routh array
TRANSCRIPT
-
8/3/2019 Routh Array
1/6
EE 422G Notes: Chapter 6 Instructor: Cheung
Page 6-26
6-5 Routh Array
Both asymptotically stability and BIBO stability require checing whetherH(s) =
N(s)/D(s) has all its poles on the open left half plane.
How to check?
Solution 1: Factorize D(s):
0)Re(
))...()((
...)(
21
011
1
-
8/3/2019 Routh Array
2/6
EE 422G Notes: Chapter 6 Instructor: Cheung
Page 6-27
Example 6-8
564114)( 23 ++= ssssD
sign: Changed once =>one pole in the r.h.p
verification:
)8)(7)(1(564114)(
23
++=
++=
sssssssD
Example 6-9
1105)( 234 ++++= sssssD
1
015
11
105
111
10
211
212
3
4
=
==
==
ds
ccs
bbs
s
s
115
0)1(115
01
0501
151
15101
15
0115
15
10115
1
2
1
2
1
=
=
=
=
=
=
=
=
=
=
d
c
c
b
b
Sign: changed twice => two poles in r.h.p.
PDF Created with deskPDF PDF Writer - Trial :: http://www.docudesk.com
-
8/3/2019 Routh Array
3/6
EE 422G Notes: Chapter 6 Instructor: Cheung
Page 6-28
FAQ about Routh Array
1.Why do number of sign changes in Routh array has anything to do with thenumber of r.h.p. poles?
Let say ))()(()(321
pspspssD =
Expand it out you get, ...)()( 2321
3+++= spppssD
In this case, the Routh array looks like this:
...
...)(
...1
321
2
3
ppp
s
s
s
++
If all the poles are on the left half plane, )( 321 ppp ++ must be positive and
there will not be a sign change between the first and second row. If there is a
sign change, we must have at least ONE pole on the open right half plane.
2.What about the rest of the rows?A detail explanation of the Routh-Hurwitz criterion is beyond the scope of this
course. An elementary proof of this criterion can be found in the paper titled
Elementary proof of the Routh-Hurwitz test by G. Meinsma in Systems &
control Letters 25 (1995) p. 237-242.
3.What about poles on the imaginary axis?It is in fact possible to also keep track of the number of pure-imaginary poles
from the Routh array as well. However, we will not consider such procedurehere.
4.What happen when there are zeros in the first column?Excellent question! There are two cases
Case 1: Only the first element of the row is 0 but the rest of the row is neither
entirely zero nor empty.
Procedure:
1. replace 0 by a small 2. Find # sign changes in the first column for either >0 and
-
8/3/2019 Routh Array
4/6
EE 422G Notes: Chapter 6 Instructor: Cheung
Page 6-29
Example 6-10 3)( 234 ++++= sssssD
Case 2: the whole row is zero or zero occurs in the very last row.
This happens, for example, when all the odd (even) powers of the polynomial are
missing:
4 2( ) ( 2 )( 2 )( 3 )( 3 ) 13 36D s s j s j s j s j s s= + + = + +
4
3
1 13 36
0 0 0
s
s
Solution:
1. Construct an auxiliary polynomial based on the row before the all-zerorow
2. Replace the all-zero row by the DERIVATIVE of the auxiliarypolynomial.
Why? As motivated by the previous example, a zero row implies that a polynomial
D(s) has only even or odd power. It turns out in this case, D(s) and D(s)+D(s)have exactly the same numbers of r.h.p. poles (proof beyond scope). As the goal is
just to find the # of r.h.p. poles, we can use D(s) as a surrogate to continue the
procedure.
PDF Created with deskPDF PDF Writer - Trial :: http://www.docudesk.com
-
8/3/2019 Routh Array
5/6
EE 422G Notes: Chapter 6 Instructor: Cheung
Page 6-30
Example 6-11
13333)( 234567 +++++++= ssssssssD
Two sign change two
poles on the open right half
plane
PDF Created with deskPDF PDF Writer - Trial :: http://www.docudesk.com
-
8/3/2019 Routh Array
6/6
EE 422G Notes: Chapter 6 Instructor: Cheung
Page 6-31
Sample Application of Routh Array: Range of system parameters.
Example: )1(33)( 23 kssssD ++++=
kcs
kbs
ks
s
+=
=
+
1
3/)8(
13
31
10
11
2
3
k
K
kkc
kkb
+=
+=
=
+=
13/)8(
03)1(3/)8(
38
3)1(133
1
1
Stable system
18
101
80
3
8
>>
>>+
>>
k
kk
kk
to ensure system stable!
PDF Created with deskPDF PDF Writer - Trial :: http://www.docudesk.com