Modeling: Continuous Systems

Arijit Mondal
Dept. of Computer Science & Engineering
Indian Institute of Technology Patna
[email protected]

IIT Patna 1
System modeling
• Mimic the real world behavior of the system
• There exist a large variety of systems
• Mechanical, electrical, chemical, biological, etc.
• Behavior of most of the system can be described using differential equations
• Continuous dynamics
• Modal models
• Used for modeling discrete systems
• For each mode, we have continuous dynamics
• Ordinary differential equation will be used to describe the system
• Properties like linearity, time invariance, stability, etc. will be considered

IIT Patna 2
Helicopter

Image source: Internet

IIT Patna 3
Helicopter

Image source: Internet

IIT Patna 4
Newtonian mechanics
• Motion of object can be represented with six degrees of freedom

IIT Patna 5
Newtonian mechanics
• Motion of object can be represented with six degrees of freedom
• Linear motion along x , y , z axis
• Angular motion θx (roll), θy (yaw), θz (pitch)

IIT Patna 5
Newtonian mechanics
• Motion of object can be represented with six degrees of freedom
• Linear motion along x , y , z axis
• Angular motion θx (roll), θy (yaw), θz (pitch)
• Position of an object may be specified using six functions f : R → R
• R denotes the time
• R denotes the distance along an axis or angle relative to an axis

IIT Patna 5
Newtonian mechanics
• Motion of object can be represented with six degrees of freedom
• Linear motion along x , y , z axis
• Angular motion θx (roll), θy (yaw), θz (pitch)
• Position of an object may be specified using six functions f : R → R
• R denotes the time
• R denotes the distance along an axis or angle relative to an axis
• Sometime this may be represented x : R → R3 , θ : R → R3

IIT Patna 5
Newtonian mechanics
• Motion of object can be represented with six degrees of freedom
• Linear motion along x , y , z axis
• Angular motion θx (roll), θy (yaw), θz (pitch)
• Position of an object may be specified using six functions f : R → R
• R denotes the time
• R denotes the distance along an axis or angle relative to an axis
• Sometime this may be represented x : R → R3 , θ : R → R3
• Change in position or orientation can be determined by Newton’s 2nd law

F (t) = M ẍ (t)
• F - force, M - mass and ẍ - second derivative ie. acceleration

IIT Patna 5
Newtonian mechanics (contd.)
Z t
• Solving the equation we get t > 0, ẋ (t) = ẋ (0) + ẍ (τ )dτ
0

IIT Patna 6
Newtonian mechanics (contd.)
Z t
• Solving the equation we get t > 0, ẋ (t) = ẋ (0) + ẍ (τ )dτ
0
Z t
• Rewriting, t > 0, ẋ (t) = ẋ (0) + 1
M
F (τ )dτ
0

IIT Patna 6
Newtonian mechanics (contd.)
Z t
• Solving the equation we get t > 0, ẋ (t) = ẋ (0) + ẍ (τ )dτ
0
Z t
• Rewriting, t > 0, ẋ (t) = ẋ (0) + 1
M
F (τ )dτ
0
Z t
• We have, x (t) = x (0) + ẋ (τ )dτ
0

IIT Patna 6
Newtonian mechanics (contd.)
Z t
• Solving the equation we get t > 0, ẋ (t) = ẋ (0) + ẍ (τ )dτ
0
Z t
• Rewriting, t > 0, ẋ (t) = ẋ (0) + 1
M
F (τ )dτ
0
Z t Z tZ τ
1
• We have, x (t) = x (0) + ẋ (τ )dτ = x (0) + t ẋ (0) + F (α) dα dτ
0 M 0 0

IIT Patna 6
Newtonian mechanics (contd.)
Z t
• Solving the equation we get t > 0, ẋ (t) = ẋ (0) + ẍ (τ )dτ
0
Z t
• Rewriting, t > 0, ẋ (t) = ẋ (0) + 1
M
F (τ )dτ
0
Z t Z tZ τ
1
• We have, x (t) = x (0) + ẋ (τ )dτ = x (0) + t ẋ (0) + F (α) dα dτ
0 M 0 0

• Rotational version of force is torque T (t) = dt I (t)θ̇(t)

 
d

• I - Moment of inertia (depends on the geometry and orientation)

IIT Patna 6
Newtonian mechanics (contd.)
Z t
• Solving the equation we get t > 0, ẋ (t) = ẋ (0) + ẍ (τ )dτ
0
Z t
• Rewriting, t > 0, ẋ (t) = ẋ (0) + 1
M
F (τ )dτ
0
Z t Z tZ τ
1
• We have, x (t) = x (0) + ẋ (τ )dτ = x (0) + t ẋ (0) + F (α) dα dτ
0 M 0 0

• Rotational version of force is torque T (t) = dt I (t)θ̇(t)

 
d

• I - Moment of inertia (depends on the geometry and orientation)

Z tZ τ
1
θ(t) = θ(0) + t θ̇(0) + T (α) dα dτ
I 0 0

IIT Patna 6
Helicopter model
• Helicopter has two rotors
• Main rotor to lift
• Tail rotor to counter balance spin
• Hence, we have
θ̈y (t) = Ty (t)/Iyy ⇒
Z t
1
θ̇y (t) = θ̇y (0) + Ty (τ ) dτ
Iyy 0

Image source: Internet

IIT Patna 7
Actor model
• Physical system can be described by input (force, torque) and output (position, orien-
tation, velocity, rotation, etc.)
x y
S

• Usually X is time (domain) and Y value of particular signal (codomain)

• S : X → Y , x, y ∈ R

IIT Patna 8
Actor model
• Physical system can be described by input (force, torque) and output (position, orien-
tation, velocity, rotation, etc.)
x y
S

• Usually X is time (domain) and Y value of particular signal (codomain)

• S : X → Y , x, y ∈ R

• Example
Ty θ̇y
Iyy , θ

IIT Patna 8
Actor model (contd.)
• Actor models are composable
x1 y1 x2 y2
S1 S2

∀t ∈ R, y1 (t) = x2 (t)

IIT Patna 9
Actor model (contd.)
• Actor models are composable
x1 y1 x2 y2
S1 S2

∀t ∈ R, y1 (t) = x2 (t)

• Example

x1 y1 x2 R y2
a i
Helicopter
Z t
• We have ∀t ∈ R y2 (t) = i + x2 (τ )dτ where a = 1/Iyy , i = θ̇y (0), x1 = Ty and y2 = θ̇y
0

IIT Patna 9
Actor model (contd.)
• Actor can have multiple inputs
x1
y1
S1
x2

• Another useful building block is signal adder

x1 y x1 y
+ +
x2 x2 −

• y (t) = x1 (t) + x2 (t), y (t) = x1 (t) − x2 (t)

IIT Patna 10
Properties of systems
• Causal system
• Memoryless systems
• Linear and time invariant
• Stability
• Feedback control

IIT Patna 11
Causal systems
• Output depends only on current and past inputs
• Consider a continuous time signal x
• Let x|t≤τ represent restriction in time defined only for t ≤ τ
• Consider a continuous time system S : X → Y , the system is causal if for all x1 , x2 ∈ X
and τ ∈ R, x1 |t≤τ = x2 |t≤τ ⇒ S(x1 )|t≤τ = S(x2 )|t≤τ
• Strictly causal ∀τ ∈ R, x1 |t<τ = x2 |t<τ ⇒ S(x1 )|t≤τ = S(x2 )|t≤τ

IIT Patna 12
Causal systems
• Output depends only on current and past inputs
• Consider a continuous time signal x
• Let x|t≤τ represent restriction in time defined only for t ≤ τ
• Consider a continuous time system S : X → Y , the system is causal if for all x1 , x2 ∈ X
and τ ∈ R, x1 |t≤τ = x2 |t≤τ ⇒ S(x1 )|t≤τ = S(x2 )|t≤τ
• Strictly causal ∀τ ∈ R, x1 |t<τ = x2 |t<τ ⇒ S(x1 )|t≤τ = S(x2 )|t≤τ
• Example
• Integrator is strictly causal
• Adder is not strictly causal but causal
• Strictly causal actors are good for continuous feedback system

IIT Patna 12
Memoryless systems
• A systems has memory if the output depends not only on the current inputs but also on
the past inputs
• Formally, S : X → Y the system is memoryless if there exist a function f : X → Y such
that for all x ∈ X , (S(x))(t) = f (x(t)) for all t ∈ R

IIT Patna 13
Memoryless systems
• A systems has memory if the output depends not only on the current inputs but also on
the past inputs
• Formally, S : X → Y the system is memoryless if there exist a function f : X → Y such
that for all x ∈ X , (S(x))(t) = f (x(t)) for all t ∈ R
• Example
• Integrator is not memoryless
• Adder is memoryless

IIT Patna 13
Linear and time invariant (LTI)
• A systems S : X → Y where X and Y are sets of signals is linear if it satisfies the
superposition property
∀x1 , x2 ∈ X and ∀a, b ∈ R S(ax1 + bx2 ) = aS(x1 ) + bS(x2 )

• Time invariance means that whether we apply an input to the system now or T seconds
from now, the output will be identical except for a time delay of T seconds.
• Let Dτ be the delay operator such that (Dτ (x))(t) = x(t − τ )
• For time invariance, S(Dτ (x)) = Dτ (S(x))

IIT Patna 14
Linear and time invariant (LTI)
• A systems S : X → Y where X and Y are sets of signals is linear if it satisfies the
superposition property
∀x1 , x2 ∈ X and ∀a, b ∈ R S(ax1 + bx2 ) = aS(x1 ) + bS(x2 )

• Time invariance means that whether we apply an input to the system now or T seconds
from now, the output will be identical except for a time delay of T seconds.
• Let Dτ be the delay operator such that (Dτ (x))(t) = x(t − τ )
• For time invariance, S(Dτ (x)) = Dτ (S(x))

Z t
1
• θ̇y (t) = Ty (τ )dτ - LTI
Iyy −∞
• Many systems are approximated to LTI

IIT Patna 14
Stability
• A system is bounded input bounded output stable if the output signal is bounded for all
inputs signals that are bounded
• Helicopter is unstable

IIT Patna 15
Feedback systems
• A system with feedback has directed cycle where an output from an actor is fed back to
affect an input of the same actor

x1 e x3 y
+ S1 S2

x2 −

IIT Patna 16
Example: No rotation
• Want to have 0 angular velocity

ψ e Ty θ̇y
+ K Iyy , θ̇(0)

IIT Patna 17
Example: No rotation (contd.)
• Our equation remains the same, only input has changed.

ψ e Ty θ̇y
+ K Iyy , θ̇(0)

IIT Patna 18
Example: No rotation (contd.)
• Our equation remains the same, only input has changed.
Z t Z t
1 1
• θ̇y (t) = θ̇y (0) + Ty (τ ) dτ = θ̇y (0) + (ψ(τ ) − θ̇y (τ )) dτ
Iyy 0 Iyy 0

ψ e Ty θ̇y
+ K Iyy , θ̇(0)

IIT Patna 18
Example: No rotation (contd.)
• Our equation remains the same, only input has changed.
Z t Z t
1 1
• θ̇y (t) = θ̇y (0) + Ty (τ ) dτ = θ̇y (0) + (ψ(τ ) − θ̇y (τ )) dτ
Iyy 0 Iyy 0
• We have, e(t) = ψ(t) − θ̇y (t), Ty (t) = Ke(t)

ψ e Ty θ̇y
+ K Iyy , θ̇(0)

IIT Patna 18
Example: No rotation (contd.)
• Our equation remains the same, only input has changed.
Z t Z t
1 1
• θ̇y (t) = θ̇y (0) + Ty (τ ) dτ = θ̇y (0) + (ψ(τ ) − θ̇y (τ )) dτ
Iyy 0 Iyy 0
• We have, e(t) = ψ(t) − θ̇y (t), Ty (t) = Ke(t)
Z t
K
• Reorganizing we get, θ̇y (t) = θ̇y (0) − θ̇y (t) dτ
Iyy 0

ψ e Ty θ̇y
+ K Iyy , θ̇(0)

IIT Patna 18
Example: No rotation (contd.)
• Our equation remains the same, only input has changed.
Z t Z t
1 1
• θ̇y (t) = θ̇y (0) + Ty (τ ) dτ = θ̇y (0) + (ψ(τ ) − θ̇y (τ )) dτ
Iyy 0 Iyy 0
• We have, e(t) = ψ(t) − θ̇y (t), Ty (t) = Ke(t)
Z t
K
• Reorganizing we get, θ̇y (t) = θ̇y (0) − θ̇y (t) dτ
Iyy 0
Z t
• We know, ae aτ dτ = e at u(t) − 1
0

ψ e Ty θ̇y
+ K Iyy , θ̇(0)

IIT Patna 18
Example: No rotation (contd.)
• Our equation remains the same, only input has changed.
Z t Z t
1 1
• θ̇y (t) = θ̇y (0) + Ty (τ ) dτ = θ̇y (0) + (ψ(τ ) − θ̇y (τ )) dτ
Iyy 0 Iyy 0
• We have, e(t) = ψ(t) − θ̇y (t), Ty (t) = Ke(t)
Z t
K
• Reorganizing we get, θ̇y (t) = θ̇y (0) − θ̇y (t) dτ
Iyy 0
Z t
• We know, ae aτ dτ = e at u(t) − 1
0
• Therefore we have, θ̇y (t) = θ̇y (0)e −Kt/Iyy u(t)
ψ e Ty θ̇y
+ K Iyy , θ̇(0)

IIT Patna 18
Example: Constant rotation
• Only input has changed. ψ(t) = au(t)

ψ e Ty θ̇y
+ K Iyy , θ̇(0)

IIT Patna 19
Example: Constant rotation
• Only input has changed. ψ(t) = au(t)
Z t Z t
1 K
θ̇y (t) = Ty (τ )dτ = (ψ(t) − θ̇y (t))dτ
Iyy 0 Iyy 0

ψ e Ty θ̇y
+ K Iyy , θ̇(0)

IIT Patna 19
Example: Constant rotation
• Only input has changed. ψ(t) = au(t)
Z t Z t
1 K
θ̇y (t) = Ty (τ )dτ = (ψ(t) − θ̇y (t))dτ
Iyy 0 Iyy 0
Z t Z t Z t
K K Kat K
= adτ − θ̇y (τ )dτ = − θ̇y (τ )dτ
Iyy 0 Iyy 0 Iyy Iyy 0

ψ e Ty θ̇y
+ K Iyy , θ̇(0)

IIT Patna 19
Example: Constant rotation
• Only input has changed. ψ(t) = au(t)
Z t Z t
1 K
θ̇y (t) = Ty (τ )dτ = (ψ(t) − θ̇y (t))dτ
Iyy 0 Iyy 0
Z t Z t Z t
K K Kat K
= adτ − θ̇y (τ )dτ = − θ̇y (τ )dτ
Iyy 0 Iyy 0 Iyy Iyy 0

• θ̇y (t) = au(t)(1 − e −Kt/Iyy )

ψ e Ty θ̇y
+ K Iyy , θ̇(0)

IIT Patna 19