Академический Документы
Профессиональный Документы
Культура Документы
1
Im a g A x is
-1
-2
-3
-4
-3 -2 . 5 -2 -1 . 5 -1 -0 . 5 0 0.5 1
R e a l A x is
THE ROOT LOCUS METHOD
Contents
Design Examples
THE ROOT LOCUS METHOD
Objective
• This lecture will help you understand what a root locus is and
how to create and use one.
• It provides definitions of terms, a step-by-step guide to
constructing a root locus, and details about how to design and
evaluate controllers using the root locus method.
• After studying these materials, you should be able to create a
root locus and use the locus to understand the closed-loop
system behavior given an open-loop system and a feedback
controller.
THE ROOT LOCUS METHOD
Introduction
• The root locus method was introduced by Evans in 1948
• The root locus is a graphical procedure for determining the poles of a closed-
loop system given the poles and zeros of a forward-loop system.
• Graphically, the locus is the set of paths in the complex plane traced by the
closed-loop poles as the root locus gain is varied from zero to infinity.
• In mathematical terms, given a forward-loop transfer function, KG(s)
where K is the root locus gain, and the corresponding closed-loop transfer
function
Introduction
Introduction
These are examples of typical root loci
THE ROOT LOCUS METHOD
• The magnitude criterion is a direct result of the definition of the root locus; it is
another way to express the locus requirements. The root locus is defined as the set
of roots that satisfy the characteristic equation
or, equivalently,
• Taking the magnitude of each side of the equation yields the magnitude criterion
THE ROOT LOCUS METHOD
-1,1 1
-2,0 1
-1,-1 1
THE ROOT LOCUS METHOD
The angle criterion is used to determine the departure angles for the parts of the root locus near the open-
loop poles and the arrival angles for the parts of the root locus near the open-loop zeros. When used with the
magnitude criterion, the angle criterion can also be used to determine whether or not a point in the s-plane is
on the root locus.
The angle criterion is defined as
on the root locus,
Note that +180° could be used rather than -180°. The use of -180° is just a convention. Since +180° and
-180° are the same angle, either produces the same result.
The angle criterion is a direct result of the definition of the root locus; it is another way to express the locus
requirements. The root locus is defined as the set of roots that satisfy the characteristic equation
or, equivalently,
Taking the phase of each side of the equation yields the angle criterion.
THE ROOT LOCUS METHOD
•The phase angle for various points in the complex plane are listed in the following
tables where the points in table 1 have a phase equal to 180 degrees and thus satisfy
the angle criterion. And those in table 2 do not satisfy the angle criterion, and thus
must not be on the root locus.
Table 1 Table 2
Coordinate Phase Coordinate Phase
0,0 180 -4,0 0
-1,1 180 -2,-2 150.3
-2,0 180 1,1 104
-1,-1 180 0,-1 206.6
THE ROOT LOCUS METHOD
•The angles at
which they
arrive/leave are a
function of the
number of loci that
approach/diverge
from the break
point.
THE ROOT LOCUS METHOD
Note that the corresponding variations in the poles of the closed-loop system
determine the root locus.
As the gain moves from 0 to infinity, the poles move from the forward-loop
poles along the locus toward forward-loop zeros or infinity.
In block diagram form, the root locus gain is located in the forward loop,
before the system, as shown below.
Constructing the Locus
An illustrative Example
Let the transfer function for the system, G(s), be
The forward-loop
transfer function is
An illustrative Example
An illustrative Example
Step 7: Draw the locus by connecting the poles with the breakpoints,
axis crossings, asymptotes, and arrival angles.
Constructing the Locus
♦Draw the poles and zeros exactly as they appear in the forward-loop system.
Include all of the poles and zeros, i.e. poles and zeros of both the controller and the
uncontrolled system. Start with the forward-loop poles and zeros. Since the locus
represents the path of the roots (specifically, paths of the closed-loop poles) as the root
locus gain is varied, we start with the forward-loop configuration, i.e. the location of
the roots when the gain of the closed-loop system is 0.
The forward-loop has 5 roots: poles at -4, -1, and 1; zeros at -3 and -1.
Many root loci have paths on the real axis. Those are determined entirely by the
number of forward-loop poles and zeros and their relative locations. .
If there are no poles or zeros on the real axis, then there will be no real axis part of
the root locus.
Step 3: Asymptotes
The asymptotes indicate where the poles will go as the gain approaches infinity.
For systems with more poles than zeros, the number of asymptotes is equal to
the number of poles minus the number of zeros.
In some systems, there are no asymptotes; when the number of poles is equal to
the number of zeros then each locus is terminated at a zero rather than
asymptotically to infinity.
The asymptotes are symmetric about the real axis, and they stem from a point
defined by the relative magnitudes of the open-loop roots. This point is called
the centroid.
Note that it is possible to draw a root locus for systems with more zeros than poles,
but such systems do not represent physical systems. In these cases, you can think of
some of the poles being located at infinity.
Constructing the Locus
Step 3: Asymptotes
The number of asymptotes is equal to the difference between the number
of poles and the number of zeros. The location of the centroid on the real
axis is given by:
where pi and zj are the poles and zeros, respectively.
Note since pi and zj are symmetric about the real axis,
their imaginary parts get cancelled out.
Once you have located the centroid, draw the asymptotes at the proper angles.
The asymptotes will leave the centroid at angles defined by
Note that because of the symmetry, the asymptotes must assume one of the following configurations,
for n-m = 1,2,3,4,5.
Constructing the Locus
Note, since there are three more poles than zeros, there will be three asymptotes.
One of the asymptotes is on the real axis, the other two start at the centroid and
leave the real axis at angles of +60 and -60 degrees, respectively.
Since there are two more poles than zeros, there will be two asymptotes spaced 180
degrees from each other and leaving the real axis at +90 and -90 degrees.
Step 4: Breakpoints
- Break points occur where two or more loci join then diverge.
- Although they are most commonly encountered on the real axis, they may also
occur elsewhere in the complex plane.
- Each break point is a point where a double (or higher order) root exists for some value of K.
- Mathematically, given the root locus equation
where the transfer function G(s) consists of a numerator, A(s), and denominator,
B(s), then the break points can be determined from the roots of
- If K is real and positive at a value s that
satisfies this equation, then the point is a break
point.
Perhaps the easiest way to find break points is by trial and error. First, determine the
characteristic equation of the system in terms of K. In the vicinity of the suspected break
point, substitute values for s in this equation. A break point occurs when the
characteristic equation is minimized.
To calculate the break point explicitly requires that you derive the derivative of the
characteristic equation in terms of s then equate the derivative to 0 and solve that equation for
K and s.
Constructing the Locus
We suspect a break point on the real axis between -2 and -4 (Why? Look at the
locus from steps 1 through 3, and remember that root locus paths cannot lie on
top of each other), so we try values in that range.
s K
Of the points we tried, 2.25 yielded the smallest 2.5 1.875
value for K, so the breakpoint is about 2.25. To get
3.0 1.5
a more accurate location of the break point, one
would try points to either side of 2.25 to locate the 2.25 1.38
minimum. 2.4 1.76
Constructing the Locus
Angle of departure: At each complex pole, add up the angles from the zeros to
the current pole, then subtract the angles from the other poles to the current pole.
Angle of arrival: At each zero, add up the angles from the poles to the current
zero, then subtract the angles from the other zeros to the current zero. In
mathematical terms, for a given zero, the angle of departure is
Constructing the Locus
By convention, the arrival and departure angles are measured relative to the real
axis, so that the positive real axis is 0. Graphically, the angles look like this
Note that single poles and zeros on the real axis will always have arrival/departure
angles equal to 0 or 180 degrees due to the symmetry of the complex conjugates
Constructing the Locus
the angle of departure for the pole located at (-5, 1.5) is equal to
The complete root locus can be drawn by starting from the forward-loop poles,
connecting the real axis section, break points, and axis crossings, then ending at
either the forward-loop zeros or along the asymptotes to infinity and beyond.
If your hand-drawn locus is not detailed enough to determine the behavior of your
system, then you may want to use MATLAB or some other computer tool to calculate
the locus exactly
Now sketch in the rest of the locus. Use the asymptotes, arrival/departure angles,
break points, and axis crossings to guide your sketch.
In general, zeros tend to 'repel' the locus, whereas poles tend to 'attract' the locus.
One locus segment tends to 'attract' another locus segment until a break point forms.
Typically, the only time one needs to determine exactly the locus shape is when the
locus is near the imaginary axis or in regions where a detailed understanding of the
system behavior in the time domain is required.
In these cases, if the previous steps did not yield locus details sufficiently accurate for
your purposes, then use a computer tool to generate the locus exactly.
Constructing the Locus
The root locus shows you graphically how the system roots will move as you change
the root locus gain.
Often, however, one must determine the gain at critical points on the locus, such as
points where the locus crosses the imaginary axis
The magnitude criterion is used to determine the value of the root locus gain, K, at
any point on the root locus.
The gain is calculated by multiplying the lengths of the distance between each pole to
the point then dividing that by the product of the lengths of the distance between each
zero and the point.
Constructing the Locus
Coordinate Gain
0.5,0 1.75
0.25,0 2.4375
0,0 4
The root locus gain for other points on the locus are given
0,0.25 4.0625
in the following table: 0,0.5 4.25
0,0.75 4.5625
0,1 5
0,2 8
0,3 13
Constructing the Locus
These are some root locus plots for a variety of systems. They include the construction marks
for arrival/departure angles, asymptotes, break points, and axis crossings. Note that in some
cases, a slight change in pole or zero coordinates can result in a markedly different locus.
Note also, however, that such small changes to the roots will not change more general locus
characteristics, such as the number of asymptotes
Locus for a system with 6 poles and 2 zeros. In Locus for a system with 3 poles and 2 zeros.
this locus, determining the points where the locus Based upon the construction steps, there is
crosses the real axis is rather important, since there no obvious indication of how the curvature
is only a small set of values for K for which the of this locus should be drawn. The curvature
system is stable. must be determined computationally.
Locus for a system with 5 poles and 2 zeros. Compare this locus
Locus for a system with 5 poles and 2 zeros.
with that in the previous image. A slight change in the position of
Compare this locus with that in the next image the zeros changes the shape of the locus, but does not change the
system behavior for large values of the root locus gain
PROPERTIES OF THE ROOT LOCUS
6. The point at which the root locus intersects the imaginary axis can
be determined by using the Routh Criterion
7. The break-away and re-entry points on the root locus are determined
from the roots of dK/ds = dG/ds=0. If r branches meet at a point, the
break away occurs at an angle of 180 / r.
8. The root locus departs from complex, open loop poles, and arrives at
complex, open loop zeros at certain angles. The angle of departure from
an open-loop (complex) pole is given by
8. The points at which the root loci intersect the imaginary axis can be
determined by letting s = jω , then
Re {KG(jω ) H(jω )} = -1
Im {KG(jω ) H(jω )} = 0
Step 2: Locate the open-loop poles and zeros of p(s) in the s-plane
Step 6: Estimate the angle of root locus departure from complex poles and the
angle of arrival at complex zeros
Step 7: Determine the points of the root locus crossing of the imaginary axis
The ROOT LOCUS Problem
Problem description
H(s)
The ROOT LOCUS Problem
Problem Formulation
EXAMPLE 1
k
1+ =0
s ( s + 4 )(s +4 + j4)(s + 4 - j4)
1. Poles of the open-loop system (k = 0) are s = 0 , -4 , -4 ± j4
2. n=4, m=0 i.e. there are 4 branches ending at infinity
3. Asymptotes center and angles are
-4-4-4
σ A
= = -3
4o
2q + 1 o o o o
ϕ A = 180 = +45 , 135 , 225 , 315
4
4. The break away points are determined from
k = - s (s + 4) (s + 4 + j4) (s + 4 - j4) = P(s) = 0
Let dk/ds = 0 and solve for s
The ROOT LOCUS Problem
EXAMPLE 1 .. continued
or search for a maximum of P(s) in the region between s = 0 and s = -4 as shown
below
s -4 -3 -2 -1.5 -1.0 0
P(s) 0 51 80 85 -1.0 0
Thus the RL breaks away at about s = - 1.5
EXAMPLE 1 .. continued
× j4
θ d1 = 225
j3.25
ϕ Α = 45
× ×
-4 -2 0
σ Α = −3
-j3.25
θ d2 = 135
× - j4
The ROOT LOCUS Problem
EXAMPLE 1 .. continued
Using the following MATLAB and Control System Toolbox script
n=[0 0 0 1]
p=[1 12 64 128 0]
rlocus(p,n)
you will get the shown graph 6
Im a g A xi s
-2
-4
-6
-6 -5 -4 -3 -2 -1 0 1 2
R e a l A xis
The ROOT LOCUS Problem
EXAMPLE 2
k
G(s) H(s) =
s ( s2 + 6s + 25 )
1. Poles of the open-loop system (k = 0) are s = 0 , - 3 ± j4
2. n = 3 , m = 0 i.e. there are 3 branches ending at infinity
3. Asymptotes center and angles are
0-3-3
σ A
= = -2 ϕ Α = 60 , 180 , 300
3
4. The break away points are determined from
k = s ( s2 + 6s + 25 ) = P(s) = 0 and letting dk/ds = 0 , and solve for s
to yield s = -2 ± j2.0817
5. The jω axis crossings occur at s = ± j5 when k= 150
6. The angles of departure are calculated as θ d1 = − 36.8 and θ d2 = 36.8
The ROOT LOCUS Problem
EXAMPLE 2 .. continued
s = 0 , - 3 ± j4
8
0-3-3
σ = = -2 6
A
3 4
ϕ Α = 60 , 180 , 300 2
Im a g A xi s
0
-4
jω axis crossings at -6
s = ± j5 -8
-4 -3 -2 -1 0 1 2
R e a l A xi s
EXAMPLE 3
k
G(s) H(s) =
s ( s2 + 6s + 10 )
1. Poles of the open-loop system (k = 0) are s = 0 , - 3 ± j1
2. n = 3 , m = 0 i.e. there are 3 branches ending at infinity
3. Asymptotes center and angles are
0-3-3
σ A
= = -2 ϕ Α = 60 , 180 , 300
3
4. The break away points are determined from
k = s ( s2 + 6s + 10 ) = P(s) = 0 and letting dk/ds = 0 , and solve for s
to yield s1 = 1.1835 and s2= -2.8165
5. The jω axis crossings occur at s = ± j√ 10 when k= 60
6. The angles of departure are calculated as θ d1 = − 71.57 and θ d2 = 71.57
The ROOT LOCUS Problem
EXAMPLE 3 .. continued
s = 0 , - 3 ± j1 6
0-3-3
σ = = -2
4
A
3
2
ϕ Α = 60 , 180 , 300
Im a g A x i s
0
s2= -2.8165 -4
jω axis crossings at
s = ± j√ 10 when k= 60
-6
-4 -3 -2 -1 0 1 2
R e a l A xis
Design Problem
A control system is needed to satisfy the following specifications :
(a) steady state error for a ramp input to be ≤ 10% of the input magnitude
(b) damping ratio of the dominant roots to be ≥ 0.707
(c) settling time to be ≤ 3 sec
G(s)
+ E(s) k
R(s) Σ 1
Y(s)
_ s(s+2)
H(s)
1+ks 2
The ROOT LOCUS Problem
1.5
0.5
s2 + (2 + k1k2) s + k1 = 0
Ima g A xis
0
-0.5
when α = k1 = 20
-1
-1.5
-2
-3 -2.5 -2 -1.5 -1 -0.5 0 0.5 1 1.5 2
Real Axis
s
1+ β =0
4
s2 + 2s + α 2
Im a g A xis
0
-4
= 20 k2 ,
and k2 = 0.215
The ROOT LOCUS Problem
1 1 p1
1 × 4 (s + p1)(s + p2)
(s + p1) -p1 p2
p1 , p2 Complex
1 .5
-z1 (s + z1)
(s + z1)
0 .5
2 × o 5
Im a g A xis
-p1
0
-1
z1 < p1 < p2 -1 .5
-6 -5 -4 -3 -2
R e a l A xi s
-1 0 1 2
1 (s + z1) -z1
× × × o ×
3 (s + p1) (s + p2) -p2 -p1 6 (s + p1)(s + p2) -p1 -p2
p1 < z1 < p2
The ROOT LOCUS Problem
(s + z1) 4
-z1 1
2
7 (s + p1)(s + p2) × × o
Im a g A xi s
10 s (s + p2)(s + p3)
0
-p1 -p2 -2
p2 < p1 < z2 -4
-6
-4 -3 -2 -1 0 1 2
Real Axis
1 3
0 .8
(s + z1)
2
0 .6
0 .4
1 1
(s + p1)(s + p2)
0 .2
8 11
Im a g A xis
(s + p)
Im a g A xi s
0
2
0
-0.2
-1
-0.4
-0.6
p1 ,p2 Complex
-2
-0.8
-1
-2 -1.5 -1 -0.5 0 0.5 1 -3
-4 -3 -2 -1 0 1 2
Real Axis
R e al A xis
2 .5 2 .5
2 2
1 .5 1 .5
1
1
0 .5
1 1
9 12
0 .5
Im a g A x is
0 0
-0 .5 -0 .5
p2, p3 Complex
-1 -1
-1 .5 -1 .5
-2
-2
-2 .5
-2 .5
-4 -3 -2 -1 0 1 2
-1 - 0 .8 -0 .6 - 0 .4 - 0 .2 0 0 .2 0 .4 0 .6 0 .8 1
R e a l A x is
R e a l A x is
The ROOT LOCUS Problem
1.5
3
(s + z1) (s + z1)
1
2
(s + z1) 1
0.5
13
Im a g A xis
Im a g A xi s
(s + p1)(s + p2)(s + p3) 16 (s + p1)(s + p2)(s + p3)
0
0
-0.5
-1
-1
-2
-1.5
-2
-8 -7 -6 -5 -4 -3 -2 -1 0 1 2
-4
-5 -4 -3 -2 -1 0 1 2 Real Axis
Real Axis
1 .5
3
1
2
0 .5
14 17
Im a g A xis
Im a g A xi s
-1
s(s + p1)(s + p2) -0 .5
-3
-4 -3 -2 -1 0 1 2 -1 .5
-7 -6 -5 -4 -3 -2 -1 0 1 2
R e al A x is
R e a l A xis
3 3
2 2
(s + z1)(s + z2)
1 1
Im a g A x i s
0 0
-3
-3
-4 -3 -2 -1 0 1 2
-1 0 -8 -6 -4 -2 0 2
R e a l A x is
R e a l A x is
The ROOT LOCUS Problem
G(S)= 1
(s + p1)(s + p2)(s + p3)(s + p4)
No Root Locus No Root Locus
3 2.5
19 1
21
1
0.5
Imag A xis
Imag A xis
0 0
-0.5
-1
-1
-1.5
-2
-2
-3 -2.5
-6 -5 -4 -3 -2 -1 0 1 2 -2 -1.5 -1 -0.5 0 0.5 1
Real Axis Real Axis
2 1.5
0.5
0.5
Im a g A xis
Im a g A xi s
0 0
-0.5
20 22
-0.5
-1
-1
-1.5
-2 -1.5
-2 -1.5 -1 -0.5 0 0.5 1 -1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1
Real Axis Real Axis
Design by Root Locus
Problem 1
For the system given, plot the root locus as the parameter K1 is varied from 0 to
∞ given λ = 2.
Solution:
We need to put the
characteristic equation for
the closed loop system in
the standard form for a root
locus plot, viz.,
The characteristic equation can be found by Block diagram reduction. Note that the
summations in the feedback path all occur at essentially the same signal point. Hence
we could think of the block diagram having the following nested feedback loop
structure:
Design by Root Locus
Problem 1
We can then successively replace the simple feedback loops with single blocks
according to the G/(1+GH) rule:
Design by Root Locus
Problem 1
This leads to
Hence, in standard form with K1 as the free parameter, the characteristic equation is
Design by Root Locus
Problem 1
•The asymptotes are determined to be ± 90° and intersect the real axis at
λ =2
Problem 1
Problem 1
Now, repeat the problem with λ = 5. What is special about this value?
Solution
For λ = 5, the characteristic equation becomes
rlocus(GH)
Design by Root Locus
Problem 1
•The branch of the root locus beginning at the open-loop pole s = -5 terminates at the zero
at the same location. That is, one closed-loop pole is "trapped" at s = -5.
•This conclusion may also be reached by examining the characteristic equation again, in the form
•Note that there is no zero in the closed-loop transfer function from R(s) to Y(s):
Design by Root Locus
Problem 2
For the system given in problem 1, plot the root locus as the parameter λ is varied from 0
to ∞ given K1 = 2.
Solution
We need to re-arrange the characteristic equation so that λ is the free parameter:
That is,
Design by Root Locus
Problem 2
Using MATLAB for the root locus:
GH = tf([1 11 0], [1 11 4 20]);
rlocus(GH)
Design by Root Locus
Problem 3
Consider the system in figure
(a) Use Routh's criterion to determine the regions in the (K1, K2) plane for which
the system is stable.
(b) Use the root-locus method to verify your answer to part (a).
Design by Root Locus
Problem 2 .. Solution
•The characteristic equation for the closed-loop system is
Problem 2 .. Solution
Confirm this with a root locus plot for
Problem 2
•The upper branch crosses the imaginary axis when s = jω .
Substituting this in the characteristic equation, and collecting
real and imaginary parts of the equation, gives
•The first equation yields the solutions ω = 0 (corresponding to
the open-loop pole, when K1 = 0), and w = ± 0.707 rad/s.
Problem 2
Design by Root Locus
Problem 2
•Now consider the case when K2 ≠ 0, where the characteristic equation can now be written as
or
•Necessary conditions for stability are that K1 > 0 and K2 > 0 and others from the R-H table
•Since K1 > 0, this means that K1 + 2.25K2 < 0.75 s1: [(0.75-K1)K1
− 2.25Κ 1Κ 2]/(1.5 − Κ 1)
•Together with K1 > 0 and K2 > 0, this defines
s0:1.5K1K2
an area in the (K1, K2) plane shown
Design by Root Locus
Problem 2
•We can use a root locus analysis to check, say, two points in this plane.
•For example, let K2 = 0.25, and draw the root locus for K1.
•The characteristic equation is
•Hence, we want to draw the root locus for the equation
Problem 2
•The second equation has solutions ω = 0 (corresponding to K1 = 0), and ω 2
=
K1/1.5.
•Substituting the latter solution in the first equation yields
•Hence, K1 = 0, or K1 = 0.1875, for which the loci cross the imaginary axis at ω = ± 0.3536.
•This confirms the Routh analysis, which required that K1 > 0 and K1 + 2.25K2 < 0.75;
i.e., K1 < 0.75 – 2.25× 0.25 = 0.1875.
•As a further test, let K1 = 0.5, say, and draw the root locus for K2.
•The characteristic equation is
That is,
Design by Root Locus
Problem 2
•The axis crossing analysis yields
∀ ω 2
= 0.5/1.5; , ω = ± 0.5774.
The Routh analysis requires K2 > 0 and K2 < (0.75 – 0.5)/2.25 = 0.1111.