KR100495826B1 - Method for diagnosing fault of roll - Google Patents

Abstract

The present invention relates to a method for diagnosing a roll shape abnormality during a rolling process, and more specifically, to trace back a signal regarding a material shape and a system operation measured during material processing in a continuous process system such as a rolling system to check whether there is an abnormality in a mechanical part. The present invention relates to a method for diagnosing a roll-shaped abnormality capable of diagnosing and identifying a degree of abnormality, thereby removing adverse effects occurring on a product due to component abnormality.

In the present invention, the step of receiving the basic conditions to be rolled, obtaining the setting value of the reduction force and the changed roll radius corresponding to the input conditions, the step of receiving the thickness of the exit material after rolling, the input material thickness in the width direction Dividing into N microstructures, calculating the actual roll radius using the thickness of each microstructure, the reduction force, and the set value of the changed roll radius, and the linear sum of roll radius information corresponding to each microstructure. There is provided a roll shape abnormality diagnosis method comprising the step of obtaining a three-dimensional roll shape.

Description

Method of diagnosing roll shape abnormality during rolling process {METHOD FOR DIAGNOSING FAULT OF ROLL}

The present invention relates to a method for diagnosing a roll shape abnormality during a rolling process, and more specifically, to trace back a signal regarding a material shape and a system operation measured during material processing in a continuous process system such as a rolling system to check whether there is an abnormality in a mechanical part. The present invention relates to a method for diagnosing a roll-shaped abnormality capable of diagnosing and identifying a degree of abnormality, thereby removing adverse effects occurring on a product due to component abnormality.

In general, continuous process systems such as rolling systems consist of many parts such as unrolling rolls, winding rolls, several idle rolls and driven rolls. If any one of these parts is defective, the web Has a direct and fatal effect on the product.

That is, the defect in the shape of the roll in the rolling process has a fatal effect on the product to be produced, which greatly reduces the quality of the product and has a problem of lowering the productivity of the product.

However, in spite of such a problem, until now, research on the fault of the roll in the rolling system (particularly the fault in the width direction of the roll) has not been studied.

Therefore, the present invention has been devised to meet the above demands, and by diagnosing the abnormality of the shape of the roll in the rolling system and identifying the degree of abnormality, it is possible to replace and repair parts for which the degree of abnormality exceeds a certain level. The purpose of the present invention is to provide a method for diagnosing roll-shaped abnormalities that can ultimately improve product quality and improve productivity.

In addition, it is intended to provide a method for diagnosing a roll shape abnormality, which can achieve cost reduction due to an irregular roll replacement by replacing only a roll in which an abnormality is found.

In order to achieve the above object, in the present invention, the step of receiving the basic conditions to be rolled, obtaining a setting value of the reduction force and the changed roll radius corresponding to the input conditions, the step of receiving the thickness of the exit material after rolling, input Dividing the received material thickness into N microstructures in the width direction, calculating the actual roll radius using the thicknesses of the microstructures, the pressing force, and the changed roll radius setting; A roll shape abnormality diagnosis method is provided, comprising: obtaining a three-dimensional roll shape by a linear sum of roll radius information.

Hereinafter, with reference to the accompanying drawings, the roll shape diagnosis method of the present invention will be described in detail.

The roll shape abnormality diagnosis method of the present invention was applied to a rolling process in particular in a continuous process system.

Rolling is the process of making a desired thickness by passing a material (usually iron) with a constant width and very long in length, between two rolls.

In the present invention, as shown in Figure 1, the whole material 10 to be rolled is divided into n strip elements 11, the whole material 10 is composed of the sum of the microstructures 11, each micro It is assumed that the thickness of the sieve 11 is constant in the width direction. In addition, in order to analyze the influence on the deformation of the length and width direction of the material 10, it is assumed that the deformation of the entire material 10 as a linear sum of the microstructures (11). In addition, by using the thickness information of the measured material 10, the mutual influence of each microstructure 11 is considered.

In the figure, the symbol W denotes the width of the material and R denotes the roll.

For rolling, information about the material, such as the thickness, Young's modulus, width, and friction coefficient of the rolled material, the pressing force (pressing force), the radius of the roller, the changed roll radius, and the tension on the material Information is required.

In the rolling system of the present invention, mathematical models for the rolling system are derived under the following assumptions.

1) friction coefficient schedule,

2) uniform compression,

3) yield stress schedule during rolling,

4) Ignore the elastic deformation of the material,

5) Contact angle is very small .

FIG. 2 is a front view schematically showing the rolling process, wherein the material is plastically deformed while passing between two rolls to increase its thickness. in It shows the process of changing to. At this time, the repulsive force of the material to be plastically deformed is called the reduction force (P), the equation for obtaining the reduction force is as follows.

here,

E Zero coefficient

f (arb): rolling force function considering the influence of tension

: Advanced rate

h: thickness of material

K: two-dimensional strain resistance

: 2D average strain resistance

P: reduction force

R: radius of work roll

R ': radius of the deformed work roll

r: reduction ratio

S: Roll spacing

T: tension of material

V: speed of material

: Speed of roll

W: width of material

Contact angle

: Coefficient of friction between material and roll

: Stress in x direction of material

Indicates.

Here, the reduction force can be obtained by other equations in addition to the above equation (1), of course, the present invention is not limited thereto.

The rolling force varies depending on the material's inlet and outlet thicknesses and the material's speed. Here, the velocity of the material affects the tension and the tension affects the stress of the material.

On the other hand, the following formula is substituted into the parameter which comprises Formula (1).

(2D average strain resistance)

(Contact angle of roller and material)

Meanwhile, FIG. 3 is a conceptual diagram for obtaining a radius of the deformed roll, and the equation regarding the radius of the deformed roll uses Hitchcock's equation.

( Mm2 / kgf)

Here, the radius of the deformed roll can be obtained by other equations in addition to the above equation (2), of course, the present invention is not limited thereto.

As shown in Equation 2, the radius change in the elastic region of the roll during rolling changes its value according to the difference between the rolling force and the inlet and outlet thicknesses of the material. As such, since the change of the roll radius is related to the reduction force, Equations 1 and 2 should be calculated simultaneously.

Here, the roll spacing is affected by the mill spring constant, the rolling force, and the exit thickness of the material. The mathematical model for this is as follows.

In addition, in the rolling system of the present invention, the "contact angle is very small. ", A mathematical model of the advancement rate is derived from the assumption.

As shown in the figure, the advance rate varies depending on the position of the midpoint, the roll radius, and the thickness of the exit side.

On the other hand, Figure 4 is the thickness of the roll passing in the direction of the arrow between the two rolls (roll of Stand # 1, Roll of Stand # 2) in the direction of the material flow Material is thick , The process of rolling is shown.

here, Material speed ,

: Steady state value of material velocity,

: Change of material speed from steady state,

: Change of material tension from steady state,

= Spring constant,

w: material width,

Is the cross-sectional area of the material,

L: length of material span,

Material thickness ,

: Steady state value of material thickness,

: Change of material thickness from steady state

to be.

In FIG. 4, equations in which tension transfer phenomena are considered can be derived from the law of preservation of mass for the entrance and exit of the stand. The following assumptions were made to derive the equation for the tension model.

1) Tension change of material occurs only in elastic deformation section.

2) There is a point where the relative speed is zero between the rolling roll and the material (plastic firing section).

3) Ignore changes in material width and thickness due to tack (elastic deformation section).

4) It is an elastic-plastic material and neglects deformation hardening in plastic deformation.

5) After rolling, the material is regarded as an isotropic material in the elastic deformation zone.

6) Stress outside the machine direction (x-direction) is considered to be zero (elastic deformation section). )

7) The degree of deformation is constant in the elastic deformation section.

8) The degree of deformation in the material is very small (much smaller than the unit length).

9) When no tension is applied, the cross section of the material does not change along the length of the material in the elastic deformation zone.

10) Ignore changes in material properties due to temperature and humidity (elastic deformation section).

11) When there is no deformation in the material, the density and modulus of elasticity of the material are constant over the entire cross-sectional area.

12) There is no change in mass (plastic firing section).

13) The width of the material does not change.

14) Ignore the effects of non-driven rolls between the stands.

15) It is assumed that roll speed and roll gap are fully controlled.

16) Assume that there is little sag of the material.

In FIG. 4, the deformation between the first roll Stand # 1 and the second roll Stand # 1 is In the exit of the second roll It is called. In order to derive a mathematical model for tension, when the material is continuously rolled, it is divided into plastic deformation zone and elastic deformation zone. In FIG. 4, it is assumed that the material thickness at the same point (1, 2) is the same as the material thickness at the roll exit side. In FIG. 4, the 01-1 section was divided into the plastic strain section, the 1-12 section was the elastic strain section, and the 12-2 section was divided into the plastic strain section. In the plastic deformation zone and the elastic deformation zone, the relation between stress, roll speed and material thickness of the material is investigated by using the stress strain relationship and the law of mass conservation, respectively.

If the rectangular abcd is set as the control volume in the elastic deformation section 1-12 of FIG. 4 and the law of mass conservation is applied, Equation 5 below.

The microvolume in the test volume in FIG. 5 is (dx, dy, dz <1).

The stress-strain relationship is as follows.

Since the mass does not change even if the material is deformed, the following equation is always satisfied.

here,

,

Therefore, the following equation can be obtained.

Substituting the above equation into Equation 5 yields the following equation.

At this time, according to the above assumptions (9), (10) and (11), the above formula can be written as the following formula.

Each term in this expression By dividing by, we get

By the assumption (7), the above equation is

By the assumption (8), deformation Is very small,

Therefore, the following equation (6) can be obtained.

It is necessary to find out the relationship between the deformation degree by rolling force and the deformation degree by tension between plastic deformation sections 01-1. The assumptions (4), (12) and (13) can be interpreted as follows. Applying the law of conservation of mass to the test volume eafc between the plastic deformation intervals 01-1, to be.

On both sides If you add or to be. Lay, Is the plastic strain due to the rolling force, and the stress strain at the entrance and point 1 of the material Can be expressed as (Figure 6)

That is, the stress strain in FIG. 6 is the sum of the elastic strain and the plastic strain because the elastic effect should be included when considering the plastic behavior of the material. But plastic deformation Since is the strain due to the pressing force, irrespective of the tension causing the strain in the x direction, the strain at the material entry side as the strain at the point 1 in Equation 6 , In other words Can be used.

Therefore, rearranging Equation 6 is as follows.

Since the above equation is non-linear, the linearization of the initial steady state value is as follows.

here,

, to be.

Equation 7 can be simplified as follows.

Equation 7 is the deformation degree of the material in the inspection volume in the elastic deformation region ( Small change of speed of both ends and material change of material () Linearized dynamic relationship between In FIG. 4, at the point 1 and the point 2, it is assumed that there is no slip between the roll and the material. ) Is the linear velocity at contact 1 with the material of the roll ( ), But the speed of the material on the second roll ( ) Is the speed of the roll ( ) Speed of the material in the second roll ( ) And the speed of the roll ( In order to find the relation between the sig- nals, the continuous equation is applied under the assumption (12) (13) to the test volume bgdh between the plastic deformation intervals 12-2.

Since this equation is nonlinear, the following equation can be obtained by linearizing it with the Perturbation Method.

only,

Using the above equation in Equation 8, the following equation can be obtained.

This equation can be written as follows, assuming that the tangential linear velocity of the roll and the velocity of the material are the same at points 1 and 2, respectively.

This equation shows the linearized dynamic relationship between the change in the strain (epsilon) of the material between the two rolls, the linear velocity change of the two rolls on both ends of the material, and the change in the thickness of the material. In the assumptions (3), (5), (7), (8), (10), and (11), the Hook's Law is used as follows.

,

only, Wow Is assumed to be constant within the elastic deformation region. Also, the Is the amount of change in tension at the initial steady state value. Substituting the above equation into Equation 9, the following equation can be obtained.

The mathematical model of thickness, speed, and tension from FIG. 4 is as follows.

Equation 10 is a linearized dynamic model showing the relationship between the change in tension between the two rolls, the tangential velocity of the rolls on both ends of the roll, and the change in the thickness of the roll. As can be seen in Equation 10, it can be seen that the tension between the two stands is the tension transfer of the material tension between the stand at both ends of the material (tension transfer) between the two next stand.

As mentioned above, the thickness of the material, the rolling force, the changed roll radius, etc. in the rolling system are closely related to various variables. Therefore, sensitivity analysis of various parameters in the rolling system was performed to reduce the computation time and to diagnose three-dimensional roll shape anomalies.

7 and 8 show the change according to the thickness change of the exit material after rolling for the rolling force function and the changed roll radius. As shown in FIG. 7 and FIG. 8, it can be seen that the reduction force function f (arb) and the changed roll radius are very small according to the thickness variation of the exit material.

Therefore, it can be seen that the sensitivity is very small.

In addition, Figure 9 shows the change in the rolling force with respect to the amount of change in the entry material thickness for a constant exit thickness. As can be seen in Figure 9, unlike the reduction force function and the changed roll radius, the reduction force is very large.

On the basis of the above results, in the simultaneous calculation of the rolling system models that are closely related to each other, parameters with low sensitivity to the thickness variation of the exit material after rolling are assumed to be constants.

Based on this information, the roll-shaped abnormality diagnosis method of this invention is demonstrated.

10 shows an algorithm according to the present invention.

First, after rolling, the thickness of the exit material is measured in the width direction to obtain shape information.

In this case, the thickness of the material in the width direction is measured using an X-ray gauge meter.

Then, the width direction velocity of the material located between the stand and the stand using Equation 10, which is a model of the thickness information and continuous equation of the width direction of the material at the entry and exit of the mill and the tension of the web between the two rolls And the tension is obtained (S2).

Equation 1 showing the relationship between the tension and velocity values obtained in Equation 10 and the thickness and tension of the material and the changed roll radius, and Equation 2 showing the relationship between the changed roll radius and the actual roll radius and the reduction force The initial set value of the changed roll radius and the rolling force and the actual roll radius value for the calculation is obtained. (S3)

Then, through the sensitivity analysis of the above-described parameters, the variables that can be assumed to be constant in the iterative operation are defined, and the set values obtained from the initial operation results and given conditions (tension, roll speed, feed rate of the material, rolling force, changed rolls). Radius, the actual rolling radius value, and the changed roll radius value are obtained (S4).

Simultaneous calculation is repeated until the rolling force, the actual roll radius value, and the changed roll radius value converge to obtain a result for each microstructure (S5).

As mentioned above, since the entire material is assumed to be a linear sum of each microelement, the final three-dimensional roll shape information is obtained as a linear sum of the results for each microelement.

Through this process, the three-dimensional shape abnormality diagnosis of the rolling roll is performed using the shape information and the rolling conditions of a given material.

11A and 11B have shown roll shapes with respect to the width direction. As a shape error of the form which is symmetrical with respect to a roll center part, as shown in FIG. 11A, the center part is concave, and as shown in FIG. 11B, the center part is classified into a convex form.

In addition, a shape abnormality diagnosis for a rolling roll having a symmetrical shape error as shown in FIGS. 11A and 11B and an eccentricity with respect to the cross-sectional direction of FIG. 12 is performed.

Computer simulations were performed on the rolling rolls under the following conditions.

Simulation parameters shame Drive-side mill constant ) 5000000kgf / mm Work-side mill constant ) 5000000kgf / mm Reference entry strip thickness 10 mm Reference delivery strip thickness 7.45 mm Back tension stress 15.3㎏f / ㎡ Front tension stress 20.5㎏f / ㎡ Work roll radius 210 mm Eccentricity of work roll in stand1 0.02mm, 0.05㎐ Distance between stands 6000 mm Strip width 1240 mm Poisson's coefficient of strip 0.12 Elastic modulus 21000 yen The number of finite strip element 40 The width of a finite strip element 31 mm Inlet strip velocity 3000 mm / s

Table 1 lists the conditions for computer simulation. The reference thickness of the rolling mill entry material is 10 mm, and the reference thickness of the exit material is 7.45 mm. Moreover, the amount of eccentricity which exists with respect to a roll cross section is 0.02 mm, and a generation period is 0.05 ms.

The finite strip element method is used to analyze the influence of each microelement in order to diagnose the abnormality of the three-dimensional shape of the roll. The number of microelements and the computational time for diagnosing the abnormalities are closely related. In consideration of this relationship, the following simulation assumes that the width of each microelement is 31 mm and the total material is a linear sum of 40 microelements.

(1) Convex type error

The thickness information of the width direction of the material obtained through the X-ray gauge meter at the entry and exit of the rolling mill is as follows. 13A to 13C show a three-dimensional shape of the rolling mill entry material, and it can be seen that a thickness error in the time domain and the width direction of the material is an ideal material. Only half the size of the material thickness is shown, and the value of the thickness surface of the material is shown in FIG. 13A. In addition, FIG. 13B shows the thickness variation with respect to the time domain, and FIG. 13C shows the thickness distribution in the width direction of the material.

14A to 14C show the three-dimensional shape of the? -Side material after rolling. In FIG. 14B, a thickness error having a constant period with respect to the time domain is shown, which means that there is an eccentricity with respect to the rolling roll cross section. In addition, in FIG. 14C, the thickness distribution is uneven with respect to the width direction of the material. This indicates that the rolling roll has radial errors in the roll width direction.

That is, it can be seen that the material thickness information of FIGS. 13 and 14 includes the shape information of the rolling rolls, and the three-dimensional shape abnormality of the final rolling roll can be diagnosed through the shape information and the rolling conditions of the material. It is.

15A to 15C are tension distributions of the exiting material after rolling obtained through the rolling mill entrance and exit material shape information. As can be seen from FIG. 14 due to the shape error of the rolling roll, the exit material has a periodic thickness error in the time domain and a nonuniform thickness distribution in the width direction of the material. As a result, the exiting material has a non-uniform tension distribution with respect to the time domain and the width direction of the material. This uneven tension distribution is expected to act as tension disturbance and adversely affect the thickness control of the material. Since the shape error with respect to the width direction of a rolling roll is assumed to be symmetrical about a center part, as shown in FIG. 15C, the tension distribution form with respect to the width direction of a raw material is also shown symmetrically about the center part of a raw material.

16A to 16C show three-dimensional shapes of the rolling rolls obtained through the shape information of the entry and exit materials and information such as tension, speed, and rolling conditions. As can be seen from this, the rolled roll has a more convex shape at the center due to wear at both ends, and shows a periodic radius change in the time domain. In other words, it can be seen that the rolled roll has an eccentricity with respect to the cross section and a radius error in the roll width direction.

(2) Concave type error

The thickness information of the width direction of the material obtained through the X-ray gauge meter at the entry and exit of the rolling mill is as follows. 17A to 17C show a three-dimensional shape of the rolling mill entry material, and it can be seen that it is an ideal material having no thickness error in the time domain and the width direction of the material. Only half the size of the material thickness is shown, and the value of the thickness surface of the material is shown in FIG. 17A. In addition, FIG. 17B shows the thickness variation in the time domain, and FIG. 17C shows the thickness distribution in the width direction of the material.

18A shows the three-dimensional shape of the exit material after rolling. In FIG. 18B, a thickness error having a constant period with respect to the time domain is shown, which means that there is an eccentricity with respect to the rolling roll cross section. In addition, in FIG. 18C, the thickness distribution is uneven with respect to the width direction of the material. This indicates that the rolling roll has radial errors in the roll width direction.

That is, it can be seen that the material thickness information of FIGS. 17 and 18 includes the shape information of the rolling rolls, and the three-dimensional shape abnormality of the final rolling roll can be diagnosed through the shape information and the rolling conditions of the material. It is.

19A to 19C are tension distributions of the exit material after rolling, which are obtained by using the rolling mill entrance and exit material shape information. Due to the shape error of the rolling roll, the exit material has a cyclic thickness error in the time domain and a nonuniform thickness distribution in the width direction of the material. As a result, the exiting material has a non-uniform tension distribution with respect to the time domain and the width direction of the material. This uneven tension distribution is expected to act as tension disturbance and adversely affect the thickness control of the material. Since the shape error with respect to the width direction of the rolling roll is assumed to be symmetrical with respect to the center part, as shown in FIG. 19C, the tension distribution form with respect to the width direction of the material also appears symmetrically with respect to the center part of the material.

20A to 20C show three-dimensional shapes of the rolling rolls obtained through the shape information of the entry and exit materials and information such as tension, speed, and rolling conditions. As can be seen in Figure 20a, the rolling roll has a concave shape in the center portion due to the wear of the center portion of the roll, and also shows a periodic radius change over the time domain. In other words, it can be seen that the rolled roll has an eccentricity with respect to the cross section and a radius error in the roll width direction.

As described above, the present invention reversely traces signals related to material shape (thickness, tension, etc.) and system operation (speed, etc.) measured during material processing through identification of the principle of material processing through each part in the rolling system. By diagnosing the abnormality of the shape of the roll and identifying the degree of abnormality, it is possible to replace and repair parts over which the degree of abnormality is over a certain level, thereby ultimately improving the product quality and improving productivity. .

In addition, by replacing only the roll in which an abnormality is found, it is possible to achieve a cost reduction due to an irregular roll replacement.

In addition, the algorithm developed in this study can be used as a finishing rolling process diagnosis system.

1 is a plan view showing a rolled material and a roller,

2 is a front view schematically showing a rolling process,

3 is a conceptual diagram for obtaining a radius of a deformed roll;

4 is a conceptual diagram showing rolling of a raw material between two rolls;

5 is a graph showing the microvolume in the test volume;

6 is a stress strain diagram of an elastic-plastic material,

7 is a graph showing a change amount of the reduction force function according to the change amount of the thickness of the exit material;

8 is a graph showing the amount of change in the roll radius in accordance with the amount of change in the thickness of the exit material;

9 is a graph showing a change amount of the reduction force according to the change amount of the thickness of the exit material;

10 is a processing flowchart of the roll shape abnormality diagnosis method according to the present invention;

11A and 11B illustrate a shape error of a roll, and FIG. 11A is a plan view showing a roll having a concave center portion, and FIG. 11B having a convex center portion.

12 is a cross-sectional view showing a rolling roll with an eccentricity with respect to the cross-sectional direction;

FIG. 13 shows a three-dimensional shape of the rolling mill entry material, FIG. 13A is a graph showing the value of the thickness surface of the material, FIG. 13B is a graph of thickness variation with respect to the time domain, and FIG. 13C is a thickness distribution graph of the width direction of the material. ,

14 is a three-dimensional shape of the exit material after rolling, Figure 14a is a graph showing the value of the thickness surface of the material, Figure 14b is a thickness error graph having a constant period for the time domain, Figure 14c is the width direction of the material Thickness distribution graph for,

FIG. 15 shows the tension distribution of the exit material after rolling obtained from the rolling mill entrance and exit material shape information, FIG. 15A is a graph showing the value of the thickness surface of the material, and FIG. 15B is a thickness error having a constant period for the time domain. 15C is a graph showing a thickness distribution of a width direction of a material;

FIG. 16 shows a three-dimensional shape of the rolled roll obtained through the shape information of the entry and exit materials and information such as tension, speed, and rolling conditions, FIG. 16A is a graph showing the value of the thickness surface of the material, and FIG. Thickness error graph having a constant period with respect to the area, FIG. 16C is a thickness distribution graph with respect to the width direction of the material,

17 is a three-dimensional shape of the rolling mill entry material, Figure 17a is a graph showing the value of the thickness surface of the material, Figure 17b is a thickness change graph over the time domain, Figure 17c is a thickness distribution graph of the width direction of the material ,

18 is a three-dimensional shape of the exit material after rolling, Figure 18a is a graph showing the value of the thickness surface of the material, Figure 18b is a thickness error graph having a constant period with respect to the time domain, Figure 18c is the width direction of the material Thickness distribution graph for,

Fig. 19 shows the tension distribution of the exit material after rolling obtained from the rolling mill entrance and exit material shape information, Fig. 19A is a graph showing the value of the thickness surface of the material, and Fig. 19B is a thickness error having a constant period for the time domain. 19C is a graph showing a thickness distribution of a width direction of a material;

20 shows a three-dimensional shape of a rolled roll obtained through the shape information of the entry and exit materials and information such as tension, speed, and rolling conditions, FIG. 20A is a graph showing values of the thickness surface of the material, and FIG. 20B is a time Thickness error graph having a constant period with respect to the area, FIG. 20C is a thickness distribution graph with respect to the width direction of the material.

♣ Explanation of symbols for main part of drawing ♣

10: the whole material 11: microstructure

W: width of material R: roll

Claims (7)

Receiving a basic condition of rolling; Receiving a thickness of the exit material after rolling; Obtaining a setting value of the reduction force and the changed roll radius corresponding to the input condition; Dividing the input material thickness into N microstructures in the width direction; Calculating an actual roll radius using the thickness of each microstructure, the reduction force, and the set value of the changed roll radius; And obtaining a three-dimensional roll shape by a linear sum of roll radius information corresponding to each microstructure. The method of claim 1, further comprising calculating the widthwise velocity (V) and the tension (T) of the material by using the thickness information of the material on the rolling mill entrance and exit, the continuous equation, and the mathematical model of the tension. Roll shape abnormality diagnosis method.  The method of claim 2, wherein the mathematical model for the tension satisfies the following equation. here, : Material speed ( ), : Steady state value of material velocity, : Change of material speed from steady state, : Change of material tension from steady state, Is the cross-sectional area of the material, L: length of material span, E : Young's modulus, Material thickness ), : Steady state value of material thickness, : Change of material thickness from steady state. The method of claim 1, wherein the pressing force satisfies the following equation in the step of obtaining the pressing force corresponding to the input condition and the set value of the changed roll radius. here, : Two-dimensional average strain resistance, : Stress in the x direction of the material, K: two-dimensional strain resistance, R 'is the radius of the deformed work roll, h: thickness of the material, f (arb): The reduction force function considering the influence of tension. The method of claim 1, wherein the changed roll radius satisfies the following formula in the step of obtaining a setting force of the reduction force and the changed roll radius corresponding to the input condition. Where R is the radius of the work roll, P: reduction force, W is the width of the material, h: thickness of the material, Mm 2 / kgf. The method of claim 1, wherein in calculating the actual roll radius, the rolling reduction and the set roll radius are assumed to be constants through a sensitivity analysis. The method of claim 1, wherein the thickness of the exit material after rolling is obtained by using an X-ray gauge meter in the step of receiving the thickness of the exit material after rolling.