Control systems

System Transient Performance

D. Jamal Siam
Linear Time invariant system in Laplace and time domains:

Poles and Zeros of LTI Systems:

Given the transfer function of a proper system i(primitive rational function):
𝛽𝑚 𝑠 𝑚 +𝛽𝑚−1 𝑠 𝑚−1 +⋯+𝛽1 𝑠 1 +𝛽0
𝑇 𝑠 = with 𝑚 < 𝑛
𝛼𝑛 𝑠 𝑛 +𝛼𝑛−1 𝑠 𝑛−1 +⋯+𝛼1 𝑠 1 +𝛼0
System Zeros:
A system zero is defined as the value 𝑠𝑧 at which 𝑇 𝑠𝑧 = 0.
A system zero can be a zero at finite or infinite.
A proper system has 𝑛 − 𝑚 zeros at infinite, that is those that satisfy the relation lim |𝑇 𝑠 | = 0 .
𝑠→∞
System Poles:
A system pole is defined as the value 𝑠𝑝 at which lim |𝑇 𝑠 | = ∞.
𝑠→𝑠𝑝
A system pole can be a pole at finite or infinite.
An improper system (improper: 𝑚 ≥ 𝑛) has 𝑚 − 𝑛 poles at infinite, that is those that satisfy the relation lim |𝑇 𝑠 | = ∞
𝑠→∞
Effect of poles on system response:
The poles number and locations (system transfer
function roots ) determine the shape and the time
performance of the transient response:
• Left-side poles generate a response that vanishes for
𝑡 → ∞, whereas right-side poles transient diverges
• Real-axe poles do not produce oscillation in the time
response.
• Imaginary axe poles produce an undamped
oscillation response.
• Complex poles produce oscillation in the response.
• Transient time performance depends on the relative distance
between the imaginary axe and the pole location. That is the
magnitude of the real part of the pole. Higher distance →
Higher performance and faster transient. The time constant
1
of a pole 𝑠𝑝 is defined as 𝜏𝑝 = − 𝑅𝑒(𝑠 )
𝑝
• Oscillation frequency depends on the relative distance
between the real axe and the pole location. That is the
magnitude of the imaginary part of the pole. Higher distance
→ Higher oscillation frequency and smaller period with higher
density of oscillation cycles.
• Poles that are located on the same line have different time
performance and oscillation frequency but equal damping
ratio and relative overshoot value.
Dominant Poles:
• The set of dominant poles are those proximal to
the imaginary axe and from the more distal poles
with min(𝜏𝑑𝑜𝑚 ) > 5 × max(𝜏𝑛𝑜𝑛−𝑑𝑜𝑚 ). The
dominant poles has slower transient response and
thus they are the objective of the control problem.
• Considering the dominant poles reduces the order
of the control system and the design of the
controllers.

Pole-Zero Cancellation: zeros and poles can be set at the same

position to cancel the effects of each other. Cancellation can be
employed in controller design to cancel undesired effects or to
reduce the order of the system(if it is a design degree of
freedom)
Effect of Zeros:
• The zeros affect the response amplitude.
• The effects of the zeros are more evident when they are more proximal to the dominant poles (zeros with smaller real part
has a higher time constant and has a more evident effect on the system response).
• The zeros affect the response phase.
• A real zero (or the real part of a complex zero) introduces a derivative and proportional effect in the response without zero.
• For more distal zeros (from the imaginary axe) the proportional effect is higher than the derivative effect (fast zero effect).
For the nearer zeros, the derivative effect is higher than the proportional one.
• Slower zeros cause higher signal overshoot because of the added positive value of the derivative.
• A left-side complex zero has a positive phase and thus an anticipation effect.
• A right-side complex zero has a negative phase and thus introduces a delay effect.
• A right-side zero with a smaller derivative effect than the proportional part may cause initial phase inversion.
• Asymptotically stable systems with only left-side zeros are said to be minimum-phase systems
Performance parameters:
Performance parameters are used to set, evaluate, and compare the behavior of
stable dynamic systems.
Time performance parameters:
• Rising time 𝑡𝑟𝑖𝑠 : the time necessary for the response to rise from 10% to 90% of
its final value.
• Delay-time 𝑡𝑑 : the time necessary for the response to reach 50% of its final value.
• Steady-state time (Settling time) 𝑡𝑠𝑒𝑡 : at p% error: the time necessary for the
response to reach and stay in ±0.0𝑝 around its final value.
• Peak time 𝑡𝑝𝑛 : the time of the local maximum and minimum values of the
response.
• Overshoot time 𝑡𝑜𝑣 : the time of the maximum deviation of the response from its
final value.
Value Performance parameters:
• Overshoot (𝑂𝑉): the maximum deviation between the response and its final steady state value.
𝑂𝑉 𝑡𝑜𝑣 = 𝑦𝑚𝑎𝑥 (𝑡𝑜𝑣 ) − 𝑦𝑓𝑖𝑛𝑎𝑙 . This parameter depends on the input value.

• Relative Overshoot (𝑂𝑉𝑟 ): the ratio of the overshoot and the response final value. That is 𝑂𝑉𝑟
𝑦𝑚𝑎𝑥 (𝑡𝑜𝑣 )−𝑦𝑓𝑖𝑛𝑎𝑙
= independent of the input value but requires the knowledge of the final value.
𝑦𝑓𝑖𝑛𝑎𝑙

• Percentage Overshoot (𝑂𝑉𝑟 %): the ratio of the overshoot and the response final value. That is
𝑦𝑚𝑎𝑥 (𝑡𝑜𝑣 )−𝑦𝑓𝑖𝑛𝑎𝑙
𝑂𝑉𝑟 = 𝑦𝑓𝑖𝑛𝑎𝑙
× 100%
First Order System Step response:

Time constant and steady-state approximation:

1
• The time constant is defined as 𝜏 = 𝑎
• Steady-state time: is the time at which the steady state
response is assumed to be reached accepting and tolerating a
defined maximum error value (because operations with the
system can not be done for 𝑡 → ∞).
• The most used in Engineering is 𝑡𝑠𝑡𝑒𝑎𝑑𝑦= 4𝜏 with approximately
𝑒𝑟𝑟𝑜𝑟𝑠𝑡𝑒𝑎𝑑𝑦= 2%
Second Order System:

The response of a stable second-order system has four shapes

according to the classification of the poles of the system (roots
of the characteristic algebraic equation.
• Overdamped oscillation response (no oscillation): real and
different poles 𝜁 > 1 (positive discriminant)
• Critically damped response (change in convexity-start of
oscillation: real and equal poles 𝜁 = 1 (discriminant=0)
• Undamped oscillation response (sustained oscillation):
imaginary poles 𝜁 = 0 (negative discriminant with
Re(pole)=0)
• Underdamped oscillation response:(damped oscillation):
complex roots 0 < 𝜁 < 1 (negative discriminant with
𝑅𝑒(𝑝𝑜𝑙𝑒) ≠ 0)
Second-order system-performance parameters
The underdamped response will be taken to determine the performance parameters because it has the
maximum number of performance parameters.
Polar and cartesian representation:
a complex pair of system poles can be represented in:
cartesian form: 𝑠1,2 = 𝛼 ± 𝑗𝜔𝑑 𝛼: 𝑎𝑡𝑡𝑒𝑛𝑢𝑎𝑡𝑖𝑜𝑛 𝑓𝑎𝑐𝑡𝑜𝑟, 𝜔𝑑 : 𝑑𝑎𝑚𝑝𝑒𝑑 𝑜𝑠𝑐𝑖𝑙𝑙𝑎𝑡𝑖𝑜𝑛 𝑓𝑟𝑒𝑞𝑢𝑒𝑛𝑐𝑦
Polar form: 𝑠 = 𝜔𝑛 𝑒 𝑗𝜃 𝜁 = 𝑐𝑜𝑠𝜃: 𝑎𝑡𝑡𝑒𝑛𝑢𝑎𝑡𝑖𝑜𝑛 𝑟𝑎𝑡𝑖𝑜, 𝜔𝑛 : 𝑛𝑎𝑡𝑢𝑟𝑎𝑙 𝑜𝑠𝑐𝑖𝑙𝑙𝑎𝑡𝑖𝑜𝑛 𝑓𝑟𝑒𝑞𝑢𝑒𝑛𝑐𝑦

Relations between polar and cartesian representations:

𝛼 = 𝜔𝑛 cos 𝜃 = 𝜔𝑛 𝜁 , 𝜔𝑑 = 𝜔𝑛 sin 𝜃 = 𝜔𝑛 1 − 𝜁 2
2 2 −1
𝜔𝑑
𝜔𝑛 = 𝛼 + 𝜔𝑑 , 𝜃 = −𝑡𝑎𝑛 ( )
𝛼
Under damped Step response:

By computing the parameters using partial fractions and adjusting the

function form to have the Laplace cosine and sine expression:
using the Laplace inverse we obtain:
Performance parameters of the second-order system:
The most used performance parameters of the second-order system are the settling time, the overshoot time, and the
overshoot value with all its variants.
Settling time at p% error:
To simplify computation it is assumed that the settling time is reaches at the first peak after the 0.0p, that is we consider
a smaller error that satisfies the requirements. Thus, considering
1
we have to solve the equation: 2
𝑒 −𝜁𝜔𝑛 𝑡𝑠𝑒𝑡 = 0.0𝑝
1−𝜁
−ln(0.𝑜𝑝 1−𝜁 2 ) 4 4
Solving for 𝑡𝑠𝑒𝑡 we obtain: 𝑡𝑠𝑒𝑡 = 𝜁𝜔𝑛
at 2% this result is approximated as 𝑡𝑠𝑒𝑡_2% = |𝜁𝜔 | = |𝛼| = 4𝜏
𝑛
Peak and overshoot time:
The peak time is periodic and obtained by computing equating the derivative of the step response c(t) to zero.
Since the underdamped response of the second order system is strictly decreasing, the overshoot time is obtained at the
first peak value. Considering:
and the inverse of the derivative

Applying Laplace inverse

Thus the overshoot time is:

 Overshoot evaluation: using and

compute

Applying 𝑐𝑓𝑖𝑛𝑎𝑙 = 1 in the percentage overshoot equation we obtain

Moreover, in the design problem, we can compute the damping ratio necessary to obtain a specific percentage overshoot by:
Example: Consider the following system and determine the moment of inertia and the damping coefficient to
20% overshoot and a 2% ERROR settling time of 2 seconds for a step torque input.