JP2011085970A - Quality prediction device, quality prediction method, program and computer-readable recording medium - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a quality prediction device for achieving quality prediction by generating a model having sufficient precision with a few number of division by using an exponent smoothing method. <P>SOLUTION: The quality prediction device is provided with: a division pattern candidate creation part for creating a division pattern candidate for dividing an overall region configured of variables included in handling data into a plurality of local regions; an activity function calculation part for calculating activity functions about each division pattern candidate; a minimum error division pattern selection part for calculating a quality prediction value in an overall region by deriving a relational expression to calculate the quality prediction value in the overall region based on an exponent smoothing function and the activity function to calculate the current quality prediction value for each logical region about each division pattern candidate, and selecting a division pattern to minimize a prediction error between the quality prediction value and quality data; a learning error evaluation part for determining the convergence state of the prediction error; a quality prediction value output part for outputting the quality prediction value from the relational expression of the division pattern whose convergence is sufficient; and an exponent smoothing coefficient calculation part for calculating an exponent smoothing coefficient. <P>COPYRIGHT: (C)2011,JPO&INPIT

Description

  The present invention relates to a quality prediction apparatus, a quality prediction method, a program, and a computer-readable recording medium in a manufacturing process, and more specifically, by changing operating conditions in a production line such as a steel process, In the manufacturing process that creates products, a quality prediction device, a quality prediction method, which is used to create a quality prediction model from time-series quality data in which the quality information of products manufactured in the past is arranged in the manufacturing order and perform quality prediction, The present invention relates to a program and a computer-readable recording medium.

  In the manufacturing process of steel, semiconductors, etc., according to the order of the product from the customer, the operating conditions for manufacturing the product are determined from the required specifications for the customer's product, and the product is manufactured. In the middle of manufacturing, whether the product satisfies the required specifications is evaluated based on the operation results in the process (upper process) that has been passed so far and the scheduled operation conditions in the future pass process (lower process). In some cases, work for changing operation conditions in the lower process and work for prioritizing inspection before shipment may be performed to avoid outflow of defective products to customers. Therefore, it is very important to predict the quality of the product during the manufacturing process. For this reason, conventionally, an operation operator or a technical person in charge of a manufacturing process has performed an operation of estimating the quality of a product from operation conditions.

  As a general method of predicting quality (hereinafter referred to as “first method”), when manufacturing the same product, for example, as defined in the operation standard, the operation condition is made constant. Operation is performed. At this time, particularly when a product is manufactured under a severe environment such as a high temperature and a high load such as a steel process, the characteristics of equipment such as a mill constant of a rolling mill gradually change. For this reason, a situation in which the quality differs depending on the production timing even under exactly the same operating conditions. In this situation, the operation operator and the person in charge refer to the results of the quality of the product manufactured immediately before, and further consider the past experience and knowledge about the manufacturing process to improve the quality of the product currently being manufactured. We are working to predict.

  In addition, even under the same manufacturing conditions (that is, input conditions) as described above, the situation in which the quality (that is, output) changes at the manufacturing timing is not limited to the final quality, for example, in the hot rolling of a steel process. Control parameters such as the learning coefficient of the cooling control model in the hot rolling runout table for controlling the scraping temperature are also generated. For this reason, it is necessary to predict the suitability value with reference to the results of products manufactured just before.

  However, this first method has a problem that the prediction accuracy is greatly influenced by the knowledge and skill level of the operation operator and the person in charge. Furthermore, when many products are manufactured in a short time, there is a problem that the amount of information to be processed is large and high-speed processing is required, which cannot be handled by humans.

  Therefore, as a second method, Patent Document 1 proposes a method of making predictions with the aid of a computer using past quality data, instead of relying solely on human memory. The second method uses an exponential smoothing method that is widely used as a practical and simple method for predicting the quality of products currently being manufactured by analyzing accumulated time-series data. In the exponential smoothing method, it is assumed that the value of new data is affected by the past value, the strength of the effect is stronger as it is closer to the present, and weaker as it goes back in the past. And calculate the predicted value of the new data. The second method is a method of predicting quality by exponential smoothing method, wherein a smoothing constant is calculated from a theoretical optimal solution using N pieces of time series data, and is defined by a difference between a predicted value and an actual result. This is a technique that makes it possible to estimate the upper and lower limits of the predicted value based on the standard deviation of the prediction error.

  Here, since the quality data of the manufacturing process generally includes data on a plurality of products, in order to obtain sufficient accuracy, the data is divided under appropriate conditions, and exponential smoothing is performed for each division. There is a need to do. However, the method of Patent Document 1 does not consider the data division at all, and when this method is applied to the quality prediction of the manufacturing process, there is a problem that the prediction accuracy is greatly deteriorated.

  As a third method, in Patent Document 2, the smoothing coefficient should be set separately in estimating the friction coefficient between the rolling roll and the material to be rolled by the exponential smoothing method for the setup calculation in hot rolling finish rolling. A method is disclosed in which a person sets rolling condition categories and performs exponential smoothing for each category. This third method can ensure relatively high accuracy and is often used in actual processes. However, since the category is determined based on the experience of the person in charge, it depends on the skill level of the person in charge. The prediction accuracy also depends on the person in charge. In addition, for humans, multiple types of rolling conditions, such as sheet thickness and target scraping temperature, are determined by focusing on one of them and mechanically combining the multiple rolling conditions to create a category. To do. For this reason, there is a problem that a segment in which almost no data exists is generated and accuracy is deteriorated. In addition, since the division is easily subdivided, there is a practical problem that a maintenance load becomes high when accuracy is deteriorated.

JP 2004-110300 A JP-A-9-29316

Keiyuki Sakamoto et al. “Information Statistics”, Kyoritsu Publishing Co., Ltd. (1983), 43 pages

  Therefore, the present invention has been made in view of the above problems, and an object of the present invention is to use an exponential smoothing method based on time-series quality data in which quality information of products manufactured in the past are arranged in manufacturing order. It is an object of the present invention to provide a new and improved quality prediction apparatus, quality prediction method, program, and computer-readable recording medium having sufficient accuracy and reduced maintenance load in a method for performing quality prediction. .

  In order to solve the above-described problems, according to an aspect of the present invention, there is provided a quality prediction apparatus in a manufacturing process, and an operation corresponding to a predetermined selection condition from a database storing operation data and quality data related to a plurality of products. A data extraction unit that extracts data and quality data, and for the extracted operation data, an area that the variable included in the operation data takes a value as an entire area, and a division pattern candidate that divides the entire area into a plurality of local areas Activity to calculate an activity function that represents a contribution ratio that is a ratio that a quality prediction value in each local region contributes to a prediction value in the entire region for each division pattern candidate For the function calculation unit and each divided pattern candidate, the current quality data and the past quality prediction value for each local area are Based on the exponential smoothing function set based on the exponential smoothing method for calculating the quality prediction value and the activity function, a relational expression for calculating the quality prediction value in the entire region is derived, and the relational expression and the past Calculates the quality prediction value for the entire area based on the quality data and the past quality prediction value, and selects the division pattern that minimizes the prediction error that is the difference between the quality prediction value and the quality data for the entire area. Whether or not the convergence of the prediction error is sufficient based on the comparison result between the quality prediction value of the division pattern selected by the division pattern selection unit and the minimum error division pattern selection unit and a preset evaluation reference value A learning error evaluation unit for determining the quality, a quality prediction value output unit for outputting a quality prediction value using a relational expression of division patterns determined to have sufficient convergence by the learning error evaluation unit, and exponential smoothing In a few, characterized in that it comprises, and exponential smoothing coefficient calculation unit for calculating an exponential smoothing coefficient that determines the strength of the effect of past quality prediction for the current quality prediction value, quality prediction apparatus is provided.

  According to the present invention, a constant pattern model that creates divided pattern candidates from the entire region and calculates the current quality prediction value from the past quality data and the past quality prediction value for each local region for each divided pattern candidate. Based on an exponential smoothing function and an activity function, a relational expression for calculating a quality prediction value in the entire region is derived. Then, based on the derived relational expression, past quality data, and past quality prediction value, a division pattern that minimizes the quality prediction error is selected. Then, it is determined whether or not the selected divided pattern is a model that can explain data with sufficient accuracy. When the model is constructed, a quality prediction value is output using the relational expression of the divided pattern. As a result, it is possible to prevent the occurrence of a local region in which almost no data exists, and it is possible to automatically search for and obtain a model that can explain data with sufficient accuracy with an appropriate minimum number of divisions. Further, since the area is not subdivided more than necessary, it is possible to reduce the maintenance load when the accuracy is deteriorated.

  The exponential smoothing coefficient calculating unit can also perform exponential smoothing calculation for a plurality of exponential smoothing coefficients and select an exponential smoothing coefficient with the smallest prediction error calculated from the calculation result.

  Alternatively, the learning error evaluation unit further includes a division exponent smoothing model determination unit that selects one division pattern from a plurality of division patterns determined to have sufficient convergence of the prediction error, and the exponential smoothing coefficient calculation unit includes a plurality of division factors An exponential smoothing coefficient is set, and for each exponential smoothing coefficient, a division pattern candidate is created by a division pattern candidate creation unit, and a prediction error is converged by an activity function calculation unit, a minimum error division pattern selection unit, and a learning error evaluation unit. May be determined, and the division exponent smoothing model determination unit may select a division pattern having the smallest number of divisions among the division patterns determined for each exponential smoothing coefficient.

  In addition, the division pattern candidate creation unit includes a numerical division creation unit that divides an operation variable space composed of operation variables and a code division creation unit that divides a group composed of code variables, and is created by the numerical division creation unit. The division pattern candidate created by the division pattern candidate and the code division creation unit can be used as a division pattern candidate for the entire region.

  The activity function can be a function configured by combining a normal distribution function having the center at the center of gravity of the local region and a binary function calculated from the code division information. In addition, the code division creation unit may divide a group of code variables using a search method.

  Alternatively, the division pattern candidate creation unit can collectively generate division pattern candidates including operation variables and code variables using an optimization method.

  In addition, the manufacturing process is a steel process. At this time, the quality data includes product surface wrinkles, mechanical strength characteristic values, shape flatness, product size, internal stress, or process values that affect these quality. can do.

  The manufacturing process is a hot rolling process of steel sheets, which are steel products. When the quality predicted by the quality prediction device is the scraping temperature at the hot-rolling runout table, the operating variables of the manufacturing process are , C amount, Si amount, Mn amount, P amount, S amount, Cu amount, Ni amount, Cr amount, Mo amount, Nb amount, V amount, Ti in the molten steel at the time when the material to be rolled has finished the refining process Amount, B amount, Al amount, N amount, O amount, Ca amount, target plate width, target plate thickness, equivalent carbon amount, finish rolling target temperature, rolling target temperature, rolling speed, cooling water At least one or more of water density, cooling water temperature, and material code may be selected.

  In order to solve the above-mentioned problem, according to another aspect of the present invention, there is provided a quality prediction method in a manufacturing process, from a database storing operation data and quality data related to a plurality of products, to a predetermined selection condition. A step of extracting corresponding operation data and quality data, and for the extracted operation data, an area taken as a value by a variable included in the operation data is defined as an entire area, and the entire area is divided into a plurality of local areas. A step of creating a plurality of candidates, a step of calculating an activity function representing a contribution rate that is a ratio of a quality predicted value in each local region to a predicted value in the entire region, for each divided pattern candidate; Exponential smoothing to calculate the current quality prediction value for each local region from time series data of quality prediction values for the division pattern candidates Calculating an exponential smoothing coefficient that determines the strength of the influence of the past quality prediction on the current quality prediction value based on the quality data corresponding to each local region, and for each divided pattern candidate, Based on an exponential smoothing function in which an exponential smoothing coefficient is set for an area and an activity function, a relational expression for calculating a quality prediction value in the entire area is derived, and the relational expression, past quality data, and past quality are calculated. Calculating a quality prediction value in the entire area based on the prediction value, selecting a division pattern that minimizes a prediction error that is a difference between the quality prediction value and the quality data in the entire area, and the selected division pattern A step of determining whether or not the convergence of the prediction error is sufficient based on a comparison result between the quality predicted value of the image and a preset evaluation reference value, and the convergence of the prediction error is sufficient. And a step of outputting a quality prediction value using a relational expression of the division pattern determined to be sufficiently converged by the step of determining whether or not the quality prediction method is provided. Is done.

  Furthermore, in order to solve the above-mentioned problem, according to another aspect of the present invention, there is provided a quality prediction method in a manufacturing process, wherein a predetermined selection condition is obtained from a database storing operation data and quality data related to a plurality of products. Extracting relevant operation data and quality data, and past quality prediction for current quality prediction value of exponential smoothing function to calculate current quality prediction value for each local region from time series data of quality prediction value A step of setting a plurality of exponential smoothing coefficients that determine the strength of the influence of the operation, and for the extracted operation data, the region that the variable included in the operation data takes as a whole region, and the whole region is a plurality of local regions A plurality of division pattern candidates and a quality prediction value in each local region for each division pattern candidate. A step of calculating an activity function representing a contribution ratio, which is a ratio contributing to a predicted value in the body region, and an exponential smoothing function in which one of the exponential smoothing coefficients for each local region is set for each divided pattern candidate And a relational expression for calculating a quality prediction value in the whole area based on the activity function, and a quality prediction value in the whole area based on the relational expression, past quality data, and a past quality prediction value. Calculating a division pattern that minimizes a prediction error that is a difference between the quality prediction value and the quality data in the entire region, a quality prediction value of the selected division pattern, and a preset evaluation criterion A step of determining whether or not the convergence of the prediction error is sufficient based on the comparison result with the value, and determining whether or not the convergence of the prediction error is sufficient for each exponential smoothing function A step of selecting a division pattern having the smallest number of divisions from division patterns determined to have sufficient convergence by the step, and a step of outputting a quality prediction value using a relational expression of the division pattern having the smallest number of divisions. The quality prediction method characterized by including these is provided.

  Moreover, in order to solve the said subject, according to another viewpoint of this invention, the program for making a computer function as said quality prediction apparatus is provided. Such a program is stored in a storage device included in the computer, and read and executed by a CPU included in the computer, thereby causing the computer to function as the quality prediction device. A computer-readable recording medium on which the program is recorded is also provided. The recording medium is, for example, a magnetic disk or an optical disk.

  As described above, according to the present invention, in the method of performing quality prediction by exponential smoothing from time-series quality data in which the quality information of products manufactured in the past is arranged in the manufacturing order, the method has sufficient accuracy, In addition, it is possible to provide a quality prediction apparatus, a quality prediction method, a program, and a computer-readable recording medium with reduced maintenance load.

It is a block diagram which shows the function structure of the quality prediction apparatus concerning the 1st Embodiment of this invention. It is a flowchart which shows the quality prediction process by the quality prediction apparatus concerning the embodiment. It is explanatory drawing which shows the production | generation process of the division pattern candidate by the division pattern candidate creation part concerning the embodiment. It is explanatory drawing which shows typically the procedure which produces a division | segmentation pattern candidate in a two-dimensional operation variable space. It is explanatory drawing which shows typically the procedure which produces a division | segmentation pattern candidate in the two-dimensional operation variable space divided into three. It is a flowchart which shows the quality prediction method by the quality prediction apparatus concerning the 2nd Embodiment of this invention. It is a block diagram which shows the function structure of the quality prediction apparatus concerning the 3rd Embodiment of this invention. It is explanatory drawing which shows the production | generation process of the division pattern candidate by the division pattern candidate creation part concerning the embodiment. It is a flowchart which shows the production | generation process of the division | segmentation pattern candidate of a code variable. It is explanatory drawing which shows the code replacement process in the preparation process of the division | segmentation pattern candidate of a code variable. It is a block diagram which shows one structural example of the hardware constitutions of the quality prediction apparatus concerning embodiment of this invention. It is explanatory drawing which shows the outline of the hot rolling process of a thin steel plate. It is a graph which shows the relationship between the number of division | segmentation and a prediction error at the time of producing a division | segmentation pattern using an operation variable and a code variable. The histogram of the error of the prediction result at the time of applying a conventional method is shown. The histogram of the error of the prediction result at the time of applying the quality prediction apparatus concerning embodiment of this invention, Comprising: A division | segmentation pattern is produced without using a code variable is shown. The histogram of the error of the prediction result at the time of applying the quality prediction apparatus concerning embodiment of this invention, and producing a division | segmentation pattern using a code variable is shown.

  Exemplary embodiments of the present invention will be described below in detail with reference to the accompanying drawings. In addition, in this specification and drawing, about the component which has the substantially same function structure, duplication description is abbreviate | omitted by attaching | subjecting the same code | symbol.

<1. First Embodiment>
[Configuration of quality prediction device]
First, based on FIG. 1, the structure of the quality prediction apparatus concerning the 1st Embodiment of this invention is demonstrated. FIG. 1 is a block diagram illustrating a functional configuration of the quality prediction apparatus 100 according to the present embodiment.

  The quality prediction apparatus 100 in the manufacturing process according to the present embodiment is an apparatus that predicts the quality of a target for which quality prediction is performed. As the quality predicted by the quality prediction apparatus 100, for example, in the case of a steel process, mechanical strength characteristic values such as the number of surface defects and internal defects, tensile strength, yield stress, and elongation rate of various products such as thin plates and thick plates. There are various quality indicators of the final product directly requested by customers, such as shape flatness such as wave height, product size such as plate thickness, plate width and plate length, and internal stress. In addition, process values that affect these final qualities are also predicted qualities.

  As shown in FIG. 1, the quality prediction apparatus 100 includes a data extraction unit 110, a division pattern candidate creation unit 120, an activity function calculation unit 130, a minimum error division pattern selection unit 140, an exponential smoothing coefficient. The calculation unit 150, the learning error evaluation unit 160, the division index smoothing model determination unit 170, the quality prediction value output unit 180, and the database 190 are included.

  The data extraction unit 110 obtains, from the database 190, a plurality of operation data and quality data corresponding to a predetermined selection condition from the database 190, based on a selection condition relating to a quality prediction target input from the outside, such as code information indicating a product type. Extract. Here, the operation data includes an operation variable whose possible value is a numerical value and a code variable whose possible value is character code information. In this embodiment, a division pattern is created using only operation variables in the operation data. For example, in the steel process, the operational variables include the amounts of various elements of molten steel measured in the refining process, the amount of molten steel surface fluctuation and casting speed in the continuous casting process, and the rolling load and scraping temperature in the hot rolling process. It is. The data extraction unit 110 outputs the data extracted from the database 190 to the division pattern candidate creation unit 120.

  The division pattern candidate creation unit 120 performs a process of creating a plurality of division pattern candidates that are divided into a plurality of M local regions from the whole region, using operation variables whose possible values are numerical values as a whole region. The division pattern candidate creation unit 120 outputs, to the activity function calculation unit 130, division coordinate information of a division pattern obtained by dividing the operation variable space into a plurality of local regions for the plurality of division pattern candidates created. The division pattern candidate creation unit 120 creates division pattern candidates before the exponential smoothing coefficient used when calculating the predicted value for the local area of each division pattern is set by the exponential smoothing coefficient calculation unit 150. Alternatively, the division pattern candidates may be created after the exponential smoothing coefficient is set (first embodiment) (second embodiment described later).

  The activity function calculation unit 130 calculates an activity function for all the division pattern candidates created by the division pattern candidate creation unit 120 based on the division coordinate information. The activity function is used to represent a relational expression representing the relationship between the operation data and the quality data.

  The minimum error division pattern selection unit 140 calculates a quality prediction value in each local region for the division pattern candidate, calculates a prediction error based on the quality prediction value, and selects a division pattern with the smallest error. The minimum error division pattern selection unit 140 of the present embodiment uses an exponential smoothing function based on an exponential smoothing method as a constant model for calculating a current quality prediction value from past quality data and a past quality prediction value, and uses a quality prediction value. Is calculated. As described in paragraph 0006, the exponential smoothing method is a method of predicting a current state by analyzing accumulated time series data. In the present embodiment, it is possible to select an appropriately divided pattern by calculating a quality prediction value using an exponential smoothing method for each divided pattern.

  When the minimum error division pattern selection unit 140 calculates the predicted value of each local region using the exponential smoothing method, an exponential smoothing coefficient for weighting past data is calculated by the exponential smoothing coefficient calculation unit 150. For example, the exponential smoothing coefficient calculation unit 150 sets a plurality of values from values in the range of 0 to 1 that can be taken by the exponential smoothing coefficient, performs exponential smoothing simulation for the set values, and determines the most predictive error from the simulation result. By adopting a small value as the exponential smoothing coefficient, the exponential smoothing coefficient can be determined.

  The exponential smoothing simulation uses quality time-series data [y1, y2,..., YN] and operation data [x1, x2,. This means calculating the quality prediction value by a simple procedure. Note that the operation data xi includes an operation variable u and a code variable v.

  First, M activity functions are calculated for the division pattern whose accuracy is to be evaluated, and the constant model initial value y (y hat) is calculated for each local region. As an initial value, there is a method of using a simple average value of quality data belonging to the local region. Next, a plurality of exponential smoothing coefficient values are set in a range of 0 to 1.

Next, one is extracted from a plurality of exponential smoothing coefficients set in the range of 0 to 1, and the following processes (A) to (F) are performed, and evaluation by simulation is started.
(A) The contribution rate is calculated from the activity function with the first time series data x1, and the quality predicted value y ^ 1 is calculated using Equation (2) described later.
(B) Using the contribution rate calculated from the quality data y1 and the quality prediction value y ^ 1 and further from the operation data x1, the constant model is updated by Equation (7) described later.
(C) For the second time series data x2, a contribution rate and a quality prediction value y ^ 2 are calculated.
(D) The constant model is updated using the contribution rate calculated from the quality data y2, the quality prediction value y ^ 2, and the operation x2.
(E) This is repeated up to data N.
(F) A prediction error is calculated from N quality prediction values and quality data.
(G) The processing of (A) to (F) is performed for another exponential smoothing coefficient set in the preparation stage.
(H) The above (A) to (G) are performed for all exponential smoothing coefficients.

  Or the exponential smoothing coefficient calculation part 150 can also determine an exponential smoothing coefficient using the theoretical method described in patent document 1, for example. When the minimum error division pattern selection unit 140 calculates the prediction value of each local region using the exponential smoothing coefficient calculated by the exponential smoothing coefficient calculation unit 150, the minimum error division pattern selection unit 140 calculates the prediction error and selects the division pattern with the smallest error. .

  The learning error evaluation unit 160 compares the prediction error of the division pattern selected by the minimum error division pattern selection unit 140 with a preset evaluation reference value to determine whether or not a relational expression having sufficient accuracy has been constructed. Determine. When the exponential smoothing coefficient is set after creating the division pattern candidate as in the present embodiment, the learning error evaluation unit 160 uses the information for expressing the obtained division pattern when it is determined that it has converged. A coefficient of a certain activity function is extracted and output to the quality predicted value output unit 180. On the other hand, when it is determined that the convergence is insufficient, the learning error evaluation unit 160 further divides the division pattern selected by the minimum error division pattern selection unit 140 into a new division pattern candidate creation unit 120 and creates a new one. An instruction to create a division pattern candidate is given.

  Or when using the quality prediction method concerning 2nd Embodiment mentioned later, an exponential smoothing coefficient is set before creating a division | segmentation pattern candidate. In such a case, when it is determined that the learning error has been converged, the learning error evaluation unit 160 extracts a coefficient of the activity function that is information for expressing the division pattern obtained for the currently set exponential smoothing coefficient. And output to the division exponent smoothing model determination unit 170. On the other hand, when it is determined that the convergence is insufficient, the learning error evaluation unit 160 instructs the division pattern candidate creation unit 120 to provide the minimum error division pattern, as in the process of the present embodiment described in paragraph 0035. The division unit selected by the selection unit 140 is further divided to instruct to create a new division pattern candidate.

  The division exponential smoothing model determination unit 170 selects the model of the exponential smoothing coefficient with the smallest number of divisions from the division pattern obtained for the exponential smoothing coefficient when using the quality prediction method according to the second embodiment described later. And decide. That is, the division exponent smoothing model determination unit 170 selects a division pattern represented by a relational expression that can explain the data with sufficient accuracy with the smallest number of divisions. The division exponent smoothing model determination unit 170 extracts the coefficient of the activity function, which is information for expressing the determined division pattern, and outputs the coefficient to the quality prediction value output unit 180.

  The quality prediction value output unit 180 includes information on the division pattern input from the learning error evaluation unit 160 or the division exponent smoothing model determination unit 170, separately input operation data, and the latest quality known at the time of prediction. Based on the data, a quality prediction value is calculated from the relational expression between the operation data and the quality data. The quality predicted value output unit 180 outputs the quality predicted value to the outside for use as guidance to the operator or as an input signal to the process control system.

  The database 190 includes past operation data and quality data in the manufacturing process, code information indicating the type of product, manufacturing time in each manufacturing process, order information such as a destination, a product number for specifying the product, and the like. Is stored in association with each other. These pieces of information are input and recorded in the database 190 from outside at a predetermined timing. The information stored in the database 190 is extracted by the database extraction unit 110 and used for quality prediction as described in paragraph 0030.

[Quality prediction processing by the quality prediction device]
Next, based on FIGS. 2-5, the quality prediction process by the quality prediction apparatus 100 concerning this embodiment is demonstrated in detail. In addition, FIG. 2 is a flowchart which shows the quality prediction process by the quality prediction apparatus 100 concerning this embodiment. FIG. 3 is an explanatory diagram showing a process of creating a division pattern candidate by the division pattern candidate creation unit 120. FIG. 4 is an explanatory diagram schematically showing a procedure for creating candidate division patterns in a two-dimensional operation variable space. FIG. 5 is an explanatory diagram schematically showing a procedure for creating a division pattern candidate in a three-dimensional two-dimensional operation variable space.

  Now, let us assume that a quality index representing the degree of quality by a numerical value is represented by a variable y. Further, it is assumed that an operation variable whose value can take a numerical value among the operation conditions when the product is manufactured is represented by a column vector composed of the following formula (1). Note that t in Equation (1) represents transposition of a matrix. Further, in the prediction data used for quality prediction in the present embodiment, information for identifying the manufacturing order, for example, manufacturing time, etc. is attached to each quality data.

・・・数式（１）

... Formula (1)

  In this embodiment, the whole area | region which consists of the operation variable of operation data is divided | segmented into a some local area, and the process which calculates a quality predicted value in each local area is performed. Then, the entire operation data is calculated by the following formula (2), which is the sum of products of the predicted value of the local region and the activity function Φi representing the ratio of the predicted value of the local region to the predicted value in the entire region. And a relational expression y between the quality data. Here, Σ represents the sum of terms, and M represents the number of local regions.

・・・数式（２）

... Formula (2)

  In the present embodiment, the predicted value in the local region is represented as a constant model using an exponential smoothing method. When such a constant model is used, for example, compared to the case where the predicted value is expressed by a linear polynomial, the convergence until obtaining a desired division pattern with sufficient accuracy is inferior, but the desired model is surely obtained. Can be earned. In addition, when a constant model is used, conventionally, it has been necessary to perform many area divisions until a desired division pattern is obtained. However, by using the quality prediction apparatus according to the present embodiment, the number of divisions can be reduced. A desired model can be acquired. The activity function is set in advance so as to satisfy the normal condition of Expression (3) in which the sum of the function values is 1 at an arbitrary position in the entire space.

・・・数式（３）

... Formula (3)

  In the quality prediction process performed by the quality prediction apparatus 100 according to the present embodiment, first, operation data and quality data relating to a target for which quality prediction is performed are extracted from the database 190 by the data extraction unit 110 (step S100). The data extraction unit 110 obtains, from the database 190, a plurality of operation data and quality data corresponding to a predetermined selection condition from the database 190, based on a selection condition relating to a quality prediction target input from the outside, such as code information indicating a product type. Extract. The data extraction unit 110 outputs the extracted operation data and quality data to the division pattern candidate creation unit 120.

  Next, the division pattern candidate creation unit 120 creates a division pattern candidate by dividing the whole area into a plurality of local areas, with the area taken as the value of the operation variable of the operation data as the whole area (step S102). That is, first, as shown in FIG. 3, the division pattern candidate creation unit 120 performs a process of dividing the space into two local regions (M = 2) for the two-dimensional operation variable space that is not divided. Here, in the division of the operation variable space, a division point is set on an axis parallel to each operation variable. That is, as shown in FIG. 4, the two-dimensional operation variable space includes two operation variables u1 and u2. Therefore, a division parallel to the operation variable u1 axis and a division parallel to the operation variable u2 axis are performed on the area 1-1. In this way, the area 1-1 is divided into two local areas 2-1 and 2-2. The division pattern candidate creation unit 120 changes the setting of the division points and creates a plurality of division pattern candidates as shown in FIG.

  In addition, when the operation variable space is already divided into several local regions, the division pattern candidate creation unit 120 uses the following equation (4), which is a weighted error evaluation function based on the activity function, for each local relational expression. The error is calculated, and the local region having the largest error is divided into two. As an example, FIG. 5 shows a procedure for dividing an operation variable space that has already been divided into three into four (M = 4). As shown in FIG. 5, if the region with the largest error is the region 3-2, the dividing points are set so that the region 3-2 is divided into two by an axis parallel to the operation variable u1 or the u2 axis. At this time, the remaining areas 3-1 and 3-3 are not divided.

・・・数式（４）

... Formula (4)

  In FIG. 3, when the operation variable space is divided into two, if the region with the largest error is the region 2-1, the region 2-1 is divided into two by the axis parallel to the operation variable u1 or u2 axis. A point is set. As described above, for the operation variable, a plurality of new division pattern candidates are generated from the current division pattern.

  The value of the division candidate point necessary for creating the operation variable division pattern candidate is, for example, extracting data of one operation variable, dividing this data into a plurality of groups, and operating variables serving as boundaries between the groups. There is a method in which the value of is calculated, calculated for all the operation variables, and used for the division candidate points. Specifically, the operation variable data is divided into a plurality of groups using, for example, a clustering method, and the minimum value and the maximum value of the operation variable values included in each group are calculated. Then, among the adjacent groups, the average value of the maximum value of the group having the smaller value of the operation variable and the minimum value of the group having the larger value of the operation variable is set as the value of the dividing point. Alternatively, if the operation operator or the person in charge can set the range of groups that can be regarded as the same operation level for the data value of the operation variable, the division candidate points set manually may be used.

When division pattern candidates are created in step S102, activity functions are calculated for all division pattern candidates (step S104). As a specific activity function, in the present embodiment, a p-dimensional normal distribution function μ _i shown in the following formula (5) is substituted into the following formula (6) as a function for smoothly connecting the boundaries between adjacent local regions. The normal membership function obtained as above is defined as the activity function. Here, c ⁱ _j represents the barycentric point of the local region, and σ ⁱ _j represents the standard deviation of the normal distribution function, and the barycentric point and the standard deviation of the normal distribution function are coefficients in the normal membership function. The regular membership function of Equation (6) has a value in the range of 0 to 1 depending on the value of the operation variable.

・・・数式（５）

・・・数式（６）
（ｉ＝１，・・・，Ｍ）、（ｊ＝１，・・・，Ｍ）

... Formula (5)

... Formula (6)
(I = 1,..., M), (j = 1,..., M)

  Thereafter, the smallest error division pattern selection unit 140 selects a division pattern having the smallest prediction error from the division pattern candidates (step S106). First, the minimum error division pattern selection unit 140 extracts one division pattern candidate from the plurality of division pattern candidates created by the division pattern candidate creation unit 120, and uses the division coordinate information of the division pattern to obtain individual data. Based on the value of the operation data, it is divided into M groups. Next, the minimum error division pattern selection unit 140 calculates a predicted value of each local region for each of the divided M groups using an exponential smoothing function expressed by the following mathematical formula (7).

Predicted value [i] = predicted value [i-1] + exponential smoothing coefficient × activity function × (actual value [i-1] −predicted value [i-1])
... Formula (7)

  Here, [i] is the value of the current step, and [i−1] is the value at the time of the previous prediction. Thus, the exponential smoothing function corrects the predicted value at the previous prediction with a correction value obtained by weighting the error at the previous prediction with a predetermined weight, and calculates the predicted value at the current step. By using the exponential smoothing method, the accumulated time series data can be analyzed and the quality prediction of the current step can be easily performed. In this embodiment, since the value of the constant model of each local region is synthesized using the activity function and prediction is performed, the constant model is also updated in accordance with the activity function when updated by the exponential smoothing method. . In Equation (7), the constant model (= predicted value) of the local region having a large activity function is greatly corrected according to the error. The exponential smoothing coefficient of Expression (7) is calculated by the exponential smoothing coefficient calculating unit 150. As a method of calculating the exponential smoothing coefficient by the exponential smoothing coefficient calculator 150, the method described in paragraph 0034 can be used.

  Then, the minimum error division pattern selection unit 140 uses the prediction value of each local region calculated by Expression (7) to calculate the overall prediction value by Expression (2) that is a relational expression between the operation data and the quality data. To do. Thereafter, the minimum error division pattern selection unit 140 calculates a prediction error, which is an error between the overall prediction value calculated by Equation (2) and the quality data, using the error evaluation equation shown in Equation (8) below. The calculation of the prediction error represented by Equation (8) is performed for all the divided patterns.

・・・数式（８）

... Formula (8)

  The minimum error division pattern selection unit 140 compares the prediction errors calculated for each division pattern, and selects the division pattern with the smallest error. As a result, the division pattern having the smallest error defined by the equation (8) is selected from the plurality of division pattern candidates created by the division pattern candidate creation unit 120.

  Next, the learning error evaluation unit 160 compares the error of the minimum error relational expression selected by the minimum error division pattern selection unit 140 with a preset evaluation reference value to determine whether learning is sufficient. It is determined whether or not the error has converged (step S108). As a method for determining convergence, for example, a method for comparing the error of the relational expression with a convergence determination factor (evaluation reference value), and a comparison of the change amount of the relational expression error with respect to the increment of the local region division with the convergence determination factor (evaluation reference value). And an index that takes into account the number of divisions and errors, for example, an information amount index of Akaike described in Non-Patent Document 1, and an index that adds not only the learning error but also the number of local regions to the evaluation. For example, a method of aborting the division when the index increases is used.

  The learning error evaluation unit 160 evaluates the error using such a convergence determination method, and if the error is larger than the evaluation reference value, a relational expression that can explain the data with sufficient accuracy has not yet been constructed. Therefore, it is determined that the error convergence is insufficient. In this case, the learning error evaluation unit 160 instructs the division pattern candidate creation unit 120 to increase the division number M of the division pattern by one and perform the process of step S102 (step S110). Then, the processes in steps S102 to S110 are repeated until it is determined in step S108 that a relational expression that can explain the data with sufficient accuracy is constructed.

  On the other hand, if the error is equal to or smaller than the evaluation reference value in step S108, the learning error evaluation unit 160 determines that a relational expression that can explain the data with sufficient accuracy is constructed, and expresses the obtained relational expression. The coefficient of the activity function, which is information on the above, is extracted and output to the quality predicted value output unit 180. The quality prediction value output unit 180 uses the information on the input relational expression, the operation data that is sequentially input, and the latest quality information that is known at the time of prediction, to calculate the quality prediction value from Equation (2). Calculate and output to the outside (step S112). The quality prediction value calculated by the quality prediction value output unit 180 can be used, for example, as guidance to the quality prediction operator or as an input signal to the process control system.

  The quality prediction device 100 according to the first embodiment of the present invention and the quality prediction method using the same have been described above. According to the present embodiment, a division pattern candidate is created from the entire region composed of a plurality of operation variables, and the activity function of each division pattern and the plurality of local regions of each division pattern calculated using the exponential smoothing method Based on the predicted value, the division pattern of the relational expression that minimizes the quality prediction error is selected. Then, it is determined whether or not the selected division pattern is a model that can explain data with sufficient accuracy, and the number of divisions of the division pattern is increased until the model is constructed. As a result, it is possible to prevent the occurrence of a local region in which almost no data exists, and it is possible to automatically search for and obtain a model that can explain data with sufficient accuracy with an appropriate minimum number of divisions. Further, since the area is not subdivided more than necessary, it is possible to reduce the maintenance load when the accuracy is deteriorated.

<2. Second Embodiment>
Next, based on FIG. 6, the quality prediction method by the quality prediction apparatus concerning the 2nd Embodiment of this invention is demonstrated. FIG. 6 is a flowchart showing a quality prediction method by the quality prediction apparatus according to the present embodiment. The quality prediction apparatus according to the present embodiment can be configured similarly to the quality prediction apparatus 100 of the first embodiment shown in FIG. For this reason, also in the description of the quality prediction apparatus of this embodiment below, it demonstrates using the code | symbol same as FIG. The quality prediction method by the quality prediction apparatus 100 of this embodiment is different from the method of the first embodiment in that an exponential smoothing coefficient is set first. In the following, the quality prediction method performed by the quality prediction apparatus 100 according to the present embodiment will be described with emphasis on processing different from that of the first embodiment.

  In the quality prediction processing by the quality prediction apparatus 100 according to the present embodiment, first, operation data and quality data relating to a target for which quality prediction is performed are extracted from the database 190 by the data extraction unit 110 (step S200). The data extraction unit 110 extracts a plurality of operation data and quality data corresponding to a predetermined selection condition from the database 190 based on a selection condition related to a target for which quality prediction is performed. The data extraction unit 110 outputs the extracted operation data and quality data to the division pattern candidate creation unit 120.

  Next, an exponential smoothing coefficient is set by the exponential smoothing coefficient calculation unit 150 (step S202). For example, the exponential smoothing coefficient calculation unit 150 can provide a plurality of values such as 0.1, 0.2,..., 0.9, for example, for exponential smoothing coefficients that can take values in the range of 0 to 1. Set. Then, the exponential smoothing coefficient calculation unit 150 extracts one from the set plurality of values and outputs it to the division pattern candidate creation unit 120.

  When the exponential smoothing coefficient is set, the processes in steps S204 to S212 are repeated for all the set values. First, the division pattern candidate creation unit 120 creates a division pattern candidate by dividing the whole area into a plurality of local areas, with the area taken as the value of the operation variable of the operation data as the whole area (step S204). The process of step S204 can be performed similarly to the process of step S102 of the first embodiment. When division pattern candidates are created in step S204, activity functions are calculated for all division pattern candidates (step S206). The process of step S206 can also be performed similarly to the process of step S104 of the first embodiment. The activity function calculation unit 130 sets the normal membership function obtained by substituting Equation (5) into Equation (6) as the activity function.

  Then, the smallest error division pattern selection unit 140 selects a division pattern with the smallest prediction error from the division pattern candidates (step S208). The minimum error division pattern selection unit 140 applies the exponential smoothing coefficient set in step S202 to each local region, calculates a quality prediction value by the exponential smoothing method, and calculates a prediction error based on the calculated quality prediction value. calculate. Then, the minimum error division pattern selection unit 140 selects a division pattern having the smallest prediction error among the calculated prediction errors.

  First, the minimum error division pattern selection unit 140 extracts one division pattern candidate from the plurality of division pattern candidates created by the division pattern candidate creation unit 120, and uses the division coordinate information of the division pattern to obtain individual data. Based on the value of the operation data, it is divided into M groups. Next, the minimum error division pattern selection unit 140 calculates a predicted value of each local region for each of the divided M groups by an exponential smoothing function expressed by the above equation (7). Here, as the exponential smoothing coefficient of Equation (7), one of the values set at step S202 is currently set.

  Then, the minimum error division pattern selection unit 140 uses the prediction value of each local region calculated by Expression (7) to calculate the overall prediction value by Expression (2) that is a relational expression between the operation data and the quality data. To do. Thereafter, the minimum error division pattern selection unit 140 calculates a prediction error using the error evaluation formula shown in Formula (8) above. The calculation of the prediction error represented by Equation (8) is performed for all the divided patterns.

  The minimum error division pattern selection unit 140 compares the prediction error calculated for each division pattern, and selects the relational expression of the division pattern with the smallest error. As a result, the relational expression of the division pattern having the smallest error defined by Equation (8) is selected from the plurality of division pattern candidates created by the division pattern candidate creation unit 120.

  Next, the learning error evaluation unit 160 compares the error of the minimum error relational expression selected by the minimum error division pattern selection unit 140 with a preset evaluation reference value to determine whether learning is sufficient. It is determined whether or not the error has converged (step S210). The process of step S210 can be performed similarly to the process of step S108 of the first embodiment. If the error is larger than the evaluation reference value, the learning error evaluation unit 160 determines that a relational expression that can explain the data with sufficient accuracy has not yet been constructed, and the error convergence is insufficient. In this case, the learning error evaluation unit 160 instructs the division pattern candidate creation unit 120 to increase the division number M of the division pattern by one and perform the process of step S204 (step S212). Then, the processes in steps S204 to S212 are repeated until it is determined in step S210 that a relational expression that can explain the data with sufficient accuracy is constructed.

  On the other hand, if the error is equal to or smaller than the evaluation reference value in step S210, the learning error evaluation unit 160 determines that a relational expression that can explain the data with sufficient accuracy is constructed, and the exponential smoothing coefficient set in step S202. It is determined whether or not the processing of steps S204 to S212 has been performed for all of the above (step S214). If it is determined in step S214 that the above processing has not been completed for all exponential smoothing coefficients, the exponential smoothing coefficient calculation unit 150 extracts one exponential smoothing coefficient that has not yet been processed. The next exponential smoothing coefficient is set (step S216). Thereafter, the processes in steps S204 to S216 are repeated.

  If it is determined in step S214 that the above processing has been completed for all exponential smoothing coefficients, the divisional exponential smoothing model determination unit 170 selects the smallest division number model having the smallest number of divisions (step S218). Through the processes in steps S204 to S216, the division pattern is determined for each of the plurality of exponential smoothing coefficients set by the exponential smoothing coefficient calculating unit 150. The divided exponential smoothing model determining unit 170 selects the model of the exponential smoothing function having the smallest number of divisions among the plurality of selected division patterns. When there are a plurality of minimum division number models, these prediction errors are compared, and the model with the smallest error may be selected. As a result, it is possible to provide a model that can explain the data with sufficient accuracy with a small number of divisions.

  Thereafter, the division index smoothing model 170 extracts the coefficient of the activity function, which is information for expressing the obtained relational expression, and outputs the coefficient to the quality prediction value output unit 180. The quality prediction value output unit 180 calculates the quality prediction value from Equation (2) using the input relational expression information and the separately input operation data, and outputs it to the outside (step S220). The quality prediction value calculated by the quality prediction value output unit 180 can be used, for example, as guidance to the quality prediction operator or an input signal to the process control system, as in the first embodiment.

  The quality prediction apparatus 100 according to the second embodiment of the present invention and the quality prediction method using the same have been described above. According to the present embodiment, first, a plurality of exponential smoothing coefficients used when calculating predicted values of a plurality of local regions of the division pattern are set. After that, for each case where each exponential smoothing coefficient is set, a plurality of division patterns are created, and each division pattern calculated using the activity function of each division pattern and the exponential smoothing function of the set exponential smoothing coefficient Based on the prediction values for a plurality of local regions, a division pattern that minimizes the quality prediction error is selected. When a division pattern is selected for each exponential smoothing coefficient, an exponential smoothing coefficient corresponding to the division pattern with the smallest number of divisions is selected from these. As a result, it is possible to prevent the occurrence of a local region in which almost no data exists, and it is possible to automatically search for and obtain a model that can explain data with sufficient accuracy with an appropriate minimum number of divisions. Further, since the area is not subdivided more than necessary, it is possible to reduce the maintenance load when the accuracy is deteriorated.

<3. Third Embodiment>
Next, based on FIGS. 7-10, the quality prediction apparatus 200 concerning the 3rd Embodiment of this invention and the quality prediction method by this are demonstrated. The quality prediction apparatus according to the present embodiment uses the quality variable according to the first and second embodiments in that the operation data used for quality prediction uses a code variable that is character code information in addition to an operation variable composed of numerical information. This is different from the prediction device 100. This makes it possible to create a model that can explain the data with sufficient accuracy with a smaller number of divisions. Hereinafter, the processing unit and the processing different from those in the first and second embodiments will be mainly described with respect to the quality prediction apparatus 200 and the quality prediction method using the quality prediction apparatus 200 according to the present embodiment. FIG. 7 is a block diagram illustrating a functional configuration of the quality prediction apparatus 200 according to the present embodiment. FIG. 8 is an explanatory diagram showing a process of creating a division pattern candidate by the division pattern candidate creation unit 220. FIG. 9 is a flowchart showing a process for creating code variable division pattern candidates. FIG. 10 is an explanatory diagram illustrating a code replacement process in a process for creating a division pattern candidate for a code variable.

[Configuration of quality prediction device]
As shown in FIG. 7, the quality prediction apparatus 200 in the manufacturing process according to the present embodiment includes a data extraction unit 210, a division pattern candidate creation unit 220, an activity function calculation unit 230, and a minimum error division pattern selection unit 240. And an exponential smoothing coefficient calculation unit 250, a learning error evaluation unit 260, a divided exponential smoothing model determination unit 270, a quality predicted value output unit 280, and a database 290. Here, the data extraction unit 210, the activity function calculation unit 230, the minimum error division pattern selection unit 240, the exponential smoothing coefficient calculation unit 250, the learning error evaluation unit 260, the division exponent smoothing model determination unit 270, and the quality prediction value output unit 280. , And the database 290, the data extraction unit 110, the activity function calculation unit 130, the minimum error division pattern selection unit 140, the exponential smoothing coefficient calculation unit 150, the learning error evaluation unit 160, the division of the embodiment described with reference to FIG. It functions in substantially the same manner as the exponential smoothing model determination unit 170, the quality prediction value output unit 180, and the database 190. In the quality prediction apparatus 200 according to the present embodiment, since operation variables and code variables are used as operation data, the process of creating the division pattern candidates by the division pattern candidate creation unit 220 is the same as the division pattern candidate creation unit 120 of the above embodiment. Is different.

  The data extraction unit 210 obtains, from the database 290, a plurality of operation data and quality data corresponding to a predetermined selection condition from the database 290 based on an externally input selection condition regarding a quality prediction target such as code information indicating a product type. Extract. As described in paragraph 0030, the operation data includes operation variables that are numerical information and code variables that are character code information. In the present embodiment, not only the operation variables of the operation data but also the code variables are targets of the division pattern. The code variable is, for example, character code information such as a product symbol and a raw material symbol. The data extraction unit 210 outputs the data extracted from the database 290 to the division pattern candidate creation unit 220.

  The division pattern candidate creation unit 220 divides the entire area composed of both operation variables whose possible values are numerical values and code variables whose possible values are character code information into a plurality of M local areas. A process for creating a plurality of files is performed. That is, the division pattern candidate creation unit 220 divides the entire area where the operation variables and code variables included in the operation data input from the data extraction unit 210 are mixed to create a plurality of division pattern candidates.

  The division pattern candidate creation unit 220 according to the present embodiment includes a numerical division creation unit 222 that divides only operation variables in the entire area as a division target, and a code division creation that performs division using only code variables in the whole area as division targets. Part 224. The division pattern candidate creation unit 220 divides a plurality of division pattern candidates generated by dividing the operation variable constituting the whole area by the numerical division creation unit 222 and the code variable constituting the whole area by the code division creation unit 224. Together with the plurality of divided pattern candidates generated in this way, it is determined as a divided pattern candidate for the entire area. Then, the division pattern candidate creation unit 220 outputs the created plurality of division pattern candidates to the activity function calculation unit 230.

  The activity function calculation unit 230 calculates the activity function for all the division pattern candidates created by the division pattern candidate creation unit 220 based on the division coordinate information of the operation variable and the grouping result of the code variable. . The activity function is used to represent a relational expression representing the relationship between the operation data and the quality data.

  The minimum error division pattern selection unit 240 calculates a quality prediction value in each local region using the exponential smoothing method for the division pattern candidate, calculates a prediction error based on the quality prediction value, and further generates a division pattern with the smallest error. Select. When the minimum error division pattern selection unit 240 calculates the predicted value of each local region using the exponential smoothing method, an exponential smoothing coefficient for weighting past data is calculated by the exponential smoothing coefficient calculation unit 250. The exponential smoothing coefficient can be calculated using, for example, the method described in the first embodiment. When the predicted value of each local region is calculated using the exponential smoothing coefficient calculated by the exponential smoothing coefficient calculation unit 250, the minimum error division pattern selection unit 240 calculates a prediction error and selects a division pattern with the smallest error. .

  The learning error evaluation unit 260 compares the prediction error of the division pattern selected by the minimum error division pattern selection unit 240 with a preset evaluation reference value to determine whether a relational expression having sufficient accuracy has been constructed. Determine. Also in the quality prediction apparatus 200 according to this embodiment, quality prediction can be performed by the two quality prediction methods shown in the first and second embodiments. When the exponential smoothing coefficient is set after creating the division pattern candidate (in the case of quality prediction method 1 described later), the learning error evaluation unit 260 expresses the obtained division pattern when it is determined that it has converged. The coefficient of the activity function, which is information for this purpose, is extracted and output to the quality prediction value output unit 280. On the other hand, when it is determined that the convergence is insufficient, the learning error evaluation unit 260 further divides the division pattern selected by the minimum error division pattern selection unit 240 into a new division pattern candidate creation unit 220 and performs a new division. An instruction to create a division pattern candidate is given.

  Alternatively, when the exponential smoothing coefficient is set before creating the division pattern candidate (in the case of quality prediction method 2 described later), the learning error evaluation unit 260 is currently set when it is determined that it has converged. The coefficient of the activity function, which is information for expressing the division pattern obtained for the exponential smoothing coefficient, is extracted and output to the divisional exponential smoothing model determination unit 270. On the other hand, when it is determined that the convergence is insufficient, the learning error evaluation unit 260 further divides the division pattern selected by the minimum error division pattern selection unit 240 with respect to the division pattern candidate creation unit 220. An instruction to create a new division pattern candidate is issued.

  The division exponent smoothing model determination unit 270 selects and determines a model of the exponential smoothing coefficient having the smallest number of divisions from the division patterns obtained for the exponential smoothing coefficient when using the quality prediction method 2 described later. The division index smoothing model determination unit 270 extracts the coefficient of the activity function, which is information for expressing the determined division pattern, and outputs the coefficient to the quality prediction value output unit 280.

  The quality prediction value output unit 280 includes information on the division pattern input from the learning error evaluation unit 260 or the division exponent smoothing model determination unit 270, operation data that is input separately, and the latest quality that is known at the time of prediction. Based on the data, a quality prediction value is calculated from the relational expression between the operation data and the quality data. The quality prediction value output unit 280 outputs the quality prediction value to the outside for use as guidance to the operator or as an input signal to the process control system.

  The database 290 includes past operation data and quality data in the manufacturing process, code information indicating the type of product, order information such as manufacturing time and destination in each manufacturing process, product number for specifying the product, and the like. Is stored in association with each other. Such information is input and recorded in the database 290 from the outside at a predetermined timing. The information stored in the database 190 is extracted by the database extraction unit 210 and used for quality prediction as described in paragraph 0080.

[Quality prediction processing by the quality prediction device]
Next, the quality prediction process by the quality prediction apparatus 200 according to the present embodiment will be described in detail. The quality prediction apparatus 200 according to the present embodiment can perform quality prediction processing in the same manner as the quality prediction method in the first embodiment shown in FIG. 2 and the quality prediction method in the second embodiment shown in FIG. . Hereinafter, processing in each case will be described. Note that a detailed description of portions that perform the same processing as in the above embodiment will be omitted.

(Quality prediction method 1)
First, processing when the quality prediction method shown in FIG. 2 is used will be described. In the quality prediction process performed by the quality prediction apparatus 200 according to the present embodiment, first, operation data and quality data relating to a target for which quality prediction is performed are extracted from the database 290 by the data extraction unit 210 (step S100). Similarly to the first embodiment, the data extraction unit 210 extracts a plurality of operation data and quality data corresponding to a predetermined selection condition from the database 290. The data extraction unit 210 outputs the extracted operation data and quality data to the division pattern candidate creation unit 220.

  Next, the division pattern candidate creation unit 220 creates a division pattern candidate by dividing the whole area into a plurality of local areas by taking the area that the operation variable of the operation data takes as the value and the area that the code variable takes as the value as the whole area. (Step S102). In the present embodiment, in order to create a division pattern candidate from the entire region in which operation variables and code variables are mixed, a numerical value division creation unit 222 that divides only based on the operation variable space, and a code that divides only based on the code variables Using the division creation unit 224, division pattern candidates are created separately for the operation variable and the code variable. Then, the divided pattern candidates created by both are combined to be a divided pattern candidate for the entire area.

  The operation variable space division processing by the numerical value division creation unit 222 can be performed in the same manner as the processing by the division pattern candidate creation unit 220 described in the first embodiment. That is, as illustrated in FIG. 8, the numerical value division creating unit 222 performs a process of sequentially dividing the local region having the largest error, and generates a plurality of new division pattern candidates from the current division pattern. However, the weighted error evaluation function by the activity function for calculating the error of the local region is calculated by the following mathematical formula (9) instead of the mathematical formula (4). Equation (9) specifies that the activity function is a function of operation variables and code variables.

・・・数式（９）

... Formula (9)

  On the other hand, the code division creation unit 224 creates code variable division pattern candidates by dividing the code variable into two groups. Here, since the possible values of the code variables are generally several hundreds, it is not realistic from the viewpoint of calculation time to evaluate all combinations that divide these into two groups. Therefore, in the present embodiment, for example, by applying a search method proposed in the field of mathematical optimization, an initial combination of division is assumed from a huge number of combinations, and then a combination is obtained so as to obtain a highly accurate relational expression. A method of calculating the division into the optimum group by searching while correcting is used. Specifically, the code division can be calculated using a local search method which is a kind of search method.

  Here, a code variable dividing method by the local search method will be described with reference to FIGS. First, the code division creation unit 224 generates a random number to initially divide the code variable, and creates a relational expression (hereinafter also referred to as “quality prediction model”) representing the relationship between the operation data and the quality data (step “step”). S1021). For example, numbers 1 to D are assigned to all types D of values that the code variable v can take. A random number generator that generates a random number between 0 and 1 is executed i = 1 to D times, and the i-th code value is determined by whether or not the value of the i-th random number exceeds a threshold value (for example, 0.5). To which group data V belongs to group V or a group of code variables other than the values of code variables belonging to V. For example, as shown in FIG. 7, when there are eight possible values of the code variable v, these data are divided so as to belong to one of the two groups V and V˜.

  Next, the code division creation unit 224 creates a quality prediction model for the data divided into two groups by the same procedure as the process in the minimum error division pattern selection unit 240 from the activity function calculation unit 230, A prediction error is calculated (step S1022). Steps 1021 and S1022 are performed as initial processing in the code variable division processing.

  When the prediction error is calculated for the quality prediction model after the initial division, the code division creation unit 224 executes the random number generator that can generate the values 1 to D once, and the code corresponding to the obtained random value DR The value data is replaced with a group different from the current one (step S1023). For example, in the example shown in FIG. 10, it is assumed that the data of the code value corresponding to the random value DR obtained by the random number generator is “2”. The data belongs to the current group V˜, but is moved to a group V different from the current group by the processing of step S1023. In this way, the code division creation unit 224 recreates the quality prediction model by exchanging data.

  Thereafter, the code division creation unit 224 creates a quality prediction model from the replaced data, and calculates a prediction error (step S1024). The process of step S1024 can be performed in the same manner as the process of step S1022.

  Next, the code division creation unit 224 compares the prediction error before replacement with the prediction error after replacement, and performs accuracy evaluation (step S1025). If the prediction error after replacement is smaller than the prediction error before replacement, the code division creation unit 224 determines the replacement performed in step S1023 and proceeds to the process of step S1027. On the other hand, when the prediction error after replacement is equal to or greater than the prediction error before replacement, the two groups are returned to the state before replacement (step S1026). Then, it returns to step S1023 and repeats a process.

  In step S1025, after confirming the group replacement in step S1023, the code division creation unit 224 determines whether or not the iterative stop condition is satisfied (step S1027). As the iterative stop condition, for example, a condition such as when the upper limit value of the number of iterations has been reached, or when the upper limit value of the number of executions of replacement that does not improve the prediction error can be set. The code division creation unit 224 stops the iterative process when the iterative stop condition is met, and ends the code variable division process. That is, the two groups determined at this time are code variable division pattern candidates. On the other hand, if it is determined that the repeated stop condition is not met, the process returns to step S1023 to repeat the process.

  The code variable division processing has been described above. When there are a plurality (p) of code variables, the above processing is performed for each code variable j = 1 to p, and division pattern candidates for q types of code variables may be created.

  The method of creating the division pattern candidate by the division pattern candidate creation unit 220 is not limited to the above-described method. For example, an optimization method such as a genetic algorithm is used for dividing a pattern candidate composed of an operation variable and a code variable. A method of generating a batch using, is also within the scope of the present invention. Specifically, a plurality of values of operation variables and code variables that are division points in division pattern candidates are stored in an array, crossover processing for exchanging a part of the arrays, and a partial value of the array A mutation process to be changed is performed, a quality prediction model is created from the activity function calculation unit 230 in the same procedure as the process in the minimum error division pattern selection unit 240, and a prediction error is calculated. Then, similarly to the above method, the size of the prediction error can be evaluated to determine the division pattern candidate.

  Returning to the description of the quality prediction method of FIG. 2 in the present embodiment, when division pattern candidates are created in step S102, activity functions are calculated for all division pattern candidates (step S104). As specific activity functions, for example, a p-dimensional normal distribution function shown in Expression (10) and a binary function expressed by Expression (11) calculated from the code division information are expressed in Expression (12). The membership function obtained by substitution can be used as the activity function.

・・・数式（１０）

・・・数式（１１）

・・・数式（１２）
（ｉ＝１，・・・，Ｍ）、（ｋ＝１，・・・，Ｍ）

... Formula (10)

... Formula (11)

... Formula (12)
(I = 1,..., M), (k = 1,..., M)

Here, c ⁱ _j is the barycentric point of the local region, σ ⁱ _j is the standard deviation of the normal distribution function, v is the code variable, V is a set of values of the code variables belonging to the group V obtained by the code division creation means. is there. As shown in Equation (12), it can be seen that the activity function according to the present embodiment is a function of an operation variable and a code variable. This membership function is 0 for any operation data if the code variable value of the operation data does not belong to the set V, and 0 to 1 depending on the value of the operation variable if it belongs. It has a range of values.

  Thereafter, the minimum error division pattern selection unit 240 selects a division pattern having the smallest prediction error from the division pattern candidates (step S106). First, the minimum error division pattern selection unit 240 extracts one division pattern candidate from the plurality of division pattern candidates created by the division pattern candidate creation unit 220, and uses the division coordinate information and code information of the division pattern to individually Is divided into M groups based on the value of the operation data. Next, the minimum error division pattern selection unit 240 calculates the predicted value of each local region for each of the divided M groups by an exponential smoothing function expressed by the above equation (7). The exponential smoothing coefficient of Equation (7) is calculated by the exponential smoothing coefficient calculation unit 150 using the exponential smoothing coefficient calculation method described in paragraph 0034.

  Then, the minimum error division pattern selection unit 240 uses the predicted value of each local region calculated by Formula (7) to calculate the overall predicted value by Formula (2) that is a relational expression between the operation data and the quality data. To do. Thereafter, the minimum error division pattern selection unit 240 calculates a prediction error using the error evaluation formula shown in the formula (8). The calculation of the prediction error represented by Equation (8) is performed for all the divided patterns. Then, the minimum error division pattern selection unit 140 compares the prediction error calculated for each division pattern, and selects the relational expression of the division pattern with the smallest error. The processing in this step can be performed in substantially the same manner as in the first embodiment.

  Next, the learning error evaluation unit 260 compares the error of the minimum error relational expression selected by the minimum error division pattern selection unit 240 with a preset evaluation reference value to determine whether learning is sufficient. It is determined whether or not the error has converged (step S108). The processing in this step can also be performed in the same manner as in the first embodiment. If the error is larger than the evaluation reference value, the learning error evaluation unit 260 determines that the error has not converged sufficiently, and sets the division pattern candidate creation unit 220 to one division pattern division number M. Increase the number and instruct to perform the process of step S102 (step S110). Then, the processes in steps S102 to S110 are repeated until it is determined in step S108 that a relational expression that can explain the data with sufficient accuracy is constructed.

  On the other hand, if the error is equal to or smaller than the evaluation reference value in step S108, the learning error evaluation unit 260 determines that a relational expression that can explain the data with sufficient accuracy is constructed, and the relational expression smoothing model 270 determines the relational expression. Is determined. Thereafter, as in the first embodiment, the division index smoothing model 270 extracts the coefficient of the activity function, which is information for expressing the obtained relational expression, and outputs the coefficient to the quality prediction value output unit 280. The quality prediction value output unit 280 uses the information on the input relational expression, the operation data that is sequentially input, and the latest quality information that is known at the time of the prediction, to calculate the quality prediction value from Equation (2). Is calculated and output to the outside (step S112).

(Quality prediction method 2)
Next, processing when the quality prediction method shown in FIG. 6 is used will be described. In the quality prediction process performed by the quality prediction apparatus 200 according to the present embodiment, first, operation data and quality data relating to a target for which quality prediction is performed are extracted from the database 290 by the data extraction unit 210 (step S200). Similarly to the second embodiment, the data extraction unit 210 extracts a plurality of operation data and quality data corresponding to a predetermined selection condition from the database 290. The data extraction unit 210 outputs the extracted operation data and quality data to the division pattern candidate creation unit 220.

  Next, an exponential smoothing coefficient is set by the exponential smoothing coefficient calculation unit 250 (step S202). As in the second embodiment, the exponential smoothing coefficient calculation unit 250 can, for example, perform an exponential smoothing coefficient that can take a value in the range of 0 to 1, for example, 0.1, 0.2,. Set multiple types of values as shown in .9. Then, the exponential smoothing coefficient calculation unit 250 extracts one from the set plurality of values and outputs it to the division pattern candidate creation unit 220.

  When the exponential smoothing coefficient is set, the processes in steps S204 to S212 are repeated for all the set values. First, the division pattern candidate creation unit 220 creates a division pattern candidate by dividing the entire area into a plurality of local areas, with the area that the operation variable of the operation data takes as the value and the area that the code variable takes as the value as the entire area. (Step S204). The process of step S204 can be performed similarly to the process of step S102 in the quality prediction method 1 of this embodiment mentioned above. When division pattern candidates are created in step S204, activity functions are calculated for all division pattern candidates (step S206). The process of step S206 can also be performed similarly to the process of step S104 in the quality prediction method 1 of this embodiment mentioned above.

  Then, the smallest error division pattern selection unit 240 selects a division pattern with the smallest prediction error from the division pattern candidates (step S208). The minimum error division pattern selection unit 240 applies the exponential smoothing coefficient set in step S202 to each local region, calculates a quality prediction value by the exponential smoothing method, and calculates a prediction error based on the calculated quality prediction value. calculate. Then, the minimum error division pattern selection unit 240 selects a division pattern having the smallest prediction error among the calculated prediction errors. The minimum error division pattern selection unit 240 extracts one division pattern candidate from the plurality of division pattern candidates created by the division pattern candidate creation unit 220, and uses the division coordinate information and code information of the division pattern to obtain individual data. Are divided into M groups based on the value of the operation data, and then the same processing as the processing in step S208 of the second embodiment is performed, so that a plurality of divisions created by the division pattern candidate creation unit 220 are performed. From the pattern candidates, the division pattern of the relational expression having the smallest error defined by Expression (8) is selected.

  Next, the learning error evaluation unit 260 compares the error of the minimum error relational expression selected by the minimum error division pattern selection unit 240 with a preset evaluation reference value to determine whether learning is sufficient. It is determined whether or not the error has converged (step S210). When the error is larger than the evaluation reference value, the learning error evaluation unit 260 determines that the relational expression that can explain the data with sufficient accuracy has not been constructed yet and the convergence of the error is insufficient. In this case, the learning error evaluation unit 260 instructs the division pattern candidate creation unit 220 to increase the division number M of the division pattern by one and perform the process of step S204 (step S212). Then, the processes in steps S204 to S212 are repeated until it is determined in step S210 that a relational expression that can explain the data with sufficient accuracy is constructed.

  On the other hand, if the error is equal to or less than the evaluation reference value in step S210, the learning error evaluation unit 260 determines that a relational expression that can explain the data with sufficient accuracy has been constructed, and the exponential smoothing coefficient set in step S202. It is determined whether or not the processing of steps S204 to S212 has been performed for all of the above (step S214). If it is determined in step S214 that the above processing has not been completed for all exponential smoothing coefficients, the exponential smoothing coefficient calculation unit 250 extracts one exponential smoothing coefficient that has not yet been processed. The next exponential smoothing coefficient is set (step S216). Thereafter, the processes in steps S204 to S216 are repeated.

  If it is determined in step S214 that the above processing has been completed for all exponential smoothing coefficients, the divisional exponential smoothing model determining unit 270 selects the smallest division number model having the smallest number of divisions (step S218). Thereafter, the division index smoothing model 270 extracts the coefficient of the activity function, which is information for expressing the obtained relational expression, and outputs the coefficient to the quality prediction value output unit 280. The quality prediction value output unit 280 calculates the quality prediction value from Expression (2) using the input relational expression information and the separately input operation data, and outputs it to the outside (step S220). Note that the processes in steps S210 to S218 described above can be performed in the same manner as in the second embodiment.

  The quality prediction apparatus 200 and the quality prediction method using the same according to the third embodiment of the present invention have been described above. According to the present embodiment, an operation variable, which is numerical information of operation data used for quality prediction, and a code variable, which is character code information, are used as an entire area, and a division pattern candidate is created from the entire area. As in the first or second embodiment, the quality prediction is performed based on the activity function of each divided pattern and the quality predicted value of each local area of each divided pattern calculated using the exponential smoothing method. The division pattern that minimizes the error is selected. As a result, it is possible to prevent the occurrence of a local region in which almost no data exists, and automatically search for and obtain a model that can explain the data with sufficient accuracy with a smaller number of divisions than in the case of the above embodiment. be able to. Further, since the area is not subdivided more than necessary, it is possible to reduce the maintenance load when the accuracy is deteriorated.

<4. Hardware configuration diagram>
The quality prediction apparatuses 100 and 200 in the manufacturing process according to the embodiment of the present invention can be realized by a computer. FIG. 11 shows a configuration example of a computer system 400 that can function as a quality prediction apparatus according to an embodiment of the present invention. The computer system 400 includes a CPU 401, a ROM 402, a RAM 403, a keyboard controller (KBC) 405, a CRT controller (CRTC) 406, a disk controller (DKC) 407, and a network interface controller (NIC) 408. They are connected to each other via 404.

  The CPU 401 executes software stored in the ROM 402 or the HD 411 or software supplied from the FD 412 and comprehensively controls each component connected to the system bus 404. That is, the CPU 401 reads out and executes a processing program according to a predetermined processing sequence from the ROM 402, the HD 411, or the FD 412, and performs control for realizing the functions of the quality prediction apparatuses 100 and 200 in the present embodiment. .

  The RAM 403 functions as a main memory or work area for the CPU 401. The KBC 405 controls instruction input from the KB 409, a pointing device (not shown), or the like. The CRTC 406 controls the display of the CRT 410 that is a display unit. The DKC 407 controls access to the HD 411 and the FD 412 that store a boot program, various applications, an editing file, a user file, a network management program, a predetermined processing program in the present embodiment, and the like. The NIC 408 exchanges data bidirectionally with devices or systems on the LAN 420.

  In addition, the program of the software which described the process for implement | achieving the function of the process of each step of the quality prediction apparatus which is embodiment of this invention and the quality prediction apparatus of this invention was supplied, and was stored in the computer What is implemented by operating various devices according to the program is also included in the scope of the present invention.

  In this case, the software program itself realizes the processing functions of the quality prediction apparatuses 100 and 200 according to the embodiment of the present invention, and the program itself is included in the scope of the present invention. As a transmission medium for the program code, a communication medium such as a computer network system that transmits the program as an electric signal can be used.

  Furthermore, means for supplying the program to the computer, for example, a storage medium storing the program is also included in the scope of the present invention. As a storage medium for storing the program code, for example, a flexible disk, a hard disk, an optical disk, a magneto-optical disk, a CD-ROM, a CD-R, a magnetic tape, a nonvolatile memory card, a ROM, or the like can be used.

  Below, for the hot rolling process of a thin steel sheet that is a steel product, the steel sheet temperature in the scraping is regarded as quality, and the operation variable includes the amount of C in the molten steel at the time when the material to be rolled has finished the refining process, Si, Mn, P, S, Cu, Ni, Cr, Mo, Nb, V, Ti, B, Al, N, O, Ca, rolled Quality prediction using material code as target plate width, target plate thickness, equivalent carbon content, finish rolling target temperature, scraping target temperature, rolling speed, cooling water volume density, cooling water temperature, and code variables Examples will be described.

  FIG. 12 is a diagram showing an outline of a hot rolling process of a thin steel plate. A base material of the material 10 to be rolled, called a slab, is extracted from a heating furnace and then rolled into a bar shape having a thickness of 40 to 50 mm by a roughing mill 510. A roughing side plate width meter 540 is installed on the exit side of the rough rolling mill 510, and the plate width of the material to be rolled 10 at this time is measured by the roughing side plate width meter 540. Next, the material to be rolled 10 is continuously rolled by a finish rolling mill 520 configured by arranging a plurality of work rolls 521 supported by a backup roll 522 in series, and then coiled by a take-off machine 530. It is wound up.

  Immediately before the scoring machine 530, a non-contact scoring thermometer 560 using a radiation thermometer is installed, and the scoring thermometer 560 has an important effect on material properties such as mechanical property values of the thin steel plate. The steel sheet temperature at the time of cutting, which is a management index, is measured. In order to control the steel sheet temperature within a predetermined range, a cooling facility for spraying water on the steel sheet is provided on the runout table 550 from the exit side of the finish rolling mill 520 to the scraper 530.

  In the hot rolling process, a prediction model for estimating the steel plate scraping temperature is created by inputting the target temperature at each stage such as the steel plate component and size, the exit side of the finish rolling mill 520 and just before scraping. Furthermore, according to the actual temperature measured with the thermometer installed on the entrance side of the finish rolling mill 520, using this prediction model, the cooling water amount density in which the cutting temperature is in a predetermined range is calculated, and the setting value of the water amount density is Change operations are performed.

  Conventionally, the cooling phenomenon when water is sprayed on the steel plate on the run-out table 550 is estimated by a physical model based on the heat transfer equation, and a correction coefficient for correcting the calculation result by this physical model is calculated by an exponential smoothing method. I was doing what I wanted. At this time, since the number of divided meshes for improving the correction accuracy reaches 4000 divisions, there is a mesh (local region) to which almost no data is input, and there is a problem that the accuracy of the quality prediction value is poor in such a mesh. .

  Therefore, the quality prediction apparatus according to the embodiment of the present invention is applied to configure a quality prediction apparatus that uses the correction coefficient for correcting the calculation result by the physical model as the quality index y. When the quality prediction apparatus according to the embodiment of the present invention is applied, the number of divisions necessary for constructing a relational expression that can explain data with sufficient accuracy was obtained. Here, the data for about 41,000 coils is used, and the standard deviation σ = 151.5 of the error of the conventional method is used as an error determination variable, and the number of divisions when the accuracy becomes better than this is sufficient. The number of divisions required to construct a relational expression that can explain the data with high accuracy was used. Since the correction coefficient is described by multiplying the value by 1000, the value 1000 means the correction coefficient 1, and is a state in which no correction is applied to the value of the prediction model. As an evaluation index, a prediction error is defined by a difference between an accurate correction coefficient value obtained by comparison with actual temperature and a correction coefficient obtained by prediction, and this standard deviation is used as an index.

  FIG. 13 shows a graph showing the relationship between the number of divisions and the prediction error when the quality prediction apparatus 200 according to the third embodiment of the present invention, that is, when a division pattern is created using operation variables and code variables. . From FIG. 13, it can be seen that the prediction error rapidly converges as the number of divisions in the entire region increases. When the entire area was divided into 50, the standard deviation of the error was σ = 149.0, and the accuracy was better than the conventional error equivalent value (σ = 151.5). Similarly, when the quality prediction apparatus 100 according to the first or second embodiment of the present invention, that is, when the division pattern is created using only the operation variable, the total number of divisions is 100 divisions. The standard deviation of the error was σ = 150.8, and the accuracy was better than the conventional error equivalent value.

  As described above, by applying the quality prediction apparatus according to the embodiment of the present invention, when the division pattern is generated without using the code variable, 100 divisions are performed, and when the division pattern is generated using the code variable, 50 divisions are performed. In contrast to the 4000 division of the conventional method, a model having the same degree of accuracy could be obtained with a small number of divisions in each stage.

  Moreover, the histogram of the error of a prediction result is shown in FIGS. 14 shows a case where the conventional method is applied, and FIG. 15 shows a case where the quality prediction apparatus according to the embodiment of the present invention is applied. In the case where a division pattern is created without using a code variable, FIG. This is a case where the quality prediction apparatus according to the embodiment is applied and a division pattern is created using a code variable.

  By using the method of the embodiment of the present invention in contrast to the error σ = 151.5 of the conventional method shown in FIG. 14, the error of the 100-division model shown in FIG. 15 is σ = 150.8, and the code variable shown in FIG. It was found that a good result of improvement with σ = 149.0 was obtained with the 50-division model used. Thus, by applying the quality prediction apparatus according to the embodiment of the present invention, an effect of improving accuracy while reducing the number of divisions was obtained. As a result, the standard deviation of the scraping temperature prediction error is improved, and as a result of continuous operation of this model, the occurrence rate of scraping temperature mismatch is reduced, and the quality due to the reduction in variation in mechanical characteristic values is further reduced. Effects such as improvement and yield improvement could be obtained.

  The preferred embodiments of the present invention have been described in detail above with reference to the accompanying drawings, but the present invention is not limited to such examples. It is obvious that a person having ordinary knowledge in the technical field to which the present invention pertains can come up with various changes or modifications within the scope of the technical idea described in the claims. Of course, it is understood that these also belong to the technical scope of the present invention.

  For example, in the above embodiments, the quality prediction apparatus is realized as a program on a computer, but the present invention is not limited to such an example. For example, it may be configured by hardware combining an arithmetic device, a memory, and the like. Further, the operation / quality relation analysis apparatus of the present invention may be composed of a plurality of devices or a single device.

  Moreover, in the said embodiment, although the division | segmentation of the local area | region by a division | segmentation pattern candidate production | generation part divided | segmented into 2 parts, this invention is not limited to this example, For example, a local area | region may be divided into 3 parts. .

100, 200 Quality prediction device 110, 210 Data extraction unit 120, 220 Division pattern candidate creation unit 130, 230 Activity function calculation unit 140, 240 Minimum error division pattern selection unit 150, 250 Exponential smoothing coefficient calculation unit 160, 260 Learning error Evaluation unit 170, 270 Division exponent smoothing model determination unit 180, 280 Quality prediction value output unit 190, 290 Database 222 Numerical division creation unit 224 Code division creation unit

Claims (13)

A quality prediction device in a manufacturing process,
A data extraction unit for extracting operation data and quality data corresponding to a predetermined selection condition from a database storing operation data and quality data related to a plurality of products;
With respect to the extracted operation data, a division pattern candidate creation unit that creates a plurality of division pattern candidates that divide the whole region into a plurality of local regions, with the region that the variable included in the operation data takes a value as a whole region, ,
For each of the divided pattern candidates, an activity function calculation unit that calculates an activity function that represents a contribution rate that is a ratio of a quality predicted value in each local region to a predicted value in the entire region;
For each divided pattern candidate, an exponential smoothing function set based on an exponential smoothing method for calculating a current quality predicted value from past quality data and a past quality predicted value for each local region, and the activity Based on the degree function, a relational expression for calculating a quality prediction value in the entire area is derived, and a quality prediction value in the entire area is calculated based on the relational expression, past quality data, and a past quality prediction value, A minimum error division pattern selection unit that selects a division pattern that minimizes a prediction error that is a difference between the quality prediction value in the entire region and the quality data;
Learning to determine whether the prediction error is sufficiently converged based on a comparison result between a quality prediction value of the division pattern selected by the minimum error division pattern selection unit and a preset evaluation reference value An error evaluation unit;
A quality prediction value output unit that outputs a quality prediction value using the relational expression of the divided pattern that is determined to have sufficient convergence by the learning error evaluation unit;
In the exponential smoothing function, an exponential smoothing coefficient calculating unit that calculates an exponential smoothing coefficient that determines the strength of the influence of the past quality prediction on the current quality prediction value;
A quality prediction apparatus comprising:   The exponent smoothing coefficient calculating unit performs the exponential smoothing calculation on a plurality of exponential smoothing coefficients, and selects an exponential smoothing coefficient having the smallest prediction error calculated from the calculation result. Quality prediction equipment. A division index smoothing model determination unit that selects one of the plurality of division patterns determined by the learning error evaluation unit to determine that the convergence of the prediction error is sufficient;
The exponential smoothing coefficient calculating unit sets a plurality of exponential smoothing coefficients,
For each exponential smoothing coefficient, a division pattern candidate is created by the division pattern candidate creation unit, and the prediction error is converged by the activity function calculation unit, the minimum error division pattern selection unit, and the learning error evaluation unit. Determining each said division pattern that is sufficient,
The quality prediction according to claim 1, wherein the division exponent smoothing model determination unit selects the division pattern having the smallest number of divisions among the division patterns determined for each exponential smoothing coefficient. apparatus. The division pattern candidate creation unit
A numerical value division creating unit for dividing an operation variable space composed of operation variables;
A code division creation unit for dividing a group of code variables;
With
The division pattern candidate created by the numerical value division creation unit and the division pattern candidate created by the code division creation unit are set as division pattern candidates for the entire region. The quality prediction apparatus according to item.   5. The activity function is a function configured by combining a normal distribution function having a center at the center of gravity of the local region and a binary function calculated from the code division information. The quality prediction apparatus described in 1.   6. The quality prediction apparatus according to claim 4, wherein the code division creation unit divides a group including the code variables using a search method.   The division pattern candidate creation unit generates a division pattern candidate consisting of the operation variable and the code variable in a lump using an optimization method. The quality prediction apparatus described in 1. The manufacturing process is a steel process,
The quality data is a surface wrinkle of a product, a mechanical strength characteristic value, a flatness of a shape, a product size, an internal stress, or a process value that affects the quality, according to any one of claims 1 to 7. The quality prediction apparatus according to claim 1. The manufacturing process is a hot rolling process of a thin steel plate that is a steel product,
When the quality to be predicted by the quality prediction device is the scraping temperature at the hot-run runout table outlet side,
The operation variables of the manufacturing process are as follows: C amount in molten steel, Si amount, Mn amount, P amount, S amount, Cu amount, Ni amount, Cr amount, Mo amount in the molten steel when the material to be rolled has finished the refining process, Nb amount, V amount, Ti amount, B amount, Al amount, N amount, O amount, Ca amount, target plate width of rolled material, target plate thickness, equivalent carbon amount, finish rolling target temperature, scraping target The quality prediction apparatus according to claim 8, wherein at least one is selected from a temperature, a rolling speed, a cooling water amount density, a cooling water temperature, and a material code. A quality prediction method in a manufacturing process,
Extracting operation data and quality data corresponding to a predetermined selection condition from a database storing operation data and quality data for a plurality of products;
With respect to the extracted operation data, a region where a variable included in the operation data takes a value as a whole region, dividing the whole region into a plurality of local regions, creating a plurality of division pattern candidates,
For each of the divided pattern candidates, calculating an activity function representing a contribution rate that is a ratio of a quality predicted value in each local region to a predicted value in the entire region;
For each division pattern candidate, the strength of the influence of the past quality prediction on the current quality prediction value in the exponential smoothing function for calculating the current quality prediction value for each local region from the time series data of the quality prediction value Calculating an exponential smoothing coefficient for determining the data based on the quality data corresponding to each local region;
For each divided pattern candidate, based on an exponential smoothing function in which the exponential smoothing coefficient is set for each local region and the activity function, a relational expression for calculating a quality prediction value in the entire region is derived, A quality prediction value in the entire area is calculated based on the relational expression, past quality data, and a past quality prediction value, and a prediction error that is a difference between the quality prediction value in the entire area and the quality data is minimized. Selecting a division pattern; and
Determining whether or not the convergence of the prediction error is sufficient based on a comparison result between a quality prediction value of the selected divided pattern and a preset evaluation reference value;
Using the relational expression of the divided pattern determined to have sufficient convergence by the step of determining whether or not the convergence of the prediction error is sufficient, and outputting a quality prediction value;
The quality prediction method characterized by including these. A quality prediction method in a manufacturing process,
Extracting operation data and quality data corresponding to a predetermined selection condition from a database storing operation data and quality data for a plurality of products;
An exponential smoothing coefficient that determines the strength of the influence of the past quality prediction on the current quality prediction value in the exponential smoothing function for calculating the current quality prediction value for each local region from the time series data of the quality prediction value A step of setting a plurality,
With respect to the extracted operation data, a region where a variable included in the operation data takes a value as a whole region, dividing the whole region into a plurality of local regions, creating a plurality of division pattern candidates,
For each of the divided pattern candidates, calculating an activity function representing a contribution rate that is a ratio of a quality predicted value in each local region to a predicted value in the entire region;
For each divided pattern candidate, a quality prediction value in the entire region is calculated based on an exponential smoothing function in which one of the plurality of exponential smoothing coefficients is set for each local region and the activity function. Deriving a relational expression, calculating a quality prediction value in the entire area based on the relational expression, past quality data, and past quality prediction value, and a difference between the quality prediction value in the whole area and the quality data Selecting a division pattern that minimizes the prediction error;
Determining whether or not the convergence of the prediction error is sufficient based on a comparison result between a quality prediction value of the selected divided pattern and a preset evaluation reference value;
For each exponential smoothing coefficient, a step of selecting the division pattern having the smallest number of divisions from the division patterns determined to have sufficient convergence by the step of determining whether or not the prediction error has sufficiently converged When,
Using a relational expression of the division pattern with the smallest number of divisions, and outputting a quality prediction value;
The quality prediction method characterized by including these. Computer
Data extraction means for extracting operation data and quality data corresponding to a predetermined selection condition from a database storing operation data and quality data related to a plurality of products;
With respect to the extracted operation data, a division pattern candidate creation that creates a plurality of division pattern candidates that divide the whole region into a plurality of local regions, with the region taken as a value included in the operation data as a whole region Means,
For each of the divided pattern candidates, an activity function calculating unit that calculates an activity function representing a contribution rate that is a ratio of a quality predicted value in each local region to a predicted value in the entire region;
For each divided pattern candidate, an exponential smoothing function set based on an exponential smoothing method for calculating a current quality predicted value from past quality data and a past quality predicted value for each local region, and the activity Based on the degree function, a relational expression for calculating a quality prediction value in the entire area is derived, and a quality prediction value in the entire area is calculated based on the relational expression, past quality data, and a past quality prediction value, Minimum error division pattern selection means for selecting a division pattern that minimizes a prediction error that is a difference between the quality prediction value in the entire region and the quality data;
Learning to determine whether or not the convergence of the prediction error is sufficient based on a comparison result between a quality prediction value of the division pattern selected by the minimum error division pattern selection unit and a preset evaluation reference value Error evaluation means;
Quality prediction value output means for outputting a quality prediction value using the relational expression of the divided pattern determined to have sufficient convergence by the learning error evaluation means;
In the exponential smoothing function, exponential smoothing coefficient calculating means for calculating an exponential smoothing coefficient that determines the strength of the influence of the past quality prediction on the current quality prediction value;
The program for functioning as a quality prediction apparatus characterized by comprising. On the computer,
Data extraction means for extracting operation data and quality data corresponding to a predetermined selection condition from a database storing operation data and quality data related to a plurality of products;
With respect to the extracted operation data, a division pattern candidate creating unit that creates a plurality of division pattern candidates that divide the whole region into a plurality of local regions, with the region that the variable included in the operation data takes as a whole region. ,
For each of the divided pattern candidates, an activity function calculating unit that calculates an activity function representing a contribution rate that is a ratio of a quality predicted value in each local region to a predicted value in the entire region;
For each divided pattern candidate, an exponential smoothing function set based on an exponential smoothing method for calculating a current quality predicted value from past quality data and a past quality predicted value for each local region, and the activity Based on the degree function, a relational expression for calculating a quality prediction value in the entire area is derived, and a quality prediction value in the entire area is calculated based on the relational expression, past quality data, and a past quality prediction value, Minimum error division pattern selection means for selecting a division pattern that minimizes a prediction error that is a difference between the quality prediction value in the entire region and the quality data;
Learning to determine whether or not the convergence of the prediction error is sufficient based on a comparison result between a quality prediction value of the division pattern selected by the minimum error division pattern selection unit and a preset evaluation reference value Error evaluation means;
Quality prediction value output means for outputting a quality prediction value using the relational expression of the divided pattern determined to have sufficient convergence by the learning error evaluation means;
In the exponential smoothing function, exponential smoothing coefficient calculating means for calculating an exponential smoothing coefficient that determines the strength of the influence of the past quality prediction on the current quality prediction value;
The computer-readable recording medium which recorded the program for functioning as a quality prediction apparatus characterized by the above-mentioned.

Images