欠阻尼谐振腔模型状态转移矩阵计算

欠阻尼谐振腔模型状态转移矩阵计算

如图所示,具有定基座的线性阻尼理想谐振腔。物体质量为 m m m,位移 δ \delta δ k s k_s ks为弹簧常数, k d k_d kd为阻力系数。
欠阻尼谐振腔模型状态转移矩阵计算_第1张图片

  1. 根据牛顿第二定律,可得:
    F ( t ) = m a ( t ) = m [ d 2 δ ( t ) d t 2 ] = − k s δ ( t ) − k d d δ ( t ) d t \\F(t)=ma(t)=m[\frac{d^2\delta(t)}{dt^2}]=-ks\delta(t)-k_d\frac{d\delta(t)}{dt} F(t)=ma(t)=m[dt2d2δ(t)]=ksδ(t)kddtdδ(t)
  2. 写为微分方程形式:
    m [ d 2 δ ( t ) d t 2 ] + k d d δ ( t ) d t + k s δ ( t ) = 0 \\m[\frac{d^2\delta(t)}{dt^2}]+k_d\frac{d\delta(t)}{dt}+ks\delta(t)=0 m[dt2d2δ(t)]+kddtdδ(t)+ksδ(t)=0
  3. 列出状态空间模型:
    x 1 = δ x_1=\delta x1=δ x 2 = δ ˙ x_2=\dot{\delta} x2=δ˙,则:
    d d t [ x 1 ( t ) x 2 ( t ) ] = F [ x 1 ( t ) x 2 ( t ) ] 令 x = [ δ δ ˙ ] , 则 d d t [ δ δ ˙ ] = F x 即 x ˙ = F x F = [ 0 1 − k s m − k d m ] , ( F 为 动 态 系 数 矩 阵 ) \\\frac{d}{dt} \left[ \begin{matrix} x_1(t) \\ x_2(t) \end{matrix} \right]=F\left[ \begin{matrix} x_1(t) \\ x_2(t) \end{matrix} \right] \\令x= \left[ \begin{matrix} \delta \\ \dot{\delta} \end{matrix} \right],则\frac{d}{dt}\left[ \begin{matrix} \delta \\ \dot{\delta} \end{matrix} \right]=Fx \\即\dot{x}=Fx \\F=\left[ \begin{matrix} 0 & 1\\ -\frac{k_s}{m} & -\frac{k_d}{m} \end{matrix} \right],(F为动态系数矩阵) dtd[x1(t)x2(t)]=F[x1(t)x2(t)]x=[δδ˙],dtd[δδ˙]=Fxx˙=FxF=[0mks1mkd],(F)
  4. F F F的特征值:
    d e t ( λ I − F ) = λ 2 + k d m λ + k s m \\det(\lambda I-F)=\lambda^2+\frac{k_d}{m}\lambda+\frac{k_s}{m} det(λIF)=λ2+mkdλ+mks
    已知一元二次方程的解为:
    λ 1 = 1 2 ( − k d m + k d 2 m 2 − 4 k s m ) λ 2 = 1 2 ( − k d m − k d 2 m 2 − 4 k s m ) \\\lambda_1=\frac{1}{2}(-\frac{k_d}{m}+\sqrt{\frac{k^2_d}{m^2}-\frac{4k_s}{m}}) \\\lambda_2=\frac{1}{2}(-\frac{k_d}{m}-\sqrt{\frac{k^2_d}{m^2}-\frac{4k_s}{m}}) λ1=21(mkd+m2kd2m4ks )λ2=21(mkdm2kd2m4ks )

因为欠阻尼,所以 k d 2 m 2 − 4 k s m < 0 \frac{k^2_d}{m^2}-\frac{4k_s}{m}<0 m2kd2m4ks<0

那么,特征方程的根是一对共轭复根 λ 1 , 2 = α ± w i \lambda_{1,2}=\alpha\pm w i λ1,2=α±wi,则微分方程的通解为:
y = e α t ( a c o s w t + b s i n w t ) , 且 λ 1 , 2 = − k d 2 m ± k s m − k d 2 4 m 2 × i \\y=e^{\alpha t}(acoswt+bsinwt),且 \\\lambda_{1,2}=-\frac{k_d}{2m}\pm\sqrt{\frac{k_s}{m}-\frac{k^2_d}{4m^2}} \times i y=eαt(acoswt+bsinwt),λ1,2=2mkd±mks4m2kd2 ×i
τ = 2 m k d , w = k s m − k d 2 4 m 2 \tau=\frac{2m}{k_d},w=\sqrt{\frac{k_s}{m}-\frac{k^2_d}{4m^2}} τ=kd2m,w=mks4m2kd2 则:
δ ( t ) = a e − t / τ c o s ( w t ) + b e − t / τ s i n ( w t ) \\ \delta(t)=ae^{-t/\tau}cos(wt)+be^{-t/\tau}sin(wt) δ(t)=aet/τcos(wt)+bet/τsin(wt)
这个解可以用实数变量 a , b a,b a,b表示为下列状态空间形式:
[ δ δ ˙ ] = e − t / τ [ c o s w t s i n w t − c o s w t τ − w s i n w t w c o s w t − s i n w t τ ] [ a b ] ( 式 1 ) \left[ \begin{matrix} \delta \\ \dot{\delta} \end{matrix} \right]=e^{-t/\tau} \left[ \begin{matrix} coswt & sinwt \\ -\frac{coswt}{\tau}-wsinwt & wcoswt-\frac{sinwt}{\tau} \end{matrix} \right] \left[ \begin{matrix} a \\ b \end{matrix} \right](式1) [δδ˙]=et/τ[coswtτcoswtwsinwtsinwtwcoswtτsinwt][ab]1
当初值为0时有:
δ ( 0 ) = a , δ ( 0 ) ˙ = − a τ + w b \\ \delta(0)=a,\dot{\delta(0)}=-\frac{a}{\tau}+wb δ(0)=a,δ(0)˙=τa+wb

[ a b ] = [ 1 0 1 w τ 1 w ] [ δ ( 0 ) δ ( 0 ) ˙ ] ( 式 2 ) \left[ \begin{matrix} a\\ b \end{matrix} \right]= \left[ \begin{matrix} 1 & 0 \\ \frac{1}{w\tau} & \frac{1}{w} \end{matrix} \right] \left[ \begin{matrix} \delta(0) \\ \dot{\delta(0)} \end{matrix} \right] (式2) [ab]=[1wτ10w1][δ(0)δ(0)˙]2
将(式2)带入(式1)得:
[ δ δ ˙ ] = e − t / τ [ c o s w t s i n w t − c o s w t τ − w s i n w t w c o s w t − s i n w t τ ] [ 1 0 1 w τ 1 w ] [ δ ( 0 ) δ ( 0 ) ˙ ] \left[ \begin{matrix} \delta \\ \dot{\delta} \end{matrix} \right]=e^{-t/\tau} \left[ \begin{matrix} coswt & sinwt \\ -\frac{coswt}{\tau}-wsinwt & wcoswt-\frac{sinwt}{\tau} \end{matrix} \right] \left[ \begin{matrix} 1 & 0 \\ \frac{1}{w\tau} & \frac{1}{w} \end{matrix} \right] \left[ \begin{matrix} \delta(0) \\ \dot{\delta(0)} \end{matrix} \right] [δδ˙]=et/τ[coswtτcoswtwsinwtsinwtwcoswtτsinwt][1wτ10w1][δ(0)δ(0)˙]
x ( t ) = ϕ ( t ) x ( 0 ) x(t)=\phi(t)x(0) x(t)=ϕ(t)x(0),故可求得状态转移矩阵:
ϕ ( t ) = e − t / τ w τ 2 [ τ [ w τ c o s w t + s i n w t ] τ 2 s i n w t − ( 1 + w 2 τ 2 ) s i n w t [ w τ 2 c o s w t − τ s i n w t ] ] \\\phi(t)=\frac{e^{-t/\tau}}{w\tau^2} \left[ \begin{matrix} \tau[w\tau coswt+sinwt] &\tau^2sinwt \\ -(1+w^2\tau^2)sinwt & [w\tau^2coswt-\tau sinwt] \end{matrix} \right] ϕ(t)=wτ2et/τ[τ[wτcoswt+sinwt](1+w2τ2)sinwtτ2sinwt[wτ2coswtτsinwt]]

你可能感兴趣的:(Kalman,Filter)