CN1945470A - Two freedom decoupling smith pre-evaluating control system of industrial multiple variable time lag process - Google Patents

Abstract

A kind of two degrees of freedom decouple Smith estimate control system with industrial multivariable process of multi-skewing, is composed of the n-dimensional set-point tracking controller, n-dimensional perturbation estimator, controlled process identification module and two multi-route signals mixer, in which, n is the dimension of the charged multivariable process. Through setting internal control loop between the process of input and output, and perturbation estimator, it effectively suppresses the loaded jamming signals in the process, to obtain a smooth process output. The given value response of system is adjusted in the way of open-loop control by the set-point tracking controller, so that the given value response and process loading jamming respond can be adjusted respectively and independently. Meanwhile, the impact of the system skewing to performance can be effectively compensated. The control system of the invention can achieve the notable decoupling between the output responses, maintain a good stability, and adapt the modeling errors of the charged process in a wide range and process parameter perturbation.

Description

The two freedom decoupling Smith Prediction Control system of industry multivariable time-lag process

Technical field

The present invention relates to a kind of system, specifically is a kind of two freedom decoupling Smith Prediction Control system of industrial multivariable process of tool multiple time delay, belongs to industrial process control technology field.

Background technology

Along with industrial expansion, production scale becomes increasingly complex, and is more and more higher to the requirement of control.In order to realize the high-quality product of High-efficient Production, a lot of processes all are configured to the higher-dimension multivariable process, and multivariable process is a modal class process in the commercial production.Because each road output of multivariable process has transmission and detects time lag, and there is crosslinked coupling between each output channel, the single argument control method that feasible great majority have developed is difficult to the multiple-input and multiple-output process, especially for the process that contains obvious time lag, coupling between system's output is very outstanding, can the output response performance of deterioration system seriously.Therefore, how implementing multiple time delay compensation and decoupling zero control is a present research and an application difficult problem.

For the coupling between the multi-variable system, first-selected method is to overcome by the suitable coupling between employing controlled variable and the manipulated variable, and wherein the most representative method is the relative gain matched pair technique.But for the comparatively serious system of association, even often adopt best variable pair relationhip can not reach satisfied decoupling zero effect, at this moment just must in system, add a decoupling zero network (or claiming corrective network) and carry out decoupling zero, make the controlling object that the strong coupling object becomes does not have coupling or weak coupling.In the current industry practice, coupling between export on each road of adopting the static decoupling device to alleviate multivariable control system usually, promptly at first a constant matrices decoupler is set at the multi-channel input place of controlled process, the inverse matrix of the steady-state gain transfer matrix that its transfer matrix form is a controlled process, the full-fledged single argument controlling Design method of controlled process transfer matrix utilization that augmentation is thus the obtained control system of constructing and adjust then.Its major defect is a coupling effect of not considering the control system dynamic response stage, makes that the Dynamic Coupling of each road system output is still serious, thereby causes control of quality not high.Aspect time lag compensation, the representative O.J.Smith of being proposes estimates structure.This structure biggest advantage is that the time lag link has been moved on to outside the closed loop, controlling performance is improved greatly, but, its maximum shortcoming is too to rely on precise math model, when estimation model and practical object had error, controlling performance can significantly worsen, even disperses, and very responsive for external disturbance, robustness is relatively poor.So conventional Smith Prediction Control structure is difficult to obtain in practice real application.For addressing this problem, a lot of scholars have proposed improving one's methods based on conventional Smith Prediction Control structure.

Find by prior art documents, the Smith Prediction Control scheme of representative industrial multivariable process at the tool multiple time delay is that internationally famous Wang Q.-G. professor is in document " DecouplingSmith predictor design for multivariable systems with multiple timedelays " (the decoupling zero Smith prediction device design of the multi-variable system of tool multiple time delay, be published in ChemicalEngineering Research and Design, Transactions of the Institute ofChemical Engineers, chemical engineering association proceedings: chemical engineering research and design, 2000,78,565-572.) in, propose to adopt earlier decoupler decoupling zero controlling object, design the method for a plurality of single argument Smith Prediction Control system then at the process after the decoupling zero, though obtained significantly improved control effect, but often to take the approximate processing of model reduction when setting up decoupler, this causes the controlled process can not be full decoupled and obtain a diagonal matrix, make between the expectation Decoupled Model of decoupling zero process and proposition and can not realize accurate coupling, thereby can't obtain perfect performance.And this control structure can not be distinguished and optimizes set-point response of each road system and load disturbance response thereof independently, yet has strong expectation to solving this difficult problem in the industrial practice at present.Recently, the S.P.Hung professor is in document " Decoupling Multivariable Controlwith Two Degrees of Freedom " (double freedom decoupling zero multivariable Control, be published in IndustrialEngineering Chemical Research chemical engineering research publication, 2006,45,3161-3173.) in a two-degree-of-freedom decoupling control system has been proposed, though realized the decoupling zero of each road system set-point response and load disturbance response thereof, but, what the controller matrix design adopted is the method for quantizing, needed data operation quantity is quite big, is not easy to actual applying and online adjusting.The Smith Prediction Control scheme at the industrial multivariable process of tool multiple time delay that it may be noted that other all satisfies under certain constraint condition at the hypothesis process model and proposes, thereby can't use in actual industrial production.

Summary of the invention

The objective of the invention is to overcome deficiency of the prior art, a kind of two freedom decoupling Smith Prediction Control system of industrial multivariable time-lag process is provided, to realize of the influence of effective bucking-out system time lag to system performance, realize the remarkable decoupling zero between each road output response of nominal system simultaneously, realize the separate adjusting of each road system set-point response and load disturbance response thereof, and realize online one-parameter adjustment control, to guarantee simple to operation and conveniently, can be widely used in various industrial multiple-input and multiple-output production process.

The present invention is achieved by the following technical solutions, and the present invention consists of the following components: n dimension setting point tracking controller, n dimension disturbance estimation device, controlled process recognition module and two multi-channel signal mixers.Wherein n is the dimension of controlled multivariable process.First multi-channel signal mixer is arranged on the n dimension input end of controlled process, it has one group of n dimension positive polarity input end, one group of n dimension negative polarity input end and one group of n dimension output terminal, its one group of positive polarity input end connects the n dimension output signal of setting point tracking controller, its one group of negative polarity input end connects the n dimension output signal of disturbance estimation device, and another group output terminal connects the n dimension input end of controlled process.Second multi-channel signal mixer is arranged on the n dimension output of controlled process, it has one group of n dimension positive polarity input end, one group of n dimension negative polarity input end and one group of n dimension output terminal, its one group of positive polarity input end connects the n dimension output measuring-signal of controlled process, its one group of negative polarity input end connects the n dimension output terminal of controlled process recognition module, and another group output terminal connects the n dimension input end of disturbance estimation device matrix.The output signal of setting point tracking controller is divided into two-way, and one the tunnel gives the positive polarity input end of first multi-channel signal mixer, and the input end of controlled process recognition module is sent on another road.

The function of setting point tracking controller is that system's set-point input signal is handled and computing, and the needed n dimension of controlled process work intake is provided, thereby makes the n dimension of controlled process export the requirement that reaches each road set-point.The function of multi-channel signal mixer is that two groups of n dimension input signals are mixed into one group of n dimension output signal according to the input channel order.The function of disturbance estimation device is that each road output bias signal of detected controlled process is handled and computing, thereby regulates the n dimension input quantity size of controlled process, the purpose that reaches elimination system output bias and suppress the load undesired signal.The executable structure of disturbance estimation device comprises: intergrade controlled process recognition module, revise disturbance wave filter and an intergrade multi-channel signal mixer.The intergrade multi-channel signal mixer is arranged on the n dimension input end of revising the disturbance wave filter, it has two groups of n dimension positive polarity input ends and one group of n dimension output terminal, its one group of positive polarity input end connects the n dimension input signal of disturbance estimation device, its another group positive polarity input end connects the n dimension output signal of intergrade controlled process recognition module, and another group output terminal connects the n dimension input end of revising the disturbance wave filter; The n dimension output of revising the disturbance wave filter is connected to the n dimension input end of process reference model.

During actual motion, at first the n dimension multichannel set-point input signal of control system is sent into the setting point tracking controller successively according to technological requirement respectively, carry out calculation process and amplification by it, the needed n dimension of controlled n dimension multivariable process work multichannel intake is provided, thereby makes the output of n dimension control system reach the requirement of n dimension set-point input signal respectively.When the load undesired signal was sneaked into controlled process, system's output changed, and produced the deviation between real system output and the controlled process recognition module output signal thus.This deviation signal is sent into the n dimension input end of disturbance estimation device, n dimension control output signal handled and produces by the disturbance estimation device to it, this output signal is sent to the n dimension input end of controlled process and regulates, thereby offset and system's output that balance is caused by the load undesired signal changes and fluctuates, reach the purpose of asymptotic elimination system output bias.

Described setting point tracking controller, be based on robust H2 optimal performance index design, can realize nominal system output response decoupling zero optimum, nominal system output response pass function satisfies the diagonal angle form, simultaneously, in every row controller of setting point tracking controller and the nominal system output response pass function corresponding diagonal element by same adjusting parameter lambda _CjjAdjust, can onlinely regulate monotonously, the time domain response index that feasible thus j corresponding system exports is by λ _CjjQuantitative setting monotonously, simultaneously, every row controller of disturbance estimation device is by same adjusting parameter lambda _FjjAdjust, can online dullness regulate, thereby the load disturbance response that feasible j corresponding system exports is by λ _FjjQuantitative setting monotonously.

Basic thought of the present invention is: open loop control mode is adopted in system's set-point response, by the controller matrix that on the forward direction input channel, adopts the reasonable canonical of the low order industrial multivariable time-lag process of calming, thereby avoided and the back is used to suppress produce coupling between the control closed loop that load disturbs, promptly realized full decoupled between system's set-point response and the load disturbance response, utilize the internal mode controller method for designing at " no time lag " partial design setting point tracking controller in the controlled process model simultaneously, thereby effectively eliminated of the influence of system's time lag, and can guarantee that the set-point response of control system reaches the decoupling zero optimum system performance.The interior ring of control that is used for process of inhibition load undesired signal is arranged between process load undesired signal input end and the output terminal, utilize the output measuring-signal of real process and the departure between the controlled process recognition module output signal feedback regulation quantity of information as process load undesired signal, send into and be arranged on the disturbance estimation device on the ring feedback channel in the control, after disturbance estimation device calculation process, the input regulating device of the estimated signal that obtains being given controlled process to be regulating, thereby realizes elimination system's output bias and the purpose that suppresses the load undesired signal.

The outstanding advantage of the two freedom decoupling Smith Prediction Control system that the present invention proposes is: 1. effectively bucking-out system time lag influence that system is exported, thereby the system performance of significantly improving; 2. the remarkable decoupling zero between the nominal system output response can be realized, and the set-point response and the load disturbance response of each the road output of optimal control system can be distinguished; 3. the every row sub-controller in setting point tracking controller matrix and the disturbance estimation device matrix is that one-parameter is adjusted and all by same parameter tuning, thereby can be implemented in line quantitatively regulating system setting point tracking performance and system's nominal performance and system's disturbance rejection performance monotonously; 4. control system can guarantee good robust stability, and it is insensitive to change for procedure parameter, can be in endoadaptation controlled process modeling error and procedure parameter perturbation in a big way.Therefore, the two freedom decoupling Smith Prediction Control system that the present invention provides has significant superiority and practicality, can show advanced control effect in practical application in industry.

Description of drawings

Fig. 1 is the frame principle figure of two freedom decoupling Smith Prediction Control of the present invention system.

Among Fig. 1, G (s) is meant that n ties up controlled multivariate multiple time delay process, and C (s) is meant n dimension setting point tracking controller, and F (s) is meant n dimension disturbance estimation device, G _m(s) be meant the controlled process recognition module, the circle node among the figure is meant multi-channel signal mixer, and r is meant that n maintains system set-point input signal, and y is meant that n maintains system output, and u is meant the n dimension output signal of C (s),

Be meant the n dimension output signal of F (s), d is meant the load undesired signal, and v is meant the n dimension output measuring-signal of actual controlled process and the deviation signal between the output of controlled process recognition module n dimension.

Fig. 2 is the executable structure synoptic diagram of the disturbance estimation device that proposes among the present invention.

Among Fig. 2, v is meant the n dimension output measuring-signal of actual controlled process and the deviation signal between the output of controlled process recognition module n dimension, Be meant the disturbance estimation signal, G _m(s) be meant intergrade controlled process recognition module, G _Mo(s) F _Fo(s) be to revise disturbance wave filter, F _Fo(s) be meant the disturbance wave filter.

Fig. 3 under nominal case is 0.1 the resulting output closed-loop response of reverse step load undesired signal synoptic diagram for two-way unit step set-point input signal and amplitude for the rectification column object.

Wherein, Fig. 3 (a) shows the response curve of first process output, and Fig. 3 (b) shows the response curve of second process output.Solid line is represented the present invention, and dotted line is represented the Hung method.

Fig. 4 is 0.1 the resulting output closed-loop response of reverse step load undesired signal synoptic diagram for the rectification column object having under the uncertain situation of the property taken advantage of for two-way unit step set-point input signal and amplitude.

Wherein, Fig. 4 (a) shows the response curve of first process output, and Fig. 4 (b) shows the response curve of second process output.Solid line represents that system under the uncertain situation of the property taken advantage of input, does not have the output response curve of adjustment control parameter, and dotted line is represented system under the uncertain situation of the property taken advantage of input, the output response curve after the adjustment control parameter.

Embodiment

Below in conjunction with accompanying drawing embodiments of the invention are elaborated: present embodiment is being to implement under the prerequisite with the technical solution of the present invention, provided detailed embodiment and concrete operating process, but protection scope of the present invention is not limited to following embodiment.

Present embodiment decoupling zero Smith Prediction Control system as shown in Figure 1 consists of the following components: controlled multivariate multiple time delay process G (s), and n ties up setting point tracking controller C (s), and n ties up disturbance estimation device F (s), controlled process recognition module G _m(s) and two multi-channel signal mixers (the circle node among the figure). wherein, first multi-channel signal mixer is arranged on the n dimension input end of controlled process G (s), it has one group of n dimension positive polarity input end, one group of n dimension negative polarity input end and one group of n dimension output terminal, its one group of positive polarity input end connects the n dimension output signal of setting point tracking controller C (s), its one group of negative polarity input end connects the n dimension output signal of disturbance estimation device F (s), and another group output terminal connects the n dimension input end of controlled process G (s).Second multi-channel signal mixer is arranged on the n dimension output of controlled process G (s), it has one group of n dimension positive polarity input end, one group of n dimension negative polarity input end and one group of n dimension output terminal, its one group of positive polarity input end connects the n dimension output measuring-signal of controlled process G (s), and its one group of negative polarity input end connects controlled process recognition module G _m(s) n dimension output terminal, another group output terminal connect the n dimension input end of disturbance estimation device matrix F (s).The output signal of setting point tracking controller C (s) is divided into two-way, and one the tunnel gives the positive polarity input end of first multi-channel signal mixer, and controlled process recognition module G is sent on another road _m(s) input end.The executable structure of disturbance estimation device comprises: intergrade controlled process recognition module G _m(s), revise disturbance wave filter G _Mo(s) F _Fo(s) and an intergrade multi-channel signal mixer.The intergrade multi-channel signal mixer is arranged on revises disturbance wave filter G _Mo(s) F _Fo(s) n dimension input end, it has two groups of n dimension positive polarity input ends and one group of n dimension output terminal, its one group of positive polarity input end connects the n dimension input signal of disturbance estimation device F (s), and its another group positive polarity input end connects intergrade controlled process recognition module G _m(s) n dimension output signal, another group output terminal connect revises disturbance wave filter G _Mo(s) F _Fo(s) n dimension input end; Revise disturbance wave filter G _Mo(s) F _Fo(s) n dimension output is connected to intergrade process reference model G _m(s) n dimension input end.

During the two freedom decoupling Smith Prediction Control system of actual motion present embodiment, at first control system set-point input signal r is sent into setting point tracking controller C (s), setting point tracking controller C (s) amplifies and level and smooth set-point input signal r, provide controlled multivariate multiple time delay process G (s) work needed intake u, thereby make the output of controlled multivariate multiple time delay process G (s) reach the requirement of set-point signal r.The output signal u of setting point tracking controller C (s) is divided into two-way, and one the tunnel gives actual controlled process G (s), and controlled process recognition module G is sent on another road _m(s) input end.Simultaneously, the detection signal of the output y of actual controlled process G (s) and controlled process recognition module G _m(s) output signal is sent into second multi-channel signal mixer and is asked difference operation, gained departure signal v sends into the disturbance estimation device F (s) of ring in the control, send into the input end of actual controlled process G (s) after the disturbance estimation device F (s) of ring handles and amplifies in controlling with degenerative form, thereby regulate the size of the input controlled quentity controlled variable u of actual controlled process G (s), reach and eliminate the load undesired signal d sneak into purpose controlled process P1 output.

Usually with following frequency domain mathematical expression form above-mentioned controlled industrial multivariable time-lag process is described in the reality

Wherein $g_{\mathrm{ij}}\left(s\right)=g_{\mathrm{oij}}\left(s\right)e^{{-\mathrm{\θ}}_{\mathrm{ij}}s},$ It is meant i transport function that is input to j output from controlled process, g _0ij(s) be the reasonable transport function part of its stable canonical, θ _IjBe its corresponding process transmission time lag, provide setting point tracking controller C (s) below, the design formula of disturbance estimation device F (s):

(1) at first the transfer function matrix identification model of the industrial multivariable process of tool multiple time delay is decomposed, decomposed form is as follows:

G(s)＝G _D(s)G _mo(s)＝G _A(s)G _N(s)G _mo(s) (2)

Wherein,

$G_A\left(s\right)=\mathrm{diag}\left\{e^{-{\mathrm{\θ}}_{\mathrm{jj}}s}\right\}\·\·\·\left(3\right)$

θ wherein _JjBe taken as G ^-1(s) maximum discreet value in the j row.

$G_N\left(s\right)=\mathrm{diag}\left\{\underset{r=1}{\overset{r_1}{\mathrm{\Π}}}{\left(\frac{-s+z_{\mathrm{rj}}}{s+z_{\mathrm{rj}}}\right)}^{k_{\mathrm{rj}}}\right\}\·\·\·\left(4\right)$

Z wherein _RjBe G ^-1(s) unsettled limit in the j row, k _IjBe G _O ^-1(s) unstable limit z in the j row _RjMaximum number.

G _D(s)＝G _A(s)G _N(s) (5)

$G_{\mathrm{mo}}\left(s\right)=G_D^{-1}\left(s\right)G\left(s\right)\·\·\·\left(6\right)$

(2) design disturbance wave filter is following form:

$F_{\mathrm{fo}}\left(s\right)=\mathrm{diag}\left\{\frac{1}{{({\mathrm{\λ}}_{\mathrm{fjj}}s+1)}^{n_j}}\right\}\·\·\·\left(7\right)$

Wherein, λ _FjjBe adjustable parameter, be used to regulate the disturbance rejection performance that system output in j road reaches actual requirement.

(3) by step (1) and (2), disturbance estimation device F (s) is designed to following form:

$F\left(s\right)={(I-G\left(s\right)G_{\mathrm{mo}}^{-1}\left(s\right)F_o\left(s\right))}^{-1}{G}_{\mathrm{mo}}^{-1}\left(s\right)F_o\left(s\right)$

$={(I-G_D\left(s\right)F_o\left(s\right))}^{-1}{G}_{\mathrm{mo}}^{-1}\left(s\right)F_o\left(s\right)\·\·\·\left(8\right)$

N wherein _jExpression G _Mo ^-1(s) maximal phase is to order.From expression formula (8) as can be seen, the disturbance estimation device of proposition can adopt the structure shown in the accompanying drawing 2 to realize easily.It may be noted that if G _Mo ^-1(s) F _o(s) can not physics realization, can use the rational approximation technology that it is similar to, the concrete analysis of face as follows.

(4) design setting point tracking filter is following form:

$F_{\mathrm{co}}\left(s\right)=\mathrm{diag}\left\{\frac{1}{{({\mathrm{\λ}}_{\mathrm{cjj}}s+1)}^{n_j}}\right\}\·\·\·\left(9\right)$

Wherein, λ _CjjBe adjustable parameter, be used to regulate the setting point tracking performance that system output in j road reaches actual requirement.

(5) by step (1) and (5), setting point tracking controller C (s) is designed to following form:

$C\left(s\right)=G_{\mathrm{mo}}^{-1}\left(s\right)F_{\mathrm{co}}\left(s\right)\·\·\·\left(10\right)$

In above-mentioned design process, there are following special circumstances, if i.e. G after the target transfer function matrix decomposition _Mo(s) still contain the time lag item in, this controller that will cause designing is unreasonable form, thereby can't physics realization, at this moment can use approximation technique that this controller is carried out rational approximation, and concrete steps are as follows: at first, and to G _Mo ^-1(s) decompose, decomposed form is as follows:

$G_{\mathrm{mo}}^{-1}\left(s\right)={\tilde{G}}_{\mathrm{mo}}^{-1}\left(s\right)G_r\left(s\right)\·\·\·\left(11\right)$

G wherein _r(s) be G _Mo ^-1(s) decomposable rational part in, it can be directly according to G _Mo ^-1(s) expression formula obtains.Then, right Carry out rational approximation, approximate form is as follows:

${\tilde{G}}_{\mathrm{mo}}^{-1}\left(s\right)=\frac{\underset{i=0}{\overset{U}{\mathrm{\Σ}}}{\mathrm{\α}}_i{s}^i}{\underset{j=0}{\overset{V}{\mathrm{\Σ}}}{\mathrm{\β}}_j{s}^j}\·\·\·\left(12\right)$

Wherein U and V are according to the design specifications predetermined parameter by the user.The value of U and V is big more, represents that the error of this rational approximation is more little, and system performance is good more.Usually U and V can be taken as 1 or 2.Parameter alpha _iAnd β _jCan determine easily according to following two equations,

$\left[\begin{array}{c}{\mathrm{\&alpha;}}_0\\ {\mathrm{\&alpha;}}_1\\ \&CenterDot;\\ \&CenterDot;\\ \&CenterDot;\\ {\mathrm{\&alpha;}}_U\end{array}\right]=\left[\begin{array}{cccc}{\mathrm{<figure-callout\; id="0"\; label="d"\; filenames="A20061011788100173.png"\; state="\{\{state\}\}">d</figure-callout>}}_0& 0& \&CenterDot;\&CenterDot;\&CenterDot;& 0\\ d_1& {\mathrm{<figure-callout\; id="0"\; label="d"\; filenames="A20061011788100173.png"\; state="\{\{state\}\}">d</figure-callout>}}_0& \&CenterDot;\&CenterDot;\&CenterDot;& 0\\ \&CenterDot;& \&CenterDot;& & \&CenterDot;\\ \&CenterDot;& \&CenterDot;& \&CenterDot;\&CenterDot;\&CenterDot;& \&CenterDot;\\ \&CenterDot;& \&CenterDot;& & \&CenterDot;\\ d_U& d_{U-1}& \&CenterDot;\&CenterDot;\&CenterDot;& d_{U-V}\end{array}\right]\left[\begin{array}{c}{\mathrm{\&beta;}}_0\\ {\mathrm{\&beta;}}_1\\ \&CenterDot;\\ \&CenterDot;\\ \&CenterDot;\\ {\mathrm{\&beta;}}_V\end{array}\right]\&CenterDot;\&CenterDot;\&CenterDot;\left(13\right)$

Wherein

$d_k\left(s\right)=\frac{1}{k!}\underset{s\&RightArrow;0}{\lim }\frac{d^k{\tilde{G}}_{\mathrm{mo}}^{-1}\left(s\right)}{{\mathrm{ds}}^k}\&CenterDot;\&CenterDot;\&CenterDot;\left(15\right)$

${\mathrm{\&beta;}}_o=\left\{\begin{array}{c}1,({\mathrm{\&beta;}}_j\&GreaterEqual;0)\\ -1,({\mathrm{\&beta;}}_j<0)\end{array}\right.\&CenterDot;\&CenterDot;\&CenterDot;\left(16\right)$

Need explanation, setting point tracking controller that provides and the Digital Discrete realization on industrial computer and single-chip microcomputer etc. easily of the design formula of disturbance estimation device, the sampling time generally can be taken at 0.01-0.1 between second.Every row controller of the setting point tracking controller matrix that design obtains in the present embodiment is by same adjusting parameter tuning, and every row controller of disturbance estimation device matrix is also by same adjusting parameter tuning.The adjustable parameter λ of on-line tuning C (s) and F (s) _CjjAnd λ _FjjRule be: λ can initially adjust _CjjAnd λ _FjjAt (5-10) θ _JjIn the scope.Turn the setting parameter λ of C (s) down _CjjCan accelerate corresponding system's output response speed, improve the nominal response performance of control system, but the output energy of corresponding required j row controller will increase, and the required energy that provides of the topworks of their correspondences also will increase, can tend to exceed their range of capacity, and when facing the not modeling dynamic perfromance of controlled process, be easy to show aggressive behavior, be unfavorable for the robust stability of control system; On the contrary, increase setting parameter λ _FjjCorresponding system's output response is slowed down, but the output energy of desired j row controller reduces, and corresponding topworks's energy needed also reduces, thereby help improving the robust stability of control system.So setting parameter λ of practical adjustments setting point tracking controller _CjjThe time, should between the output capacity of the nominal performance of system's each road output response and every row controller and topworks thereof, weigh.Similarly, the setting parameter λ of practical adjustments disturbance estimation device _FjjThe time, should between disturbing the output capacity of rejection and robust stability and every line control unit and topworks thereof, the load of control system weigh.

Investigate the chemical industry hydrocarbonylation thing fractionator process that a broad research adopts

$G\left(s\right)=\left[\begin{array}{cc}\frac{{12.8e}^{-s}}{16.7s+1}& \frac{{-18.9}^{-3s}}{21s+1}\\ \frac{{6.6e}^{-7s}}{10.9s+1}& \frac{{-19.4e}^{-3s}}{14.4s+1}\end{array}\right]$

Use decoupling zero Smith Prediction Control structure of the present invention, at first according to the structure of the structured flowchart shown in the accompanying drawing 1 control system; Carry out the design of controller then and adjust: the first step, application of formula (2)-(6) obtain factoring, and are specific as follows:

$G_D\left(s\right)=\left[\begin{array}{cc}{e}^{-\mathrm{<figure-callout\; id="0"\; label="s"\; filenames="A20061011788100173.png"\; state="\{\{state\}\}">s</figure-callout>}}& 0\\ 0& e^{-3s}\end{array}\right]$

$G_{\mathrm{mo}}\left(s\right)=\left[\begin{array}{cc}\frac{12.8}{\mathrm{16,7}s+1}& \frac{{-18.9e}^{-2s}}{21s+1}\\ \frac{{6.6e}^{-3s}}{10.9s+1}& \frac{-19.4}{14.4s+1}\end{array}\right]$

Second step, application of formula (7) obtain the disturbance estimation device:

$F\left(s\right)=\frac{F_o\left(s\right)}{1-G\left(s\right)F_o\left(s\right)}$

Wherein

$F_o\left(s\right)=D^{-1}\left(s\right)\left[\begin{array}{cc}\frac{(16.7s+1)}{12.8({\mathrm{\&lambda;}}_{f11}s+1)}& \frac{18.9(16.7s+1)(14.4s+1)}{(21s+1)({\mathrm{\&lambda;}}_{f22}s+1)}{e}^{-2s}\\ \frac{-6.6(16.7s+1)(14.4s+1)}{(10.9s+1)({\mathrm{\&lambda;}}_{f11}s+1)}{e}^{-4s}& \frac{(14.4s+1)}{-19.4({\mathrm{\&lambda;}}_{f22}s+1)}\end{array}\right]$

$D\left(s\right)=1-\frac{0.5023(14.4s+1)(16.7s+1)}{(21s+1)(10.9s+1)}{e}^{-6s}$

Application of formula (8) obtains the setting point tracking controller:

$C\left(s\right)=G^{-1}\left(s\right)T_s\left(s\right)$

$=\frac{\left[\begin{array}{cc}\frac{(16.7s+1)}{12.8({\mathrm{\&lambda;}}_{c1}s+1)}& \frac{18.9(16.7s+1)(14.4s+1)}{(21s+1)({\mathrm{\&lambda;}}_{c2}s+1)}{e}^{-2s}\\ \frac{-6.6(16.7s+1)(14.4s+1)}{(10.9s+1)({\mathrm{\&lambda;}}_{c1}s+1)}{e}^{-4s}& \frac{(14.4s+1)}{-19.4({\mathrm{\&lambda;}}_{c2}s+1)}\end{array}\right]}{D\left(s\right)}$

To implement this controller as can be seen from above controller form, must adopt very complicated control structure, calculate and the maintenance designing requirement so use the linear-apporximation technology to simplify, according to approximate formula (9)-(14), irrational part D (s) can be similar to as follows in the controller:

$D\left(s\right)=\frac{{74.9789\mathrm{<figure-callout\; id="2"\; label="s"\; filenames="A20061011788100173.png,A20061011788100181.png"\; state="\{\{state\}\}">s</figure-callout>}}^2+16.066s+0.4977}{36.6518{\mathrm{<figure-callout\; id="2"\; label="s"\; filenames="A20061011788100173.png,A20061011788100181.png"\; state="\{\{state\}\}">s</figure-callout>}}^2+25.419s+1}$

Above process is 2 to carry out depression of order according to order, and big more if order is selected, design accuracy is high more, but causes the controller complexity big more.

The 3rd the step, set up a closed-loop control system according to the closed loop controlling structure figure shown in the accompanying drawing 1, wherein the disturbance estimation device is set up according to Fig. 2.During this decoupling zero of actual motion Smith Prediction Control system, only n need be tieed up multichannel set-point input signal sends into the setting point tracking controller successively according to job requirement respectively, the setting point tracking controller can be to its calculation process and amplification, and offer the needed n dimension of controlled multivariable time-lag process G (s) work intake, thereby make the output of n dimension multi-way control system reach n dimension set-point tracer request respectively.When the load undesired signal is sneaked into controlled process G (s), system's output changes, deviation between the reference output signal consequent and that provided by the controlled process recognition module can be admitted to the n dimension input end of disturbance estimation device F (s), F (s) produces respective change and n is tieed up the disturbance estimation signal and is sent to the n dimension input end of controlled process G (s) to regulate, thereby progressively offset and system's output that balance is caused by the load undesired signal changes and fluctuates, reach the purpose of asymptotic elimination system output bias.

The 4th step, online adjusting setting point tracking controller and disturbance estimation device, the observing system closed loop response is determined the optimizing controller parameter thus.During emulation experiment, add unit step input: r for constantly the first via input quantity second at t=0 earlier ₁=1/s, and the second road input signal is r ₂=0, also all add unit rank input signal for constantly the second tunnel input in t=100 second then, all adding amplitude for during second two-way controlled process input end at t=200 again is 0.1 reverse step input.For obtain with the Hung method in the identical response speed of response curve that obtains, set λ here _C11=3.8, λ _C22=3.5, λ _F11=2.0 and λ _F22=2.0, resulting system closed-loop response curve is shown in Fig. 3.As can be seen from Figure 3, the two freedom decoupling Smith Prediction Control system (solid line) that provides of present embodiment has realized intimate full decoupled between the output response of nominal system.Can see simultaneously that the set-point response of system's output and the rejection of load undesired signal obviously are better than the two-freedom Prediction Control system (dotted line) of Hung method.

There is the uncertain Δ of the property taken advantage of input in the actual controlled process G of hypothesis (s) now _I=diag{ (s+0.3)/(s+1), (s+0.3)/(s+1) }.Here Δ _IPhysical interpretation is approx, and two input control valves of controlled process have uncertainty up to 100% at high band, and has 30% uncertainty nearly in the low-frequency range working range.Carry out emulation experiment as mentioned above under this serious process input uncertainty, the computer artificial result of the resulting process output of the setting method of the controller that present embodiment provides response as shown in Figure 4.Can see that by Fig. 4 the setting method of the controller that present embodiment provides (solid line) can guarantee the set-point response of system and the robust stability of load disturbance response well.Can see the controlled variable (λ in the dull increase setting point tracking controller in addition _C11=6.8, λ _C22=6.5), the vibration of system's set-point response reduces, and shown in the dotted line among Fig. 4, simultaneously, increases the adjusting parameter (λ in the disturbance estimation controller monotonously _F11=6.2 and λ _F22=6.9), the disturbance rejection response speed of process output is slack-off, and shown in the dotted line among Fig. 4, but the robust stability nargin of system becomes big at this moment.

It may be noted that can do all distortion to it under the prerequisite that does not exceed the related scope of technical solution of the present invention content is implemented.Provided the method for designing of setting point tracking controller and disturbance estimation device owing to the present invention is directed to the industrial multivariable process identification model of general tool multiple time delay, so be applicable to the various multiple-input and multiple-output production run that contains time lag.The double freedom decoupling zero Smith Prediction Control system that the present invention provides can be widely used in the production run of industries such as petrochemical industry, metallurgy, medicine, building materials and weaving.

Claims (3)

1, a kind of two freedom decoupling Smith Prediction Control system of industrial multivariable time-lag process, comprise: n dimension setting point tracking controller, n dimension disturbance estimation device, controlled process recognition module and two multi-channel signal mixers, it is characterized in that: wherein n is the dimension of controlled multivariable process, first multi-channel signal mixer is arranged on the n dimension input end of controlled process, it has one group of n dimension positive polarity input end, one group of n dimension negative polarity input end and one group of n dimension output terminal, its one group of positive polarity input end connects the n dimension output signal of setting point tracking controller, its one group of negative polarity input end connects the n dimension output signal of disturbance estimation device, and another group output terminal connects the n dimension input end of controlled process; Second multi-channel signal mixer is arranged on the n dimension output of controlled process, it has one group of n dimension positive polarity input end, one group of n dimension negative polarity input end and one group of n dimension output terminal, its one group of positive polarity input end connects the n dimension output measuring-signal of controlled process, its one group of negative polarity input end connects the n dimension output terminal of controlled process recognition module, another group output terminal connects the n dimension input end of disturbance estimation device matrix, the output signal of setting point tracking controller is divided into two-way, one the tunnel gives the positive polarity input end of first multi-channel signal mixer, and the input end of controlled process recognition module is sent on another road; The setting point tracking controller is handled and computing system's set-point input signal, the needed n dimension of controlled process work intake is provided, thereby the n dimension output that makes controlled process reaches the requirement of each road set-point, the disturbance estimation device is handled and computing each road output bias signal of detected controlled process, thereby regulate the n dimension input quantity size of controlled process, the purpose that reaches elimination system output bias and suppress the load undesired signal, multi-channel signal mixer is mixed into one group of n dimension output signal with two groups of n dimension input signals according to the input channel order.

2, the two freedom decoupling Smith Prediction Control system of industrial multivariable time-lag process as claimed in claim 1, it is characterized in that: described disturbance estimation device, its executable structure comprises: intergrade controlled process recognition module, revise disturbance wave filter and an intergrade multi-channel signal mixer, the intergrade multi-channel signal mixer is arranged on the n dimension input end of revising the disturbance wave filter, it has two groups of n dimension positive polarity input ends and one group of n dimension output terminal, its one group of positive polarity input end connects the n dimension input signal of disturbance estimation device, its another group positive polarity input end connects the n dimension output signal of intergrade controlled process recognition module, and another group output terminal connects the n dimension input end of revising the disturbance wave filter; The n dimension output of revising the disturbance wave filter is connected to the n dimension input end of process reference model.

3, the two freedom decoupling Smith Prediction Control system of industrial multivariable time-lag process as claimed in claim 1 is characterized in that: described setting point tracking controller, be based on robust H2 optimal performance index design,

Every row controller of setting point tracking controller comprises same adjusting parameter lambda _Cjj, be used to regulate the setting point tracking performance that system output in j road reaches actual requirement, by the online λ that regulates monotonously _CjjCan realize quantitatively regulating the time domain response index of j system's output; Every row controller of described disturbance estimation device comprises same adjusting parameter lambda _Fjj, be used to regulate the disturbance rejection performance that system output in j road reaches actual requirement, by the online λ that regulates monotonously _FjjCan realize quantitatively regulating the load disturbance response of j system's output, on-line tuning adjustable parameter λ _CjjAnd λ _FjjRule be: λ initially adjusts _CjjAnd λ _FjjAt (5-10) θ _JjIn the scope, turn the adjusting parameter lambda down _CjjCan accelerate corresponding system's output response speed, improve the nominal response performance of control system, but be unfavorable for the robust stability of control system; Increase setting parameter λ _FjjCorresponding system's output response is slowed down, but help improving the robust stability of control system.

Images