一、问题描述
设控制对象的传递函数为
W
p
(
s
)
=
k
1
T
1
2
s
2
+
2
T
1
ξ
1
s
+
1
(1)
W_{p}(s)=\frac{k_{1}}{T_{1}^{2} s^{2}+2 T_{1} \xi_{1} s+1} \tag{1}
Wp(s)=T12s2+2T1ξ1s+1k1(1)
参数
k
1
k_1
k1,
T
1
T_{1}
T1 和
ξ
1
\xi_{1}
ξ1 随时间而变的变化规律为
k
1
(
t
)
=
1.12
−
0.008
t
,
T
1
(
t
)
=
0.036
+
0.004
t
,
ξ
1
(
t
)
=
0.8
−
0.01
t
(2)
k_{1}(t)=1.12-0.008 t, \quad T_{1}(t)=0.036+0.004 t, \quad \xi_{1}(t)=0.8-0.01 t \tag{2}
k1(t)=1.12−0.008t,T1(t)=0.036+0.004t,ξ1(t)=0.8−0.01t(2)
设参考模型的传递函数为
W
m
(
s
)
=
1
0.0
8
2
s
2
+
2
×
0.08
×
0.75
s
+
1
(3)
W_{m}(s)=\frac{1}{0.08^{2} s^{2}+2 \times 0.08 \times 0.75 s+1} \tag{3}
Wm(s)=0.082s2+2×0.08×0.75s+11(3)
用超稳定性理论设计模型参考自适应系统。
假定系统参考输入:
r ( t ) r(t) r(t) 是方波信号,周期为4s,振幅为 ± 2 \pm 2 ±2。
设计自适应规律,给出仿真结果。
二、问题建模
本次仿真实验主要针对的是带状态变量滤波器(情况1)下的并联模型参考自适应控制系统。设参考模型方程为:
(
∑
i
=
0
n
a
m
i
p
i
)
y
m
=
(
∑
i
=
0
m
b
m
i
p
i
)
r
,
a
m
n
=
1
(4)
\left(\sum_{i=0}^{n} a_{m i} p^{i}\right) y_{m}=\left(\sum_{i=0}^{m} b_{m i} p^{i}\right) r, a_{m n}=1 \tag{4}
(i=0∑namipi)ym=(i=0∑mbmipi)r,amn=1(4)
接在参考模型输出端的状态变量滤波器的方程为:
(
∑
i
=
0
n
−
1
c
i
p
i
)
y
m
f
=
y
m
,
c
n
−
1
=
1
(5)
\left(\sum_{i=0}^{n-1} c_{i} p^{i}\right) y_{m f}=y_{m}, c_{n-1}=1 \tag{5}
(i=0∑n−1cipi)ymf=ym,cn−1=1(5)
接在可调系统输入端的状态变量滤波器的方程为:
(
∑
i
=
0
n
−
1
c
i
p
i
)
r
f
=
r
,
c
n
−
1
=
1
(6)
\left(\sum_{i=0}^{n-1} c_{i} p^{i}\right) r_{f}=r, c_{n-1}=1 \tag{6}
(i=0∑n−1cipi)rf=r,cn−1=1(6)
可调系统方程为:
(
∑
i
=
0
n
a
s
i
(
v
,
t
)
p
i
)
y
s
f
=
(
∑
i
=
0
m
b
s
i
(
v
,
t
)
p
i
)
r
f
,
a
s
n
(
v
,
t
)
=
1
(7)
\left(\sum_{i=0}^{n} a_{s i}(v, t) p^{i}\right) y_{s f}=\left(\sum_{i=0}^{m} b_{s i}(v, t) p^{i}\right) r_{f}, a_{s n}(v, t)=1 \tag{7}
(i=0∑nasi(v,t)pi)ysf=(i=0∑mbsi(v,t)pi)rf,asn(v,t)=1(7)
广义输出误差为:
ε
f
=
y
m
f
−
y
s
f
(8)
\varepsilon_{f}=y_{m f}-y_{s f} \tag{8}
εf=ymf−ysf(8)
为保证等价的前向方块严格正实,引入串联补偿器:
v
=
D
(
p
)
ε
f
=
(
∑
i
=
0
n
−
1
d
i
p
i
)
ε
f
(9)
v=D(p) \varepsilon_{f}=\left(\sum_{i=0}^{n-1} d_{i} p^{i}\right) \varepsilon_{f} \tag{9}
v=D(p)εf=(i=0∑n−1dipi)εf(9)
对于可调系统中的可调参数
a
s
i
(
v
,
t
)
a_{si}(v,t)
asi(v,t) 及
b
s
i
(
v
,
t
)
b_{si}(v,t)
bsi(v,t),采取PI控制,则自适应规律为
a
s
i
(
v
,
t
)
=
∫
0
t
φ
1
i
(
v
,
t
,
τ
)
d
τ
+
φ
2
i
(
v
,
t
)
+
a
s
i
(
0
)
,
i
=
0
,
1
,
⋯
,
n
−
1
b
s
i
(
v
,
t
)
=
∫
0
t
ψ
1
i
(
v
,
t
,
τ
)
d
τ
+
ψ
2
i
(
v
,
t
)
+
b
s
i
(
0
)
,
i
=
0
,
1
,
⋯
,
m
(10)
\begin{aligned} a_{s i}(v, t)&=\int_{0}^{t} \varphi_{1 i}(v, t, \tau) d \tau+\varphi_{2 i}(v, t)+a_{s i}(0), i=0,1, \cdots, n-1 \\ b_{s i}(v, t)&=\int_{0}^{t} \psi_{1 i}(v, t, \tau) d \tau+\psi_{2 i}(v, t)+b_{s i}(0), i=0,1, \cdots, m \end{aligned} \tag{10}
asi(v,t)bsi(v,t)=∫0tφ1i(v,t,τ)dτ+φ2i(v,t)+asi(0),i=0,1,⋯,n−1=∫0tψ1i(v,t,τ)dτ+ψ2i(v,t)+bsi(0),i=0,1,⋯,m(10)
将参考模型与状态变量滤波器互换位置,可得到形式同(7)式的参考模型方程如下:
(
∑
i
=
0
n
a
m
i
p
i
)
y
m
f
=
(
∑
i
=
0
m
b
m
i
p
i
)
r
f
,
a
m
n
=
1
(11)
\left(\sum_{i=0}^{n} a_{m i} p^{i}\right) y_{m f}=\left(\sum_{i=0}^{m} b_{m i} p^{i}\right) r_{f}, a_{m n}=1 \tag{11}
(i=0∑namipi)ymf=(i=0∑mbmipi)rf,amn=1(11)
结合(7)式,(11)式和(8)式,可推导出
(
∑
i
=
0
n
a
m
i
p
i
)
ε
f
=
[
∑
i
=
0
n
(
a
s
i
−
a
m
i
)
p
i
]
y
s
f
+
[
∑
i
=
0
n
(
b
m
i
−
b
s
i
)
p
i
]
r
f
(12)
\left(\sum_{i=0}^{n} a_{m i} p^{i}\right) \varepsilon_{f}=\left[\sum_{i=0}^{n}\left(a_{s i}-a_{m i}\right) p^{i}\right] y_{s f}+\left[\sum_{i=0}^{n}\left(b_{m i}-b_{s i}\right) p^{i}\right] r_{f} \tag{12}
(i=0∑namipi)εf=[i=0∑n(asi−ami)pi]ysf+[i=0∑n(bmi−bsi)pi]rf(12)
令
ω
1
=
[
∑
i
=
0
n
(
a
s
i
−
a
m
i
)
p
i
]
y
s
f
+
[
∑
i
=
0
n
(
b
m
i
−
b
s
i
)
p
i
]
r
f
(13)
\omega_{1}=\left[\sum_{i=0}^{n}\left(a_{s i}-a_{m i}\right) p^{i}\right] y_{s f}+\left[\sum_{i=0}^{n}\left(b_{m i}-b_{s i}\right) p^{i}\right] r_{f} \tag{13}
ω1=[i=0∑n(asi−ami)pi]ysf+[i=0∑n(bmi−bsi)pi]rf(13)
则(12)式变成
(
∑
i
=
0
n
a
m
i
p
i
)
ε
f
=
ω
13
(14)
\left(\sum_{i=0}^{n} a_{m i} p^{i}\right) \varepsilon_{f}=\omega_{13} \tag{14}
(i=0∑namipi)εf=ω13(14)
将可调系统中的可调参数选定的自适应规律(10)式代入(13)式中,可得反馈方框的输出量
ω
\omega
ω 的形式如下:
ω
=
−
ω
1
=
−
{
∑
i
=
0
n
−
1
[
∫
0
t
φ
1
i
(
v
,
t
,
τ
)
d
τ
+
φ
2
i
(
v
,
t
)
+
a
s
i
(
0
)
−
a
m
i
]
p
i
}
y
s
f
+
{
∑
i
=
0
n
−
1
[
∫
0
t
ψ
1
i
(
v
,
t
,
τ
)
d
τ
+
ψ
2
i
(
v
,
t
)
+
b
s
i
(
0
)
−
b
m
i
]
p
i
}
r
f
(15)
\begin{aligned} \omega=-\omega_{1}=&-\left\{\sum_{i=0}^{n-1}\right. {\left.\left[\int_{0}^{t} \varphi_{1 i}(v, t, \tau) d \tau+\varphi_{2 i}(v, t)+a_{s i}(0)-a_{m i}\right] p^{i}\right\} y_{s f} } \\ &+\left\{\sum_{i=0}^{n-1}\left[\int_{0}^{t} \psi_{1 i}(v, t, \tau) d \tau+\psi_{2 i}(v, t)+b_{s i}(0)-b_{m i}\right] p^{i}\right\} r_{f} \end{aligned} \tag{15}
ω=−ω1=−{i=0∑n−1[∫0tφ1i(v,t,τ)dτ+φ2i(v,t)+asi(0)−ami]pi}ysf+{i=0∑n−1[∫0tψ1i(v,t,τ)dτ+ψ2i(v,t)+bsi(0)−bmi]pi}rf(15)
由波波夫积分不等式,及引理1和引理2,可得自适应规律
φ
1
i
=
−
k
a
i
(
t
−
τ
)
v
(
τ
)
p
i
y
s
f
(
τ
)
,
τ
≤
t
,
i
=
0
,
1
,
⋯
,
n
−
1
φ
2
i
=
−
k
a
i
′
(
t
)
v
(
t
)
p
i
y
s
f
(
t
)
,
i
=
0
,
1
,
⋯
,
n
−
1
ψ
1
i
=
k
b
i
(
t
−
τ
)
v
(
τ
)
p
i
r
f
(
τ
)
,
τ
≤
t
,
i
=
0
,
1
,
⋯
,
m
ψ
2
i
=
k
b
i
′
(
t
)
v
(
t
)
p
i
r
f
(
t
)
,
i
=
0
,
1
,
⋯
,
m
(16)
\begin{aligned} \varphi_{1 i}&=-k_{a i}(t-\tau) v(\tau) p^{i} y_{s f}(\tau), \quad \tau \leq t, i=0,1, \cdots, n-1 \\ \varphi_{2 i}&=-k_{a i}^{\prime}(t) v(t) p^{i} y_{s f}(t), \quad i=0,1, \cdots, n-1 \\ \psi_{1 i}&=k_{b i}(t-\tau) v(\tau) p^{i} r_{f}(\tau), \quad \tau \leq t, i=0,1, \cdots, m \\ \psi_{2 i}&=k_{b i}^{\prime}(t) v(t) p^{i} r_{f}(t), \quad i=0,1, \cdots, m \end{aligned} \tag{16}
φ1iφ2iψ1iψ2i=−kai(t−τ)v(τ)piysf(τ),τ≤t,i=0,1,⋯,n−1=−kai′(t)v(t)piysf(t),i=0,1,⋯,n−1=kbi(t−τ)v(τ)pirf(τ),τ≤t,i=0,1,⋯,m=kbi′(t)v(t)pirf(t),i=0,1,⋯,m(16)
式中,
k
a
i
(
t
−
τ
)
k_{a i}(t-\tau)
kai(t−τ) 和
k
b
i
(
t
−
τ
)
k_{b i}(t-\tau)
kbi(t−τ) 是正定标量积分核,它们的拉普拉斯变化式为在
s
=
0
s=0
s=0 处有一极点的正实传递函数;
k
a
i
′
k_{a i}^{\prime}
kai′ 和
k
b
i
′
k_{b i}^{\prime}
kbi′ 在
t
≥
0
t\ge0
t≥0 时为非负标量增益。
三、问题求解
将原题中给定的参考模型和可调系统的传递函数写成输入输出方程的形式:
(
0.0
8
2
p
2
+
2
×
0.08
×
0.75
p
+
1
)
y
m
=
r
(
T
1
2
p
2
+
2
T
1
ξ
1
p
+
1
)
y
s
f
=
k
1
r
f
(17)
\begin{gathered} &\left(0.08^{2} p^{2}+2 \times 0.08 \times 0.75 p+1\right) y_{m}=r \\ &\left(T_{1}^{2} p^{2}+2 T_{1} \xi_{1} p+1\right) y_{s f}=k_{1} r_{f} \end{gathered} \tag{17}
(0.082p2+2×0.08×0.75p+1)ym=r(T12p2+2T1ξ1p+1)ysf=k1rf(17)
再将上式写成首一古尔维兹多项式的形式:
(
p
2
+
2
×
0.08
×
0.75
0.0
8
2
p
+
1
0.0
8
2
)
y
m
=
1
0.0
8
2
r
(
p
2
+
2
T
1
ξ
1
T
1
2
+
1
T
1
2
)
y
s
f
=
k
1
T
1
2
r
f
(18)
\begin{gathered} \left(p^{2}+\frac{2 \times 0.08 \times 0.75}{0.08^{2}} p+\frac{1}{0.08^{2}}\right) y_{m}=\frac{1}{0.08^{2}} r \\ \left(p^{2}+\frac{2 T_{1} \xi_{1}}{T_{1}^{2}}+\frac{1}{T_{1}^{2}}\right) y_{s f}=\frac{k_{1}}{T_{1}^{2}} r_{f} \end{gathered} \tag{18}
(p2+0.0822×0.08×0.75p+0.0821)ym=0.0821r(p2+T122T1ξ1+T121)ysf=T12k1rf(18)
对照(7)式和(11)式,可知相关参数如下:
a m 1 = 2 × 0.08 × 0.75 0.0 8 2 = 18.75 a m 0 = 1 0.0 8 2 = 156.25 b m 0 = 1 0.0 8 2 = 156.25 a s 1 ( v , t ) = 2 T 1 ( t ) ξ 1 ( t ) T 1 2 ( t ) = 2 ( 0.8 − 0.01 t ) ( 0.036 + 0.004 t ) a s 0 ( v , t ) = 1 T 1 2 ( t ) = 1 ( 0.036 + 0.004 t ) 2 b s 0 ( v , t ) = k 1 ( t ) T 1 2 ( t ) = 1.12 − 0.008 t ( 0.036 + 0.004 t ) 2 (19) \begin{aligned} &a_{m 1}=\frac{2 \times 0.08 \times 0.75}{0.08^{2}}=18.75 \\ &a_{m 0}=\frac{1}{0.08^{2}}=156.25 \\ &b_{m 0}=\frac{1}{0.08^{2}}=156.25 \\ &a_{s 1}(v, t)=\frac{2 T_{1}(t) \xi_{1}(t)}{T_{1}^{2}(t)}=\frac{2(0.8-0.01 t)}{(0.036+0.004 t)} \\ &a_{s 0}(v, t)=\frac{1}{T_{1}^{2}(t)}=\frac{1}{(0.036+0.004 t)^{2}} \\ &b_{s 0}(v, t)=\frac{k_{1}(t)}{T_{1}^{2}(t)}=\frac{1.12-0.008 t}{(0.036+0.004 t)^{2}} \end{aligned} \tag{19} am1=0.0822×0.08×0.75=18.75am0=0.0821=156.25bm0=0.0821=156.25as1(v,t)=T12(t)2T1(t)ξ1(t)=(0.036+0.004t)2(0.8−0.01t)as0(v,t)=T12(t)1=(0.036+0.004t)21bs0(v,t)=T12(t)k1(t)=(0.036+0.004t)21.12−0.008t(19)
进而可知, a s 1 ( 0 ) ≈ 44.4 a_{s 1}(0) \approx 44.4 as1(0)≈44.4, a s 0 ( 0 ) ≈ 771.6 a_{s 0}(0) \approx 771.6 as0(0)≈771.6, b s 0 ( 0 ) ≈ 864.2 b_{s 0}(0) \approx 864.2 bs0(0)≈864.2。
设输出的广义误差为
ε
f
=
y
m
f
−
y
s
f
(20)
\varepsilon_{f}=y_{m f}-y_{s f} \tag{20}
εf=ymf−ysf(20)
串联补偿器方程为
v
=
D
(
p
)
ε
f
=
(
d
1
p
+
d
0
)
ε
f
(21)
v=D(p) \varepsilon_{f}=\left(d_{1} p+d_{0}\right) \varepsilon_{f} \tag{21}
v=D(p)εf=(d1p+d0)εf(21)
选取的自适应规律如下
a
s
i
(
v
,
t
)
=
∫
0
t
φ
1
i
(
v
,
t
,
τ
)
d
τ
+
φ
2
i
(
v
,
t
)
+
a
s
i
(
0
)
,
i
=
0
,
1
b
s
0
(
v
,
t
)
=
∫
0
t
ψ
10
(
v
,
t
,
τ
)
d
τ
+
ψ
20
(
v
,
t
)
+
b
s
0
(
0
)
(22)
\begin{aligned} a_{s i}(v, t)&=\int_{0}^{t} \varphi_{1 i}(v, t, \tau) d \tau+\varphi_{2 i}(v, t)+a_{s i}(0), i=0,1 \\ b_{s 0}(v, t)&=\int_{0}^{t} \psi_{10}(v, t, \tau) d \tau+\psi_{20}(v, t)+b_{s 0}(0) \end{aligned} \tag{22}
asi(v,t)bs0(v,t)=∫0tφ1i(v,t,τ)dτ+φ2i(v,t)+asi(0),i=0,1=∫0tψ10(v,t,τ)dτ+ψ20(v,t)+bs0(0)(22)
参考(16)式的形式,可得可调参数的自适应规律如下:
φ
10
=
−
k
a
0
(
t
−
τ
)
v
(
τ
)
y
s
f
(
τ
)
,
τ
≤
t
φ
20
=
−
k
a
0
′
(
t
)
v
(
t
)
y
s
f
(
t
)
φ
11
=
−
k
a
1
(
t
−
τ
)
v
(
τ
)
p
y
s
f
(
τ
)
,
τ
≤
t
φ
21
=
−
k
a
1
′
(
t
)
v
(
t
)
p
y
s
f
(
t
)
ψ
10
=
k
b
0
(
t
−
τ
)
v
(
τ
)
r
f
(
τ
)
,
τ
⩽
t
ψ
20
=
k
b
0
′
(
t
)
v
(
t
)
r
f
(
t
)
(23)
\begin{aligned} \varphi_{1 0}&=-k_{a 0}(t-\tau) v(\tau) y_{sf}(\tau), \quad \tau \le t \\ \varphi_{2 0}&=-k_{a 0}^{\prime}(t) v(t) y_{s f}(t) \\ \varphi_{1 1}&=-k_{a 1}(t-\tau) v(\tau) p y_{sf}(\tau), \quad \tau \le t \\ \varphi_{2 1}&=-k_{a 1}^{\prime}(t) v(t) p y_{s f}(t) \\ \psi_{1 0}&=k_{b 0}(t-\tau) v(\tau) r_{f}(\tau), \quad \tau \leqslant t \\ \psi_{2 0}&=k_{b 0}^{\prime}(t) v(t) r_{f}(t) \end{aligned} \tag{23}
φ10φ20φ11φ21ψ10ψ20=−ka0(t−τ)v(τ)ysf(τ),τ≤t=−ka0′(t)v(t)ysf(t)=−ka1(t−τ)v(τ)pysf(τ),τ≤t=−ka1′(t)v(t)pysf(t)=kb0(t−τ)v(τ)rf(τ),τ⩽t=kb0′(t)v(t)rf(t)(23)
式中, k a 0 ( t − τ ) k_{a 0}(t-\tau) ka0(t−τ)、 k a 1 ( t − τ ) k_{a 1}(t-\tau) ka1(t−τ) 和 k b 0 ( t − τ ) k_{b 0}(t-\tau) kb0(t−τ) 为正定积分核, k a 0 ′ ( t ) k_{a 0}^{\prime}(t) ka0′(t)、 k a 1 ′ ( t ) k_{a 1}^{\prime}(t) ka1′(t) 和 k b 0 ′ ( t ) k_{b 0}^{\prime}(t) kb0′(t) 对 ∀ t ≥ 0 \forall t \ge 0 ∀t≥0 均为非负标量增益。
下面再讨论一下引入的串联补偿器中的参数
d
0
d_0
d0 和
d
1
d_1
d1 的取值范围,系统的等价前向线性方块传递函数为:
h
(
s
)
=
d
1
(
s
)
+
d
0
s
2
+
a
m
1
s
+
a
m
0
(24)
h(s)=\frac {d_1(s)+d_0} {s^2+a_{m1}s+a_{m0}} \tag{24}
h(s)=s2+am1s+am0d1(s)+d0(24)
其对应的能控标准型如下:
e
˙
=
A
m
e
+
b
ω
1
v
=
d
T
e
(25)
\begin{aligned} \boldsymbol{\dot e} &= \boldsymbol{A_m} \boldsymbol{e} + b \omega_1 \\ v &= d^T \boldsymbol{e} \end{aligned} \tag{25}
e˙v=Ame+bω1=dTe(25)
式中,
e
=
[
ε
ε
˙
]
\boldsymbol{e}=\left[ \begin{matrix} \varepsilon \\ \dot \varepsilon \end{matrix} \right]
e=[εε˙],
A
m
=
[
0
1
−
a
m
0
−
a
m
1
]
\boldsymbol{A_m}=\left[ \begin{matrix} 0 & 1 \\ -a_{m0} & -a_{m1} \end{matrix} \right]
Am=[0−am01−am1],
b
=
[
0
1
]
b=\left[ \begin{matrix} 0 \\ 1 \end{matrix} \right]
b=[01],
d
=
[
d
0
d
1
]
d=\left[ \begin{matrix} d_0 \\ d_1 \end{matrix} \right]
d=[d0d1]。
要求
h
(
s
)
h(s)
h(s) 是一个严格正实传递函数,则必定存在正定对称矩阵
P
P
P 和
Q
Q
Q,使方程式(26)成立:
{
P
A
m
+
A
m
T
P
=
−
Q
P
b
=
d
(26)
\left\{ \begin{aligned} &P A_m + A_m^T P = -Q\\ &P b = d \end{aligned} \right. \tag{26}
{PAm+AmTP=−QPb=d(26)
由此可解得:
d
0
>
0
,
d
1
d
0
>
1
a
m
1
=
0.053
(27)
d_0 > 0, \quad \frac {d_1} {d_0} > \frac {1} {a_{m_1}} =0.053 \tag{27}
d0>0,d0d1>am11=0.053(27)
最终搭建的仿真模型框图如 图1 所示:
具体的 Simulink 仿真文件我已上传至百度网盘中,链接如下:experiment_3.slx_免费高速下载|百度网盘-分享无限制 (baidu.com)
输入信号与广义输出误差信号如 图2 所示:
输入信号与增益信号如 图3 所示:
输入信号与可调参数1的变化曲线如 图4 所示:
输入信号与可调参数2的变化曲线如 图5 所示:
参考书目
李言俊, 张科. 自适应控制理论及应用[M]. 西北工业大学出版社, 2005.