CN103353764B - Electromechanical model discrimination method in position closed loop kinetic control system - Google Patents

Abstract

The present invention relates to the position-force control device in position closed loop kinetic control system.Particularly relate to a kind of based on electromechanical model discrimination method in the position closed loop kinetic control system of Orthogonal Decomposition alternative manner.By building orthogonal vector basis function, Controlling model being said at basis function in the orthogonal vector space built and projects, adopting the method for iterative learning to carry out electromechanical model parameter iteration identification in position closed loop kinetic control system along basis function direction of principal axis.The inventive method significantly improves tracking accuracy to position and high speed response property, meets the requirement that kinetic control system high speed is high-precision.

Description

Electromechanical model discrimination method in position closed loop kinetic control system

Technical field

The present invention relates to the position-force control device in position closed loop kinetic control system.Particularly relate to a kind of based on electromechanical model discrimination method in the position closed loop kinetic control system of Orthogonal Decomposition alternative manner.

Background technology

In the position-force control system of Mechatronic Systems, positioner that must be reasonable in design is with quick and precisely tracing positional instruction.Current kinetic control system position adopts the PID controller design method based on experience, and the positioner optimized needs to carry out system analysis and design based on dynamo-electric plant model, therefore, the design of model on positioner of controlled dynamo-electric object has direct, important impact, is necessary that design attitude ring controller obtains good control effects on the basis of the rational dynamo-electric object model picked out.

Summary of the invention

Consider above description, an object of the present invention is to provide a kind of based on electromechanical model discrimination method in the position closed loop kinetic control system of Orthogonal Decomposition alternative manner.

Based on electromechanical model discrimination method in the position closed loop kinetic control system of Orthogonal Decomposition alternative manner, it is characterized in that: by building orthogonal vector basis function, Controlling model is said at basis function in the orthogonal vector space built and projects, adopt the method for iterative learning to carry out electromechanical model parameter iteration identification in position closed loop kinetic control system along basis function direction of principal axis;

Comprise the following steps:

1), by input signal r (t), projection matrix A is calculated;

2), impulse response matrix [G is calculated _s], [G _p];

3), M is calculated according to projection matrix A _p, M _s, M _p=A ^t[G _p] A, M _s=A ^t[G _s] A

4), make iterations k=0, and set iterative initial value

5), the projective parameter of the error of calculation wherein H is permanent gain matrix;

6), new parameter is calculated

7), judge whether be less than given iteration error, if be less than iteration error, then iteration terminates, otherwise repeats 5-7.

Further, step (1) is made up of following steps:

(1.1) Q matrix is obtained, $Q=\left[\begin{array}{cccc}r\left(0\right)& \stackrel{.}{r}\left(0\right)& \·\·\·& r^{\left(n\right)}\left(0\right)\\ r\left(T_s\right)& \stackrel{\·}{r}\left(T_s\right)& \·\·\·& r^{\left(n\right)}\left(T_s\right)\\ \·& \·& & \\ \·& \·& \·\·\·& \·\·\·\\ \·& \·& & \\ r\left({\mathrm{NT}}_s\right)& \stackrel{\·}{r}\left({\mathrm{NT}}_s\right)& \·\·\·& r^{\left(n\right)}\left({\mathrm{NT}}_s\right)\end{array}\right],$ Wherein Ts is the sampling period, and N is sampling number;

(1.2) Orthogonal Decomposition is carried out to Q matrix, Q=UR, wherein U ^tu=I, R are upper triangular matrix;

(1.3) n+1 the column vector choosing matrix U builds vector space, then projection matrix A=U=[f as orthogonal vector basis function ₁(t), f ₂(t) ..., f _n+1(t)].

Further, step (2) is made up of following steps:

(2.1) system output signal Y (s)=G _p(s) R (s)+G _s(s) F (s) R (s)+G _ss () W (s), carries out employing discretize vector mode to this formula and describes each signal and can obtain following equation:

$\left[\begin{array}{c}{y}^k\left(0\right)\\ y^k\left(1\right)\\ \·\\ \·\\ \·\\ y^k(N-1)\end{array}\right]=\left[G_P\right]\left[\begin{array}{c}{r}^{*}\left(0\right)\\ r^{*}\left(1\right)\\ \·\\ \·\\ \·\\ r^{*}(N-1)\end{array}\right]-[G_S]\left[\begin{array}{c}{u}_f^{*}\left(0\right)\\ u_f^{*}\left(1\right)\\ \·\\ \·\\ \·\\ u_f^{*}(N-1)\end{array}\right]+[G_S]\left[\begin{array}{c}{\mathrm{\ω}}^k\left(0\right)\\ {\mathrm{\ω}}^k\left(1\right)\\ \·\\ \·\\ \·\\ {\mathrm{\ω}}^k(N-1)\end{array}\right]$

Wherein, y ^kwhen () represents kth time iteration j, jth+1 element of output signal y, r ^*for the instruction of input ideal trajectory, N=T/T _s, the time span t ∈ [0, T] of each iteration identification, the sampling period is T _s.

The invention has the beneficial effects as follows: the inventive method significantly improves tracking accuracy to position and high speed response property, meets the requirement that kinetic control system high speed is high-precision.

Accompanying drawing explanation

Fig. 1 Control system architecture figure

Fig. 2 is based on iteration territory Control loop control block diagram

Embodiment

Below by embodiment, and by reference to the accompanying drawings, technical scheme of the present invention is described in further detail.

Based on electromechanical model discrimination method in the position closed loop kinetic control system of Orthogonal Decomposition alternative manner, it is characterized in that: by building orthogonal vector basis function, Controlling model is said at basis function in the orthogonal vector space built and projects, adopt the method for iterative learning to carry out electromechanical model parameter iteration identification in position closed loop kinetic control system along basis function direction of principal axis;

Comprise the following steps:

1), by input signal r (t), projection matrix A is calculated;

2), impulse response matrix [G is calculated _s], [G _p];

3), M is calculated according to projection matrix A _p, M _s, M _p=A ^t[G _p] A, M _s=A ^t[G _s] A

4), make iterations k=0, and set iterative initial value

5), according to the projective parameter of Fig. 2 error of calculation wherein H is permanent gain matrix;

6), new parameter is calculated ${\stackrel{\→}{\mathrm{\λ}}}^k{|}_{k=0}={\stackrel{\→}{\mathrm{\λ}}}^0;$

7), judge whether be less than given iteration error, if be less than iteration error, then iteration terminates, otherwise repeats 5-7.

Further, step (1) is made up of following steps:

(1.1) Q matrix is obtained, $Q=\left[\begin{array}{cccc}r\left(0\right)& \stackrel{.}{r}\left(0\right)& \·\·\·& r^{\left(n\right)}\left(0\right)\\ r\left(T_s\right)& \stackrel{\·}{r}\left(T_s\right)& \·\·\·& r^{\left(n\right)}\left(T_s\right)\\ \·& \·& & \\ \·& \·& \·\·\·& \·\·\·\\ \·& \·& & \\ r\left({\mathrm{NT}}_s\right)& \stackrel{\·}{r}\left({\mathrm{NT}}_s\right)& \·\·\·& r^{\left(n\right)}\left({\mathrm{NT}}_s\right)\end{array}\right],$ Wherein Ts is the sampling period, and N is sampling number;

(1.2) Orthogonal Decomposition is carried out to Q matrix, Q=UR, wherein U ^tu=I, R are upper triangular matrix;

(1.3) n+1 the column vector choosing matrix U builds vector space, then projection matrix A=U=[f as orthogonal vector basis function ₁(t), f ₂(t) ..., f _n+1(t)].

Further, step (2) is made up of following steps:

(2.1) Y (s)=G is known by Fig. 1 _p(s) R (s)+G _s(s) F (s) R (s)+G _ss () W (s), carries out employing discretize vector mode to this formula and describes each signal and can obtain following equation:

$\left[\begin{array}{c}{y}^k\left(0\right)\\ y^k\left(1\right)\\ \·\\ \·\\ \·\\ y^k(N-1)\end{array}\right]=\left[G_P\right]\left[\begin{array}{c}{r}^{*}\left(0\right)\\ r^{*}\left(1\right)\\ \·\\ \·\\ \·\\ r^{*}(N-1)\end{array}\right]-\left[G_S\right]\left[\begin{array}{c}{u}_f^{*}\left(0\right)\\ u_f^{*}\left(1\right)\\ \·\\ \·\\ \·\\ u_f^{*}(N-1)\end{array}\right]+\left[G_S\right]\left[\begin{array}{c}{\mathrm{\ω}}^k\left(0\right)\\ {\mathrm{\ω}}^k\left(1\right)\\ \·\\ \·\\ \·\\ {\mathrm{\ω}}^k(N-1)\end{array}\right]$

Wherein, y ^kwhen () represents kth time iteration j, jth+1 element of output signal y, r ^*for the instruction of input ideal trajectory, N=T/T _s, the time span t ∈ [0, T] of each iteration identification, the sampling period is T _s.

The inventive method significantly improves tracking accuracy to position and high speed response property.

Content described in this instructions embodiment is only enumerating the way of realization of inventive concept; protection scope of the present invention should not be regarded as being only limitted to the concrete form that embodiment is stated, protection scope of the present invention also and conceive the equivalent technologies means that can expect according to the present invention in those skilled in the art.

Claims (3)

1. electromechanical model discrimination method in position closed loop kinetic control system, it is characterized in that: by building orthogonal vector basis function, Controlling model is projected in the orthogonal vector space constructed by basis function, adopts the method for iterative learning to carry out electromechanical model parameter iteration identification in position closed loop kinetic control system along basis function direction of principal axis;

Comprise the following steps:

1), by input signal r (t), projection matrix A is calculated;

2), impulse response matrix [G is calculated _s], [G _p];

3), M is calculated according to projection matrix A _p, M _s, M _p=A ^t[G _p] A, M _s=A ^t[G _s] A

4), make iterations k=0, and set iterative initial value

5), the projective parameter of the error of calculation wherein H is permanent gain matrix;

6), new parameter is calculated

7), judge whether be less than given iteration error, if be less than iteration error, then iteration terminates, otherwise repeats 5-7.

2. electromechanical model discrimination method in position closed loop kinetic control system as claimed in claim 1, is characterized in that: step (1) is made up of following steps:

(1.1) Q matrix is obtained, $Q=\left[\begin{array}{cccc}r\left(0\right)& \stackrel{\·}{r}\left(0\right)& \mathrm{...}& r^{\left(n\right)}\left(0\right)\\ r\left(T_s\right)& \stackrel{\·}{r}\left(T_s\right)& \mathrm{...}& r^{\left(n\right)}\left(T_s\right)\\ .& .& & .\\ .& .& \mathrm{...}& .\\ .& .& & .\\ r\left({\mathrm{NT}}_s\right)& \stackrel{\·}{r}\left({\mathrm{NT}}_s\right)& \mathrm{...}& r^{\left(n\right)}\left({\mathrm{NT}}_s\right)\end{array}\right],$ Wherein Ts is the sampling period, and N is sampling number;

(1.2) Orthogonal Decomposition is carried out to Q matrix, Q=UR, wherein U ^tu=I, R are upper triangular matrix;

(1.3) n+1 the column vector choosing matrix U builds vector space, then projection matrix A=U=[f as orthogonal vector basis function ₁(t), f ₂(t) ..., f _n+1(t)].

3. electromechanical model discrimination method in position closed loop kinetic control system as claimed in claim 1, is characterized in that: step (2) is made up of following steps:

(2.1) system output signal Y (s)=G _p(s) R (s)+G _s(s) F (s) R (s)+G _ss () W (s), carries out employing discretize vector mode to this formula and describes each signal and can obtain following equation:

$\left[\begin{array}{c}{y}^k\left(0\right)\\ y^k\left(1\right)\\ .\\ .\\ .\\ y^k\left(N-1\right)\end{array}\right]=\[G_P\]\left[\begin{array}{c}{r}^{*}\left(0\right)\\ r^{*}\left(1\right)\\ .\\ .\\ .\\ r^{*}\left(N-1\right)\end{array}\right]-\[G_S\]\left[\begin{array}{c}{u}_f^{*}\left(0\right)\\ u_f^{*}\left(1\right)\\ .\\ .\\ .\\ u_f^{*}\left(N-1\right)\end{array}\right]+\[G_S\]\left[\begin{array}{c}{\mathrm{\ω}}^k\left(0\right)\\ {\mathrm{\ω}}^k\left(1\right)\\ .\\ .\\ .\\ {\mathrm{\ω}}^k\left(N-1\right)\end{array}\right]$

Wherein, y ^kwhen () represents kth time iteration j, jth+1 element of output signal y, r ^*for the instruction of input ideal trajectory, N=T/T _s, the time span t ∈ [0, T] of each iteration identification, the sampling period is T _s.