GB2278464A

GB2278464A - Reverse rolling control system of pair cross rolling mill

Info

Publication number: GB2278464A
Application number: GB9412120A
Authority: GB
Inventors: Tetsuya Ogawa; Yuri Hori
Original assignee: Nippon Steel Corp
Current assignee: Nippon Steel Corp
Priority date: 1992-11-10
Filing date: 1993-11-10
Publication date: 1994-11-30
Anticipated expiration: 2013-11-10
Also published as: GB9412120D0; SE9402305D0; WO1994011129A1; KR0148612B1; SE505470C2; GB2278464B; SE9402305L; KR940703722A

Abstract

In a pair cross rolling mill, a true rolling load free from influences of a thrust load occurring during rolling and an actual difference in load between roll ends are calculated, and sheet thickness control and wedge control are carried out through learning on the basis of detection values and calculated values. To accomplish a predetermined sheet crown, an actual sheet crown is estimated and calculated on a real time basis and is used for feedback control so that an actual roll bending quantity of a previous pass is reflected on a cross angle set value of next and subsequent passes at the time of completion of a given pass, and the cross angle set value is updated through learning. Further, the range of a mechanical crown for satisfying the roll shape in each pass and setup capacity is calculated, and a pass schedule is decided so as to attain the smallest number of passes. <IMAGE>

Description

Reference Numerals of Figures 1, 6 ... load cell 2 ... backup roll (BUR) 3 ... backup roll bearing chock 4 ... work roll bearing chock 5 ... work roll (WR) 7 ... rolling material 8 ... reduction device 8A ... hydraulic reduction device 8B ... hydraulic setter 9 ... work roll bending apparatus 9A ... roll bending apparatus 9B ... bending setter 10 ... sequencer 11 ... process computer llA ... finish pass schedule calculation unit llB ... adaptive control calculation unit 12 ... business computer 13 ... plant controller 14 ... housing 15T ... radiation thermometer lSH ... gamma-ray thickness gauge 16 ... table 17 ... cross angle setter 18 ... bearing support frame 19 ... pull-back cylinder METHOD FOR REGULATING REVERSE ROLLING OF PAIR CROSS MILL FIELD OF THE INVENTION This invention relates to a reverse rolling control method of a pair cross rolling mill having the function of crossing upper and lower rolls as pairs, and more particularly to a plate thickness control and shape control of a steel plate and a pass schedule determination method.

BACKGROUND OF THE INVENTION In a pair cross rolling mill for reverse rolling a rolling material by relatively crossing roll sets each comprising a pair of backup rolls and work rolls inside a plane parallel to the rolling material, a method of independently controlling the sheet thickness on the right and left by eliminating the influences as result of providing thrust load cells has been proposed in the past because unbalance occurs in the loads applied to right and left load cells due to the influence of thrust force occurring during rolling (Japanese Examined Patent Publication (Kokoku) No. 63-23851). A method which disposes load cells at upper and lower positions and controls the sheet thickness by reducing the influence of hysteresis has also been proposed (Japanese Examined Patent Publication (Kokoku) No. 63-1128). Further, a method which calculates a true rolling load (the difference and sum of right and left loads) by offsetting the unbalance of the loads by upper and lower and right and left load cells, and which controls the sheet thickness has been proposed. Rolling has been carried out by detecting the thrust load occurring in a roll axis direction by a single load cell and using it for checking a withstand load of a setup.

According to the prior art which off sets the load unbalance by the upper and lower and right and left load cells, control can be made on the premise that the upper and lower mill rigidity is uniform. In an actual rolling mill, however, a detection load error occurs depending on the condition of a hit surface of a roll chock, a maintenance condition of bearings, and so forth, and there remains the problem that it is not possible to discriminate whether the unbalance of the detection load results from an eccentric load due to the thrust or from an eccentric load due to deformation of the to-be-rolled material.

When any difference exists between the upper and lower mill spring quantities, the influence of the load unbalance due to a thrust moment is not uniform.

Consequently, a mill stretch difference of the right and left mills changes. For this reason, properties reverse between normal rotation and reverse rotation particularly in reverse rolling, and wedge control of the to-be-rolled material becomes difficult.

On the other hand, to accomplish automatic crown shape control in the pair cross rolling mill, a schedule of a cross angle per slab has been subjected to preset setting control.

Since this prior art technology is the preset control of one slab before rolling, there is the problem that correction is not possible even when no difference occurs from an estimation value due to disturbance of rolling, if re-calculation is not made with the progress of a pass while the sheet material is actually rolled.

A determination method of a pass schedule in the conventional rolling mill is described in Japanese Unexamined Patent Publication (Kokai) No. 62-259605.

This method determines the reduction schedule under the condition where the rolling load is limited to the maximum setup allowance capacity in order to limit the change of a sheet crown for each pass and to achieve flatness of the rolling shape.

Japanese Examined Patent Publication (Kokoku) No. 63-123 describes a pass schedule determination method which rolls a rolling material at the full load in an upstream pass having little shape influence and limits the load only in downstream passes where the shape is susceptible to the change of the sheet crown, so as to highly efficiently roll the rolling material.

On the other hand, when the number of passes is fixed from the number of stands as in continuous hot rolling, a method which determines in advance the plate thickness schedule of all the passes and then determines the schedule of a roll cross angle so as to satisfy the shape has been proposed in No. 120th Conference of Japan Iron & Steel Society CAMP-ISIJ Vol 13 (1990), p1388.

According to the conventional rolling methods described above, the change of the plate crown ratio for each pass must be restricted within a predetermined range in order to make the shape of the rolling material flat.

Therefore, the rolling load as a governing factor of a mechanical crown is limited and rolling must be carried out at a load by far smaller than the setup capacity, so that the number of passes becomes great and rolling efficiency drops.

CONSTRUCTION OF THE INVENTION The present invention is completed in view of the problems described above. In other words, it is a first object of the present invention to make it possible to accurately control the plate thickness and the wedge by correctly separating and offsetting the load unbalance described above when the plate thickness is automatically controlled.

It is the second object of the present invention to make it possible to accurately control the plate crown and the shape by correctly predicting the shape of the plate material during each pass and making a correction calculation for each pass when the plate crown and the shape are automatically controlled.

It is the third object of the present invention to obtain a pass schedule which fully exploits the capacity of the rolling setup throughout all the passes, accomplishes high efficiency rolling minimizing the number of passes and optimizes the rolling shape on the premise that the cross angle is changeable in each pass.

The gist of the present invention resides in the following points.

(1) In a rolling mill for reverse rolling by relatively crossing roll sets each comprising a pair of backup rolls and work rolls inside a plane parallel to a rolling material, an automatic plate thickness control method in reverse rolling of a pair cross rolling mill characterized in that a true rolling actual record load and a load difference between right and left loads, free from the influences of a thrust loads, are calculated by the use of detection values of upper and lower and right and left load cells for detecting a rolling load and a load cell for detecting a thrust load occurring in a roll axial direction, and the plate thickness and a wedge are automatically controlled on the basis of these detection values and the calculated values.

(2) In rolling by a reverse rolling mill which relatively crosses roll setts each comprising a pair of backup rolls and work rolls inside a plane parallel to a rolling material, and which includes a roll bending controller at both end portions of a roll, a shape control method in reverse rolling of a pair cross rolling mill characterized in that an actual record crown is calculated on the real time basis in accordance with a rolling load fluctuation inside a plate in order to accomplish a plate crown determined in advance for making the final shape of the plate flat, roll bending control is executed, a plate crown calculation quantity is corrected by an actual record roll bending quantity of a previous pass at the end of a pass, and is reflected on a cross angle set value of next and subsequent passes, and the shape of the plate is thus controlled.

(3) In a rolling operation by a reverse rolling mill which relatively crosses roll sets each comprising a pair of backup rolls and work rolls inside a plane parallel to a rolling material, and which includes roll bending controllers at both end portions of a roll, a shape control method in reverse rolling of a pair of cross rolling characterized in that, when a roll bending quantity is corrected in accordance with an exit side shape of each pass, a shape estimation value of the pass is corrected from a roll bending actual record value of the pass, a plate crown estimation value is further corrected, a mechanical crown quantity as a target value in next and subsequent passes is calculated again, and is reflected on a cross angle set value, and the shape of the plate is thus controlled.

(4) When a pass schedule is determined at the time of rolling of a plate material by a reverse rolling mill including a controller for relatively crossing an upper roll set and a lower roll set each comprising a pair of backup rolls and work rolls inside a plane parallel to a rolling material, a reverse rolling schedule determination method of a pair of cross rolling mill characterized in that a region which simultaneously satisfies a mechanical crown allowable range judged from a shape in each pass and a mechanical crown allowable range judged from a setup capacity is calculated, a plate thickness schedule is determined by sequentially building up the values so as to attain an allowable maximum rolling load among them, and reduction of the shortest pass number satisfying the shape and a cross angle schedule are simultaneously determined.

BRIEF DESCRIPTION OF THE DRAWINGS Fig. 1 is a front view showing the outline of an example of a rolling mill to which the present invention is applied; Fig. 2 is a block diagram showing a rolling control system of the rolling mill shown in Fig. 1; Fig. 3 is a flowchart showing the content of a pass schedule determination method of the present invention; Fig. 4 is a graph showing camberes occurring in thick plate rolling according to the prior art and camberes occurring in an example of the present invention; Fig. 5 is a block diagram showing constituent elements of a rolling mill and a controller in an embodiment of the present invention; Fig. 6 is a side view showing the outline of the mechanism of a reverse rolling mill shown in Fig. 5; Fig. 7 is a flowchart showing the contents of previous pass actual record calculation and learning calculation processing of an adaptive control calculation processing unit llB shown in Fig. 5; Fig. 8 is a flowchart showing a next pass setting calculation process of the adaptive control calculation processing unit llB shown in Fig. S; Fig. 9 is a block diagram showing the content of a dynamic shape control processing of a plant controller 13 shown in Fig. 5; Fig. 10 is a block diagram showing a dynamic shape control functional construction of the plant controller 13 shown in Fig. S; Fig. 11 is a time chart showing the timing of each of various calculation processings contained in the shape control according to the present invention; Figs. 12(a) and 12(b) are graph showing a crown ratio actual record value and a cross angle set value in the shape control according to the present invention and to the prior art; Figs. 13(a) and 13(b) are graph showing the occurrence frequency of a crown ratio change quantity of the final pass in the shape control according to the present invention and the prior art; and Figs. 14(a), 14(b) and 14(c) are graph showing the number of times of passes in the present invention and in the prior art method, and a rolling load, a cross angle and a crown ratio.

BEST MODE FOR PRACTICING THE INVENTION To accomplish the objects described above, the first invention of the present invention calculates a true rolling actual record load free from the influences of a thrust load and the difference of right and left loads from detection values of a vertical load detection load cell and a thrust load detection load cell, and controls a plate thickness and a chamber on the basis of these detection values and calculation values.

Hereinafter, the first embodiment of the present invention will be explained in detail. First of all, the content of the calculation of the true rolling actual record load and the difference of the right and left loads will be explained. Fig. 1 shows the outline of the front surface of a pair cross rolling mill and loads acting on a rolling mill and on rolls during rolling of a material. When the material is rolled by this rolling mill, the following loads occur: rolling load in a vertical direction from the material: P, and thrust load in a roll axial direction: F These loads P and F occurring from the material are detected by load cells 1 and 6 through a work roll 5 to backup roll 2, or the work roll 5 to a work roll chock 4, respectively.

In the actual rolling operation, P is not always distributed uniformly to the right and left because the material 7 to be rolled deviates by a from the center position in the roll axial direction and a plastic resistance error occurs in the right and left portions of the material. At this time, the formulas (1) to (8) of the fundamental equilibrium condition can be established from the equilibrium formula of the load and the formula of moment of rotation with a point A between the right and left load cells as the center.

[Rolling load] At the time of: ON load up: PTWS + PTDS = P + PFT ... (1) down: PBs + PBDS = P - PFT ... (2) OFF load up: PTWS + PTDS = P - PFT ... (3) down: PBWS + PBDS = P + PFT ... (4) [Thrust load] up: F = FTWR + FTBUR + FTHC ... (5) down: F = FBn + FBBUR + FBHG ... (6) [Equilibrium of moment of rotation] up: F[d + (DB/2) + DW] - FTWR[d + (DB/2) + (DW/2)] - FTBURd - PTos(L/2) + PTus(L/2) - Pa = 0 ... . . (7) down: F[e + (DB/2) + DW] - FB[e + (DB/2) + (OW/2)] - FBBURe - PBos(L/2) + PWus(L/2) - Pa = 0 ... . . (8) Symbols used in the formulas (1) to (8) are shown in Fig. 1. Next, a method of detecting a true load P received from the material 7 from load cell loads (detection values of the load cells 1, 6) which can actually be measured, by the use of these load unbalance formulas will be explained below.

In the thrust load formulas (5) and (6), the components of force acting on the work roll 5 and on the backup roll 2 can be expressed as given below using distribution coefficients ac and , respectively: FTWR = αF ... (9) FTBUR = F ... (10) FTHC = (1 - a - )F ... (11) When these formulas (9), (10) and (11) are substituted for the formula (7) and are put in order: F[d + (DB/2) + DW] - aF[d + (DB/2) + (DW/2)] F.d - (L/2)(PTDs - PTWS) - Pa = 0 PTDS - PTWS = F.(1/L)[2d(1 - α - ) + DB(1 - α) + DW(2 - α)] + (2Pa/L) ... (7') Here, if the thrust load is regarded as the force FTHC @ 0 which the housing itself receives, when DB @ 2DW, ~ PTDS - PTws = F.(DW/L)(4 - 3a) + (2Pa/L) ... (12) When the thrust load acting on the work roll 5 (WR) can be observed, for example, the following expression can be made: PTDS - PTWS = FTWR(DW/L)(4 - 3α)/α + (2Pa/L) ... . (13) When the lower side, too, is put in order, PBDS + PBWS = -F(1/L)[2e(1 - α - ) + DB(1 - α) + DW(2 - a)] + 2Pa/L ... (8') from PBDS - PB, = -F.DW(4 - 3a)/L + (2Pa/L) ... (14) from PBDS - PBWS = - FBWR(DW/L)(4 - 3α/α) + (2Pa/L) (15) When the formulas (12) and (14) are combined, the first term of the right side (the thrust load influence term) can be offset and the following equation can be obtained: (PTDS - PTWS) + (PBDs - PBWS) = (4Pa/L) . . (16) The true off-center rolling load difference can be calculated from the formula (16).

Further, the second term of the right side (the offcenter influence term) can be offset from the difference between the formulas (12) and (14), and the following equation can be obtained: (PTDS - PTWS) - (PBDS + PBWS) = 2F.[DW(4 - 3α)/L] ... (17) In this way, the unbalance force due to the thrust force can be extracted.

When the thrust load FWR acting on the work roll 5 (WR) can be detected or calculated, for example, [(PTDS - PTWS) - (PBDS - PBWS)]/2 = FWR.(DW/L)(4 - 3α/α) ... (18) Therefore, the a value can be estimated.

When the work roll diameter DW = 985 and the distance L = 5918 between the right and left bearings are substituted, and the left side @ aPm from the mill dimension, for example, - a = 4/[(L.#Pm)/(FWR.DW + 3) - 4/[(6#Pm)/(FWR + 3)) ... (19) For examples, the following values are given in this case: when FWR = 200T and #Pm = 150T, α = 0.533 Further, the true thrust load F received from the material can be expressed from the formulas (19) and (9) as follows: F = (3#Pm/2) + (3F/4) . . (20) In the case described above, F becomes approximately 375 tons.

The thrust load applied to the backup roll 2 (BUR) is given as follows: FBUR = (3aPm/2) - (1Fu/4) . . . (21) Further, the true rolling load sum devoid of the influences of the mill hysteresis can easily be recognized by combining the formulas (1) to (4).

The relationships described above can be summarized as follows.

The true rolling load sum P devoid of the influences of the mill rolling system frictional force (hysteresis) is: P = (PTDS + PTWS + PBDS + PBWS)/2 ... (22) The true right and left rolling load difference PREF from the material, which is devoid of the thrust load influences, is: PREF = (PTDS - PTWS + PBDS - PBWS)/2 ... (23) The true thrust load received from the material 7 is: F = (3aPm/2) + (3Fu/4) . . . (24) The thrust load which the backup roll 2 receives is: = = (3#Pm/2) - (1F/4) . . . (25) The thrust load distribution ratio which the work roll S receives is: a = 4/[6(#Pm/FWR) + 3] . . . (26) These values can be obtained by such a calculation. The present invention detects the vertical load and the transverse load by the use of the rolling load detection load cells 1 disposed both vertically and transversely and the thrust load detection load cells 6 for detecting the thrust load acting on the work roll 2, calculates the true load P and the right and left load difference PREF received from the material 7 in accordance with these detection values and with the formulas (22) to (26), and also calculates the thrust load F, etc.

Next, the influences of the load unbalance on the right and left mill stretch difference will be explained below. Assuming that the spring constants of the backup roll support portions of the upper and lower portions are KTD, KTW, KBD and KBW on the work side (WS)/drive side (DS), respectively, the rolling load detection single load cell load is as such reflected on the stretch: DS stretch: SD = (PTDS/KTD) + (PBDS/KBD) WS stretch: SW = (PTWS/KTW) + (PBWS/KBW) stretch difference DS - WS: Sup = SD SREF = [[PTDS/KTD) - (PTWS/KTW)] + [(PBDS/KBD) (PBWS/KBW)] = [1(PTDS - PTWS)KTD] + [1(PBDS - PTDS)/KBD] + PTws[(1/KTo) - (l/KTu)] + PBWS[(1/KBD) (1/KBW)] . . . ( 27) Assuming that the right and left mill rigidities are equal to each other, KT = KTD = KTW and KB = KBD = KBW Accordingly, the formula (27) becomes as follows: SREF = [1(PTDS - PTWS)KT] + [1(PBDS - PTDS)/KB] SREF = [PREF(1/KT) + (1/KB)] + [#Pm(1/KT) - (l/KB)] . . , (28) Here, aPm is given as follows from the formula (24): aPm = [(2/3) - a]F . . . (29) Therefore, when (1/KT) + (1/KB) = 1/K, S = (P=Z/K) + [F(2/3) - a][(l/KT) - (1/KB)] (30) In other words, if the spring constants are known beforehand, and the real load difference PEr and the trust load F can be estimated, the difference S between the right and left mill stretch quantities can be estimated.

Next, reflection (feedback: FB) of the load, etc., thus calculated on the plate thickness and wedge control will be explained.

Feedback to the control can be classified broadly into the following two groups from the control timing.

(1) Reflection on preset control between passes (2) Reflection on dynamic control in roll bites Hereinafter, reflection on the preset control between the passes described in the item (1) will be explained. Fig. 2 shows a system configuration for executing this preset control. A process computer 11 receives data necessary for rolling from a business computer 12, and decides in advance reduction schedules for all the passes (pass schedule calculation unit).

Next, the preset data for controlling the rolling mill for each pass at the timing at which the plate is actually rolled is calculated at an "adaptive control calculation unit", and the data is transferred to a sequencer 10. The sequencer 10 receives the set value for each pass and converts it to a signal for practically executing a reduction position control by a process computer 11, drives a hydraulic apparatus and a motor of the rolling mill, and sets them to predetermined positions and predetermined pressures. The above explains the outline of the preset control. By the way, the process computer 11 includes a learning calculation unit which stores the rolling records of the passes inclusive of the pass just before or a further previous pass or passes as well as the detection values of the sensors, and reflects them on the set calculation of this pass, and a role profile calculation unit which estimates the change with time resulting from the wear of the rolls and to thermal expansion.

Next, reflection on the preset control of the item (1) described above in the set calculation of this process computer 11 will be explained in detail.

The process computer 11 which receives the detection data of the load cells 1, 6 in the previous pass calculates the true load, etc., from the material in the previous pass in accordance with the afore-mentioned formulas (22) to (26).

Here, true rolling load sum: Piece (= P) true rolling load difference: #Pmact (= PDS - PU) thrust load from material: Fact (= F) On the other hand, the load in the previous pass is estimated and calculated from the record reduction quantity of the previous pass and the rolling record data of the temperature, etc., and the difference between the estimated and calculated load and the load calculated on the basis of the detection values of the load cells 1, 6 is calculated, and the load estimation value of the next pass is corrected on the basis of this difference value.

In other words, the load estimation is learned and updated.

In other words, when the calculation values (the values obtained by estimating and calculating the load) are expressed as Pm cal, aP cal and F cal, respectively, load sum error: Ep = Pmact/Pm cal load difference error: E. = oP, - oP cal thrust load error: Ep = Fact/F cal are calculated, and correction is applied to the estimation values PESTS #PEST and FEST of the next pass in accordance with the following formulas. Here, symbol s represents the value after smoothing of the error. When the error is 1.10, for example, it means that the error can be be reduced to 1.05 in the case of 50% learning reflection: P!ST = PEST X SEP APB = APEST x SE,p PAT = FEST X sSEp The above represents the content of learning and correction of the set calculation from the pass data (the rolling record values and the detection values) to the next material. In this way, the load which the material practically applies to the rolling mill (particularly, the right and left load difference) is fed back to the plate thickness control and reliability of the set values for the plate thickness control of the next and subsequent passes becomes high, so that accuracy of the rolled plate thickness becomes high.

On the other hand, the right and left mill stretch difference of the previous pass is calculated from the detection signals of the load cells in accordance with the formula (27). Therefore, the upper and lower and right and left mill rigidities (spring constants) in the formula (27) are learned and corrected from the right and left mill stretch difference and the wedge quantity or the camber quantity of the plate thickness actually measured. Next, Pup: real rolling load difference F: thrust load predicted in the next pass are calculated at the time of the calculation of the roll gap setting of the next pass, and the mill stretch difference due to the load unbalance is calculated from the formula (30) and the right and left roll gap is set in such a manner as to offset in advance the mill stretch difference by preset. In this way, the camber can be drastically suppressed.

Next, the second and third embodiments of the present invention will be explained.

When the roll gap set value is re-calculated in rolling the plate material, the second and third inventions of the present invention calculate a record plate crown on the real time basis in accordance with the rolling load fluctuation in the plate so as to accomplish a plate crown which is determined in advance so as to make the final shape of the plate material flat, execute the roll bending control, correct the plate crown calculation value from the record roll bending quantity of the previous pass at the end of the pass, reflects it on the cross angle set values of the next and subsequent passes, and control the shape of the plate.

When an operator corrects the roll bending quantity in accordance with the exit shape of each pass, the shape estimation value of this pass is corrected by the roll bending record quantity of this pass, the plate crown estimation value is further corrected, a mechanical crown quantity as the target of the subsequent pass is again calculated, and the shape of the plate is controlled by reflecting the mechanical crown quantity on the cross angle set value.

In other words, according to the second and third inventions, the crown estimation error by the preset control of the cross angle by the estimation load can be absorbed and corrected by the control inside the bar by roll bending while measuring the actually measured load on a real time basis, and accurate crown control becomes possible.

It becomes also possible to grasp the difference between the estimation value and the actually measured crown value by the use of the record roll bending quantity in the previous pass, and to reflect this difference on the cross angle preset control of the next and subsequent passes for learning. Therefore, feedback control having high followup property can be made.

Furthermore, when the operator intervenes and corrects the exit side plate during roll bite so as to accomplish flatness of the plate shape, feed-forward control to the next pass becomes possible by adding correction learning to the estimation of the shape and crown on the basis of the record correction quantity. Accordingly, the plate shape can be made flat more correctly.

Next, the fourth embodiment of the present invention will be explained.

When a pass schedule for rolling the plate material by a reverse rolling mill having the function of crossing an upper and lower roll pair is determined, the fourth embodiment of the present invention eliminates the concept according to the prior art in which a load control pass for regulating the shape" and "a pass for rolling at the full load" are separated, or the concept that the number of passes is determined in advance and load distribution for the shape regulation is regulated, simultaneously calculates and determines the shape and crown of the rolling material for each pass and the l simultaneously determined for each pass from the downstream pass to the upstream pass in accordance with thefollowing calculation. Namely, the temperature and the rolling assumption speed on the entrance side (clamp side) are first assumed (525, S26).

Here, the upper and lower limit range values of allowable sharpness of this pass and its aim value (hereinafter referred to as #max, #min and #sim, respectively) are given from the plate width of the rolling material and the exit side plate thickness. This #sim value is 0 in principle, and the XThu and #min value are the parameters which express the shape allowable range of the size of each rolling material, and are empirically determined in accordance with the operating condition or with the required flatness of the steel plate.

The allowable stretch strain a difference and the target stretch strain difference are calculated using this value in accordance with the following formula (31): AE = (#/2).# ... (31) Furthermore, max, min and aim of of are given in accordance with the formula (32), and the allowable range of the plate crown ratio on the entry side and the aim value are calculated (S27): Cin/Hin = Cout/Hout - AE/ + a . . . (32) where Via : plate crown on entry side Hit : plate thickness on entry side C : plate crown on exit side : : plate thickness on exit side stretch strain difference influence coefficient of the crown ratio on the shape (hereinafter called "the shape change coefficient") a : shape change correction coefficient.

Here, the allowable range of the mechanical crown limited from the shape of the present pass and its aimed value are calculated (S28) from the max, min and aim of (Cin/Hin) in accordance with the following formula (33): MCX = [1/(1 - 8)] x [Cxt - #.Hout.(Cin/Hin)] . . . (33) where MCK: mechanical crown from shape crown inheritance coefficient On the other hand, the mechanical crown from the rolling load and the setup allowable capacity can be estimated and calculated from the rolling load P, the roll bending load F3, the roll cross angle and the roll profile in accordance with the following formula (34): MCh = cl.P + c2.F3 + E + c3 ... (34) where MCh : mechanical crown from setup load P : rolling load F3 : roll bending load E : mechanical crown quantity generated by roll cross angle cl : mechanical crown influence coefficient due to rolling load c2 : mechanical crown influence coefficient due to bending load c3 : mechanical crown quantity generated by roll profile.

The mechanical crown MCh from the setup load can be calculated by omitting the second term in the formula described above when the roll bending controller does not exist, and by omitting the third term of the formula when the roll cross apparatus does not exist.

In the formula (34), MCh becomes maximal at the time of the maximum rolling load Pmax and the minimum cross angle 2*d , and MCh becomes minimal at the time of the minimum rolling load Pmin and the maximum cross angle 2e=, on the contrary, and in this way, the mechanical crown allowable range from the setup load can be determined (S29). Here, when the mechanical crown quantity E formed by the roll cross angle is determined, the limit range of the mechanical crown can be determined from the rolling load limit range at that time by adding the limit condition so that the cross angle reaches the minimum in the final stage pass.

The range which simultaneously satisfies the mechanical crown limit range from the shape according to the formula (33) and the mechanical crown allowable range from the setup load according to the formula (34) is determined as the true mechanical crown allowable range in this pass. Further, the true mechanical crown aim value MC. > is determined by making correction so that MCK., exists within this range (530).

Subsequently, the optimum combination of the reduction ratio r, the roll cross angle 2e and the bending load FB is simultaneously determined on the premise that MC, is established.

In other words, assuming that P = fP(r) E = fe(2e), then, since MC,i, is as follows from the formula (34), MCsim = cl.fP(r) + c2oF1 + fe(2e) + c3, r can be expressed as r = fr(MCsim.2#.FB) ... (35) In other words, the reduction ratio r can be retrieved and determined from 2e and the bending load F8 under the condition that MC, is constant.

Since high efficiency rolling is generally desired from the aspect of the operation factor, 2e and the bending load FB are determined in the formula (35) so that the reduction ratio r becomes maximal. Other operating conditions are. also caused to reflect by the use of evaluation functions, and the like, and the optimum combination can be determined by a linear planning method, and so forth.

After the pass reduction ratio r of this pass is determined (S31), the entry side plate thickness is calculated, and the temperature drop on the pass exit side inclusive of the temperature change inside the roll bite is estimated and calculated (S32) so as to again calculate the plate temperature at the time of gripping.

Next, a more accurate rolling load (S33) and a rolling torque (S34) are calculated using this temperature, and after the load is checked (S35), the temperature, the load and the crown in the next upstream pass are repeatedly calculated.

The pass schedule can be sequentially determined by cumulatively making the calculation for each pass described above from the downstream pass to the upstream pass and finally, repetition of the calculation is completed in the pass in which the pass entry side thickness exceeds the scheduled thickness at the start of rolling.

During the generation of this pass schedule, if the scheduled thickness at the start of rolling cannot be changed, the load distribution correction calculation is made, whenever necessary, and after the plate thickness schedule is corrected, the calculation is completed and the entire pass schedule is determined.

The shape of the rolling material, the crown and the load (reduction) schedule are simultaneously calculated and determined for each pass by the pass schedule determination method according to the present invention in accordance with the shape control capacity as the premise for minimizing the cross angle at the final stage, and a cumulative calculation is made of the optimum values. In this way, the pass schedule which satisfies the shape throughout the full passes and which can roll at the maximum values of the rolling setup capacity can be determined. Since the number of times of passes can be automatically regulated in accordance with the shape control capacity described above, the capacity of changing the number of times of passes in the reversible rolling mill can be sufficiently exhibited.

Next, the present invention will be explained in further detail on the basis of examples thereof.

EXAMPLES Example 1 When the record detection loads of the previous pass are as follows: PTDS PTgs PBDs Plus FR 2222 ton 2101 ton 2105 ton 2231 ton 182 ton true rolling load is obtained as follows from the formula (22): P = 4330 ton true rolling load difference is obtained from the formula (23): PREF = -S ton true thrust load received from the material is obtained from the formula (24): oPm = 124 ton Therefore, F = 323 ton.

Furthermore, the thrust load share which the work roll receives is a = 0.563.

The set calculation of the next pass is learned and calculated for the true load group received from the material described above from the record reduction quantity of the previous pass. That is, when P cal = 4250 ton AP cal = 0 F cal = 350 ton, from E; = 0.981, sEpB = 0.99 from E.p = 5, sE@p = 3 from Ep = 0.923, sEF = 0.95 In this way, the learning values are determined.

Next, learning and correction are made for the estimated values of the next pass.

PEST = 4110 ton 4069 ton #PEST = 0 3 ton E= = 315 ton 299 ton On the other hand, as to the mill stretch difference due to the load unbalance of the previous pass, KT = KTB = KTU = 1509 ton/mm KB = KBD = KTW = 2708 ton/mm K = 969 SREF previous pass = (2222 - 2101)/1509 + (2105 - 2230)/2708 Sup = 0.034 mm Next, SREF in the next pass is given as follows from the formula (30): Sup = (3/969) + [299 x {(2/3) - 0.563} {(1/1509) - (1 -/2708)}] x (-1) = = -0.029 mm It is one of the characterizing features of reverse rolling that the second term in the formula given above becomes negative.

The load and the roll gap record for each pass when rolling is practically carried out in the embodiment described above are illustrated in Table 1.

Table 1

Pass PTDS PTWS PBDS PBWS P #P PREF F SREF number 1 2252 2132 2140 2260 4392 120 -1 184 +0.033 2 @ 2229 2254 2235 2285 4539 13 -37 20 -0.035 3 2450 2263 2260 2450 4713 189 1 283 +0.056 4 2363 2414 2400 2410 4824 -21 -29 -17 -0.039 5 2464 2273 2280 2470 4743 191 3 286 +0.031 6 2366 2409 2405 2415 4824 -17 -25 -21 -0.055 As tabulated in Table 1, the practical rolling load exhibits the behavior such that the difference load of DS-WS reverses on the upper side and the lower side during the normal rotation and the reverse rotation, and due to this influence, the mill stretch difference alternately occurs as S > in positive and reverse directions.

Accordingly, the camber occurs in the direction in which the front portion of the plate is bent towards the wS side during the normal rotation of the mill, and it is bent towards the DS side during the reverse rotation, on the contrary, and this operation is repeated. As the pass of rolling proceeds, the camber disperses in the bend maximum direction, and according to the rolling technique of the prior art, the camber becomes great as indicated by a solid line in Fig. 4.

In contrast, the camber can be drastically limited as indicated by a dotted line in Fig. 4 by estimating Sup tabulated in Table 1 immediately before each pass as in this embodiment and deciding this quantity in advance as the roll gap unbalance value.

Example 2 Hereinafter, the second and third embodiments of the present invention will be explained with reference to the drawings. Fig. 5 is a block diagram showing the construction of a control system for accomplishing the shape control of the present invention. First of all, the construction of Fig. 5 will be explained.

The process computer 11 comprises a finish pass schedule calculation unit 11A and a finish adaptive control calculation unit llB. The finish pass schedule calculation unit llA receives the steel plate data from the business computer 12, calculates in advance the schedule of the plate thickness, the temperature, etc., for each pass before the start of rolling, decides the processing contents of the finish rolling pass as a whole, feeds forward the record data between the passes to the next pass and thus makes a learning calculation, and the finish adaptive control calculation unit 11B practically rolls the plate for each pass in accordance with the schedule obtained by the pass schedule calculation unit IlA and controls on a real time basis the work roll bending in such a manner as to correspond to the detection value during rolling and to the operator input values.

Fig. 6 is a side view of the rolling mill shown in Fig. 5. Referring to Figs. 5 and 6, a radiation thermometer 1ST (see Fig. 5) detects the surface temperature of a to-be-rolled material 7 during the rolling operation, and a gamma-ray thickness gauge lSH measures the crown. Table rolls 16 are positioned on the front and rear faces of the rolling mill and transfers the to-be-rolled material 7 in synchronism with the rolling mill speed. The peripheral surface of the work roll 5 is supported by the backup roll 2, while the backup roll 2 is supported by the bearing 3. A work roll bending apparatus (hereinafter called "WRB") 9A adjusts the distance of a work roll bearing 4 to the bearing 3 and bends the work roll. A setter 9B is a controller of an apparatus 9A for setting the bending quantity. A hydraulic reduction apparatus (hereinafter called "AGC") 8A (see Fig. 5) determines the distance between the upper and lower backup roll bearings 3, that is, the roll gap (the gap between the upper and lower work rolls). A reduction setter 8B determines a reduction position. In other words, the setter 8B is a controller of the apparatus 8A. While the to-be-rolled material 7 is under the roll bite state by the rolling mill, the load cell 1 detects the rolling load. Reference numeral 18 denotes a bearing support frame of a cross apparatus which combines the backup roll and the work roll into a set on both upper and lower sides, and determines the crossing angle (twice the cross angle) of the rotary shafts of the upper and lower sets. The support frame 18 is combined with a screw 8 and is driven in a horizontal direction (transverse direction in Fig. 6) by the rotation of 'the screw 8. A pull-back cylinder 19 always applies a force in the move-back direction to the support frame 18, and suppresses a position error resulting from back-lash during driving of the support frame 18 by the screw 8.

When the upper and lower support frames 18 are driven in mutually opposite directions, the upper and lower rolls cross each other as shown in Fig. 5. A cross angle setter 17 (see Fig. 5) is a controller for energizing this cross angle regulation mechanism.

Fig. 7 shows the previous pass actual record calculation as the premise of the next pass set calculation using the shape control of the present invention and the content of learning calculation based on the previous pass actual record calculation, and Fig. 8 shows the content of the next pass set calculation.

First of all, the contents of the previous pass actual record calculation and the learning calculation will be explained with reference to Fig. 7. Before the previous pass actual record calculation is started, the plate thickness schedules for the full passes are determined in advance, and the draft (plate thickness) schedule set hereby is generally determined in such a manner as to satisfy the rolling load and the plate shape after a suitable load distribution is made. Symbol S in Fig. 7 represents the step.

First, receiving a 1-pass metal off signal, the finish pass schedule calculation processing unit llA of the process computer 11 extracts the sensor detection values of this pass (hereinafter referred to as the "previous pass") held by the adaptive control calculation unit llB and each actual record value (rolling condition and rolling result) (S1). The schedule calculation processing unit llA then calculates the plate thickness Sg". from the actual record roll gap S and the actual measurement load Pet in accordance with the gauge meter formula given below (S2): Pact Hgage = Sact + + Hofs M Here, Hofs is a plate thickness correction term by learning.

Next, the calculated thickness obtained at S2 and the measured value Hout are compared so as to grasp the difference, and Hcfs up to the previous pass is corrected (S3). Further, the calculation of the actual record value and updating of the learning value are effected similarly for the crown, too (54, S5). In other words, the mechanical crown quantity Cm formed during roll bite is determined in accordance with the following formula, and then, the actual record calculation crown quantity Cg,g is calculated by taking the hereditary influences on the entry side crown in accordance with the following formula: Cm = p x Pact + f x Fact + r x Rut + E + Cmofs C" = nC + (1 - #)Cm Next, the difference between the actual measurement crown Cact and C., is calculated at the step S5, and this difference is converted to the error of the mechanical crown Cmofs so as to effect crown learning.

Furthermore, after correction learning calculation of the temperature is made at the steps S6 and S7, the actual record calculation load and the load learning coefficient are calculated at the steps S8 and S9.

Finally, shape learning as the characterizing feature of the present invention is carried out at the steps S10 and S11. In other words, the shape of the tobe-rolled material is evaluated by a wave height/wave pitch, and when ramda (#) is expressed as the degree of sharpness, the shape is generally expressed as a sinusoidal wave shape in accordance with the following formula: # = (2/#) # {[(Cin/Hin) - (Cout/Hout + α] #} + #ofs where #: influence coefficient of the crown ratio on the shape (hereinafter called "the shape change coefficient") Here, the following is known by the calculation up to the step S9: Cia : entry side crown : : entry side plate thickness Hout : Hgage Cout : Cgage Therefore, the calculation actual record shape #gage of the previous pass can be calculated (S10).

Here, when the actual record exit side shape is flat, actual record sharpness #act = 0 Therefore, a difference occurs from gag. which is recognized by the calculation.

Quite naturally, the shape recognition error described above is preferably evaluated using the actual record shape sensor. In the practical rolling operation, however, the operator applies correction inside the bar to the WRB (work rolling bending) load by judging the shape by eye so as to make the shape flat. In other words, the operator plays the role of the sensor and feeds back (FB) the result to the WRB operation terminal.

As a result, since correction intervention of #act - 0 is made, the control FB becomes possible by approximating the error AX to #gage, at the step S11: #ofs' = a## + (1 - where a: learning smoothing term.

Next, a content of the next pass set calculation will be explained with reference to Fig. 8 showing the calculation processing flow for actually reflecting the learning result determined by Fig. 7 on the next pass set calculation. First of all, the plate thickness and the crown shaft by the schedule calculation as the target are in advance set (S12).

At the same time, the set values are given so that the initial set WRB load of the next pass is at a neutral point, and the shape and crown control is made by setting a large cross angle having a large control capacity so that they can be reflected. Next, the gripping temperature of the next pass is estimated from the estimated time till the next pass (S13), and the estimated load P., of the next pass is estimated using the load learning value Pofs' till the previous pass on the premise of the temperature described above (S14).

Next, the target crown value of the next pass is corrected from the crown ratio of the previous pass by taking the shape allowance into consideration (Step 15).

That is, Cin # out = Hout[ - {( )(#sim - #ofs')} Htn 2 Generally, #sim = 0, but when #sim is within ##crt (critical sharpness), the initial schedule Cain = Cout is as such accepted, and when it exceeds Acrt, Csim is corrected in accordance with the formula given above.

Next, the mechanical crown quantity Cmsin necessary for the next pass is calculated in accordance with the following formula, and the cross angle necessary for accomplishing Cm is determined (S16): Cmsin = (Csim - #Cin)/(1 - #) RSET = (Cm - p x PEST - f x FSET - E - Cmofs')/R Here, PEST is the estimation load at the step S14, FES? is the WRB load as the neutral point, and Cmofs' is the mechanical crown correction quantity which is learnt.

According to the formula given above, Ql Reflection of the crown learning result till the previous pass; and Reflection of the operator's correction of WRB till the previous pass; are absorbed by the cross angle having a large control capacity.

Finally, the next pass set roll gap value is calculated by the gauge meter formula given below at the step S17, and the next pass set calculation is completed: Fast ae: = Hain + - + Hof.

M Fig. 9 shows the outline of the dynamic shape control processing of the plant controller 13 shown in Fig. 5, and Fig. 10 shows the functional construction for executing this processing. Next, the dynamic control method of the plate shape during roll bite of WRB will be explained with reference to Figs. 9 and 10. The following actual record values are grasped on the real time basis during roll bite: (1) actual record instantaneous rolling load: P (2) actual record instantaneous WRB load : (3) WRB correction value by the operator AFO In the process computer 11, on the other hand, the finish pass schedule calculation processing unit llA transfers the following data immediately before rolling of this pass to the adaptive control calculation unit 11B: (1) next pass target crown : Cain (2) crown influence coefficient of load: #c/#p (3) crown influence coefficient of WRB : (4) other crown term offsets : e (5) preset WRB load : Fo Here, since Cm = p x P + f x Fw + E C = Cin# + (i - ii)Cm as given already, C can be expressed as follows: #c C = (#c/#p)P + ( )FWB + e #FWB Accordingly, when the pass schedule calculation processing unit 11A sends #c/#p, #c/#FWA and e as the crown control constants to the adaptive control calculation unit 11B, the adaptive control calculation unit lIB can estimate (calculate) the actual record crown quantity using the formula given above.

The adaptive control calculation unit llB calculates on a real time basis the actual record calculation crown quantity in roll bite by the internal arithmetic operation, and sequentially recognizes the error aC with the target value Cain A tuning gain a is applied to oC (error) so as to convert the error to the feedback correction value oFX of WRB. In this way, interference with AGC (Automatic Gauge Control) is eliminated, and correction is automatically applied to the set value of WRB. In other words, the processing goes round the feedback loop which repeats processings comprising the calculation of the actual record crown quantity calculation of the error AC - calculation of the correction quantity oFW of WRB - correction of the WRB set value.

Further, the shape correction function by the operator is added by the arrangement such that the correction quantity by the operator can be reflected outside the feedback loop described above ("wB correction quantity" in Fig. 10).

Fig. 11 shows the execution timings of the various operations of the process computer 11 described above.

As described above, the present invention makes it possible to reflect the correction by the operator without disturbance in addition to the closed loop of the automatic control of WRB, and furthermore, in the present between the passes after roll bite, the present invention lets the result be reflected on the cross angle control of the next pass.

Fig. 12 shows the cross angle set value and the crown ratio actual record value in this embodiment in comparison with those of the prior art method.

Incidentally, the prior art method has the automatic control function of WRB, but represents the case where feedback of the actual values inclusive of the operator's correction is not reflected on the cross angle setting.

According to the prior art method, the crown ratio change for each pass is unstable and is particularly great in the final pass stage and eventually, a rolling wave occurs. In contrast, according to the method of the present invention, rolling can be made with a constant crown ratio throughout all the passes, and the degree of flatness and the shape can be improved.

Further, Fig. 13 shows the frequency distribution table of the crown ratio change quantity in the last pass which exerts the sharpest influence on the rolling shape in the present invention in comparison with the prior art method. When the present invention is applied to the control, crown control accuracy can be drastically improved.

Example 3 Another embodiment as the fourth invention of the present invention relating to the schedule determination method will be explained hereinafter.

The pass schedule of the rolling material is calculated under the following premise condition: final target thickness: 6.0 mm final pass exit side crown quantity: 0.02 mm 'plate width: 3500 mm finish temperature of final pass: 7500C (finish in rear face direction) descaling execution pass: 1st and 3rd passes from the initial pass maximum cross angle: 0.585 final stage cross angle limitation: 0.0000 The premise condition described above in the calculation by the process computer in the actual on-line is transferred as the rolling material data from the host business computer, or is given as patterned data in accordance with the operating conditions.

The reduction schedule and the schedule of the roll crossing angle are simultaneously determined sequentially from the downstream pass to the upstream pass by the following calculation. Here, an example of numerical values of the calculation process of the final one pass (calculation start pass) will be given.

[Assumption of entry side (bite side) temperature] The entry side bite assumption temperature is set to 7800C by assuming the temperature drop as 300C.

[Assumption of rolling schedule speed] The standard mill speed is set to 100 rpm from the plate width of the rolling material and the exit side plate rolling thickness.

[Upper and lower limit range of allowable sharpness and target values] It is assumed that #max = 0.4%, Ain = -0.4%, #sim = 0.

These are parameters representing the shape allowable range by each rolling material size, and are empirically determined from the table values in accordance with the operating conditions.

[Calculation of allowable stretch strain difference and target stretch difference] From the formula (31), #E = (#/2).# = = 0.004% #Emin = -0.004% again = 0 [Calculation of allowable range of entry side plate crown ratio and target value] From the formula (32), Cin/Hin = Cout/Hout - #E/# + α a = 0.15 i = 0.61 Cout/Hout = 0.33% (Cin/Hin)max = 0.49% min = 0.48% sim = 0.48% [Calculation of allowable range of mechanical crown limited by shape and target value] From the formula (33), MCK = 1/(1 - #) [Cout - #.Hout.(Cin/Hin)] n = 0.702 (hereditary coefficient) MCKmax = 0.00 mm min = 0.00 mm ail = 0.00 mm [Allowable range of mechanical crown from rolling load and setup allowable capacity] Pmax = = 6500 ton Pmin = = 2200 ton #max = 0.000 #min = 0.000 In other words, the final pass is provided with the limitation of "cross angle = 0 ", and while FB is set to 130 ton, from the formula (34), MCh = clP + c2.F1 + E + c3 MChmax = 0.72 mm MCh = +0.01 mm Here, the roll bending load is set to a fixed value on the premise of preset.

[Determination of target value MCsim of true mechanical crown allowable range] MC = 0.01 mm MCmin = O.01 mm MCsim = 0.01 mm [Retrieval and determination of reduction ratio r, roll crossing angle 2e and bending load F3] In the formula (35), that is, r = fr(MCsim, 2e, FB), when the maximum reduction ratio is searched so as to obtain high efficiency rolling, r = 0.1328 (r = seek) can be obtained at the maximum value e = e= = 0.0000 under the condition where F3 is fixed.

[Calculation of entry side plate thickness] from H/Hxt/(1 Hin = 6.84 mm [Estimation calculation of temperature drop on pass exit side] temperature drop = 150C plate temperature at the time of biting = 7650C [Calculation of rolling load and rolling torque] P = 3120 ton torque = 98 ton.m Both of them are within the range of the setup capacity.

[Estimation calculation of temperature drop on pass entry side] temperature drop = 210C sheet temperature on exit side of previous pass = 7860C In the manner described above, calculation of the temperature, the load and the crown for one pass is completed.

The pass schedule is sequentially determined by cumulatively effecting the calculation for each pass from the downstream passes to the upstream passes, and this calculation is completed when the thickness on the pass entry side finally exceeds the scheduled thickness at the time of the start of rolling. In this example, the pass schedules for all the passes is calculated by setting the scheduled thickness at the time of start of rolling to 45 mm. The result is shown in Table 2 and Fig. 14 in comparison with the prior art method.

Table 2

This Invention Prior Art Method pass exit plate rolling cross exit plate rolling cross number side crown load angle side crown load angle thick- thickness ness 1 30.82 0.10(0.32) 4087 0.466 33.07 0.48(1.45) 3202 0.261 2 21.78 0.07(0.32) 4653 0.576 25.86 0.36(1.39) 3365 0.321 3 16.32 0.05(0.31) 4523 0.529 18.99 0.27(1.42) 3331 0.399 4 12.15 0.04(0.33) 4255 0.369 14.82 0.19(1.28) 3370 0.385 5 8.05 0.03(0.37) 3798 0.285 12.35 0.13(1.05) 3302 0.360 6 6.82 0.02(0.30) 3333 0.145 10.16 0.09(0.88) 3349 0.365 7 6.00 0.02(0.33) 3120 0.000 8.24 0.06(0.73) 3365 0.338 8 6.92 0.04(0.58) 3395 0.324 9 6.00 0.02(0.33) 3353 0.323 The shape regulation function by the roll cross function is the same between the present invention and the prior art. According to the pass schedule of the prior art, however, the rolling load and the setup capacity on the upstream side do not become maximal and eventually, the number of times of passes increases. In contrast, according to the present invention, since the schedule calculation is executed by seeking the allowable maximum value of the rolling load under the cross limit condition, the smallest number of times of passes can be accomplished while securing the shape even when the cross angle at the final stage is set to the minimum value.

Example 4 The pass schedule of the rolling material is calculated under the following premise condition. final target thickness: 20.0 mm plate crown on final pass exit side: 0.00 mm plate width: 3500 mm finish temperature of final pass: 8500C (finish in rear face direction) descaling execution pass: first pass from initial pass maximum cross angle: 0.600 final stage pass cross angle limit: 0.2000 maximum rolling load: 6000 ton scheduled thickness at start of rolling: 93 mm The pass schedules for all the passes in this example are calculated. The result is shown in Table 3.

Table 3

number exit side exit side rolling cross of pass thickness plate crown load angle 1 84.30 -0.05 5088 0.350 2 75.25 -0.04 5258 0.370 3 65.25 -0.04 5570 0.385 4 54.92 -0.04 5782 0.386 5 45.29 -0.04 5850 0.403 6 36.25 -0.03 6000 0.424 7 28.95 -0.02 5944 0.383 8 24.70 -0.01 5544 0.256 9 20.00 0.00 5328 0.175 ExamPle 5 The pass schedule of the rolling material is calculated under the following premise condition. final target thickness: 45.0 mm plate crown on final pass exit side: -0.20 mm plate width: 1,500 mm finish temperature of final pass: 8500C (finish in rear face direction) descaling execution pass: first, third and fifth passes from initial pass maximum cross angle: 0.5000 final stage pass cross angel limit: 0.0000 maximum rolling load: 4200 ton maximum torque: 420 ton scheduled thickness at start of rolling: 157 mm The pass schedules of all the passes of this example are calculated. The result is shown in Table 4.

Table 4

pass exit side plate crown rolling cross number thickness 1 load angle 1 134.58 0.08 3182 0.489 2 113.10 +0.05 3336 0.452 3 94.04 +0.04 3337 0.233 4 76.57 -0.16 3696 0.158 5 60.00 -0.18 3943 0.103 6 45.00 +0.20 4184 0.000 POSSIBILITY OF UTILIZING THE INVENTION IN THE INDUSTRY As is obvious from the examples given above, the load which is practically applied by the material to the rolling mill is fed back to the plate thickness and wedge control in the present invention, so that accuracy of the thickness of the rolled plate becomes high and the camber can be drastically restricted. Further, according to the present invention, rolling can be made with a constant crown ratio throughout the full passes, and the degree of flatness and the shape can be improved. Furthermore, the present invention makes it possible to simultaneously determine the shape of the rolling material, and the crown and load (reduction) schedules for each pass, to satisfy the shape throughout the entire passes and to determine the pass schedule facilitating rolling at the maximum value of the setup capacity.

Claims

1. In a rolling mill for effecting reverse rolling of a rolling material by relatively crossing roll sets each comprising a pair of backup rolls and work rolls inside a plane parallel to said rolling material, an automatic plate thickness control method in reverse rolling of a pair cross rolling mill characterized in that a true rolling actual record load and a load difference between right and left loads, free from the influences of a thrust load, are calculated by the use of detection values of upper and lower and right and left load cells for detecting a rolling load and a load cell for detecting a thrust load occurring in a roll axial direction, and the plate thickness and a wedge are automatically controlled on the basis of these detection values and the calculated values.

2. In a reverse rolling mill which relatively crosses roll sets each comprising a pair of backup rolls and work rolls inside a plane parallel to a rolling material, and which includes roll bending controllers at both end portions of a roll, a shape control method in reverse rolling of a pair cross rolling mill characterized in that an actual record crown is calculated on the real time basis in accordance with a rolling load fluctuation inside a plate in order to accomplish a plate crown determined in advance for making the final shape of the plate flat, roll bending control is executed, a plate crown calculation quantity is corrected by an actual record roll bending quantity of a previous pass at the end of a pass, and is reflected on a cross angle set value of next and subsequent passes, and the shape of the plate is thus controlled.

3. In a rolling operation by a reverse rolling mill which relatively crosses roll sets each comprising a pair of backup rolls and work rolls inside a plane parallel to a rolling material, and which includes roll bending controllers at both end portions of a roll, a shape control method in reverse rolling of a pair cross rolling mill characterized in that, when a roll bending quantity is corrected in accordance with an exit side shape of each pass, a shape estimation value of said pass is corrected from a roll bending actual record value of said pass, a plate crown estimation value is further corrected, a mechanical crown quantity as a target value in next and subsequent passes is calculated again, and is reflected on a cross angle set value, and the shape of the plate is thus controlled.

4. When a pass schedule is determined at the time of rolling of a plate material by a reverse rolling mill including a controller for relatively crossing an upper roll set and a lower roll set each comprising a pair of backup rolls and work rolls inside a plane parallel to a rolling material, a reverse rolling schedule determination method of a pair cross rolling mill characterized in that a region which simultaneously satisfies a mechanical crown allowable range judged from a shape in each pass and a mechanical crown allowable range judged from a setup capacity is calculated, a plate thickness schedule is determined by sequentially building up the values so as to attain an allowable maximum rolling load among them, and reduction of the shortest pass number satisfying the shape and a cross angle schedule are simultaneously determined.