CN103418619B - Cold-rolled strip steel plate shape prediction control method - Google Patents

Abstract

The invention provides a cold-rolled strip steel plate shape prediction control method, comprising the following steps: for cold-rolled strip steel of the same specification, establishing a corresponding plate shape prediction control fuzzy reasoning model; setting fuzzy membership functions of various parameters in the plate shape prediction control fuzzy reasoning model by being combined with the characteristics of influence of rolling force change on the strip steel plate shape; establishing plate shape fuzzy prediction control models by utilizing a Takagi-Sugeno fuzzy model modeling rule; selecting a corresponding plate shape fuzzy prediction control model to carry out online adjustment on a working roller bending device. A dynamic relationship is established among rolling force variation, forward pull variation of a roll mill, backward pull variation of the roll mill, and online adjustment variable of the working roller bending device by using a fuzzy modeling method, the adverse effect of transmission time lag existing between the roll mill body and a plate shape instrument on the plate shape control at the outlet of the cold-rolled strip steel, and two familiar plate shape defects of intermediate waves and edge waves existing in the cold-rolled strip steel products are effectively overcome.

Description

A kind of cold-rolled strip steel shape forecast Control Algorithm

Technical field

The invention belongs to cold-strip steel field, particularly relate to a kind of cold-rolled strip steel shape forecast Control Algorithm.

Background technology

In recent years, along with the continuous expansion of the cold-rolled steel strip products scope of application, the requirement of user to belt plate shape quality is also more and more higher, and thus cold-rolled strip steel shape control technology has become an emphasis research topic especially paid attention to by modern steel enterprise and large-scale research institution.The principle of Strip Shape Control is the regulated quantity being calculated each Strip Shape Control actuator by layout board shape control algolithm, and each plate shape control measures are cooperatively interacted, farthest to eliminate the deviation between plate shape desired value and actual measured value.

Strip Shape Control algorithm is the core content of cold-rolled strip steel shape control system, and its superiority-inferiority directly determines the quality of Strip Shape Control effect.The plate shape Multi-variables optimum design control method based on feedback control idea extensively adopted at present is by configuring the real-time Shape signal of contact plate profile instrument on-line measurement in milling train exit, then with the quadratic sum of the deviation between plate shape desired value and real-time measurement values for evaluation index function, each plate shape regulation device on-line control amount when cycle calculations makes this evaluation index function obtain minimum of a value.It is to be noted, owing to there is transmission time lag between rolling mill body and plate profile instrument, the Shape signal of a certain moment plate profile instrument actual measurement is not the instant Shape signal of milling train outlet, that is, tradition board-shape control method utilizes Shape signal in the past to control current plate shape, and this also determines and is difficult to obtain high-precision Strip Shape Control effect.Particularly cause tension force acute variation before and after roll-force fluctuation and milling train when being with steel acceleration and deceleration, these factors all can cause outlet belt plate shape defect.Therefore, even if modern cold rolling enterprise production line is configured with state-of-the-art flatness detection device, also still fail effectively to eliminate the flatness defect that cold-rolled steel strip products exists, the unrestrained defect of wave and limit particularly.

Summary of the invention

The technical problem to be solved in the present invention is: provide a kind of cold-rolled strip steel shape forecast Control Algorithm, to solve the technical problem causing cold-rolled steel strip products Strip Shape Control of low quality owing to there is transmission time lag between rolling mill body and plate profile instrument.

The present invention for solving the problems of the technologies described above taked technical scheme is:

A kind of cold-rolled strip steel shape forecast Control Algorithm, is characterized in that: it comprises the following steps:

1) for the cold-strip steel of same specification, set up corresponding shape prediction and control Fuzzy Inference Model:

IFΔF is FP and IFΔT1is T1P and IFΔT2is T2P and，THEN U=U ₁；

IFΔF is FP and IFΔT1is T1P and IFΔT2is T2Z and，THEN U=U ₂；

IFΔF is FP and IFΔT1is T1P and IFΔT2is T2N and，THEN U=U ₃；

IFΔF is FP and IFΔT1is T1Z and IFΔT2is T2P and，THEN U=U ₄；

IFΔF is FP and IFΔT1is T1Z and IFΔT2is T2Z and，THEN U=U ₅；

IFΔF is FP and IFΔT1is T1Z and IFΔT2is T2N and，THEN U=U ₆；

IFΔF is FP and IFΔT1is T1N and IFΔT2is T2P and，THEN U=U ₇；

IFΔF is FP and IFΔT1is T1N and IFΔT2is T2Z and，THEN U=U ₈；

IFΔF is FP and IFΔT1is T1N and IFΔT2is T2N and，THEN U=U ₉；

IFΔF is FZ and IFΔT1is T1P and IFΔT2is T2P and，THEN U=U ₁₀；

IFΔF is FZ and IFΔT1is T1P and IFΔT2is T2Z and，THEN U=U ₁₁；

IFΔF is FZ and IFΔT1is T1P and IFΔT2is T2N and，THEN U=U ₁₂；

IFΔF is FZ and IFΔT1is T1Z and IFΔT2is T2P and，THEN U=U ₁₂；

IFΔF is FZ and IFΔT1is T1Z and IFΔT2is T2Z and，THEN U=U ₁₄；

IFΔF is FZ and IFΔT1is T1Z and IFΔT2is T2N and，THEN U=U ₁₅；

IFΔF is FZ and IFΔT1is T1N and IFΔT2is T2P and，THEN U=U ₁₆；

IFΔF is FZ and IFΔT1is T1N and IFΔT2is T2Z and，THEN U=U ₁₇；

IFΔF is FZ and IFΔT1is T1N and IFΔT2is T2N and，THEN U=U ₁₈；

IFΔF is FN and IFΔT1is T1P and IFΔT2is T2P and，THEN U=U ₁₉；

IFΔF is FN and IFΔT1is T1P and IFΔT2is T2Z and，THEN U=U ₂₀；

IFΔF is FN and IFΔT1is T1P and IFΔT2is T2N and，THEN U=U ₂₁；

IFΔF is FN and IFΔT1is T1Z and IFΔT2is T2P and，THEN U=U ₂₂；

IFΔF is FN and IFΔT1is T1Z and IFΔT2is T2Z and，THEN U=U ₂₃；

IFΔF is FN and IFΔT1is T1Z and IFΔT2is T2N and，THEN U=U ₂₄；

IFΔF is FN and IFΔT1is T1N and IFΔT2is T2P and，THEN U=U ₂₅；

IFΔF is FN and IFΔT1is T1N and IFΔT2is T2Z and，THEN U=U ₂₆；

IFΔF is FN and IFΔT1is T1N and IFΔT2is T2N and，THEN U=U ₂₇；

Wherein, Δ F is the difference of the roll-force measured value of current control period and the roll-force measured value of a upper control cycle, FP, FZ, FN be respectively describe single order plate shape deviation be just, zero, negative fuzzy number; Δ T1 is the difference of the milling train forward pull measured value of current control period and the milling train forward pull measured value of a upper control cycle, T1P, T1Z, T1N be respectively describe second order plate shape deviation be just, zero, negative fuzzy number; Δ T2 is the difference of the milling train backward pull measured value of current control period and the milling train backward pull measured value of a upper control cycle, T2P, T2Z, T2N be respectively describe quadravalence plate shape deviation be just, zero, negative fuzzy number; U is the on-line control amount at working-roller bending device; U _jfor in jth bar fuzzy rule bottom working roll roll-bending device on-line control amount reference value, obtain by manual operation Heuristics, j=1,2 ..., 27;

2) in conjunction with roll-force change to the following fuzzy membership functions of the influencing characterisitic of belt plate shape setting about Δ F:

Δ F is about the fuzzy membership functions of FP:

$f_{\mathrm{FP}}\left(\mathrm{\&Delta;F}\right)=\left\{\begin{array}{c}1,\mathrm{\&Delta;F}>\mathrm{\&alpha;}\\ \frac{1}{\mathrm{\&alpha;}}\mathrm{\&Delta;F},0\&le;\mathrm{\&Delta;F}\&le;\mathrm{\&alpha;}\\ 0,\mathrm{\&Delta;F}<0\end{array}\right.,$

Here, α is Δ F is positive threshold value, and controlling to think in fuzzy reasoning Controlling model that Δ F is just when being greater than α at shape prediction described in step 1), is negative when Δ F is less than-α, lower same;

Δ F is about the fuzzy membership functions of FZ:

$f_{\mathrm{FZ}}\left(\mathrm{\&Delta;F}\right)=\left\{\begin{array}{c}0,\mathrm{\&Delta;F}>\mathrm{\&alpha;}\\ 1-\frac{1}{\mathrm{\&alpha;}}\mathrm{\&Delta;F},0\&le;\mathrm{\&Delta;F}\&le;\mathrm{\&alpha;}\\ 1+\frac{1}{\mathrm{\&alpha;}}\mathrm{\&Delta;F},-\mathrm{\&alpha;}\&le;\mathrm{\&Delta;F}<0\\ 0,\mathrm{\&Delta;F}<-\mathrm{\&alpha;}\end{array}\right.,$

Δ F is about the fuzzy membership functions of FN:

$F_{\mathrm{FN}}\left(\mathrm{\&Delta;F}\right)=\left\{\begin{array}{c}0,\mathrm{\&Delta;F}>0\\ -\frac{1}{\mathrm{\&alpha;}}\mathrm{\&Delta;F},-\mathrm{\&alpha;}\&le;\mathrm{\&Delta;F}\&le;0\\ 1,\mathrm{\&Delta;F}<-\mathrm{\&alpha;}\end{array}\right.;$

3) in conjunction with the change of milling train forward pull to the following fuzzy membership functions of the influencing characterisitic of belt plate shape setting about Δ T1:

Δ T1 is about the fuzzy membership functions of T1P:

$f_{T1P}\left(\mathrm{\&Delta;T}1\right)=\left\{\begin{array}{c}1,\mathrm{\&Delta;T}1>\mathrm{\&beta;}\\ \frac{1}{\mathrm{\&beta;}}\mathrm{\&Delta;T}1,0\&le;\mathrm{\&Delta;T}1\&le;\mathrm{\&beta;}\\ 0,\mathrm{\&Delta;T}1<0\end{array}\right.,$

Here, β is Δ T1 is positive threshold value, and controlling to think in fuzzy reasoning Controlling model that Δ T1 is just when being greater than β at shape prediction described in step 1), is negative when Δ T1 is less than-β, lower same;

Δ T1 is about the fuzzy membership functions of T1Z:

$f_{T1Z}\left(\mathrm{\&Delta;T}1\right)=\left\{\begin{array}{c}0,\mathrm{\&Delta;T}1>\mathrm{\&beta;}\\ 1-\frac{1}{\mathrm{\&beta;}}\mathrm{\&Delta;T}\mathrm{1,0}\&le;\mathrm{\&Delta;T}1\&le;\mathrm{\&beta;}\\ 1+\frac{1}{\mathrm{\&beta;}}\mathrm{\&Delta;T}1,-\mathrm{\&beta;}\&le;\mathrm{\&Delta;T}1<0\\ 0,\mathrm{\&Delta;T}1<-\mathrm{\&beta;}\end{array}\right.,$

Δ T1 is about the fuzzy membership functions of T1N:

$f_{T1N}\left(\mathrm{\&Delta;T}1\right)=\left\{\begin{array}{c}0,\mathrm{\&Delta;T}1>0\\ -\frac{1}{\mathrm{\&beta;}}\mathrm{\&Delta;T}1,-\mathrm{\&beta;}\&le;\mathrm{\&Delta;T}1\&le;0\\ 1,\mathrm{\&Delta;T}1<-\mathrm{\&beta;}\end{array}\right.;$

4) in conjunction with the change of milling train backward pull to the following fuzzy membership functions of the influencing characterisitic of belt plate shape setting about Δ T2:

Δ T2 is about the fuzzy membership functions of T2P

$f_{T2P}\left(\mathrm{\&Delta;T}2\right)=\left\{\begin{array}{c}1,\mathrm{\&Delta;T}2>\mathrm{\&gamma;}\\ \frac{1}{\mathrm{\&gamma;}}\mathrm{\&Delta;T}\mathrm{2,0}\&le;\mathrm{\&Delta;T}2\&le;\mathrm{\&gamma;}\\ 0,\mathrm{\&Delta;T}2<0\end{array}\right.,$

Here, γ is Δ T2 is positive threshold value, and controlling to think in fuzzy reasoning Controlling model that Δ T2 is just when being greater than γ in the described shape prediction of step (1), is negative when Δ T2 is less than-γ, lower same;

Δ T2 is about the fuzzy membership functions of T2Z:

$f_{T2Z}\left(\mathrm{\&Delta;T}2\right)=\left\{\begin{array}{c}0,\mathrm{\&Delta;T}2>\mathrm{\&gamma;}\\ 1-\frac{1}{\mathrm{\&gamma;}}\mathrm{\&Delta;T}\mathrm{2,0}\&le;\mathrm{\&Delta;T}2\&le;\mathrm{\&gamma;}\\ 1+\frac{1}{\mathrm{\&gamma;}}\mathrm{\&Delta;T}2,-\mathrm{\&gamma;}\&le;\mathrm{\&Delta;T}2<0\\ 0,\mathrm{\&Delta;T}2<-\mathrm{\&gamma;}\end{array}\right.,$

Δ T2 is about the fuzzy membership functions of T2N:

$f_{T2N}\left(\mathrm{\&Delta;T}2\right)=\left\{\begin{array}{c}0,\mathrm{\&Delta;T}2>0\\ -\frac{1}{\mathrm{\&gamma;}}\mathrm{\&Delta;T}2,-\mathrm{\&gamma;}\&le;\mathrm{\&Delta;T}2\&le;0\\ 1,\mathrm{\&Delta;T}2<-\mathrm{\&gamma;}\end{array}\right.;$

5) Takagi-Sugeno fuzzy model modeling rule is utilized to set up as lower plate shape Fuzzy Predictive Control model:

$U=\underset{j=1}{\overset{27}{\mathrm{\&Sigma;}}}{h}_j\&times;U_j,$

Wherein h _ifor i-th fuzzy membership of plate shape Fuzzy Predictive Control model, and:

h ₁=f _FP(ΔF)×f _T1P(ΔT1)×f _T2P(ΔT2)，h ₂=f _FP(ΔF)×f _T1P(ΔT1)×f _T2Z(ΔT2)，

h ₃=f _FP(ΔF)×f _T1P(ΔT1)×f _T2N(ΔT2)，h ₄=f _FP(ΔF)×f _T1Z(ΔT1)×f _T2P(ΔT2)，

h ₅=f _FP(ΔF)×f _T1Z(ΔT1)×f _T2Z(ΔT2)，h ₆=f _FP(ΔF)×f _T1Z(ΔT1)×f _T2N(ΔT2)，

h ₇=f _FP(ΔF)×f _T1N(ΔT1)×f _T2P(ΔT2)，h ₈=f _FP(ΔF)×f _T1N(ΔT1)×f _T2Z(ΔT2)，

h ₉=f _FP(ΔF)×f _T1N(ΔT1)×f _T2N(ΔT2)，h ₁₀=f _FZ(ΔF)×f _T1P(ΔT1)×f _T2P(ΔT2)，

h ₁₁=f _FZ(ΔF)×f _T1P(ΔT1)×f _T2Z(ΔT2)，h ₁₂=f _FZ(ΔF)×f _T1P(ΔT1)×f _T2N(ΔT2)，

h ₁₃=f _FZ(ΔF)×f _T1Z(ΔT1)×f _T2P(ΔT2)，h ₁₄=f _FZ(ΔF)×f _T1Z(ΔT1)×f _T2Z(ΔT2)，

h ₁₅=f _FZ(ΔF)×f _T1Z(ΔT1)×f _T2N(ΔT2)，h ₁₆=f _FZ(ΔF)×f _T1N(ΔT1)×f _T2P(ΔT2)，

h ₁₇=f _FZ(ΔF)×f _T1N(ΔT1)×f _T2Z(ΔT2)，h ₁₈=f _FZ(ΔF)×f _T1N(ΔT1)×f _T2N(ΔT2)，

h ₁₉=f _FN(ΔF)×f _T1P(ΔT1)×f _T2P(ΔT2)，h ₂₀=f _FN(ΔF)×f _T1P(ΔT1)×f _T2Z(ΔT2)，

h ₂₁=f _FN(ΔF)×f _T1P(ΔT1)×f _T2N(ΔT2)，h ₂₂=f _FN(ΔF)×f _T1Z(ΔT1)×f _T2P(ΔT2)，

h ₂₃=f _FN(ΔF)×f _T1Z(ΔT1)×f _T2Z(ΔT2)，h ₂₄=f _FN(ΔF)×f _T1Z(ΔT1)×f _T2N(ΔT2)，

h ₂₅=f _FN(ΔF)×f _T1N(ΔT1)×f _T2P(ΔT2)，h ₂₆=f _FN(ΔF)×f _T1N(ΔT1)×f _T2Z(ΔT2)，

h ₂₇=f _FN(ΔF)×f _T1N(ΔT1)×f _T2N(ΔT2)；

6) for the cold-strip steel determining specification, corresponding plate shape Fuzzy Predictive Control model is selected to carry out the on-line control of working-roller bending device.

By such scheme, described step 6) comprises following steps:

6.1) judge that shape prediction controls current control period k and whether arrives, if do not arrive, continuous wait;

6.2), when current control period k arrives, the relevant actual measurement parameter in current control period is read: roll-force measured value F (k-1), milling train forward pull measured value T1 (k-1) of a upper control cycle, milling train backward pull measured value T2 (k-1) of a upper control cycle of control cycle before milling train backward pull measured value T2 (k), upper of roll-force measured value F (k) of current control period, milling train forward pull measured value T1 (k) of current control period, current control period;

6.3) difference of the roll-force measured value of current control period and the roll-force measured value of a upper control cycle is calculated respectively: Δ F=F (k)-F (k-1), the difference of the milling train forward pull measured value of current control period and the milling train forward pull measured value of a upper control cycle: Δ T1=T1 (k)-T1 (k-1), the difference of the milling train backward pull measured value of current control period and the milling train backward pull measured value of a upper control cycle: Δ T2=T2 (k)-T2 (k-1);

6.4) the Δ F, Δ T1 and the Δ T2 that obtain are substituted into the on-line control amount U that plate shape Fuzzy Predictive Control model obtains working-roller bending device;

6.5) the on-line control amount U of working-roller bending device is sent to the PLC controlled for working-roller bending device, realize the On-line Control of working-roller bending device;

6.6) shape prediction is controlled current control period k and is set as k+1, repeat step 6.1) to 6.5), realize the online loop control of working-roller bending device.

Beneficial effect of the present invention is: the present invention uses fuzzy Modeling Method to establish roll-force variable quantity, milling train forward pull variable quantity, dynamic relationship between milling train backward pull variable quantity and working-roller bending device on-line control amount, set up plate shape Fuzzy Predictive Control model, overcome adverse effect cold-strip steel outlet Strip Shape Control produced owing to there is transmission time lag between rolling mill body and plate profile instrument, decrease roll-force change and tension variation to the negative effect of belt plate shape quality, effectively can eliminate the unrestrained two kinds of common flatness defects of middle wave and limit that cold-rolled steel strip products exists, improve the quality of products.

Accompanying drawing explanation

Fig. 1 is the method flow diagram of one embodiment of the invention.

Fig. 2 is milling train exit plate shape distribution map when not using the inventive method.

Fig. 3 is the milling train exit plate shape distribution map after using the inventive method.

Detailed description of the invention

Below in conjunction with the drawings and specific embodiments, the invention will be further described, and to make those skilled in the art the present invention may be better understood and can be implemented, but illustrated embodiment is not as a limitation of the invention.

Fig. 1 is the method flow diagram of one embodiment of the invention, and it comprises the following steps:

1) for the cold-strip steel of same specification, set up corresponding shape prediction and control Fuzzy Inference Model:

IFΔF is FP and IFΔT1is T1P and IFΔT2is T2P and，THEN U=U ₁；

IFΔF is FP and IFΔT1is T1P and IFΔT2is T2Z and，THEN U=U ₂；

IFΔF is FP and IFΔT1is T1P and IFΔT2is T2N and，THEN U=U ₃；

IFΔF is FP and IFΔT1is T1Z and IFΔT2is T2P and，THEN U=U ₄；

IFΔF is FP and IFΔT1is T1Z and IFΔT2is T2Z and，THEN U=U ₅；

IFΔF is FP and IFΔT1is T1Z and IFΔT2is T2N and，THEN U=U ₆；

IFΔF is FP and IFΔT1is T1N and IFΔT2is T2P and，THEN U=U ₇；

IFΔF is FP and IFΔT1is T1N and IFΔT2is T2Z and，THEN U=U ₈；

IFΔF is FP and IFΔT1is T1N and IFΔT2is T2N and，THEN U=U ₉；

IFΔF is FZ and IFΔT1is T1P and IFΔT2is T2P and，THEN U=U ₁₀；

IFΔF is FZ and IFΔT1is T1P and IFΔT2is T2Z and，THEN U=U ₁₁；

IFΔF is FZ and IFΔT1is T1P and IFΔT2is T2N and，THEN U=U ₁₂；

IFΔF is FZ and IFΔT1is T1Z and IFΔT2is T2P and，THEN U=U ₁₂；

IFΔF is FZ and IFΔT1is T1Z and IFΔT2is T2Z and，THEN U=U ₁₄；

IFΔF is FZ and IFΔT1is T1Z and IFΔT2is T2N and，THEN U=U ₁₅；

IFΔF is FZ and IFΔT1is T1N and IFΔT2is T2P and，THEN U=U ₁₆；

IFΔF is FZ and IFΔT1is T1N and IFΔT2is T2Z and，THEN U=U ₁₇；

IFΔF is FZ and IFΔT1is T1N and IFΔT2is T2N and，THEN U=U ₁₈；

IFΔF is FN and IFΔT1is T1P and IFΔT2is T2P and，THEN U=U ₁₉；

IFΔF is FN and IFΔT1is T1P and IFΔT2is T2Z and，THEN U=U ₂₀；

IFΔF is FN and IFΔT1is T1P and IFΔT2is T2N and，THEN U=U ₂₁；

IFΔF is FN and IFΔT1is T1Z and IFΔT2is T2P and，THEN U=U ₂₂；

IFΔF is FN and IFΔT1is T1Z and IFΔT2is T2Z and，THEN U=U ₂₃；

IFΔF is FN and IFΔT1is T1Z and IFΔT2is T2N and，THEN U=U ₂₄；

IFΔF is FN and IFΔT1is T1N and IFΔT2is T2P and，THEN U=U ₂₅；

IFΔF is FN and IFΔT1is T1N and IFΔT2is T2Z and，THEN U=U ₂₆；

IFΔF is FN and IFΔT1is T1N and IFΔT2is T2N and，THEN U=U ₂₇；

Wherein, Δ F is the difference of the roll-force measured value of current control period and the roll-force measured value of a upper control cycle, unit is KN, FP, FZ, FN be respectively describe single order plate shape deviation be just, zero, negative fuzzy number; Δ T1 is the difference of the milling train forward pull measured value of current control period and the milling train forward pull measured value of a upper control cycle, unit is KN, T1P, T1Z, T1N be respectively describe second order plate shape deviation be just, zero, negative fuzzy number; Δ T2 is the difference of the milling train backward pull measured value of current control period and the milling train backward pull measured value of a upper control cycle, unit is KN, T2P, T2Z, T2N be respectively describe quadravalence plate shape deviation be just, zero, negative fuzzy number; U is the on-line control amount at working-roller bending device, and unit is KN; U _jfor in jth bar fuzzy rule bottom working roll roll-bending device on-line control amount reference value, obtain by manual operation Heuristics, unit is KN, j=1,2 ..., 27;

2) in conjunction with roll-force change to the following fuzzy membership functions of the influencing characterisitic of belt plate shape setting about Δ F:

Δ F is about the fuzzy membership functions of FP:

$f_{\mathrm{FP}}\left(\mathrm{\&Delta;F}\right)=\left\{\begin{array}{c}1,\mathrm{\&Delta;F}>\mathrm{\&alpha;}\\ \frac{1}{\mathrm{\&alpha;}}\mathrm{\&Delta;F},0\&le;\mathrm{\&Delta;F}\&le;\mathrm{\&alpha;}\\ 0,\mathrm{\&Delta;F}<0\end{array}\right.,$

Here, α is Δ F is positive threshold value, and unit is KN, and controlling to think in fuzzy reasoning Controlling model that Δ F is just when being greater than α at shape prediction described in step 1), is negative when Δ F is less than-α, lower same;

Δ F is about the fuzzy membership functions of FZ:

$f_{\mathrm{FZ}}\left(\mathrm{\&Delta;F}\right)=\left\{\begin{array}{c}0,\mathrm{\&Delta;F}>\mathrm{\&alpha;}\\ 1-\frac{1}{\mathrm{\&alpha;}}\mathrm{\&Delta;F},0\&le;\mathrm{\&Delta;F}\&le;\mathrm{\&alpha;}\\ 1+\frac{1}{\mathrm{\&alpha;}}\mathrm{\&Delta;F},-\mathrm{\&alpha;}\&le;\mathrm{\&Delta;F}<0\\ 0,\mathrm{\&Delta;F}<-\mathrm{\&alpha;}\end{array}\right.,$

Δ F is about the fuzzy membership functions of FN:

$F_{\mathrm{FN}}\left(\mathrm{\&Delta;F}\right)=\left\{\begin{array}{c}0,\mathrm{\&Delta;F}>0\\ -\frac{1}{\mathrm{\&alpha;}}\mathrm{\&Delta;F},-\mathrm{\&alpha;}\&le;\mathrm{\&Delta;F}\&le;0\\ 1,\mathrm{\&Delta;F}<-\mathrm{\&alpha;}\end{array}\right.;$

3) in conjunction with the change of milling train forward pull to the following fuzzy membership functions of the influencing characterisitic of belt plate shape setting about Δ T1:

Δ T1 is about the fuzzy membership functions of T1P:

$f_{T1P}\left(\mathrm{\&Delta;T}1\right)=\left\{\begin{array}{c}1,\mathrm{\&Delta;T}1>\mathrm{\&beta;}\\ \frac{1}{\mathrm{\&beta;}}\mathrm{\&Delta;T}1,0\&le;\mathrm{\&Delta;T}1\&le;\mathrm{\&beta;}\\ 0,\mathrm{\&Delta;T}1<0\end{array}\right.,$

Here, β is Δ T1 is positive threshold value, and unit is KN, and controlling to think in fuzzy reasoning Controlling model that Δ T1 is just when being greater than β at shape prediction described in step 1), is negative when Δ T1 is less than-β, lower same;

Δ T1 is about the fuzzy membership functions of T1Z:

$f_{T1Z}\left(\mathrm{\&Delta;T}1\right)=\left\{\begin{array}{c}0,\mathrm{\&Delta;T}1>\mathrm{\&beta;}\\ 1-\frac{1}{\mathrm{\&beta;}}\mathrm{\&Delta;T}\mathrm{1,0}\&le;\mathrm{\&Delta;T}1\&le;\mathrm{\&beta;}\\ 1+\frac{1}{\mathrm{\&beta;}}\mathrm{\&Delta;T}1,-\mathrm{\&beta;}\&le;\mathrm{\&Delta;T}1<0\\ 0,\mathrm{\&Delta;T}1<-\mathrm{\&beta;}\end{array}\right.,$

Δ T1 is about the fuzzy membership functions of T1N:

$f_{T1N}\left(\mathrm{\&Delta;T}1\right)=\left\{\begin{array}{c}0,\mathrm{\&Delta;T}1>0\\ -\frac{1}{\mathrm{\&beta;}}\mathrm{\&Delta;T}1,-\mathrm{\&beta;}\&le;\mathrm{\&Delta;T}1\&le;0\\ 1,\mathrm{\&Delta;T}1<-\mathrm{\&beta;}\end{array}\right.;$

4) in conjunction with the change of milling train backward pull to the following fuzzy membership functions of the influencing characterisitic of belt plate shape setting about Δ T2:

Δ T2 is about the fuzzy membership functions of T2P

$f_{T2P}\left(\mathrm{\&Delta;T}2\right)=\left\{\begin{array}{c}1,\mathrm{\&Delta;T}2>\mathrm{\&gamma;}\\ \frac{1}{\mathrm{\&gamma;}}\mathrm{\&Delta;T}\mathrm{2,0}\&le;\mathrm{\&Delta;T}2\&le;\mathrm{\&gamma;}\\ 0,\mathrm{\&Delta;T}2<0\end{array}\right.,$

Here, γ is Δ T2 is positive threshold value, and unit is KN, and controlling to think in fuzzy reasoning Controlling model that Δ T2 is just when being greater than γ in the described shape prediction of step (1), is negative when Δ T2 is less than-γ, lower same;

Δ T2 is about the fuzzy membership functions of T2Z:

$f_{T2Z}\left(\mathrm{\&Delta;T}2\right)=\left\{\begin{array}{c}0,\mathrm{\&Delta;T}2>\mathrm{\&gamma;}\\ 1-\frac{1}{\mathrm{\&gamma;}}\mathrm{\&Delta;T}\mathrm{2,0}\&le;\mathrm{\&Delta;T}2\&le;\mathrm{\&gamma;}\\ 1+\frac{1}{\mathrm{\&gamma;}}\mathrm{\&Delta;T}2,-\mathrm{\&gamma;}\&le;\mathrm{\&Delta;T}2<0\\ 0,\mathrm{\&Delta;T}2<-\mathrm{\&gamma;}\end{array}\right.,$

Δ T2 is about the fuzzy membership functions of T2N:

$f_{T2P}\left(\mathrm{\&Delta;T}2\right)=\left\{\begin{array}{c}0,\mathrm{\&Delta;T}2>0\\ -\frac{1}{\mathrm{\&gamma;}}\mathrm{\&Delta;T}2,-\mathrm{\&gamma;}\&le;\mathrm{\&Delta;T}2\&le;0\\ 1,\mathrm{\&Delta;T}2<-\mathrm{\&gamma;}\end{array}\right.;$

5) Takagi-Sugeno fuzzy model modeling rule is utilized to set up as lower plate shape Fuzzy Predictive Control model:

$U=\underset{j=1}{\overset{27}{\mathrm{\&Sigma;}}}{h}_j\&times;U_j,$

Wherein h _ifor i-th fuzzy membership of plate shape Fuzzy Predictive Control model, and:

h ₁=f _FP(ΔF)×f _T1P(ΔT1)×f _T2P(ΔT2)，h ₂=f _FP(ΔF)×f _T1P(ΔT1)×f _T2Z(ΔT2)，

h ₃=f _FP(ΔF)×f _T1P(ΔT1)×f _T2N(ΔT2)，h ₄=f _FP(ΔF)×f _T1Z(ΔT1)×f _T2P(ΔT2)，

h ₅=f _FP(ΔF)×f _T1Z(ΔT1)×f _T2Z(ΔT2)，h ₆=f _FP(ΔF)×f _T1Z(ΔT1)×f _T2N(ΔT2)，

h ₇=f _FP(ΔF)×f _T1N(ΔT1)×f _T2P(ΔT2)，h ₈=f _FP(ΔF)×f _T1N(ΔT1)×f _T2Z(ΔT2)，

h ₉=f _FP(ΔF)×f _T1N(ΔT1)×f _T2N(ΔT2)，h ₁₀=f _FZ(ΔF)×f _T1P(ΔT1)×f _T2P(ΔT2)，

h ₁₁=f _FZ(ΔF)×f _T1P(ΔT1)×f _T2Z(ΔT2)，h ₁₂=f _FZ(ΔF)×f _T1P(ΔT1)×f _T2N(ΔT2)，

h ₁₃=f _FZ(ΔF)×f _T1Z(ΔT1)×f _T2P(ΔT2)，h ₁₄=f _FZ(ΔF)×f _T1Z(ΔT1)×f _T2Z(ΔT2)，

h ₁₅=f _FZ(ΔF)×f _T1Z(ΔT1)×f _T2N(ΔT2)，h ₁₆=f _FZ(ΔF)×f _T1N(ΔT1)×f _T2P(ΔT2)，

h ₁₇=f _FZ(ΔF)×f _T1N(ΔT1)×f _T2Z(ΔT2)，h ₁₈=f _FZ(ΔF)×f _T1N(ΔT1)×f _T2N(ΔT2)，

h ₁₉=f _FN(ΔF)×f _T1P(ΔT1)×f _T2P(ΔT2)，h ₂₀=f _FN(ΔF)×f _T1P(ΔT1)×f _T2Z(ΔT2)，

h ₂₁=f _FN(ΔF)×f _T1P(ΔT1)×f _T2N(ΔT2)，h ₂₂=f _FN(ΔF)×f _T1Z(ΔT1)×f _T2P(ΔT2)，

h ₂₃=f _FN(ΔF)×f _T1Z(ΔT1)×f _T2Z(ΔT2)，h ₂₄=f _FN(ΔF)×f _T1Z(ΔT1)×f _T2N(ΔT2)，

h ₂₅=f _FN(ΔF)×f _T1N(ΔT1)×f _T2P(ΔT2)，h ₂₆=f _FN(ΔF)×f _T1N(ΔT1)×f _T2Z(ΔT2)，

h ₂₇=f _FN(ΔF)×f _T1N(ΔT1)×f _T2N(ΔT2)；

6) for the cold-strip steel determining specification, select corresponding plate shape Fuzzy Predictive Control model to carry out the on-line control of working-roller bending device, in the present embodiment, it specifically comprises following steps:

6.1) judge that shape prediction controls current control period k and whether arrives, if do not arrive, continuous wait;

6.2), when current control period k arrives, the relevant actual measurement parameter in current control period is read: roll-force measured value F (k-1), milling train forward pull measured value T1 (k-1) of a upper control cycle, milling train backward pull measured value T2 (k-1) of a upper control cycle of control cycle before milling train backward pull measured value T2 (k), upper of roll-force measured value F (k) of current control period, milling train forward pull measured value T1 (k) of current control period, current control period;

6.3) difference of the roll-force measured value of current control period and the roll-force measured value of a upper control cycle is calculated respectively: Δ F=F (k)-F (k-1), the difference of the milling train forward pull measured value of current control period and the milling train forward pull measured value of a upper control cycle: Δ T1=T1 (k)-T1 (k-1), the difference of the milling train backward pull measured value of current control period and the milling train backward pull measured value of a upper control cycle: Δ T2=T2 (k)-T2 (k-1);

6.4) the Δ F, Δ T1 and the Δ T2 that obtain are substituted into the on-line control amount U that plate shape Fuzzy Predictive Control model obtains working-roller bending device;

6.5) the on-line control amount U of working-roller bending device is sent to the PLC controlled for working-roller bending device, realize the On-line Control of working-roller bending device;

6.6) shape prediction is controlled current control period k and is set as k+1, repeat step 6.1) to 6.5), realize the online loop control of working-roller bending device.

Four rollers, six roller single chassis or multi-frame tandem mills is can be used for based on cold-rolled strip steel shape thickness of slab integrated control method of the present invention.Below for a single chassis six-high cluster mill, six-high cluster mill the product of rolling can comprise common plate, high-strength steel, part stainless steel and silicon steel etc.The present embodiment rolling be middle high grade silicon steel, type is UCM milling train, and Strip Shape Control means comprise roller declination, the positive and negative roller of working roll, the positive roller of intermediate calender rolls, intermediate roll shifting and emulsion section cooling etc.Wherein intermediate roll shifting carries out presetting according to strip width, and Adjustment principle is alignd with steel edge portion at intermediate calender rolls body of roll edge, and also can consider interpolation correction by operation side, after being transferred to position, holding position is constant; Emulsion section cooling has larger characteristic time lag.Thus the Strip Shape Control means of on-line control mainly contain roller declination, the positive and negative roller of working roll, the positive roller of intermediate calender rolls three kinds.The basic mechanical design feature index of this unit and device parameter are:

Mill speed: Max900m/min, draught pressure: Max18000KN, maximum rolling force square: 140.3KN × m, coiling tension: Max220KN, main motor current: 5500KW;

Supplied materials thickness range: 1.8 ~ 2.5mm, supplied materials width range: 850 ~ 1280mm, outgoing gauge scope: 0.3mm ~ 1.0mm;

Work roll diameter: 290 ~ 340mm, working roll height: 1400mm, intermediate calender rolls diameter: 440 ~ 500mm, intermediate calender rolls height: 1640mm, backing roll diameter: 1150 ~ 1250mm, backing roll height: 1400mm;

Every side work roll bending power :-280 ~ 350KN, every side intermediate calender rolls bending roller force: 0 ~ 500KN, intermediate calender rolls is traversing amount :-120 ~ 120mm axially, auxiliary hydraulic system pressure: 14MPa, balance bending system pressure: 28MPa, press down system pressure: 28MPa.

Plate profile instrument adopts ABB AB's plate shape roller of Sweden, the roller footpath 313mm of this plate shape roller, be made up of single solid steel axle, a measured zone is divided in the width direction every 52mm or 26mm, surrounding vertically at measuring roller in each measured zone is uniform-distribution with four grooves to place magnetoelasticity force snesor, the outside of sensor wrap up by steel loop.Plate shape roller often rotates a circle, and can measure four times to Strip Shape.

Measuring cell is installed be used for measuring mill rolling force in real time, tensiometer is installed at inlet of rolling mill place and is used for measuring milling train forward pull, tensiometer is installed in milling train exit and is used for measuring milling train backward pull.

For the present embodiment, Fig. 2 give sheet shape prediction and control method of the present invention put into operation before milling train exit plate shape distribution map, now adopt regulative mode manually to carry out cold-rolled strip steel shape control.As seen from Figure 2, the plate shape deviation maximum of milling train exit plate shape has even exceeded 20I, has obviously limit wave flatness defect, have impact on product quality and economic benefit; This also illustrates the necessity of actual production to advanced Strip Shape Control technical research.Milling train exit plate shape distribution map after the control method of the present invention that gives Fig. 3 puts into operation.As seen from Figure 3, the method of the invention effectively eliminates plate shape deviation, and the plate shape deviation Maximum constraint of milling train exit plate shape, within 5I, does not have obvious limit wave or middle unrestrained flatness defect, significantly improve belt steel product exit plate shape, improve the strip shape quality of band.

Above embodiment, only for illustration of technological thought of the present invention and feature, its object is to enable those skilled in the art understand content of the present invention and implement according to this.The scope of the claims of the present invention is not limited to above-described embodiment, and all equivalent variations of doing according to disclosed principle, design philosophy or modification, all within the scope of the claims of the present invention.

Claims (2)

1. a cold-rolled strip steel shape forecast Control Algorithm, is characterized in that: it comprises the following steps:

1) for the cold-strip steel of same specification, set up corresponding shape prediction and control Fuzzy Inference Model:

IF ΔF is FP and IF ΔT1 is T1P and IF ΔT2 is T2P and,THEN U＝U ₁；

IF ΔF is FP and IF ΔT1 is T1P and IF ΔT2 is T2Z and,THEN U＝U ₂；

IF ΔF is FP and IF ΔT1 is T1P and IF ΔT2 is T2N and,THEN U＝U ₃；

IF ΔF is FP and IF ΔT1 is T1Z and IF ΔT2 is T2P and,THEN U＝U ₄；

IF ΔF is FP and IF ΔT1 is T1Z and IF ΔT2 is T2Z and,THEN U＝U ₅；

IF ΔF is FP and IF ΔT1 is T1Z and IF ΔT2 is T2N and,THEN U＝U ₆；

IF ΔF is FP and IF ΔT1 is T1N and IF ΔT2 is T2P and,THEN U＝U ₇；

IF ΔF is FP and IF ΔT1 is T1N and IF ΔT2 is T2Z and,THEN U＝U ₈；

IF ΔF is FP and IF ΔT1 is T1N and IF ΔT2 is T2N and,THEN U＝U ₉；

IF ΔF is FZ and IF ΔT1 is T1P and IF ΔT2 is T2P and,THEN U＝U ₁₀；

IF ΔF is FZ and IF ΔT1 is T1P and IF ΔT2 is T2Z and,THEN U＝U ₁₁；

IF ΔF is FZ and IF ΔT1 is T1P and IF ΔT2 is T2N and,THEN U＝U ₁₂；

IF ΔF is FZ and IF ΔT1 is T1Z and IF ΔT2 is T2P and,THEN U＝U ₁₂；

IF ΔF is FZ and IF ΔT1 is T1Z and IF ΔT2 is T2Z and,THEN U＝U ₁₄；

IF ΔF is FZ and IF ΔT1 is T1Z and IF ΔT2 is T2N and,THEN U＝U ₁₅；

IF ΔF is FZ and IF ΔT1 is T1N and IF ΔT2 is T2P and,THEN U＝U ₁₆；

IF ΔF is FZ and IF ΔT1 is T1N and IF ΔT2 is T2Z and,THEN U＝U ₁₇；

IF ΔF is FZ and IF ΔT1 is T1N and IF ΔT2 is T2N and,THEN U＝U ₁₈；

IF ΔF is FN and IF ΔT1 is T1P and IF ΔT2 is T2P and,THEN U＝U ₁₉；

IF ΔF is FN and IF ΔT1 is T1P and IF ΔT2 is T2Z and,THEN U＝U ₂₀；

IF ΔF is FN and IF ΔT1 is T1P and IF ΔT2 is T2N and,THEN U＝U ₂₁；

IF ΔF is FN and IF ΔT1 is T1Z and IF ΔT2 is T2P and,THEN U＝U ₂₂；

IF ΔF is FN and IF ΔT1 is T1Z and IF ΔT2 is T2Z and,THEN U＝U ₂₃；

IF ΔF is FN and IF ΔT1 is T1Z and IF ΔT2 is T2N and,THEN U＝U ₂₄；

IF ΔF is FN and IF ΔT1 is T1N and IF ΔT2 is T2P and,THEN U＝U ₂₅；

IF ΔF is FN and IF ΔT1 is T1N and IF ΔT2 is T2Z and,THEN U＝U ₂₆；

IF ΔF is FN and IF ΔT1 is T1N and IF ΔT2 is T2N and,THEN U＝U ₂₇；

Wherein, Δ F is the difference of the roll-force measured value of current control period and the roll-force measured value of a upper control cycle, FP, FZ, FN be respectively describe single order plate shape deviation be just, zero, negative fuzzy number; Δ T1 is the difference of the milling train forward pull measured value of current control period and the milling train forward pull measured value of a upper control cycle, T1P, T1Z, T1N be respectively describe second order plate shape deviation be just, zero, negative fuzzy number; Δ T2 is the difference of the milling train backward pull measured value of current control period and the milling train backward pull measured value of a upper control cycle, T2P, T2Z, T2N be respectively describe quadravalence plate shape deviation be just, zero, negative fuzzy number; U is the on-line control amount at working-roller bending device; U _jfor in jth bar fuzzy rule bottom working roll roll-bending device on-line control amount reference value, obtain by manual operation Heuristics, j=1,2 ..., 27;

2) in conjunction with roll-force change to the following fuzzy membership functions of the influencing characterisitic of belt plate shape setting about Δ F:

Δ F is about the fuzzy membership functions of FP:

$f_{\mathrm{FP}}\left(\mathrm{\&Delta;F}\right)=\left\{\begin{array}{cc}1,& \mathrm{\&Delta;F}>\mathrm{\&alpha;}\\ \frac{1}{\mathrm{\&alpha;}}\mathrm{\&Delta;F},& 0\&le;\mathrm{\&Delta;F}\&le;\mathrm{\&alpha;}\\ 0,& \mathrm{\&Delta;F}<0\end{array}\right.,$

Here, α is Δ F is positive threshold value, in step 1) described shape prediction controls to think in fuzzy reasoning Controlling model that Δ F is just when being greater than α, is negative when Δ F is less than-α, lower same;

Δ F is about the fuzzy membership functions of FZ:

$f_{\mathrm{FZ}}\left(\mathrm{\&Delta;F}\right)=\left\{\begin{array}{cc}0,& \mathrm{\&Delta;F}>\mathrm{\&alpha;}\\ 1-\frac{1}{\mathrm{\&alpha;}}\mathrm{\&Delta;F},& 0\&le;\mathrm{\&Delta;F}\&le;\mathrm{\&alpha;}\\ 1+\frac{1}{\mathrm{\&alpha;}}\mathrm{\&Delta;F},& -\mathrm{\&alpha;}\&le;\mathrm{\&Delta;F}<0\\ 0,& \mathrm{\&Delta;F}<-\mathrm{\&alpha;}\end{array}\right.,$

Δ F is about the fuzzy membership functions of FN:

$F_{\mathrm{FN}}\left(\mathrm{\&Delta;F}\right)=\left\{\begin{array}{cc}0,& \mathrm{\&Delta;F}>0\\ -\frac{1}{\mathrm{\&alpha;}}\mathrm{\&Delta;F},& -\mathrm{\&alpha;}\&le;\mathrm{\&Delta;F}\&le;0\\ 1,& \mathrm{\&Delta;F}<-\mathrm{\&alpha;}\end{array}\right.;$

3) in conjunction with the change of milling train forward pull to the following fuzzy membership functions of the influencing characterisitic of belt plate shape setting about Δ T1:

Δ T1 is about the fuzzy membership functions of T1P:

$f_{T1P}\left(\mathrm{\&Delta;T}1\right)=\left\{\begin{array}{cc}1,& \mathrm{\&Delta;T}1>\mathrm{\&beta;}\\ \frac{1}{\mathrm{\&beta;}}\mathrm{\&Delta;T}1& 0\&le;\mathrm{\&Delta;T}1\&le;\mathrm{\&beta;}\\ 0,& \mathrm{\&Delta;T}1<0\end{array}\right.,$

Here, β is Δ T1 is positive threshold value, in step 1) described shape prediction controls to think in fuzzy reasoning Controlling model that Δ T1 is just when being greater than β, is negative when Δ T1 is less than-β, lower same;

Δ T1 is about the fuzzy membership functions of T1Z:

$f_{T1Z}\left(\mathrm{\&Delta;T}1\right)=\left\{\begin{array}{cc}0,& \mathrm{\&Delta;T}1>\mathrm{\&beta;}\\ 1-\frac{1}{\mathrm{\&beta;}}\mathrm{\&Delta;T}1,& 0\&le;\mathrm{\&Delta;T}1\&le;\mathrm{\&beta;}\\ 1+\frac{1}{\mathrm{\&beta;}}\mathrm{\&Delta;T}1,& -\mathrm{\&beta;}\&le;\mathrm{\&Delta;T}1<0\\ 0,& \mathrm{\&Delta;T}1<-\mathrm{\&beta;}\end{array}\right.,$

Δ T1 is about the fuzzy membership functions of T1N:

$f_{T1N}\left(\mathrm{\&Delta;T}1\right)=\left\{\begin{array}{cc}0,& \mathrm{\&Delta;T}1>0\\ -\frac{1}{\mathrm{\&beta;}}\mathrm{\&Delta;T}1,& -\mathrm{\&beta;}\&le;\mathrm{\&Delta;T}1\&le;0\\ 1,& \mathrm{\&Delta;T}1<-\mathrm{\&beta;}\end{array}\right.;$

4) in conjunction with the change of milling train backward pull to the following fuzzy membership functions of the influencing characterisitic of belt plate shape setting about Δ T2:

Δ T2 is about the fuzzy membership functions of T2P

$f_{T2P}\left(\mathrm{\&Delta;T}2\right)=\left\{\begin{array}{cc}1,& \mathrm{\&Delta;T}2>\mathrm{\&gamma;}\\ \frac{1}{\mathrm{\&gamma;}}\mathrm{\&Delta;T}2,& 0\&le;\mathrm{\&Delta;T}2\&le;\mathrm{\&gamma;}\\ 0,& \mathrm{\&Delta;T}2<0\end{array}\right.,$

Here, γ is Δ T2 is positive threshold value, and controlling to think in fuzzy reasoning Controlling model that Δ T2 is just when being greater than γ in the described shape prediction of step (1), is negative when Δ T2 is less than-γ, lower same;

Δ T2 is about the fuzzy membership functions of T2Z:

$f_{T2Z}\left(\mathrm{\&Delta;T}2\right)=\left\{\begin{array}{cc}0,& \mathrm{\&Delta;T}2>\mathrm{\&gamma;}\\ 1-\frac{1}{\mathrm{\&gamma;}}\mathrm{\&Delta;T}2,& 0\&le;\mathrm{\&Delta;T}2\&le;\mathrm{\&gamma;}\\ 1+\frac{1}{\mathrm{\&gamma;}}\mathrm{\&Delta;T}2,& -\mathrm{\&gamma;}\&le;\mathrm{\&Delta;T}2<0\\ 0,& \mathrm{\&Delta;T}2<-\mathrm{\&gamma;}\end{array}\right.,$

Δ T2 is about the fuzzy membership functions of T2N:

$f_{T2N}\left(\mathrm{\&Delta;T}2\right)=\left\{\begin{array}{cc}0,& \mathrm{\&Delta;T}2>0\\ -\frac{1}{\mathrm{\&gamma;}}\mathrm{\&Delta;T}2& -\mathrm{\&gamma;}\&le;\mathrm{\&Delta;T}2\&le;0\\ 1,& \mathrm{\&Delta;T}2<-\mathrm{\&gamma;}\end{array}\right.;$

5) Takagi-Sugeno fuzzy model modeling rule is utilized to set up as lower plate shape Fuzzy Predictive Control model:

$U=\underset{j=1}{\overset{27}{\mathrm{\&Sigma;}}}{h}_j\&times;U_j,$

Wherein h _jfor a jth fuzzy membership of plate shape Fuzzy Predictive Control model, and:

h ₁＝f _FP(ΔF)×f _T1P(ΔT1)×f _T2P(ΔT2),h ₂＝f _FP(ΔF)×f _T1P(ΔT1)×f _T2Z(ΔT2),

h ₃＝f _FP(ΔF)×f _T1P(ΔT1)×f _T2N(ΔT2),h ₄＝f _FP(ΔF)×f _T1Z(ΔT1)×f _T2P(ΔT2),

h ₅＝f _FP(ΔF)×f _T1Z(ΔT1)×f _T2Z(ΔT2),h ₆＝f _FP(ΔF)×f _T1Z(ΔT1)×f _T2N(ΔT2),

h ₇＝f _FP(ΔF)×f _T1N(ΔT1)×f _T2P(ΔT2),h ₈＝f _FP(ΔF)×f _T1N(ΔT1)×f _T2Z(ΔT2),

h ₉＝f _FP(ΔF)×f _T1N(ΔT1)×f _T2N(ΔT2),h ₁₀＝f _FZ(ΔF)×f _T1P(ΔT1)×f _T2P(ΔT2),

h ₁₁＝f _FZ(ΔF)×f _T1P(ΔT1)×f _T2Z(ΔT2),h ₁₂＝f _FZ(ΔF)×f _T1P(ΔT1)×f _T2N(ΔT2),

h ₁₃＝f _FZ(ΔF)×f _T1Z(ΔT1)×f _T2P(ΔT2),h ₁₄＝f _FZ(ΔF)×f _T1Z(ΔT1)×f _T2Z(ΔT2),

h ₁₅＝f _FZ(ΔF)×f _T1Z(ΔT1)×f _T2N(ΔT2),h ₁₆＝f _FZ(ΔF)×f _T1N(ΔT1)×f _T2P(ΔT2),

h ₁₇＝f _FZ(ΔF)×f _T1N(ΔT1)×f _T2Z(ΔT2),h ₁₈＝f _FZ(ΔF)×f _T1N(ΔT1)×f _T2N(ΔT2),

h ₁₉＝f _FN(ΔF)×f _T1P(ΔT1)×f _T2P(ΔT2),h ₂₀＝f _FN(ΔF)×f _T1P(ΔT1)×f _T2Z(ΔT2),

h ₂₁＝f _FN(ΔF)×f _T1P(ΔT1)×f _T2N(ΔT2),h ₂₂＝f _FN(ΔF)×f _T1Z(ΔT1)×f _T2P(ΔT2),

h ₂₃＝f _FN(ΔF)×f _T1Z(ΔT1)×f _T2Z(ΔT2),h ₂₄＝f _FN(ΔF)×f _T1Z(ΔT1)×f _T2N(ΔT2),

h ₂₅＝f _FN(ΔF)×f _T1N(ΔT1)×f _T2P(ΔT2),h ₂₆＝f _FN(ΔF)×f _T1N(ΔT1)×f _T2Z(ΔT2),

h ₂₇＝f _FN(ΔF)×f _T1N(ΔT1)×f _T2N(ΔT2)；

6) for the cold-strip steel determining specification, corresponding plate shape Fuzzy Predictive Control model is selected to carry out the on-line control of working-roller bending device.

2. cold-rolled strip steel shape forecast Control Algorithm according to claim 1, is characterized in that: described step 6) comprise following steps:

6.1) judge that shape prediction controls current control period k and whether arrives, if do not arrive, continuous wait;

6.2), when current control period k arrives, the relevant actual measurement parameter in current control period is read: roll-force measured value F (k-1), milling train forward pull measured value T1 (k-1) of a upper control cycle, milling train backward pull measured value T2 (k-1) of a upper control cycle of control cycle before milling train backward pull measured value T2 (k), upper of roll-force measured value F (k) of current control period, milling train forward pull measured value T1 (k) of current control period, current control period;

6.3) difference of the roll-force measured value of current control period and the roll-force measured value of a upper control cycle is calculated respectively: Δ F=F (k)-F (k-1), the difference of the milling train forward pull measured value of current control period and the milling train forward pull measured value of a upper control cycle: Δ T1=T1 (k)-T1 (k-1), the difference of the milling train backward pull measured value of current control period and the milling train backward pull measured value of a upper control cycle: Δ T2=T2 (k)-T2 (k-1);

6.4) the Δ F, Δ T1 and the Δ T2 that obtain are substituted into the on-line control amount U that plate shape Fuzzy Predictive Control model obtains working-roller bending device;

6.5) the on-line control amount U of working-roller bending device is sent to the PLC controlled for working-roller bending device, realize the On-line Control of working-roller bending device;

6.6) shape prediction is controlled current control period k and is set as k+1, repeat step 6.1) to 6.5), realize the online loop control of working-roller bending device.