Module 12:
Non-Canonical Systems
Non-Canonical Systems
Rick Hill, PhD
Dept of Mechanical Engineering
Module 12:
Non-Canonical Systems
Time Response
Module 10: First-order systems
Module 11: Second-order systems
Non-Canonical Systems
Module 12: Non-canonical systems
1. Higher-order systems
2. Model reduction
3. Effect of zeros
4. Nonlinear systems
Higher-Order Systems
• Higher-order systems (more than two poles)
can be considered as the sum of 1st- and
2nd- order responses
• Exact time response can be determined by
Non-Canonical Systems
partial fraction expansion
• Ex: K
A
B
( s a)( s 2 2n s n2 ) sa s 2 2n s n2
• Higher-order systems can often be
approximated by 1st- or 2nd-order models
Model Reduction
Y ( s) 20 2 10
• Ex: G( s)
U ( s) ( s 2)( s 10) s 2 s 10
Non-Canonical Systems
10 2
s + 10 s+2
• If one pole is much slower than another, it
will dominate the dynamic response
Model Reduction
• Rule of thumb: the dynamics of a pole may
be neglected if its real part is 5 times greater
than the real part of the other poles
• Its steady-state contribution must still be
Non-Canonical Systems
accounted for though (replace by a static
gain)
Y ( s) 20
G( s) pole at -2 dominates
U ( s) ( s 2)( s 10) the pole at -10
Model Reduction
20 K
G( s) Gred ( s)
( s 2)( s 10) ( s 2)
• Need K so that G(s) and Gred(s) have same
steady-state value
Non-Canonical Systems
1
DC gain lims0 sY ( s) lim s 0 sG ( s) G (0)
s
• For G(s), DC gain = G(0) = 1
• To achieve Gred(0) = 1, need K = 2 →
2
Gred ( s)
( s 2)
Effect of Zeros
Zeros are values of s that make numerator equal 0
Interpretation #1
• Zeros affect the contribution of each pole,
Non-Canonical Systems
that is, the coefficients of the partial fraction
expansion
20 2.5 2.5
( s 2)( s 10) s 2 s 10
whereas
s 20 2.25 1.25
( s 2)( s 10) s 2 s 10
Effect of Zeros
Interpretation #2
s 1 s 1
2 2
s s 1 s s 1 s s 1
2
Non-Canonical Systems
derivative of system
system without zero without zero
d
step response
dt
Effect of Zeros
y(t)
Non-Canonical Systems
t
• Zeros tend to have the effect of increasing
overshoot and decreasing rise/peak time
Effect of Zeros
• If zero has positive real part, then the
response will initially move in the reverse
direction, called a nonminimum phase zero
Non-Canonical Systems
y(t)
t
Effect of Zeros
• An exception to the preceding rules is when
a zero is near a pole
If z = p, then pole and zero cancel
sz exactly and H (s) H (s)
Non-Canonical Systems
H ( s) H ( s)
s p If z ≈ p, then effect of pole and zero
are minimal and H (s) H (s)
• Example: s 1 1
( s 1.1)( s 10) s 10
Nonlinear Systems
• Most real world systems are nonlinear in some
respect
• Friction, air drag, saturation, backlash
• Nonlinear differential equations are difficult, if
Non-Canonical Systems
not impossible, to solve analytically
• Transfer functions model only linear systems
• Previously in the course we have numerically
simulated nonlinear systems to determine
behavior … difficult to design in this manner,
can lack insight
Nonlinear Systems
• Nonlinear functions can approximated by a
linear function in a neighborhood about an
operating point
df f ( x) f ( x )
dx x x x
df
f ( x) f ( x ) x
First two terms of a dx x x
Taylor series expansion
y k x
Nonlinear Systems
• This same idea can be
used for providing
approximate models of
nonlinear systems
Non-Canonical Systems
• Nonlinear diff eq Linear
diff eq in terms of deviation
from the operating point (Δ)
• These approximate models
are only valid in a small
neighborhood about the
y k x
operating point
Example
• Car with nonlinear air drag:
F bv 2 mv
• Replace each term with its linear
approximation:
f (v) bv 2 bv 2 2bv v
F F F , v v v
f(v) F
• Linear approx:
F 2bv v mv
Example (continued)
• Simulink model of linearized equation
Summary of Module 12
• Higher-order systems
• Effect of zeros
Non-Canonical Systems
• Model reduction
• Fast dynamics
• Pole-zero cancellation
• Linearization