20.01.2015 Views

10 5 transient response specifications - KFUPM Open Courseware

10 5 transient response specifications - KFUPM Open Courseware

10 5 transient response specifications - KFUPM Open Courseware

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

ME 413 Systems Dynamics & Control<br />

Chapter <strong>10</strong>: Time‐Domain Analysis and Design of Control Systems<br />

Chapter <strong>10</strong>:<br />

Time‐Domain Analysis of and Design<br />

of Control Systems<br />

A. Bazoune<br />

<strong>10</strong>.5 Transient Response Specifications of a Second Order System<br />

Because systems that stores energy cannot respond instantaneously, they exhibit a <strong>transient</strong><br />

<strong>response</strong> when they are subjected to inputs or disturbances. Consequently, the <strong>transient</strong> <strong>response</strong><br />

characteristics constitute one of the most important factors in system design.<br />

In many practical cases, the desired performance characteristics of control systems can be<br />

given in terms of <strong>transient</strong>‐<strong>response</strong> <strong>specifications</strong>. Frequently, such performance characteristics are<br />

specified in terms of the <strong>transient</strong> <strong>response</strong> to unit‐step input, since such an input is easy to generate<br />

and is sufficiently drastic. (If the <strong>response</strong> of a linear system to a step input is known, it is<br />

mathematically possible to compute the system’s <strong>response</strong> to any input).<br />

The <strong>transient</strong> <strong>response</strong> of a system to a unit step‐input depends on initial conditions. For<br />

convenience in comparing the <strong>transient</strong> <strong>response</strong>s of various systems, it is common practice to use<br />

standard initial conditions: The system is at rest initially, with its output and all time derivatives<br />

thereof zero. Then the <strong>response</strong> characteristics can be easily compared.<br />

Transient‐Response Specifications.<br />

The <strong>transient</strong> <strong>response</strong> of a practical<br />

control system often exhibits damped oscillations before reaching a steady state. In specifying the<br />

<strong>transient</strong>‐<strong>response</strong> characteristics of a control system to a unit‐step input, it is common to name the<br />

following:<br />

1. Delay time, T<br />

d<br />

2. Rise time, T<br />

r<br />

3. Peak time, T<br />

p<br />

4. Maximum overshoot, M<br />

p<br />

5. Settling time, T<br />

s<br />

These <strong>specifications</strong> are<br />

defined next and are shown in<br />

graphically in Figure <strong>10</strong>‐21.<br />

1/9


ME 413 Systems Dynamics & Control<br />

Chapter <strong>10</strong>: Time‐Domain Analysis and Design of Control Systems<br />

Delay Time. The delay time T<br />

d<br />

is the time needed for the <strong>response</strong> to reach half of its<br />

final value the very first time.<br />

Rise Time. The rise time T<br />

r<br />

is the time required for the <strong>response</strong> to rise from <strong>10</strong>% to<br />

90%, 5% to 95%, or 0% to <strong>10</strong>0% of its final value. For underdamped second order systems, the 0% to<br />

<strong>10</strong>0% rise time is normally used. For overdamped systems, the <strong>10</strong>% to 90% rise time is common.<br />

Peak Time. The peak time T<br />

p<br />

is the time required for the <strong>response</strong> to reach the first peak<br />

of the overshoot.<br />

Maximum (percent Overshoot). The maximum percent overshoot M<br />

p<br />

is the<br />

c ∞ . If<br />

maximum peak value of the <strong>response</strong> curve [the curve of c( t ) versus t ], measured from ( )<br />

c ( ∞ ) = 1, the maximum percent overshoot is M × <strong>10</strong>0% . If the final steady state value ( )<br />

p<br />

c ∞ of<br />

the <strong>response</strong> differs from unity, then it is common practice to use the following definition of the<br />

maximum percent overshoot:<br />

( p ) −C( ∞)<br />

C t<br />

Maximum percent overshoot = × <strong>10</strong>0%<br />

C<br />

Settling Time. The settling time T<br />

s<br />

is the time required for the <strong>response</strong> curve to reach and<br />

stay within 2% of the final value. In some cases, 5% instead of 2%, is used as the percentage of the<br />

final value. The settling time is the largest time constant of the system.<br />

Comments. If we specify the values of<br />

curve is virtually fixed as shown in Figure <strong>10</strong>.22.<br />

T<br />

d<br />

,<br />

T<br />

r<br />

,<br />

T<br />

p<br />

,<br />

( ∞)<br />

T<br />

s<br />

and<br />

M<br />

p<br />

, the shape of the <strong>response</strong><br />

Figure <strong>10</strong>‐22<br />

Specifications of <strong>transient</strong>‐<strong>response</strong> curve.<br />

A Few Comments on Transient Response‐Specifications.<br />

In addition of requiring a dynamic system to be stable, i.e., its <strong>response</strong> does not increase unbounded<br />

with time (a condition that is satisfied for a second order system provided that ζ ≥ 0 , we also<br />

require the <strong>response</strong>:<br />

• to be fast<br />

• does not excessively overshoot the desired value (i.e., relatively stable) and<br />

• to reach and remain close to the desired reference value in the minimum time possible.<br />

2/9


ME 413 Systems Dynamics & Control<br />

Chapter <strong>10</strong>: Time‐Domain Analysis and Design of Control Systems<br />

Second‐Order Systems and Transient‐Response‐Specifications.<br />

The <strong>response</strong> for a unit step input of an underdamped second order system ( 0 ζ 1)<br />

or<br />

ζ −ζω<br />

t<br />

−ζω<br />

t<br />

n<br />

n<br />

c()<br />

t = 1− e sinω<br />

t −e cosω<br />

t<br />

d<br />

d<br />

2<br />

1 − ζ<br />

⎧<br />

−ζω<br />

⎪ ζ<br />

⎫<br />

t<br />

⎪<br />

n<br />

= 1− e ⎨ sinω<br />

t + cosω<br />

t<br />

d<br />

d ⎬<br />

2<br />

⎩⎪<br />

1 − ζ<br />

⎭⎪<br />

e<br />

< < is given by<br />

(<strong>10</strong>‐13)<br />

−ζω<br />

2<br />

1<br />

()<br />

t n<br />

−<br />

c t = 1 − sin ⎨ω<br />

t + tan<br />

d<br />

⎬<br />

(<strong>10</strong>‐14)<br />

2<br />

1 − ζ<br />

⎧⎪<br />

⎩⎪<br />

1 − ζ ⎫⎪<br />

ζ ⎭⎪<br />

A family of curves c( t ) plotted against t with various values of ζ is shown in Figure <strong>10</strong>‐24.<br />

1.6<br />

1.4<br />

ζ = 0.2<br />

Step Response<br />

1.2<br />

0.5<br />

ut ( )<br />

1<br />

0.7<br />

1<br />

1<br />

−−−−− 142 43<br />

Input<br />

t<br />

s<br />

ω<br />

2<br />

n<br />

2 2<br />

+ 2ζ ωns+<br />

ωn<br />

Amplitude<br />

0.8<br />

0.6<br />

0.4<br />

2<br />

5<br />

0.2<br />

0<br />

0 2 4 6 8 <strong>10</strong> 12 14 16 18 20<br />

Output<br />

Time (sec)<br />

144424443<br />

_______________<br />

Figure <strong>10</strong>‐24<br />

Unit step <strong>response</strong> curves for a second order system.<br />

Delay Time. We define the delay time by the following approximate formula:<br />

Rise Time. We find the rise time<br />

r<br />

T<br />

d<br />

1+<br />

07 . ζ<br />

=<br />

ω<br />

n<br />

T by letting cT ( ) 1<br />

r<br />

= in Equation (<strong>10</strong>‐13), or<br />

⎧<br />

−ζω<br />

⎪ ζ<br />

⎫⎪<br />

n r<br />

c( T)<br />

= 1= 1− T<br />

r<br />

e ⎨ sinωT + cosω<br />

d r<br />

T<br />

d r⎬<br />

(<strong>10</strong>‐15)<br />

2<br />

⎩⎪<br />

1 − ζ<br />

⎭⎪<br />

−ζωn Since e<br />

t<br />

≠ 0, Equation (<strong>10</strong>‐15) yields<br />

3/9


ME 413 Systems Dynamics & Control<br />

Chapter <strong>10</strong>: Time‐Domain Analysis and Design of Control Systems<br />

or<br />

Thus, the rise<br />

T<br />

r<br />

is<br />

ζ<br />

2<br />

1 − ζ<br />

sinωT + cosωT<br />

= 0<br />

tanω<br />

T<br />

d<br />

r<br />

d<br />

r<br />

d<br />

2<br />

1 − ζ<br />

=−<br />

ζ<br />

r<br />

r<br />

⎛<br />

2<br />

−1 1 − ⎞ π −<br />

1 ζ β<br />

= tan −<br />

=<br />

ω ⎜ ζ ⎟ ω<br />

d<br />

⎝ ⎠<br />

d<br />

T (<strong>10</strong>‐16)<br />

where β is defined in Figure <strong>10</strong>‐25. Clearly to obtain a large value of T<br />

r<br />

we must have a large value<br />

of β .<br />

jω<br />

ωn<br />

2<br />

1−<br />

ζ<br />

−σ<br />

ω n<br />

β<br />

jω d<br />

β = cos<br />

σ<br />

−1<br />

( ζ)<br />

( ζ )<br />

or β = sin 1−<br />

or<br />

β<br />

−1 2<br />

⎛<br />

1−ζ<br />

⎞<br />

ζ ⎟<br />

⎠<br />

2<br />

−1<br />

= tan ⎜ ⎟<br />

⎜<br />

⎝<br />

ζω n<br />

Figure <strong>10</strong>‐25<br />

Definition of angle β<br />

Peak Time. We obtain the peak time T<br />

p<br />

by differentiating c( t ) in Equation (<strong>10</strong>‐13), with<br />

respect to time and letting this derivative equal zero. That is,<br />

dc ( t ) ωn<br />

−ζωn = e<br />

t<br />

sinω<br />

= 0<br />

2<br />

dt<br />

dt 1−<br />

ζ<br />

It follows that<br />

sinω t d<br />

= 0<br />

or<br />

ω t = 0, π, 2 π, 3 π,... = n d<br />

π, n = 0,1,2.....<br />

Since the peak time<br />

T corresponds to the first peak overshoot ( n = 1)<br />

, we have ω T = π<br />

p<br />

π π<br />

= =<br />

ω<br />

2<br />

d ω 1−<br />

ζ<br />

T (<strong>10</strong>‐17)<br />

p<br />

n<br />

d<br />

p<br />

. Then<br />

4/9


ME 413 Systems Dynamics & Control<br />

Chapter <strong>10</strong>: Time‐Domain Analysis and Design of Control Systems<br />

The peak time T<br />

p<br />

corresponds to one half‐cycle of the frequency damped oscillations.<br />

Maximum Overshoot M<br />

p<br />

The maximum overshoot M<br />

p<br />

occurs at the peak<br />

T = π ω . Thus, from Equation (<strong>10</strong>‐13),<br />

p<br />

d<br />

or<br />

⎧<br />

⎫<br />

n ( d )<br />

Mp<br />

c( Tp)<br />

1 e ζω πω ζ − ⎪<br />

⎪<br />

= − = − ⎨ sinπ + cos<br />

2<br />

{ π ⎬<br />

⎪ 1−ζ<br />

1442443 = −1<br />

⎪<br />

⎩<br />

⎪<br />

= 0<br />

⎭<br />

⎪<br />

p<br />

− 1 − 2<br />

= (<strong>10</strong>‐18)<br />

M e πζ ζ<br />

Since c ( ∞ ) = 1, the maximum percent overshoot is<br />

=<br />

− 1 − 2<br />

×<br />

Mp% e πζ ζ <strong>10</strong>0%<br />

The relationship between the damping ratio ζ and the maximum percent overshoot is shown in<br />

Figure <strong>10</strong>‐26. Notice that no overshoot for ζ ≥ 1 and overshoot becomes negligible for ζ > 07 . .<br />

Figure <strong>10</strong>‐26 Relationship between the maximum percent overshoot M p<br />

% and damping ratio ζ .<br />

Settling Time T<br />

s<br />

Based on 2%criterion the settling time T<br />

s<br />

is defined as:<br />

−<br />

e ζω<br />

nTs<br />

= 002 .<br />

( 002 . ) 4<br />

ln<br />

− ζωnTs = ln ( 002 . ) ⇒ Ts<br />

= ≈<br />

−ζω<br />

ζω<br />

n<br />

n<br />

5/9


ME 413 Systems Dynamics & Control<br />

Chapter <strong>10</strong>: Time‐Domain Analysis and Design of Control Systems<br />

T s<br />

4<br />

ζω<br />

= ( 2%Criterion)<br />

(<strong>10</strong>‐19)<br />

n<br />

Similarly for 5% we can get<br />

3<br />

T s =<br />

ζω<br />

( 5%Criterion)<br />

(<strong>10</strong>‐20)<br />

n<br />

REVIEW AND SUMMARY<br />

TRANSIENT RESPONSE SPECIFICATIONS OF A SECOND ORDER SYSTEM<br />

TABLE 1.<br />

Useful Formulas and Step Response Specifications for the Linear<br />

Second‐Order Model m & x&+ c x&<br />

+ k x = f (t)<br />

where m, c, k constants<br />

2<br />

− c ± c − 4mk<br />

1. Roots<br />

s1,2<br />

=<br />

2m<br />

2. Damping ratio or ζ = c / 2 mk<br />

3. Undamped natural frequency<br />

ω<br />

n<br />

=<br />

2<br />

4. Damped natural frequency<br />

ω d<br />

= ω n<br />

1−ζ<br />

5. Time constant τ = 2 m / c = 1/<br />

ζω<br />

n<br />

if ζ ≤ 1<br />

k<br />

m<br />

6. Stability Property Stable if, and only if, both roots have negative real parts, this occurs if<br />

and only if , m, c, and k have the same sign.<br />

7. Maximum Percent Overshoot: The maximum % overshoot M<br />

p<br />

is the maximum peak value of the<br />

<strong>response</strong> curve.<br />

M p<br />

= <strong>10</strong>0e<br />

2<br />

−πζ / 1−ζ<br />

8. Peak time: Time needed for the <strong>response</strong> to reach the first peak of the overshoot<br />

2<br />

T = π / ω 1−<br />

ζ<br />

9. Delay time: Time needed for the <strong>response</strong> to reach 50% of its final value the first time<br />

1+<br />

0.<br />

7ζ<br />

Td<br />

≈<br />

ω<br />

p<br />

n<br />

<strong>10</strong>. Settling time: Time needed for the <strong>response</strong> curve to reach and stay within 2% of the final value<br />

4<br />

Ts<br />

=<br />

ζω<br />

n<br />

11. Rise time: Time needed for the <strong>response</strong> to rise from (<strong>10</strong>% to 90%) or (0% to <strong>10</strong>0%) or (5% to 95%) of<br />

π − β<br />

its final value Tr<br />

= (See Figure <strong>10</strong>‐25)<br />

ω<br />

d<br />

n<br />

6/9


ME 413 Systems Dynamics & Control<br />

Chapter <strong>10</strong>: Time‐Domain Analysis and Design of Control Systems<br />

SOLVED PROBLEMS<br />

█ Example 1<br />

Figure 4‐20 (for Example 1)<br />

7/9


ME 413 Systems Dynamics & Control<br />

Chapter <strong>10</strong>: Time‐Domain Analysis and Design of Control Systems<br />

█ Example 2<br />

Figure 4‐21 (for Example 2)<br />

█<br />

Solution<br />

First The transfer function of the system is<br />

8/9


ME 413 Systems Dynamics & Control<br />

Chapter <strong>10</strong>: Time‐Domain Analysis and Design of Control Systems<br />

█ Example 3 (Example <strong>10</strong>‐2in the Textbook Page 520‐521)<br />

T<br />

d<br />

,<br />

Determine the values of<br />

subject to a unit step input<br />

T<br />

r<br />

,<br />

T<br />

p<br />

,<br />

T<br />

s<br />

when the control system shown in Figure <strong>10</strong>‐28 is<br />

Rs ( )<br />

1<br />

s ( s + 1)<br />

C( s)<br />

Figure <strong>10</strong>‐28<br />

Control System<br />

█<br />

Solution<br />

The closed‐loop transfer function of the system is<br />

1<br />

C( s)<br />

s( s+<br />

1)<br />

1<br />

= =<br />

2<br />

R( s)<br />

1<br />

1+<br />

s + s+<br />

1<br />

s s+<br />

1<br />

( )<br />

Notice that ω<br />

n<br />

= 1 rad/s and ζ = 05 . for this system. So<br />

Rise Time.<br />

T<br />

r<br />

π − β<br />

=<br />

ω<br />

where<br />

−1 −<br />

β = sin ( ω ) 1 d<br />

ωn<br />

= sin ( 0. 866 1)<br />

= 1.<br />

05rad<br />

or<br />

−1 −1 −<br />

β = cos ( ζω ) cos ( ) cos<br />

1 n<br />

ωn<br />

= ζ = ( 05 . )<br />

= <strong>10</strong>5 . rad<br />

Therfore,<br />

π −1.05<br />

T<br />

r<br />

= = 2.41 s<br />

0.866<br />

π π<br />

Peak Time. Tp<br />

= = = 363 . s<br />

ωd<br />

0.<br />

866<br />

Delay Time.<br />

1+<br />

0.<br />

7ζ<br />

1+<br />

0. 7( 0.<br />

5)<br />

Td<br />

= = = 135 . s<br />

ω 1<br />

n<br />

d<br />

ωn<br />

ω = ω − ζ = − =<br />

d<br />

2<br />

1−<br />

ζ<br />

Maximum Overshoot : Mp<br />

= e e e<br />

4 4<br />

Settling time: Ts<br />

= = = 8s<br />

ζω 05 . × 1<br />

n<br />

−σ<br />

n<br />

2 2<br />

1 1 0. 5 0.<br />

866<br />

ω n<br />

β<br />

ζω n<br />

jω<br />

jω d<br />

−πζ 1−ζ 2 − π× 05 . 1−05<br />

. 2<br />

−181<br />

.<br />

= = = =<br />

σ<br />

0. 163 16. 3%<br />

9/9

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!