in Euclidean geometry) by the
Greek philosophers, and the discovery of the
possibility to nd out causal relationships by
systematic experiment (Renaissance)”

25
fi
The weather:
what causes precipitation?

26

Di erent questions → di erent models!

How much will it rain in the morning? Models don’t have to be perfect
Will it rain in the next 5-10 days? 27 → feedback provides robustness
ff
ff

The model you use depends on


the questions you want to answer

28
Terminology
1. State captures e ects of the past
• independent quantities
• determine future evolution
2. Inputs describe external excitation
• extrinsic to the system dynamics
3. Dynamics describe state evolution
• update rule for system state
• function of current state + inputs
4. Outputs describe measured quantities
• function of state + inputs (not independent variables)
• often subset of the state
29
ff

Modeling Properties
Choice of state is not unique
• many choices of variables can act as the state
Choice of inputs and outputs depend on point of view
• inputs: factors that are external to the model you
are building
• outputs: what variables can you measure:
- what you can sense
- what parts of the component model interact
with other component models

30

Types of models
• Ordinary di erential equations
• Di erence equations
• Discrete event
• Partial di erential equations
• Hybrid models
• Cellular automata

31
ff
ff
ff

Second order model


Example #1: Spring Mass System
u(t)
Applications
q2
! Flexible structures (many apps)
q1
! Suspension systems (eg, “Alice”)
m1 m2
! Molecular
Questions and quantum
we wantdynamics
to answer:
k1 k2
Questions we want to answer
k3
• How
! How much much
do masses domovemassesas a move as a
function of the forcing frequency?
b ! What frequency of thethe
happens if I change forcing
values force?
of the masses?
• What
! Will happens
Alice fly into the air ifif I change
I take that the
speed bump at 25 mph?
values of the masses
Modeling assumptions
! Mass, spring, and damper constants
Willand
• fixed
are it known
y into the air if I take a
speed
! Springs bump
satisfy atlaw
Hooke’s 30 km/h
! Damper is (linear) viscous force,
proportional to velocity

8 Oct 07 R. M. Murray, Caltech CDS 4

32
fl

Second order model


Example #1: Spring Mass System
Newton's Second Law of Motion
u(t)
Applications
q2
·· = ( − ) −
! Flexible structures (many apps)
q1
·· =systems
! Suspension
( −(eg,) −
“Alice”)
( − )− ·
m1 m2
! Molecular and quantum dynamics
Questions we want to answer
k1 k2 k3 Dynamics (state evolution)
! How much do masses move as a
function of the forcing frequency?·
b ! What happens if I change the values ·
of the masses?
! Will Alice· fly =
into the air if I (take−that) −
speed bump at 25 mph?
·
Modeling assumptions ( − )− ( − )− ·
! Mass, spring, and damper constants
are fixed and known

( viscous
)
! Springs satisfy Hooke’s law
=
! Damper is (linear) force,
=
𝟤
𝟤
𝟤
𝗆
𝗆
𝗆
proportional to velocity
𝟤
𝟤
𝟣
𝟤
𝗎
𝗊
𝗊
𝗊
𝗊
𝟤
𝟤
=
𝟥
𝖻
𝗄
𝗄
𝗊
4
∈ℝ: =·
𝟣
𝟣
𝟦
𝟤
𝖽
𝗍
𝗆
𝗆
𝟣
De ne state vector
𝗑
𝗊
𝟤
𝟣
𝟣
𝗊
𝗊
𝗊
𝗊
𝟤
𝟣
8 Oct 07 R. M. Murray, Caltech CDS 4
𝗄
𝗄

𝟥
𝟣
𝟤
𝖽
𝗑
𝗊
𝗊
𝟤
𝗊
𝟤
𝟣
𝗊
𝟤
𝟤
𝗊
𝗑
𝗊
𝗒
33
𝟣
𝟣
𝟤
𝟣
𝟤
𝟤
𝟥
𝟤
𝟤
𝟣
𝟤
𝟣
𝟤
𝟣
𝟣
𝟤
𝟣
𝗊
𝗑
𝗆
𝗆
𝗊
𝗊
𝗄
𝗄
𝗊
𝗎
𝗊
𝗊
𝗄
𝗄
𝗊
𝗊
𝗊
𝖻
𝗊
𝗊
𝟣
𝟣
𝗑
𝗊
fi
Second order model
Example #1: Spring Mass System
Newton's Second Law of Motion
u(t)
Applications
q2
·· = ( − ) −
! Flexible structures (many apps)
q1
·· =systems
! Suspension
( −(eg,) −
“Alice”)
( − )− ·
m1 m2
! Molecular and quantum dynamics
Questions we want to answer
k1 k2 k3 Dynamics (state evolution)
! How much do masses move as a
function of the forcing frequency?·
b ! What happens if I change the values ·
of the masses?
! Will Alice· fly =
into the air if I (take−that) −
speed bump at 25 mph?
·
Modeling assumptions ( − )− ( − )− ·
! Mass, spring, and damper constants
are fixed and known
! Springs satisfy Hooke’s law
! Damper is (linear) viscous force,
𝟤
𝟤
𝟤
𝗆
𝗆
𝗆
proportional to velocity
𝟤
𝟤
𝟣
𝟤
𝟤
𝟤
𝟤
=
𝗆
𝗆
𝗆
𝗎
𝗊
𝗊
𝗊
𝗊
( − )−
𝟤
𝟤
𝟣
𝟦
𝗎
𝗑
𝗑
𝗑
𝗑
𝟤
𝟤
𝟥
𝖻
𝗄
𝗄
𝗊
𝟦
𝟤
𝟥
𝖻
𝗄
𝗄
𝗑
𝟣
𝟣
𝖽
𝗍
𝗆
𝗆
𝟣
𝖽
𝗍
𝟤
𝟣
𝟣
𝟣
𝟣
𝟥
𝗊
𝗆
𝗆
𝗊
𝗊
𝗊
( − )− ( − )−
𝗑
𝟤
𝟣
𝟣
𝗑
𝗑
𝗑
𝟤
𝟣
8 Oct 07 R. M. Murray, Caltech CDS 4
𝗄
𝗄
𝟤
𝟣
𝟤
𝖽
𝟤
𝗄
𝗄
𝖽
𝗊
𝗑
𝟤
𝗊
𝟦
𝟣
𝟣
𝗑
𝗊
𝗑
34
𝟣
𝟣
𝟤
𝟣
𝟤
𝟤
𝟥
𝟤
𝟤
𝟣
𝟤
𝟣
𝟤
𝟣
𝟣
𝟤
𝟥
𝗊
𝗆
𝗆
𝗊
𝗊
𝗄
𝗄
𝗊
𝗎
𝗊
𝗊
𝗄
𝗄
𝗊
𝗊
𝗊
𝖻
𝗊
𝗑
Second order model
Example #1: Spring Mass System
Newton's Second Law of Motion
u(t)
Applications
q2
·· = ( − ) −
! Flexible structures (many apps)
q1
·· =systems
! Suspension
( −(eg,) −
“Alice”)
( − )− ·
m1 m2
! Molecular and quantum dynamics
Questions we want to answer
k1 k2 k3 Dynamics (state evolution)
! How much do masses move as a
function of the forcing frequency?·
b ! What happens if I change the values ·
of the masses?
! Will Alice· fly =
into the air if I (take−that) −
speed bump at 25 mph?
·
Modeling assumptions ( − )− ( − )− ·
! Mass, spring, and damper constants
are fixed and known
! Springs satisfy Hooke’s law
n
= ( , )
! Damper is (linear) viscous force, ∈ℝ , ∈ℝ
𝟤
𝟤
𝟤
𝗆
𝗆
𝗆
proportional to velocity
𝟤
𝟤
𝟣
𝟤
𝗎
𝗊
𝗊
𝗊
𝗊
𝟤
𝟤
𝟥
𝖻
𝗄
𝗄
𝗊
p
∈ℝ
𝟣
𝟣
= ( )
𝖽
𝗍
𝗆
𝗆
𝟣
𝟤
𝟣
𝟣
𝗊
𝗊
𝗊
𝗊
𝟤
𝟣
8 Oct 07 R. M. Murray, Caltech CDS 4
𝗄
𝗄
𝟤
𝖽
𝗊
𝟤
𝗊
𝖽
𝗍
𝟣
𝗊
𝖿
𝗑
𝗎
35
𝟣
𝟣
𝟤
𝟣
𝟤
𝟤
𝟥
𝟤
𝟤
𝟣
𝟤
𝟣
𝟤
𝟣
𝟣
𝟤
𝗊
𝗆
𝗆
𝗒
𝗑
𝗒
𝗎
𝗊
𝗊
𝗁
𝗑
𝗄
𝗄
𝗊
𝗎
𝗊
𝗊
𝗄
𝗄
𝗊
𝗊
𝗊
𝖻
𝗊
𝖽
𝗑
Electric circuit
coil L resistor R

+ + +
v v2 v1
- capacitor c2 - capacitor c1 -

States: variables that capture the energy storage mechanisms in the system

dx1
c1 = c1 ẋ1 : current through c1
dt
dx2
c2 = c2 ẋ2 : current through c2
dt
dx3
L = c3 ẋ3 : voltage though L
dt
<latexit sha1_base64="xEYRmd66N/dclWtG4h3wk2ggN9o=">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</latexit>

36

Electric circuit
coil L resistor R

+ + +
v v2 v1
- capacitor c2 - capacitor c1 -

Using Kirchho Law: Combining these equations we get:

x2 x1
c1 ẋ1 = 2 3 2 1 x2 x1 3 2 3
R ẋ1 c1 R 0
x2 x1 ẋ = 4 ẋ2 5 = 4 1
x3 x2 Rx1 5+4 0 5
c2 ẋ2 = x3 c2
1
R ẋ3 1
Lv
L x2
c3 ẋ3 = v x2
<latexit sha1_base64="Ww0Jr9W3GB7u0YCf+29tnraNTLY=">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</latexit>

<latexit sha1_base64="eNkmxl6zV00B1hfXOPh2trploUE=">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</latexit>

37

ff
Consider the di erential equation…
dy 2 dy
2
= 7 12y 2u
<latexit sha1_base64="qYkPtIGU/mhTtrs5fWg8X/PpvEM=">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</latexit>
dt dt

Writing it as the state space representation:

x1 = q
<latexit sha1_base64="AN4HdwcDh8bK2rZ0jvcHOr4xde8=">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</latexit>
x2 = dy
dt

ẋ1 = x2
<latexit sha1_base64="JP7siGwswso4bnBPpO+JTFCEpBQ=">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</latexit>
ẋ2 = 7x2 12x1 2u

 
0 1 0
ẋ = x+ u
<latexit sha1_base64="BECUstMWb36MMmMfU46vNBWyVBk=">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</latexit>
12 7 2

What happens if the di erential equation includes a derivative of the input?

38
ff
ff
d 2y dy du
2
= 7 12y 2u
<latexit sha1_base64="l18m1F3q/EWKGCPYNKK3fkJpAwg=">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</latexit>
dt dt dt

Take:

x1 = y
d 2y du
x2 = dy
dt + u
→ ẋ2 = +
<latexit sha1_base64="/JHnKyG+mheqKHhOFmpLn8p7Q1Q=">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</latexit>

<latexit sha1_base64="/nKrCyDdI9tPaQTKq8KWFU0b/kY=">AAACc3icbVDLihNBFK20rzG+Moqr2RQTBUHNdMeFuhgYEMTlCCYzkG5DdfXtpJh6NFW3dULRnzE7twPzC+JWP8IPcW91koWT8cKFc899HO7JKykcxvHvTnTt+o2bt7Zud+/cvXf/QW/74diZ2nIYcSONPc6ZAyk0jFCghOPKAlO5hKP85F3bP/oC1gmjP+GigkyxmRal4AwDNe3tpYVBf9pM/bDZT0vLuC8+B0wXjS8oNs/XHK1X9bTXjwfxMuhVkKxB/yB5//j7+cW3w+l250WQ4LUCjVwy5yZJXGHmmUXBJTTdtHZQMX7CZjAJUDMFLvPLzxr6NDAFLY0NqZEu2X83PFPOLVQeJhXDudvsteR/exJKFMXphjyWbzIvdFUjaL5SL2tJ0dDWO1oICxzlIgDGrQgPUD5nwR8MDl86z1VuxWyO4b6Gr9woxXSRamPVJMl82qqncgwW+0m6HExtWzXd4G+y6eZVMB4OkleD4cdg9Fuyii2yQ3bJM5KQ1+SAfCCHZEQ4OSM/yE/yq/Mn2ol2oyer0aiz3nlELkX08i9W4cSy</latexit>
dt dt

Then

ẋ1 = x2 u
du du
<latexit sha1_base64="hdwPOUXlL5kbFu+TyjzEB6j5cE4=">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</latexit>
ẋ2 = 7 (x2 u) 12x1 dt 2u + dt
= 7 (x2 u) 12x1 2u
=
<latexit sha1_base64="a2z11BrOYod0VbbDuKR6G/8UogQ=">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</latexit>
7x2 12x1 2u + 7u

39
Summary: system modeling

Model = state + inputs + outputs + dynamics

dx
= f (x, u) xk+1 = f (xk , uk )
dt
y = h(x) yk+1 = h(xk+1 )

Choice of model depends on questions you want answer!

40

You might also like