JP6601104B2 - Parameter setting device, roll position change amount setting device, roll position change method, and program - Google Patents

Description

  The present invention relates to a parameter setting device, a roll position change amount setting device, a roll position change method, and a program, and is particularly suitable for use in setting a roll gap in a universal rolling mill.

  FIG. 12 is a diagram showing an example of rolling a material to be rolled by a universal rolling mill. Specifically, FIG. 12 (a) is a view of the universal rolling mill viewed from above, FIG. 12 (b) is a view of the universal rolling mill viewed from the side, and FIG. FIG. 3 is a view of a universal rolling mill as viewed from the front (a view taken along a conveyance direction of a material to be rolled). In each figure, the x, y, and z coordinates indicate the orientation of each figure. Here, a case where an H-section steel is manufactured will be described as an example.

  12 (a) and 12 (b), the universal rolling mill used when manufacturing the shape steel has four rolls of a set of horizontal rolls 1201a and 1201b and a set of trough rolls 1202a and 1202b. Have. FIG. 13 is a diagram illustrating the rotation axes of the horizontal rolls 1201a and 1201b and the saddle rolls 1202a and 1202b. As shown in FIG. 13, the horizontal rolls 1201 a and 1201 b are positioned on the upper side and the lower side of the material to be rolled S, respectively, with the rotation shafts 1203 a and 1203 b facing the horizontal direction with respect to the ground. When there is no material to be rolled S, the outer peripheral surfaces of the horizontal rolls 1201a and 1201b can contact each other. On the other hand, the scissors rolls 1202a and 1202b are positioned on the right side and the left side of the material S to be rolled, respectively, with the rotation shafts 1204a and 1204b facing the direction perpendicular to the ground. In the absence of the material to be rolled S, the outer peripheral surfaces of the rolls 1202a and 1202b can contact the side surfaces of the horizontal rolls 1201a and 1201b.

  FIG. 14 is a diagram illustrating the roll gap. The gap between the horizontal rolls 1201a and 1201b during rolling or the distance between the gaps (the sum of distances A and B) is called the roll gap of the horizontal roll. Further, the gap between the rolls 1202a and 1202b during rolling or the distance between the gaps (the sum of the distances C and D) is referred to as the roll gap of the roll. In the following description, the term “roll gap” refers to the distance between the horizontal rolls 1201a, 1201b and the saddle rolls 1202a and 1202b, not the gap between the horizontal rolls 1201a, 1201b and the saddle rolls 1202a and 1202b. To do.

  A method for measuring correct values of the roll gaps of the horizontal rolls 1201a and 1201b and the roll gaps of the saddle rolls 1202a and 1202b has not been put into practical use. Therefore, the positions at which the roll gap is assumed to be zero are designated as zero points 1206a, 1206b, 1206c, the sum of the relative changes in the positions of the horizontal rolls 1201a, 1201b from the zero point 1206a, and the saddle roll from the zero point 1206b The sum of the relative change in the position of 1202a and the relative change in the position of the saddle roll 1202b from the zero point 1206c is regarded as a roll gap. In the following description, this roll gap is referred to as “relative roll gap” as necessary. On the other hand, the actual roll gap in the horizontal rolls 1201a and 1201b and the actual roll gap in the saddle rolls 1202a and 1202b are referred to as “absolute roll gap” as necessary.

  When the roll is changed, it is necessary to adjust the zeros 1206a, 1206b, and 1206c. This adjustment of the zero points 1206a, 1206b, 1206c is called zero point adjustment. When performing the zero point adjustment, the rolling load and the number of rotations of the horizontal rolls 1201a and 1201b / the rolls 1202a and 1202b are made constant. For example, if the horizontal rolls 1201a, 1201b, the saddle rolls 1202a, 1202b are both 300 [rpm] and the rolling load is 100 [ton], the roll gap is assumed to be zero. At this time, the relative roll gaps of the horizontal rolls 1201a and 1201b and the saddle rolls 1202a and 1202b are each 0 (zero). For example, if the saddle rolls 1202a and 1202b are separated from each other by 0.5 [mm] from this state, the distances C and D described above become 0.5 [mm], respectively. Therefore, the relative roll gap between the saddle rolls 1202a and 1202b is 1 [mm]. The zero point adjustment may be performed even when the roll is not changed. In the present specification, a period from when the zero adjustment is performed until the next zero adjustment is performed is referred to as a roll chance.

As described above, when rolling by the universal rolling mill, unless the roll gap setting value (the set value of the relative roll gap) is appropriately determined, the dimensional accuracy of the product may be lowered.
Therefore, Patent Document 1 discloses that a roll gap setting value and a product dimension are used as rolling results, and a rough gap theory is used to obtain a roll gap setting value that achieves a target product dimension. Has been.
Also, in Patent Document 2, a set of roll gap setting values and product dimensions is collected as rolling results, and the contribution ratio of each principal component of the profile component matrix obtained by performing principal component analysis of the roll gap setting value matrix It is disclosed to minimize the sum of squares of the correction amount of the roll gap setting value weighted by the reciprocal of.

JP 2002-282913 A Japanese Patent No. 4203223

  If the roll gap setting value is set by the techniques described in Patent Documents 1 and 2, the accuracy of the product dimensions may not be good. The present inventors examined the reason. As a result, the following findings were obtained.

  In the universal rolling mill, in addition to the horizontal rolls 1201a and 1201b, the zero point adjustment is performed with the side rolls 1202a and 1202b simultaneously in contact with each other from the side. For this reason, even if the rotation axis and contact surface of the roll are slightly shifted, the pressure balance of the horizontal rolls 1201a, 1201b and the saddle rolls 1202a, 1202b is lost, and the rate of change in shape due to the load and speed of the rolls changes. Since it is difficult to suppress such balance loss, it is difficult to make the difference between the relative roll gap and the absolute roll gap zero even if zero adjustment is performed. In the following description, the difference between the relative roll gap and the absolute roll gap generated by performing the zero point adjustment is referred to as “zero tone error for each roll chance” as necessary.

  In addition, it is rare that the zero adjustment error per roll chance generated by the zero adjustment performed at a certain roll chance and the zero adjustment error per roll chance generated by the zero adjustment performed at another roll chance are the same. . In the following description, the difference between the zero-tone errors at each roll chance is referred to as “zero-tone error between roll chances” as necessary. In principle, it is not easy to suppress the zero-tone error between roll chances.

FIG. 15 is a diagram for explaining a zero-tone error between roll chances of the horizontal rolls 1201a and 1201b.
The state on the left side of FIG. 15 shows a state in which zero adjustment is performed on the horizontal rolls 1201a and 1201b in a certain roll chance, and the state on the right side of FIG. 15 shows the horizontal roll in a roll chance different from the roll chance. A state in which zero point adjustment is performed on 1201a and 1201b is shown. In any state, the rolling control is performed on the assumption that the relative roll gap between the horizontal rolls 1201a and 1201b is 0 (zero). The zero roll error for each roll chance of the horizontal rolls 1201a and 1201b in the left state of FIG. 15 is 0.1 [mm], and the zero roll error for each roll chance of the horizontal rolls 1201a and 1201b in the right state of FIG. In the case of 0.2 [mm], the zero adjustment error E between roll chances of the horizontal rolls 1201a and 1201b when the state on the left side of FIG. 15 is used is −0.3 [mm].

  Since the zero-tone error at each roll chance and the zero-tone error between roll chances have a range close to the adjustment value when adjusting the roll gap, the disturbance increases when the rolling record across the roll chance is used. On the other hand, the number is not enough in the rolling performance which does not straddle the roll chance. Therefore, the techniques described in Patent Documents 1 and 2 cannot use a sufficient number of rolling records. Further, Patent Documents 1 and 2 do not consider the effects of the zero tone error for each roll chance and the zero tone error between roll chances, so the effects of the zero tone error for each roll chance and the zero tone error between roll chances cannot be removed. The roll gap setting value will be calculated. From the above, it is not easy to calculate the roll gap setting value with which the accuracy of product dimensions is good with the techniques described in Patent Documents 1 and 2. For this reason, the actual situation is that the operator (expert) sets the roll gap setting value based on his / her own experience (past knowledge). It is not easy to obtain a roll gap setting value with good accuracy.

  The present invention has been made in view of the above-described problems, and an object of the present invention is to enable automatic setting of a roll gap setting value at which product dimension accuracy is good for a universal rolling mill.

The parameter setting device of the present invention is a parameter setting device for setting a parameter for changing the position of a roll in a universal rolling mill, the target dimensions of a product rolled and manufactured by the universal rolling mill, and the product Is the set value of the relative roll gap, which is the distance between the two rolls at positions facing each other, the type of the product, the type of the product, and the distance from a predetermined zero point. 2 different from data acquisition means for acquiring operation schedule / actual data including, as data items, a roll gap set value and a roll chance that is a period from the adjustment of the zero point to the adjustment of the zero point. A combination of two operation schedules / results data extracted one by one from the operation schedule / results data in one roll opportunity, A pair group creating means for creating a regular pair that is a combination of two operation schedule / actual data having the same target size of the product and the type of the product, and an actual value and a target value of the product dimension. The coefficient in a function in which the change amount of the relative roll gap is determined according to the dimensional change amount that is a difference and a coefficient multiplied by the dimensional change amount is derived as the parameter by using an algorithm of an optimization problem. Parameter derivation means, wherein the parameter derivation means derives the dimensional change amount from each of the two operation schedule / actual data constituting the regular pair, and the dimensional change Using the dimensional change amount derived by the quantity deriving means and the function, the target dimension in the two operation schedule / actual data constituting the same regular pair A coefficient deriving means for deriving said coefficient variation of the relative zero adjustment error which is the difference relative roll gap to achieve is to be the smallest in the algorithm of the optimization problem as the parameter, and further comprising a To do.

  The roll position change amount setting device of the present invention is a roll position change amount setting device that sets a change amount of the roll position using the function in which the coefficient derived as the parameter is set by the parameter setting device. The dimensional deviation deriving means for deriving the difference between the actual dimension of the product different from the product in the operation schedule / actual data and the target dimension, and the actual dimension and the target of the product derived by the dimensional deviation deriving means. The roll position change amount for deriving the change amount of the relative roll gap by substituting the difference derived from the dimension as the dimensional change amount into the function in which the coefficient derived as the parameter by the parameter setting device is set. And a deriving unit.

The roll position changing method of the present invention is a roll position changing method for changing the position of a roll in a universal rolling mill, the target dimension of a product produced by rolling with the universal rolling mill, and the actual dimension of the product, The roll gap setting value, which is the distance between the two types of rolls at the positions facing each other and the type of the product, which is a relative roll gap setting value based on a predetermined zero point. And a data acquisition step of acquiring operation schedule / actual data including a roll opportunity as a data item after the zero point is adjusted until the zero point is adjusted next, and in two different roll chances A combination of the two operation schedule / result data extracted one by one from the operation schedule / result data, the target size of the product, and the product A pair group creation step of creating a normal pair that is a combination of two operation schedules / actual data of the same type, a dimensional change amount that is a difference between the actual value and the target value of the product dimension, and A parameter deriving step for deriving the coefficient in the function in which the change amount of the relative roll gap is determined according to a coefficient multiplied by the dimensional change amount using an algorithm of an optimization problem, and a product in the operation schedule / actual data The dimensional deviation derivation step for deriving the difference between the actual dimension of the product different from the target dimension and the target dimension, and the difference between the actual dimension of the product and the target dimension derived by the dimensional deviation derivation step as the dimensional change amount, A roll position change amount derivation step for deriving a change amount of the relative roll gap by substituting the coefficient derived in the parameter derivation step into the set function. A roll position change step derived from the roll position change amount deriving step, and changing the position of the roll based on the change amount of the relative roll gap. Using the dimensional change amount deriving step for deriving the dimensional change amount from the two operation schedule / actual data constituting the same, using the dimensional change amount and the function derived by the dimensional change amount deriving step, respectively. The variation of the relative zero-tone error, which is the difference in the relative roll gap that realizes the target dimension in the two operation schedule / actual data constituting the normal pair, is supposed to be minimized in the optimization problem algorithm. A coefficient deriving step for deriving the coefficient.

  The program of the present invention causes a computer to function as each means of the parameter setting device.

  ADVANTAGE OF THE INVENTION According to this invention, the roll clearance setting value from which the precision of a product dimension becomes favorable can be set automatically with respect to a universal rolling mill.

It is a figure which shows an example of a structure of a rolling control system. It is a figure which shows an example of the content of operation schedule and performance data in a table format. It is a figure which shows an example of H-section steel. It is a figure explaining an example of a regular pair. It is a figure which shows an example of a regular pair in a table format. It is a figure which shows web thickness variation | change_quantity and flange thickness variation | change_quantity in a tabular form. It is a figure explaining an example of the method of deriving | releasing the relative roll gap | interval of a horizontal roll and a reed roll which implement | achieve a target dimension. It is a figure explaining an example of the derivation method of the zero tone error between roll chances of a horizontal roll and a reed roll. It is a flowchart explaining an example of operation | movement of a parameter setting apparatus. It is a flowchart explaining an example of operation | movement of a roll position change amount setting apparatus. It is a figure which shows the relationship between the roll gap setting value of a horizontal roll, the roll gap setting value of a roll, and the actual dimension of web thickness and flange thickness, and the rolling frequency. It is a figure which shows an example of a mode that a to-be-rolled material is rolled with a universal rolling mill. It is a figure which shows the rotating shaft of a horizontal roll and a reed roll. It is a figure explaining a roll gap. It is a figure explaining the zero tone error between roll chances of a horizontal roll.

Hereinafter, an embodiment of the present invention will be described with reference to the drawings. In this embodiment, a case where a product manufactured using a universal rolling mill is an H-section steel will be described as an example.
FIG. 1 is a diagram illustrating an example of a configuration of a rolling control system.
In FIG. 1, the rolling control system includes a parameter setting device 100, a roll position change amount setting device 110, a roll gap control device 120, and a universal rolling mill 130.

(Parameter setting device 100)
The parameter setting device 100 includes a data acquisition unit 101, a pair group creation unit 102, a parameter derivation unit 103, and a parameter output unit 104, and operates offline. The hardware of the parameter setting device 100 is realized by using, for example, an information processing device provided with a CPU, ROM, RAM, HDD, and various interfaces, or dedicated hardware.

<Data acquisition unit 101>
The data acquisition unit 101 acquires operation schedule / result data.
FIG. 2 is a diagram showing an example of the contents of the operation schedule / result data in a table format.
In FIG. 2, the operation schedule / result data includes a roll opportunity, manufacturing setting information, manufacturing result information, and operation setting information as data items. A set of a roll chance, manufacturing setting information, manufacturing performance information, and operation setting information becomes one operation schedule / result data. FIG. 2 shows six operation schedule / result data.

As described above, the roll chance is a period from when the zero adjustment is performed until before the next zero adjustment is performed.
The production setting information includes a target dimension and a steel type. The steel type is an example of the type of product. The type of product is determined based on the chemical composition of the product (rolled material), for example.
A target dimension is a dimension requested | required as a product. In the present embodiment, a case where the web thickness and the flange thickness are product dimensions will be described as an example.
FIG. 3 is a diagram illustrating an example of an H-section steel. FIG. 3 shows a cross section of the H-section steel cut in a direction perpendicular to the axial direction (longitudinal direction).

  The H-shaped steel is a steel material having a shape that is generally H-shaped when cut in a direction perpendicular to the axial direction, and has a shape that is long in the axial direction. The H-shaped steel has a web 310 and flanges 320a and 320b. The web 310 is a portion formed when the material to be rolled passes through a region between the horizontal rolls 1201a and 1201b, and the thickness thereof is referred to as a web thickness WT. The flanges 320a and 320b are portions formed when the material to be rolled passes through a region between the horizontal rolls 1201a and 1201b and the rolls 1202a and 1202b, and the thickness thereof is referred to as a flange thickness FT. The length of the flanges 320a and 320b is referred to as the flange width, and the length of the web 310 including the flange thickness is referred to as the web height.

Manufacturing performance information is the actual size of the product. In this embodiment, since the web thickness and the flange thickness are product dimensions, the actual value of the web thickness and the actual value of the flange thickness are the manufacturing result information.
The operation setting information is a roll gap setting value. The roll gap setting value includes a horizontal gap roll gap setting value and a heel roll roll gap setting value. In FIG. 2, the roll gap setting value of the horizontal roll is expressed as a horizontal roll gap, and the roll gap setting value of the heel roll is expressed as a heel roll gap. As described above, the roll gap setting value of the horizontal roll is the set value of the sum of the relative change amounts of the positions of the horizontal rolls 1201a and 1201b from the zero point 1206a (the sum of the distances A and B in FIG. 14A). (Target value). The roll gap setting value of the heel roll is the sum of the relative change amount of the position of the heel roll 1202a from the zero point 1206b and the relative change amount of the position of the heel roll 1202b from the zero point 1206c (of FIG. 14B). This is a set value (target value) of the sum of the distances C and D.

  In FIG. 2, for example, the operation schedule / result data of the first line (No. = 1) is the H-section steel of the steel type P in the roll chance A, the target dimension of the web thickness is 16 [mm], and the flange thickness In order to manufacture an H-section steel having a target dimension of 24 mm, roll roll setting values of the scissors rolls 1202a and 1202b and roll roll setting values of the horizontal rolls 1201a and 1201b are 15.5 [mm] and 23, respectively. When rolled to 0.5 mm, an H-section steel having a web thickness of 16.1 [mm] and a flange thickness of 23.8 [mm] is produced.

  The data acquisition unit 101 acquires and stores a plurality of operation schedule / result data for each of a plurality of different roll opportunities. For example, at least one of operation of the user interface of the parameter setting device 100, communication with an external device, and reading from an external storage medium may be adopted as a form of acquiring a plurality of operation schedule / result data. it can.

<Pair group creation unit 102>
The pair group creation unit 102 extracts operation schedule / result data for two roll opportunities from the operation schedule / result data acquired by the data acquisition unit 101. For example, the pair group creation unit 102 extracts all the operation schedule / result data for the roll chances A and B.

  Next, the pair group creation unit 102 has a combination of (one) operation schedule / result data in one roll opportunity and two (1) operation schedule / result data in the other roll chance. Extract all possible combinations. In the example shown in FIG. 2, for example, one roll chance is a roll chance A, and the other roll chance is a roll chance B.

  In this case, the pair group creation unit 102 selects No. No. 1 operation schedule / result data and No. 1 No. 4 operation schedule / result data combination, No. 1 operation schedule / result data and No. 1 No. 5 operation schedule / result data combination, No. 1 operation schedule / result data and No. 1 No. 6 operation schedule / result data combination, No. 2 operation schedule / result data and No. 2 No. 4 operation schedule / result data combination, No. 2 operation schedule / result data and No. 2 No. 5 operation schedule / result data combination, No. 2 operation schedule / result data and No. 2 No. 6 operation schedule / result data combination, No. 3 operation schedule / result data and No. 3 No. 4 operation schedule / result data combination, No. 3 operation schedule / result data and No. 3 No. 5 operation schedule / result data combination, No. 3 operation schedule / result data and No. 3 6 combinations of operation schedules and performance data are extracted.

  If the operation schedule / actual data in roll chances A and B are other than those shown in FIG. 2, the pair group creation unit 102 includes the operation schedule / actual data in the roll chance A (one ) All possible combinations of the operation schedule / result data and (one) operation schedule / result data in the roll chance B are extracted.

Next, the pair group creation unit 102 extracts, as normal pairs, combinations having the same target dimension and steel type from among the combinations extracted as described above.
FIG. 4 is a diagram illustrating an example of a regular pair.
In the example shown in FIG. No. 1 operation schedule / result data target dimensions and steel types, The target dimensions and steel types of the operation schedule / result data of 4 are the same. No. No. 3 operation schedule / result data target dimensions and steel type, The target dimensions and steel types of the operation schedule / result data of 5 are the same. Therefore, these combinations are regular pairs.
FIG. 5 is a diagram showing the normal pairs X and Y in a table format. 5A shows the normal pair X, and FIG. 5B shows the normal pair Y. The normal pair X is No. No. 1 operation schedule / result data and No. 1 No. 4 operation schedule / result data combination, and the regular pair Y is No. No. 3 operation schedule / result data and No. 3 This is a combination of 5 operational schedules and performance data.

  On the other hand, for example, no. No. 1 operation schedule / result data and No. 1 The combination of operation schedule and performance data of 5 is not a regular pair because the target dimensions are different. No. No. 2 operation schedule / result data and No. 2 The combination of operation schedule and performance data of 4 is not a regular pair because the steel grades are different.

<Parameter derivation unit 103>
The parameter deriving unit 103 subtracts the target dimension from the actual dimension for each of the web thickness and the flange thickness in the operation schedule / actual data constituting the regular pair. In the following description, in the operation schedule / actual data constituting the regular pair, the result obtained by subtracting the target dimension from the actual dimension of the web thickness is referred to as “web thickness change d _w ” as necessary. In the operation schedule / actual data constituting the regular pair, the value obtained by subtracting the target dimension from the actual dimension of the flange thickness is referred to as “flange thickness change amount d _f ” as necessary.

Figure 6 is a normal pair X shown in FIG. 5, a web thickness variation d _w and a flange thickness variation d _f in operational planning and performance data constituting Y is a diagram illustrating in tabular form. For example, No. which becomes a regular pair X. The web thickness change amount d _w in the operation schedule / result data of No. 1 is No. 1 operation schedule, actual size of actual data, web thickness (= 16.1 [mm]), No. 1 No. 1 target schedule / actual dimension column of No. 1 in FIG. 5 (a) and the target dimension / web thickness (= 16 [mm]) of 1 operation schedule / actual data subtracted. (See the column of web thickness change No. 1 in FIG. 6 (a)). In addition, No. which becomes the regular pair X. Flange Thickness variation d _f in the first operation plan, actual data, No. No. 1 operation schedule / actual data from actual data / flange thickness (= 23.8 [mm]). No. 1 target dimension / actual dimension column of No. 1 in FIG. 5A is obtained by subtracting the target dimension / flange thickness (= 24 [mm]) of the operation schedule / actual data of No. 1 And the column of No. 1 flange thickness change amount in FIG.
Parameter deriving section 103 for all operations planned and results data constituting a regular pair, as described above derives a web thickness variation d _w and a flange thickness variation d _f.

Next, the parameter deriving unit 103 derives a coefficient f _{i, j} shown in the following expression (1) and a coefficient g _{i, j} shown in the expression (2). The coefficients f _{i, j} and g _{i, j} are real numbers.

(1) In the formula and (2), n is the order to a web thickness variation d _w and a flange thickness variation d _f, is selected from among the natural numbers. d _H is the amount of change in the relative roll gap of the horizontal roll, and d _V is the amount of change in the relative roll gap of the saddle roll.
As shown in the equations (1) and (2), the functions F (d _w , d _f ) and G (dw, d _f ) have the web thickness variation d _w and the flange thickness variation d _f as variables. It is an n-dimensional line format. For example, 3 can be adopted as n.

(1), the equation (2), the coefficients f _{i, j,} g _i, if the value of _j is the correct value, from the web thickness variation d _w and a flange thickness variation d _f, the target product dimensions This means that it is possible to derive the change amount d _H of the relative roll gap of the horizontal roll and the change amount d _V of the relative roll gap of the saddle roll for realizing. In other words, to obtain actual dimensions of the web thickness and flange thickness, web thickness variation d _w and a flange thickness variation d _f at that time determined, (1) and (2), the relative roll gap of the horizontal roll of derives the change amount d _H and vertical change amount d _V of the relative roll gap of the roll, the roll gap set value of the current horizontal roll vertical rolls, the change amount of the relative roll gap of the horizontal roll vertical rolls d _H, by adding the d _V, it is possible to derive the roll gap set value for realizing the target product dimensions. The initial value of the roll gap setting value is set based on past knowledge, for example, without using the equations (1) and (2). The initial value of the roll gap setting value is reset every time at least one of the steel type of the product to be manufactured and the size range of the target dimension is changed. In this embodiment, the dimension range is determined from the range of the web thickness WT and the range of the flange thickness FT.

In the following description, the functions F (d _w , d _f ) and G (d _w , d _f ) in the expressions (1) and (2) are referred to as “roll gap effect functions F and G” as necessary.
In the present embodiment, a case where the coefficients f _{i, j} and g _{i, j} of the roll gap effect functions F and G are derived using genetic algorithms (GA) will be described as an example. Hereinafter, an example of a method for deriving the coefficients f _{i, j} and g _{i, j} will be described.

The parameter deriving unit 103 creates a predetermined number of candidates as candidates for the coefficients f _{i, j} and g _{i, j} using random numbers. Here, 100 candidates for the coefficients f _{i, j} and g _{i, j} are created.
Next, the parameter derivation unit 103, a web thickness variation d _w and a flange thickness variation d _f in operational planning and performance data constituting a regular pair, coefficients f _{i, j,} g _i, and a candidate of the _j ( Substituting into the formulas (1) and (2), the relative roll gap of the horizontal roll and the relative roll gap of the vertical roll to achieve the target dimensions (that is, the roll gap set value of the horizontal roll and the roll gap set value of the vertical roll) Is derived.

  FIG. 7 is a diagram for explaining an example of a method for deriving the relative roll clearance of the horizontal roll and the relative roll clearance of the eaves roll that realizes the target dimension. In FIG. 7, the relative roll clearance of the horizontal roll that achieves the target dimension and the relative roll clearance of the heel roll that achieves the target dimension are expressed as the horizontal roll clearance that realizes the target dimension, and the heel roll clearance that realizes the target dimension, respectively. is doing.

For example, no. Web thickness variation d _w in the first operation plan and results data, the flange thickness variation d _f, as shown in FIG. 6 (a), respectively, 0.1 [mm], - 0.2 [mm] It is. No. As shown in FIGS. 5A, 6A, and 7A, the roll gap setting value of the saddle roll in the operation schedule / result data of 1 is 15.5 [mm]. Therefore, no. In the operation schedule / result data of 1, the relative roll gap (roll roll setting value of the roll) that achieves the target dimension is 15.5 + F (+0.1, −0.2).

No. The roll gap setting value of the horizontal roll in the operation schedule / result data of 1 is 23.5 [mm] as shown in FIGS. 5 (a), 6 (a), and 7 (a). Therefore, no. In the operation schedule / result data of 1, the relative roll gap (horizontal roll gap setting value) of the horizontal roll realizing the target dimension is 23.5 + F (+0.1, −0.2).
As described above, the parameter deriving unit 103 achieves the target dimensions when the candidate coefficients f _{i, j} and g _{i, j} are used for each of the operation schedule / actual data constituting the normal pair. The relative roll gap of the horizontal roll to be achieved and the relative roll gap of the saddle roll to achieve the target dimension are derived. That is, with respect to one coefficient f _{i, j} , g _{i, j} candidate, the relative roll gap of the horizontal roll that realizes the target dimension and the relative roll gap of the eaves roll that realizes the target dimension will constitute a normal pair.・ Derived by the number of actual data.

Next, the parameter deriving unit 103 extracts two operation schedule / actual data constituting the regular pair one by one. The parameter deriving unit 103 derives the extracted two operation schedule / actual data using the same coefficients f _{i, j} and g _{i, j} candidates, and the relative roll gap of the horizontal roll that realizes the target dimension. And the zero tone error between roll roll and horizontal roll chances is derived from the relative roll gap between roll and horizontal roll to achieve the target dimension. The parameter deriving unit 103 derives the zero tone error between roll rolls / horizontal rolls for each of the two operation schedules / actual data constituting the normal pair.

  FIG. 8 is a diagram for explaining an example of a method for deriving a zero-tone error between roll chances of the horizontal roll and the saddle roll. FIG. 8 shows the zero-tone error between roll chances in the roll chance B when the zero-tone error for each roll chance occurring at the roll chance A is used as a reference. In FIG. "Error". In the present embodiment, it is assumed that the zero tone error between the roll chances of the horizontal roll and the saddle roll does not change within the roll chance.

  As shown in FIG. 8 (a), the zero-tone error between roll chances of the saddle roll in the normal pair X is the relative roll gap (= 16.0 + F (+0.3, +0.1)), a value obtained by subtracting the relative roll gap (15.5 + F (+0.1, −0.2)) of the saddle roll that achieves the target dimension at roll chance A (= 0.5 + F (+0.3) , +0.1) −F (+0.1, −0.2)).

  Further, the zero-tone error between the roll chances of the horizontal roll in the normal pair X is calculated from the relative roll gap (= 23.8 + F (+0.3, +0.1)) of the horizontal roll that achieves the target dimension in the roll chance B. A value obtained by subtracting the relative roll gap (23.5 + F (+0.1, −0.2)) of the horizontal roll that achieves the target dimension at chance A (= 0.3 + F (+0.3, +0.1) −F ( +0.1, -0.2)).

As described above, the parameter deriving unit 103 derives the zero-tone error between roll chances of the saddle roll and the zero-tone error between roll chances of the horizontal roll for each of the normal pair candidates. As described above, for one candidate f _{i, j} and g _{i, j} , the relative roll gap of the horizontal roll that realizes the target dimension and the relative roll gap of the eaves roll that realizes the target dimension constitute a normal pair. It is derived only for the number of operation schedules and performance data to be performed. Accordingly, with respect to one candidate of the coefficients f _{i, j} and g _{i, j} , the zero-tone error between roll chances of the saddle roll and the zero-tone error between roll chances of the horizontal roll are derived by the number of regular pairs.

Next, the parameter deriving unit 103 derives the variances σ _V and σ _H of the zero tone error between roll roll / horizontal roll chances derived from the candidates for the coefficients f _{i, j} and g _{i, j} . The parameter deriving unit 103 derives the variances σ _V and σ _H of the zero-tone error between roll roll / horizontal roll chances for each of the candidate sets of coefficients f _{i, j} and g _{i, j} . As described above, for each of the candidates for coefficients f _{i, j} and g _{i, j} , the variances σ _V and σ _{H of the} zero-tone error between roll roll and horizontal roll chances are derived (one by one). .

Next, the parameter deriving unit 103 calculates, for each candidate of the coefficients f _{i, j} and g _{i, j} , the average value of the variances σ _V and σ _H of the zero tone error between roll rolls / horizontal rolls (= (σ _V + σ _H ) / 2) is derived as a score s (fitness).
Next, the parameter deriving unit 103 determines whether or not there is a score s that is lower than a preset reference value. As a result of this determination, when there is a score s that is lower than a preset reference value, the parameter deriving unit 103 selects candidates for coefficients f _{i, j} and g _{i, j} corresponding to the minimum score s among the scores s. _Are determined as the coefficients f _{i, j} and g _{i, j} in the equations (1) and (2). As described above, in the present embodiment, the coefficients f _{i, j} and g _{i, j} that reduce the variation in the zero-tone error between the roll chances of the saddle roll and the horizontal roll in each normal pair are searched.

On the other hand, when there is no score s below the preset reference value, the parameter deriving unit 103 determines whether the number of repetitions of the process of calculating the score s and comparing it with the preset reference value exceeds a predetermined value. judge. As a result of this determination, when the number of repetitions exceeds a predetermined value, the parameter deriving unit 103 determines the smallest score s of the scores s derived in the current process and the smallest score s derived in the previous process. coefficient f _i corresponding to the smaller score s of the score _{s, j,} g _i, candidates _j, (1) formula, is determined as a coefficient f _{i, j,} g _{i, j} in equation (2) .

On the other hand, if the number of repetitions does not exceed the predetermined value, the parameter deriving unit 103 rearranges the candidates for the coefficients f _{i, j} and g _{i, j} in ascending order of the score s derived in the current process, The candidates for the new coefficients f _{i, j} and g _{i, j} are obtained by performing crossover, mutation, copying, and the like on the rearranged coefficients f _{i, j} and g _{i, j} according to the genetic algorithm. Create 100. The parameter deriving unit 103 performs the above-described processing using the new coefficients f _{i, j} and g _{i, j} candidates. The parameter deriving unit 103 repeats the above processing until the coefficients f _{i, j} and g _{i, j} in the equations (1) and (2) are determined.
The parameter setting device 100 determines the coefficients f _{i, j} and g _{i, j} as described above for each preset steel type and each size range. In this embodiment, the dimension range is determined from the range of the web thickness WT and the range of the flange thickness FT.

<Parameter output unit 104>
The parameter output unit 104 outputs information indicating the coefficients f _{i, j} and g _{i, j} determined by the parameter deriving unit 103. As a form of output of information indicating the coefficients f _{i, j} and g _{i, j} , for example, among transmission to the roll position change amount setting device 110, storage in a portable storage medium, and display on a computer display, At least one of them can be adopted. The parameter output unit 104 can store the coefficients f _{i, j} and g _{i, j} determined by the parameter deriving unit 103 in a storage medium inside the parameter setting device 100.

(Roll position change amount setting device 110)
In FIG. 1, a roll position change amount setting device 110 includes a parameter acquisition unit 111, a dimensional deviation derivation unit 112, a roll position change amount derivation unit 113, a roll gap set value derivation unit 114, and a roll gap set value output unit. 115 and operates online.

<Parameter acquisition unit 111>
The parameter acquisition unit 111 acquires and stores information on the coefficients f _{i, j} and g _{i, j} output from the parameter output unit 104. Acquiring the form of information indicating coefficients f _{i, j,} g _i, a _j is the coefficient f _{i, j,} g _i, determined in accordance with the form of output information _j. As described above, the coefficients f _{i, j} and g _{i, j} are determined and output for each steel type and each dimension range. Therefore, the parameter acquisition unit 111 acquires and stores the coefficients f _{i, j} and g _{i, j} for each steel type and each size range. For example, the parameter acquisition unit 111 creates a table that stores a steel type, a size range, and coefficients f _{i, j} and g _{i, j} in association with each other.

<Dimension deviation deriving unit 112>
When the rolling by the universal rolling mill 130 is finished and the H-section steel is manufactured, the dimension deviation deriving unit 112 obtains the actual dimension h _wr of the web thickness WT and the actual dimension h _fr of the flange thickness FT of the H-section steel. The actual dimension h _wr of the web thickness WT and the actual dimension h _fr of the flange thickness FT are obtained online (before the next rolling of the rolling is performed). Further, the dimensional deviation deriving unit 112 obtains the target dimension h _wt of the web thickness WT and the target dimension h _ft of the flange thickness FT in the H-section steel online. As these acquisition forms, for example, at least one of operation of a user interface of the roll position change amount setting device 110, communication with an external device, and reading from an external storage medium can be employed.

Next, the dimension deviation deriving unit 112 subtracts the target dimension h _wt from the actual dimension h _{wr of} the web thickness WT to derive the dimension deviation d _w of the web thickness WT. Furthermore, dimensional deviations deriving unit 112 derives the dimensional deviations d _f of the flange thickness FT from actual dimensions h _fr flange thickness FT by subtracting the target dimension h _ft. Incidentally, web thickness WT of the dimensional deviations d _w, dimensional deviations d _f of the flange thickness FT, respectively, (1) and (2) which corresponds to the web thickness variation d _w, a flange thickness variation d _f in formula It is.

Next, dimensional deviations deriving unit 112, at least one of dimensional deviations d _f of dimensional deviations d _w and a flange thickness FT of web thickness WT determines whether exceeds a predetermined value. Note that the predetermined value includes a value for the dimensional deviations d _f values and the flange thickness FT for dimensional deviations of the web thickness WT. Thus, dimensional deviations deriving unit 112 of whether dimensional deviations d _w of the web thickness WT exceeds a predetermined value for the dimensional deviations d _w of the web thickness WT, flange thickness FT dimensional deviations dimensional deviations d _f is the flange thickness FT of and whether exceeds a predetermined value for d _f, it will be determined.

As a result of the determination, when at least one of the dimensional deviations d _f of dimensional deviations d _w and a flange thickness FT of web thickness WT exceeds a predetermined value, the dimension deviation obtaining unit 112, their size deviation d _w, d _f Information on the steel type and target dimensions of the H-section steel having As an output form of the information, for example, at least one of transmission to the parameter setting device 100, storage on a portable storage medium, and display on a computer display can be employed.

Parameter setting apparatus 100, a large dimensional deviations than the predetermined value d _w, acquires the information of the steel grade and the target dimensions of the H-shaped steel having a d _f, the coefficient corresponding to the said steel grades, the size range of the target dimension belongs Update f _{i, j} and g _{i, j} . That is, the parameter setting device 100 is different from the combination of two roll chances included in the operation schedule / result data used when the coefficients f _{i, j} and g _{i, j} are derived as the operation schedule / result data. The coefficients f _{i, j} and g _{i, j} are derived as described above using the operation schedule / actual data in the two roll chances of the combination, and the coefficients corresponding to the steel type and the dimension range to which the target dimension belongs. f _{i, j} and g _{i, j} are updated to the derived coefficients f _{i, j} and g _{i, j} and output. As described with reference to FIG. 2 and FIG. 4 to FIG. 8, two rolls in which at least one of the roll chances A and B is different when the operation schedule / result data of the roll chances A and B are used. The operation schedule and performance data in the opportunity will be used.

<Roll position change amount deriving unit 113>
The roll position change amount deriving unit 113 uses the parameter acquisition unit 111 to calculate the coefficients f _{i, j} and g _{i, j} corresponding to the steel type and target dimensions of the H-section steel manufactured after the rolling by the universal rolling mill 130 is completed. Extracted from the obtained coefficients f _{i, j} and g _{i, j} .
Next, the roll position change amount deriving unit 113 includes the extracted coefficients f _{i, j} and g _{i, j} , the web thickness WT dimensional deviation d _w and the dimensional deviation of the flange thickness FT derived by the dimensional deviation deriving unit 112. by substituting the d _f (1) and equation (2) below, to derive the change amount d _V of the relative roll gap change amount d _H and the vertical rolls of the relative roll gap of the horizontal rolls.

<Roll gap set value deriving unit 114>
The roll gap set value deriving unit 114 adds the amount of change d _H of the relative roll gap of the horizontal roll derived by the roll position change amount deriving unit 113 to the current roll gap set value of the horizontal roll, Update the roll gap setting. Further, the roll gap setting value deriving unit 114 adds the change amount d _V of the relative roll gap of the heel roll to the current roll gap setting value of the heel roll to update the roll gap setting value of the heel roll.

<Roll gap set value output unit 115>
The roll gap set value output unit 115 outputs the horizontal roll roll gap set value and the heel roll roll gap set value derived by the roll gap set value derivation unit 114. As an output form of these roll gap setting values, for example, at least one of transmission to the roll gap control device 120, storage in a portable storage medium, and display on a computer display may be adopted. it can.

  As described above, the initial value of the roll gap setting value is set based on past knowledge. Therefore, the roll gap set value deriving unit 114 stores the initial value of the roll gap set value for each steel type and each dimension range. The roll gap set value deriving unit 114 changes the at least one of the steel grade of the product (H-section steel) and the target dimension size range, and first manufactures (rolls) the product having the changed content. In this case, the initial value of the roll gap setting value corresponding to the steel type and the dimension range is adopted without performing the calculations of the above-described equations (1) and (2).

(Roll gap control device 120)
The roll gap control device 120 is configured to make the relative rolls of the horizontal rolls 1201a and 1201b of the universal rolling mill 130 based on the roll gap set value of the horizontal roll and the roll gap set value of the vertical roll output from the roll gap set value output unit 115. The gap and the relative roll gap between the heel rolls 1202a and 1202b are changed. The hardware of the roll gap control device 120 can be realized by using, for example, a PLC (Programmable Logic Controller). Further, the roll gap control device 120 performs zero point adjustment by a known method.
Note that the universal rolling mill 130 can be realized by a known technique as described in the background art, and thus detailed description thereof is omitted here.

(Operation flowchart)
Next, an example of the operation of the parameter setting device 100 will be described with reference to the flowchart of FIG.
First, in step S901, the data acquisition unit 101 acquires operation schedule / result data (see FIG. 2).
Next, in step S902, the pair group creation unit 102 extracts the operation schedule / actual data in two different roll opportunities from the operation schedule / actual data acquired in step S901, and sets the target dimensions and steel types. Are extracted as regular pairs (see FIGS. 4 and 5).

Next, in step S903, the parameter deriving unit 103 derives candidates for the coefficients f _{i, j} and g _{i, j} . The parameter deriving unit 103 first derives candidates for the coefficients f _{i, j} and g _{i, j using} random numbers. When the candidates for the coefficients f _{i, j} and g _{i, j} are already derived and the process returns from step S910 to step S903, which will be described later, the parameter deriving unit 103 follows the genetic algorithm to determine the coefficients f _{i, j} that have already been derived. , G _{i, j} candidates are updated.

Next, in step S904, the parameter derivation unit 103 derives the web thickness variation d _w and a flange thickness variation d _f in operational planning and performance data constituting a regular pair extracted in step S902. The parameter derivation unit 103 derives and a web thickness variation d _w and a flange thickness variation d _f has coefficients f _i is derived in step _{S903, j,} g _i, and a candidate of the _j (1) formula, Substituting into the equation (2), the relative roll gap of the horizontal roll and the relative roll gap of the heel roll that realize the target dimension are derived (see FIGS. 6 and 7). As described above, for one candidate f _{i, j} and g _{i, j} , the relative roll gap of the horizontal roll that realizes the target dimension and the relative roll gap of the eaves roll that realizes the target dimension constitute a normal pair. It is derived only for the number of operation schedules and performance data to be performed.

Next, in step S905, the parameter deriving unit 103 determines whether the roll roll / horizontal roll roll chance is zero from the horizontal roll relative roll gap and the horizontal roll relative roll gap that are derived in step S904. The adjustment error is derived (see FIG. 8). As described above, with respect to one coefficient f _{i, j} , g _{i, j} , zero roll error between roll chances and roll roll zero roll errors between horizontal rolls are derived by the number of regular pairs. The

Next, in step S906, the parameter deriving unit 103 derives the variances σ _V and σ _H of the zero tone error between roll roll / horizontal roll chances derived from the candidates for the coefficients f _{i, j} and g _{i, j.} The average value is derived as a score s. The score s is derived (one by one) for each of the candidates for the coefficients f _{i, j} and g _{i, j} .

Next, in step S907, the parameter deriving unit 103 determines whether the score s derived in step S906 has a score s that is lower than a preset reference value. As a result of this determination, if there is a score s that falls below a preset reference value, the process proceeds to step S908. In step S908, the parameter deriving unit 103 selects the candidates for the coefficients f _{i, j} and g _{i, j} corresponding to the minimum score s among the scores s that are lower than the preset reference value, as shown in equation (1). , (2) are determined as coefficients f _{i, j} and g _{i, j} .

Next, in step S909, the parameter output unit 104 outputs information indicating the coefficients f _{i, j} and g _{i, j} determined by the parameter deriving unit 103. And the process by the flowchart of FIG. 9 is complete | finished.

  If the result of determination in step S907 is that there is no score s below the preset reference value, processing proceeds to step S910. In step S910, the parameter deriving unit 103 determines whether or not the number of repetitions of the processes in steps S903 to S907 has exceeded a predetermined value. As a result of the determination, if the number of repetitions exceeds a predetermined value, the process proceeds to step S911.

In step S911, the parameter deriving unit 103 calculates the minimum score s among the scores s derived in the current processing in step S906 and the minimum score s among the scores s derived in the previous processing in step S906. of the smaller coefficient f _i corresponding to the score s _{of, j,} g _i, candidates _j, (1) formula, is determined as a coefficient f _{i, j,} g _{i, j} in equation (2). In step S909, the parameter output unit 104 outputs information indicating the coefficients f _{i, j} and g _{i, j} determined by the parameter deriving unit 103.

On the other hand, if the number of repetitions exceeds the predetermined value as a result of the determination in step S910, the process returns to step S903. Returning to step S903, as described above, the parameter deriving unit 103 updates the already derived candidates for the coefficients f _{i, j} and g _{i, j} according to the genetic algorithm. Then, the processes of steps S903 to S907 and S910 are repeated until there is a score s that is lower than a preset reference value or the number of repetitions of the processes of steps S903 to S907 exceeds a predetermined value.

Next, an example of the operation of the roll position change amount setting device 110 will be described with reference to the flowchart of FIG.
First, in step S <b> 1001, the parameter acquisition unit 111 acquires and stores information on the coefficients f _{i, j} and g _{i, j} for each steel type and size range output by the parameter output unit 104.

Next, in step S1002, the dimension deviation deriving unit 112 determines the actual dimension h _wr of the web thickness WT and the actual dimension h _fr of the flange thickness FT in the H-section steel that has been rolled by the universal rolling mill 130, and the H-section steel. The target dimension h _wt of the web thickness WT and the target dimension h _ft of the flange thickness FT are obtained online. The dimensional deviations deriving unit 112 derives the dimensional deviations d _f of dimensional deviations d _w and a flange thickness FT of web thickness WT.

Next, in step S1003, the dimension deviation obtaining unit 112, whether at least one dimension deviation d _w and a flange thickness FT of dimensional deviations d _f of the derived web thickness WT in step S1002 exceeds a predetermined value Determine. If the result of this determination is that at least one of the dimensional deviations d _f of dimensional deviations d _w and a flange thickness FT of web thickness WT exceeds a predetermined value, the process proceeds to step S1004. Proceeding to step S1004, the dimensional deviations deriving unit 112, their size deviation d _w, and outputs information of the steel grade and the target dimensions of the H-shaped steel having a d _f, the coefficient f _{i, j,} g _{i, j-update} To the parameter setting device 100.

Large dimensional deviations than the predetermined value d _w, acquires the information of the steel grade and the target dimensions of the H-shaped steel having a d _f, in step S902 of FIG. 9 described above, the pair group creation unit 102, steels of the H-shaped steel And operation in two roll chances with a combination different from the combination of the two roll chances included in the operation schedule / result data used when deriving the coefficients f _{i, j} and g _{i, j} of the dimension range including the target dimensions Extract planned / actual data. And the pair group production | generation part 102 extracts the same combination as what is contained in the request | requirement whose steel grade and target dimension were mentioned among the extracted combinations as a regular pair. Then, the coefficients f _{i, j} and g _{i, j} are updated by performing the processing of steps S903 to S911 in FIG. 9 using the regular pairs extracted in this way. Then, the process proceeds to step S1005.

On the other hand, the result of the determination in step S1003, if at least one of the dimensional deviations d _f of dimensional deviations d _w and a flange thickness FT of web thickness WT does not exceed the predetermined value, the step S1005 is omitted step S1004 move on.
In step S1005, the roll position change amount deriving unit 113 calculates the coefficients f _{i, j} and g _{i, j} corresponding to the steel type and target dimensions of the H-section steel manufactured after the rolling by the universal rolling mill 130 is completed. , Extracted from the coefficients f _{i, j} and g _{i, j} acquired in step S1001.

Next, in step S1006, the roll position change amount deriving unit 113 calculates the coefficients f _{i, j} and g _{i, j} extracted in step S1005, the dimensional deviation d _w of the web thickness WT derived in step S1002, and the flange thickness. by substituting FT of the dimensional deviations d _f (1) to the formula and (2) below, to derive the change amount d _V of the relative roll gap change amount d _H and the vertical rolls of the relative roll gap of the horizontal rolls.

Next, in step S1007, the roll gap set value derivation unit 114, derived in the step S1006, on the basis of the change amount d _V of the relative roll gap change amount d _H and the vertical rolls of the relative roll gap of the horizontal roll, Update the roll gap setting value for the horizontal roll and the roll gap setting value for the heel roll.

Next, in step S1008, the roll gap set value output unit 115 outputs the horizontal roll roll gap set value and the heel roll roll gap set value derived in step S1007. And the process by the flowchart of FIG. 10 is complete | finished.
As described above, when at least one of the steel type of the product (H-section steel) and the size range of the target dimension is changed, and when the product having the changed contents is first manufactured (rolled), FIG. Without performing the processing according to the flowchart, the roll gap set value output unit 115 outputs the initial value of the roll gap set value corresponding to the steel type and the dimension range.

(Example)
In this example, when four (four) H-section steels having the same steel type and target dimensions are manufactured (rolled), the setting of the roll gap setting value by the method of the present embodiment and the skill in actual operation It was verified whether the roll gap setting value that can achieve the target dimension can be predicted first when the operator sets the roll gap setting value.

  FIG. 11 is a diagram showing the relationship between the roll gap setting value of the horizontal roll, the roll gap setting value of the saddle roll, the actual dimension of the web thickness / flange thickness, and the number of rolling operations. In addition, in FIG. 11, the roll gap setting value of the horizontal roll, the roll gap setting value of the saddle roll, and the actual dimension values of the web thickness and the flange thickness are shown as relative values.

  As described above, in the method of the present embodiment, the initial value of the roll gap setting value is adopted for the first rolling. In this example, the same rolling roll set value set by the operator was adopted, and the first rolling was performed.

In FIG. 11A, a graph 1101 shows the roll gap setting value of the horizontal roll according to the method of the present embodiment, and a graph 1102 shows the roll gap setting value of the horizontal roll set by the operator.
Further, in FIG. 11B, a graph 1111 shows the roll gap setting value of the saddle roll according to the method of this embodiment, and a graph 1112 shows the roll gap setting value of the saddle roll set by the operator.
Moreover, in FIG.11 (c), the graph 1121 is the actual dimension of web thickness, and the graph 1122 is the actual dimension of flange thickness. The actual dimensions of web thickness and flange thickness in the fourth H-section steel are closest to the target position.

  As shown in FIGS. 11 (a) and 11 (b), if the operator does not repeat the adjustment of the roll set value based on the past knowledge, the actual dimensions of the web thickness and the flange thickness are brought closer to the target positions. It is not possible to set a proper roll gap setting value. On the other hand, in the example shown in FIGS. 11A and 11B, when the method of this embodiment is used, if the roll gap setting value is adjusted once, the actual dimensions of the web thickness and the flange thickness can be obtained. It can be seen that a roll gap setting value that approaches the target position can be derived.

(Summary)
As described above, in this embodiment, the amount of change d _H and d _V of the relative roll gap between the horizontal roll and the heel roll is an n-dimensional line format with the web thickness change d _w and the flange thickness change _df as variables. Represented by The operation schedule / actual data for two different roll opportunities with the same steel grade and target dimensions are extracted one by one to create a regular pair, and the web thickness change amount d from the two operation schedule / actual data constituting the regular pair. to derive the _w and a flange thickness change amount d _f. Then, the linear equation of the coefficient f _{i n-dimensional, j,} g _i, and candidates _j, the web thickness variation d _w, giving a flange thickness variation d _f to the linear equation of the n-dimensional, horizontal roll vertical Relative rolls of horizontal rolls and reed rolls that achieve the target dimensions when the change amounts d _H and d _V of the roll relative roll gap are derived and the candidates for the coefficients f _{i, j} and g _{i, j} are used. Deriving gaps. Then, from the relative roll gap between the horizontal roll and the reed roll that achieve the target dimension in the regular pair, a zero tone error between roll chances is derived, and a coefficient f _i, Search _j , g _{i, j} using a genetic algorithm (GA).

As described above, in this embodiment, the n-dimensional line _format having the coefficients f _{i, j} and g _{i, j} determined by using the zero-tone error between roll chances as an evaluation index is changed to change the relative roll gap between the horizontal roll and the reed roll. A statistical model for determining the quantities d _H and d _V is created. Therefore, since the rolling results at a plurality of roll chances are used, more rolling results can be used as the rolling results used for the creation of the statistical model, and the roll gap setting value in consideration of the zero tone error at each roll chance. Can be derived. Therefore, it is possible to automatically derive a more accurate roll gap setting value.

(Modification)
<Modification 1>
In the present embodiment, a case where the change amounts d _H and d _V of the relative roll gap between the horizontal roll and the saddle roll are expressed in an n-dimensional line format with the web thickness change d _w and the flange thickness change _df as variables. This has been described by way of example (see equations (1) and (2)). However, a web thickness variation d _w and a flange thickness variation d _f, the coefficient is multiplied to them f, g and changing the amount d _H of the relative roll gap of the horizontal roll vertical rolls in accordance with, so that d _V is determined Therefore, it is not always necessary to use the expressions (1) and (2). For example, the change amount d _H of the relative roll gap of the horizontal roll vertical rolls, d _V increases monotonically with an increase in flange thickness variation d _f in case of fixing the web thickness variation d _w, and the flange As a function of monotonically increasing the amount of change d _H , d _V of the relative roll gap between the horizontal roll and the reed roll as the web thickness change d _w increases when the thickness change _df is fixed, Functions other than the linear form can be adopted. Note that this function, web thickness variation d _w and a flange thickness variation d _f is 0 change amount of the relative roll gap of the horizontal roll vertical rolls when a (zero) d _H, d _V is 0 (zero) It may or may not be.

<Modification 2>
In the present embodiment, the case where information on the coefficients f _{i, j} and g _{i, j} is output from the parameter setting device 100 has been described as an example. However, it is not always necessary to do this. For example, the functions of the expressions (1) and (2) in which the determined coefficients f _{i, j} and g _{i, j} are set may be output.

<Modification 3>
In this embodiment, the case where the roll gap set value is derived and output has been described as an example. However, it is not always necessary to do this. For example, the horizontal roll relative roll gap change amount d _H and the heel roll relative roll gap change amount d _V derived by the roll position change amount deriving unit 113 are output. May be.

<Modification 4>
In the present embodiment, the case where a genetic algorithm is used has been described as an example. However, an algorithm that solves a known optimization problem other than the genetic algorithm may be used. For example, metaheuristics other than the genetic algorithm (for example, simulated annealing (SA) method, tabusearch method, particle swarm optimization (PSO), particle swarm optimization) may be used. You may use a hit method or a hill-climbing method.

<Modification 5>
In the present embodiment, the case where the parameter setting device 100 and the roll position change amount setting device 110 are different devices has been described as an example. However, for example, the function of the parameter setting device 100 may be included in the roll position change amount setting device 110.

<Modification 6>
In this embodiment, the case where the product is an H-section steel has been described as an example. However, the product is not limited to the H-section steel as long as it is rolled (manufactured) by a universal rolling mill. For example, a shape steel other than an H-shaped steel such as a grooved steel may be used as the product.

<Other variations>
The embodiment of the present invention described above can be realized by a computer executing a program. Further, a computer-readable recording medium in which the program is recorded and a computer program product such as the program can also be applied as an embodiment of the present invention. As the recording medium, for example, a flexible disk, a hard disk, an optical disk, a magneto-optical disk, a CD-ROM, a magnetic tape, a nonvolatile memory card, a ROM, or the like can be used.
In addition, the embodiments of the present invention described above are merely examples of implementation in carrying out the present invention, and the technical scope of the present invention should not be construed as being limited thereto. Is. That is, the present invention can be implemented in various forms without departing from the technical idea or the main features thereof.

(Relationship with claims)
An example of the relationship between this embodiment and the claims is shown below. It should be noted that the claims are not limited to those described in the present embodiment, as shown in the modified examples.
The data acquisition unit (process) is realized by, for example, the data acquisition unit 101 (which performs the process of step S901).
The pair group creation means (process) is realized by, for example, the pair group creation unit 102 (which performs the process of step S902).
The parameter deriving means (process) is realized by, for example, the parameter deriving unit 103 (which performs the processes of steps S903 to S908, S910, and S911).
Dimensional change is realized by, for example, web thickness variation d _w and a flange thickness variation d _f.
The coefficient multiplied by the dimensional change amount is realized by, for example, the coefficients f _{i, j} and g _{i, j} .
The change amount of the relative roll gap is realized by, for example, the change amount d _H of the relative roll gap of the horizontal roll and the change amount d _V of the relative roll gap of the saddle roll. Further, the change amount of the relative roll gap for a set of horizontal rolls is realized by the change amount d _H of the relative roll gap of the horizontal roll, and the change amount of the relative roll gap for a set of rolls is This is realized by the change amount d _V of the relative roll gap of the roll.
The function in which the change amount of the relative roll gap is determined according to the dimensional change amount that is the difference between the actual value of the product dimension and the target value and the coefficient that is multiplied by the dimensional change amount is, for example, Equation (1), This is realized by using equation (2).
The dimensional change amount deriving means (process) is realized by, for example, the parameter deriving unit 103 (which performs the process of step S904).
The coefficient deriving means (process) is realized by, for example, the parameter deriving unit 103 (which performs the processing of steps S906 to S908, S910, and S911).
The relative zero tone error is realized by, for example, the zero tone error between the A → B roll chances of the saddle roll and the zero tone error between the A → B roll chances of the horizontal roll shown in FIG.
The variation of the relative zero tone error is realized by, for example, variances σ _V and σ _H of the zero tone error between roll chances of the saddle roll and horizontal roll.
The coefficient candidate deriving unit is realized by, for example, the parameter deriving unit 103 (which performs the process of step S903).
The relative roll gap deriving unit is realized by, for example, the parameter deriving unit 103 (which performs the process of step S904).
The relative roll gap that realizes the target dimension in the operation schedule / actual data includes, for example, a vertical roll gap (relative roll gap of vertical roll) and a horizontal roll gap (relative of horizontal roll) that realize the target dimensions of FIGS. Roll gap).
The relative zero tone error deriving means is realized by, for example, the parameter deriving unit 103 (which performs the process of step S905).
The dimension deviation deriving means (process) is realized by, for example, the dimension deviation deriving unit 112 (the process of step S1002 is performed).
The roll position change amount deriving unit (process) is realized by, for example, the roll position change amount deriving unit 113 (which performs the process of step S1006).
The roll position changing step is realized by, for example, processing of the roll gap control device 120.

  DESCRIPTION OF SYMBOLS 100: Parameter setting apparatus 101: Data acquisition part 102: Pair group preparation part 103: Parameter derivation part 104: Parameter output part 110: Roll position change amount setting apparatus 111: Parameter acquisition part 112: Dimensional deviation Deriving unit, 113: Roll position change amount deriving unit, 114: Roll gap setting value deriving unit, 115: Roll gap setting value output unit, 120: Roll gap setting device, 130: Universal rolling mill

Claims (8)

A parameter setting device for setting parameters for changing the position of a roll in a universal rolling mill,
A target dimension of a product that is rolled and manufactured by the universal rolling mill, an actual dimension of the product, a type of the product, and a distance between the two rolls at positions facing each other. A roll gap setting value that is a setting value of a relative roll gap that is a distance when the zero point is a reference, and a roll chance that is a period from when the zero point is adjusted until the zero point is adjusted next. Data acquisition means for acquiring operation schedule / result data included as data items;
A combination of two operation schedule / actual data extracted one by one from the operation schedule / actual data at two different roll chances, wherein the target dimension of the product and the type of the product are the same 2 A pair group creating means for creating a normal pair which is a combination of the two operation schedules / actual data;
Optimize the coefficient in a function in which the change amount of the relative roll gap is determined according to a dimensional change amount that is a difference between the actual value of the product dimension and a target value, and a coefficient multiplied by the dimensional change amount . Parameter deriving means for deriving as the parameter using a problem algorithm ;
The parameter derivation means includes dimensional change amount derivation means for deriving the dimensional change amount from the two operation schedule / actual data constituting the regular pair,
Using the dimensional change amount derived by the dimensional change amount deriving means and the function, the relative roll gap that realizes the target dimension in the two operation schedule / actual data constituting the same regular pair. parameter setting device variation of a difference relative zero tone error, characterized by further comprising a, a coefficient deriving means for deriving said coefficient as the parameter that is to be the smallest in the algorithm of the optimization problem. The parameter derivation means includes coefficient candidate derivation means for creating the coefficient candidates,
By substituting the coefficient candidate derived by the coefficient candidate deriving unit and the dimensional variation derived by the dimensional variation deriving unit into the function, the target dimension in the operation schedule / actual data is obtained. The amount of change of the relative roll gap to be realized is derived for each of the normal pair and the coefficient candidate, and the amount of change of the relative roll gap derived and the roll gap set value included in the operation schedule / result data. Relative roll gap derivation means for deriving a value obtained by adding as a relative roll gap that realizes the target dimension in the operation schedule / actual data, for each of the normal pair and the coefficient candidates,
Relative zero tone error deriving means for deriving the difference between the relative roll gaps in the same normal pair, derived by the relative roll gap deriving means, as the relative zero tone error, for each of the coefficient candidates; Further comprising
2. The parameter according to claim 1, wherein the coefficient deriving unit derives, as the parameter, the coefficient candidate that minimizes the variation of the relative zero tone error derived by the relative zero tone error deriving unit. Setting device.   3. The function according to claim 1, wherein the function is a function that represents the amount of change in the relative roll gap in an n-dimensional (n is a natural number) linear format with the amount of change in the roll gap as a variable. Parameter setting device. Each of the rolls is oriented in a direction in which the respective rotation axes are horizontal to the ground, and is a pair of horizontal rolls positioned on the upper side and the lower side of the material to be rolled, and directions in which the respective rotation axes are perpendicular to the ground. With a set of scissors rolls located on the right and left sides of the material to be rolled,
The roll gap setting value included in the operation schedule / result data includes the roll gap setting value for the set of horizontal rolls, and the roll gap setting value for the set of saddle rolls,
The function is a coefficient by which the amount of change in the relative roll gap for the set of horizontal rolls is a deviation between the actual value of the product dimension and the target value, and the dimensional change amount. And a change amount of the relative roll gap with respect to the set of reed rolls, a dimensional change amount that is a deviation between an actual value of the product dimension and a target value, and the dimensional change amount. And a function determined according to the coefficient to be multiplied,
The relative zero-tone error is a relative zero-tone error that is a difference in the relative roll gap between the set of horizontal rolls that realizes the target dimension in the two operation schedule / actual data constituting the normal pair; A relative zero-tone error that is a difference between the relative roll gaps for the set of saddle rolls that realize the target dimensions in the two operation schedule / actual data constituting the regular pair,
The variation in relative zero tone error includes a variation in relative zero tone error for the set of horizontal rolls and a variation in relative zero tone error for the set of saddle rolls. The parameter setting device according to any one of the above.   The parameter setting apparatus according to claim 1, wherein the coefficient deriving unit derives the coefficient using metaheuristics. A roll position change amount setting device for setting a change amount of the roll position using the function in which the coefficient derived as the parameter is set by the parameter setting device according to claim 1. Because
Dimensional deviation derivation means for deriving the difference between the actual dimension of the product different from the product in the operation schedule / actual data and the target dimension;
The difference between the actual dimension of the product and the target dimension derived by the dimension deviation deriving unit is substituted as the dimensional change amount into the function in which the coefficient derived as the parameter by the parameter setting device is set. And a roll position change amount deriving means for deriving the change amount of the relative roll gap. A roll position changing method for changing the position of a roll in a universal rolling mill,
A target dimension of a product that is rolled and manufactured by the universal rolling mill, an actual dimension of the product, a type of the product, and a distance between the two rolls at positions facing each other. A roll gap setting value that is a setting value of a relative roll gap that is a distance when the zero point is a reference, and a roll chance that is a period from when the zero point is adjusted until the zero point is adjusted next. A data acquisition process to acquire operation schedule / result data included as data items,
A combination of two operation schedule / actual data extracted one by one from the operation schedule / actual data at two different roll chances, wherein the target dimension of the product and the type of the product are the same 2 A pair group creation step of creating a regular pair that is a combination of the two operation schedules / actual data;
Optimization of the coefficient in a function in which the amount of change in the relative roll gap is determined according to a dimensional change amount that is a difference between the actual value of the product dimension and a target value, and a coefficient multiplied by the dimensional change amount A parameter derivation process derived using the algorithm of
A dimensional deviation derivation step for deriving a difference between the actual dimension and the target dimension of the product different from the product in the operation schedule / actual data;
By substituting the difference between the actual dimension of the product and the target dimension derived in the dimension deviation deriving step as the dimensional change amount into the function in which the coefficient derived in the parameter deriving step is set, A roll position change amount deriving step for deriving a change amount of the relative roll gap;
A roll position change step for changing the position of the roll based on the change amount of the relative roll gap, derived by the roll position change amount derivation step;
The parameter derivation step includes a dimensional change amount derivation step for deriving the dimensional change amount from the two operation schedule / actual data constituting the regular pair, and
Using the dimensional change amount derived by the dimensional change amount deriving step and the function, the relative roll gap that realizes the target dimension in the two operation schedule / actual data constituting the same regular pair. roll position changing method characterized by further comprising a coefficient deriving step of deriving the coefficients of variations of a difference relative zero tone errors are minimized in the algorithm of the optimization problem, the.   A program that causes a computer to function as each unit of the parameter setting device according to any one of claims 1 to 5.

Images