AME455 - Midterm Exam 02 - 03/11/11
TO GET POINTS IN EACH PROBLEM, YOU MUST SHOW THE DETAILS OF YOUR WORK.
PROBLEM 1: . . . . . . . . . . . .10 points
PROBLEM 2: . . . . . . . . . . . .10 points
PROBLEM 3: . . . . . . . . . . . .10 points
PROBLEM 4: . . . . . . . . . . . .10 points
TOTAL POINTS: . . . . . . . . . 40 points
NOTATION:
1(t) denotes the step input.
A controller D and plant G in unitary negative feedback interconnection are given by
W
+
R
Controller
P lant
G
+
P +
USEFUL EQUATIONS/FACTS:
Roots of the homogeneous polynomial as2 + bs + c = 0 are s1 =
The Laplace transform of 1(t) is 1s , of t 1(t) is
1
,
s2
b+ b2 4ac
2a
and of t2 1(t) is
and s2 =
b b2 4ac
.
2a
2
.
s3
The Laplace transform of a generic P ID controller D is given by D(s) = kp +
are constant.
ki
s
+ kd s, where kp , ki , kd
The Final Value Theorem states that the asymptotic (or steady state) value of a function of time e(t)
is given by
ess = lim sE(s),
s0
where E(s) is the Laplace transform of e, when sE(s) has all poles in the left-half plane.
For root locus plots, the characteristic equation is written as 1 + KL(s) = 0, where K is the changing
parameter and L(s) = b(s)/a(s).
The rise time for the time response to a step input for a second order system H(s) =
approximately
1.8
n ,
2
n
2
s2 +2n s+n
is
where n is the natural damping frequency. The overshoot Mp (not in percent - that
is, if 10% overshoot
r is desired, then Mp = 0.1) is given by Mp = e
ln2 (Mp )
.
Equivalently, = 2 +ln2 (M
)
p
1 2
, where is the damping factor.
Problem 1: Characterize the properties of closed-loop systems with pole/zero locations given in the splane
plots below. Determine:
1a) Which one(s) of the closed-loop system(s) is(are) stable? Why?
1b) Which one(s) of the closed-loop system(s) is(are) unstable? Why?
1c) Which one(s) of the closed-loop system(s) is(are) marginally stable? Why?
1d) For the system(s) that you determine as stable, which one(s) does(do) lead to oscillatory impulse responses?
1e) For the system(s) that you determine as stable, which one has smallest overshoot Mp for step input?
Why?
Im(s)
Im(s)
Re(s)
Re(s)
(a) Two real poles at same location
(b) Two simple poles
Im(s)
Im(s)
Re(s)
Re(s)
(c) One pair of complex conjugate poles
(d) Two poles at origin and one zero
(use this page to answer Problem 1 if needed)
Solution:
(a) is stable, because all poles are in LHP .
(b) is marginally stable, because one pole at origin and one pole at LHP .
(c) is stable, because all poles are in LHP .
(d) is unstable, because pole at origin has multiplicity 2.
1d) system (a) and (c) are stable, (c) will lead oscillatory impulse response, because poles have imaginary
parts.
1e) (a) has smallest overshoot, because its = 1, Mp = e
1 2
=0
Problem 2: Consider a plant
G(s) =
s2
4
+ 4s + 4
1
in unitary negative feedback interconnection with a P I controller D(s) = 1 + .
s
(2a) Knowing that the poles of sE(s) are in the left-hand plane, determine the steady-state error for tracking
a ramp input r(t) = t1(t). Can the system track a ramp perfectly?
Solution: by FVT,
ess = lim sE(s) = lim s
s0
s0
s0 s 1 +
= lim
= lim
s0
1
1
s2 1 + (1 + 1s )
4
s2 +4s+4
1
4(s+1)
s(s2 +4s+4)
1
s+
4(s+1)
s2 +4s+4
=1
The system cannot track ramp input perfectly.
(2b) What is the steady-state value of the plants output when a step r(t) = 1(t) is applied?
Solution:
Method 1:
yss
4
s+1
1
s
s2 +4s+4
= lim sY (s) = lim s
4
s0
s0
s 1 + s2 +4s+4 s+1
s
= lim
s0 (s2
4(s + 1)
+ 4s + 4)s + 4(s + 1)
=1
Method 2:
ess = lim s
s0
s 1+
1
4
s2 +4s+4
s+1
s
s(s2 + 4s + 4)
s0 (s2 + 4s + 4)s + 4(s + 1)
= 0 = rss yss
= lim
Since rss = 1, yss = 1.
Hint: Apply FVT.
4
Problem 3: Given a plant
G(s) =
s4
2s3
1
+ 3s2 + 4s + 5
(3a) Are there any poles of the plant G in the right-half plane?
Solution: using Rouths test
s4 1 3 5
s3 2 4 0
s2 a1 a2 0
s1 a3 0
2314
=1
2
2510
a2 =
=5
2
a1 4 a2 2
a3 =
= 6
a1
a1 =
Since a sign change has occurred to the first column, there is at least one pole located on RHP .
(3b) Consider adding a P controller D(s) = K in unitary negative feedback interconnection with the plant.
Compute the range of K for which the system is stable.
Solution: the characteristic equation for unitary negative feedback system is 1 + D(s)G(s) = 0, therefore,
1
=0
+
+ 3s2 + 4s + 5
s4 + 2s3 + 3s2 + 4s + 5 + K = 0
1+K
using Rouths test
s4
2s3
s4
1
3
5+K
s3
2
4
0
s2
1
5+K
0
1
s 6 2K
0
s0
5+K
In order to make the system stable,
6 2K > 0,
5 + K > 0,
5 < K < 3 .
Hint: Apply Rouths test. For part (3a), you dont need to compute the entire array.
Problem 4: Given the characteristic equation
1+
2s + 1
=0
(s + K)s
determine which one of the following plots is its root locus when K varies from 0 to +. Justify your answer.
Solution:
2s + 1
=0
(s + K)s
s2 + Ks + 2s + 1 = 0
1+
s2 + 2s + 1 + Ks = 0
s
=0
1 + K 2
s + 2s + 1
s
L(s) = 2
s + 2s + 1
Poles of L(s) are 1 with multiplicity 2, zero of L(s) is at 0, by Rule 1,
2 branches emanate from poles of L(s), 1.
1 branches end at finite zero of L(s), 0.
Only plot (b) fulfill this condition.
Hint: Put the characteristic in the form 1 + KL(s) = 0.
