JPH05329516A - Automatic controller for brush roll - Google Patents

Abstract

PURPOSE:To always maintain a coated layer in the best condition by automatically controlling brush rolls to remove a coated layer generated on the surface of a rolling work roll on a product coil manufacturing process of a hot rolling line. CONSTITUTION:The state of the coated layer of the roll surface which is generated when a metal strip 4 is rolled is read and digitized by a picture reader 11 and divided by the neuro arithmetic processing means 13 into a limited number of patterns and the optimum controlled variable to a brush roll 7 is inferred by the fuzzy inference processing means 14 from the pattern. The brush roll controlling means 8 drives and controls the actual brush roll 7 through the inferred result.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a brush roll automatic control device for controlling a brush roll for removing a coating layer generated on the surface of a rolled work roll in a product coil manufacturing process of a hot rolling line for a metal strip such as aluminum. ..

[0002]

2. Description of the Related Art In a hot rolling line for aluminum, for example, as shown in FIG. 4, a distance d between a pair of rolling work rolls 1a, 1b is set to be wide so that the rolling work rolls 1a, 1b in a rotating state. When aluminum ingots are brought into contact with each other, the ingots are caught between the rolling work rolls 1a and 1b and rolled in the process of passing between the rolling work rolls 1a and 1b.
Then, when the distance d between the pair of rolling work rolls 1a and 1b is narrowed by repeating forward rotation and reverse rotation, an aluminum strip 2 as shown in the drawing is obtained. When the aluminum strip 2 reaches the specified thickness t, the aluminum strip 2 is wound into a coil.

In a hot rolling line in which such an operation is performed, a film mainly called aluminum oxide, which is mainly composed of aluminum oxide, is formed on the roll surfaces 3a and 3b of the rolling work rolls 1a and 1b. And, the surface quality of the final product coil is greatly influenced by this roll coating.

This roll coating, which greatly affects the surface quality of the product coil, is closely related to the material (steel type) of aluminum to be rolled and rolling oil. The aluminum strip 2 serving as a product coil is rolled work roll 1 through the coating layer formed on the roll surfaces 3a and 3b.
Since it is rolled at a and 1b, if the coating layer becomes too thick, a part of the coating layer will fall off to form scratches on the roll surfaces 3a and 3g, resulting in a significant defect in the surface quality of the aluminum strip 2.

On the contrary, if the coating layer is too thin, the leading edge of the aluminum strip 2 will not be smoothly meshed between the rotating work rolls 1a and 1b, and a defective biting will occur, or a slip after the biting will occur. Therefore, the surface of the aluminum strip 2 may be damaged. The thickness of the coating layer is also related to the surface gloss, which is one of the important items of surface quality for aluminum products.

Therefore, in order to prevent defects occurring on the surface of the aluminum strip 2 as such a product and ensure a good surface quality, it is necessary to remove this coating layer moderately.

Generally, a brush roll is used as a device for removing the coating layer. A steel brush is used to remove a large amount of the coating layer, and a nylon brush or the like is used to uniformly smooth the coating layer.

Then, depending on the state of the coating layer,
The following two methods have been proposed as brush roll control methods for selectively using a steel brush and a nylon brush.

(1) Based on past rolling results, inside the control computer, the type of brush roll used for each type of metal such as aluminum to be rolled, whether or not each brush roll is used, and use of the designated type of brush roll Each condition such as time and brush roll rotation speed is registered in advance. Then, the type of brush roll designated by the control computer is operated under the designated condition to remove the coating layer. (2) The operator controls the amount of coating on the industrial TV (IT
While observing in V), the brush roll is manually operated to remove the coating layer.

[0010]

However, the brush roll control methods described in (1) and (2) above still have the following problems to be solved.

That is, in general, the rolling work rolls 1a,
Rolling oil is interposed between 1b and the aluminum strip 2 to be rolled, but when the control computer shown in (1) is used, the operating conditions of the brush rolls are obtained offline from past rolling experience values. Therefore, the relationship between operating conditions and rolling oil is not taken into consideration. Further, even if it is considered, the change (deterioration) due to the circulating use of rolling oil cannot be completely absorbed within the operating condition.

Therefore, as a result, the brush roll control shown in (2) is artificially carried out, but in this case, empirical support is required for the operation, and even if a highly skilled person It is difficult to set constant operating conditions. Therefore, there are large variations in product quality depending on the operator. Further, since the operator's judgment itself is probably based on his sense and is not quantitatively established, it is difficult to inherit the technique and standardize the work.

The present invention has been made in view of the above circumstances, and evaluates the state of the coating layer on the roll surface generated during the rolling of a metal strip by a neuro, and the result is fuzzy reasoning By setting it as the optimum control amount, the coating layer can be removed reasonably, and the surface quality of the product coil can be made uniform and reliable without fail, relying on the experience of the operator. An object of the present invention is to provide a brush roll automatic control device.

[0014]

In order to achieve the above object, in the brush roll automatic control device of the present invention, a coating generated on the roll surface of a rolling work roll for rolling a metal strip in a hot rolling line for the metal strip. A brush roll that removes layers, an image reading device that reads the image of the roll surface of the rolling work roll, an image processing unit that digitizes the image data read by this image reading device, and a numerical value by this image processing unit Neuro arithmetic processing means for recognizing the converted image data by a neural network method, and fuzzy inference for inferring a control amount for a brush roll from a pattern obtained by the neuro arithmetic processing means using a predetermined inference rule. The brush roll is processed by the processing means and the control amount inferred by the fuzzy inference processing means. That is obtained by a brush roll drive control means for controlling the amount removed relative to the coating layer.

[0015]

With the brush roll automatic control device configured as described above, a two-dimensional image showing the color, brilliance, etc. of the coating layer generated on the roll surface of the rolling work roll for rolling the metal strip can be displayed by the image reading device. It is read and digitized by the image processing means.

Then, this value is pattern-recognized as the state of the coating layer by the neuro arithmetic processing means. Based on this recognition result, the fuzzy inference processing means infers the optimum control amount of the brush roll and sets it in the brush roll drive control means. Therefore, it is possible to easily make the surface quality of the rolled product uniform and improve the surface quality without depending on the experience of the operator.

[0017]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described below with reference to the drawings. FIG. 1 is a block diagram showing a schematic configuration of a brush roll automatic control device of an embodiment.

Rolling work rolls 5a, 5b are arranged so as to sandwich the conveying path of the aluminum strip 4 as a metal strip,
Rolling backup rolls 6a and 6b are arranged above and below the rolling work rolls 5a and 5b. The rolling backup rolls 6a, 6b are rolling work rolls 5a, 5b.
It is a roll that supports b. Rolling backup roll 6
The rolls a, 6b and the rolling work rolls 5a, 5b are driven and controlled by the rolling drive device 10 in the forward and reverse directions, and the distance d between the rolling work rolls 5a, 5b is also variably controlled.

A brush roll 7 is arranged so as to face the upper work roll 5a. This brush roll 7
Is a roll for removing the coating layer generated on the rolling work roll 5a, and is composed of a steel brush 7a located on the upper side and a nylon brush 7b located on the lower side. Then, the brush roll 7 is pressed against the roll surface of the rolling work roll 5a by the set pressing pressure by the brush roll driving device 8 to remove the coating layer. Further, oscillation in the width direction is also possible.

The coolant nozzles 9a, 9b and 9c are composed of a rolling backup roll 6a and a rolling work roll 5a.
Rolling oil that has both cooling and lubrication properties is sprayed onto the roll surface. Whether or not the coolant nozzles 9a, 9b, 9c are used is determined by automatic setting by a computer and manual intervention by an operator. There are several types of rolling oils used, but since they are repeatedly used, deterioration due to aging occurs, which adversely affects the surface quality.

The image reading device 11 composed of, for example, an industrial camera detects a two-dimensional image of the color and volatility of the coating layer generated on the roll surface of the rolling work roll 5a in the width direction of the rolling work roll 5a. .. Image reading device 11
The two-dimensional image data read by the
Sent to.

The image processing device 12, as shown in FIG.
The input two-dimensional image data is equally divided in the circumferential direction and the axial direction of the roll, and the chromaticity and luminance of the coating layer in each area 12a are digitized to 0.0 to 1.0. The maximum value of the numerical values in the respective regions 12a belonging to each the same axial position is defined as a coating amount K ₁ ~K _n in the relevant axial position. Then, the coating amounts K _{1 to} K _n are calculated each time the rolling work roll 5 a rotates N times preset by the processing speed and the required resolution of each processing device,
The data is transmitted to the next neural processing unit 13.

The neuro-calculation processing unit 13 uses a large number of learning data stored in the database 13a, and based on the coating amounts K _{1 to} K _n at each axial position input from the image processing apparatus 12, The pattern is recognized in the axial width direction (width direction) of the roll.

Specifically, in the database 13a,
As shown in FIG. 3, the numerical coating amount K of the image data for each type of past metal strip and each rolling oil
And a plurality of relations between the operator and the judgment value are collected as learning data. The neuro operation processing unit 13 determines the current coating state by learning them by, for example, the back propulsion method. In the apparatus of the embodiment, the judgment is made into three patterns of thin, good and thick. The pattern obtained by the neuro calculation processing unit 14 is sent to the next fuzzy inference processing unit 14.

The fuzzy inference processing section 14 follows, for example, a number of inference rules stored in the database 14a so as to obtain an optimum roll coating state for each steel type / rolling oil obtained in advance from past rolling results. Inference processing such as is performed. It should be noted that the operator previously inputs each condition other than the pattern indicating the coating amount, such as material and quality designation, using the man-machine interface device 14b. For example, the following inference rules are stored in the database 14a by the IF-THEN method.

(1) IF (a steel type that is difficult to bite) THEN (thick coating thickness) IF (rolling oil has changed over time) THEN (thick coating thickness) IF (steel type with strong adhesion THEN (decrease the coating thickness) These rules are rules that determine the optimum coating amount in consideration of the deterioration of the rolling stock and rolling oil and rolling oil due to aging.

(2) IF (current coating thickness> optimum coating thickness) THEN (need to use) IF (material not capable of removing coating) THEN (no need to use) These rules are based on comparison between the optimum coating amount and the current coating amount. Is a rule that determines whether or not to use. In addition, if it is necessary to use a brush roll in (2), then the control amounts in (3), (4), and (5) below are determined. (3) Brush material selection

IF (a steel type that is easily coated) THEN (steel is used) IF (quality is prioritized over removal time) THEN (nylon is used) (4) Pressing pressure and time selection

IF (a large amount of removal and nylon) THEN (high pressure, long time) IF (small removal amount) THEN (low pressure, short time) (5) Oscillation amount in the width direction (position , Speed) selection IF (large deviation in width direction) THEN (low speed) IF (small deviation in width direction) THEN (high speed)

As described above, it is easy to use the fuzzy inference method for the control amount which the operator conventionally judges based on the state of the rolling work roll 5a, the state of use of the rolling oil, and the type of steel based on experience and intuition. It becomes possible to decide.

Each of the above-mentioned inference results obtained by the fuzzy inference processing device 14 is sent to the brush roll drive device 8 as an optimum control amount. The brush roll driving device 8 follows the control amount when removing the coating layer.
Either one of the rolls 7a and 7b is selected, the selected rolls 7a and 7b are rotationally driven, and further moved forward to be brought into contact with the roll surface of the rolling work roll 5a at a predetermined pressure. As a result, the coating layer generated on the roll surface of the rolling work roll 5a is started to be removed. And
When the time required for coating removal set in the brush roll driving device 8 has elapsed, the operation is stopped and the brush roll 7 is retracted.

In the case of the brush roll automatic control device configured as described above, the database 13a of the neuro arithmetic processing device 13 is preliminarily stored in the database 13a of each type of metal strip and each rolling oil, and the numerical coating amount K of the image data and the skill. The relationship with the judgment result of thinness, goodness, thickness, etc. is set as learning data. The larger the number of sets of learning data, the more appropriate the pattern determination result.

Similarly, the inference rule for each condition is set in the database 14a of the fuzzy inference processing device 14. As the number of inference rules set increases, the inference accuracy can be improved.

In this way, each database 13a, 14
If the learning data and the inference rule are set in a, the rolling work roll 5 can be set only by setting the conditions such as the material from the man-machine interface device 14b during the actual operation.
The thickness of the coating layer on the roll surface of a is always controlled to the optimum value.

[0035]

As described above, according to the brush roll automatic control device of the present invention, the state of the coating layer on the roll surface generated during the rolling of the metal strip is read by the image reading device and digitized. The pattern is evaluated by a neuron, and the result is set as the optimum control amount of the brush roll by fuzzy reasoning. Therefore, the coating layer can be removed moderately, and the surface quality of the product coil can be made uniform and the quality of the product coil can be improved reliably and easily without depending on the experience of the operator.

[Brief description of drawings]

FIG. 1 is a block diagram showing a schematic configuration of a brush roll automatic control device according to an embodiment of the present invention,

FIG. 2 is a diagram showing two-dimensional image data read by the image reading apparatus of the apparatus of the embodiment,

FIG. 3 is a schematic diagram showing an example of learning data stored in a database of a neuro calculation processing unit,

FIG. 4 is a perspective view showing a relationship between a general rolling work roll and a metal strip.

[Explanation of symbols]

4 ... Aluminum strip, 5a, 5b ... Rolled work roll, 6a,
6b ... Rolling backup roll, 7 ... Brush roll, 7
a ... Steel brush, 7b ... Nylon brush, 8 ... Brush roll drive device, 10 ... Rolling roll drive control device, 1
DESCRIPTION OF SYMBOLS 1 ... Image reading device, 12 ... Image processing device, 13 ... Neuro arithmetic processing device, 14 ... Fuzzy arithmetic processing device.

─────────────────────────────────────────────────── ─── Continuation of the front page (51) Int.Cl. ⁵ Identification code Internal reference number FI Technical display location G06F 15/18 8945-5L 15/62 400 9287-5L

Claims (1)

[Claims] 1. A brush roll for removing a coating layer generated on the roll surface of a rolling work roll for rolling the metal strip in a metal strip hot rolling line, and an image reading for reading an image of the roll surface of the rolling work roll. A device, an image processing means for digitizing the image data read by the image reading device, and a neuro arithmetic processing means for recognizing the pattern of the image data digitized by the image processing means by using a neural network method. A fuzzy inference processing means for inferring a control amount for the brush roll from a pattern obtained by the neuro calculation processing means using a predetermined inference rule, and a control amount inferred by the fuzzy inference processing means Brush roll drive control hand for controlling the removal amount for the coating layer in the brush roll Brush roll automatic control device with steps.