Hardware Reference
In-Depth Information
Substituting equation 3.72 into equation 3.73 yields
λ 1 λ q−1
a i,q
···
λ 1
1
1
λ 2 λ q−1
a i,q−1
.
a i,0
···
λ 2
1
2
.
λ q m+n λ q−1
m+n ,
··· λ m+n
1
−N a 1 ) −1 N p 1 )d m 1 )
−N a 2 ) −1 N p 2 )d m 2 )
.
−N a m+n ) −1 N p m+n )d m m+n )
=
.
(3.74)
If q + l = m + n d +1andifD does not have repeated roots, then the
first term in the above equation is a non-singular square matrix, and n i and n
can be easily calculated by solving equation (3.74). Furthermore, d has been
specified by equation (3.70). Hence we have the I V = d .
Example: Let the actuator transfer function be G p (s)= 1×10 4
s 2 ,whichis
controlled by a PD controller
G c (s)= 0.014s +1
0.00014s +1 .
(3.75)
Following the above equations,
s 3 + 714.3s 2
s 3 + 714.3s 2 +10 5 s +7.143 × 10 6 ,
S =
100000s +7.143 × 10 6
s 3 + 714.3s 2 +10 5 s +7.143 × 10 6 .
T =
(3.76)
The poles of the closed-loop system are:−848.97,−75.516 + 77.95i,−75.516−
77.95i.
We wish to inject the IVC signal from the reference signal point r.Letthe
transfer function from the initial position value to plant output considering I v :
y =(S + I v T )y(0).
(3.77)
Solution 1:
Following the equations given above, we have N a =(s 2 (s+714.2857). The
roots of N are λ i = 100 × [−5.5802,−0.7813 + 0.8183i,−0.7813−0.8183i]. d m
is to be 20 times faster than these poles which can be determined. Substituting
into equation 3.74 we have a i =9.7843×10
−6 , 6.6852×10
−3 , 3.1529×10
−2 , 0.
Hence
5.591 × 10 5 s 3 +3.82 × 10 8 s 2 +1.802 × 10 9 s
s 4 +1.304 × 10 4 s 3 +2.625 × 10 7 s 2 +5.895 × 10 10 s +4.082 × 10 12 .
(3.78)
I v = n
d =
Search WWH ::




Custom Search