JP7348488B2 - Physical property data prediction method and physical property data prediction device - Google Patents

Description

The present invention relates to a physical property data prediction method and a physical property data prediction device for predicting values of physical property data of a vulcanized rubber composition.

In recent years, technologies have been actively proposed that allow computers to perform machine learning to make various predictions from input data. It is also conceivable that the above technique can be applied to predicting the values of physical property data of a vulcanized rubber composition produced by blending a plurality of rubber materials, fillers, oils, etc. as raw materials.
BACKGROUND ART Conventionally, a vulcanized rubber composition has been prepared as a prototype by blending a plurality of rubber materials, fillers, oils, etc. through trial and error, and physical property data has been experimentally measured. For this reason, a large amount of data linking the formulation information of vulcanized rubber compositions and the values of physical property data has been accumulated. Using this accumulated data, it is possible to use machine learning on a computer to predict the values of physical property data.

For example, using a neural network method, we can learn the mapping relationship between a group of factors and a group of characteristics in experimental data such as design and formulation, and estimate characteristic values from factor conditions. A technique is known that provides a method for efficiently and easily determining the optimal value of factor data that produces the following (Patent Document 1).

Japanese Patent Application Publication No. 2003-58582

In neural network learning using this technology, all prepared data is uniformly read and used for multiple pieces of training data. For example, as the plurality of learning data, a large number of accumulated experimental data sets in which information on the blending ratio of raw materials of the vulcanized rubber composition is linked to values of physical property data are used. This accumulated experimental data set is a collection of a large number of data with different experiment dates and times. In the experimental data, each time an experiment is conducted, the blending ratio is determined from the blending ratio of the reference vulcanized rubber composition using the information on the blending ratio of the reference vulcanized rubber composition and the value of the physical property data as a standard for comparison. It is common practice to obtain values of physical property data of the vulcanized rubber composition that has been changed. Therefore, in the plurality of learning data, there are many pairs of information on the blending ratio of the reference vulcanized rubber composition and values of physical property data.

However, the information on the blending ratio of the vulcanized rubber composition and the values of the physical property data, which serve as a reference, are not always constant and vary. Although the blending ratio in the standard vulcanized rubber composition is within a preset range, there may be variations between the past blending ratio and the current blending ratio. The values of physical property data are also data obtained through experiments, so there are variations due to differences in date and time.
Therefore, even if the vulcanized rubber compositions are the same, different blending ratio information and physical property data values may be included in the learning data. This is not preferable for accurately predicting the values of physical property data.

Therefore, the present invention aims to accurately predict the physical property data of a vulcanized rubber composition produced from an unvulcanized rubber composition prepared by combining a plurality of preset raw materials. It is an object of the present invention to provide a physical property data prediction method and a physical property data prediction device that can predict the values of physical property data of a sulfur rubber composition.

One aspect of the present invention is a physical property data prediction method that causes a computer to predict physical property data values of a vulcanized rubber composition produced from an unvulcanized rubber composition blended by combining a plurality of preset raw materials. be. The prediction method is
The values of physical property data regarding a plurality of vulcanized rubber compositions prepared using a plurality of raw materials used in the vulcanized rubber composition are used as learning output data, and the blending ratio of each of the raw materials in the vulcanized rubber composition is calculated. and information on the processing conditions of the vulcanized rubber composition as input data for learning, the relationship between the physical property data, the blending ratio , and the processing conditions is calculated in a computer. A step of machine learning the predictive model;
Information on the blending ratio of constituent raw materials constituting the unvulcanized rubber composition before vulcanization of the vulcanized rubber composition to be predicted, and processing conditions in processing for producing the vulcanized rubber composition to be predicted. using the prediction model to predict the value of the physical property data of the vulcanized rubber composition to be predicted;
Before the predictive model is subjected to machine learning, a plurality of sets of the vulcanized rubber compositions are included, with the value of the physical property data, information on the blending ratio of each of the raw materials, and information on the processing conditions being set as a set. The original data is acquired, and the information on the blending ratio in the vulcanized rubber composition in the original data and the value of the physical property data in the vulcanized rubber composition corresponding to the information are determined from any other vulcanization in the original data. A group of vulcanized rubber compositions satisfying the above conditions is determined by determining whether the information on the blending ratio in the vulcanized rubber composition and the value of the physical property data corresponding to the information satisfy the same condition. creating the learning output data from the original data by extracting at least one and unifying the values of the physical property data in the group of extracted vulcanized rubber compositions to one value.

In the extracted group of vulcanized rubber compositions, for any pair of vulcanized rubber compositions, the ratio of the values of corresponding compounding ratios in the pair is within the first range, and It is preferable that the ratio of values of corresponding physical property data is within a second range, and the first range and the second range are different.

Preferably, the first range is narrower than the second range.

Preferably, the first range is within a range of 0.95 to 1.05.

The conditions include a condition in which the information on the mixing ratio is considered to be the same, and a condition in which the value of the physical property data is considered to be the same,
The condition under which the values of the physical property data are considered to be the same is the value obtained by multiplying the total number of vulcanized rubber compositions that satisfy the conditions under which the information on the mixing ratio is considered to be the same by a predetermined number greater than 0 and less than or equal to 1. It is preferable to set conditions under which the values of the physical property data are considered to be the same, such that the total number of groups of the vulcanized rubber compositions is the set number.

It is preferable that the value to be unified as the value of the physical property data is an average value or an intermediate value of the values of the physical property data in the group of the vulcanized rubber compositions.

Another aspect of the present invention is a computer configured to predict values of physical property data of a vulcanized rubber composition produced from an unvulcanized rubber composition blended by combining a plurality of preset raw materials. This is a physical property data prediction device. The physical property data prediction device is
Output for learning values of physical property data regarding multiple vulcanized rubber compositions produced by blending multiple raw materials used for vulcanized rubber compositions at a set blending ratio and processing under set processing conditions. The value of the physical property data is predicted using learning data in which information on the blending ratio of each of the raw materials in the vulcanized rubber composition and information on each of the processing conditions are input data for learning. A machine learning department that performs machine learning on the model,
Information on the blending ratio of constituent raw materials constituting the unvulcanized rubber composition before vulcanization of the vulcanized rubber composition to be predicted, and processing conditions in processing for producing the vulcanized rubber composition to be predicted. a prediction unit comprising a prediction model that uses the prediction model to predict the value of the physical property data of the vulcanized rubber composition to be predicted;
Before the prediction model is subjected to machine learning, the value of the physical property data, the information on the blending ratio of each of the raw materials, and the information on the processing conditions are set as a set, and the set is included in a plurality of the vulcanized rubber compositions. The original data is prepared, and the information on the blending ratio of the vulcanized rubber composition in the original data and the value of the physical property data of the vulcanized rubber composition corresponding to the information are calculated from any other data in the original data. A group of vulcanized rubber compositions satisfying the above conditions is determined by determining whether the information on the blending ratio in the vulcanized rubber composition and the value of the physical property data corresponding to the information satisfy the same condition. a preprocessing unit that creates the learning output data from the original data by extracting the physical property data of the group of extracted vulcanized rubber compositions and unifying the values of physical property data in the extracted group of vulcanized rubber compositions to one value;
Equipped with

According to the above-described physical property data prediction method and physical property data prediction device, when a computer predicts the values of physical property data of a vulcanized rubber composition, it is possible to accurately predict the values of physical property data of the vulcanized rubber composition. I can do it.

FIG. 2 is a diagram illustrating an example of a method for predicting physical property data by a computer. FIG. 1 is a diagram illustrating an example of the main configuration of a physical property data prediction device according to an embodiment. (a), (b) is a figure explaining each raw material used in the physical property data prediction method of one embodiment.

Hereinafter, a physical property data prediction method and a physical property data prediction device of one embodiment will be described in detail.

The physical property data prediction method of one embodiment causes a computer to machine learn the association between physical property data values and factor data regarding a plurality of vulcanized rubber compositions, and uses a machine learned computer prediction model to This is a method for predicting the values of physical property data of sulfur rubber compositions.
FIG. 1 is a diagram showing an example of a method for predicting physical property data using a computer. FIG. 2 is a diagram illustrating an example of the main configuration of a physical property data prediction device according to an embodiment that performs a physical property data prediction method. The physical property data prediction device 100 is composed of a computer including a CPU 102 and a memory 104. The physical property data prediction device 100 forms at least a preprocessing unit 106, a machine learning unit 108, and a prediction unit 110 as software modules, which function by activating a program stored in the memory 104. The physical property data prediction device 100 is connected to a display (not shown), and is further connected to input operation devices including a mouse and a keyboard.

For machine learning, the preprocessing unit 106 of the physical property data prediction device 100 calculates the values of the physical property data of the vulcanized rubber composition and the blending ratio of each raw material constituting the vulcanized rubber composition, as shown in FIG. Original data including a plurality of sets of vulcanized rubber compositions is obtained by combining the information and processing condition information (ST10 in FIG. 1). The original data is stored and held in the memory 104. The original data is acquired by the physical property data prediction device 100 from a data library in which a large amount of experimental data including physical property data values obtained through experiments is stored in correspondence with information on the vulcanized rubber composition (compounding ratio and processing conditions). It may be transferred to.

The information on the blending ratio includes the name of each raw material in the vulcanized rubber composition and the blending ratio of each raw material. Processing conditions include, for example, the type of mixing machine that mixes raw materials containing unvulcanized rubber, the temperature during mixing, the mixing rotation speed of the mixing machine, the mixing pressure, and the amount of raw materials input to the mixing machine. or vulcanization conditions including at least one of vulcanization temperature and vulcanization time when vulcanizing the mixed unvulcanized rubber. Physical property data regarding the vulcanized rubber composition include, for example, rubber elasticity, tan δ, specific gravity, Mooney Vis viscosity, Mooney scorch, rheometer measurement results, Lübke JIS hardness, modulus, breaking strength, breaking energy, filler dispersity, rebound modulus, constant It includes at least one of strain test results, wear test results, fatigue test results, and crack growth characteristics.

Next, the preprocessing unit 106 determines whether the information on the blending ratio in the vulcanized rubber composition in the original data and the value of the physical property data in the vulcanized rubber composition corresponding to this information are the same as any other in the original data. It is determined whether the conditions for being considered to be the same as the information on the blending ratio in the vulcanized rubber composition and the value of the physical property data corresponding to this information are satisfied (ST12 in FIG. 1). This extracts a group of vulcanized rubber compositions that satisfy the conditions that are considered to be the same. The conditions that are considered to be the same can be input and set by an operator operating the physical property data prediction device 100 using an input operation device (not shown). As a result of the determination, information (blending ratio, physical property data) on a group of vulcanized rubber compositions that satisfy the conditions that are considered to be the same is displayed on a display (not shown).

To determine whether the conditions for being considered to be the same are satisfied, it is possible to determine whether the information on the mixture ratio and the value of the physical property data are the same, but it is also possible to determine whether the information on the mixture ratio is the same or not. The determination and the determination of whether the values of the physical property data are the same may be performed separately. The conditions for information on mixture ratios to be considered the same may be, for example, that there are multiple mixture ratios and the ratios of the mixture ratios to be compared are all within a predetermined range, or there may be a correlation coefficient between the mixture ratios to be compared. The correlation coefficient may be greater than or equal to a predetermined value. The condition for the values of physical property data to be considered to be the same may be that when there is only one physical property data, the ratio of the values of the physical property data to be compared is within a predetermined range, and when there are multiple physical property data, the condition for the physical property data to be compared to be the same. It may be determined that the ratio between the data values is within a predetermined range, or it may be determined that the correlation coefficient between the values of the physical property data to be compared is calculated and that this correlation coefficient is greater than or equal to a predetermined value. Good too. The conditions for the information on the mixture ratio and the values of the physical property data to be considered to be the same may be that the ratios of the mixture ratios to be compared and the values of the physical property data are both within a predetermined range, or the conditions for the ratio of the mixture ratios to be compared and the values of the physical property data to be within a predetermined range. A correlation coefficient between values of physical property data may be calculated, and the correlation coefficient may be equal to or greater than a predetermined value.

Next, the preprocessing unit 106 creates learning output data from the original data by unifying the physical property data values of the group of extracted vulcanized rubber compositions into one value (ST14 in FIG. 1). The value of the physical property data to be unified is, for example, a representative value of the physical property data of vulcanized rubber compositions satisfying the conditions considered to be the same, and may be an average value or an intermediate value.

Thereafter, the machine learning unit 108 uses the values of physical property data regarding the vulcanized rubber composition as output data for learning, and learns information on the blending ratio of each raw material in the vulcanized rubber composition and information on each processing condition. Using the learning data used as input data, the prediction model in the prediction unit 110 in the computer learns the relationship between the mixture ratio information, processing condition information, and physical property data values (Fig. 1 ST16). As the machine learning of the predictive model, for example, deep learning using a neural network is used. Additionally, a random forest using a tree structure can be used. As the model, a known model such as a convolutional neural network or a stag autoencoder can be used. In this way, the machine learning of the predictive model is completed.

Next, the prediction unit 110 provides information on the blending ratio of the constituent raw materials constituting the unvulcanized rubber composition before vulcanization of the vulcanized rubber composition to be predicted, and Using the processing conditions in the processing, the prediction model predicts the physical property data of the vulcanized rubber composition to be predicted (FIG. 1 ST18).

In this way, the learning data when the prediction model performs machine learning is not the original data itself, but the information on the blending ratio of the vulcanized rubber composition in the original data and the information on the vulcanized rubber composition corresponding to this information. This data is created by unifying the physical property data values of a group of vulcanized rubber compositions that satisfy the condition that the physical property data values are considered to be the same to one value. As a result, the values of the physical property data of the vulcanized rubber compositions, which should originally be considered the same, are unified into one value, so the machine learning prediction unit can accurately calculate the values of the physical property data of the vulcanized rubber compositions to be predicted. Can be predicted well.

FIG. 3(a) is a diagram illustrating an example in which a part of the original data is simplified and described. As shown in FIG. 3(a), as a vulcanized rubber composition, No. 1~No. 4 is shown.
For the vulcanized rubber composition, raw materials A to D and the value of one physical property data are shown. Here, the raw materials A and B are, for example, different rubber materials, and the raw materials C and D are, for example, filler materials such as carbon and silica. In addition, a vulcanization accelerator and a plasticizer such as oil can be added as raw materials.

Vulcanized rubber composition No. The weight-based compounding ratio of raw material A in No. 1 is 0.430, taking the total weight as 1. The weight-based blending ratio of raw material B in No. 1 means 0.350.

The physical property data is, for example, the rubber elastic modulus at a predetermined elongation at a predetermined temperature. In the example shown in FIG. 3(a), one physical property data is provided, but a plurality of physical property data may be provided. Vulcanized rubber composition No. The value of the physical property data for No. 1 is 10.0.
On the other hand, vulcanized rubber composition No. The blending ratio in vulcanized rubber composition No. 2 is as follows. The blending ratio is slightly different from that in vulcanized rubber composition No. 1, and the physical property data values are also different from those in vulcanized rubber composition No. 1. The blending ratio is slightly different from that in 1. Similarly, vulcanized rubber composition No. The blending ratio and physical property data values for vulcanized rubber composition No. 3 and 4 are also the same as those for vulcanized rubber composition No. The values of the blending ratio and physical property data in 1 and 2 are slightly different.

FIG. 3(b) is a diagram showing an example of data extracted as learning data from the original data.
Vulcanized rubber composition No. shown in FIG. 3(b). 1 and no. 2 is a group of vulcanized rubber compositions that are considered to be the same. Here, the conditions that are considered to be the same regarding the blending ratio and the conditions that are considered to be the same regarding the values of physical property data are different from each other, and as an example, the conditions that are considered to be the same regarding the blending ratio are such that the ratio of the corresponding blending ratios is 0. The following description will be made on the assumption that the ratio of the physical property data values is in the range of 0.93 to 1.07. It is more preferable that the conditions regarding the blending ratios to be considered the same are that the ratio between the corresponding blending ratios is in the range of 0.97 to 1.03. It is more preferable that the ratio between the physical property data values is in the range of 0.95 to 1.05.
Sulfur rubber composition No. shown in FIG. 3(a). 1 and vulcanized rubber composition No. 1. 2 satisfies the condition that the corresponding blending ratios are considered to be the same (ratio is 0.95 to 1.05). However, vulcanized rubber composition No. 3 is vulcanized rubber composition No. 3. 1, the condition that the ratios of the corresponding blending ratios are considered to be the same (ratio of 0.95 to 1.05) is not satisfied. Specifically, vulcanized rubber composition No. Vulcanized rubber composition No. 3 with a blending ratio of raw material B of 0.370. The ratio of 1 to the blending ratio of raw material B of 0.350 is 1.057. Vulcanized rubber composition No. Vulcanized rubber composition No. 3 with a blending ratio of raw material B of 0.370. Even if the ratio of raw material B of No. 2 to the blending ratio of 0.362 is within the range of 0.95 to 1.05, vulcanized rubber composition No. 1 and no. It does not satisfy the conditions for being considered the same as Group 2.
On the other hand, vulcanized rubber composition No. 4 is vulcanized rubber composition No. 4. In contrast to vulcanized rubber composition No. 1 and 2, the corresponding compounding ratios satisfy the condition that the ratios are considered to be the same (ratio is 0.95 to 1.05). 1, the ratio of the corresponding physical property data value is 1.08, which does not satisfy the condition for the physical property data values to be considered to be the same (ratio of 0.93 to 1.07).
Therefore, the vulcanized rubber compositions satisfying the conditions considered to be the same are No. 1 and no. It becomes 2.
A group of vulcanized rubber compositions that satisfy such conditions that are considered to be the same is determined based on the conditions that are considered to be the same, using cluster analysis in advance. Groups can be set.

In addition, in the actual original data, the values of the physical property data of the standard vulcanized rubber composition obtained by conducting experiments at different dates and times vary, and the blending ratio of raw materials also varies, but originally, Since the experiments were conducted using the same vulcanized rubber composition, the physical property data values and blending ratios are distributed around a certain value, although there are variations. Therefore, for one vulcanized rubber composition that is originally the same, a group of vulcanized rubber compositions that satisfy the conditions for being considered the same is created as one group, and multiple groups are created. There isn't.
The thus extracted vulcanized rubber composition No. The physical property data values of No. 1 and No. 2 are unified into one value. At this time, the blending ratio is not unified. In the example shown in FIG. 3(b), vulcanized rubber composition No. The average value of the physical property data values of No. 1 and No. 2 is unified to 10.3 (=(10.0+10.5)/2).
In this way, the values of physical property data of a group of vulcanized rubber compositions that are considered to be the same are unified to one value, and learning data is created.

In this way, the learning data is not the original data itself, but the information on the blending ratio of the vulcanized rubber composition in the original data and the values of the physical property data corresponding to this information satisfy the condition that they are considered to be the same. This data is created by unifying the physical property data values for a group of vulcanized rubber compositions into one value. Therefore, it is possible to obtain information on the blending ratio of the reference vulcanized rubber composition and learning data that suppresses fluctuations in the values of physical property data, and the physical property data of the target vulcanized rubber composition to be predicted from the machine learned prediction model. The value of can be predicted with high accuracy. In particular, by unifying the values of the physical property data to one value, it is possible to accurately predict the value of the physical property data of the vulcanized rubber composition to be predicted.

As mentioned above, if the conditions for considering the vulcanized rubber compositions to be compared as the same include conditions regarding the blending ratio and conditions regarding the value of physical property data, then in the group of extracted vulcanized rubber compositions, , for any pair of vulcanized rubber compositions, the ratio of the values of the corresponding compounding ratios in the pair is within the first range, and the ratio of the values of the corresponding physical property data in the pair is: If it is within the second range, it is preferable that the first range and the second range are set to be different. Variations in the blending ratio of the vulcanized rubber composition and variations in the values of physical property data are often different, and the first range and the second range are set to be different in consideration of the degree of variation. By doing so, it is possible to accurately extract a group of vulcanized rubber compositions that are considered to be the same.

Preferably, the first range is narrower than the second range. Variations in the values of physical property data are determined through experiments, but they tend to change and have large fluctuations if the experiment dates and times differ. On the other hand, the target blending ratio is predetermined according to design guidelines, and the variation thereof is relatively small. Therefore, the first range is preferably narrower than the second range in order to accurately extract the reference vulcanized rubber composition.
Preferably, the first range is within the range of 0.95 to 1.05. By setting the first range within this range, the reference vulcanized rubber composition can be extracted with high accuracy.

In addition, if the conditions under which the vulcanized rubber compositions to be compared are considered to be the same include the conditions under which the information on the blending ratio is considered to be the same and the conditions under which the values of physical property data are considered to be the same, the information on the blending ratio The set number is the value obtained by multiplying the total number of vulcanized rubber compositions that satisfy the conditions that are considered to be the same by a predetermined number greater than 0 and less than or equal to 1, and the total number of groups of vulcanized rubber compositions to be extracted is It is preferable that the conditions for the values of the physical property data to be considered to be the same are set so that the above set number is achieved. For example, while the first range is fixed, a second range is set by variously changing the second range from among vulcanized rubber compositions that satisfy the condition that the information on the blending ratio is considered to be the same. The second range can be set by determining the number of vulcanized rubber compositions that satisfy the conditions for having the same physical property data values. Thereby, the second range can be freely set depending on the magnitude of variation in the values of physical property data, and the total number of vulcanized rubber compositions that are considered to be the same can be secured as an appropriate number.

The unified value set for the physical property data in a group of vulcanized rubber compositions that are considered to be the same is preferably an average value or an intermediate value of the physical property data values in the group of vulcanized rubber compositions.

(Example, comparative example)
The above-mentioned physical property data prediction method was performed using a vulcanized rubber composition as the tread rubber used in tires, and its effects were confirmed.
The blending ratio of the original data for tread rubber is the blending ratio of rubber raw materials containing at least butadiene rubber, styrene-butadiene rubber, and natural rubber as rubber raw materials, and the blending ratio of 50 types of carbon black with different specific surface areas and specific surface areas as filler raw materials. The blending ratio of filler raw materials containing at least 30 types of silica with different values, the blending ratio of 20 types of vulcanization accelerators as vulcanization accelerators, and the blending ratio of 200 types of plasticizers containing at least oil and resin as plasticizers. Contains.
The original data includes 30,000 sets of information on the blending ratio and processing conditions of the vulcanized rubber composition and physical property data values.
The physical property data includes the elastic modulus at 2% elongation at 20° C. and tan δ (loss tangent) at a frequency of 20 Hz at 20° C.

As an example, the conditions under which the vulcanized rubber compositions are considered to be the same are divided into the conditions under which the blending ratios are considered to be the same and the conditions under which the values of physical property data are considered to be the same, and the conditions under which the blending ratios are considered to be the same are as described above. The first range was set to 0.97 to 1.03, and the second range was set to 0.9 to 1.1 as a condition for the values of physical property data to be considered to be the same.

By pre-processing the original data under these conditions, 200 groups of vulcanized rubber compositions were extracted as standards. That is, it shows that the original data includes 200 pieces of data for the vulcanized rubber composition that serves as a reference. Therefore, the values of the physical property data of the group of reference vulcanized rubber compositions were unified to one representative value for each group. The average value was used as the representative value. Using the learning data created from the original data through preprocessing in this manner, the prediction model of the prediction unit 110 was subjected to machine learning.
On the other hand, as a comparative example, the prediction model of the prediction unit 110 was subjected to machine learning using original data as learning data without preprocessing.

Each of the machine learning-based prediction models of Examples and Comparative Examples was used to predict values of physical property data of four types of tread rubber having different blending ratios. The values of the physical property data (the above elastic modulus and the above tan δ) of these four types of tread rubber were obtained in advance through experiments.

Among the errors between the experimental values of the elastic modulus and tan δ of the four types of tread rubber and the corresponding predicted values predicted by the prediction model, the largest error expressed as a percentage is 6 to 6 in the case of the example. 7%, whereas in the case of the comparative example it was 8 to 15%, and the error in the example was smaller than that in the comparative example.
From this, it can be seen that the above-mentioned physical property data prediction method can accurately predict the values of the physical property data of the vulcanized rubber composition when the computer predicts the values of the physical property data of the vulcanized rubber composition. .

Although the physical property data prediction method and the physical property data prediction device of the present invention have been described in detail above, the present invention is not limited to the above embodiments, and various improvements and changes may be made without departing from the spirit of the present invention. Of course it's good.

100 Physical property data prediction device 102 CPU
104 Memory 106 Preprocessing unit 108 Machine learning unit 110 Prediction unit

Claims (7)

A physical property data prediction method that causes a computer to predict physical property data values of a vulcanized rubber composition produced from an unvulcanized rubber composition blended by combining a plurality of preset raw materials, the method comprising:
The values of physical property data regarding a plurality of vulcanized rubber compositions prepared using a plurality of raw materials used in the vulcanized rubber composition are used as learning output data, and the blending ratio of each of the raw materials in the vulcanized rubber composition is calculated. and information on the processing conditions of the vulcanized rubber composition as input data for learning, the relationship between the physical property data, the blending ratio , and the processing conditions is calculated in a computer. A step of machine learning the predictive model;
Information on the blending ratio of constituent raw materials constituting the unvulcanized rubber composition before vulcanization of the vulcanized rubber composition to be predicted, and processing conditions in processing for producing the vulcanized rubber composition to be predicted. using the prediction model to predict the value of the physical property data of the vulcanized rubber composition to be predicted;
Before the predictive model is subjected to machine learning, a plurality of sets of the vulcanized rubber compositions are included, with the value of the physical property data, information on the blending ratio of each of the raw materials, and information on the processing conditions being set as a set. The original data is acquired, and the information on the blending ratio in the vulcanized rubber composition in the original data and the value of the physical property data in the vulcanized rubber composition corresponding to the information are determined from any other vulcanization in the original data. A group of vulcanized rubber compositions satisfying the above conditions is determined by determining whether the information on the blending ratio in the vulcanized rubber composition and the value of the physical property data corresponding to the information satisfy the same condition. creating the learning output data from the original data by extracting at least one and unifying the values of the physical property data in the group of extracted vulcanized rubber compositions to one value;
A physical property data prediction method characterized by the following. In the extracted group of vulcanized rubber compositions, for any pair of vulcanized rubber compositions, the ratio of the values of corresponding compounding ratios in the pair is within the first range, and 2. The physical property data prediction method according to claim 1, wherein a ratio of values of corresponding physical property data is within a second range, and the first range and the second range are different. The physical property data prediction method according to claim 2, wherein the first range is narrower than the second range. The physical property data prediction method according to claim 2 or 3, wherein the first range is within a range of 0.95 to 1.05. The conditions include a condition in which the information on the mixing ratio is considered to be the same, and a condition in which the value of the physical property data is considered to be the same,
The condition under which the values of the physical property data are considered to be the same is the value obtained by multiplying the total number of vulcanized rubber compositions that satisfy the conditions under which the information on the mixing ratio is considered to be the same by a predetermined number greater than 0 and less than or equal to 1. is a set number, and conditions are set for the values of the physical property data to be considered to be the same so that the total number of groups of the vulcanized rubber composition is the set number. Physical property data prediction method described in . The physical property data prediction according to any one of claims 1 to 5, wherein the value to be unified as the physical property data value is an average value or an intermediate value of the physical property data values in the group of the vulcanized rubber compositions. Method. A physical property data prediction device configured with a computer that predicts physical property data values of a vulcanized rubber composition produced from an unvulcanized rubber composition compounded by combining a plurality of preset raw materials,
Output for learning values of physical property data regarding multiple vulcanized rubber compositions produced by blending multiple raw materials used for vulcanized rubber compositions at a set blending ratio and processing under set processing conditions. The value of the physical property data is predicted using learning data in which information on the blending ratio of each of the raw materials in the vulcanized rubber composition and information on each of the processing conditions are input data for learning. A machine learning department that performs machine learning on the model,
Information on the blending ratio of constituent raw materials constituting the unvulcanized rubber composition before vulcanization of the vulcanized rubber composition to be predicted, and processing conditions in processing for producing the vulcanized rubber composition to be predicted. a prediction unit comprising a prediction model that uses the prediction model to predict the value of the physical property data of the vulcanized rubber composition to be predicted;
Before the prediction model is subjected to machine learning, the value of the physical property data, the information on the blending ratio of each of the raw materials, and the information on the processing conditions are set as a set, and the set is included in a plurality of the vulcanized rubber compositions. The original data is prepared, and the information on the blending ratio of the vulcanized rubber composition in the original data and the value of the physical property data of the vulcanized rubber composition corresponding to the information are calculated from any other data in the original data. A group of vulcanized rubber compositions satisfying the above conditions is determined by determining whether the information on the blending ratio in the vulcanized rubber composition and the value of the physical property data corresponding to the information satisfy the same condition. a preprocessing unit that creates the learning output data from the original data by extracting the physical property data of the group of extracted vulcanized rubber compositions and unifying the values of physical property data in the extracted group of vulcanized rubber compositions to one value;
A physical property data prediction device comprising:

Images