JP2011039763A - Output value prediction method, output value prediction device, and program for the method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an output value prediction method, device, and program, obtaining variations of an output value. <P>SOLUTION: In the output value prediction method, an error parameter A consisting of a plurality of parameter elements α is prepared beforehand, M pieces of error parameters A<SB>j</SB>(j=1 to M) calculated from M pieces of past result data (X, y) are used, and thereby M kinds of predicted values y<SB>0</SB>for prediction target data X<SB>0</SB>are calculated, and then a weighted frequency F<SB>w</SB>for the predicted value y<SB>0</SB>is calculated according to similarity w<SB>j</SB>to the prediction target data X<SB>0</SB>. Further, a probability density is calculated from the weighted frequency F<SB>w</SB>as the variation of the predicted value. <P>COPYRIGHT: (C)2011,JPO&INPIT

Description

  The present invention relates to an output value prediction technique for predicting an output value from given data, and more particularly to an output value prediction technique for obtaining a variation in predicted output values.

  In various fields, future predictions are often made when determining future actions. In particular, in products manufactured through various manufacturing processes in a relatively large-scale manufacturing plant, such as manufacturing of steel products and chemical products, for example, depending on input amount, operation input amount, time passage, etc. In many cases, the output value in each manufacturing process and the output value of the final process directly connected to the product change every moment. For this reason, in order to control the output value, the prediction of the output value is important.

  Such prediction is generally performed by analyzing a factor related to a prediction target and statistically analyzing past performance data of the factor.

  For example, the steel material estimation apparatus disclosed in Patent Document 1 includes a material storage unit that accumulates material component results, operation results, and material results for each product manufactured in the past, and a product storage device from among a large number of input variables. An input variable that restricts an input variable according to the rule by using an input variable limiting rule storage unit that stores a rule for selecting an input variable having a large influence on the material, and input material component information and operation information. Defining a distance function using a weighting factor as an influence of the input value on the output value for calculating the distance between each data in the material storage means and the input value using the limited input variable. And a material estimation calculating means for extracting data close to the input value based on the distance calculated using the distance function, calculating an estimated value of the material from the extracted data, and outputting the estimated value.According to Patent Document 1, the material estimation apparatus having such a configuration prevents an estimation error caused by a divergence between a model structure and a target structure, and improves estimation accuracy in all areas of the input space. It becomes possible to do.

  In addition, for example, the steel product material predicting apparatus disclosed in Patent Document 2 is fitted with data of factors affecting the material of each process of the product of the steel product manufactured over a plurality of processes. An apparatus for predicting using a material prediction model, comprising means for displaying past data as a histogram. In the material predicting apparatus having such a configuration, according to Patent Document 2, a user can easily recognize past abnormal data by referring to a histogram of past data.

  Further, for example, the plant monitoring device disclosed in Patent Document 3 is a plant monitoring device that predicts a state quantity at a future time of an item based on various types of plant data, and stores various types of plant data over a plurality of different times. The history information storage means to store, the time series data stored in the performance information storage means as input, the state quantity prediction means for predicting the state quantity at a future time by a predetermined method, and the state quantity prediction means in the past Prediction error calculation means for calculating a prediction error included in the predicted state quantity prediction value, and a relationship between the state quantity prediction value calculated in the past by the state quantity prediction means and the prediction error at that time is stored. Error information storage means, information stored in the result information storage means, state quantity predicted value calculated by the state quantity prediction means, and error information storage Based on the relationship between the past estimated value and the error stored in the stage, and a prediction error estimation means for estimating a prediction error included in the state quantity prediction values in the future time. In the plant monitoring apparatus having such a configuration, according to Patent Document 3, when performing control or information generation using a predicted value, the range of the prediction error is identified, so the user is more accurate and appropriate. Judgment can be made.

  Further, for example, the operation support device in the continuous treatment facility for metal strip disclosed in Patent Document 4 is an entry side of the continuous treatment facility for metal strip that continuously treats a plurality of metal strips, or a pre-process of the continuous treatment facility. Shape detecting means for continuously or intermittently measuring the shape of the metal strip on the outlet side and outputting a quantitative shape index corresponding to the position of the metal strip through plate, and the width of the metal strip with respect to the shape index Using the direction center as a reference axis, weighting is performed by a predetermined function of the position in the width direction of the metal strip, and the shape index left and right is calculated as actual weight data of the weighted shape index that is an index of the left-right difference of the weighted metal strip. A difference calculating means and a meandering amount in a direction orthogonal to a traveling direction when the metal strip passes through the continuous processing facility is detected. For a new metal strip, a line detection unit, a meandering amount actual data for a plurality of metal strips, and a meandering amount estimation model creating unit for creating a meandering estimation model from the weighted shape index difference data Meandering amount estimation means for estimating a meandering amount when the new metal strip passes through the continuous processing equipment using the meandering estimation model based on the weighted shape index left-right difference obtained in the above, Based on the estimated value of the meandering amount, a production failure risk rate estimating means for estimating a risk rate at which the new metal strip generates a preset type of production failure, and presenting the risk rate to the operator of the continuous processing facility And a risk factor display means. In the aspect, the meandering amount estimation means may calculate the weighted shape index left-right difference of the new metal strip from the weighted shape index left-right difference record data and the meandering amount result data for a plurality of metal strips. Is selected, and based on the frequency distribution of the meandering amount of the selected result data, the range in which the meandering amount of the new metal strip occurs at a preset probability value is estimated. According to the above-mentioned Patent Document 4, the operation support device having such a configuration may impair the processing capacity more than necessary by estimating the amount of meandering of the strip and the risk factor that leads to production failure, and presenting it to the operation operator in advance. Operational support can be provided to avoid commercial troubles and increase productivity as much as possible.

  Also, for example, the converter's blowing end temperature target setting method disclosed in Patent Document 5 is a converter's blowing end temperature target setting method in a steelmaking process having a converter, a secondary refining device, and a continuous casting machine. , The required molten steel temperature at the time of continuous casting, the converter, the secondary refining equipment and the continuous casting machine, the scheduled operation start time, the scheduled operation time, the transport time, and the secondary refining equipment and the continuous casting machine Collecting heat history and actual temperature rise / fall of molten steel, and based on these collected data, the temperature drop of molten steel until the casting time of the continuous caster after the end of converter blowing according to changes in steelmaking process and transportation The amount is calculated, and the sum of the calculated molten steel temperature drop and the continuous casting pouring target temperature is set as the converter's blowing end temperature target. Here, the temperature drop ΔT of molten steel is a change in molten steel temperature that occurs from the end of converter blowing to the start of casting in the continuous casting machine, and the change in molten steel temperature that occurs from the end of converter blowing to the end of ladle receiving steel. The amount of change in molten steel temperature that occurs from the end of the ladle receiving steel to the end of the secondary refining treatment is assumed to be ΔTBP, and the molten steel temperature that occurs from the end of the treatment in the secondary refining device to the start of casting in the continuous casting machine. When the amount of change is ΔTBC and the fluctuation risk term of temperature prediction is δ, it is the sum of ΔTCL, ΔTBP, ΔTBC and δ (ΔT = ΔTCL + ΔTBP + ΔTBC + δ). According to Patent Document 5, the method for setting the end point temperature of the converter in such a configuration omits the temperature measurement at the arrival of the secondary refining equipment after the end of the ladle receiving after the end of the converter blowing. It is possible to set an end point temperature target for the converter in consideration of the fluctuation risk of temperature prediction.

Japanese Patent No. 393441 JP 2006-277374 A JP 09-204218 A JP 2007-046131 A JP 2009-007631 A

  By the way, in the techniques disclosed in Patent Document 1 and Patent Document 2, both predict the predicted value from one point of data. For this reason, it is good if this predicted value is correct, but if this predicted value deviates from the true value, the operations and judgments performed based on this predicted value will be incorrect and appropriate. Can not get the correct output value.

  In particular, depending on the direction of deviation of the true value with respect to the predicted value, that is, the true value is likely to shift upward relative to the predicted value, or the true value is likely to shift downward relative to the predicted value. Therefore, when the operation or determination performed based on the predicted value is different, it is difficult to perform an appropriate operation or determination only with the predicted value. For example, in the manufacturing process of steel products, in the case where it is sufficient if it is not more than a certain standard value, such as impurities in molten steel, the impurity may be added if there is a high possibility that the true value is shifted upward from the predicted value. While it is necessary to perform an operation for removing, if there is a high possibility that the true value is shifted downward with respect to the predicted value, it is not necessary to perform the operation for removing impurities. For example, in the case of a steel product manufacturing process, in the case of the molten steel temperature in the ladle transported from the molten steel processing facility to the continuous casting facility, if the true value is likely to shift downward relative to the predicted temperature, In addition, since there is an operational risk such as casting stoppage due to solidification of the molten steel, the operating conditions for avoiding a decrease in the molten steel temperature are selected, while the true value is likely to shift upward with respect to the predicted temperature. In such a case, so-called breakout in continuous casting is likely to occur, so that the casting speed is adjusted.

  In Patent Document 1, Patent Document 3, Patent Document 4, and Patent Document 5, not only the predicted value but also the prediction error is calculated. The reliability of the predicted value can be understood from this prediction error. However, if the operation or judgment performed based on the predicted value differs depending on the direction of deviation of the true value from the predicted value, an appropriate value depends on the predicted value and the predicted error. It is difficult to make operations and judgments.

In Patent Document 5, the fluctuation risk term δ composed of ΔTBC, ΔTCP, and ΔTCL is taken into consideration, and the fluctuation risk term δ is determined based on each standard deviation δ1 in ΔTBC, ΔTCP, and ΔTCL, The square root of the sum of squares of δ2 and δ3 (δ = 3 × (δ1 ² + δ2 ² + δ3 ² ) ^1/2 ). Therefore, in the said patent document 5, it is assumed that the error of a model follows a normal distribution. However, the model error does not always follow the normal distribution, and the actual variation risk cannot be appropriately expressed by the standard deviation depending on the model error distribution. For example, in the case of the molten steel temperature, the error to the side where the temperature is difficult to decrease is small, and conversely, the error to the side where the temperature is easy to decrease is large, and usually does not have a normal distribution. In addition, when the error of each process is independent of each other, the fluctuation risk term δ can be appropriately expressed based on the sum of squares of the standard deviations δ1, δ2, and δ3. However, the error of each process is not necessarily independent of each other. Not necessarily. For example, when considering the intermediate of a product in the manufacturing process, if the temperature of the intermediate decreases in the previous process and approaches the ambient temperature, in the next process, the intermediate is increased by the amount close to the ambient temperature in the previous process. The temperature of the film becomes difficult to decrease, and errors in each process are not independent of each other. More specifically, for example, when transferring the molten steel to the ladle and the heat transfer from the molten steel to the ladle is good, in this step, the molten steel temperature drops sharply. As a result, since the difference between the molten steel temperature and the ladle temperature is reduced, the decreasing method of the molten steel temperature is slowed down.

  In addition, although the above-mentioned patent document 2 has a description of a histogram, in order to recognize abnormal data, only past data is displayed as a histogram, and variations in predicted values are not displayed. Further, in Patent Document 4, a range in which the new metal strip meander amount occurs at a preset probability value based on the frequency distribution of the meandering amount of the actual data is obtained. This is obtained using the amount of meandering data. Furthermore, since the performance data for obtaining the frequency distribution is selected as data similar to the data of the new metal strip, if there is no data similar to the data of the new metal strip, the frequency distribution of the meandering amount of the performance data Is not required.

  The present invention has been made in view of the above-described circumstances, and its object is to take into account an error parameter composed of a plurality of elements, and thereby to predict the variation of the predicted value regardless of the distribution mode in the variation of the predicted value. An output value prediction method, an output value prediction device, and an output value prediction program that can take into account variations in predicted values when performing operations, determinations, and the like based on predicted values. Is to provide.

  As a result of various studies, the present inventor has found that the above object is achieved by the present invention described below. That is, the output value prediction method according to one aspect of the present invention includes a plurality of past performance data acquired in the past, including prediction target data to be predicted, a predetermined output and a first factor that can be quantified related to the output. A similarity calculation step for calculating each of the plurality of past performance data based on a predetermined second factor for obtaining the similarity in the prediction target data and the past performance data; When a predetermined model is used as an output variable and a predetermined third factor for obtaining a predicted value to be described later is used as an input variable, and the first model representing the relationship between the output variable and the input variable is generated, the input variable The difference between the value obtained by giving the input value to the first model and the output value of the output variable corresponding to the input value of the input variable, A parameter calculating step of calculating an error parameter composed of a first element of each of the plurality of past performance data based on an input variable and an output variable in the past performance data, and using the input variable and the error parameter Generating a second model representing a relationship between the output variable and the input variable, and providing the second model with a value corresponding to the third factor in the prediction target data and a value of the error parameter; A prediction value calculation step for calculating each of a plurality of error parameters calculated by the parameter calculation step using an output value of the prediction target data as a prediction value, a plurality of similarities calculated by the similarity calculation step, and the prediction Based on a plurality of predicted values calculated by the value calculating step, the prediction target data Characterized in that it comprises a variation calculation step of calculating the variation of the output value.

  Moreover, the output value prediction apparatus according to another aspect of the present invention includes measured data that stores a plurality of past performance data acquired in the past, which includes a predetermined output and a first factor that can be quantified related to the output. Based on a predetermined second factor for obtaining the similarity between the storage unit and the prediction target data to be predicted and the past result data, based on a predetermined second factor for obtaining the similarity in the prediction target data and the past result data, the plurality of past results A similarity calculation unit for calculating each of the data, the predetermined output as an output variable, a predetermined third factor for obtaining a predicted value described later as an input variable, and a relationship between the output variable and the input variable When the first model is generated, the value obtained by giving the input value of the input variable to the first model and the output of the output variable corresponding to the input value of the input variable are generated. A parameter which is a difference from a value and which is calculated for each of the plurality of past performance data based on an input variable and an output variable in the past performance data, which is an error parameter including a plurality of first elements set in advance. Using the calculation unit, the input variable and the error parameter to generate a second model representing the relationship between the output variable and the input variable, the value corresponding to the third factor in the prediction target data and the A prediction value calculation unit that calculates each of a plurality of error parameters calculated by the parameter calculation unit using an output value of the prediction target data as a prediction value by giving an error parameter value to the second model; A plurality of similarities calculated by the degree calculating unit and a plurality of predicted values calculated by the predicted value calculating unit Based on, characterized in that it comprises a variation calculation unit for calculating a fluctuation of the output value of the prediction target data.

  An output value prediction program for causing a computer to execute according to another aspect of the present invention includes prediction target data to be predicted, a predetermined output, and a first factor that can be numerically related to the output. For each of the plurality of past performance data based on a predetermined second factor for obtaining the similarity in the prediction target data and the past performance data, the similarity with the plurality of past performance data acquired in A first model representing the relationship between the output variable and the input variable, using the similarity calculation step to calculate, and the predetermined output as an output variable and a predetermined third factor for obtaining a predicted value to be described later as an input variable Is generated, the value obtained by giving the input value of the input variable to the first model and the output of the output variable corresponding to the input value of the input variable. A parameter which is a difference from a value and which is calculated for each of the plurality of past performance data based on an input variable and an output variable in the past performance data, which is an error parameter including a plurality of first elements set in advance. Generating a second model representing a relationship between the output variable and the input variable using the calculation step, the input variable and the error parameter; and a value corresponding to the third factor in the prediction target data and the A prediction value calculation step of calculating each of a plurality of error parameters calculated by the parameter calculation step using the output value of the prediction target data as a prediction value by giving an error parameter value to the second model; A plurality of similarities calculated by the degree calculating step and a plurality of similarities calculated by the predicted value calculating step. Based on the predicted value, characterized in that it comprises a variation calculation step of calculating the variation of the output value of the prediction target data.

  Even if the predetermined output (predicted value) is predicted by an assumed third factor related to the predetermined output, an unexpected uncertain element or model such as a fluctuation, disturbance, or a factor that cannot be clarified at the present time. An error α exists between the predicted value and the true value due to uncertain factors such as a conversion error. That is, even if the predetermined output y is modeled by the first model; y = f (Z, Θ) by the third factor Z related to the predetermined output y, which is an assumed factor, Z and Θ An error α exists between the predicted value and the true value due to an uncertain element that cannot express the output y alone. Note that Θ is a predetermined adjustment parameter.

  This uncertain element is a factor of variation. In the output value prediction method, the output value prediction apparatus, and the output value prediction program having the above-described configuration, an error related to this uncertain element is an error composed of a plurality of first elements. The error parameter A is calculated for each of the plurality of past performance data based on the past performance data, and a second model is created with the error parameter A taken into account. Are calculated for each of the plurality of error parameters A as predicted values. And the dispersion | variation in the output value of prediction object data is calculated based on several similarity and several predicted value.

  Therefore, in the output value prediction method, the output value prediction apparatus, and the output value prediction program configured as described above, the predicted value variation is calculated with higher accuracy regardless of the distribution mode of the predicted value variation, and as a result, based on the predicted value. Thus, it is possible to take into account variations in predicted values when performing operations and determinations.

  Here, the factors that can be quantified include not only physical quantities that can be measured by a measuring instrument, but also, for example, each individual of an operation group that executes a process, each individual of equipment that is used to execute a process, and the like. Such quantification of each individual is, for example, a predetermined value a (for example, 1) when involved in the process and another predetermined value b (for example, 0) different from the predetermined value when not involved in the process. To be executed. For example, when there are four teams A, B, C, and D and the team A is involved, the data (input value) of the team A becomes the predetermined value a (for example, 1) and the other team B Or each data (input value) of the D group becomes the other predetermined value b (for example, 0).

  Of course, the difference between the value obtained by giving the input value of the input variable to the first model and the output value of the output variable corresponding to the input value of the input variable is, of course, the input variable and the output. If the first model for a variable is expressed exponentially, its logarithmic difference is also included.

  In another aspect, in the above-described output value prediction method, the predicted value calculation step generates a second model that represents the relationship between the output variable and the input variable using the input variable and the error parameter. Then, for each of the plurality of first elements of the error parameter, the value of the factor corresponding to the input variable among the factors of the prediction target data and the first of the plurality of first elements of the error parameter. The value of the element is given to the second model to calculate the value of the second model, and the A1 step corresponding to each of the plurality of first elements of the error parameter is calculated. A second step of calculating the predicted value by performing a weighted average of a plurality of values, wherein the A1 step and the A2 step are performed by the parameter calculation step. Characterized in that it is performed for each of a plurality of error parameters issued.

According to this configuration, the value of the second model is _obtained for each of the plurality of first elements (parameter elements) α _{Nα in} the error parameter A, and the plurality of values of the obtained second model are weighted averaged. For this reason, this weighted average process ensures the relevance of the plurality of parameter elements α _Nα , and even when the error parameter is composed of a plurality of parameter elements including parameter elements that are related to each other, the variation in the predicted value becomes more accurate. It is possible to take into account variations in the predicted value when an operation or determination is performed based on the calculated value and the predicted value.

Here, the sum of each weight ν _k in the weighted average process is fixed to a predetermined constant value.

  According to another aspect, in the above-described output value prediction method, the variation is a histogram, and the variation calculation step calculates the similarity to a plurality of prediction values calculated by the prediction value calculation step. A first B1 step that associates each of the plurality of similarities calculated in the step, a second B2 step that divides a range including at least the plurality of predicted values into a finite number of sections, and a predicted value included in the section. B3 process which calculates the frequency of the said area | region about each of these several area | regions by adding up all the corresponding similarities, It is characterized by the above-mentioned.

  According to this configuration, the variation is indicated by the histogram, and the appearance frequency of the predicted value can be easily known.

  According to another aspect, in the output value prediction method described above, the variation is a probability density, and the variation calculation step further sets the scale of the frequency so that the area of the histogram is 1. It is characterized by including the B4 process to adjust.

  According to this configuration, the variation is indicated by the probability density, and the appearance probability of the predicted value can be easily known.

  Further, in another aspect, in the above-described output value prediction method, the similarity calculation step further includes a step of converting a predetermined constant or a predetermined constant ratio of similarity into 0 from the smaller one, or A step of converting similarity below a predetermined threshold value to 0 is provided.

  According to this configuration, in obtaining the predicted value, it is possible to prevent the past performance data that is not very similar to the prediction target data from being considered more than necessary. In addition, past performance data that is not very similar to the prediction target data is excluded, and subsequent calculation processing becomes unnecessary, and as a result, the calculation processing amount can be reduced (calculation processing time can be reduced).

  In another aspect, in the above-described output value prediction method, in the similarity calculation in the similarity calculation step, the past performance data closer in time to the prediction target data is increased in similarity. It is characterized by containing predetermined evaluation items.

  According to this configuration, it is possible to perform robust prediction when the characteristics of a system in which variations in output values are predicted change due to, for example, fluctuations in equipment or operation, or changes with time.

  In another aspect, in the above-described output value prediction method, the first factor that can be numerically related to the output includes a plurality of second elements, and includes at least time as the second element. And

  According to this configuration, it is possible to obtain a predicted value of an output value in a process in which the output changes from time to time, and to obtain a variation in the predicted value.

  According to another aspect, the above-described output value prediction method further includes a presentation step of presenting the variation.

  According to this configuration, the user can know the variation and can also consider the variation in the predicted value when performing an operation, a determination, or the like based on the predicted value.

  According to another aspect, in the output value prediction method described above, in the presentation step of presenting the variation, an auxiliary display representing a portion outside the predetermined management range set in the predetermined output is presented as the variation. It is characterized by presenting it according to.

  According to this configuration, the user can know the relationship between the variation and the portion outside the predetermined management range, and when performing an operation or determination based on the predicted value, the variation in the predicted value is taken into consideration. However, it is possible to determine conditions relating to the above factors such as operating conditions. For this reason, it is possible to make more appropriate decision making, and it is possible to grasp information that is difficult to understand simply by looking at the shape of the variation by the auxiliary display.

  Here, the auxiliary display includes various expression methods, and includes, for example, a numerical value representing a portion outside the predetermined management range, a line segment (boundary line), a texture, and the like.

  Moreover, in another aspect, in the output value prediction method described above, the auxiliary display is the above-described predetermined upper limit value or more when the variation is a histogram or a probability density and is presented by a graph. It is the 1st area display regarding the area of a graph, and / or the 2nd area display regarding the area of the said graph below the predetermined predetermined lower limit.

  According to this configuration, the user can know a portion outside the predetermined management range as a risk, for example, by the first area display and / or the second area display, and perform an operation, a determination, or the like based on the predicted value. In this case, it is possible to determine a condition related to the factor such as an operating condition in consideration of, for example, a risk that deviates from the predetermined management range while considering a variation in predicted values.

  Here, A and / or B means at least one of A and B.

  According to another aspect, in the output value prediction method described above, the auxiliary display is the histogram or the probability density, and when the variation is presented by a graph, the auxiliary display is based on the total area As in the variation graph. It is a boundary value corresponding to an area ratio as / As of a predetermined area as set in advance from one end of the variation graph.

  According to this configuration, the user can know the deviation rate, which is a rate outside the predetermined management range, from the boundary value, and when performing an operation or a determination based on the predicted value, the user can take the risk of the predicted value being off. It is possible to determine conditions relating to the factors such as operating conditions so that the deviation rate becomes an appropriate value while taking into consideration.

  According to another aspect, in the output value prediction method described above, the auxiliary display is the histogram or the probability density, and when the variation is presented by a graph, the auxiliary display is based on the total area As in the variation graph. It is a difference between a boundary value corresponding to an area ratio as / As of a predetermined area as set in advance from one end of the variation graph and a management value set in advance.

  According to this configuration, the user can know the deviation of the deviation rate from the difference between the boundary value and the management value, and considers the risk of deviation of the predicted value when performing an operation or determination based on the predicted value. However, it is possible to determine conditions relating to the factors such as operating conditions so that the deviation rate becomes an appropriate value.

According to another aspect, in the output value prediction method described above, when the variation is a histogram or a probability density and is presented by a graph, the auxiliary display is preliminarily applied to the total area As in the variation graph. The first area ratio au1 / As, which is the ratio of the first area au1 of the graph above the set predetermined upper limit value, and the predetermined area below the predetermined lower limit value with respect to the total area As in the variation graph. characterized in that it is a sum au1 / as + au2 / as or square sum ^{(au1 / as) 2 + (} au2 / as) 2 and the second area ratio au2 / as is the ratio of the second area au2 graphs.

According to this configuration, the user can set the predetermined target as the sum au1 / As + au2 / As of the first area ratio au1 / As and the second area ratio au2 / As or the degree of detachment of the portion outside the predetermined management range. The deviation rate can be determined by taking into account the risk of divergence due to variations in predicted values when performing operations, judgments, etc. based on the predicted values, which can be known from the sum of squares (au1 / As) ² + (au2 / As) ² For example, it is possible to determine conditions relating to the above factors such as operation conditions so that the value becomes an appropriate value.

  According to another aspect, in the output value prediction method described above, the auxiliary display is calculated by the predicted value calculation step in a coordinate system in which the variation has a predicted value on one coordinate axis and a similarity on the other axis. When the plurality of predicted values and the plurality of similarities calculated in the similarity calculating step are plotted, the difference between the management value set in advance and the predicted value is calculated for the plurality of predicted values. The sum of the values obtained by multiplying the absolute value by the similarity of the predicted value (management value reference moment) or the sum of the squares of the difference between the managed value and the predicted value multiplied by the similarity of the predicted value (managed value) Reference moment) or a sum of values obtained by multiplying the difference between the management value and the prediction value by the similarity of the prediction value (management value reference moment).

  According to this configuration, the user can know the deviation of the variation in the management value from the management value reference moment, and consider the variation in the prediction value when performing an operation or a determination based on the prediction value. On the other hand, it is possible to determine conditions relating to the factors such as operation conditions so that the predetermined output is more appropriately distributed around the management value.

  According to another aspect, in the output value prediction method described above, the auxiliary display is calculated by the predicted value calculation step in a coordinate system in which the variation has a predicted value on one coordinate axis and a similarity on the other axis. When the plurality of predicted values and the plurality of similarities calculated by the similarity calculating step are plotted, the prediction is performed with an upper limit value equal to or higher than a preset upper limit value among the plurality of predicted values. For the value, the sum of values obtained by multiplying the absolute value of the difference between the upper limit value and the upper limit value or more predicted value by the similarity of the upper limit value or higher predicted value (upper limit reference moment) or the upper limit value and the upper limit value or higher It is a sum (upper limit reference moment) of a value obtained by multiplying the square of the difference from the predicted value by the similarity between the predicted value and the predicted value.

  According to this configuration, the user can know the risk of deviating from the upper limit value and the degree of detachment from the upper limit value reference moment, and consider the variation of the predicted value when performing an operation or determination based on the predicted value. However, it is possible to determine conditions relating to the factors such as operating conditions so that the risk of deviating the predetermined output from the upper limit value is more appropriate.

  According to another aspect, in the output value prediction method described above, the auxiliary display is calculated by the predicted value calculation step in a coordinate system in which the variation has a predicted value on one coordinate axis and a similarity on the other axis. When the plurality of predicted values and the plurality of similarities calculated by the similarity calculating step are plotted, prediction below the lower limit value below the preset lower limit value among the plurality of predicted values For the value, the sum of values obtained by multiplying the absolute value of the difference between the lower limit value and the predicted value below the lower limit value by the similarity of the predicted value below the lower limit value (lower limit reference moment) or the lower limit value and the lower limit value or less It is the sum (lower limit reference moment) of the value obtained by multiplying the square of the difference from the predicted value by the similarity of the predicted value below the lower limit value.

  According to this configuration, the user can know the risk of deviating from the lower limit value and the degree of deviation from the lower limit reference moment, and take into account variations in the predicted value when performing operations, determinations, etc. based on the predicted value. However, it is possible to determine conditions relating to the factors such as operating conditions so that the risk of deviating the predetermined output from the lower limit value is more appropriate.

  According to another aspect, in the output value prediction method described above, the auxiliary display is calculated by the predicted value calculation step in a coordinate system in which the variation has a predicted value on one coordinate axis and a similarity on the other axis. A plurality of predicted values and a plurality of similarities calculated by the similarity calculating step are plotted, and an upper limit value equal to or higher than a preset upper limit value among the plurality of predicted values With respect to the predicted value, the sum of values obtained by multiplying the absolute value of the difference between the upper limit value and the predicted value equal to or higher than the upper limit value by the similarity of the predicted value equal to or higher than the upper limit value or the upper limit value and the predicted value equal to or higher than the upper limit value. The sum of values obtained by multiplying the square of the difference by the similarity between the upper limit value and the predicted value is set as the upper limit value reference moment, and the predicted value below the lower limit value below the preset lower limit value among the plurality of predicted values. The lower limit and the lower The sum of values obtained by multiplying the absolute value of the difference between the lower limit value and the predicted value below the lower limit value or the square of the difference between the lower limit value and the predicted value below the lower limit value is calculated by multiplying the absolute value of the difference between the lower limit value and the predicted value below the lower limit value. When the sum of the values multiplied by the similarity is used as the lower limit reference moment, it is the sum of the upper limit reference moment and the lower limit reference moment.

  According to this configuration, the user can know the risk of deviating from the upper and lower limit values and the degree of deviation from the sum of the upper limit value reference moment and the lower limit value reference moment. When performing the above, it is possible to determine conditions relating to the factors such as operating conditions so that the risk that the predetermined output deviates from the upper and lower limit values becomes more appropriate in consideration of variations in predicted values. .

  Moreover, in another one aspect | mode, in these above-mentioned output value prediction methods, the said predetermined output is in the ladle or in a tundish in the process from a converter steeling process through a molten steel treatment process to a continuous casting process. It is a molten steel temperature.

  According to this configuration, the molten steel temperature in the ladle or in the tundish is predicted in the process from the converter steelmaking process through the molten steel treatment process to the continuous casting process. It is possible to provide an output value prediction method that can be obtained.

  Moreover, in another aspect, in the above-described output value prediction method, the predetermined output is a molten steel component or a molten steel temperature according to an integrated amount of blown blown oxygen in the converter process. To do.

  According to this structure, in the converter process, the molten steel component or molten steel temperature corresponding to the integrated amount of blown blown oxygen is predicted, and the output value prediction method that can more appropriately determine the variation in the predicted value. Can be provided.

  Moreover, in another one aspect | mode, in these above-mentioned output value prediction methods, the said predetermined output is the steel material temperature of the said steel material according to the integration amount of the heating time or the heating calorie | heat amount in the heating furnace process of steel materials. Features.

  According to this configuration, the steel material temperature of the steel material is predicted according to the heating time or the integrated amount of heating heat in the steel heating furnace process, and the output value that can more appropriately determine the variation of the predicted value predicted. A prediction method can be provided.

  The output prediction method, the output prediction apparatus, and the output prediction program according to the present invention can more appropriately determine the variation in the prediction value regardless of the distribution mode in the variation in the prediction value. When making a determination or the like, it is possible to take into account variations in predicted values.

It is a block diagram which shows the structure of the output value prediction apparatus in 1st Embodiment. It is a flowchart which shows operation | movement of the output value prediction apparatus in 1st Embodiment. It is a figure which shows the data memorize | stored in the actual measurement data memory | storage part in 1st Embodiment. It is a figure for demonstrating the Euclidean distance of prediction object data and each past performance data. It is a figure which shows the data memorize | stored in the intermediate data storage part in 1st Embodiment. It is a figure which shows the data memorize | stored in the predicted value memory | storage part in 1st Embodiment. It is a figure for demonstrating the calculation procedure of the dispersion | variation in the predicted value in 1st Embodiment. It is a figure for demonstrating the method of calculating | requiring the probability density curve shown in FIG.7 (C) from the histogram shown in FIG.7 (B). It is a figure for demonstrating the auxiliary display of a 1st aspect. It is a figure for demonstrating the auxiliary display of a 2nd aspect. It is a figure for demonstrating the auxiliary display of a 3rd aspect. It is a figure which shows the relationship between the temperature fall amount of an object, and elapsed time. It is a figure which shows the data memorize | stored in the intermediate data storage part in 2nd Embodiment. It is a figure which shows the model of the past performance data in 2nd Embodiment. It is a figure which shows the data memorize | stored in the predicted value memory | storage part in 2nd Embodiment. It is a figure which shows the prediction model in 2nd Embodiment. It is a figure for demonstrating the calculation procedure of the dispersion | variation in the predicted value in 2nd Embodiment. It is FIG. (1) which shows the relationship between the weighted distance and similarity in 3rd Embodiment. It is a figure (the 2) which shows the relationship between the weighted distance and similarity in 3rd Embodiment. It is a figure which shows the model of the past performance data in 3rd Embodiment. It is a figure which shows the probability density in each estimated value in 3rd Embodiment. It is a block diagram which shows the structure of an output value prediction system.

DESCRIPTION OF EXEMPLARY EMBODIMENTS Hereinafter, an embodiment of the invention will be described with reference to the drawings. In addition, the structure which attached | subjected the same code | symbol in each figure shows that it is the same structure, The description is abbreviate | omitted.
(First embodiment)
First, the configuration of the output value prediction apparatus S will be described. FIG. 1 is a block diagram illustrating a configuration of an output value prediction apparatus according to the first embodiment. In FIG. 1, the output value prediction apparatus S includes an arithmetic control unit 1, an input unit 2, a presentation unit 3, and a storage unit 4.

  The input unit 2 outputs various commands such as a command for starting an output value prediction program for predicting an output value by using the method of the present invention from previously given data, and various data necessary for predicting the output value. A device that inputs to the prediction device S, such as a keyboard or a mouse. The presenting unit 3 is a device that presents (outputs) the command and data input from the input unit 2 and the output value (predicted value) predicted by the output value predicting apparatus S. For example, the presenting unit 3 includes a CRT display, LCD, A display device such as an organic EL display and a plasma display, and a printing device such as a printer.

  The storage unit 4 is functionally measured data storage that stores a plurality of past performance data acquired in the past and prediction target data for which an output value is to be predicted, which is functionally composed of a predetermined output and a quantifiable factor related to the output. Unit 41, intermediate data storage unit 42 for storing intermediate data generated in the process of predicting an output value based on past performance data from prediction target data, and prediction based on past performance data from prediction target data A prediction value storage unit 43 that stores (calculated) output values (prediction values), and a variation storage unit 44 that stores variation in prediction values, and various programs such as an output value prediction program; It is a device that stores various data such as data necessary for execution and data generated during the execution. The storage unit 4 is, for example, a volatile storage element such as a RAM (Random Access Memory) serving as a so-called working memory of the arithmetic control unit 1, a ROM (Read Only Memory), or a rewritable EEPROM (Electrically Erasable Programmable Read Only Memory). And the like, and a hard disk for storing various programs and various data.

  The arithmetic control unit 1 includes, for example, a microprocessor and its peripheral circuits, and functionally includes a distance calculation unit 11, a similarity calculation unit 12, a parameter calculation unit 13, and a predicted value calculation unit 14. And a variation calculating unit 15 for controlling the input unit 2, the presenting unit 3 and the storage unit 4 according to the function according to the control program.

  The distance calculation unit 11 calculates a predetermined distance between the prediction target data and the past performance data for each of the plurality of past performance data based on the factors of the prediction target data and the past performance data.

  The similarity calculation unit 12 calculates the similarity between the prediction target data and the past performance data for each of the plurality of past performance data based on the plurality of distances calculated by the distance calculation unit 11. .

  The parameter calculation unit 13 uses the input variable as the output variable y and the input variable when the predetermined output is used as an output variable and some or all of the quantifiable factors related to the predetermined output are used as input variables. The first model representing the relationship with Z; when y = f (Z, Θ) is generated, the value obtained by giving the input value of the input variable to the first model and the output corresponding to the input value of the input variable An error parameter A, which is a difference from the output value of the variable, is calculated for each of a plurality of past result data based on the input variable and output variable of the past result data.

Here, it should be noted that the error parameter A includes a predetermined plurality of predetermined Nα elements (parameter elements) α (α _Nα ) (A _j = [α _j1 , α _j2). ,..., Α _jNα ]). The plurality of parameter elements α _i in the error parameter A may include parameter elements that are related to each other. The plurality of parameter elements α _i may have a relationship in which all parameter elements are associated with one or more other parameter elements, and some parameter elements are associated with one or more other parameter elements. You may be in a relationship. Further, from the viewpoint of more appropriately obtaining the variation of the predicted value y ₀ , the plurality of parameter elements α are preferably elements (predictable parameters) that can be predicted by using, for example, physical laws, empirical rules, or statistical methods. Element) and other elements (residual parameter elements). This residual parameter element includes an unpredictable element. The residual parameter element may include such a predictable element because it is not recognized as a predictable element, for example.

The predicted value calculation unit 14 generates a second model representing the relationship between the output variable y and the input variable Z using the input variable Z and the error parameter A; y = (Z, Θ, A), and A plurality of error parameters α calculated by the parameter calculation unit 13 using the output value of the prediction target data as a predicted value by giving the value of the factor corresponding to the input variable of the factors and the value of the error parameter A to the second model. _This is calculated for each _i .

As described above, since the error parameter A includes a plurality of parameter elements α _ji , for example, in the present embodiment, the predicted value calculation unit 14 determines each of the plurality of parameter elements α _i of the error parameter A. , The value of the factor corresponding to the input variable among the factors of the prediction target data and the value of the parameter element α of the plurality of parameter elements α _i of the error parameter A are expressed as the second model; y = (Z, Θ, A ), That is, more specifically, by leaving only the value of the parameter element α as it is and setting the others to 0, the value of the second model is calculated, and a plurality of parameters of the error parameter A are calculated. corresponding to each element alpha _i, to calculate the predicted value y ₀ by weighted averaging the calculated plurality of values.

  The variation calculating unit 15 calculates the variation in the output value of the prediction target data based on the plurality of similarities calculated by the similarity calculating unit 12 and the plurality of predicted values calculated by the predicted value calculating unit 14. is there.

  The arithmetic control unit 1, the input unit 2, the presentation unit 3, and the storage unit 4 are connected by a bus 5 so that signals can be exchanged with each other.

  The arithmetic control unit 1, the input unit 2, the presentation unit 3, the storage unit 4, and the bus 5 can be configured by, for example, a computer, more specifically, a personal computer such as a notebook type or a desktop type.

  Note that the output value prediction apparatus S may further include an external storage unit as necessary. The external storage unit is, for example, a flexible disk, a CD-ROM (Compact Disc Read Only Memory), a CD-R (Compact Disc Recordable), a DVD-R (Digital Versatile Disc Recordable), and a Blu-ray Disc (registered trademark). For example, a flexible disk drive, a CD-ROM drive, a CD-R drive, a DVD-R drive, a Blu-ray disk drive, and the like are devices that read data from and / or write data to / from a recording medium.

  Here, when the output value prediction program or the like is not stored, the output value prediction program is installed in the storage unit 4 via the external storage unit from the recording medium on which the output value prediction program or the like is recorded. The output value prediction device S may be configured. Alternatively, the output value prediction device S may be configured such that data such as past performance data and data for predicting an output value is recorded on a recording medium via an external storage unit.

Next, the operation of the output value prediction apparatus S will be described. FIG. 2 is a flowchart showing the operation of the output value prediction apparatus in the first embodiment. FIG. 3 is a diagram illustrating data stored in the actual measurement data storage unit according to the first embodiment. FIG. 4 is a diagram for explaining the Euclidean distance between the prediction target data and each past performance data. FIG. 5 is a diagram illustrating data stored in the intermediate data storage unit in the first embodiment. FIG. 6 is a diagram illustrating data stored in the predicted value storage unit according to the first embodiment. FIG. 7 is a diagram for explaining a procedure for calculating a variation in predicted values in the first embodiment. FIG. 7A shows the relationship between the similarity w and the output predicted value y ₀ , the horizontal axis is the similarity w, and the vertical axis is the predicted value y ₀ . FIG. 7B shows a histogram of the predicted value y ₀ , the horizontal axis is the weighted frequency Fw, and the vertical axis is the predicted value y ₀ . FIG. 7C shows the probability density distribution of the predicted value y ₀ , the horizontal axis is the probability density P (y ₀ ), and the vertical axis is the predicted value y ₀ .

  For example, when an output command is received from the input unit 2 by a user operation, the output value prediction device S executes an output value prediction program. By executing this output prediction program, the calculation control unit 1 functionally includes a distance calculation unit 11, a similarity calculation unit 12, a parameter calculation unit 13, a predicted value calculation unit 14, and a variation calculation unit 15. And the output value prediction apparatus S estimates an output value (predicted value) from prediction object data based on past performance data by the following operations.

In the prediction of the output value, the actual measurement data storage unit 41 in the storage unit 4 of the output value prediction apparatus S stores, for example, past performance data (X, y) and prediction target data in the table format (table format) shown in FIG. _x0n is stored in advance. Here, the predetermined output y is a set of M elements (output elements) y _j , that is, y = {y _j | j = 1, 2,..., M}, The factor X involved in the output y is a set of N elements x _ji , that is, X = {x _ji | i = 1, 2,..., N}.

The actual measurement data table 51 shown in FIG. 3 includes an output field 511 for registering the actually measured output value y, and a data field 512 for registering the data x _ji of the factor X related to the predetermined output y. is configured to include, includes a record for each historical track record data (X, y), further comprises a record of the predicted target data X _0. Incidentally, the predicted target data X _0, in order to distinguishably distinguished historical performance data (X, y) and, 0 is assigned as the first subscript (left subscript), historical performance data (X, y ), 1 to M are assigned as first subscripts to distinguish the M pieces of data in an identifiable manner. Then, the predicted target data X ₀ and historical performance data (X, y), said predetermined first or for the second N data items distinguishably distinguished from each other in the factor X associated with the output y is the N elements 1 to N are added as second subscripts (the right side of the subscripts). X ₀ = [x ₀₁ , x ₀₂ ,..., X _0N ], and X _ji = [x _j1 , x _j2 ,..., X _jN ]. For example, y ₃ represents the third output value in the past performance data (X, y), and for example, x ₂₃ represents the second third data item in the past performance data (X, y). represents the value, also for example, x ₀₄ represents the value of the fourth data item in the prediction target data X _0.

As described above, the factor X related to the predetermined output y is configured to include a plurality of N elements (data items, factor elements). For this reason, the data field 512 includes data corresponding to the number N of elements. It has an item subfield. In the example shown in FIG. 3, the predetermined output y involves at least N elements (first to Nth data items). Therefore, the data item subfield has a first or data item subfield 5121~512N for registering each respective data _x j1 _{~x jN} of the N data items correspond to each element _{x ji} factor X . The past performance data (X, y) is data obtained by actual measurement or the like at different times (time points) t _j in the past under different conditions in the past. In the example shown in FIG. It consists of data. Predicted target data X _0, the a predetermined data x _0i for which you want to predict the value y ₀ in the output y, for example, and the data x _0i that is actually measured to predict the time t _0, the value x ₀ Ya operation input , and the value x ₀ of the operating conditions, and operation date and time x _0, the data x ₀ like prepared for simulation.

Here, the output value prediction apparatus S uses the entire Z or a part Z of the data values x _0i of the prediction target data X ₀ and outputs the output value (predicted value) y ₀ based on the past performance data (X, y). It predicted, and requests a variation of the predicted value y _0. In this case, in the output value prediction apparatus S of the present embodiment, an error parameter described later is assumed to be composed of a plurality of elements (parameter elements), and the plurality of parameter elements include parameter elements that are related to each other. Is done.

In the case where such a historical record data (X, y) and the predicted target data X ₀ is stored in the measured data storage unit 41, an output value prediction device S, as shown in FIG. 2, initially, historical performance data In order to evaluate the relationship between (X, y) and the prediction target data X ₀ , the distance between the two data is calculated by the distance calculation unit 11 of the calculation control unit 1, and the calculated distance is stored in the middle of the storage unit 4. The data is stored in the data storage unit 42 (S11). The distance is defined so as to represent the relationship between both data, and for example, Euclidean distance, weighted Euclidean distance, normalized Euclidean distance, or the like is used. The data x used to calculate this distance is all or part x of the factor X related to the predetermined output y, and all or part of the factor X related to the predetermined output y. It may coincide with a certain Z, may not coincide, or may partially coincide (partly disagree).

More specifically, in the present embodiment, the distance calculation unit 11 calculates the Euclidean distance d between the prediction target data X ₀ (♦) and the past performance data X (◯) in the data item space, as shown in FIG. It calculates about each past performance data (X, y). For example, in this aspect, the data x used to obtain the Euclidean distance d is all the factors X related to the predetermined output y, and the data item space has N data items. It becomes a dimensional space. In the present embodiment, the Euclidean distance d is a weighted distance, and is obtained by, for example, Expression 1-1. In Formula 1-1, the weighted Euclidean distance d _j between the prediction target data X ₀ and the j-th past performance data X _j is the j-th past performance in the data x used to obtain the Euclidean distance d. The square of the difference between the i-th data item x _ji in the data (X, y) and the i-th data item x _0i in the prediction target data X ₀ is set to the weight a _i of the i-th data item (weights a _i and a related to the distance). _i ≧ 0) is multiplied by the square of the first data item to the Nth data item, and the square root of the result is calculated.

Further, for example, the Euclidean distance d may be obtained by Expression 1-2. In Formula 1-2, the weighted Euclidean distance d _j between the prediction target data X ₀ and the j-th past performance data X _j is the j-th past performance in the data x used to obtain the Euclidean distance d. data (X, y) to the difference between the i-th data item _{x 0i} in the i data item _{x ji} and the prediction target data _{X 0} in the absolute value of the value obtained by multiplying the weight _{a i} of the i data item, first Calculated by taking the sum from the data item to the Nth data item.

Further, for example, the Euclidean distance d may be obtained by Expression 1-3. In Expression 1-3, the weighted Euclidean distance d _j between the prediction target data X ₀ and the j-th past performance data X _j is the first data item from the first data item in the data x used to obtain the Euclidean distance d. For each of up to N data items, the difference between the i-th data item x _ji in the j-th past performance data (X, y) and the i-th data item x _0i in the prediction target data X ₀ is The absolute value of the value obtained by multiplying the weight a _i is obtained, and the maximum value is obtained from the obtained absolute values. Note that max (A, B) is an operator for obtaining the maximum value in the set {A, B}.

Here, when each variable x _i is normalized, the weight a _i is the importance (degree of importance) of the i-th data item in the first to N-th data items in obtaining the Euclidean distance d. Which represents the degree to which the factor (i.sup.th data item _x.sub.i ) contributes to the magnitude of variation in the predetermined output y. For example, when the factor element (i-th data item x _i ) is standardized or normalized, the weight a _i increases as the data item that has a large degree of influence on the predetermined output y increases. It is set so that the smaller the data item that affects the output y, the smaller the data item. This weight a _i can be determined by using a known method, for example, can be obtained by multiple regression analysis as disclosed in Japanese Patent No. 393441, and for example, Japanese Patent No. 3912215. As disclosed in the document, each data item can be obtained by normalizing it with its statistical value (for example, average value or standard deviation).

Then, the distance calculation unit 11 stores the distances d _j which is the calculated intermediate data storage unit 42 of the storage unit 4.

Next, in order to evaluate how much the output value prediction device S is similar to the prediction target data X ₀ , the output value prediction device S uses the first to M-th past results as the similarity (degree of similarity) w between the two data x. Each of the data (X, y) is calculated by the similarity calculation unit 12 of the calculation control unit 1, and each calculated similarity w is stored in the intermediate data storage unit 42 of the storage unit 4 (S12). The similarity w _j is assumed to be a condition that the prediction target data X ₀ and the j-th past performance data X _j are more similar as the weighted Euclidean distance d _j is smaller, for example. The smaller the added Euclidean distance d _j is, the larger the similarity is and the positive value is defined.

More specifically, the similarity calculation unit 12 calculates, for example, the similarity w _j using Equation 2-1. Further, for example, the similarity calculation unit 12 calculates the similarity w _j by using the expression 2-2. Further, for example, the similarity calculation unit 12 calculates the similarity w _j by using the expression 2-3, for example. In calculating the similarity w _j , for example, a calculation method disclosed in Japanese Patent Application Laid-Open No. 2004-355189 can be referred to.

Here, w _j is the similarity of the j-th past performance data X with respect to the prediction target data X ₀ , σ is a normalization parameter, specifically, d _j (j = 1 to M) Is a standard deviation, c ₁ is an adjustment parameter of a real number greater than or equal to 0, c ₂ is an adjustment parameter of a real number (which may be a negative value), and g and r are adjustment parameters of a positive real number .

Then, the similarity calculation unit 12 stores the calculated similarities w _j in the intermediate data storage unit 42 of the storage unit 4.

In addition, when the upper limit value of the similarity score w _j and / or the lower limit value thereof is provided and the similarity score w calculated by any one of the formulas 2-1 to 2-3 exceeds the upper limit value, the similarity score When w _j is replaced with the upper limit value and / or the similarity score w _j calculated by any one of Formula 2-1 to Formula 2-3 exceeds the lower limit value, the similarity score w _j is The similarity calculation unit 12 may be configured to be replaced with a lower limit value. With such a configuration, it is possible to prevent only the specific past performance data X from excessively increasing the similarity w _j or conversely decreasing. If only the specific past performance data X has an excessive similarity w _j , if there is an error in the data measurement value, it will be pulled by that error, and the wrong variation will be predicted. Will go. For this reason, as described above, setting the upper limit value has an effect of becoming stronger in error.

Further, for example, when a predetermined threshold value is provided in advance and the similarity score w _j calculated by any one of the formulas 2-1 to 2-3 is equal to or less than the threshold value, the similarity score w _j is replaced with 0. As described above, the similarity calculation unit 12 may be configured. Alternatively, ordered similarity w _j is less calculated by any of formulas 2-1 or Formula 2-3, the similarity w _j to a preset predetermined number from the smallest (or a predetermined rate) 0 The similarity calculation unit 12 may be configured to be replaced with By being constructed as described above, when obtaining the prediction values y _0, it is possible to prevent to consider more than necessary historical performance data X which is not much similar to the predicted target data X _0. Also, it excluded historical performance data X which is not much similar to the predicted target data X ₀ is, the arithmetic processing is not required to be described below, as a result, the arithmetic processing amount reduction (shortening of processing time) becomes possible.

Further, in the case of calculating the similarity w _j, the predicted target data X ₀ and temporally past as actual data X near causes increase its similarity w _j (raised to) evaluation items in the calculation of the similarity w _j It is preferable to enter. For example, one or more evaluation items for increasing the similarity w _j may be included in the data x used for obtaining the Euclidean distance d, and, for example, the equation 2 for calculating the similarity w -1 to Formula 2-3. More specifically, the correction term exp (− (wday × Δday _j ) ² ) is multiplied by the expressions 2-1 to 2-3. Here, wday is the weight of the adjustment parameter, △ day _j is an operational days difference between the j historical performance data _{X j} obtained at time _{t j} and the predicted target data _{X 0.} By configuring in this way, for example, when the characteristics of a system (for example, a manufacturing plant or a manufacturing process) in which variations in output values are predicted are changed due to a change in equipment or operation or a change over time, it is robust. Prediction can be performed.

  Next, the output value prediction device S calculates an error parameter A representing an indeterminate element by the parameter calculation unit 13 of the arithmetic control unit 1 for each of the first to M-th past performance data (X, y), The calculated error parameters A are stored in the intermediate data storage unit 42 of the storage unit 4 (S13).

  More specifically, the parameter calculation unit 13 obtains a prediction model (second model) for predicting the output value y based on the M past performance data (X, y), and uses the obtained model. Each error parameter A is obtained for each of the first to Mth past performance data (X, y).

  This prediction model is expressed by, for example, the function f in Expression 3. In this case, the parameter calculation unit 13 obtains Θ of the function formula 3 based on M pieces of past performance data (X, y), and uses the obtained function formula 3 to obtain the first to Mth past results. Each error parameter A is obtained for each of the data (X, y). Θ may include a physical constant, a physical property value, a value obtained from other data, an empirically determined value, and a theoretically determined value.

Here, Z is an output value y ₀ of a factor X that can be quantified related to the predetermined output y, such as each condition of operation conditions (manufacturing conditions) and each measurement item in each process of the manufacturing process. This is a factor used for prediction, and is configured to include a plurality of L elements z (Z _j = [z _j1 , z _j2 ,..., Z _jL ]). Z is constituted by a part or all of data items (X _j = [x _j1 , x _j2 ,..., X _jN ]) which are a plurality of elements in the factor X related to the predetermined output y, for example. The The Z may further include elements other than the data item X. As described above, factors Z used to predict the output value y ₀ is a data x that is used to determine the Euclidean distance d, may be coincident, also be a mismatch Also, it may be partially coincident (partly inconsistent). This Z is set in advance when the expression 3 of the function f is determined. Θ is a predetermined adjustment parameter such as a coefficient of the function formula 3, and is obtained by identification calculation based on M pieces of past performance data (X, y), for example. In the case of this identification, the actual value y _j and the predicted value y _{0 of the} predetermined output y such as the least square method, the maximum likelihood estimation method, the partial least square method, the quadratic programming method, and PSO (Particle Swarm Optimization). Is used so that the error is minimum (or maximum) under a predetermined evaluation function (within a predetermined constraint). A is an error parameter representing an uncertain element, and represents a factor (variation factor) in which the output y cannot be expressed by Θ and Z alone. The predetermined output y is predicted using Θ and Z. corresponding to the error between the predicted value y ₀ and actual y in the case where. That is, the error parameter A is also a parameter for matching the predicted value y ₀ in actual value y.

When predicting the output value y ₀ by multiple regression equation, function formula 3, for example, it can be used Equation 4-1, the j-th historical performance data (X _{j, y j)} error parameters in A _j is given by Equation 4-2 (Equation 4-2-1, Equation 4-2-2, etc.). When _obtaining each parameter element α _jNα of the error parameter A _j in the j-th past performance data (X _j , y _j ), for each of the first parameter element α _j1 to the _Nα parameter element α _jNα , By _{assuming that} only one parameter element α _j has a variation and setting all other (residual) parameter elements α _j to 0, the one parameter element α _j is obtained. For example, _assuming that there is a variation only in the first parameter element α _j1 and setting all other (residual) parameter elements α _j to 0, the first parameter element α ₁ can be obtained by Expression 4-2-1. . Further, for example, _{assuming that} only the second parameter element α _j2 has a variation, and all other (remaining) parameter elements α _j are set to 0, the second parameter element α ₂ is obtained by Expression 4-2-2. It is done. The same _{applies to} the third parameter element α _{j3 and thereafter} .

In addition, for example, when the output value y ₀ is predicted by a multiple regression equation, for example, Equation 5-1 can be used as the function equation 3, and the j-th past performance data (X _j , y _j ) The error parameter A _j is given by Equation 5-2-1 or Equation 5-2-2.

Further, for example, when the predicted value y ₀ is predicted by a multiple regression equation, for example, Equation 6-1 can be used as the function equation 3, and the j-th past performance data (X _j , y _j ) The error parameter α _j is given by Equation 6-2-1 or Equation 5-2-2.

In the above description, a mathematical expression representing the function f is used. However, an arithmetic program including a table representing the function f, a convergence calculation algorithm, an if-then rule, a fuzzy reasoning, a neural network, a JIT (Just in Time) model, and the like. May be used. Here, when the error parameter A _j cannot be obtained from Z _j , Θ, and y _j by back calculation, the value of the error parameter A _j is obtained by, for example, a binary search method, a carpet bombing method, or PSO (Particle Swarm Optimization). And an error parameter A _j is obtained so that the output value y substantially matches y _j .

  The parameter calculation unit 13 stores the calculated error parameters A in the intermediate data storage unit 42 of the storage unit 4.

The weighted Euclidean distance d _j , similarity w _j, and error parameter A _j calculated by each of the processes S11 to S13 are, for example, the intermediate data storage unit 42 in a table format (table format) as shown in FIG. Is remembered. The intermediate data table 52 shown in FIG. 5 includes an output field 521 for registering the actually measured output value y _j, and a similarity calculation data field 522 for registering data x of the data item used for calculating the similarity w _j. , the predicted value output calculation data field 523 for registering the data x of the data items used for calculation of y _0, the weighted distance field 524 for registering a weighted Euclidean distance d _j of the historical performance data X, the historical record The past record data (X, y) includes a similarity field 525 for registering the similarity degree w of the data X and an error parameter field 526 for registering the error parameter A _j of the past record data X. It includes a record for every further comprises a record of the predicted target data X _0. The similarity calculation data field 522 includes data item subfields 5221 to 522N corresponding to the respective data items used for the calculation of the similarity w _j . Similarly, the output calculation data field 523, a data item subfield 5231~523L corresponding to each data item that is used to calculate the predicted value _{y 0.} Further, the error parameter field 526 includes error parameter subfields 5261 to 526Nα corresponding to a plurality of parameter elements α _{j1 to} α _jNα in the error parameter A _j . The output field 521 and the similarity calculation field 522 in the intermediate data table 52 shown in FIG. 5 correspond to the output field 511 and the data field 512 in the actual measurement data table 51 shown in FIG.

Then, the output value prediction device S, by the predictive value calculating unit 14 of the arithmetic and control unit 1, using the prediction model obtained in process S13, the first to the data value of the L data item in the prediction target data X ₀ A predicted value y _0j is calculated for each error parameter A _j obtained in step S13 based on x _{01 to} x _0L, and each predicted value y _0j is calculated by weighted averaging the calculated predicted values y _0jk. And each of the calculated predicted values _y0j is stored in the predicted value storage unit 43 of the storage unit 4 (S14).

More specifically, first, in the error parameter A _j = [α _j1 , α _j2 ,..., Α _jNα ], an error in which other parameter elements α _j other than the kth parameter element α _jk are 0. parameter _{a j} vector _{a jk} and _a jk = the [0, ···, 0, α jk, 0, ···, 0] when the respect of these respective error Paramera _{a jk,} the k by equation 7 A predicted value y _0jk corresponding to the th parameter element α _jk is obtained. Next, as shown in Expression 8, these predicted values y _0jk (k = 1 to Nα) are weighted and averaged to obtain the predicted value y _0j .

Here, ν _k is a weight in the case of performing the weighted average, and is a real number of 0 or more, and the sum thereof is a predetermined constant value B, for example, B = 1 as shown in Equation 9. Fixed.

The weight ν _k is set in advance according to the degree of influence on the variation, for example. For example, as the degree of influence on the variation increases (the more equivalent to the dominant variation factor), the weight ν _k corresponding to the parameter element α _jk is also increased, and the error parameter A _jk = [α _j1 , α _j2, · · ·, among the parameter elements alpha _jk of _α jNα], weights [nu _k corresponding to the parameter elements alpha _jk that most affect the variation is largest. Further, for example, as shown in Expression 10, the weights ν _k are equal and the weights ν _k corresponding to the parameter elements α _jk may be the same. That is, the predicted value y _0j is obtained by simple averaging of the predicted values y _0jk (k = 1 to Nα).

In this process S14, the prediction model is changed to the error parameter A _j obtained in the process S13.

  In the processing S14, as in the processing S13, an arithmetic program including a table representing the function f, a convergence calculation algorithm, an if-then rule, a fuzzy inference, a neural network, a JIT model, and the like may be used.

Each such prediction value y ₀₁ ~y _0M calculated by the step S14 is stored in, for example, the predicted value storing unit 43 by the table format table format as shown in FIG. Predicted value data table 53 shown in FIG. 6, the data of the predicted value field 531, first to L data items used for calculation of the predicted value y ₀ to register the predicted value y ₀ calculated by the step S14 The output prediction data field 532 for registering the values x _{01 to} x _0L , the error parameter field 533 for registering the error parameter A calculated by the processing S13, and the past used for calculating the error parameter of the error parameter field 533 Each field of the similarity field 534 for registering the similarity w _j in the record data (X, y) is provided, and a record is provided for each error parameter A _j . The error parameter field 533 includes error parameter subfields 5331 to 533Nα corresponding to a plurality of parameter elements α _{j1 to} α _jNα in the error parameter A _j . Here, each prediction value y ₀₁ ~y _0M calculated by the step S14, as shown in FIG. 6, also the similarity w _j corresponding to the predicted value y _0, the predicted value storing section to correspond to each other 43.

Then, the output value prediction device S, the variation calculation unit 15 of the arithmetic and control unit 1, with each of the prediction value _y 01 _{~y 0M} obtained in processing S14, calculates the variance of the predicted value _{y 0j,} the calculated The variation of the predicted value _y0j is stored in the variation storage unit 44 of the storage unit 4 (S15).

In the present embodiment, based on the rule of thumb that results in similar in similar conditions, the higher the similarity between the data x ₀ of the prediction target data X ₀ and data x _j of the j-th historical performance data X ( If the similarity w _j is large), the error parameter A ₀ (A ₀ = [α ₀₁ , α ₀₂ ,..., Α _0Nα ]) of the prediction target data X ₀ is also considered to have high similarity. For this reason, the predicted value y ₀ of the prediction target data X ₀ is considered to be the predicted value y _0j predicted using the error parameter α _j with the similarity w _j .

Based on this concept, and more specifically, the variation calculation unit 15, as shown in FIG. 7 (A), taking the similarity w on the horizontal axis with taking the prediction value y ₀ on the vertical axis, First, M error parameters A _j (A _j = [α _j1 , α _j2 ,..., Α _jNα ]) (j = 1) calculated from the M past performance data (X, y), respectively. _˜M ), the similarity w _{j is made} to correspond to each of the M predicted values y _0j respectively corresponding to. Next, the variation calculation unit 15 _divides a range y _0j including at least each predicted value on the vertical axis y ₀ in FIG. _7A into a finite number of sections (classes, grades), and is included in each section. A weighted frequency F _w is generated by adding all the similarities w _j of the predicted value y _0j , and a histogram representing the variation of the predicted value y ₀ is generated as shown in FIG.

That is, as shown in Equation 11, if the set of j the predicted value _{y 0j} is included in the k-th interval _{(Y k} or _{Y k + 1} less than the section) and _{S k (S k = {j} | Y k ≦ y _0j <Y _{k + 1} }) and the sum of the similarities w _j for j included in the set S _k is the weighted frequency F _w in the k-th interval.

This variation of the predicted value y ₀ as is indicated by the histogram, it is possible to easily know the frequency of occurrence of the predicted value y _0.

As described above, the histogram illustrated in FIG. 7B may be the variation of the predicted value y ₀ , but in the present embodiment, the variation calculation unit 15 further has an area of the histogram illustrated in FIG. Normalize so that This normalized histogram is taken as the probability density of the predicted value y ₀ in the prediction target data X _0, and is regarded as the variation of the predicted value y ₀ . Furthermore, the variation calculation unit 15 may represent the histogram shown in FIG. 7B with a curve as shown in FIG. 7C while maintaining the area at 1. This curve is a probability density of the predicted value y ₀ in the prediction target data X _0, are variations of the predicted value y _0.

Note that the normalization is performed when, for example, the vertical axis y ₀ in FIG. 7A is divided into a finite number of sections and is divided into an equal width h = | Y _{k + 1} −Y _k |. , According to Equation 12.

  Further, when obtaining the curve shown in FIG. 7C from the histogram shown in FIG. 7B, a known probability distribution such as a lognormal distribution or a Weibull distribution may be used.

  FIG. 8 is a diagram for explaining a method for obtaining the probability density curve shown in FIG. 7C from the histogram shown in FIG. 7B. 8A shows a state in which the center points of the histogram are connected by a broken line, FIG. 8B shows the cumulative probability density of FIG. 8A, and FIG. 8C shows FIG. FIG. 8D shows a state where the cumulative probability density shown in B) is smoothed, and FIG. 8D shows the smoothed probability density (probability density curve).

First, as shown in FIG. 8 (A), in the histogram normalized shown in FIG. 7 (B), connecting the center positions of the respective frequencies of (y ₀ direction of the center) in a line. At both ends, a point separated from the end by a half (h / 2) of the section width h is also set to 0 and connected to the broken line. The area surrounded by the broken line is also 1.

Next, as shown in FIG. 8 (B), the cumulative probability density _SwN is obtained from FIG. 8 (A) by Expression 13-1.

Next, as shown in FIG. 8C, the cumulative probability density _SwN of the broken line in FIG. 8B is smoothed by using, for example, Equation 13-2.

  Then, as shown in FIG. 8D, the smoothed probability density (probability density curve) is obtained from the smoothed cumulative probability density shown in FIG. It is done.

Thus, the variation of the predicted value y ₀ is indicated by the probability density, and the appearance probability of the predicted value y ₀ can be easily known.

Further, when calculating the weighted frequency F _w , the similarity w of the M past performance data (X, y) is arranged in descending order (in descending order), and a predetermined value set in advance from the larger one is set. The past performance data (X, y) up to the number (predetermined ratio) may be extracted, and the weighted frequency F _w may be obtained by using only the extracted data. By removing the past performance data (X, y) having a low similarity w _j in advance, it is possible to reduce the amount of computation processing for computing the weighted frequency F _w (shortening the computation processing time). In the case where the similarity w _j is calculated by the above-described Expression 2 (Expression 2-1 to Expression 2-3), the past performance data (X, y) having a low similarity w _j with the prediction target data X ₀ is calculated. However, since the similarity w _j does not become 0, the weighted frequency F _w is affected. For this reason, the width of the weighted frequency F _w shown in FIG. 7B matches the width of the M predicted values y ₀ , and when the function f is Equation 4, the width is the prediction target data. always constant irrespective of the conditions of X _0. As a result, the base of probability density shown in FIG. 7C may spread more than necessary. However, as described above, by excluding past performance data (X, y) having a small similarity w _j , it is possible to prevent the base of probability density from being excessively widened, and the predicted value y ₀ in the prediction target data X ₀ . The feature of the distribution shape is remarkably expressed.

The output value prediction device S, the arithmetic control unit 1, the variation of the predicted value y ₀ calculated by the variation calculation unit 15 in the processing S15 is presented to the presentation unit 3 (S16), the processing is terminated. This variation of the predicted value y ₀ is presented to the presentation unit 3 as the user is able to know the variation of the predicted value y _0, the prediction in the case of performing the operation or judgment, etc. based on the predicted value y ₀ variation value y ₀ it becomes possible to consider.

Here, in the case of presenting the variation of the predicted value y ₀ in the presentation unit 3, the combined auxiliary display that represents the portion outside the preset predetermined control range in the predetermined output y to the presentation of the variation presented The output value prediction device S may be configured to do so. With such a configuration, when the user, which can know the relationship between the portion outside the predetermined control range with the variation, performing operations and judgments, etc. on the basis of the predicted value y _0, the predicted value y It is possible to determine conditions relating to the above factors (data items) such as operation conditions while taking into account the variation of ₀ . For this reason, it is possible to make more appropriate decision making, and it is possible to grasp information that is difficult to understand simply by looking at the shape of the variation by the auxiliary display.

  Here, the auxiliary display includes various expression methods, and includes, for example, a numerical value representing a portion outside the predetermined management range, a line segment (boundary line), a texture, and the like.

  More specifically, the auxiliary display can include the following displays.

FIG. 9 is a diagram for explaining a first mode of auxiliary display. FIG. 9A shows a case where the variation of the predicted value y ₀ is indicated by a histogram, and FIG. 9B shows a case where the variation of the predicted value y ₀ is indicated by the probability density. FIG. 10 is a diagram for explaining a second mode of auxiliary display. FIG. 10A shows a case where the variation of the predicted value y ₀ is indicated by a histogram, and FIG. 10B shows a case where the variation of the predicted value y ₀ is indicated by the probability density. FIG. 11 is a diagram for explaining a third mode of auxiliary display. FIG. 11A shows the case of the management value reference moment, FIG. 11B shows the case of the upper limit reference moment, and shows the case of the lower limit reference moment.

For example, the auxiliary display, as shown in FIG. 9 (A), when the variation of the predicted value y ₀ is presented by the graph a histogram, the graph in preset or predetermined upper limit value yth _max It may be a first area display relating to the area. For example, as shown in FIG. 9A, the first area display may be a texture display in which a hatched texture is applied to an area portion that deviates above the upper limit value yth _{max in the} graph. It may be a numerical display that numerically represents the area of the area that deviates above the upper limit value yth _max in the graph, and for example, deviates above the upper limit value yth _{max in the} graph with respect to the total area As in the graph. It may be a numerical display in which the area ratio (area outside the upper limit) of the area of the portion is expressed numerically. Such auxiliary display, in the example shown in FIG. 9 (A), but the variation of the predicted value y ₀ is a histogram, the variation of the predicted value y ₀ may be a probability density. In addition, for example, the auxiliary display, as shown in FIG. 9 (B), when the variation of the predicted value y ₀ is presented by the graph a probability density, definitive below a predetermined lower limit value yth _min a preset It may be a second area display relating to the area of the graph. For example, as shown in FIG. 9B, the second area display may be a texture display in which a texture is applied to an area portion that falls below the lower limit value yth _{min in the} graph. It may be a numerical display in which the area of the area portion that falls below the lower limit value yth _min in the graph is expressed numerically. Further, for example, numerical display may be used to express the area ratio (area outside the lower limit) of the area of the portion that falls below the lower limit value yth _{min in the} graph with respect to the total area As in the graph. In such an auxiliary display, in the example shown in FIG. 9B, the variation of the predicted value y ₀ is a probability density, but the variation of the predicted value y ₀ may be a histogram, and the first area display And the second area display may be presented together.

With this configuration, when the user wants to keep the output value within a certain management range, or when he / she wants to manage the rate out of the management value or the rate out of the management value, can know Do management target by the first area display and / or the second area display, also, if the predicted value y ₀ in any relation with the management target, for example, the management target prediction value y ₀ You can know how far you are off. For this reason, when performing an operation or determination based on the predicted value y ₀ , the user changes a condition related to the factor (data item) such as an operation input or an operation condition while referring to such an auxiliary display. By taking into account the variation of the predicted value y ₀ and taking into account the variation of the predicted value, the condition relating to the factor (data item) such as the operating condition is taken into account, for example, taking the risk out of the predetermined management range. It becomes possible to determine more appropriately.

When the variation is a histogram, the area ratio outside the upper limit is obtained by dividing the value _obtained by adding the similarity w _k to only the value equal to or greater than the upper limit value yth _max by the sum of the similarities w _k for all values. It is required by doing. That is, if a set of values equal to or greater than the upper limit value yth _max is U _j (U _j = {j | y _0j > yth _max }), the area ratio outside the upper limit is expressed by Expression 14. Similarly, the area ratio outside the lower limit is also obtained.

When the variation is probability density, the total area is 1. Therefore, the area outside the upper limit is obtained by obtaining the area outside the upper limit value yth _max , and the lower limit value yth _min is reduced. The area ratio outside the lower limit is obtained by obtaining the area deviating to the side.

Further, for example, contrary to the above first embodiment, the auxiliary display, in a case where the variation of the predicted value y ₀ as shown in FIG. 10 (A) is presented by the graph a histogram, in the graph of the variation It may be a boundary value corresponding to an area ratio as / As of a predetermined area as set in advance from one end of the variation graph with respect to the total area As. Wherein one end, graph axis is the point of intersection with the (predicted value y ₀ axis (the output y ₀ axis) in this case), the point of the upper end side (direction in which the predicted value y ₀ increases) (upper point) and bottom side ( there is a point of the predicted value y ₀ decreases direction) (lower point). Therefore, for example, as shown in FIG. 10 (A), this boundary value is a predetermined area set in advance from the upper end point y _0max in the variation graph with respect to the total area As in the variation graph (FIG. 10A). This is a value (= upper boundary value y _0up ) corresponding to the area ratio as / As (= R _up ) of as). That is, the boundary value (upper boundary value) y _0up is a predicted value y ₀ that is an area ratio as / As (= R _up ). Further, for example, the boundary value corresponds to an area ratio as / As (= R _dn ) of a predetermined area as set in advance from the lower end point y _{0 min} in the variation graph with respect to the total area As in the variation graph. (= Lower boundary value y _0low ) (not shown). Such auxiliary display, in the example shown in FIG. 10 (A), but the variation of the predicted value y ₀ is a histogram, the variation of the predicted value y ₀ may be a probability density. This auxiliary display is shown in FIG. _{10B in} the case where the boundary value is the lower boundary value y _0low .

  Such a boundary value is obtained by using, for example, a binary search method while changing the value of the boundary value because the relationship between the boundary value and the area ratio as / As is monotonically increasing.

With this configuration, the user can know a predetermined management index such as a rate (allowable deviation rate) that can be allowed to deviate upward or downward from the boundary value, and the boundary value and the management range. by comparing the bets, whether the predicted value y ₀ in any relation with the management indicator, for example, can be predicted value y ₀ know disengaging degree from the management indicator, appropriateness of such operating conditions You can know the sex. For this reason, when performing an operation or determination based on the predicted value y ₀ , the user changes a condition related to the factor (data item) such as an operation input or an operation condition while referring to such an auxiliary display. Then, by comparing the boundary value with the predicted value y ₀ , the factor (data item) such as the operating condition is set so that the allowable deviation rate becomes an appropriate value while considering the risk of the predicted value y ₀ being deviated. It becomes possible to determine the conditions related to.

In the example shown in FIG. 10, the boundary value is graphically displayed. However, the boundary value may be numerically displayed as a numerical value thereof, or a difference from a preset management value may be numerically displayed. Good. That is, for example, in the case where the variation of the predicted value y ₀ is a histogram or a probability density and is presented by a graph, the auxiliary display is displayed from one end of the variation graph with respect to the total area As in the variation graph. It may be a difference between a boundary value corresponding to an area ratio as / As of a predetermined area as set in advance and a management value set in advance. Also with this configuration, the user can know the deviation of the allowable deviation rate from the difference between the boundary value and the management value, and when performing an operation or determination based on the predicted value y ₀ , the prediction It is possible to determine the conditions related to the factors (data items) such as operation conditions so that the allowable deviation rate becomes an appropriate value while considering the risk of deviating values. In particular, in such a configuration, a user such as an operator only has to change (adjust), for example, an operation input or an operation condition in advance so that the difference becomes zero. Easy to understand. Furthermore, by adjusting using an optimization method such as PSO (Particle Swarm Optimization) so that the difference becomes 0, it is possible to reduce the burden of trial and error of the user.

Further, for example, the auxiliary display, in a case where the variation of the predicted value y ₀ is presented by the graph a histogram or probability density, to the total area As in the graph of this variation, preset predetermined upper limit value or more The ratio of the first area ratio au1 / As, which is the ratio of the first area au1 of the graph, to the total area As of the variation graph, and the second area au2 of the graph below a predetermined lower limit set in advance. sum au1 / as + au2 / as or square sum ^{(au1 / as) 2 + (} au2 / as) may be ^two of the second area ratio au2 / as is. The sum or the square sum includes a weighted sum or a weighted square sum.

Even with such a configuration, the user can set the degree of deviation of the portion outside the predetermined management range as the sum au1 / As + au2 / As of the first area ratio au1 / As and the second area ratio au2 / As or the sum of squares ( ^{au1 / As) 2 + (au2} / As) can be known by ^2, if on the basis of the predicted value _{y 0} performing operations and judgments, etc., the off-rate, taking into account the risk of deviating the variation of the predicted value _{y 0} For example, it is possible to determine conditions relating to the above factors (data items) such as operation conditions so that the value becomes an appropriate value.

In particular, in such a configuration, when the predicted value exceeds the upper limit value and / or exceeds the lower limit value from a predetermined management range set in advance, or the upper limit for the preset management value is set. The sum au1 / As + au2 / As or the sum of squares (au1 / As) ² + (au2 / As) when adjusting the degree of detachment (degree of detachment) exceeding the value and the degree of detachment exceeding the lower limit value ) Conditions relating to the above factors (data items) such as operation conditions may be determined so that ² is minimized. In other words, evaluate the weighted sum of the first area ratio au1 / As and the second area ratio au2 / As of formula 15-1 and Equation 15-2 function _J A, as _{J B,} the evaluation function _{J A} , J _B may be determined such that conditions relating to the factors (data items) such as operation conditions are determined.

Here, w _U and w _L are weight coefficients set in advance, and are real numbers of 0 or more. For example, in the case of smaller second area ratio au2 / As, the value of _{w L} is larger than the value of _{w U.}

In particular, in the configuration, the ratio of the first area ratio au1 / As to the sum of the first area ratio au1 / As and the second area ratio au2 / As is a predetermined ratio R _UL (0 ≦ R _UL ≦ 1). In the case where it is desired to set the ratio, the ratio of the first area ratio au1 / As to the sum of the first area ratio au1 / As and the second area ratio au2 / As, expressed by Expression 16, and the predetermined ratio R _UL as an evaluation function J _C the absolute value of the difference, as the evaluation function J _C is minimized, for example, may be determined conditions related to the factors such as operating conditions (data item).

Furthermore, when there are a plurality of conditions related to the factors (data items) that minimize these evaluation functions J _A , J _B , and J _C , costs associated with changes in other conditions, for example, the operation input and operation conditions And the one with the smallest change amount may be selected. There are various methods for minimizing these evaluation functions J _A , J _B , and J _C , and examples include an optimization method such as PSO (Particle Swarm Optimization).

Further, for example, the auxiliary display, for example, as shown in FIG. 11 (A), the coordinate system taking the similarity w to the other shaft with taking the prediction value y ₀ variations whereas the axes of the predicted value y ₀ When _{plotting a} plurality of predicted values y _0j (y _j ) and a plurality of similarities w _j , the preset management represented by Expression 17-1 for the plurality of predicted values y _{0j Sum} of values _obtained by multiplying the absolute value of the difference between the value y _aim and the predicted value y _{0j by} the similarity w _j of the predicted value y _0j (management value reference moment) μ _1abs , or expressed by Expression 17-2 The sum (management value reference moment) μ ₂ of the difference between the management value y _aim and the prediction value y _0j multiplied by the similarity w _j of the prediction value y _0j may be used. Incidentally, instead of equation 17-1 is represented by the formula 17-3, the preset control value _{y aim} and value obtained by multiplying the similarity _{w j} of the predicted value _{y 0j} the difference between the predicted value _{y 0j} The sum (control value reference moment) μ ₁ may be used.

With this configuration as well, the user can know the degree of deviation of the management value from the management value reference moment μ, and when performing an operation or determination based on the predicted value y ₀ , the prediction it is possible to determine the condition relating to the factors such as for example operational conditions such variation is more appropriate in the surrounding of the control value of the predetermined output while considering the variation of the values y ₀ (data item) .

Further, for example, the auxiliary display, for example, as shown in FIG. 11 (B), the coordinate system taking the similarity w to the other shaft with taking the prediction value y ₀ variations whereas the axes of the predicted value y ₀ When a plurality of predicted values y _0j (y _j ) and a plurality of similarities w _j are plotted, the upper limit value equal to or greater than a preset upper limit value yth _max among the plurality of predicted values y _0j For the predicted value, the sum of the values obtained by multiplying the absolute value of the difference between the upper limit value yth _max and the predicted value equal to or higher than the upper limit value by the similarity w _j of the predicted value equal to or higher than the upper limit value represented by Expression 18-1 ( Upper limit value reference moment) μ _U1abs or the difference between the upper limit value yth _max and the upper limit value and the predicted value expressed by the equation 18-2 is set to the similarity w _j between the upper limit value and the predicted value. the sum of the multiplied values (upper limit reference moment) mu _U It may be.

Here, if the set of index values exceeding the upper limit value _{yth max} and _{U j, U j} = | a _{{j y j> yth max}} .

Even with this configuration, the user can know the risk and the degree of deviation from the upper limit value by the upper limit value reference moment μ, and when performing an operation or a determination based on the predicted value y ₀ , it is possible to determine the condition relating to the factors such as for example operational conditions such risks outside the predetermined output while considering the variation of the predicted value y ₀ from the upper limit value is made more appropriate (data item) .

Further, for example, the auxiliary display, for example, as shown in FIG. 11 (C), the coordinate system taking the similarity w to the other shaft with taking the prediction value y ₀ variations whereas the axes of the predicted value y ₀ When a plurality of predicted values y _0j (y _j ) and a plurality of similarities w _j are plotted, predictions of a lower limit value less than or equal to a preset lower limit value yth _min of the plurality of predicted values y _0j The sum of the values obtained by multiplying the absolute value of the difference between the lower limit value yth _min and the lower limit value or less predicted value by the similarity w _j of the lower limit value or less predicted value represented by the equation 19-1 (lower limit) Value reference moment) μ _L1abs , or the square of the difference between the lower limit y _0low and the predicted value below the lower limit represented by Expression 19-1 is multiplied by the similarity w _j of the predicted value below the lower limit the sum of the value (lower limit reference moment) mu _L2 der It may be.

Here, if the set of index values below the lower limit _{yth min} and _{L j, L j} = | a _{{j y j <yth min}} .

Even with this configuration, the user can know the risk and the degree of deviation from the lower limit value by the lower limit reference moment μ, and when performing an operation or a determination based on the predicted value y ₀ , it is possible to determine the condition relating to the factors such as for example operational conditions such risks outside the predetermined output from the lower limit value taking into account the variation of the predicted value y ₀ is made more appropriate (data item) .

Multiple addition, for example, the auxiliary display, the predicted value y ₀ of the variation Meanwhile coordinate axes of the plurality of coordinate system taking the similarity w to the other shaft with taking the prediction value y ₀ predicted value y _0j and _{(y j)} When the similarity w _j is plotted, it may be the sum of the upper limit value reference moment μ and the lower limit value reference moment μ represented by the expressions 20-1 and 20-2. . The sum includes a weighted sum.

With such a configuration, the user can know the disengaged condition and the risk of outside the upper and lower limit values by the sum of said lower limit reference moment mu and the upper limit reference moment mu, based on the predicted value y ₀ Te when performing operations and judgments, etc., conditions relating to the factors such as for example operational conditions such risks deviating from the upper limit value of the predetermined output while considering the variation of the predicted value y ₀ is more appropriate It becomes possible to decide.

Here, w _U and w _L are weight coefficients set in advance, and are real numbers of 0 or more. For example, if the degree deviates below the lower limit _{yth min} than the degree deviating beyond the upper limit value _{yth max} is small, the value of _{w L} is larger than the value of _{w U.}

As the output value prediction device S operates in this manner, the error parameter A is assumed to be composed of a plurality of parameter elements, and M error parameters A _j (calculated from M past performance data (X, y)). _{a j = [α j1, α} j2, ···, α jNα]) (j = 1~M) by using the predicted value _{y 0} of the prediction target data _{X 0} is calculated as M, and candidate prediction A weighted frequency F _w for the predicted value y ₀ is calculated according to the similarity w _j with the data X ₀ . Furthermore, the probability density is calculated from the weighted frequency F _w. Therefore, regardless of the manner of distribution of the variations of the predicted value y _0, historical performance data (X, y) and the predicted value y ₀ in the prediction target data X ₀ of similarity is considered the predicted target data X ₀ Variation is appropriately determined with high accuracy. Accordingly, the output value prediction unit S may present variations of the predicted value y _0, and thus also the variation of the predicted value y ₀ if based on the predicted value y ₀ performing operations and judgments, etc. can be considered It becomes.

Next, another embodiment will be described.
(Second Embodiment)
For example, in the case of products manufactured through various manufacturing processes in a relatively large-scale manufacturing plant, such as the manufacture of steel products and chemical products, for example, according to the input amount, operation input amount, time passage, etc. The output value in the manufacturing process and the output value in the final process directly connected to the product often change every moment. For example, in the manufacturing process of steel products, the relationship between the hot metal temperature in the topped car and the elapsed time, the relationship between the molten steel temperature in the ladle and the elapsed time, and the carbon concentration in the molten steel and the integrated oxygen injection value in the converter blowing. The relationship and the relationship between the molten steel temperature and the blown oxygen integrated value in converter blowing are mentioned.

FIG. 12 is a diagram illustrating the relationship between the temperature drop amount of the object and the elapsed time. FIG. 12 shows the relationship between the temperature drop amount (deviation from the initial temperature) y of the object left in the atmosphere and the elapsed time (temperature measurement time) t plotted with ○ for each past performance data. Assume that it was a result. Here, when predicting the temperature drop amount y (t ₀ ) at the predetermined time t ₀ , the following problem may arise when the probability density is obtained by using past performance data near the time t ₀ . That is, firstly, in the time region where the past performance data is small (or does not exist), there is very little data that can be utilized, and the data that is utilized is only a part of the past performance data. For this reason, it is difficult to predict the distribution of the temperature drop amount y (t ₀ ) of the prediction target data with high accuracy. Secondly, past performance data having a high degree of similarity with the prediction target data is not necessarily near the predetermined time t ₀ , and a past performance having a high degree of similarity with the prediction target data at a location away from the time t _0. If there is data, the past performance data is not used.

Therefore, for such a problem, as shown for a part of past performance data by a thin broken line in FIG. 12, a prediction model representing the relationship between the temperature drop amount y and the elapsed temperature t in each past performance data is constructed, By past each past performance data at a predetermined time t ₀ (that is, by determining the amount of temperature drop y (t ₀ ) at the time t ₀ of the constructed prediction model), the past away from the predetermined time t ₀ actual data can also be utilized for the estimation of the probability density in the prediction value y (t _0), it is possible to determine the variation of the predicted value y of the predicted target data (t ₀₎ with higher accuracy.

  In the second embodiment, the output value prediction apparatus S of the first embodiment is applied to the above-described case. Therefore, in the output value prediction apparatus S of the second embodiment, the output value prediction apparatus S of the first embodiment includes parameter calculation processing (S13) for calculating parameters and prediction value calculation processing (S14) for calculating prediction values. Since it is the same as that of the output value prediction apparatus S in 1st Embodiment except the point which performs a process as follows, description of the same point is abbreviate | omitted.

FIG. 13 is a diagram illustrating data stored in the intermediate data storage unit in the second embodiment. FIG. 14 is a diagram illustrating a model of past performance data in the second embodiment. FIG. 15 is a diagram illustrating data stored in a predicted value storage unit according to the second embodiment. FIG. 15 (A) shows the data at an arbitrary time t, FIG. 15 (B) shows the data at a predetermined time _{t 0.} FIG. 16 is a diagram illustrating a prediction model in the second embodiment. FIG. 17 is a diagram for explaining a procedure for calculating a variation in predicted values in the second embodiment. FIG. 17A shows the predicted value y (t ₀ ) at a predetermined time t ₀ , and FIG. 17B shows the relationship between the similarity w and the output predicted value y (t ₀ ). The horizontal axis is the similarity w, and the vertical axis is the predicted value y (t ₀ ).

  In the output value prediction apparatus S of the second embodiment, the past record data and the prediction target data are stored in advance in the table format (table format) in the actual measurement data storage unit 41 of the storage unit 4 as in the first embodiment. ing. In the second embodiment, the past performance data and the prediction target data include the temperature drop amount y, the actual measurement time t at which the temperature drop amount y is measured, and the factor data x related to the temperature drop amount y. Is done. The temperature drop amount y corresponds to the output value y, and the actual measurement time t can be regarded as one of factors in the factor related to the output value y. That is, the factor X related to the output value y includes at least the time t as an element. The origin of the actual measurement time t is the time when the temperature drop amount y = 0, that is, the time of the initial temperature of the object (the time when the measurement of the temperature of the object is started). Note that the output value prediction apparatus S may be configured such that the time t is not included in the factor X but is included in the factor Z.

The historical performance data ((X, t), y) the predicted target data _(X 0, _{t 0)} the output value from the (predicted value) prediction of _{y 0} is started on the basis of, in the processing S11, first Similar to the embodiment, the distance calculation unit 11 calculates the distance d _j between the past performance data (X, t) and the prediction target data (X ₀ , t ₀ ) in the first to Nth data item spaces. The calculated distance _dj is stored in the intermediate data storage unit 42 of the storage unit 4.

Next, the process S12, as in the first embodiment, the similarity calculating unit 12, the similarity w _j between the predicted target data X ₀ and historical performance data X _j, first through M past achievements Each of the data (X, y) is calculated, and each calculated similarity w _j is stored in the intermediate data storage unit 42 of the storage unit 4.

Next, in the process S13, by using the formula 21 instead of the formula 3 and otherwise performing the same process as in the first embodiment, the parameter calculation unit 13 sets the error parameter A _j representing the uncertain element as Each of the first to M-th past performance data ((X, t), y) is calculated, and each calculated error parameter A _j is stored in the intermediate data storage unit 42 of the storage unit 4. Here, y _j (t) is a temperature drop amount at the actual measurement time t, and Z _j , Θ, and A _j are the same as those in Expression 3 of the first embodiment. t is the elapsed time from the initial time (the time of the initial temperature of the object). In this process, as in the first embodiment, in the error parameters A _j = [α _j1 , α _j2 ,..., Α _jNα ], the other parameter elements α _j excluding the kth parameter element α _jk are When the vector A _jk of the error parameter A _j that is 0 is A _jk = [0,..., 0, α _jk , 0,..., 0], these error parameters A _jk are _expressed by Equation 21. By substituting, Equation 22 is obtained. By substituting the first to M-th past performance data ((X, t), y) into this equation 22, the k-th parameter element α _jk is obtained. Here, Θ is a known model parameter, and since y _j (t _j ), Z _j , and t _j are also known from each past performance data ((X, t), y), the k-th parameter Element α _jk is determined.

FIG. 13 corresponds to FIG. 5, and FIG. 13 shows data stored in the intermediate data storage unit 42, and each parameter element α _jk thus obtained is shown. In FIG. 14, Expression 22 is indicated by a thick broken line, and in FIG. 14, the model of each past performance data ((X, t), y) is indicated by a thin broken line. In FIG. 12, an example of Expression 22 is indicated by a thick broken line.

Next, the process S14, the prediction time t ₀ is used, the output value prediction device S, by the predictive value calculating unit 14 of the arithmetic and control unit 1, using the prediction model (Equation 22) which has been determined by the processing S13, As shown in FIG. 13, the prediction time t ₀ in the prediction target data (X ₀ , t ₀ ) and the data values z _{01 to} z _0L of the first to Lth data items are stored in the intermediate data storage unit 42. Based on the prediction value y _0jk (t ₀ ) (k = 1 to Nα) based on each error parameter A _j obtained in step S13 (Equation 23), the calculated prediction value y _0jk is weighted. Each predicted value _y0j is calculated by _averaging (Equation 24).

Here, ν _k is a weight in the case where the weighted average is performed as in Equation 9, and the sum thereof is fixed to a predetermined constant value B, for example, B = 1 as shown in Equation 25. The The weight ν _k is set in advance in accordance with, for example, the degree of influence on the variation, as in the case of Expression 9. Here, the weight ν _k is equal, for example, as shown in Expression 26. The weights ν _k corresponding to the parameter elements α _jk are the same.

Then, each predicted value _y 01 calculated _{(t 0) ~y 0M (t} 0) , as shown in FIG. 15 (B), and the error parameter _A 1 to A _M and the similarity _w 1 to w _M The associated values are stored in the predicted value storage unit 43 of the storage unit 4. FIG. 15A shows a prediction model at an arbitrary time t in a table format, and FIG. 16 shows a graph of the prediction model. This prediction model is obtained by substituting the time t and the data values z _{01 to} z _0L of the first to Lth data items into Expression 22 for each of the error parameters A _j .

Next, the process S15, as in the first embodiment, the variation calculation unit 15 uses the respective predicted value _y 01 _{~y 0M} obtained in the processing S14, the variation of the predicted value _{y 0 (t} 0) And the variation of the calculated predicted value y ₀ (t ₀ ) is stored in the variation storage unit 44 of the storage unit 4. More specifically, as shown in FIG. 17, the variation calculation unit 15 takes the predicted value y (t ₀ ) on the vertical axis and the similarity w on the horizontal axis. For each of M predicted values y _0j (t ₀ ) respectively corresponding to the M error parameters A _j (j = 1 to M) calculated from the data ((t, x), y), The similarity w _{j is made} to correspond. Next, the variation calculation unit 15 _divides a range including at least each predicted value y _0j (t ₀ ) on the vertical axis y (t ₀ ) in FIG. 17B into a finite number of sections (classes, grades). Then, the weighted frequency F _w is generated by adding all the similarities w _j of the predicted values y _0j (t ₀ ) included in each section, and a histogram representing the variation of the predicted values y ₀ (t ₀ ) is generated. To do. The histogram may be a variation of the predicted value y ₀ (t ₀ ), but in the present embodiment, the variation calculation unit 15 further normalizes the area of the histogram to be 1. This normalized histogram is the probability density of the predicted value y ₀ (t ₀ ) in the prediction target data (t ₀ , x ₀ ), and is the variation of the predicted value y ₀ (t ₀ ). Or the variation calculation part 15 calculates | requires the said curve similarly to the above-mentioned from this histogram, further maintaining the area at 1. This curve is the probability density of the predicted value y ₀ (t ₀ ) in the prediction target data (t ₀ , x ₀ ), and is the variation of the predicted value y ₀ (t ₀ ).

By such an operation, the output value predicting device S of the second embodiment, regardless of the manner of distribution of the variations of the predicted value y _0, the predicted value of the output in the process output changes every moment with the lapse of time y ₀ can be determined more appropriately, and the variation of the predicted value y ₀ can be determined more appropriately. Further, in the output value prediction device S of the second embodiment, it is possible to obtain the predicted value y ₀ even in a time region where the past performance data is small (or does not exist), and it is also possible to obtain the variation of the predicted value y _0. It becomes. Further, in the output value prediction apparatus S of the second embodiment, past performance data separated from a predetermined time t ₀ can also be used for estimation of variation in the predicted value y ₀ (t ₀ ), and prediction of prediction target data is performed. Variations in the value y ₀ (t ₀ ) can be obtained with higher accuracy. Moreover, in the output value prediction device S of the second embodiment, as can be seen from FIG. 15A and FIG. 17, the predicted values y ₀ (t) at a plurality of different times t can be obtained. The variation of ₀ (t) can also be obtained. Therefore, it is possible to determine the processing end timing with the least risk by comparing the variation of the predicted value y ₀ (t) at each time t. For example, in a heating furnace of a steel product manufacturing process, it is possible to terminate the heat treatment at a low risk timing, considering not only whether the steel was heated as intended but also considering the probability of temperature deviation. It becomes. Further, for example, in converter blowing, it is possible to terminate the blowing at a timing with less risk in consideration of the molten steel temperature and the probability that the components in the molten steel deviate from the target. In the case of this converter blowing, the horizontal axis in FIG. 17 is the integrated value of the blowing oxygen amount.

Next, another embodiment will be described.
(Third embodiment)
In the steel product manufacturing process, after the converter is blown, the molten steel is transferred from the converter to the ladle, and after the molten steel is processed, the molten steel is transported to the continuous casting equipment. Therefore, it is preferable that the molten steel temperature is slightly higher than the solidification temperature when the ladle arrives at the continuous casting facility. If the molten steel temperature is too low, the molten steel solidifies, which is not preferable. If the molten steel temperature remains high, the casting speed must be reduced. Depending on the charge, the molten steel composition, the amount of molten steel, the type of ladle, the initial state of the ladle (the refractory melting condition, the temperature distribution inside the ladle (cooling condition)), and the ladle when receiving steel from the converter The amount of temperature drop varies depending on the amount of alloy and type of alloy placed in advance. For this reason, it is difficult to predict the molten steel temperature that changes from moment to moment at a single point. Therefore, it is important to accurately estimate the variation in molten steel temperature in the ladle for the charge.

  In the third embodiment, the predetermined output is the molten steel temperature in the ladle in the process from the converter steelmaking process through the molten steel treatment process to the continuous casting process, and the output value prediction device S of the first embodiment is applied. The output value predicting apparatus S in the third embodiment estimates the probability distribution of the temperature drop of the molten steel until the molten steel transferred from the converter to the ladle is transported to the molten steel processing facility. It is. Therefore, the output value prediction apparatus S in the third embodiment is the same as the output value prediction apparatus S in the first embodiment, in the distance calculation process (S11) for calculating the distance, the parameter calculation process (S13) for calculating the parameter, and the predicted value. The predicted value calculation process (S14) for calculating is the same as the output value prediction apparatus S in the first embodiment except that the process is executed as follows, and the description of the same points is omitted.

18 and 19 are diagrams illustrating the relationship between the weighted distance and the similarity in the third embodiment. The horizontal axis of FIG. 18 and FIG. 19 is the weighted distance d _j , and the vertical axis is the similarity w _j . FIG. 20 is a diagram illustrating a model of past performance data in the third embodiment. FIG. 21 is a diagram showing the probability density at each predicted value in the third embodiment. 20 and FIG. 21, the horizontal axis represents the elapsed time t expressed in units of minutes (min), and the vertical axis thereof represents the temperature drop amount y (t) expressed in degrees (° C.).

In the output value prediction apparatus S of the third embodiment, the past record data and the prediction target data are stored in advance in a table format (table format) in the actual measurement data storage unit 41 of the storage unit 4 as in the first embodiment. ing. In the third embodiment, the past performance data and the prediction target data include the temperature drop amount y, the actual measurement time t when the temperature drop amount y is measured, and the factor data x related to the temperature drop amount y. Is done. The temperature drop amount y corresponds to the output value y, and the actual measurement time t can be regarded as one of the factors in the factor related to the output value y. The origin of the actual measurement time t is the time when the temperature drop amount y = 0, that is, the time of the initial temperature of the object (the time when the measurement of the temperature of the object is started). The data items used for calculating the distance d _j and the similarity w _j are the number of times the ladle is received steel, the type of deoxidizer, the concentration of molten steel, the state of the ladle in the ladle, the temperature at which the steel is discharged, and the solidification temperature. And each item of the operation team. Here, in the present embodiment, the number of times the steel is received by the ladle is converted by a nonlinear function so as to be, for example, the square root of the number of times the steel is received. The type of deoxidizer is quantified according to the strength of deoxidation. When the ladle is in an empty pan state (the state in which no molten steel is contained), the standing time, the heat retaining time, the standing time after the heat retaining, and the like are quantified by a nonlinear function. The operation group is given an identifier for each group.

Then, when the prediction of the output value (predicted value) y ₀ is started from the prediction target data (X ₀ , t ₀ ) based on the past performance data ((X, t), y), The distance calculation unit 11 uses the distance d _j defined in the above to perform the same processing as in the first embodiment, so that the distance calculation unit 11 predicts the past performance data (X, t) in the first to Nth data item spaces. A distance d _j between the target data (X ₀ , t ₀ ) is calculated, and the calculated distance d _j is stored in the intermediate data storage unit 42 of the storage unit 4.

Here, f _d (X _ji , x _0i ) is a function that takes 0 when x _ji and x _0i are the same, and takes 1 when x _ji and x _0i are different. In this embodiment, a _i (i = 1 to N) = 1. N is the number of data items. Further, k <N.

When comparing the operating conditions of the relevant charge with the operating conditions of each past charge, it may be subtracted, such as the operating team or equipment number (number of equipment used for processing among multiple equipment). can not data items or, some data items insignificant to the subtraction itself, a distance d _j defined by equation 27, valid for data items such data item is significant whether the same It is.

Further, date and time or date may be added as a data item when calculating the similarity w _j . Some processes have characteristics that change over time, such as aging and seasonal variation. In such a case, even if the operating conditions are the same, the result may be different if the date is different. By adding the date as a data item, the similarity W _j of the old data can be reduced, and the prediction can be made in consideration of the secular change. Note that the date may be expressed as the number of days elapsed from a reference date (for example, January 1, 1900).

Next, the process S12, as in the first embodiment, the similarity calculating unit 12, the similarity w _j between the predicted target data X ₀ and historical performance data X _j, first through M past achievements Each of the data (X, y) is calculated, and each calculated similarity w _j is stored in the intermediate data storage unit 42 of the storage unit 4.

Here, as the similarity _{w j,} instead of Formula 2-1 Formula 2-3, where the similarity _{w j} defined by Equation 28 is used.

Here, μ is an average value of the distance d _j (j = 1 to M), and σ is a standard deviation of the distance d _j (j = 1 to M). In the present embodiment, g = 1 is set. FIG. 18 shows the relationship between the weighted distance d _j according to the equation 27 and the similarity w _j according to the equation 28.

In the present embodiment, as shown in FIG. 19, the similarity w _j smaller than a predetermined threshold set in advance is set to zero. By previously removing past performance data ((X, t), y) having a low degree of similarity w, for example, the following calculation processing amount such as calculation processing amount for calculating the weighted frequency F _w is reduced (calculation processing) Time reduction).

Next, in the process S13, the parameter calculation unit 13 uses the expression 29-1 and the expression 29-2 instead of the expression 3 and the other processes in the same manner as in the first embodiment, so that the parameter calculating unit 13 determines the uncertain element. An error parameter A _j to be expressed is calculated for each of the first to M-th past performance data ((X, t), y) (= ((t, x), y)), and each of the calculated error parameters A _j Is stored in the intermediate data storage unit 42 of the storage unit 4.

Here, y _j (t) is the amount of temperature drop at the actual measurement time t, T _0j is the converter steel temperature that is the molten steel temperature at the time when the molten steel is transferred from the converter to the ladle, and T _∽j is the solidification temperature of the molten steel, q _j is a value obtained by converting the amount of heat to a temperature previously entered by that alloy (put every alloy) rob the ladle, when placed every alloy no q _j = 0. Here, in this embodiment, an equation estimated by multiple regression models from past performance data ((t, x), y) is used for q _j , and Equation 30 is used for the heat transfer coefficient. α _j1 is a variation in heat transfer coefficient, and α _j2 is a variation in the initial temperature of the molten steel. In this case, Nα = 2.

Here, p _ji is a value set according to the number of received steels and the molten steel carbon concentration, λ ₀ is a constant term, N _λ is the number of coefficients, and is 2 in this embodiment. is there.

In the present embodiment, the data items for predicting the output include the converter steel temperature T _0j , the solidification temperature T _{∽j of the} molten steel, and the _{input of the detained} alloy necessary for estimating the temperature drop q _j due to the _detained alloy. Amount.

When the predicted value y ₀ is predicted by a multiple regression equation, the parameter element of the error parameter A _j in the j-th past performance data ((X _j , t _j ), y _j ) based on the function equation 29 α _j1 is given by Equation 31. Here, α _j2 = 0.

Note that the predicted value y _0j (t) corresponding to the error parameter α _j1 obtained from the j-th past performance data is given by Expression 32-1 and Expression 32-2.

Here, T ₀₀ is the converter steel temperature in the prediction target charge, T _∽0 is the solidification temperature of the molten steel in the prediction target charge, and q ₀ is the _detained alloy in the prediction target charge. It is a value obtained by converting the amount of heat taken into temperature.

Next, with α _j1 = 0, Equation 33 is used instead of Equation 31. That is, the element α _j2 of A _j in the j-th data is given by Expression 33. Also, the predicted value y _0j (t) corresponding to α _j2 is given by Equation 34-1 and Equation 34-2.

FIG. 20 shows a model of each past performance data. FIG. 20 (A) shows the case of y _0j1 (t), and FIG. 20 (A) shows the case of y _0j2 (t). ing.

Then, the processing S14, the prediction time t ₀ is used, the output value prediction device S, by the predictive value calculating unit 14 of the arithmetic and control unit 1, using the prediction model which has been determined by the processing S13, the prediction target data ( Based on the predicted time t _{0 at} X ₀ , t ₀ ) and the data values z _{01 to} z _0L of the first to Lth data items, the predicted value y _0jk (t ₀ ) (k = 1 to Nα) is processed in step S13. Each prediction parameter y _0j is calculated by calculating each of the error parameters A _j obtained in step (1) and calculating the weighted average of the calculated prediction values y _0jk . Here, y _0j (t) = 0.7 × y _0j1 (t) + 0.3 × y _0j2 (t) (FIG. 21). Then, each prediction value and the calculated _{y 01 (t 0) ~y 0M} (t 0) , the error parameter _A 1 to A _M and the similarity _w 1 to w _M and association with the predicted value storing memory 4 Stored in the unit 43.

Next, the process S15, as in the first embodiment, the variation calculation unit 15 uses the respective predicted value _y 01 _{~y 0M} obtained in the processing S14, the variation of the predicted value _{y 0 (t} 0) And the variation of the calculated predicted value y ₀ (t ₀ ) is stored in the variation storage unit 44 of the storage unit 4. More specifically, the variation calculation unit 15 takes the predicted value y (t ₀ ) on the vertical axis and the similarity w _j on the horizontal axis, and first, M pieces of past performance data ((t, x ), for each error parameter of M calculated respectively from _{y) α j (j = 1~M} ) to the corresponding M number of each prediction value _y 0j _{(t 0),} the corresponding degree of similarity _{w j} Let Next, the variation calculation unit 15 _divides a range including at least each predicted value y _0j (t ₀ ) of the vertical axis y (t ₀ ) into a finite number of sections (classes, grades), and A weighted frequency F _w is generated by adding all the similarities w _j of the included predicted values y _0j (t ₀ ), and a histogram representing the variation of the predicted values y ₀ (t ₀ ) is generated. The histogram may be a variation of the predicted value y ₀ (t ₀ ), but in the present embodiment, the variation calculation unit 15 further normalizes the area of the histogram to be 1. This normalized histogram is the probability density of the predicted value y ₀ (t ₀ ) in the prediction target data (t ₀ , x ₀ ), and is the variation of the predicted value y ₀ (t ₀ ). Or the variation calculation part 15 calculates | requires the said curve similarly to the above-mentioned from this histogram, further maintaining the area at 1. This curve is the probability density of the predicted value y ₀ (t ₀ ) in the prediction target data (t ₀ , x ₀ ), and is the variation of the predicted value y ₀ (t ₀ ). FIG. 21 shows the probability density of the predicted value y ₀ (t ₀ ) of the temperature drop amount every 10 minutes. In other words, the predicted time point is every 10 minutes. In FIG. 21, ◯ indicates the actually measured temperature, ◯ at time 0 is the temperature at which the steel is discharged from the converter, and ◯ after about 50 minutes is the start of molten steel processing in the prediction target data (t ₀ , x ₀ ). This is the previous measured temperature. The measured temperature indicated by ○ after about 50 minutes is used to calculate the predicted value y ₀ (t ₀ ) of the temperature drop amount or the probability density of the predicted value y ₀ (t ₀ ) of the temperature drop amount shown in FIG. Is not used, but is displayed in FIG. 21 for reference. Further, the broken line connecting both circles in FIG. 21 is a calculation of the molten steel temperature of the charge to be predicted by Equation 29, which is obtained by fitting the error parameter α ₀ of Equation 31. Similarly, the temperature drop amount the calculation of the probability density of the predicted value y _{0 (t 0)} the amount of temperature drop of the predicted value y ₀ shown in calculating and FIG 21 _{(t 0)} is not used, for reference, it is displayed in Figure 21 ing. Further, in FIG. 21, the scale of the horizontal axis of probability density (corresponding to the horizontal axis of FIG. 7C) is enlarged for easy viewing.

By operating in this way, in the output value predicting apparatus S of the third embodiment, regardless of the distribution mode in the variation of the predicted value y ₀ , the converter steelmaking process goes through the molten steel treatment process and the continuous casting process. In the process, it becomes possible to predict the molten steel temperature in the ladle of the charge and more appropriately determine the variation in the predicted molten steel temperature in the ladle.

  In the third embodiment described above, the predetermined output is the molten steel temperature in the ladle in the process from the converter steelmaking process through the molten steel treatment process to the continuous casting process, but the predetermined output is It may be the molten steel temperature in the tundish in the process from the converter steelmaking process through the molten steel treatment process to the continuous casting process. By being configured in this way, it is possible to predict the molten steel temperature in the tundish in the process from the converter steelmaking process through the molten steel treatment process to the continuous casting process, and obtain the variation in the predicted value. It becomes.

  In the third embodiment described above, the predetermined output may be a molten steel component or a molten steel temperature corresponding to the integrated amount of blown blown oxygen in the converter process. By comprising in this way, it becomes possible to predict the molten steel component or molten steel temperature according to the integration amount of blown blowing oxygen in a converter process, and to obtain | require the dispersion | variation in this estimated predicted value.

  In the third embodiment described above, the predetermined output may be the steel material temperature of the steel material in accordance with the heating time or the integrated amount of heat in the steel material heating furnace step. By comprising in this way, it becomes possible to estimate the steel material temperature of the steel material according to the heating time or the integrated amount of the heating heat amount in the heating furnace process of the steel material, and to obtain the variation of the predicted value.

  In the first to third embodiments described above, the distance calculation unit 11 functionally includes a weight calculation unit, and uses the A weight calculated by the weight calculation unit to predict data and past results. The predetermined distance from the data may be calculated for each of the plurality of past performance data based on the factors of the prediction target data and the past performance data.

  The weight calculation unit is configured to calculate the factor as the Ath weight to which the factor contributes to the magnitude of variation in the predetermined output. More specifically, this weight calculation unit is based on a plurality of past performance data when the magnitude of variation in the predetermined output is an A output variable and the variable related to the factor is an A input variable. An A model representing the relationship between the A input variable and the A output variable is generated, and the A weight is calculated based on the A model. The A-th model is represented by, for example, a function equation having two variables or three or more variables (a regression equation such as a single regression equation or a multiple regression equation), and is regressed from the A input variable and the A output variable. Calculated by calculation. Examples of the regression calculation include a least square method and a partial least square method (PLS, Partial Least Square).

  Here, the degree to which the factor contributes to the magnitude of the variation in the predetermined output is large when the factor greatly contributes to the magnitude of the variation in the predetermined output, and the factor is the magnitude of the variation in the predetermined output. If it does not contribute much to, it will be a small value. That is, when the range of variation in the predetermined output (range of the range) is relatively dependent on the value of the factor, the factor greatly contributes to the magnitude of variation in the predetermined output, and the factor is the predetermined output. The magnitude of the influence on the magnitude of variation in the output is large, and if the range of variation in the predetermined output (range) is relatively independent of the factor value, the factor is the magnitude of variation in the predetermined output. In this case, the influence of the factor on the magnitude of the variation in the predetermined output is a small value. In other words, the magnitude of the A-th weight is determined according to the magnitude of the influence exerted on the predetermined output variation by the factor.

  Further, the weight calculation unit may calculate a B-th weight according to the number of past past record data, and generate the A-th model using the B-th weight. With this configuration, it is possible to reduce the influence of the error included in the factor on the Ath weight, and to generate the Ath model with higher accuracy. As a result, the Ath weight with higher accuracy. Can be calculated.

In addition, the weight calculation unit uses the degree to which the factor contributes to the magnitude of variation in the predetermined output and the degree to which the factor contributes to the magnitude of the absolute value in the predetermined output as the A weight. It may be calculated. With this configuration, the Ath weight a _i can be obtained in consideration of the influence of the factor on the variation in the predetermined output and the influence of the factor on the absolute value of the predetermined output. , The accuracy of the variation of the predicted value is improved.

  In addition, although illustrated also above, in the above-described embodiment, specific examples of x, X, and Z are as follows. In this example, X is a combination of x and Z.

  In the case of the temperature of a predetermined object accommodated in a predetermined container, x can include the volume of the object (for example, liquid state), the number of times the container is used, the solidification temperature of the object, etc. The initial temperature of the object, the initial temperature of the container, the specific heat / density / volume of the object, the specific heat / density / volume of the container, the contact area between the object and the container, and the calculated heat transfer coefficient (for example, physical property values) And the like.

  In the case of the temperature of the molten steel in the ladle, x is, for example, the number of times the ladle is received, the type of deoxidizer, the concentration of molten steel, the state of the ladle in the empty ladle, the ladle transport time, the outgoing steel temperature, the solidification temperature. , Various alloy amounts, steel types, operation groups, treatment dates (for example, the number of days elapsed since January 1, 1900), and Z is, for example, the number of times the ladle receives steel, the molten steel carbon concentration, Examples include a steeling temperature, a solidification temperature, and various alloy amounts.

  In the case of the molten steel temperature in the tundish, x is, for example, the number of times the ladle is received, the type of deoxidizer, the molten steel carbon concentration, the state of the ladle empty, the number of times the tundish is used, and the ladle transport time. Steel temperature, solidification temperature, various alloy amounts, steel type, operation team, molten steel treatment type, treatment date (e.g., number of days since January 1, 1900), molten steel treatment time, temperature rise in molten steel treatment, Casting time etc. can be mentioned, for example, Z is the number of times the ladle is received, ladle transport time, molten steel carbon concentration, outgoing steel temperature, solidification temperature, amount of various alloys, molten steel temperature after molten steel treatment, tundish fire resistance The object temperature, casting time, etc. can be mentioned.

  Moreover, in the case of the molten steel temperature and molten steel component in converter blowing, x is, for example, the amount of steel output, molten iron temperature, molten iron component (C, Si, Mn, P, S, etc.), blowing target temperature, blowing Target components (target molten steel components, C, Mn, P, S, etc.), hot metal compounding ratio, various auxiliary raw material inputs, alloy inputs, slag basicity, slag amount, acid feed rate, number of furnaces, number of lances, lance height In addition, furnace shutdown time, sublance measurement temperature, sublance measurement component, treatment date (for example, the number of days elapsed since January 1, 1900), number of converters, operation group, pre-charge information, oxygenated amount integrated value, etc. Z is, for example, main raw material (hot metal, cold metal, scrap) input amount, hot metal temperature, hot metal component, various auxiliary material input amounts, various auxiliary material compositions, alloy input amount, acid feed amount, blowing rate Stop target component, sublance measurement temperature, sublance measurement component, pre-charge information A oxygen-flow amount integrated values or the like.

  As described in the first to third embodiments, the output value predicting apparatus S can predict the output value in each process of the operation process and the manufacturing process and the output value of the final process directly connected to the product together with variations. Here, a configuration example of an output value prediction system in which the output value prediction device S is applied to an operation process or a manufacturing process will be described.

  FIG. 22 is a block diagram illustrating a configuration of the output value prediction system. 22, the output value prediction system includes a performance data collection device 101 that collects performance data from the operation process / manufacturing process 201, and a past operation data storage that stores the performance data collected by the performance data collection device 101 as past performance data. Apparatus 105, parameter fitting arithmetic device 102 for calculating error parameter A based on past actual data collected by actual data collecting device 101, and parameter estimated value storage device for storing error parameter A calculated by parameter fitting arithmetic device 102 106, a prediction target data operation condition collection device 104 that collects prediction target data from the operation process / manufacturing process 201, a similarity calculation device 108 that calculates a similarity between the prediction target data and each past performance data, and a prediction target Data operating condition collection equipment The prediction target data output prediction calculation device 103 that predicts the output value (prediction value) of the prediction target data based on the calculated error parameter A from the prediction target data collected in 104, and prediction by the prediction target data output prediction calculation device 103 An output predicted value probability distribution estimation device 107 that calculates the probability density of the predicted value of the prediction target data based on the output value (predicted value) of the prediction target data and the similarity calculated by the similarity calculation device 108; And a probability distribution display device 109 for displaying the probability density of the predicted value of the prediction target data calculated by the probability distribution estimation device 107.

  When the output value prediction system shown in FIG. 22 is compared with the output value prediction device S shown in FIG. 1, the similarity calculation device 108 is substantially the same as the distance calculation unit 11, the similarity calculation unit 12, and the intermediate data storage unit 42. The parameter fitting calculation device 102 has substantially the same function as the parameter calculation unit 13, and the prediction target data output prediction calculation device 103 is substantially the same as the prediction value calculation unit 14 and the prediction value storage unit 43. The output predicted value probability distribution estimation device 107 has substantially the same function as the variation calculation unit 15 and the variation storage unit 44, and the past operation data storage device 105 has substantially the same function as the actual measurement data storage unit 41. The parameter estimated value storage device 106 has substantially the same function as the intermediate data storage unit 42.

  In the output value prediction system having such a configuration, when the prediction target data is collected during the execution of the process, the prediction value and the probability density in the prediction target data can be obtained and displayed. . For this reason, a user such as an operator can appropriately adjust the process based on the predicted value in the prediction target data and its probability density, and can execute the process.

  In order to express the present invention, the present invention has been properly and fully described through the embodiments with reference to the drawings. However, those skilled in the art can easily change and / or improve the above-described embodiments. It should be recognized that this is possible. Therefore, unless the modifications or improvements implemented by those skilled in the art are at a level that departs from the scope of the claims recited in the claims, the modifications or improvements are not covered by the claims. It is interpreted that it is included in

S output value prediction device 1 calculation control unit 4 storage unit 11 distance calculation unit 12 similarity calculation unit 13 parameter calculation unit 14 prediction value calculation unit 15 variation calculation unit 41 actual measurement data storage unit 42 intermediate data storage unit 43 prediction value storage unit 44 Variation storage unit 102 Parameter fitting calculation device 103 Prediction target data output prediction calculation device 105 Past operation data storage device 107 Output predicted value probability distribution estimation device 108 Similarity calculation device

Claims (22)

Prediction target data to be predicted, a predetermined output and a first factor that can be quantified relating to the output, and the similarity between a plurality of past performance data acquired in the past, the prediction target data and the past performance data A similarity calculation step for calculating each of the plurality of past performance data based on a predetermined second factor for obtaining the similarity in
When the predetermined output is used as an output variable and a predetermined third factor for obtaining a predicted value to be described later is used as an input variable, and the first model representing the relationship between the output variable and the input variable is generated, the input A difference between a value obtained by giving an input value of a variable to the first model and an output value of the output variable corresponding to the input value of the input variable, and a plurality of predetermined first elements set in advance A parameter calculation step for calculating an error parameter for each of the plurality of past performance data based on input variables and output variables in the past performance data;
A second model representing a relationship between the output variable and the input variable is generated using the input variable and the error parameter, and a value corresponding to the third factor in the prediction target data and a value of the error parameter A predicted value calculation step for calculating each of the plurality of error parameters calculated by the parameter calculation step using the output value of the prediction target data as a predicted value by
A variation calculating step of calculating variations in the output value of the prediction target data based on the plurality of similarities calculated by the similarity calculating step and the plurality of predicted values calculated by the predicted value calculating step. An output value prediction method characterized by The predicted value calculation step includes
A second model representing a relationship between the output variable and the input variable is generated using the input variable and the error parameter, and a factor of the prediction target data is determined for each of the plurality of first elements of the error parameter. A value of the second model is calculated by giving a value of the factor corresponding to the input variable and a value of the first element of the plurality of first elements of the error parameter to the second model. A1 step, and A2 step of calculating the predicted value by weighted averaging the plurality of values calculated in the first A1 step corresponding to each of the plurality of first elements of the error parameter,
The output value prediction method according to claim 1, wherein the first step A1 and the second step A2 are performed for each of a plurality of error parameters calculated by the parameter calculation step. The variation is a histogram,
The variation calculating step includes
A first B1 step in which a plurality of similarities calculated by the similarity calculating step correspond to a plurality of predicted values calculated by the predicted value calculating step, respectively;
Step B2 for dividing a range including at least the plurality of predicted values into a finite number of sections;
The B3 step of calculating the frequency of the section for each of the plurality of sections by adding all the similarities corresponding to the predicted values included in the section. The output value prediction method according to 2. The variation is probability density,
The variation calculation step further includes:
The output value prediction method according to claim 3, further comprising a B4 step of adjusting the scale of the frequency so that the area of the histogram is 1. 5. The similarity calculation step further includes:
A step of converting a predetermined constant or a predetermined constant ratio of similarity into 0 from the smallest one, or converting a similarity below a predetermined threshold into 0 is provided. 4. The output value prediction method according to any one of 3 above. The calculation of similarity in the similarity calculation step includes a predetermined evaluation item for increasing the degree of similarity of past performance data that is closer in time to the prediction target data. The output value prediction method according to claim 5. The first factor that can be quantified related to the output includes a plurality of second elements, and includes at least time as the second element. Output value prediction method. The output value prediction method according to any one of claims 1 to 7, further comprising a presentation step of presenting the variation. The presenting step for presenting the variation includes:
The output value prediction method according to claim 8, wherein an auxiliary display representing a portion outside the predetermined management range set in advance in the predetermined output is presented in accordance with the presentation of the variation. The auxiliary display is
In the case where the variation is a histogram or a probability density and is presented by a graph, a first area display relating to the area of the graph above a predetermined upper limit value and / or a predetermined lower limit value set in advance The output value prediction method according to claim 9, wherein the display is a second area display relating to an area of the graph in the following. The auxiliary display is
When the variation is a histogram or a probability density and is presented by a graph, the area ratio as / of a predetermined area as set in advance from one end of the variation graph to the total area As in the variation graph The output value prediction method according to claim 9, wherein the output value prediction method is a boundary value corresponding to As. The auxiliary display is
When the variation is a histogram or a probability density and is presented by a graph, the area ratio as / of a predetermined area as set in advance from one end of the variation graph to the total area As in the variation graph The output value prediction method according to claim 9, wherein the output value prediction method is a difference between a boundary value corresponding to As and a preset management value. The auxiliary display is
In the case where the variation is a histogram or a probability density and is presented by a graph, a ratio of the first area au1 of the graph over a predetermined upper limit value set in advance to the total area As in the variation graph. The sum au1 of the 1 area ratio au1 / As and the second area ratio au2 / As which is the ratio of the second area au2 of the graph below a predetermined lower limit set in advance to the total area As in the variation graph It is / As + au2 / As or a square sum (au1 / As) ² + (au2 / As) ^2. The output value prediction method according to claim 9, wherein: The auxiliary display is
A plurality of predicted values calculated by the predicted value calculating step and a plurality of similarities calculated by the similarity calculating step in a coordinate system in which the variation takes a predicted value on one coordinate axis and the similarity is taken on the other axis. For the plurality of predicted values, the sum of values obtained by multiplying the absolute value of the difference between a preset target value and the predicted value by the similarity of the predicted value or the target value and the predicted value It is the sum of values obtained by multiplying the square of the difference from the value by the similarity of the predicted value, or the sum of values obtained by multiplying the difference between the management value and the predicted value by the similarity of the predicted value. The output value prediction method according to claim 9. The auxiliary display is
A plurality of predicted values calculated by the predicted value calculating step and a plurality of similarities calculated by the similarity calculating step in a coordinate system in which the variation takes a predicted value on one coordinate axis and the similarity is taken on the other axis. For the predicted value greater than or equal to a preset upper limit value or more among the plurality of predicted values, the absolute value of the difference between the upper limit value and the predicted value greater than or equal to the upper limit value The sum of the values obtained by multiplying the similarity between the upper limit value and the predicted value, or the sum of the values obtained by multiplying the similarity between the upper limit value and the predicted value, and the sum of the values obtained by multiplying the similarity between the upper limit value and the predicted value. The output value prediction method according to claim 9. The auxiliary display is
A plurality of predicted values calculated by the predicted value calculating step and a plurality of similarities calculated by the similarity calculating step in a coordinate system in which the variation takes a predicted value on one coordinate axis and the similarity is taken on the other axis. For the lower limit value below the preset lower limit value of the plurality of predicted values, the absolute value of the difference between the lower limit value and the lower limit value predicted value. The sum of the values obtained by multiplying the similarity between the predicted values below the lower limit value or the sum of the values obtained by multiplying the square of the difference between the lower limit value and the predicted value below the lower limit value by the similarity between the predicted values below the lower limit value The output value prediction method according to claim 9. The auxiliary display is
A plurality of predicted values calculated by the predicted value calculating step and a plurality of similarities calculated by the similarity calculating step in a coordinate system in which the variation takes a predicted value on one coordinate axis and the similarity is taken on the other axis. The absolute value of the difference between the upper limit value and the predicted value greater than or equal to the upper limit value for the predicted value greater than or equal to the upper limit value set in advance or not among the plurality of predicted values. Or a sum of values obtained by multiplying the similarity between the upper limit value and the predicted value by the square of the difference between the upper limit value and the predicted value greater than the upper limit value. With respect to the lower limit value less than or equal to a preset lower limit value among the plurality of predicted values, the absolute value of the difference between the lower limit value and the lower limit value or less predicted value is the lower limit value or less. Multiplied by the similarity of the predicted values Or the sum of values obtained by multiplying the square of the difference between the lower limit value and the predicted value below the lower limit value by the similarity between the lower limit value and the predicted value is used as the lower limit value reference moment. The output value prediction method according to claim 9, wherein the output value prediction moment is a sum of the lower limit value reference moment. The predetermined power is a molten steel temperature in a ladle or a tundish in a process from a converter steeling process to a continuous casting process through a molten steel treatment process. The output value prediction method according to any one of the above items. The output according to any one of claims 1 to 17, wherein the predetermined output is a molten steel component or a molten steel temperature corresponding to an integrated amount of blown blown oxygen in a converter process. Value prediction method. The said predetermined output is the steel material temperature of the said steel material according to the integration amount of the heating time or the amount of heating heat in the steel furnace process, The one of Claim 1 thru | or 17 characterized by the above-mentioned. Output value prediction method. An actual measurement data storage unit for storing a plurality of past performance data acquired in the past, which includes a predetermined output and a first factor that can be quantified related to the output;
Based on a predetermined second factor for determining the similarity between the prediction target data to be predicted and the past performance data, and the similarity in the prediction target data and the past performance data, for each of the plurality of past performance data A similarity calculation unit to calculate,
When the predetermined output is used as an output variable and a predetermined third factor for obtaining a predicted value to be described later is used as an input variable, and the first model representing the relationship between the output variable and the input variable is generated, the input A difference between a value obtained by giving an input value of a variable to the first model and an output value of the output variable corresponding to the input value of the input variable, and a plurality of predetermined first elements set in advance An error parameter comprising: a parameter calculation unit that calculates each of the plurality of past performance data based on input variables and output variables in the past performance data;
A second model representing a relationship between the output variable and the input variable is generated using the input variable and the error parameter, and a value corresponding to the third factor in the prediction target data and a value of the error parameter A predicted value calculation unit for calculating each of the plurality of error parameters calculated by the parameter calculation unit using the output value of the prediction target data as a predicted value by
A variation calculating unit that calculates a variation in the output value of the prediction target data based on the plurality of similarities calculated by the similarity calculating unit and the plurality of predicted values calculated by the predicted value calculating unit. An output value prediction apparatus characterized by the above. An output value prediction program for causing a computer to execute,
Prediction target data to be predicted, a predetermined output and a first factor that can be quantified relating to the output, and the similarity between a plurality of past performance data acquired in the past, the prediction target data and the past performance data A similarity calculation step for calculating each of the plurality of past performance data based on a predetermined second factor for obtaining the similarity in
When the predetermined output is used as an output variable and a predetermined third factor for obtaining a predicted value to be described later is used as an input variable, and the first model representing the relationship between the output variable and the input variable is generated, the input A difference between a value obtained by giving an input value of a variable to the first model and an output value of the output variable corresponding to the input value of the input variable, and a plurality of predetermined first elements set in advance A parameter calculation step for calculating an error parameter for each of the plurality of past performance data based on input variables and output variables in the past performance data;
A second model representing a relationship between the output variable and the input variable is generated using the input variable and the error parameter, and a value corresponding to the third factor in the prediction target data and a value of the error parameter A predicted value calculation step for calculating each of the plurality of error parameters calculated by the parameter calculation step using the output value of the prediction target data as a predicted value by
A variation calculating step of calculating variations in the output value of the prediction target data based on the plurality of similarities calculated by the similarity calculating step and the plurality of predicted values calculated by the predicted value calculating step. An output value prediction program characterized by