Professional Documents
Culture Documents
Brief recap
We did...
Michela Mulas
Disturbance
Manipulated Controlled
Reference Error Variable
Variable
Controller System
Measurement
We did... We did...
I A system is an “object” in which variables of different kinds interact and produce I A system is an “object” in which variables of different kinds interact and produce
observable signals. observable signals.
I If the system is dynamic these variables (or signals) evolve with time. I If the system is dynamic these variables (or signals) evolve with time.
u1 (t) y1 (t)
Today’s lecture is about ...
.. ..
Input-output representation . S .
I Describe dynamic systems ur (t) yp (t)
I Describe and understand mathematical models
Inputs (causes) Outputs (effects)
I Define the main properties of a mathematical model
I The inputs (external causes) are developed outside the system, their evolution
Reading list influence the system behaviour.
They are the manipulated variables and the disturbances.
Nise, Control Systems Engineering (6th Edition) I The outputs (effects) depend on the inputs and on the nature of the system itself.
Ogata and Severo Engenharia de Controle Moderno (3th Ed., or newer), in I The system (S) can be considered as an operator that assigns a specific
Portuguese. behaviour to the outputs given every possible behaviour of the inputs.
I The output at any t > t0 depends on the values of q1 (t) and q2 (t) during the
interval [t0 , t].
I The state variable is the link between the inputs and the output.
The IO model for a SISO system is given by one differential equation as: The IO model for a MIMO system with p output and r inputs is given by p differential
equations as:
h y(t), ẏ(t), . . . , yn (t), u(t), u̇(t), . . . , um (t) = 0
| {z }| {z }
output input
h1 y1 (t), y˙1 (t), . . . , yn1 (t), u1 (t), u˙1 (t), . . . , um m
1 (t), . . . , ur (t), u˙r (t), . . . , ur (t), t = 0
| {z }| {z } | {z }
d (n) d n output 1 input 1 input r
I ẏ(t) = dt y(t), . . . , ẏ (t) = dt y (t);
(t), ˙2 (t), . . . , yn2 (t), u1 (t), u˙1 (t), . . . , um (t), . . . , (t), ˙r (t), . . . , um (t) , =0
h2 y 2 y 1 u r u r t
h is a function of parameter that depends on the system under study;
I
| {z }| {z } | {z }
output 2 input 1 input r
I n is the higher derivation order of the outputs and coincides with the system order. .. .
. ..
I m is the higher derivation order of the inputs.
hp yp (t), y˙p (t), . . . , ynp (t), u1 (t), u˙1 (t), . . . , um m
1 (t), . . . , ur (t), u˙r (t), . . . , ur (t), t = 0
}| {z } | {z }
Example:
| {z
√
output p input 1 input r
2ẏ(t)y(t) + 2 tu(t)ü(t) = 0
I hi , with i = 1, . . . , p are functions that depend on the system.
I ni is the maximum degree of differentiation of the output yi (t).
I mi is the maximum degree of differentiation of the input ui (t).
The state-variable model for a SISO system is given by n differential equations that The state-variable model for a SISO system is given by n differential equations that
link the derivative of each state variable to the input and the output to the state link the derivative of each state variable to the input and the output to the state
variables and the input: variables and the input:
ẋ(t) = f1 (x1 , . . . , xn (t), u(t), t) Given ẋ as
ẋ(t)
.. ..
d
.
. . ẋ(t) = x(t) = ..
dt
ẋ (t) = fn (x1 , . . . , xn (t), u(t), t)
n ẋn (t)
y(t) = g(x1 , . . . , xn (t), u(t), t)
we can simply write the system as:
(
I fi , i = 1, . . . , n and g are functions of different parameters given by the dynamics of ẋ(t) = f(x(t), u(t), t)
the system under study. y(t) = g(x(t), u(t), t)
q 3 = k · h2 q 3 = k · h2
By definition, y(t) = d(t) = h1 (t) − h2 (t): The SV model is given by:
1
ẏ(t) = ḋ(t) = ḣ1 (t) − ḣ2 (t) = q1 (t) − 2q2 (t) + k · h2 (t) ẋ1 (t) = A · u1 (t) − A · u2 (t)
A ẋ2 (t) = −k · x2 (t) + A · u2 (t)
1
= u1 (t) − 2u2 (t) + k(h1 (t) − y(t)) y(t) = x1 (t) − x2 (t)
A
I x1 (t) = h1 (t) and x2 (t) = h2 (t) are the state-variables;
Taking the second derivative and rearranging, the IO model is:
I u1 (t) = q1 (t) and u2 (t) = q2 (t) are the inputs;
k 1 2 k k
ÿ(t) + ẏ(t) = u̇1 (t) − u̇2 (t) + 2 u1 (t) − 2 u2 (t)
A A A A A I y(t) = d(t) is the output.
h(y(t), u(t), t) = 0 The linear system exhibits property of additivity and satisfies the property of
homogeneity.
Analogously, for a MIMO system with r inputs and p outputs: I Nonlinear, otherwise.
h1 (y1 (t), u1 (t), . . . , ur (t), t) = 0 An IO model is said to be linear if and only if the input-output relation is given by a
h2 (y2 (t), u1 (t), . . . , ur (t), t) = 0
linear differential equation:
..
. a0 (t)y(t) + a1 ẏ(t) + · · · + an (t)y(n) (t) = b0 (t)u(t) + b1 ẏ(t) + · · · + bm (t)y(n) (t)
hp (yp (t), u1 (t), . . . , ur (t), t) = 0
where the linear combinations of the IO coefficients are functions of time.
y(t) = u(t) + 1
I Stationary (or time-invariant) if it obeys to the cause-effect translation
principle in time.
The system responds always with the same effect to the same cause,
Is this a linear or non-linear system? regardless of the time.
It is a non-linear system because the equation is a non-linear algebraic equation where I Non stationary (or time-variant), otherwise
the non-linearity is given by the term +1.
A state-variable system is said to be time-invariant if and only if the state equation and
the output transformation equations are time independent:
(
ẋ(t) = f(x(t), u(t))
y(t) = g(x(t), u(t))
An IO model for a SISO system is proper if and only if in the IO relationship: A SV model described by: (
ẋ(t) = f(x(t), u(t), t)
h( y(t), ẏ(t), . . . , y(n) , u(t), u̇, . . . , um (t), t = 0 y(t) = g(x(t), u(t), t)
the output order is greater or equal to the input order (n ≥ m). is always a proper system.
I The system is strictly proper if y(t) = g(x(t), t).
If n > m the system is said to be strictly proper.
I The SV model for a linear and stationary strictly proper system becomes:
A MIMO system is said to be proper if ni ≥ maxj=1,...,r mi,j .
(
ẋ(t) = Ax(t) + Bu(t)
y(t) = Cx(t)
For a SV model,
I the state vector of a lumped parameter system has a finite number of components;
I the state vector of a distributed parameter system has an infinite number of
components.
Exercises Exercises
Exercise L1E2: Given the following mathematical models of dynamic systems: Exercise L1E3: Two tanks have a cross-sectional area A1 and A2 [m2 ], respectively,
and are arranged as shown in the Figure below.
ÿ(t) + y(t) = 5u̇(t)u(t)
The liquid heights on the two tanks are, respectively, h1 (t) and h2 (t) [m] and v1 (t) and
t2 ÿ(t) + tẏ(t) + y(t) = 5sin(t)ü(t) − 1 v2 (t) are the liquid volumes in the tanks.
" # " #" # " #
ẋ1 (t) −2 t2 x1 (t) 1
= + u(t)
The first tank is fed by a flow-rate q(t) [m3 /s] and
ẋ (t) −1 x2 (t)
2 0 1
h
"
i x (t)
# has an output flow-rate given by q1 (t) = K1 h1 (t)
1 [m3 /s].
y(t) = 2 1 x (t) + 3u(t)
2
The first tank feeds the second one whose
y(t) = u̇(t − T) output flow-rate is given by
1. Classify the models as IO or SV, defining the significant parameters such as q2 (t) = K2 h2 (t) [m3 /s]
derivation order, dimension of the state vector, inputs and outputs.
2. Define the properties of the models: linear or non-linear, stationary or time-variant, The mass conservation law for an
dynamic or instantaneous, with or without time delay, proper (strictly or not) or incompressible fluid states that the derivatives
improper. Discuss your answers. of the liquid fluid v(t) in the tank is given:
Exercises
Exercise L1E2: Two tanks have a cross-sectional area A1 and A2 [m2 ], respectively,
and are arranged as shown in the Figure below.
The liquid heights on the two tanks are, respectively, h1 (t) and h2 (t) [m] and v1 (t) and
v2 (t) are the liquid volumes in the tanks.
It is required to:
1. Define a mathematical model for the
system where x1 (t) = v1 (t) and
x2 (t) = v2 (t) are the state variables,
u(t) = q(t) as input and y(t) = h2 (t) as
output. Define the A, B, C, D matrixes.
2. Define the general properties of the model.
3. Define the mathematical model as IO
model.