KR20010060945A - Automatic Setup Method of Tension Leveller by Fuzzy Expert System - Google Patents

Abstract

PURPOSE: An automatic set-up method of a tension leveler is provided to maintain an optimum shape quality by integrating an operation pattern. CONSTITUTION: An operation rule is composed of a kind and a thickness of a strip, an intermesh of a tension leveler, an elongation, a tension and a widthwise inflection height. In inputting an operating condition to be deviated from the operation rule, a new set-up value is offered by a fuzzy inference. The set-up value is reset according to an inflection direction and the inflection height when the inflection height is over a set value though the new set-up value is applied.

Description

Automatic Setup Method of Tension Leveler by Fuzzy Expert System

The present invention relates to a method for automatically setting up the tension leveler, and in particular, after creating a knowledge base based on the existing operation data, the fuzzy expert system presents the operator with an appropriate leveler setup value for minimizing the width bending in accordance with each steel type. Is a method of automatic setup of a tension leveler using a fuzzy expert system to easily find the optimal leveler setup.

The shape of the cold rolled coil or steel sheet has not only a visible problem but also greatly influences the automation task in the post-process. Therefore, the shape quality of the final product is a major determinant of the product grade along with the tensile stress value. The work for shape can be largely divided into two types: rolling work in hot rolled and cold rolled tandem cold rolling mill (TCM) combined with other work (thickness change) and skin pass mill (SPM) Mill, tempering mill (TPM), tension leveler (Tension Leveler), etc. can be divided into shape correction work. From the basic shape of the thick plate, shape rolling of hot rolled and cold rolled TCM, shape correction between processes, operation considering the shape in continuous annealing line (CAL) and plating process and final shape correction should be performed organically. Therefore, the optimal shape can be obtained only through comprehensive analysis.

In recent years, the shape control technology in the steel production technology has been further advanced, and the importance of shape correction has been added as the requirements of the shape demand of the customer become strict. Shape correction can be said to remove shape defects, which are caused by uneven stretching of the rolling process, material, temperature, and mechanical conditions, and are rejected as defective products during the processing of the product. These shape defects are divided into two-wave (cyclic shape defects occurring at the edge of the steel sheet), medium waves (cyclic shape defects occurring at the center of the steel sheet) and bending, which not only damages the appearance of the product, but also reduce reliability and price. . In addition, since these defects are a detrimental factor in processing defects, mailing, automation, etc. in various processing processes, shape correction work is required to correct them.

In addition, shape correction using the tension leveler of the coil preparation line (CPL), which is the final stage of cold rolled steel sheet product processing, is one of the most important processes that influence the shape quality. On the other hand, leveler hardware is supplied by the manufacturer, and the model development, configuration and control necessary for actual operation is performed by the supplied steel company, so the subtle shape compensation technology is left to the operator. Since the know-how about these operating standards is severe, the quality control is difficult when there is a great variation in each operator. Therefore, it is necessary to make the initial setup work of the leveler consistent with the knowledge base of the operator-specific operating standards.

Currently, operator experience inputs elongation and reference tension settings and roll intermesh, which is the interval between each leveler roll, at the cab terminal. Therefore, when the intermesh is set, the widthwise halftone, which is a shape defect reference value, exhibits a wide standard distribution due to a worker's tendency or pattern difference.

In addition, in the same steel grade, thickness, and roughness, the preferred elongation, tension, and intermesh setting according to the same appear in the operation pattern of each operator, so that it is difficult to control the widthwise bending for quality improvement. In particular, when the steel sheet needs to work on the conditions such as new steel sheet thickness, width and roughness demanded by the demand for each steel grade, the leveler setup is made only by the operator's sense, resulting in greater variation among the operators. Therefore, there is a need for leveler setup including intermesh setting in terms of cold roll quality control.

Accordingly, the present invention has been made to solve the problems of the prior art as described above, and after creating a knowledge base based on the existing operation data, fuzzy expert system appropriate leveler set-up value to minimize the widthwise bending according to each steel type It is an object of the present invention to provide a tension leveler automatic setup method that maintains optimal shape quality by consistently operating patterns.

1 is a configuration diagram showing the structure of the tension leveler and the arrangement of the rolls,

2 is a view showing a method of obtaining the height of the bend in the width direction,

3 is a flowchart of a method for automatically setting up a tension leveler using a fuzzy expert system according to an embodiment of the present invention.

♠ Explanation of symbols on the main parts of the drawing ♠

1 steel plate advancing direction 2 exit bridle roll

3: entrance bridle roll 4: first drawing roll

5: 2nd drawing roll 6: 1st bending correction roll

7: Second Bending Calibrating Roll 8: First Drawing Roll Offset

9: 2nd drawing roll offset 10: 1st bending correction roll offset

11: second half calibration roll offset 12: strip pass line

13 strip shape

14: primary trend line of strip cross section

15: The point farthest from the trend line in the (+) direction

16: The point farthest away from the trend line in the negative direction

According to the present invention for achieving the above object, consisting of the steel grade and thickness of the strip, the operating conditions, the intermesh of the tension leveler (tension leveler), the elongation, tension, and the widthwise bending height of the operation evaluation variable When the operating condition is entered using the operation rule, the new setup value is proposed by fuzzy inference and the new setup value is applied by the fuzzy inference. However, if the bending height exceeds the setting value, the width direction bending height is set. A tension leveler automatic setup method is provided for resetting the setup according to the bending direction and height.

In the following, a preferred embodiment of a tension leveler automatic setup method using a fuzzy expert system according to the present invention will be described in detail with reference to the accompanying drawings.

1 is a block diagram showing the structure of the tension leveler and the arrangement of the roll, Figure 2 is a view showing a method for obtaining the height of the bending in the width direction, Figure 3 according to an embodiment of the present invention This is a flowchart of a method for automatically setting up a tension leveler using a fuzzy expert system.

Flatness of the strip or cutting plate, ie shape correction to improve flatness, is usually achieved by tensioning and bending. Tension levelers with this combined action generally consist of a number of leveling rolls for bending-strain deformation and bridal rolls 2 and 3 for applying tension.

The structure of the overall tension leveler and the arrangement of the rolls are shown in FIG.

Intermesh means the movement amount with respect to the reference position of a roll, and the reference position of a roll makes the pass line 12 which a steel plate move to zero. The leveler is composed of a first drawing roll 4, a second drawing roll 5, a first bending correction roll 6, and a second bending correction roll 7, and the first and second drawing rolls 4, 5 Is applied to the strip, and the first and second half straightening rolls 6 and 7 are aimed at half width straightening in the width direction and in the longitudinal direction.

The operator looks at the steel grade and thickness of each strip before calibrating the strip and sets the appropriate elongation, tension, and intermesh reference values, which are databased over long operating history.

In the present invention, after devising a basic rule based on the above operation data, a new conclusion is inferred when the operation conditions are changed minutely. In other words, the operation data is a knowledge base and the new conditions thereafter are presented using general rules using the existing rules. After the leveler setting based on the result of the inference, the width of the bend in the width direction is monitored to suggest a new set-up in case of a shape defect. Unlike the steepness (percentage of the bending height over the unit width), the width of the bending height directly indicates the deformation of the strip, and it is easy to quantify the shape defect.

The method for obtaining the widthwise bending height is shown in FIG. 2. First, the first-order trend is obtained from the cross-sectional curve obtained from the strip half-measuring device. The difference between the two points farthest from the trend equation is the half height. Since the first-order trend is fast to calculate, the measurement of the half-height height does not put a load on the system operating in real time.

The width direction bounce height is used as a judgment variable as a basis for extracting a formal operation rule from operation data. Multiple setups are possible for the same operating conditions, but the setup that best removes the shape defects of the strip is limited. Therefore, to determine the combination of these setups, operation judgment variables such as the width in the bounce height are required.

In the drawings, reference numeral 1 denotes a steel sheet traveling direction, 8 denotes a first draw roll offset, 9 denotes a second stretch roll offset, 10 denotes a first bending correction roll offset, 11 denotes a second bending correction roll offset, and 13 a strip cross section. Shape 14 denotes a first-order trend line of strip cross-sectional shape, 15 indicates a point farthest from the trend line in the (+) direction, and 16 indicates a point farthest from the trend line in the (-) direction.

3 is a flowchart of the leveler setup using the fuzzy expert system. FIG. 3 contains the contents of resetting the setup value determined according to the strip steel grade and thickness according to the widthwise bending height. This reset is to prepare for the case that the shape deteriorates due to the bending state of the strip or the difference in the friction coefficient of the surface.

The operating rule for fuzzy inference consists of strip steel grade and operation conditions, leveler intermesh, elongation, tension, and width factor of the bending width, which is an operation evaluation variable. Representative operating rules are as follows.

Steel grade T2 (softest steel), with a thickness of 0.2mm, elongation 0.04%, 1st stretching roll 6mm, 2nd stretching roll 6mm, 1st bending correction roll 11.2mm, 2nd bending correction roll 7.5mm intermesh When is added, the width direction bending height becomes 1.0 mm or less.

When operating conditions that deviate from the above operating rules are entered, a new setup value is suggested by fuzzy inference. If the new setup value is applied but the bending height exceeds 1.0 mm, the second bending calibration roll 7 which is most sensitive to the widthwise bending height is reset according to the bending direction and height.

In the following, the fuzzy inference applied to the present invention will be described.

The inference method used in the present invention introduces the fuzzy synthesis law using the minimum operation proposed by Mamdani as the fuzzy association function. The fuzzy variables are mainly continuous or discrete.

For example, consider the case of a controller consisting of the following n fuzzy control rules.

R1: If x1 is A11 and x2 is A12 then w is C1

or R2: If x1 is A21 and x2 is A22 then w is C2

.....

or Ri: If x1 is Ai1 and x2 is Ai2 then w is Ci

.....

or Rn: If x1 is An1 and x2 is An2 then w is Cn

(Where x1∈U, x2∈V, w∈W)

Then, the fuzzy relation Ri is a subset of the whole set U x V x W as follows.

Ri = (Ai1 x Ai2) x Ci

At this time, the entire control rule may be expressed as follows.

R = R1 ∪ R2 ∪ ... ∪ Rn = ∪i = 1 Ri

Here, if the inputs are Ao1 and Ao2, the inference result C is as follows.

C = R (Ao1 x Ao2) = (RAo1) Ao2

= max _{x1, x2} [R (x1, x2, w) ∧ Ao1 (x1) ∧ Ao2 (x2)]

On the other hand, the above reasoning can be expressed as follows. That is, the i th rule makes the following decision.

μ _{Ci '} (w) = αi ∧ μ _Ci (w)

Here, α i represents the degree (compatibility) that the i th control rule contributes to the control operation, Ci is a fuzzy set at the conclusion of the control rule, and Ci 'is a conclusion inferred from rule i. In addition, w is a variable in the whole set C representing control.

Thus, if there are n rules, conclusion C can be obtained as

μ _c (w) = μ _{C1 '} (w) ∨ μ _C2' (w) ∨. ∨ μ _{Cn '} (w)

= [α1 μμ _C1 (w)] ∨ [α2 μμ _C2 (w)] ∨. Α [αn ∧ μ _Cn (w)]

In the above method, fuzzy inference is performed to find Co and ends. However, in control, an appropriate control input must be converted from a fuzzy set Co to a single numerical value. This is called non-fuzzy operation, and the weight-centered method, which is the most widely used method of non-fuzzy, obtains the control input wo as follows.

As described in detail above, the automatic tension leveler setup method using the fuzzy expert system of the present invention creates a knowledge base based on the existing operation data, and then the fuzzy expert system adjusts an appropriate leveler set-up value for minimizing the width bending in each steel type. As suggested by the supplier, it is possible to minimize the bending height in the width direction to improve the quality, to reduce the variation among the operators when working with the new steel sheet thickness, width and roughness, and to easily find the optimum leveler setup value.

In the above description, the technical idea of the automatic tension leveler setup method using the fuzzy expert system according to the present invention has been described with the accompanying drawings. However, the exemplary embodiments of the present invention are described by way of example and not limitation of the present invention.

In addition, it is obvious that any person skilled in the art can make various modifications and imitations without departing from the scope of the technical idea of the present invention.

Claims (1)

The operation rule is composed of steel grade and thickness of strip, which is the operating condition, intermesh of the tension leveler, which is the operating variable, the amount of movement to the reference position of the roll, elongation, tension, and the width of the bending height, which is the evaluation parameter of the width using, When operating conditions that deviate from the operation rules are input, a new setup value is presented by fuzzy inference and a new setup value is applied. However, when the bending height exceeds the setting value, it is most sensitive to the width direction bending height. Tension leveler automatic setup method using a fuzzy expert system, characterized in that for resetting the set value according to the bending direction and height.