JP5003362B2 - Product quality control method and control device - Google Patents

Description

  The present invention relates to a method for controlling the quality of products represented by steel products and the like, and a product quality control device used for controlling the product quality.

  The number of defects on the surface defects and internal defects of the product, and the number of defective products are count values represented by discrete random variables. The number of defects can be modeled as following a Poisson distribution and the number of defective products according to a binomial distribution. Conventionally, the number of defects in accordance with the Poisson distribution is managed by the c control chart, the u control chart, etc., and the number of defective products in accordance with the binomial distribution is managed by the np control chart, the p control chart, etc., and exceeds the control limit line. If the control limit line is likely to be exceeded from the time trend of the graph, a method of investigating abnormal manufacturing conditions has been taken (JIS Z 9020: 1999).

  According to the product quality control based on these control charts, it can be detected that the quality tends to deteriorate, but it is unclear which manufacturing condition is bad and in which direction it is improved. For this reason, methods for predicting product quality indicators from manufacturing condition data and methods for controlling manufacturing conditions based on the prediction results have been proposed so far in order to identify manufacturing conditions that cause quality deterioration and improve quality. ing.

  For example, Patent Document 1 discloses a technique relating to a result prediction apparatus, method, and computer-readable storage medium that predicts a product quality result based on the similarity of manufacturing condition data with past cases. Further, Patent Document 2 discloses a quality influence factor analysis method for clarifying the relationship between quality data of actual manufacturing conditions and quality using multivariate analysis, a quality prediction method using the analysis result, and a quality control method. And a technology related to a quality control device and the like. Furthermore, Non-Patent Document 1 discloses a method called a generalized linear model that can be applied to quality prediction.

JP 2001-290508 A JP 2005-242818 A P. McCullagh and J. A. Nelder, Generalized Linear Models .: Chapman and Hall, (1989)

  However, in the technique disclosed in Patent Document 1, since the current situation is predicted based on the similarity to the past case without being based on actual manufacturing condition data, for example, when the representativeness of the past case is low However, there is a problem that the prediction accuracy may be lowered, and the quality control accuracy may be lowered.

  In the technique disclosed in Patent Document 2, each product quality is classified as good or defective, and discriminant analysis is performed on the classification using the principal component analysis result of the manufacturing condition data. In discriminant analysis, good / bad is classified by threshold determination as a result of projecting the principal component analysis result onto a one-dimensional discriminant space. Furthermore, in the technique disclosed in Patent Document 2, the projection result in discriminant analysis is used as a product quality index, and the quality index is improved by regressing it on the manufacturing condition data. Seeking and controlling quality. Therefore, the quality index in the technique disclosed in Patent Document 2 is to convert manufacturing condition data into a position in the determination space so that the quality can be determined most accurately. The probability of poor quality under conditions cannot be estimated. In addition, since the technology disclosed in Patent Document 2 expresses quality only by classification into two categories, how the count value quality index according to the Poisson distribution such as the number of wrinkles changes Cannot be estimated using condition data. In the technique disclosed in Patent Literature 2, the quality index cannot be correctly linked to the manufacturing conditions of the product, and it is impossible to estimate how much the quality will be improved by changing which manufacturing conditions, so the quality is controlled. There was a problem that the optimum manufacturing conditions required above could not be specified.

  Therefore, the present invention provides a product quality control method capable of controlling the product quality based on the specified optimum manufacturing conditions, and a product quality control apparatus capable of executing the control method. The issue is to provide.

  The present invention will be described below. In addition, in order to make an understanding of this invention easy, the reference sign of an accompanying drawing may be appended in parentheses, but this invention is not limited to the form of illustration by it.

The first aspect of the present invention includes a regression equation defining unit (1) that defines a linear regression equation, a coefficient calculating unit (2) that calculates a coefficient of the linear regression equation, and a target value calculating unit that calculates a target value of manufacturing conditions. (3) A method for controlling the quality of a product using a control device (10) comprising a manufacturing condition changing unit (4) for changing a manufacturing condition based on a target value, according to the manufacturing condition The quality related to the count value of the identified product is approximated by a discrete probability distribution, and the regression equation defining step (1) is defined by the linear regression equation, the coefficient of the linear regression equation is the actual data of quality and The coefficient calculation step calculated in the coefficient calculation unit (2) using the actual data of the manufacturing conditions, the coefficient of the linear regression equation calculated in the coefficient calculation step, and the target value of the manufacturing conditions using the actual data of the manufacturing conditions calculated in the target value calculation unit (3) A target value calculating step, based on the target value of the calculated production conditions, and manufacturing condition change step of changing the production conditions changing unit production conditions (4), characterized in that it comprises a product quality control method Thus, the above problem is solved.

  In the first invention and the invention shown below (hereinafter simply referred to as “the present invention”), specific examples of “quality related to product count” are represented by surface defects, internal defects, etc. of steel materials. The number of defects and the number of defective products can be listed. Furthermore, specific examples of “manufacturing conditions” include the set values of the manufacturing equipment, the physical quantities (temperature, shape, composition, etc.) of products or intermediate products represented by steel materials, and the time of each process for manufacturing products. Can be mentioned.

In the first aspect of the present invention, the manufacturing condition includes a changeable manufacturing condition and a non-changeable manufacturing condition, and further, based on a statistic called deviance in a generalized linear model based on a discrete probability distribution, A manufacturing condition extraction step for extracting a manufacturing condition having a large influence on the quality from the changeable manufacturing conditions is provided. In the target value calculation step, the target value is calculated by using the manufacturing condition extracted in the manufacturing condition extraction step. It is preferable.

  In the first aspect of the present invention, the product is preferably a steel material.

  In the present invention, specific examples of steel products include steel products.

The second aspect of the present invention is a control device (10) used for controlling the quality of a product, and approximates the quality related to the count value of a product specified according to manufacturing conditions by a discrete probability distribution, and performs linear regression. The regression equation definition unit (1) defined by the equation, the coefficient calculation unit (2) for calculating the coefficient of the linear regression equation using the actual result data of quality and the actual result data of manufacturing conditions, and the coefficient calculation unit (2) Based on the calculated linear regression equation coefficient and the actual data of the manufacturing condition, the target value calculation unit (3) for calculating the target value of the manufacturing condition and the manufacturing based on the calculated target value of the manufacturing condition The above-mentioned problem is solved by a product quality control device (10) comprising a manufacturing condition changing unit (4) for changing conditions.

In the second aspect of the present invention, the manufacturing condition includes a changeable manufacturing condition and a non-changeable manufacturing condition, and further, based on a statistic called deviance in a generalized linear model based on a discrete probability distribution, A production condition extraction unit (5) for extracting production conditions having a large influence on quality from changeable production conditions is provided, and the production conditions extracted by the production condition extraction unit (5) in the target value calculation unit (3) are provided. It is preferable that the target value is calculated.

  In the second aspect of the present invention, the product is preferably a steel material.

  ADVANTAGE OF THE INVENTION According to this invention, it becomes possible to construct | assemble the model explaining the product quality data based on a count value by product manufacturing conditions about a product manufacturing process. Further, according to the present invention, it is possible to construct a quality estimation model in a product manufacturing process by identifying model parameters from data accumulated in the past. In addition, according to the present invention, it is possible to stably obtain a product of good quality by calculating the manufacturing condition target value for obtaining the optimum product quality and controlling the manufacturing condition.

  Embodiments of the present invention will be described below.

1. Product Quality Control Method Product manufacturing conditions in the manufacturing process of industrial products are measured results of physical quantities (temperature, shape, composition, etc.) related to products or intermediate products in the manufacturing process, measured results of physical quantities (temperature, pressure, etc.) related to manufacturing equipment, These physical quantities are configured by one or a plurality of manufacturing conditions selected from a group consisting of control target values, operating condition setting values, values measured and set between manufacturing apparatuses, and the like. Each product manufacturing condition item is represented by x _i , and a vector obtained by combining the entire manufacturing conditions is x = [x ₁ x ₂ ... X _K ] ^T. Moreover, the measurement data regarding quality represented by the number of defects, the number of defective products, etc. is represented by y. In the present invention, the quality model is constituted by a regression model in which the measurement data y related to quality is an objective variable and the manufacturing condition x is an explanatory variable. In the present invention, in constructing the quality model, a discrete probability distribution such as a binomial distribution or a Poisson distribution is assumed as the probability model of the objective variable, and the expected value of the quality y with respect to the manufacturing condition x is converted by a monotonically increasing function to obtain a linear model. A method called a generalized linear model for modeling with a regression equation is used. In the configuration of the quality model in the present invention, the coefficient of the linear regression equation is estimated by the maximum likelihood method.

1.1. First Embodiment FIG. 1 is a flowchart showing steps provided in a product quality control method of the present invention according to a first embodiment (hereinafter referred to as “control method according to the first embodiment”). As shown in FIG. 1, the control method according to the first embodiment includes a data aggregation step (step S11), a regression equation definition step (step S12), a coefficient calculation step (step S13), and a target value calculation step ( Step S14) and a manufacturing condition changing step (Step S15).

<Step S11>
In step S11, manufacturing condition data or quality data in which individual products are associated with manufacturing conditions and quality realization values are created, and a set of manufacturing condition data or quality data is created. In the manufacturing condition data, product numbers are expressed as _xn , where n = 1, 2,. The set of manufacturing condition data is a set having one or a plurality of manufacturing condition data as elements, and is represented by a matrix X = [x ₁ x ₂ ... X _N ] ^T in which the vector x _n is transposed and arranged in the row direction. Product quality represented by the focused count and y, representing the product quality data and y _n. A set of product quality data is a set having one or a plurality of product quality data as elements, and is represented by a vector Y = [y ₁ y ₂ ... Y _N ] ^T obtained by transposing y _n and arranging them in the row direction.

<Step S12>
Step S12 is a step of defining the quality of the product according to the manufacturing conditions by a linear regression equation. When the regression count vector is c = [c ₀ c ₁ ... C _N ] ^T , the linear regression equation can be expressed by the following equation 1.

<Step S13>
Step S13 is a step of calculating the coefficient of the linear regression equation based on the quality measurement result and / or the production result data. When the quality of the product to be controlled is an object with a low probability of occurrence of defects, the quality can be approximated with high accuracy by a discrete probability distribution represented by a Poisson distribution, a binomial distribution, or the like. When the quality of the product to be controlled is the number of defects of surface defects or internal defects of a long steel product, if the average number of defects per unit is λ, the number of defects y in the target quantity W is a random variable that follows a Poisson distribution with an average λW The probability can be expressed by the following equation 2.

欠点数は、単位当たりの個数で比較されるものなので、単位当たり欠点数平均λを製造条件の線形回帰式で表す品質モデルは下記式３により表すことができ、製造条件に対する対象量Ｗにおける欠点数の確率分布は下記式４で表すことができる。

Since the number of defects is compared by the number per unit, the quality model that expresses the average number of defects per unit λ by the linear regression equation of the production condition can be expressed by the following equation 3, and the defect in the target amount W with respect to the production condition: The probability distribution of the numbers can be expressed by the following formula 4.

  The estimation of the regression parameter c by the maximum likelihood method is performed by defining the log likelihood L by the following equation 5 using the product quality data set and the manufacturing condition data set, and obtaining the regression parameter c that maximizes this.

  The regression parameter c for maximizing the log likelihood L expressed by Equation 6 is obtained by finding a solution satisfying the requirements expressed by the following Equation 7 by the Newton method or the like, and the log likelihood L is maximized from the solution. This can be obtained by a method of selecting a target to be selected or by a nonlinear optimization method such as successive quadratic programming.

  On the other hand, when the quality of the product to be controlled is the number of defective products, the number of defectives y for the target product number M is the number of occurrences when the probability of occurrence of ρ when trying once is tried M times. The probability distribution can be expressed by the following equation 8 where ρ is the probability that one product will be defective.

また、式８において、

Also, in Equation 8,

は、相異なるＭ個の中からｙ個を抽出する組合せの数である。

Is the number of combinations that extract y out of the different M.

  The occurrence probability of a defective product can be expressed by the following formula 9 expressed by a linear regression formula of manufacturing conditions. Further, the probability distribution of the number of defective products with respect to the target number M with respect to the manufacturing conditions can be expressed by the following formula 10.

  When the quality of the controlled product is the number of defective products, the regression parameter c is estimated by the maximum likelihood method. The log likelihood L is defined by the following equation 11 using the product quality data set and the manufacturing condition data set, This is done by determining the regression parameter c to be maximized.

式１１において、Ｍ＝[Ｍ_１ Ｍ_２ … Ｍ_Ｎ]^Ｔである。

In Equation 11, M = [M ₁ M ₂ ... M _N ] ^T.

  From the above Equation 4, Equation 9, and Equation 10, the log likelihood L can be expressed by the following Equation 12.

  The regression parameter c that maximizes the log likelihood L expressed by Equation 12 can be obtained by the same method as that when the quality of the controlled product is the number of defects.

<Step S14>
Step S14 is a step of calculating the optimum target value of the manufacturing conditions based on the linear regression equation S (x, c). Λ (x, c) in Equation 3 and ρ (x, c) in Equation 9 correspond to a linear regression equation S (x, c) converted by a monotonically increasing function. Therefore, λ (x, c) and S (x, c) and ρ (x, c) and S (x, c) correspond to each other one to one. Further, as λ (x, c) and ρ (x, c) are smaller, the quality of the product is improved. Therefore, for any quality of the number of defects and the number of defective products, the quality improves if x _i changes in a direction in which each term c _i x _i of the linear regression equation expressed by the above equation 1 decreases. In the present embodiment, the manufacturing condition calculated using the elements of the manufacturing condition data set as the optimum target value

を最適目標値とする。製造条件データは実際の製品製造時に得られたものであるから、製造条件データに基づいて得られる最適操業目標値は、製造条件データから大きく外れることがなく、実際に適用する場合に信頼性が高い。

Is the optimum target value. Since the manufacturing condition data is obtained at the time of actual product manufacture, the optimum operation target value obtained based on the manufacturing condition data is not greatly deviated from the manufacturing condition data, and is reliable when actually applied. high.

  Specific examples of the method for calculating the optimum target value include the following.

A method for obtaining x _n that minimizes the value of S (x _n , c) and obtaining the optimum target value from the following equation (13).

  ・ Represented by Equation 13 above

と製造条件データｘ_ｎとを用いて表される

And manufacturing condition data x _n

の値が小さい順に予め定めた、ν_２個のｘ_ｎの平均値（下記式１４で表される

The average value of ν ₂ x _n (represented by the following formula 14)

）を求めることにより、最適目標値を得る方法。

) To obtain the optimal target value.

  ・ Represented by Equation 13 above

と、製造条件データ集合における平均値

And the average value in the production condition data set

とについて、成分毎に比較して、回帰係数と各条件項目との積が品質を改善する方向にある方を選択して組み合わせた、下記式１５で表される

For each of the components, the product of the regression coefficient and each condition item is selected and combined in the direction of improving the quality, and is expressed by the following formula 15.

を求めることにより、最適目標値を得る方法。

A method to obtain the optimal target value by obtaining

上記式１５において、

In Equation 15 above,

の場合（ｊ＝１、２）、

(J = 1, 2),

であり、

And

の場合（ｊ＝１、２）、

(J = 1, 2),

である。

It is.

<Step S15>
Step S15 is a step of changing manufacturing conditions based on the target value calculated in step S14. The manufacturing condition change mode is not particularly limited as long as the manufacturing condition after the change is specified so that the manufacturing condition is closer to the target value than before the change, but for example, calculated in step S14 above. The form etc. which make the set target value the manufacturing conditions after a change can be mentioned.

  Thus, according to the control method concerning 1st Embodiment provided with process S11-process S15, manufacturing conditions are changed using the target value calculated based on the manufacturing condition data obtained at the time of actual product manufacture. To control the quality of the product. Therefore, according to the control method according to the first embodiment, the quality of the product can be controlled based on the reliable target value, so that a product with a good quality can be stably obtained. A product quality control method can be provided.

1.2. Second Embodiment FIG. 2 is a flowchart showing steps provided in a product quality control method of the present invention according to a second embodiment (hereinafter referred to as “control method according to the second embodiment”). As shown in FIG. 2, the control method according to the second embodiment includes a data aggregation step (step S21), a regression equation definition step (step S22), a coefficient calculation step (step S23), and a manufacturing condition extraction step ( Step S24), a target value calculation step (Step S25), and a manufacturing condition change step (Step S26).

<Step S21>
In step S21, manufacturing condition data or quality data in which individual products are associated with manufacturing conditions and quality realization values are created, and a set of manufacturing condition data or quality data is created. Since step S21 is the same as step S11, description thereof is omitted.

<Step S22>
Step S22 is a step of defining the quality of the product according to the manufacturing conditions by a linear regression equation. Since step S22 is the same as step S12, description thereof is omitted.

<Step S23>
Step S23 is a step of calculating the coefficient of the linear regression equation based on the quality measurement result and / or the production result data. Since step S23 is the same as step S13, description thereof is omitted.

<Step S24>
In step S24, all of the explanatory variables in the quality model are not changeable in the manufacturing apparatus, or even if they can be changed, the influence on the quality may be small and the influence on the quality may be changed. This is a process aimed at selecting a variable (manufacturing condition) having a large value and specifying an optimum target value using the selected manufacturing condition. In step S24, a variable having a large influence on quality is selected based on a statistic called deviance in the generalized linear model. Debiansu d _v, the optimum regression coefficient calculated production condition data and quality data in the maximum likelihood method

を用いた最大尤度を基に、下記式１６で定義される。ただし、下記式１６では、記号が煩雑になるため欠点数における対象量や不良品数における対象個数は省略している。

Based on the maximum likelihood using However, in the following Expression 16, since the symbols are complicated, the target amount in the number of defects and the target number in the number of defective products are omitted.

式１６より、デビアンスが小さいほど、回帰係数

From Equation 16, the smaller the deviance, the regression coefficient

を用いた品質モデルは、データに対して当てはまりが良いと言える。

It can be said that the quality model using is suitable for data.

  Deviance in all explanatory variables of the quality model

と、ｋ番目の説明変数を一つだけ除外した品質モデルのデビアンス

And the deviance of the quality model that excludes only one k-th explanatory variable.

とを用いて、デビアンス増分Δｄ_ｖｋを下記式１７で定義する。

And the deviance increment Δd _vk is defined by Equation 17 below.

式１７において、Ｘ_ｋはｋ番目の説明変数を一つだけ除外した製造条件データ集合を表し、
ｃ_ｋはＸ_ｋとＹから求めた一般化線形モデルの回帰係数である。デビアンス増分Δｄ_ｖｋは、除外した変数を説明変数に含めることで一般化線形モデルの当てはまりが改善される程度を表す指標である。したがって、デビアンス増分Δｄ_ｖｋが全ての説明変数の中で上位にある変数（Δｄ_ｖｋ）は、製品の品質の変動に与える影響が大きいとみなすことができる。このようにして選択した製造条件から製造装置の運転により操作可能なものを選んだものを操作可能変数と呼び、製品の製造条件から操作可能変数を除いた変数を外乱変数と呼ぶ。

In Equation 17, X _k represents a manufacturing condition data set excluding only one k-th explanatory variable,
c _k is a regression coefficient of the generalized linear model obtained from X _k and Y. The deviance increment Δd _vk is an index representing the degree to which the fit of the generalized linear model is improved by including the excluded variable in the explanatory variable. Therefore, it can be considered that the variable (Δd _vk ) in which the deviance increment Δd _vk is higher than all the explanatory variables has a large influence on the fluctuation of the product quality. A variable that can be operated by operating the manufacturing apparatus from the selected manufacturing conditions is referred to as an operable variable, and a variable obtained by removing the operable variable from the manufacturing conditions of the product is referred to as a disturbance variable.

When the linear regression equation S (x, c) in the quality model is expressed as the sum of the operable term S _C (x _C , c _C ) and the disturbance term S _D (x _D , c _D ), the following equation 18 is obtained. become.

ここで、操作可能項Ｓ_Ｃ(ｘ_Ｃ,ｃ_Ｃ)は、操作可能変数ｘ_Ｃとこれに乗ぜられる係数ｃ_Ｃ及び定数項からなり、下記式１９で表される。また、外乱項Ｓ_Ｄ(ｘ_Ｄ,ｃ_Ｄ)は、外乱変数ｘ_Ｄとこれに乗ぜられる係数ｃ_Ｄからなり、下記式２０で表される。

Here, the manipulable term S _C (x _C , c _C ) is composed of the manipulable variable x _C , a coefficient c _C multiplied by the variable, and a constant term, and is represented by the following Expression 19. The disturbance term S _D (x _D , c _D ) is composed of a disturbance variable x _D and a coefficient c _D multiplied by the disturbance variable x _D, and is expressed by the following equation 20.

<Step S25>
Step S25 is a step of calculating a target value of the manufacturing conditions based on the information obtained in the above steps S22 to S24. In step S25, the optimal target value of the operable variable (target value of the manufacturing condition) is calculated according to the following procedure.
(1) Average value of S _C (x _Cn , c _C ) in the product manufacturing condition data set

とＳ_Ｄ(ｘ_Ｄｎ,ｃ_Ｄ)の平均値

And the average value of S _D (x _Dn , c _D )

とを各々算出する。
（２）製品製造条件データ集合において、

Are calculated respectively.
(2) In the product manufacturing condition data set,

である製造条件データｘ_ｎを用いて、Ｓ_Ｃ(ｘ_Ｃｎ,ｃ_Ｃ)が小さく、製品品質が良好になる製造条件

Manufacturing conditions in which S _C (x _Cn , c _C ) is small and product quality is good using manufacturing condition data x _n

を算出する。ここで、εは予め定めた正の定数で、例えば、製造条件データ集合全体での｜Ｓ(ｘ,ｃ)｜の平均値

Is calculated. Here, ε is a predetermined positive constant, for example, the average value of | S (x, c) | over the entire manufacturing condition data set

の５％程度に設定することができる。製造条件

Can be set to about 5%. Manufacturing conditions

の算出方法の具体例としては、以下のものを挙げることができる。

As specific examples of the calculation method, the following can be cited.

Select _xn that minimizes S _C (x _Cn , c _C ),

を算出する方法。かかる方法で算出される製造条件の目標値

How to calculate Target value of manufacturing conditions calculated by this method

は、下記式２１で表すことができる。

Can be represented by the following formula 21.

  ・ Represented by Equation 21 above

と変更可能製造条件データｘ_Ｃｎとを用いて表される

And changeable manufacturing condition data x _Cn

の値が小さい順に予め定めたν_２個のｘ_Ｃｎに関する平均値

The average value for ν ₂ x _Cn determined in ascending order

を目標値とする方法。かかる方法で算出される製造条件の目標値

A method for setting the target value. Target value of manufacturing conditions calculated by this method

は、下記式２２で表すことができる。

Can be represented by the following formula 22.

  ・ Represented by Equation 21 above

及び上記式２２で表される

And represented by Formula 22 above

と、製造条件データ集合における平均値

And the average value in the production condition data set

とについて、成分毎に比較して、回帰係数と各条件項目との積が品質を改善する方向にある方を取り出して組み合わせた

Compared for each component, the product of the regression coefficient and each condition item is in the direction of improving quality and combined.

を製造条件の目標値とする方法。かかる方法で算出される製造条件の目標値

Is the target value of manufacturing conditions. Target value of manufacturing conditions calculated by this method

は、下記式２３で表すことができる。

Can be represented by the following Equation 23.

上記式２３において、

In Equation 23 above,

の場合（ｊ＝１、２）、

(J = 1, 2),

であり、

And

の場合（ｊ＝１、２）、

(J = 1, 2),

である。

It is.

<Step S26>
Step S26 is a step of changing the manufacturing conditions based on the target value calculated in step S25. The change mode of the manufacturing condition is not particularly limited as long as the manufacturing condition after the change is specified so that the manufacturing condition is closer to the target value than before the change, but for example, calculated in the above step S25. The form etc. which make the set target value the manufacturing conditions after a change can be mentioned.

  Thus, according to the control method concerning 2nd Embodiment provided with process S21-process S26, the target value of the manufacturing conditions which have big influence on quality is calculated | required, and manufacturing conditions are changed based on the calculated | required target value. To control the quality of the product. Therefore, according to the control method according to the second embodiment, it is possible to reliably improve the quality of the product, and thus it is possible to provide a product quality control method capable of stably obtaining a good quality product. can do.

2. Product quality control device 2.1. First Embodiment FIG. 3 is a conceptual diagram showing a configuration example of a product quality control apparatus (hereinafter referred to as “control apparatus according to a first embodiment”) according to the first embodiment of the present invention. As shown in FIG. 3, the control device 10 according to the first embodiment includes a regression equation defining unit 1, a coefficient calculating unit 2, a target value calculating unit 3, a manufacturing condition changing unit 4, a calculating unit 6, And a result display unit 7.

  Information relating to the manufacturing condition data and quality data input by the operator is sent to the calculation unit 6, and the operation of the above-described step S11 is performed based on the information. Information on the numerical values calculated by the calculation unit 6 is sent to the regression equation definition unit 1 and the coefficient calculation unit 2. Then, the regression equation defining unit 1 defines the linear regression equation by performing the operation of the step S12, and the coefficient calculating unit 2 performs the operation of the step S13 so that the coefficient of the linear regression equation is determined. Calculated. Information regarding the linear regression equation defined by the regression equation defining unit 1 and information regarding the coefficient of the linear regression equation identified by the coefficient calculating unit 2 are then sent to the target value calculating unit 3. Then, using these pieces of information, the target value calculation unit 3 calculates the target value of the manufacturing condition by performing the operation of the above-described step S14, and the information regarding the calculated target value is the manufacturing condition changing unit 4 And sent to the result display unit 7. In the manufacturing condition changing unit 4, the process S15 is performed based on the target value sent from the target value calculating unit 3, and the changed manufacturing condition is specified. Information about the changed manufacturing conditions is sent to the result display unit 7. The result display unit 7 displays information sent from the target value calculation unit 3 and the manufacturing condition change unit 4, and the information displayed on the result display unit 7 is confirmed by the operator.

  As described above, according to the control device 10, the control method according to the first embodiment can be performed. Therefore, according to the present invention, a product capable of stably obtaining a product of good quality can be obtained. A quality control device 10 can be provided. Note that the regression equation defining unit 1, the coefficient calculating unit 2, the target value calculating unit 3, and the manufacturing condition changing unit 4 in the control device 10 have their functions provided to a central processing unit (CPU) of a personal computer or a process computer. You can carry it.

2.2. Second Embodiment FIG. 4 is a conceptual diagram showing an example of a product quality control apparatus (hereinafter referred to as “control apparatus according to a second embodiment”) of the present invention according to a second embodiment. 4, components having the same configuration as in FIG. 3 are denoted by the same symbols as those used in FIG.
As shown in FIG. 4, the control device 20 according to the second embodiment includes a regression equation defining unit 1, a coefficient calculating unit 2, a target value calculating unit 3, a manufacturing condition changing unit 4, and a manufacturing condition extracting unit 5. And a calculation unit 6 and a result display unit 7.

  The calculation unit 6 performs the process S21, the regression equation definition unit 1 performs the process S22, and the coefficient calculation unit 2 performs the process S23. In the manufacturing condition extraction unit 5, the operation of the above-described step S <b> 24 is performed, and the manufacturing condition extraction unit 5 extracts manufacturing conditions that greatly affect the quality of the product. Information on the manufacturing conditions extracted by the manufacturing condition extracting unit 5 and information obtained by the regression equation defining unit 1 and the coefficient calculating unit 2 are sent to the target value calculating unit 3, and the target value calculating unit 3 The operation of step S25 is performed. By performing the operation of step S25, information on the target value calculated by the target value calculation unit 3 is sent to the manufacturing condition change unit 4 and the result display unit 7. In the manufacturing condition changing unit 4, the process S26 is performed based on the target value sent from the target value calculating unit 3, and the changed manufacturing condition is specified. Information about the changed manufacturing conditions is sent to the result display unit 7. The result display unit 7 displays information sent from the target value calculation unit 3 and the manufacturing condition change unit 4, and the information displayed on the result display unit 7 is confirmed by the operator.

  Thus, according to the control apparatus 20, since the control method concerning the said 2nd Embodiment can be implemented, according to this invention, the quality of product which can improve the quality of a product reliably is improved. A control device 20 can be provided. Note that the regression equation defining unit 1, the coefficient calculating unit 2, the target value calculating unit 3, the manufacturing condition changing unit 4, and the manufacturing condition extracting unit 5 in the control device 20 are a central processing unit (CPU) of a personal computer or a process computer. And so on.

  The present invention will be further described with reference to the results of the examples.

  FIG. 5 shows an example of the manufacturing process of the steel product taken up in this embodiment. Bloom slabs cast by a continuous casting machine after secondary refining are heated and slabbed in a heating furnace, and then steel strip products such as bar and wire rods are manufactured through the slabbing and strip rolling processes. .

  In the control apparatus in the present embodiment, the product surface defect rate is selected as a quality index, and a total of 14 items such as molten steel components, manufacturing conditions in continuous casting, and heating furnace temperature in block rolling are selected as manufacturing conditions. did. The calculation unit normalized each manufacturing condition so that the average was 0 and the variance 1 for each item, and the coefficient calculation unit calculated the coefficient of the linear regression equation of the quality model in this example. Table 1 also shows the coefficients of the linear regression equation of the quality model and the deviance increment in this example. FIG. 6 shows the relationship between the value of the linear regression equation and the estimated value of the defective product rate in the calculation of the coefficient of the linear regression equation using past manufacturing performance data. FIG. 6 is a graph plotting the relationship between the linear regression value S of the manufacturing conditions and the defective product incidence ρ as a result of determining the coefficients of the regression formula of Table 1 by the maximum likelihood method. Represents the estimated accuracy.

Figure 7 shows the relationship between the operational section S _C and the disturbance term S _D of the data shown in this embodiment. As shown in FIG. 7, the operational section S _C on the horizontal axis, taking each disturbance term S _D on the vertical axis and plotting the actual values for each lot of the product, a linear regression by the optimum conditions of the operational variables The reached value of the formula is easy to understand. A plurality of broken lines shown in FIG. 7 are contour lines of the quality index and represent the same value on the line. The solid line in the horizontal direction is the average value of the disturbance term. Further, S in FIG. 7 is S (x, c)
Means. Calculated based on Equation 21 above

を、図７の丸で囲んだ点で示し、

Is indicated by the circled points in FIG.

の近傍６個から上記式２２に基づいて算出した

Calculated based on the above formula 22 from 6 neighborhoods

を、図７の菱形で示す。

Is shown by the diamonds in FIG.

The production was carried out with reference to the optimum target value determined in this way. From the above equation 18, S (x, c) is expressed as the sum of the manipulable term and the disturbance term, and among manipulable variables constituting the manipulable term, the manipulable variable ((c), which has the greatest effect on the quality fluctuation) (Manufacturing conditions) is a manufacturing condition represented by _xC1 from Table 1. Therefore, in this example, the manufacturable variable x _C1 was changed to a manufacturing condition that was almost the target value, and the product was manufactured. FIG. 8 shows the relationship between the average value of target values for each operable variable (after standardization) and the average value of each operable variable (after standardization) under the changed manufacturing conditions. The results of the incidence are shown in FIG. As shown in FIG. 9, according to this example, the defective product generation rate, which was about 10%, could be reduced to less than 5%. Therefore, according to this invention, the quality of a product can be improved reliably.

It is a flowchart which shows the process with which the control method concerning 1st Embodiment is equipped. It is a flowchart which shows the process with which the control method concerning 2nd Embodiment is equipped. It is a conceptual diagram which shows the form example of the control apparatus concerning 1st Embodiment. It is a conceptual diagram which shows the form example of the control apparatus concerning 2nd Embodiment. It is a figure which shows the example of a manufacturing process of steel products. It is the figure which plotted the relationship between the linear regression value S of manufacturing conditions, and the defect-product incidence ρ as a result of determining the coefficient of the regression equation of Table 1 by the maximum likelihood method. It is a figure which shows the relationship between an operable term and a disturbance term. It is a figure which shows the average value of an operable variable. It is a figure which shows the result of the defective product incidence before and behind manufacture condition change.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 ... Regression formula definition part 2 ... Coefficient calculation part 3 ... Target value calculation part 4 ... Manufacturing condition change part 5 ... Manufacturing condition extraction part 6 ... Calculation part 7 ... Result display part 10, 20 ... Control apparatus of product quality

Claims (6)

A regression equation defining unit that defines a linear regression equation, a coefficient calculating unit that calculates a coefficient of the linear regression equation, a target value calculating unit that calculates a target value of manufacturing conditions, and the manufacturing condition based on the target value A manufacturing condition changing unit to change, and a method of controlling the quality of the product using a control device comprising :
A regression equation defining step of approximating the quality related to the count value of the product specified according to manufacturing conditions by a discrete probability distribution, and defining the linear regression equation in the regression equation defining unit ;
A coefficient calculation step of calculating the coefficient of the linear regression equation in the coefficient calculation unit using the actual result data of the quality and the actual result data of the manufacturing conditions;
A target value calculating step of calculating a target value of the manufacturing condition in the target value calculating unit using the coefficient of the linear regression equation calculated in the coefficient calculating step and the actual data of the manufacturing condition;
Based on the calculated target value of the manufacturing condition, the manufacturing condition changing step of changing the manufacturing condition in the manufacturing condition changing unit ;
A method for controlling product quality, comprising: The manufacturing conditions include changeable manufacturing conditions and non-changeable manufacturing conditions,
Furthermore, based on a statistic called deviance in the generalized linear model based on the discrete probability distribution , a manufacturing condition extraction step is provided for extracting a manufacturing condition having a large influence on the quality from the changeable manufacturing condition,
2. The method for controlling product quality according to claim 1, wherein, in the target value calculating step, the target value is calculated using the manufacturing conditions extracted in the manufacturing condition extracting step. The product quality control method according to claim 1, wherein the product is a steel material. A control device used to control the quality of the product,
A regression equation defining unit that approximates the quality related to the count value of the product specified according to manufacturing conditions by a discrete probability distribution, and defines a linear regression equation;
A coefficient calculating unit that calculates the coefficient of the linear regression equation using the actual result data of the quality and the actual result data of the manufacturing conditions;
A target value calculating unit that calculates a target value of the manufacturing condition using the coefficient of the linear regression equation calculated by the coefficient calculating unit and the actual data of the manufacturing condition;
A manufacturing condition changing unit that changes the manufacturing condition based on the calculated target value of the manufacturing condition;
A product quality control device comprising: The manufacturing conditions include changeable manufacturing conditions and non-changeable manufacturing conditions,
Furthermore, based on a statistic called deviance in the generalized linear model based on the discrete probability distribution , a manufacturing condition extraction unit is provided that extracts a manufacturing condition having a large influence on the quality from the changeable manufacturing condition,
5. The product quality control apparatus according to claim 4, wherein the target value calculation unit calculates the target value using the manufacturing conditions extracted by the manufacturing condition extraction unit. 6. The product quality control device according to claim 4, wherein the product is a steel material.

Images