Automatic Control
System Modeling and Representation
23-Sep-16
Control System Design Process
Rule of Thumb
Dot forget to choose
proper
1. Sensors!!
2. Mechanical Sys.
3. Electrical Sys.
control theories
&
analysis techniques
Aims for this chapter
1. Recognize that differential equations can describe the dynamic behavior
of physical systems.
2. Be able to utilize linearization approximations through the use of Taylor
series expansions.
3. Understand the application of Laplace transforms and their role in
obtaining transfer functions.
4. Dynamic behaviors modelling
5. Be aware of block diagrams (and signal-flow graphs) and their role in
analyzing control systems.
6. Transfer function and state space form of the dynamic system.
System Description
Mathematical models of physical systems are key elements in the
design and analysis of control systems.
The dynamic behavior is generally described by ordinary differential
equations.
Therefore, for a given physical system, one must point out the final
design goal and derive its dynamic behavior before the controller
design.
One must understand
what the target you are trying to control.
what the nature responses are.
what the possible constraints/limitations are.
System Description
Dynamic modelling procedure
1. Define the system and its components.
2. Formulate the mathematical model
3. Make fundamental necessary assumptions based on basic principles.
3. Derive the differential equations representing the mathematical
model.
4. Solve the equations for the desired output variables.
5. Examine the solutions and the assumptions.
6. If necessary, re-analyze or redesign the system.
Modeling of Mechanical Systems
Mechanical have three passive, linear components. Two of them,
the spring and the mass, are energy-storage elements; one of them,
the viscous damper, dissipates energy.
Modeling of Mechanical Systems
Consider the following simple spring systems
Based on Hooks law, one has
For parallel spring configuration, the resulting equivalent spring
constant becomes larger.
Modeling of Mechanical Systems
For n degree of freedom parallel spring configuration, the total force
is
F = F1 + F2 + F3 + + Fn
= k1 x + k2 x + k3 x + + ki x, i = 1,, n
one can derive
F = (k1 + k2 + k3 + + ki )x
= keq x
It is easy to show that the equivalent spring constant is
keq = k1 + k2 + k3 + + ki
Modeling of Mechanical Systems
Consider the following simple spring systems
Based on Hooks law, one has
For serial springs configuration, the resulting equivalent spring
constant becomes smaller.
Modeling of Mechanical Systems
For n degree of freedom serial spring configuration, the force
distribution for each spring is the same, thus one can derive
F = k1 x1 = k2 x2 = k3 x3 = = ki xi , i = 1,, n
x1 =
F
F
F
, x2 = , x3 = ,
k2
k3
k1
F F F
F
+ + +
kn
k1 k2 k3
And the total displacement is x = x1 + x2 + x3 + =
It is easy to show that the equivalent spring constant is
F=
1
x = keq x
1 1 1
1
+ + +
kn
k1 k2 k3
keq =
1
1 1 1
1
+ + +
kn
k1 k2 k3
Modeling of Mechanical Systems
Any relative motion between the piston rod and the cylinder is
resisted by the oil because the oil must flow around the piston from
one side of the piston to the other.
Let us obtain the equivalent viscous-friction coefficient beq for each
of the parallel damper systems shown as follows
The force f due to the dampers is
It is usually referred to as damping force (velocity relevant).
Modeling of Mechanical Systems
Let us obtain the equivalent viscous-friction coefficient beq for each
of the serial damper systems shown as follows
Again since the force distribution for each damper is the same, it
gives
In terms of the equivalent viscous-friction coefficient, the damping
force is given by
Modeling of Mechanical Systems
For mass (translational) system, Newtons second law states that
x (position)
v (velocity)
a (acceleration)
where
F = u-f is the resultant force acting on the mass in the direction of the
acceleration a.
u is the applied force
f is other external forces.
Modeling of Mechanical Systems
Here is a standard mass-damper-spring system
mass displacement
The governing equation is
damping force
spring force
my + by + ky = u
my = F = u f = u (by + ky )
Modeling of Mechanical Systems
Consider a mass-damper-spring system attaching on a massless cart.
Note that u is the displacement
Not the applied force!!
Applying Newtons second law to the present system and noting
that the cart is massless
Modeling of Mechanical Systems
Consider the higher degree of freedom mass-damper-spring system
The equations of motion for the system are
Or
Modeling of Mechanical Systems
Practical Control Engineering
A balance system is a mechanical system in which the center of mass
is balanced above a pivot point.
Lets derive the mathematical model!
Modeling of Mechanical Systems
Consider an inverted pendulum system
(xG,yG)
Define the angle of the rod from the vertical line as .
Modeling of Mechanical Systems
To derive the equations of motion for the
system, consider the free-body diagram
For the rod dynamics, it contains two
governing equations:
rotational dynamics
translational dynamics
(xG,yG)
The rotational dynamics (about its center of
gravity) can be described by
applied torque
CW:positive direction
The translational dynamics can be described by
Horizontal:
Vertical:
H: horizontal force
V: vertical force
Modeling of Mechanical Systems
For the cart dynamics, it contains only one
governing equation: translational dynamics
The translational dynamics can be described by
CW:positive direction
(xG,yG)
Therefore, the overall system dynamics are
Rod Dynamics
Cart Dynamics
H: horizontal force
V: vertical force
Modeling of Mechanical Systems
It is clear that the rod dynamics are nonlinear.
To get linear model, system linearization is
applied by considering
CW:positive direction
(xG,yG)
One can obtain
Rod
Dynamics
Cart
Dynamics
H & V removal
Laplace Transform
Up to the present, we have illustrated a couple of mechanical model
and represented the system in the form of ODE.
Now, we are going to transform the time-domain system
representation into a s-domain transfer function (TF) representation
via Laplace transform.
Why Laplace transform?
The transfer function blocks can be organized into block diagrams or
signal-flow graphs to graphically depict the interconnections.
Block diagrams (and signal-flow graphs) are very convenient and natural
tools for designing and analyzing complicated control systems
Open-loop
Closed-loop
Laplace Transform
Let us define
casual system
frequency domain
Then the Laplace transform of f(t) is given by
The inverse Laplace transformation can is give by
Laplace Transform
Transfer Function
In control theory, functions called transfer functions are commonly used to
characterize the input-output relationships of components or systems that
can be described by linear, time-invariant, differential equations.
In other words, the applicability of the concept of the transfer function is
limited to
A. linear
B. time-invariant
C. zero initial condition
There doesnt exist TF for nonlinear system.
Unless system linearization is applied.
differential equation systems
The transfer function of a linear, time-invariant, differential equation
system is defined as the ratio of the Laplace transform of the output
(response function) to the Laplace transform of the input (driving function)
under the assumption that all initial conditions are zero.
Transfer Function
Consider the linear time-invariant system defined by the following
differential equation
where y is the output of the system and x is the input.
The transfer function of this system is the ratio of the Laplace
transformed output to the Laplace transformed input when all initial
conditions are zero
Modeling of Mechanical Systems
Here is a standard mass-damper-spring system
mass displacement
dynamic equation
my + by + ky = u
Taking the Laplace transform gives
The transfer function is
L ( y) = s 2Y (s ) sY (0 ) Y (0 ) = s 2Y (s )
L ( y ) = sY (s ) Y (0 ) = sY (s )
L ( y ) = Y (s )
L (u ) = U (s )
Y (s ) ms 2 + bs + k = U (s )
1
Y (s )
= 2
U (s ) ms + bs + k
Transfer Function
Consider again the a mass-damper-spring system
dynamic equation
my + by + ky = bu + ku
Taking the Laplace transform gives
The transfer function is
L ( y) = s 2Y (s ) sY (0 ) Y (0 ) = s 2Y (s )
L ( y ) = sY (s ) Y (0 ) = sY (s )
L ( y ) = Y (s )
L (u ) = sU (s ) U (0 ) = sU (s )
L (u ) = U (s )
Transfer Function
Consider again the higher degree of freedom mass-damper-spring
systemapplied
output1
output2
dynamic equation
force
Taking the Laplace transform gives
Suppose the input is U(s) and outputs are X1(s) and X2(s),
respectively, we have
Transfer Function
The linearized inverted pendulum model
Based on the dynamic equations, one has
m 2l 2
ml
2
I + ml
mgl =
u
(
)
(
)
+
M
m
M
+
m
From the control point of view, we are going
to find the input-output relationship, where
the input is u and the output is .
( s )
ml
=
U (s ) (MI + Mml 2 + mI )s 2 (M + m )mgl
=
1
if I = 0
2
Mls (M + m )g
I=0
Transfer Function
For the translational TF
Taking the Laplace transform gives
(M + m )X (s )s 2 + ml(s )s 2 = U (s )
(I + ml 2 )(s )s 2 + mlX (s )s 2 = mgl(s )
One can derive
I=0
mlX (s )s 2
( s ) =
I + ml 2 s 2 mgl
Then the input U(s) to output X(s) is
)
)
X (s )
I + ml 2 s 2 mgl
=
U (s ) MI + Mml 2 + mI s 4 (M + m )mgls 2
ml 2 s 2 mgl
=
if I = 0
Mml 2 s 4 (M + m )mgls 2
Modeling of Electrical Systems
Basic laws governing electrical circuits are
Ohm's law (V=IxR)
Kirchhoffs current law
Kirchhoffs voltage law.
Kirchhoffs current law (i.e., node law) states that the sum of currents
entering a node is equal to the sum of currents leaving the same node.
Kirchhoffs voltage law (i.e., loop law) states that the sum of the voltage
drops is equal to the sum of the voltage rises around a loop.
ik = 0
k =1
i1 + ( i2 ) + ( i3 ) + i4 = 0
i1 + i4 = i2 + i3
vk = 0
k =1
( v1 ) + ( v2 ) + ( v3 ) + (v4 ) = 0
v4 = v1 + v2 + v3
Modeling of Electrical Systems
Property of the basic electrical components:
Voltage-current, voltage-charge, and impedance relationships for
capacitors, resistors, and inductors
Together with the use of the Kirchhoffs, the above table is very
useful for deriving electrical dynamic equation.
Modeling of Electrical Systems
Consider a simple RC (low-pass filter) circuit consists of a resistor in
series with a load, and a capacitor in parallel with the load.
i1
i2
According to Kirchhoff's laws, it gives that i1 = i2.
The governing equation is v v
dvout
in
out
R
=C
dt
Taking the Laplace transform gives
Vout (s )
1
=
Vin (s ) RCs + 1
Modeling of Electrical Systems
Consider another RC (high-pass filter) circuit consists of
a capacitor in series with a load, and a resistor in parallel with the
load:
i1
i2
According to Kirchhoff's laws, it gives that i1 = i2.
The governing equation is
d (vout vin ) vout
C
dt
Taking the Laplace transform gives
Vout (s )
RCs
=
Vin (s ) RCs + 1
Modeling of Electrical Systems
Consider the electrical RLC circuit.
Input: applied voltage
Output : dropped voltage
The circuit consists of an inductance L (henry), a resistance R (ohm),
and a capacitance C (farad).
Applying Kirchhoffs voltage law to the system, one can obtain the
following equations:
dynamic equation
Modeling of Electrical Systems
For the RLC dynamic equation, assuming zero initial conditions, the
transfer function of the RLC circuit can be represented by
I (s ) = CsEo (s )
As a result, one can obtain the TF
One can transform the s-domain TF back into time domain ordinary
differential equation
This is the standard 2nd order RLC oscillation circuit.
Modeling of Electrical Systems
Consider an RLC parallel circuit. Find the TF between current and the
dropped voltage on the capacity.
Apply the Kirchhoffs current law to the system, one can obtain the
following equations:
is = iR + iL + iC
i=
vR 1 t
dv
+ 0 vL d + C c
R L
dt
Moreover, since vR = vL = vC = vout
one has
1 t
v
dv
i = out + 0 vout d + C out
R L
dt
Thus
1 1
I (s ) = + + Cs Vout (s )
R Ls
Vout (s )
RLs
=
I (s )
RLCs 2 + Ls + R
Modeling of Electrical Systems
Consider the cascade RC circuit, where ei is the input and eo is the
output.
Applying the Kirchhoffs current law leads to
Modeling of Electrical Systems
Based on the system equations, taking the Laplace transforms gives
As a result, the TF from Ei(s) to Eo(s) can be easily find
Modeling of Electrical Systems
Operational Amplifiers (OP), often called op amps, are frequently
used to amplify signals in actuator/sensor circuits.
Drivers for amplifying
control signals (OP amp inside)
1. Control algorithm realization
2. Command generation
Driving Force
Encoder Feedback
Modeling of Electrical Systems
Consider the simplest usage of OP amp.
K
The input-output relation to the amplifier is
where K denotes as a differential gain.
Note that the OP amp amplifies the difference in voltages e1 and e2 .
Such an amplifier is commonly called a differential amplifier.
In the following, we make the assumption that the OP amps are
ideal (ex. Infinity differential gain/bandwidth, zero bias current).
Modeling of Electrical Systems
Inverting Amplifier
Basic properties
Since the OP amp is supposed to be ideal, the current flows into the
amplifier is negligible. Therefore, we have
Based on the feature of OP amp, e' 0
It is clear that the input and output have the same signs, that is why
it is called non-inverting amplifier.
Modeling of Electrical Systems
Non-inverting Amplifier
K
Since
, the above equation can be approximated by
R1
R
eo eo = 1 + 2 ei
ei
R1
R1 + R2
It is clear that the input and output have the same signs, that is why
it is called non-inverting amplifier.
Modeling of Electrical Systems
Inverting 1st-order lag circuit
Basic properties
Noting that the current flowing into the amplifier is negligible, we
have
Again e' 0 , one has
Comparison of Mecanical/Electrical Systems
Laplace Transformation Table
Frequently used Laplace transformation table
Laplace Transformation Table
Laplace Transformation Table
Laplace Transformation Table
Inverse Laplace Transformation Practice
From the previous examples, it is obviously that
for higher order TF, to find the inverse LT, we must convert the function
of a summation of simpler terms for which the LT of each term is known
(or can be easily found via Look-Up table)
Let a TF by
F (s ) =
N (s )
D (s )
where N(s) is the numerator and D(s) is the denominator.
For the polynomial order
If N(s)>=D(s), then N(s) must be divided by D(s), and then take partialfraction expansion.
If N(s)<D(s), then take partial-fraction expansion directly.
Inverse Laplace Transformation Practice
Find the inverse Laplace transform of
The partial-fraction expansion of F(s) is
where a1 and a2 are found as
We have
Inverse Laplace Transformation Practice
Obtain the inverse Laplace transform of
Note that since the degree of the numerator polynomial is higher
than that of the denominator polynomial, it should be represented
by
Note that the LT of the unit-impulse function is 1, it follows the
following inverse LT
Inverse Laplace Transformation Practice
Find the inverse Laplace transform of
Notice that the denominator polynomial can be factored as
Recall that
One can obtain
Homework
Q1. Consider a 1-DOF automobile
suspension system, please derive the
transfer function
X o (s )
= G (s ), find G (s )
X i (s )
Q2. For the OP am RC circuit, please derive
the transfer function
Eo (s )
= G (s ), find G (s )
Ei (s )
