CN103433295B - Single-frame double-coiling aluminium hot-rolling mill convex degree control method - Google Patents

Abstract

nullThe present invention relates to single-frame double-coiling aluminium hot-rolling mill convex degree control method，Feature according to milling train and the Crown control strategy of on-the-spot rolling experience formulation single-frame double-coiling hot rolling，Set up the convexity preset model being applicable to single chassis、Flatulence of heat type computation model、The feedforward of convexity and feedback control system，Realize Aluminum Plate and Strip and well control with glacing flatness in the full-automatic convexity of this kind of mill milling，Alloy feature according to band formulates corresponding Crown control strategy，Roughing is given by convexity preset model and flatulence of heat type model、The bending roller force setting value of two rolling sequences of finish rolling and chilling spray pattern ensure head convex value，The feedforward compensation of bending roller force is carried out to reduce the roll-force fluctuation impact on convexity at finishing pass，The passage third from the bottom or the extreme trace time that carry out convex measuring carry out bending roller force closed loop control to ensure total length convexity and plate shape according to alloy feature situation，Carry out bending roller force and the self study of chilling spray model after the end of rolling，Improve preset model precision.

Description

Single-frame double-coiling aluminium hot-rolling mill convex degree control method

Technical field

The present invention relates to the convex degree control method that a kind of aluminothermy is rolled in production, particularly to a kind of single chassis Convexity and board-shape control method in the range of the full belt length of double-coiling aluminium hot-rolling mill.

Background technology

Single-frame double-coiling aluminium hot-rolling mill becomes numerous aluminum processing enterprise and produces the first-selected machine of hot-rolled product Type.Unlike iron and steel domain, owing to Aluminum sheets production scale is less, its technological level, equipment Automatization level is the most on the low side, and the hot-rolling mill that current most enterprises uses generally exists product quality not The phenomenon that stable, control device falls behind.Convexity is one critically important quality index of hot-rolled product, with Improving constantly of customer requirement, it is convex that increasing Aluminum sheets manufacturing enterprise starts to pay attention to hot-rolled product The control of degree.

Convexity and glacing flatness reflect band shape of cross section and longitudinal extension size respectively, and the two exists Incidence relation.In Aluminum sheets production process, glacing flatness often becomes top-priority quality index, This is because current most aluminothermy rolls convexity and Strip Shape Control is entirely the experience by feat of operative employee Realizing, glacing flatness size visually adjusts operation of rolling parameter, until band is straight.Lean on Artificial experience method is carried out panel shape and is had the disadvantages that it is first can not be in the situation controlling glacing flatness Under take into account control convexity；Next to that lack rational Strip Shape Control strategy, profile regulation device can not be sent out Shoot maximum effect, often occur profile regulation device reached ultimate value can't panel shape good Good phenomenon；Three products strip shape quality can fluctuate, no along with manipulator's experienced degree difference It is beneficial to obtain stable product quality.

For the deficiency of artificial experience method panel shape, external relevant enterprise is on Aluminum sheets hot continuous rolling Develop set of Crown control system, but its control method and strategy are not suitable for single chassis hot rolling Machine, essentially consists in: be first that continuous hot-rolling mill has multi-frame roller regulating power, real according to outlet side Survey convexity and adjust the bending roller force distribution of each frame, and the regulation of the bending roller force of single chassis hot-rolling mill can only be Completing in institute's mill train time, the setup algorithm of bending roller force is just had higher requirement by this；Next to that it is hot Tandem mill can only measure the band convexity at last rack outlet, and comes for single chassis hot-rolling mill Say, measure the band outlet convexity of two passages, after actual measurement convexity information therefore can be utilized to revise The target convex value of continuous passage, and then improve setting accuracy；3rd, frame each with continuous hot-rolling mill is born Blaming a pressure passage different, single chassis hot-rolling mill to be responsible for all of rolling pass, so can cause The flatulence of heat type change of working roll is more frequent, and real-time prediction work roll thermal crown is for convexity setting accuracy Raising is necessary.

Summary of the invention

It is an object of the invention to the deficiency overcoming prior art to exist, it is provided that a kind of single-frame double-coiling aluminum Hot-rolling mill convex degree control method, is set calculating and feedback control to material to be rolled volume convexity, it is ensured that On the premise of outlet strip flatness is good, it is achieved the concordance of convexity in band length range.

The purpose of the present invention is achieved through the following technical solutions:

Single-frame double-coiling aluminium hot-rolling mill convex degree control method, comprises the following steps:

(1) overall Crown control policy development；

(2) the export goal convexity of each rolling pass band is determined；

(3) D difference point-score is used to come the temperature field of evaluation work roller each point；

(4) setting value of each actuator is calculated；

(5) bending roller force feedforward gain calculates；

(6) crown feedback control strategy is carried out with bending roller force；

(7) convexity setting accuracy is improved by Model Self-Learning.

Further, above-mentioned single-frame double-coiling aluminium hot-rolling mill convex degree control method, described the most convex Degree control strategy is formulated, and divides aluminium strip by aluminium alloy kind, hardness, prescription and finished product thickness The alloy feature of material, and formulate corresponding Crown control strategy: alloy species is different, its rolling strategy And finish to gauge requires also different, alloy soft or hard degree is different, and its Crown control strategy is the most otherwise varied；Eventually Rolling thickness different, rolled plate or its Crown control strategy of coiled material are the most different；

1a) formulation of Crown control strategy need to consider the product quality emphasis of variety classes alloy, produces Amount, temperature, surface quality and plate shape are 4 important quality index of board rolling, when with yield Or temperature for paying the utmost attention to index time, for convexity control require reduce；When being excellent with surface quality When first considering target, convexity closed loop control needs passage in advance to carry out, to reduce extreme trace time convexity regulation And affect surface quality of strips；When with plate shape for when paying the utmost attention to index, then need to optimize corresponding bearing Lotus is distributed to ensure exit plate shape；When being common priority target with several quality index sometimes, then need Formulate corresponding Crown control strategy；

1b) at rough rolling step, each passage band outlet flakiness ratio is smaller, and band horizontal mobility is relatively big, Employing experience bending roller force rolls；From the beginning of C4 passage, i.e. inverse the 4th passage of the operation of rolling, according to Alloy soft or hard degree implements different Crown control strategies with finished product thickness；For thin material, it rolled The passage that journey camber instrument comes into operation has 2, therefore implements the passage of crown feedback control be up to Two, depend on the soft or hard degree of alloy；Mild alloy has higher convexity regulating power than hard alloy, Therefore mild alloy is implemented 2 crown feedbacks and is controlled passage strategy, and hard alloy only carries out convex in C3 passage Degree feedback control；For medium hardness alloy, in the operation of rolling, use hard alloy according to practical situation Or the Crown control strategy of mild alloy；For thickness material, the passage come into operation due to profile gauge Only 1, therefore only carry out crown feedback control in last passage.

Further, above-mentioned single-frame double-coiling aluminium hot-rolling mill convex degree control method, described determine The export goal convexity of each rolling pass band:

2a) determine each passage export goal convexity, Cn represents the n-th passage reciprocal:

$C_{\mathrm{aim}1}=\left\{\begin{array}{c}{C}_{\mathrm{rn}}\\ C_{\mathrm{real}3}\·\frac{h_1}{h_3}\end{array}\right.,$ When C3 passage outlet convexity can be measured

$C_{\mathrm{aim}2}=\left\{\begin{array}{c}{C}_{\mathrm{real}3}\·\frac{h_2}{h_3}\\ C_{\mathrm{rn}}\·\frac{h_2}{h_1}\end{array}\right.,$ When C3 passage outlet convexity can be measured

$C_{\mathrm{aimi}}=C_{\mathrm{rn}}\·\frac{h_i}{h_1},$ As i=3 or i=4

C_aimiRepresent the export goal convexity of i-th passage reciprocal, μm, C_rnRepresent enterprise requirements Finished product target convexity, μm, C_realiRepresent Ci passage actual measurement outlet convexity, μm, h_iRepresent The exit thickness of the i-th passage reciprocal, mm；

2b) when finishing pass is more than 4 passage, from finish rolling the first passage to the mesh of the 5th passage reciprocal Mark convexity is the most uncertain.

Further, above-mentioned single-frame double-coiling aluminium hot-rolling mill convex degree control method, described work The real-time roll thermal crown of roller calculates:

3a) use the temperature field of D difference point-score evaluation work roller each point；

3b) according to the temperature field of working roll, obtain the roll thermal crown value of each moment working roll axial unit；

3c) the comprehensive convexity of working roll is made up of working roll initial cut convexity and work roll thermal crown, front Person is just it has been determined that keep constant in the operation of rolling before roll change, according to the work calculated in real time The roller roll thermal crown real-time evaluation work comprehensive convexity of roller.

Further, above-mentioned single-frame double-coiling aluminium hot-rolling mill convex degree control method, described calculating The setting value of each actuator:

4a) passages several before the finish rolling stage are had

C_out=E_p·P+E_B·BF+E_WR·C_WR

E_P=f_P(B), E_B=f_B(B)

C_outRepresent band outlet convexity, μm, E_pRepresent that roll-force affects coefficient, μm/kN, P table Show roll-force, kN,

E_BRepresenting that bending roller force affects coefficient, μm/kN, BF represent bending roller force, kN, E_WRRepresent work The comprehensive convexity of roller affects coefficient, C_WRRepresenting the comprehensive convexity of working roll, μm, B represents rolled piece width, Mm, f_P(B)、f_B(B)、E_WRRepresent that roll-force and bending roller force affect coefficient calculations formula, by emulation data Recurrence obtains；

4b) passages several for finish rolling stage Mo have

C_out=E_p·P+E_C·C_in+E_B·BF+E_WR·C_WR

E_P=f_P(B,D_WR), E_C=f_C(B,H,h,D_WR), E_B=f_B(B,D_WR)

C_outRepresent band outlet convexity, μm, E_pRepresent that roll-force affects coefficient, μm/kN, P table Show roll-force, kN, E_CRepresent that supplied materials convexity affects coefficient, C_inRepresent supplied materials convexity, μm, E_BTable Showing that bending roller force affects coefficient, μm/kN, BF represent bending roller force, kN, E_WRRepresent that working roll is the most convex Degree affects coefficient, C_WRRepresenting the comprehensive convexity of working roll, μm, B represents rolled piece width, mm, D_WRTable Show that work roll diameter, mm, H, h represent rolled piece entrance and exit thickness, mm, E_P、E_C、E_BTable Reach formula and E_WRObtained by emulation data regression；

4c) single chassis aluminium hot-rolling mill entrance side and outlet side are provided with cooling down jet beam, entrance side cold But liquid is mainly used in basis cooling, and nozzle is distributed as being uniformly distributed, according to the different selections of alloy rigidity Different cooling grades；Coolant cooling grade classification be by force, in, weak three kinds, then different hardness Alloy use cooling grade be: hard alloy correspondence cools down by force grade, in medium hardness correspondence cool down The corresponding weak cooling grade of grade, mild alloy；

4d) nozzle of outlet side is distributed as quadratic distribution, low to adapt to high both sides in the middle of working roll flatulence of heat type Feature；It is fixed according to carrying out flitch width that nozzle opens number；In order to improve edge regulating power, edge Cooling water inflow is set as 0, for ease of regulation injection grade, during the injection grade of intermediate noxzzle is set as Deng cooling grade, it follows that the injection grade of each nozzle:

$L_N=\mathrm{round}\left\{L_C\right[1-{\left(\frac{X_N}{X_E}\right)}^2]+E_{\mathrm{Lrn}}\}$

L_NRepresent from the nozzle injection grade that edge is N, L_CRepresent the injection grade of intermediate noxzzle, X_N Represent the nozzle abscissa of numbered N, X_ERepresent the abscissa of intermediate noxzzle, E_LrnRepresent cooling spray Drench self study coefficient；

4e) for a batch of material of same specification, from the beginning of the 2nd block, the injection grade of each of which nozzle For:

$L_N^{\mathrm{SET}}=L_{N,\mathrm{Tail}}^{\mathrm{ACT}}$

Represent from the nozzle injection grade that edge is N,Represent that previous rolling off edge is N Nozzle afterbody actual ejection grade.

Yet further, above-mentioned single-frame double-coiling aluminium hot-rolling mill convex degree control method, described roller Power feedforward gain calculates:

5a) bending roller force quickly regulates the convexity of roll gap, determines curved according to the fluctuation size of actual measurement roll-force The regulated quantity of roller power:

△ BF=K_BF·△P

△ BF represents the bending roller force size that needs regulate, kN, K_BFRepresent bending roller force gain coefficient, △ P Represent actual measurement roll-force undulate quantity；

5b) when AGC comes into operation, bending roller force gain is tried to achieve by following formula:

$\mathrm{\ΔBF}=K_F\·\frac{\mathrm{CR}}{h}\·\mathrm{\ΔS}+K_F(\frac{1}{C_P}\·\frac{\mathrm{CR}}{h}-\frac{1}{K_P})\mathrm{\ΔP}$

K_FThe expression bending roller force lateral stiffness on the flexural deformation impact of roller system, kN/mm,Represent outlet Ratio convexity, △ S represents that roll gap is poor, mm, C_PRepresent milling train longitudinal rigidity, kN/mm, K_PExpression is rolled The power processed lateral stiffness on the flexural deformation impact of roller system, kN/mm.

Yet further, above-mentioned single-frame double-coiling aluminium hot-rolling mill convex degree control method, described convexity Feedback control strategy is formulated:

6a) formulate crown feedback control strategy according to the deviation of profile gauge measured value with convexity desired value； Bending roller force and chilling spray are two kinds of means of Crown control, and bending roller force regulation is swift in motion, and is suitable for Line regulation convexity；Owing to single-frame double-coiling aluminium hot-rolling mill is equipped with bikini profile gauge, by presetting Mode control convexity；

6b) outlet convexity and the target convexity deviation measured when profile gauge is in the convexity fluctuation range of permission Time interior, it is not necessary to plate shape is adjusted；When deviation value more than allow convexity fluctuation range time and little When bending roller force regulation and control high-low limit, then use bending roller force regulation convexity deviation；Otherwise use and ensure to put down Straight degree is good and sacrifices the strategy adjustment bending roller force of convexity；

6c) crown feedback of single chassis hot-rolling mill controls only to carry out in current pass, can only regulate and work as The bending roller force of front passage, adjusts chilling spray flow and adjusts convexity, to improve the bending roller force of next block Regulation and control effect；

6d) using single chassis hot-rolling mill to roll, be up to 2 passages carry out crown feedback control；

6e) crown feedback control in affect coefficient value by convexity setting model according to alloy, thickness, Width interval gives.

Yet further, above-mentioned single-frame double-coiling aluminium hot-rolling mill convex degree control method, described in pass through Model Self-Learning raising convexity setting accuracy:

7a) before carrying out self study, carry out self adaptation according to measured value, through certain number of times from Just starting self study after adaptation, bending roller force model coefficient self adaptation formula is:

$K_{\mathrm{BF}\_\mathrm{new}}=K_{\mathrm{BF}\_\mathrm{old}}+(\frac{{\mathrm{BF}}_m}{{\mathrm{BF}}_c}-K_{\mathrm{BF}\_\mathrm{old}})\×\mathrm{VAD}$

K_{BF_new}For new bending roller force model coefficient, K_{BF_old}For old bending roller force model coefficient, BF_mFor Bending roller force measured value, BF_cFor by measured value calculated roller force value, VAD is total gain amplifier；

7b) carrying out self study after 15～20 self adaptations, bending roller force self learning model is

L_{BF_new}=L_{BF_old}+(K_{BF_new}-L_{BF_old})×VP

L_{BF_new}For new bending roller force self study coefficient, L_{BF_old}For old not yet through adaptive roller Power self study coefficient, K_{BF_new}For through adaptive model coefficient, VP is self study amplification coefficient；

7c) chilling spray pattern self study computational methods are similar with bending roller force self study coefficient, specifically count Formula is:

$E_{\mathrm{LrnN}}^{\mathrm{Cur}}=L_N^{\mathrm{ACT}}-L_N^{\mathrm{SET}}$

$E_{\mathrm{LrnN}}^{\mathrm{New}}=E_{\mathrm{LrnN}}^{\mathrm{Old}}+(E_{\mathrm{LrnN}}^{\mathrm{Cur}}-E_{\mathrm{LrnN}}^{\mathrm{Old}})\·\mathrm{\β}$

Represent the self study coefficient that n-th nozzle is current；Represent the actual spray of n-th nozzle Drench flow,Represent n-th nozzle initially sets flow,Represent the new of n-th nozzle Self study coefficient,Representing the old self study coefficient of n-th nozzle, β represents self study gain system Number.

Substantive distinguishing features and significantly progress that technical solution of the present invention is prominent are mainly reflected in:

1. can to carry out rolling pass convexity the most pre-for single-frame double-coiling aluminium hot-rolling mill Crown control system Set and feedback control, compensate for the deficiency of manually experience panel shape；Also reduce behaviour simultaneously The dependence of workmanship's experience, it is possible to while improving product strip shape quality, it is ensured that stablizing of strip shape quality Property；

2. use head presetting and in the middle part of band, the strategy of crown feedback control is capable of in outlet flat The concordance of full belt length outlet convexity is ensured on the premise of straight degree is good；

3. derive in the setup algorithm formula obtaining good glacing flatness and convexity according to Shape theory, fully Consider the chilling spray impact for Gap crown, be distributed by setting the nozzle of coolant, spray The size of flow and cooling water inflow controls roll crown, thus alleviates the regulation burden of bending roller force, The each control measures occurred when avoiding manually experience rolling all reach the unfavorable condition of production of the limit, The service life of equipment is extended while improving product quality.

Accompanying drawing explanation

Below in conjunction with the accompanying drawings technical solution of the present invention is described further:

Fig. 1 be overall Crown control policy map by different qualified kinds and specification institution (explanation: C1～ C4 passage represents that rolling passage last is to fourth from the last passage respectively)；

Fig. 2 is the two-dimensional grid division figure of evaluation work roller flatulence of heat type；

Fig. 3 is crown feedback control block diagram.

Detailed description of the invention

Material to be rolled volume convexity is set calculating and feedback control by the present invention, it is ensured that put down at outlet band On the premise of straight degree is good, it is achieved the concordance of convexity in band length range, thus overcome artificial warp The shortcoming testing panel shape.Main process steps is as follows:

(1) overall Crown control policy development.

Overall Crown control strategy is on the premise of existing Strip Shape Control means, solves how to play this The maximum effect of a little control devices thus avoid bending roller force and chilling spray to use irrational problem, from And reach the purpose of the present invention.

Owing to the Crown control of the present invention is to ensure that outlet glacing flatness is carried out, therefore on the premise of good The crown feedback passage that comes into operation is limited by factors such as product thickness, product materials.

1a) alloy soft or hard degree is different, and its Crown control strategy also can be otherwise varied.In the present invention, 5xxx system alloy is defined as hard alloy, 1xxx system, 8xxx system and part 3xxx system alloy It is defined as mild alloy, remaining 3xxx system alloy is defined as medium hardness alloy.

1b) finish to gauge thickness is different, and its Crown control strategy also can be different.In the present invention, in code It is defined as thin material containing 3 material volumes batching passage, and code batches passage containing 1 or 2 Material volume be defined as thick material.

1c) the overall Crown control strategy of single-frame double-coiling aluminium hot-rolling mill is as shown in Figure 1.In roughing In the stage, each passage band outlet flakiness ratio is smaller, and band horizontal mobility is relatively big, will not produce plate shape Problem, therefore can use experience bending roller force (generally using balance bending roller force) to roll.With heat even Unlike rolling, single chassis aluminium hot-rolling mill several passages outlet band before finish rolling still has less Flakiness ratio, use the method for experience bending roller force to carry out rolling and do not have plate shape problem.

1d) start from C4 passage (inverse the 4th passage of the operation of rolling, lower same), soft according to alloy Hard degree implements different Crown control strategies with finished product thickness.For thin material, its operation of rolling The passage that camber instrument comes into operation has 2, and the passage therefore implementing crown feedback control is up to two Individual, this depends on the soft or hard degree of alloy.In general, mild alloy has higher convex than hard alloy Degree adjustability, therefore mild alloy is implemented 2 crown feedbacks and is controlled passage strategy, and hard alloy is only Crown feedback control is carried out in C3 passage.For medium hardness alloy, can basis in the operation of rolling Practical situation uses hard alloy or the Crown control strategy of mild alloy.For thickness material, due to convex The passage that degree instrument comes into operation only has 1, therefore only carries out crown feedback control in last passage.

(2) the export goal convexity of each rolling pass band is determined.

Its final goal of the work that the present invention is done is on the premise of ensureing that glacing flatness is good, makes finished product Passage band outlet convexity hit value in length range.For single chassis aluminium hot-rolling mill, The overall process being worked into finished product aluminum volume from supplied materials aluminium ingot all completes, therefore whole rolling on this milling train Process is divided into roughing and two stages of finish rolling, and roughing shears stub bar material tail passage for dividing with finish rolling with heavy pruning Boundary line.Owing to rough rolling step is not related to Crown control problem, the therefore convex degree control method of the present invention It is served only for the passage in finish rolling stage.

Convexity and glacing flatness are the two indices of plate shape, the most relevant, and Crown control needs Ensureing that glacing flatness is carried out on the premise of good.In general, in order to ensure to export glacing flatness well, Need to meet the condition that band entrance cross-section shape is shaped like, namely inlet ratio with exit cross-section Example convexity is equal with export ratio convexity.But for aluminothermy is rolled, due to passages several before finish rolling Having less flakiness ratio, metal can produce certain horizontal mobility in the direction of the width, even if band There is certain difference in material entrance ratio convexity and export ratio convexity, as long as this difference is at the special formula that continues In the range of regulation, flatness defects would not be produced.The target that can make each passage accordingly is convex Degree.

Method 2a) determining each passage (Cn represents the n-th passage reciprocal) export goal convexity:

$C_{\mathrm{aim}1}=\left\{\begin{array}{c}{C}_{\mathrm{rn}}\\ C_{\mathrm{real}3}\·\frac{h_1}{h_3}\end{array}\right.,$ When C3 passage outlet convexity can be measured.

$C_{\mathrm{aim}2}=\left\{\begin{array}{c}{C}_{\mathrm{real}3}\·\frac{h_2}{h_3}\\ C_{\mathrm{rn}}\·\frac{h_2}{h_1}\end{array}\right.,$ When C3 passage outlet convexity can be measured.

$C_{\mathrm{aimi}}=C_{\mathrm{rn}}\·\frac{h_i}{h_1},$ As i=3 or i=4.

C_aimiRepresent the export goal convexity of i-th passage reciprocal, μm, C_rnRepresent enterprise requirements Finished product target convexity, μm, C_realiRepresent Ci passage actual measurement outlet convexity, μm, h_iRepresent reciprocal the The exit thickness of i passage, mm.

2b) when finishing pass is more than 4 passage, from finish rolling the first passage to the mesh of the 5th passage reciprocal Mark convexity all need not determine.

(3) the real-time roll thermal crown of working roll calculates.

Working roll flatulence of heat type is with number for impact and the bending roller force of Gap crown for the impact of Gap crown Magnitude, and owing to single chassis hot-rolling mill needs the rolling all passages from aluminium ingot to finished product, its work Make the change of roller flatulence of heat type frequently, and working roll thermal expansion value can not be obtained by measurement means, therefore can only The real-time roll thermal crown of working roll is obtained by the method calculated.

In circular cylindrical coordinate, 3a) set up the parted pattern of working roll, due to roll each rotation, circumference The boundary condition of each unit is the same, therefore can working roll model simplification be only consider axially and The radially two dimensional model of unit, as shown in Figure 2.D difference point-score is used to carry out evaluation work roller each point Temperature field, so can also improve calculating speed while ensureing precision, be suitable in line computation.

3b) according to the temperature field of working roll, the heat that can obtain each moment working roll axial unit is convex Angle value.

3c) the comprehensive convexity of working roll is made up of working roll initial cut convexity and work roll thermal crown, front Person is just it has been determined that keep constant in the operation of rolling before roll change, therefore according to calculating in real time Work roll thermal crown just can the real-time comprehensive convexity of evaluation work roller.

(4) setting value of each actuator is calculated.

Start within a period of time that profile gauge puts into feedback control (head) from threading, can only be by respectively holding The setting value of row mechanism ensures the convexity of this section of band.The whether accurate of setting value directly affects head Convexity precision, it is therefore desirable to the operation of rolling is simulated emulation, obtains Gap crown and affects roll gap Relational expression between convexity factor.

4a) passages several before the finish rolling stage are had

C_out=E_p·P+E_B·BF+E_WR·C_WR ⑴

E_P=f_P(B), E_B=f_B(B)

C_outRepresent band outlet convexity, μm, E_pRepresent that roll-force affects coefficient, μm/kN, P table Show roll-force, kN,

E_BRepresenting that bending roller force affects coefficient, μm/kN, BF represent bending roller force, kN, E_WRRepresent work The comprehensive convexity of roller affects coefficient, C_WRRepresenting the comprehensive convexity of working roll, μm, B represents rolled piece width, Mm, f_P(B)、f_B(B)、E_WRRepresent that roll-force and bending roller force affect coefficient calculations formula, by emulation data Recurrence obtains.

4b) passages several for finish rolling stage Mo have

C_out=E_p·P+E_C·C_in+E_B·BF+E_WR·C_WR ⑵

E_P=f_P(B,D_WR), E_C=f_C(B,H,h,D_WR), E_B=f_B(B,D_WR)

C_outRepresent band outlet convexity, μm, E_pRepresent that roll-force affects coefficient, μm/kN, P table Show roll-force, kN, E_CRepresent that supplied materials convexity affects coefficient, C_inRepresent supplied materials convexity, μm, E_BTable Showing that bending roller force affects coefficient, μm/kN, BF represent bending roller force, kN, E_WRRepresent that working roll is the most convex Degree affects coefficient, C_WRRepresenting the comprehensive convexity of working roll, μm, B represents rolled piece width, mm, D_WRTable Show that work roll diameter, mm, H, h represent rolled piece entrance and exit thickness, mm, E_P、E_C、E_BTable Reach formula and E_WRObtained by emulation data regression.

After each passage target convexity determines, it is possible to utilize (1) and (2) formula determines the curved of each passage Roller power setting value.(1) before, (2) the roll-force rolling force model in formula is calculated, supplied materials convexity is The outlet convexity of a time.

4c) single chassis aluminium hot-rolling mill entrance side and outlet side are provided with cooling down jet beam, entrance side cold But liquid is mainly used in basis cooling, and nozzle is distributed as being uniformly distributed, according to the different selections of alloy rigidity Different cooling grades.Coolant cooling grade classification be by force, in, weak three kinds, then different hardness Alloy use cooling grade be: hard alloy correspondence cools down by force grade, in medium hardness correspondence cool down The corresponding weak cooling grade of grade, mild alloy.

4d) nozzle of outlet side is distributed as quadratic distribution, low to adapt to high both sides in the middle of working roll flatulence of heat type Feature.Nozzle is opened number and is determined according to carrying out flitch width.In order to improve edge regulating power, edge Cooling water inflow be set as 0, for the ease of regulation injection grade, the injection grade of intermediate noxzzle can set It is set to medium cooling grade, therefore deduces that the injection grade of each nozzle:

$L_N=\mathrm{round}\left\{L_C\right[1-{\left(\frac{X_N}{X_E}\right)}^2]+E_{\mathrm{Lrn}}\}---\left(3\right)$

L_NRepresent from the nozzle injection grade that edge is N, L_CRepresent the injection grade of intermediate noxzzle, X_N Represent the nozzle abscissa of numbered N, X_ERepresent the abscissa of intermediate noxzzle, E_LrnRepresent cooling spray Drench self study coefficient.

4e) for a batch of material of same specification, from the beginning of the 2nd block, the injection of each of which nozzle Grade is:

$L_N^{\mathrm{SET}}=L_{N,\mathrm{Tail}}^{\mathrm{ACT}}$

Represent from the nozzle injection grade that edge is N,Represent that previous rolling off edge is N's Nozzle afterbody actual ejection grade.

(5) bending roller force feedforward gain calculates.

Can ensure that the plate shape of band head by setting value, but the Crown control in full belt length also needs to Control to realize by bending roller force feedforward and crown feedback.

In the operation of rolling, owing to the supplied materials temperature difference, alloying component skewness etc. can cause roll-force Fluctuation, after especially AGC comes into operation, if any process is done in the fluctuation not to roll-force, just Gap crown can be made constantly to fluctuate along with the fluctuation of roll-force, thus cause the continuous of outlet convexity Change.Bending roller force feedforward can be used to compensate the roll-force fluctuation impact for Gap crown.

5a) bending roller force can quickly regulate the convexity of roll gap, therefore can be according to the ripple of actual measurement roll-force Dynamic size determines the regulated quantity of bending roller force:

△ BF=K_BF·△P ⑷

△ BF represents the bending roller force size that needs regulate, kN, K_BFRepresent bending roller force gain coefficient, △ P table Show actual measurement roll-force undulate quantity.

5b) when AGC comes into operation, bending roller force gain can be tried to achieve by following formula:

$\mathrm{\ΔBF}=K_F\·\frac{\mathrm{CR}}{h}\·\mathrm{\ΔS}+K_F(\frac{1}{C_P}\·\frac{\mathrm{CR}}{h}-\frac{1}{K_P})\mathrm{\ΔP}---\left(5\right)$

K_FThe expression bending roller force lateral stiffness on the flexural deformation impact of roller system, kN/mm,Represent outlet Ratio convexity, △ S represents that roll gap is poor, mm, C_PRepresent milling train longitudinal rigidity, kN/mm, K_PExpression is rolled The power processed lateral stiffness on the flexural deformation impact of roller system, kN/mm.

(6) crown feedback control strategy is formulated.

Deviation according to profile gauge measured value Yu convexity desired value formulates crown feedback control strategy.Curved Roller power and chilling spray are two kinds of means of Crown control, and bending roller force regulation is swift in motion, and are suitable for online Regulation convexity.Be commonly equipped with due to single-frame double-coiling aluminium hot-rolling mill is bikini profile gauge, therefore Feedback control cannot be carried out by chilling spray, convexity can only be controlled by the way of presetting.

6a) convexity bias adjustment strategy, the outlet convexity measured when profile gauge exists with target convexity deviation Time in the convexity fluctuation range allowed, it is not necessary to plate shape is adjusted；When deviation value is more than permission Time during convexity fluctuation range less than bending roller force regulation and control high-low limit, then bending roller force is used to regulate convexity Deviation；Otherwise use and ensure that glacing flatness well sacrifices the strategy of convexity to regulate bending roller force.

6b) carry out crown feedback control strategy as shown in Figure 3 with bending roller force.Use curved with continuous hot-rolling mill Roller power carries out unlike Crown control, and the crown feedback of single chassis hot-rolling mill controls can only be in current road Carry out in secondary, the bending roller force of current pass can only be regulated, it is possible to do not compensate upstream passage accumulation Convexity error, thus need adjust chilling spray flow to adjust convexity, to improve the curved of next block Roller power regulates and controls effect；Additionally use single chassis hot-rolling mill to roll, can have up to 2 passages and enter Row crown feedback controls, and this is not available for continuous hot-rolling mill.

6c) crown feedback control in affect coefficient value by convexity setting model according to alloy, thickness, Width interval gives.

(7) convexity setting accuracy is improved by Model Self-Learning.

Owing to each coefficient of convexity equation is obtained, therefore when model is applied by emulation data regression When certain a time, calculate setting value and there will be deviation, in order to eliminate this deviation for Crown control Adverse effect, need to adjust according to measured data the self study coefficient of model, thus improve follow-up The Crown control precision of material volume.

7a) before carrying out self study, first to carry out self adaptation according to measured value, through certain time Just starting self study after the self adaptation of number, bending roller force model coefficient self adaptation formula is:

$K_{\mathrm{BF}\_\mathrm{new}}=K_{\mathrm{BF}\_\mathrm{old}}+(\frac{{\mathrm{BF}}_m}{{\mathrm{BF}}_c}-K_{\mathrm{BF}\_\mathrm{old}})\×\mathrm{VAD}---\left(6\right)$

K_{BF_new}For new bending roller force model coefficient, K_{BF_old}For old bending roller force model coefficient, BF_mFor Bending roller force measured value, BF_cFor by measured value calculated roller force value, VAD is total gain amplifier.

7b) carrying out self study after 15～20 self adaptations, bending roller force self learning model is

L_{BF_new}=L_{BF_old}+(K_{BF_new}-L_{BF_old})×VP ⑺

L_{BF_new}For new bending roller force self study coefficient, L_{BF_old}For old not yet through adaptive roller Power self study coefficient, K_{BF_new}For through adaptive model coefficient, VP is self study amplification coefficient.

7c) chilling spray pattern self study computational methods are similar with bending roller force self study coefficient, specifically count Formula is:

$E_{\mathrm{LrnN}}^{\mathrm{Cur}}=L_N^{\mathrm{ACT}}-L_N^{\mathrm{SET}}$

$E_{\mathrm{LrnN}}^{\mathrm{New}}=E_{\mathrm{LrnN}}^{\mathrm{Old}}+(E_{\mathrm{LrnN}}^{\mathrm{Cur}}-E_{\mathrm{LrnN}}^{\mathrm{Old}})\·\mathrm{\β}$

Represent the self study coefficient that n-th nozzle is current；Represent the reality of n-th nozzle Spray flow,Represent n-th nozzle initially sets flow,Represent n-th nozzle New self study coefficient,Representing the old self study coefficient of n-th nozzle, β represents self study Gain coefficient.

For below as a example by the product that 2400 single reversing hot mill with coilers located on each side of the mills of certain aluminum fabrication plant produce Bright embodiments of the present invention.

The material volume information chosen is: alloy designations 3003, plate width 1290mm, and finish to gauge thickness 7mm rolls Making total passage is 20, and finishing pass 7 wherein batches passage 3.Finish rolling stage supplied materials thickness For 100mm.

Artificial experience is used to set the roller force value of last several rolling pass, as shown in table 1.

Table 1

Table 2

It is 5.9 μ from the total length measured value of the C1 passage (passage last, lower same) shown in table 2 M, hence it is evident that less than target convex value (35 μm).It will be seen that the total length of C1 passage is curved from table 1 Roller power measured value is-1831kN, is sufficiently close to the least limit (-2200kN) of bending roller force, and this explanation is such as The outlet convexity of fruit C1 passage to be improved, the regulating measure of bending roller force is very limited, by the tune of bending roller force Joint does not still reach target convexity.And the total length actual measurement bending roller force of C4 is 900kN, the biggest is curved Roller regulating power (also exists relation: bending roller force is the biggest, Gap crown between bending roller force and Gap crown The least, rolling band convexity out is the least), this illustrates that current convexity allocation strategy is also not Reasonably.

When C1 passes, the distribution curve of milling train outlet actual measurement convexity is it will be seen that in length range The concordance of interior convexity is poor, and its fluctuation range is-50 μm～30 μm, especially at head, and its fluctuation Scope is bigger.

A kind of single-frame double-coiling aluminium hot-rolling mill Crown control system approach that the present invention provides, it is concrete Implementing step is:

1) overall Crown control policy development.It is 3003 that material to be rolled intervolves gold, and belong in strategy is soft Alloy species.Finish to gauge exit thickness is 7mm, has 3 to batch passage, belong to thin material in the operation of rolling Category, therefore Crown control strategy should perform the overall Crown control strategy of the thin material of mild alloy.

2) the target convexity of each passage is determined.It will be noted from fig. 1 that the most last 4 passages Needs are set calculating, and the most also have only to determine the target convexity of these 4 passages.Due to C3 The convexity of passage can be measured, and the target convexity allocation result of the most last 4 passages is as shown in table 3.

Table 3

3) work roll thermal crown calculates.The calculating of the real-time roll thermal crown of working roll is completed by flatulence of heat type program. Program read rolling state and the change of working roll boundary condition every 2 seconds, calculated one action roller real Time roll thermal crown, the roll thermal crown value calculated each time is stored in work roll thermal crown and calculates in journal file. Therefore, when carrying out convexity setup algorithm, required roll thermal crown data can be read from journal file.

4) setting value of each actuator is calculated.After obtaining working roll real-time roll thermal crown value, just Can according to (1) formula and (2) formula calculate bending roller force setting value, determine the injection of each nozzle according to (3) formula Rank.

In C4 and C3 passage, flakiness ratio is respectively 46 and 76, respectively less than 100, therefore according to (1) formula Calculating bending roller force, the expression formula of bending roller force is

$\mathrm{BF}=\frac{C_{\mathrm{out}}-(E_p\·P+E_{\mathrm{WR}}\·C_{\mathrm{WR}})}{E_B}$

In C2 and C1 passage, flakiness ratio is respectively 112 and 184, is all higher than 100, therefore according to (2) Formula calculates bending roller force, and the expression formula of bending roller force is

$\mathrm{BF}=\frac{C_{\mathrm{out}}-(E_p\·P+E_C\·C_{\mathrm{in}}+E_{\mathrm{WR}}\·C_{\mathrm{WR}})}{E_B}$

If the more stable roller force value formed after self study meets

BF>BF_max80% or BF < BF_min80%

BF_max、BF_minRepresent equipment maximum roller force value and minimum roller force value respectively.

Then illustrate that now bending roller force is bigger than normal, owing to chilling spray is a slow process for the impact of convexity, It is thus desirable to reduced the burden of follow-up material volume bending roller force by regulation work roll thermal crown.Determining The cooling nozzle regularity of distribution, relation between cooling flow size and cooling water inflow and work roll thermal crown Afterwards, it is possible to change these conditions affecting work roll thermal crown, reduce bending roller force setting value, Thus control, for crown feedback, the bending roller force regulated quantity that offer is bigger.

5) bending roller force feedforward gain calculates.In the operation of rolling, roll-force can be measured, Thus roller force value can be adjusted according to actual measurement roll-force difference, when AGC puts into and does not comes into operation, Bending roller force gain calculating formula is that institute is distinguishing.

After threading, it is possible to carrying out the measurement of roll-force, now bending roller force setting value does not acts as With, therefore can adjust setting roller force value according to the bending roller force deviation calculated；When convexity is anti- After feedback control is come into operation, the purpose of the feedforward mainly compensates curved according to roll-force undulating value Roller power deviation value, thus keep stablizing of roll gap lateral stiffness, it is thus achieved that good convexity and glacing flatness.

6) crown feedback control realization.Feedback control only meets the passage of condition and is just performed, In this example, perform crown feedback control strategy in C3 passage.In C3 passage, special public according to continuing Formula, the entrance and exit ratio convexity difference of this passage meets

$-0.0254\&le;\frac{C_{\mathrm{in}}}{H}-\frac{C_{\mathrm{out}}}{h}\&le;0.0127$

The range of accommodation of bending roller force must assure that entrance and exit ratio convexity difference within the scope of this, no Then it is possible that the situation of deformed steel strip.

7) Model Self-Learning realizes.In the operation of rolling, bending roller force, roll-force, outlet convexity etc. Process data can be measured, and therefore according to actual measurement data, model can be carried out self study, To improve the precision of model.

At the head of rolled strip, (10 milliseconds) sampled point can be taken at regular intervals, adopt Collect 10 sampled points, and these 5 sampled points are averaged, the sample data sampled as this, even Continuous 10～15 groups of sample datas sample as model adaptation that gathers, often group sample calculates according to (6) formula One adaptation value, after 10～15 self adaptations, calculates the self study coefficient of this volume by (7) formula, Then the same specification alloy Constant in self study table is updated, when rolling for follow-up same specification alloy.

Along with being continuously increased of rolling material volume, the self study coefficient value in self study table gradually tends towards stability, Head convexity rolled up by rolling material out and strip shape quality the most just can be continuously available improvement.

By step as above, solve problems with present in producing at present:

1) the bending roller force computation model derived according to Shape theory is lacked.

2) cannot Accurate Prediction work roll thermal crown, the Strip Shape Control ability of chilling spray cannot be abundant Bring into play.

3) manually adjustable plate shape, it is impossible to reach glacing flatness and the best purpose of convexity.

The present invention formulates single-frame double-coiling hot rolling according to feature and the on-the-spot rolling experience of this type milling train Crown control strategy, set up be applicable to the convexity preset model of single chassis, flatulence of heat type computation model, The feedforward of convexity and feedback control system, it is achieved Aluminum Plate and Strip is convex in the full-automation of this kind of mill milling Degree well controls with glacing flatness, i.e. according to the alloy feature of band (by alloy species, hardness, matter Amount requirement and finished product thickness etc. divide) formulate corresponding Crown control strategy, convexity preset cover half Type and flatulence of heat type model provide roughing, the bending roller force setting value of two rolling sequences of finish rolling and chilling spray mould Formula ensures head convex value, carries out the feedforward compensation of bending roller force to reduce roll-force fluctuation at finishing pass Impact on convexity, special according to alloy in the passage third from the bottom or extreme trace time that can carry out convex measuring The situation of levying carries out bending roller force closed loop control to ensure total length convexity and plate shape, carries out curved after the end of rolling Roller power and the self study of chilling spray model, improve preset model precision.Present invention can ensure that aluminum Strip reaches target convex value when single reversing hot mill with coilers located on each side of the mill rolls in head and length range, Keep good plate shape simultaneously, improve the economic benefit of enterprise.

It is to be understood that: the above is only the preferred embodiment of the present invention, for this technology For the those of ordinary skill in field, under the premise without departing from the principles of the invention, it is also possible to if making Dry improvements and modifications, these improvements and modifications also should be regarded as protection scope of the present invention.

Claims (6)

1. single-frame double-coiling aluminium hot-rolling mill convex degree control method, it is characterised in that comprise the following steps:

(1) overall Crown control policy development；

(2) the export goal convexity of each rolling pass band is determined；

(3) the real-time roll thermal crown of working roll calculates；Specifically include following step:

A) temperature field of D difference point-score evaluation work roller each point is used；

B) according to the temperature field of working roll, the roll thermal crown value of each moment working roll axial unit is obtained；

C) the comprehensive convexity of working roll is made up of working roll initial cut convexity and work roll thermal crown, the former be before roll change the most Determine, in the operation of rolling, keep constant, according to the work roll thermal crown real-time evaluation work comprehensive convexity of roller calculated in real time；

(4) setting value of each actuator is calculated；

(5) bending roller force feedforward gain calculates；

(6) crown feedback control strategy is formulated；Specifically include following step:

A) crown feedback control strategy is formulated according to the deviation of profile gauge measured value with convexity desired value；Bending roller force and chilling spray Being two kinds of means of Crown control, bending roller force regulation is swift in motion, and is suitable for on-line control convexity；Due to single-frame double-coiling aluminium hot Milling train is equipped with bikini profile gauge, controls convexity by the way of presetting；

B) when the outlet convexity that profile gauge is measured and target convexity deviation are in the convexity fluctuation range allowed, it is not necessary to plate Shape is adjusted；Time when deviation value is more than the convexity fluctuation range allowed less than bending roller force regulation and control high-low limit, then use curved Roller power regulation convexity deviation；Otherwise use and ensure that glacing flatness well sacrifices the strategy adjustment bending roller force of convexity；

C) crown feedback of single-frame double-coiling aluminium hot-rolling mill controls only to carry out in current pass, can only regulate current pass Bending roller force, adjusts chilling spray flow and adjusts convexity, regulate and control effect improving the bending roller force of next block；

D) using single-frame double-coiling aluminium hot-rolling mill to roll, be up to 2 passages carry out crown feedback control；

E) coefficient value that affects during crown feedback controls is given according to alloy, thickness, width interval by convexity setting model；

(7) convexity setting accuracy is improved by Model Self-Learning.

Single-frame double-coiling aluminium hot-rolling mill convex degree control method the most according to claim 1, it is characterised in that: described totally Crown control policy development is pressed aluminium alloy kind, hardness, prescription and finished product thickness and is divided the alloy feature of aluminum strip, and Formulating corresponding Crown control strategy: alloy species is different, its rolling strategy and finish to gauge require also different, and alloy soft or hard degree is not With, its Crown control strategy is the most otherwise varied；Finish to gauge thickness is different, and rolled plate or its Crown control strategy of coiled material are the most different；

A) formulation of Crown control strategy need to consider the product quality emphasis of variety classes alloy, yield, temperature, surface quality It is 4 important quality index of board rolling with plate shape, when with yield or temperature for when paying the utmost attention to index, for convexity Control to require to reduce；When with surface quality for when paying the utmost attention to target, convexity closed loop control needs passage in advance to carry out, to reduce Extreme trace time convexity regulates and affects surface quality of strips；When with plate shape for when paying the utmost attention to index, then need to optimize corresponding load Distribution is to ensure exit plate shape；When being common priority target with several quality index sometimes, then need to formulate corresponding Crown control Strategy；

B) at rough rolling step, each passage band outlet flakiness ratio is smaller, and band horizontal mobility is relatively big, uses experience bending roller force to roll System；From the beginning of C4 passage, i.e. inverse the 4th passage of the operation of rolling, implement different according to alloy soft or hard degree with finished product thickness Crown control strategy；For thin material, the passage that its operation of rolling camber instrument comes into operation has 2, therefore implements convexity anti- The passage that feedback controls is up to two, depends on the soft or hard degree of alloy；Mild alloy has higher convexity regulation energy than hard alloy Power, therefore mild alloy is implemented 2 crown feedbacks and is controlled passage strategy, and hard alloy only carries out crown feedback control in C3 passage； For medium hardness alloy, in the operation of rolling, use hard alloy or the Crown control strategy of mild alloy according to practical situation；Right For thickness material, the passage come into operation due to profile gauge only has 1, therefore only carries out crown feedback control in last passage System.

Single-frame double-coiling aluminium hot-rolling mill convex degree control method the most according to claim 1, it is characterised in that: described determine The export goal convexity of each rolling pass band:

A) determining each passage export goal convexity, Cn represents the n-th passage reciprocal:

When C3 passage outlet convexity can be measured

When C3 passage outlet convexity can be measured

As i=3 or i=4

C_aimiRepresenting the export goal convexity of i-th passage reciprocal, μm, Crn represents the finished product target convexity of enterprise requirements, μm, C_realiRepresenting the i-th passage actual measurement outlet convexity, μm, h i represents the exit thickness of the i-th passage reciprocal, mm；

B) when finishing pass is more than 4 passage, from finish rolling the first passage, the target convexity to the 5th passage reciprocal is the most uncertain.

Single-frame double-coiling aluminium hot-rolling mill convex degree control method the most according to claim 1, it is characterised in that: described calculating The setting value of each actuator:

A) passages several before the finish rolling stage are had

C_out=E_p·P+E_B·BF+E_WR·C_WR

E_P=f_P(B), E_B=f_B(B)

C_outRepresent band outlet convexity, μm, E_pRepresenting that roll-force affects coefficient, μm/kN, P represent roll-force, kN,

E_BRepresenting that bending roller force affects coefficient, μm/kN, BF represent bending roller force, kN, E_WRRepresent that the impact of working roll comprehensive convexity is Number, C_WRRepresenting the comprehensive convexity of working roll, μm, B represents rolled piece width, mm, f_P(B)、f_B(B)、E_WRRepresent roll-force Affect coefficient calculations formula with bending roller force, emulation data regression obtain；

B) passages several for finish rolling stage Mo have

C_out=E_p·P+E_C·C_in+E_B·BF+E_WR·C_WR

E_p=f_P(B,D_WR), E_C=f_C(B,H,h,D_WR), E_B=f_B(B,D_WR)

C_outRepresent band outlet convexity, μm, E_pRepresenting that roll-force affects coefficient, μm/kN, P represent roll-force, kN, E_C Represent that supplied materials convexity affects coefficient, C_inRepresent supplied materials convexity, μm, E_BRepresenting that bending roller force affects coefficient, μm/kN, BF represent Bending roller force, kN, E_WRRepresent that the comprehensive convexity of working roll affects coefficient, C_WRRepresenting the comprehensive convexity of working roll, μm, B represents and rolls Part width, mm, D_WRRepresenting work roll diameter, mm, H, h represent rolled piece entrance and exit thickness, mm, E_p、E_C、 E_BExpression formula and E_WRObtained by emulation data regression；

C) single chassis aluminium hot-rolling mill entrance side and outlet side are provided with cooling down jet beam, and it is cold that the coolant of entrance side is mainly used in basis But, nozzle is distributed as being uniformly distributed, and selects different cooling grades according to the difference of alloy rigidity；Coolant cooling grade is drawn Be divided into strong, in, weak three kinds, then the cooling grade that the alloy of different hardness uses is: hard alloy correspondence cools down by force grade, medium Hardness correspondence cools down grade, the corresponding weak cooling grade of mild alloy；

D) nozzle of outlet side is distributed as quadratic distribution, to adapt to the feature that in the middle of working roll flatulence of heat type, high both sides are low；Nozzle is opened individual It is fixed that several evidences carry out flitch width；In order to improve edge regulating power, the cooling water inflow of edge is set as 0, for ease of regulation injection etc. Level, the injection grade of intermediate noxzzle is set as medium cooling grade, it follows that the injection grade of each nozzle:

$L_N=round\{L_C\&lsqb;1-{\left(\frac{X_N}{X_E}\right)}^2\&rsqb;+E_{Lrn}\}$

N represents that nozzle is numbered, L_NRepresent from the nozzle injection grade that edge is N, L_CRepresent the injection grade of intermediate noxzzle, X_NRepresent the nozzle abscissa of numbered N, X_ERepresent the abscissa of intermediate noxzzle, E_LrnRepresent chilling spray self study coefficient；

E) for a batch of material of same specification, from the beginning of the 2nd block, the injection grade of each of which nozzle is:

$L_N^{SET}=L_{N,Tail}^{ACT}$

Represent from the nozzle injection grade that edge is N,Represent that previous coiled strip when afterbody rolls from edge is The nozzle actual ejection grade of N.

Single-frame double-coiling aluminium hot-rolling mill convex degree control method the most according to claim 1, it is characterised in that: described bending roller force Feedforward gain calculates:

A) bending roller force quickly regulates the convexity of roll gap, determines the regulated quantity of bending roller force according to the fluctuation size of actual measurement roll-force:

Δ BF=K_BF·ΔP

Δ BF represents the bending roller force size that needs regulate, kN, K_BFRepresenting bending roller force gain coefficient, Δ P represents actual measurement roll-force Undulate quantity；

B) when AGC comes into operation, bending roller force gain is tried to achieve by following formula:

$\mathrm{\&Delta;}BF=K_F\&CenterDot;\frac{CR}{h}\&CenterDot;\mathrm{\&Delta;}S+K_F(\frac{1}{C_P}\&CenterDot;\frac{CR}{h}-\frac{1}{K_P})\mathrm{\&Delta;}P$

K_FRepresenting the bending roller force lateral stiffness on the flexural deformation impact of roller system, kN/mm, CR represent export ratio convexity, Δ S table Show that roll gap is poor, mm, C_PRepresent milling train longitudinal rigidity, kN/mm, K_PRepresent that the flexural deformation of roller system is affected by roll-force horizontal Rigidity, kN/mm, h represent exit thickness.

Single-frame double-coiling aluminium hot-rolling mill convex degree control method the most according to claim 1, it is characterised in that pass through described in: Model Self-Learning raising convexity setting accuracy:

A) before carrying out self study, substitute into, according to bending roller force measured value and this measured data of use, the bending roller force that model recalculates Value carries out self adaptation, just starts self study after the self adaptation of certain number of times, and bending roller force model coefficient self adaptation formula is:

$K_{BF\_new}=K_{BF\_old}+(\frac{{\mathrm{BF}}_m}{{\mathrm{BF}}_c}-K_{BF\_old})\&times;VAD$

K_{BF_new}For new bending roller force model coefficient, K_{BF_old}For old bending roller force model coefficient, BF_mFor bending roller force measured value, BF_cRecalculating, for substituting into model by measured data, the roller force value obtained, VAD is total gain amplifier；

B) carrying out self study after 15～20 self adaptations, bending roller force self learning model is

L_{BF_new}=L_{BF_old}+(K_{BF_new}-L_{BF_old})×VP

L_{BF_new}For new bending roller force self study coefficient, L_{BF_old}For old not yet through adaptive bending roller force self study coefficient, K_{BF_new}For through adaptive model coefficient, VP is self study amplification coefficient；

C) chilling spray pattern self study computational methods are similar with bending roller force self study coefficient, and calculating formula is:

$E_{LrnN}^{Cur}=L_N^{ACT}-L_N^{SET}$

$E_{LrnN}^{New}=E_{LrnN}^{Old}+(E_{LrnN}^{Cur}-E_{LrnN}^{Old})\&CenterDot;\mathrm{\&beta;}$

Represent the self study coefficient that n-th nozzle is current；Represent the actual spray flow of n-th nozzle, Represent n-th nozzle initially sets flow,Represent the new self study coefficient of n-th nozzle,Represent N The old self study coefficient of individual nozzle, β represents self study gain coefficient.