EE 312/ leecture no.

4/
4 Steady state
s error derivation

Univeersity of Gary younis

Facuulty of Engineeering
Electrrical and Elecctronic Engin neering Deparrtment
-----------------------------------------------------------------------------------------------------------------------------------
Con ntrol systems EE312 Fall 2011/2012
Cou urse Format and a Assessmeent
The course will be delivered in 56 hours,, with 4 hom me works. Fin nal exam willl be in Feb 2012. 2
Cou urse material will include handouts.h
-----------------------------------------------------------------------------------------------------------------------------------
Instrructor:
Dr. Awad
A Shameekh
-----------------------------------------------------------------------------------------------------------------------------------
Lectture no. (4) Date: 188/12/2011
Learrning Outcom mes of this leccture
The student will understand:
• Definitioon of Steady State S error.
• Steady State S error deerivation in teerms of open and closed looop system.
-----------------------------------------------------------------------------------------------------------------------------------
Refeerences:
- Modern control systeems, R.C. Dorrf
- Control systems Engiineering, Norrman Nise
- Linear control
c systemms engineerin ng, Morris Drriels
----------------------------------------------------------------------------------------------------------------------------------

Definition of
o Steady
y State Errror and Test
T Inpu
uts
Steaady-state errror is the difference
d b
between me input and the
t output ffor a prescrribed
test input as t → ∞ , see Figure (1). Test inputts used for steady-statee error anallysis
and design are summarized in Table (1).
(

Tabble(1)
--------------------------------------------------------------------------------------------------------------

1
 EE 312/ leecture no. 4/
4 Steady state
s error derivation

Figgure (1), Stteady state error. (a) step input, (b)

( ramp inpput

p inputs represent
Step r constant position and thus are useful
u in ddetermining
g the
abillity of me control sysstem to possition itselff with respect to a staationary tarrget,
suchh as a satellite in geosstationary orbit
o

2
 EE 312/ lecture no. 4/ Steady state error derivation

Ramp inputs represent constant-velocity inputs to a position control system by

their linearly increasing amplitude. These waveforms can be used to test a system
ability to follow a linearly increasing input or, equivalently, to track a constant-
velocity target. For example, a position control system that tracks a satellite that
moves across the sky at a constant angular velocity

parabolas, whose second derivatives are constant, represent constant acceleration

inputs to position control systems and can be used to represent accelerating
targets, such as the missile.

Application to stable systems

Since the error is defined as the difference between the input and the output of a
feedback control system after the steady state has been reached as t → ∞ .

Figure (2), closed loop control system error: (a) general representation;
(b) representation for unity feedback systems

In Figure (2) if c (t ) equals r (t ) , e(t ) will be zero. But with a pure gain (K) the
error e(t ) cannot be zero if c (t ) is to be finite and nonzero. Thus with a pure gain
of K in the forward path, an error must exist.

The larger the value of K the smaller the value of steady state ( ess ) would have

to be to yield a similar value of C ss . The conclusion that can be drawn is that

with a pure gain in me forward path, there will always be a steady-state error for
a step input. The error diminishes as the value of K increases.

3
 EE 312/ lecture no. 4/ Steady state error derivation

Figure (3)

lf the forward-path gain is replaced by an integrator as shown in Figure (3), there

will be zero error in the steady state for a step input. The reasoning is as follows:
As c (t ) increases e(t ) will decrease, since e(t ) = r (t ) - c (t ) . This decrease will
continue until there is zero error, but there will still be a value for c(t) since an
integrator can have a constant output without any input. For example, a motor can
be represented simply as an integrator. A voltage applied to the motor will cause
rotation. When the applied voltage is removed, the motor will stop and remain at
its present output position. Since it does not return to its initial position. Then there
is an angular displacement output without an input to the motor.

Steady state error for unity feedback systems

ess can be calculated from system's closed loop transfer function, T (s ) , or the
open loop transfer function, G(s) , for the unity feedback systems.

E ( s) = R( s) − C ( s) (1)
C ( s) = R( s)T ( s) (2)

Substitute eq. (2) into (1), this gives:

E ( s) = R( s) − R( s)T ( s) (3)
or
E ( s) = R( s)(1 − T ( s)) (4)

4
 EE 312/ lecture no. 4/ Steady state error derivation

Final value theorem

[ ] ∫ f& (t )e
∞

L f& (t ) = − st
dt = SF ( s ) − F (0−) (5)
0−

As s → 0
∞

∫ f& (t )dt =
0−
f (∞) − f (0−) = limSF ( s ) − F (0−)
s →0
(6)

Or
f (∞) = lim SF ( s ) (7)
s→ 0

e(∞ ) = lim e(t) = lim SE ( s ) (8)

t →∞ s →0

e(∞ ) = lim S R ( s ) [1 − T ( s )] (9)

s →0

Example:
Find the steady state error for the system given by:

5
T (s) =
S + 7 S + 10
2

If the input is a unit step.

Solution

the steady state error can be evaluated by eq. (9), and R(s) is a unit step then

e(∞ ) = lim S R ( s ) [1 − T ( s )]
s →0

1 5
Or e(∞) = lim S [1 − 2 ]
s →0 S S + 7 S + 10
S 2 + 7 S + 10 5
Or e(∞) = lim [ 2 − 2 ]
s →0 S + 7 S + 10 S + 7 S + 10
S 2 + 7S + 5
Or e(∞ ) = lim [ 2 ]
s → 0 S + 7 S + 10

5 1
Then e( ∞ ) = = #
10 2

5
 EE 312/ lecture no. 4/ Steady state error derivation

Steady state error in terms of G(s), open loop system transfer

function
Expressing the steady state error in terms of G(s) rather than T(s) gives more
insight for the design and analysis. Consider the system shown in the figure (4)

Figure (4)

Since H(s) in the above system is unity, then

E ( s) = R( s ) − C ( s) (10)
and

C ( s) = E ( s)G( s) (11)

Substituting eq.(11) into eq. (10):

E ( s) = R( s) − E ( s)G( s) (12)
Or
E ( s)[1 + G ( s)] = R( s) (13)
Or
R( s )
E ( s) = (14)
1 + G( s)

Now apply the final value theorem

R( s)
e(∞) = lim S E ( s ) = lim S (15)
s →0 s →0 1 + G(s)

The last equation allows to calculate the e (∞) given the input signal and the
ss

open loop system transfer function, G(s). Three different input signals will be
tested to draw conclusions about the relationships that exist between the open loop
system, G(s), and the nature of the steady state error, e (∞) .
ss

1
Step function input ( R ( s ) = )
S

6
 EE 312/ lecture no. 4/ Steady state error derivation

1
S 1 1
e (∞ ) = e (∞) = lim S = lim =
ss step s → 0 1 + G(s) s → 0 1 + G ( s ) 1 + lim G ( s )
s →0

In order to have a zero steady state error lim G ( s ) = ∞ . This should satisfy the
s →0

following:
1 1 1
e (∞ ) = = = =0#
step 1 + lim G ( s ) 1 + ∞ ∞
s →0

1
Ramp input ( R ( s ) = )
S2
1
2 1 1
e (∞ ) = e (∞) = lim S S = lim =
ss ramp s →0 1 + G( s ) s→0 S + SG( s) lim SG( s)
s →0

To have a zero steady state error for a ramp input, lim SG ( s ) = ∞ , that yields
s →0

1 1
e (∞ ) = = =0
ramp lim SG ( s) ∞
s →0

1
Parabolic input ( R ( s ) = )
S3

1
3 1 1
e (∞ ) = e (∞) = lim S S = lim 2 =
ss parabola s →0 1 + G ( s) s→0 S + S G ( s) lim S 2 G ( s)
2
s →0

To have a zero steady state error lim S G ( s ) = ∞ , which gives

2
s →0

1 1
e (∞ ) = = =0#
parabola lim S G ( s) ∞
2
s →0

System type and steady sate error evaluation

The general form of

( s + z1 )( s + z 2 )........
G (s) ≅ (16)
s ( s + p1 )( s + p 2 )........
n

7
 EE 312/ lecture no. 4/ Steady state error derivation

Note: n in equation (16) defines the system type such that:

™ If n = 0 then the system is called type zero;
™ If n = 1 then the system is called type one;
™ If n = 2 then the system is called type two;
™ If n = 3 then the system is called type three;
and so no.
9 Remark : system type denoted by number of pure integrations in the
forward path.

For the limit of G(s) to be infinite, the denominator must be equal to zero as S
goes to zero. Thus n ≥ 1; that is at least one pole must be at the origin. Since
division by S in the frequency domain is integration in the time domain.

For the following input signals :

™ the step function input
1
e (∞ ) = (17)
step 1 + lim G ( s)
s →0

If n=0 in equation (17) then :

z1 z 2 ....
lim G ( s ) = ( is finite)
s →0 p1 p 2 ...
• Therefore for a step input to a unity feedback system, the steady
state error will be zero if there is at least one pure integrator in the
forward path ( that means n ≥ 1, the system type must be one or
larger).
• If there is no pure integration in the path, then there will be a
nonzero finite error.

™ the Ramp input

1
e (∞ ) = (18)
ramp lim SG ( s )
s →0

8
 EE 312/ lecture no. 4/ Steady state error derivation

• To have lim SG ( s ) = ∞ ( an infinite value ) there must be at least 2 poles (

s →0

the system must be type 2 or larger, n ≥ 2 ) at the origin.

• If only one integration in the forward path then

z1 z 2 ...
lim SG ( s ) = and the error is finite.
s →0 p1 p 2 ...
• If there are no integrators in the path then lim SG ( s ) = 0 , which leads to an
s →0

infinite e .
ss

™ the parabolic input

1
e (∞ ) = (19)
parabola lim S 2G ( s)
s →0

• to have a zero steady state error there must be at least 3 poles at the origin
(the system must be type three or larger; n ≥ 3 ).
z1 z 2 ...
• If n = 2 , then lim S 2 G ( s ) = , that means e = finite value.
s →0 p1 p 2 .. ss

• If n = 1 or less then lim S 2 G ( s ) = 0 , which leads to e = ∞.

s →0 ss

Example:
For the system given in Figure (5), find the steady state errors for inputs of 5u (t )
, 5tu (t ) and 5t 2 u (t ) . The function u (t ) is the unit step.

Figure (5)

Solution
Note: there is no pole (no pure integrator) at the origin, then system is type zero.

9
 EE 312/ lecture no. 4/ Steady state error derivation

5
™ For 5u (t ) → ( step input)
s
sR ( s ) s (5 / s )
e(∞) = estep (∞) = lim =
s →0 1 + G ( s ) 1 + lim G ( s )
s →0

120(0 + 2) 240
lim G ( s ) = = = 20
s →0 (0 + 3)(0 + 4) 12
Then
5 5
e( ∞ ) = = #
1 + 20 21
5
™ For 5tu (t ) → (ramp input)
s2

sR ( s ) s (5 / s 2 )
e(∞) = eramp (∞) = lim = lim
s →0 1 + G ( s ) s →0 1 + G ( s )

5 5 5
e(∞) = lim = = =∞ #
s →0 s + sG ( s ) lim sG ( s ) 0
s →0

10
™ For 5t 2 u (t ) → (parabola input)
s3

sR ( s ) s (10 / s 3 )
e(∞) = e parabola (∞) = lim = lim
s →0 1 + G ( s ) s →0 1 + G ( s )

10 10 10
e(∞) = lim 2 = = =∞ #
s →0 s + s G ( s )
2 2
lim s G ( s ) 0
s →0

Example:

Evaluate the steady state errors for the system given in the next block diagram,
Figure (6), if the input signal:
5
™ 5u (t ) → ( step input)
s
5
™ 5tu (t ) → 2 (ramp input)
s
10
™ 5t 2 u (t ) → 3 (parabola input)
s

10
 EE 312/ lecture no. 4/ Steady state error derivation

Figure(6)

Solution
Note: the system has a pole at the origin ( a pure integrator in the forward path ).
Then the system is type one.
sR( s ) s (5 / s ) 5
e(∞) = estep (∞) = lim = lim = lim
s →0 1 + G ( s ) s →0 1 + G ( s ) s →0 1 + lim G ( s )
s →0

Provided
100(0 + 2)(0 + 6) 1200
lim G ( s ) = = =∞
s →0 0(0 + 3)(0 + 4) 0

Then
5 5
e(∞) = estep (∞) = = =0#
s →0 1 + lim G ( s ) ∞
s →0

9 (remember that type one system and step input yields a zero ess )

*************************

sR ( s ) s (5 / s 2 )
e(∞) = eramp (∞) = lim = lim
s →0 1 + G ( s ) s →0 1 + G ( s )

5 5
e(∞) = lim =
s →0 s + sG ( s ) lim sG ( s )
s →0

Since
s100( s + 2)( s + 6) 1200
lim sG ( s ) = = = 100
s →0 s ( s + 3)( s + 4) 12
Then
5 5 5 1
e(∞) = lim = = = #
s →0 s + sG ( s) lim sG ( s) 100 20
s →0

9 (remember that type one system with ramp input gives a finite ess )

*************************************

11
 EE 312/ lecture no. 4/ Steady state error derivation

sR( s) s(10 / s 3 )
e(∞) = e parabola (∞) = lim = lim
s →0 1 + G ( s ) s →0 1 + G ( s )

10 10
e(∞) = lim 2 =
s →0 s + s G ( s )
2
lim s 2 G ( s)
s →0

Where:

s 2 100( s + 2)( s + 6) 0 × 100(0 + 2)(0 + 6) 0

lim s 2 G ( s ) = lim = = =0
s →0 s →0 s ( s + 3)( s + 4) (0 + 3)( s + 4) 12

Then
10 10 10
e(∞) = lim = = = ∞#
s →0 s + s G ( s ) lim s G ( s ) 0
2 2 2
s →0

9 (remember that for type one system excited by parabola input results an
infinite ess )

Static error constant

Three different static error constants are generally defined in ess calculation.
These are as follows:
™ Position constant ( K P ), this constant is defined for step input and given
by
1 1
e(∞) = e step (∞) = =
1 + lim G ( s ) 1 + K p
s →0

Then
K p = lim G ( s )
s →0

™ Velocity constant ( K v ), which is derived when the system excited by

ramp signal. The following formula shows how K V can be determined:
1 1
e(∞) = e ramp (∞) = =
lim sG ( s) K v
s →0

Or

12
 EE 312/ lecture no. 4/ Steady state error derivation

™ Acceleration constant ( K a ), when the system is supplied by parabola

input function K a can be expressed as:
1 1
e(∞) = e parabola (∞) = 2
=
lim s G ( s ) K a
s →0

Then

K a = lim s 2 G ( s )
s →0

Example:
For the system shown in figure (7) find the static error constants and expected
error for the standard step, ramp, and parabolic inputs.

Figure (7)

Solution
9 Remember this system is type zero.

Static error constants

500(0 + 2)(0 + 5) 125
ƒ K p = lim G ( s ) = = = 5.208 #
s →0 (0 + 8)(0 + 10)(0 + 12) 24
(0)(500)(0 + 2)(0 + 5)
ƒ K v = lim sG ( s ) = = 0#
s →0 (0 + 8)(0 + 10)(0 + 12)
(0)(0)(500)(0 + 2)(0 + 5)
ƒ K a = lim s 2 G ( s ) = = 0#
s →0 (0 + 8)(0 + 10)(0 + 12)
Steady state errors

1 1
ƒ Step input: e(∞) = estep (∞) = = = 0.161
1 + lim G ( s) 1 + K p
s →0

1 1
ƒ Ramp input: e(∞) = e ramp (∞) = = =∞
lim sG ( s ) K v
s →0

1 1
ƒ Parabolic input e(∞) = e parabola (∞) = 2
= =∞
lim s G ( s) K a
s →0

13
 EE 312/ lecture no. 4/ Steady state error derivation

Table (2) finalizes the general conclusion for the general system form that shown
in Figure (8).

Figure (8)

Table (2), shows the relationships between input, system type, static error
constants, and steady state errors

(ess ) Type 0, (n=0) Type 1, (n=1) Type 2, (n=2)

Input error error error
formula Static const. Static const. Static const.
1 1
step K p = const . Kp = ∞ 0 Kp = ∞ 0
1+ K p 1+ K p
1 1
Ramp Kv = 0 ∞ K v = const. Kv = ∞ 0
Kv Kv
1 1
parabola Ka = 0 ∞ Ka = 0 ∞ K p = const .
Ka Ka

Steady state error specifications

Static error constants can be used to specify the steady-stare error characteristics
of control systems that denoted by damping ratio,( ζ ), settling time, ( Ts ), peak
time, ( T p ), and percentage of over shoot,( OS % ). The static error constants can
provide in control system design a wealth information.

Example
If a control system has the K v = 1000 , the following points about the system can be
concluded:
• The system is stable.
• The system type 1 since K v = const.
• A ramp input is the test signal.
1 1
• ess = = = 0.001
K v 1000

14
 EE 312/ lecture no. 4/ Steady state error derivation

Example
Given the control system in Figure (9), find the value of K so that there is 10% error
in the steady state.

Figure (9)
(From the figure it is clear that the system is type one).

Solution

Since the system is type 1 the error stated in the problem must apply to a ramp
input; only a ramp yields a finite error in a type 1system. Thus
1
e(∞) = e ramp (∞) = = 0. 1
Kv
( S ) K (0 + 5) ( K )(5)
K v = 10 = lim sG ( s ) = =
s →0 ( s )(0 + 6)(0 + 7)(0 + 8) (6)(7)(8)

K = 672 #

By applying the Routh-Hurwitz criterion, it is possible to confirm that the system

is stable at this gain (K=672). However, this does not mean that the gain yields the
desirable transient response specifications.

Steady state error for disturbances

Feedback systems are used to compensate for disturbances or unwanted inputs that
come in a system. Figure (10) illustrates the plant, input, controller, and disturbance.

Figure (10)

15
 EE 312/ lecture no. 4/ Steady state error derivation

In the above system:

C ( s ) = E ( s )G1 ( s )G2 ( s ) + D ( s )G2 ( s ) (20)
But
C ( s ) = R( s ) − E ( s ) (21)

Substitute eq.(21) into eq.(21):

R ( s ) − E ( s ) = E ( s )G1 ( s )G2 ( s ) + D ( s )G2 ( s ) (22)

Or
E ( s )(1 + G1 ( s )G2 ( s )) = R ( s ) − D ( s )G2 ( s ) (23)
Or
R(s) G2 ( s )
E (s) = − D( s) (24)
1 + G1 ( s )G2 ( s ) 1 + G1 ( s )G2 ( s )

To find e(∞) apply the following formula:

S SG2 ( s )
e(∞) = lim SE ( s ) = lim R ( s ) − lim D( s ) (25)
s →0 s → 0 1 + G ( s )G ( s ) s → 0 1 + G ( s )G ( s )
1 2 1 2

The first term in eq.(25) is the ess due to the input signal, R(s), whereas the second
term is the ess due to the disturbance, D(s).
Equation (25) can be expressed as:

e(∞ ) = e R (∞ ) + e D (∞ ) (26)

Where

S
eR (∞) = lim R(s) (27)
s → 0 1 + G ( s )G ( s )
1 2
And
SG2 ( s )
eD (∞) = − lim D( s) (28)
s → 0 1 + G ( s )G ( s )
1 2

Equation (27) is the steady state error due to R(s), which we have already
obtained. Equation (28) denotes the steady state error due to the disturbance D(s).
At this point, some facts should be highlighted in order to show how e D (∞ ) can
be minimized.
Assume D(s) is a step disturbance this allows to have e D (∞ ) as follows:

sG2 ( s ) 1 1
eD (∞) = − lim =− (29)
s →0 1 + G ( s )G ( s ) s 1
1 2 lim + lim G1 ( s )
s →0 G ( s ) s →0
2

16
 EE 312/ lecture no. 4/ Steady state error derivation

Obviously, from eq.(29) the e D (∞ ) can be reduced by increasing the dc gain of

G1 ( s ) or deceasing the dc gain of G2 ( s ) .

Example:
Find the steady state error component due to a step disturbance for the system of
Figure (11).

Figure (11)

Solution
1 1 1 1
e D (∞ ) = − =− =− =− #
1 1 0 + 1000 1000
lim + lim G1 ( s ) + 1000
s →0 G ( s )
2
s →0 0(0 + 25)

Remark :
the results show that the e D (∞) induced by the step disturbance is
inversely proportional to the dc gain of G1 ( s ) . In this example the dc gain
of G2 ( s ) is infinite.

17