Abnormality detection method and system of physical property measurement data of vulcanization rubber

Abstract

To provide an abnormality detection method and a system for easily determining an abnormality caused by measurement of physical property measurement data of vulcanization rubber, and having high versatility.SOLUTION: An abnormality detection method is configured to: set a standard S of pieces of data of a prescribed physical property of a test sample 13 of vulcanization rubber, the data being continuously acquired for a prescribed period in once measurement by a physical property measurement device 8; use as the test sample 13, a suitable sample which has suitable quality suitable for measurement of the prescribed physical property of the vulcanization rubber; by the physical property measurement device 8, acquire physical property measurement data D about the prescribed physical property of the suitable sample 13; by a calculation unit 2, compare the acquired physical property measurement data D with the standard S, and if the difference of the comparison exceeds an allowable range Ar, determine that the physical property measurement data D has an abnormality due to the measurement and output the determination result to an output part 7.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present invention relates to a method and system for detecting abnormalities in physical property measurement data of vulcanized rubber, and more specifically, to a method and system for detecting abnormalities in physical property measurement data of vulcanized rubber that can easily detect abnormalities caused by measurement of physical property measurement data of vulcanized rubber obtained by various physical property tests and is highly versatile. [Background technology]

[0002] Physical property tests such as tensile tests and viscoelastic tests are known as methods for evaluating the properties of vulcanized rubber. The physical property measurement data from these physical property tests may contain abnormal data due to the settings of the measuring device or the measurement environment. Therefore, in order to ensure the reliability of these physical property tests, it is necessary to understand whether the physical property measurement data is abnormal and to eliminate the abnormal physical property measurement data.

[0003] A device has been proposed that detects abnormalities caused by measurement by setting acceptable and unacceptable ranges for the deviation of X-Y coordinates (coordinates on a plane perpendicular to the tensile direction) between grippers during tensile strength measurement and judging whether the measurement is acceptable or unacceptable (see Patent Document 1). The invention proposed in Patent Document 1 requires modification of the measurement device, such as by configuring the gripper of the measurement device to be freely movable in the XY directions. In addition, since this invention focuses only on tensile testing, it cannot be applied to other physical property tests such as viscoelasticity testing and friction and wear testing. Therefore, there is room for improvement in order to more easily detect abnormalities in the physical property measurement data of vulcanized rubber obtained by various physical property tests. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] JP 2006-145394 A Summary of the Invention [Problem to be solved by the invention]

[0005] An object of the present invention is to provide a method and system for detecting anomalies in physical property measurement data of vulcanized rubber that can easily detect anomalies caused by measurement of physical property measurement data of vulcanized rubber obtained by various physical property tests, and that is highly versatile. [Means for solving the problem]

[0006] To achieve the above-mentioned object, the method for detecting abnormalities in physical property measurement data of vulcanized rubber of the present invention detects abnormalities in physical property measurement data of a test sample of vulcanized rubber, which is obtained by a physical property measuring device for a predetermined period of time during a single measurement, and is characterized in that a standard for the data of the specified physical property of the vulcanized rubber is set, the physical property measurement data is obtained using a conforming sample having a quality suitable for measurement of the specified physical property of the vulcanized rubber as the test sample, and the presence or absence of an abnormality in the physical property measurement data of the conforming sample is determined based on a comparison of the obtained physical property measurement data of the conforming sample with the standard.

[0007] The system for detecting abnormalities in physical property measurement data of vulcanized rubber of the present invention detects abnormalities in physical property measurement data of a test sample of vulcanized rubber obtained by a physical property measuring device for a predetermined period of time during a single measurement for a predetermined physical property of the vulcanized rubber, and is characterized in that the system has a calculation device to which a standard for the predetermined physical property of the vulcanized rubber and an acceptable range for this standard are input, and which inputs the physical property measurement data obtained by the physical property measuring device using a conforming product sample having a quality suitable for the measurement of the predetermined physical property of the vulcanized rubber as the test sample, compares the input physical property measurement data of the conforming product sample with the standard value and determines whether the difference between the two exceeds the acceptable range, and an output unit to which the result of the determination made by the calculation device is output. Effect of the Invention

[0008] According to the present invention, the conforming product sample is used as the test sample to obtain the physical property measurement data, so that a non-conforming test sample is not used for this measurement. Therefore, when determining whether or not there is an abnormality in the physical property measurement data, an abnormality caused by the test sample used is eliminated. Therefore, in the present invention, since it is assumed that there is no abnormality caused by the individual test sample used, it is possible to grasp whether or not there is an abnormality caused by the measurement of the physical property measurement data of the test sample (conforming product sample). Then, based on a comparison between the set standard of the data of the predetermined physical property of the vulcanized rubber and the physical property measurement data of the conforming product sample, it is determined whether or not there is an abnormality in the physical property measurement data of the conforming product sample, so that it is possible to apply even if the physical property measurement data is physical property measurement data of vulcanized rubber obtained by various physical property tests. Therefore, the present invention can easily detect an abnormality caused by the measurement of the physical property measurement data of the vulcanized rubber and has high versatility. [Brief description of the drawings]

[0009] [Figure 1] 1 is an explanatory diagram illustrating an embodiment of an anomaly detection system for physical property measurement data of vulcanized rubber; [Diagram 2] 1 is a flow diagram illustrating the procedure of an embodiment of a method for detecting anomalies in physical property measurement data of vulcanized rubber. [Diagram 3] FIG. 2 is an explanatory diagram illustrating an example of physical property measurement data obtained by a physical property measuring device. [Figure 4] FIG. 11 is an explanatory diagram illustrating another example of physical property measurement data obtained by a physical property measuring device. [Diagram 5] FIG. 5 is an enlarged view of a portion of the physical property measurement data D4 shown in FIG. [Figure 6] FIG. 5 is an enlarged view of a portion of the physical property measurement data D5 shown in FIG. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0010] The method and system for detecting anomalies in physical property measurement data of vulcanized rubber according to the present invention will be described below with reference to the embodiment shown in the drawings. For example, this embodiment is applied to vulcanized rubber used in various rubber products such as tires, conveyor belts, and hoses.

[0011] 1 is used to detect an abnormality caused by measurement of physical property measurement data D of a vulcanized rubber test sample 13 acquired by a physical property measuring device 8. This physical property measurement data D is data regarding a predetermined physical property of vulcanized rubber, which is acquired by the physical property measuring device 8 continuously for a predetermined period of time during a single measurement of one test sample 13.

[0012] In this embodiment, a vulcanized rubber tensile tester is used as the physical property measuring device 8. A tensile test specified in JIS K6251:2017 is performed by this physical property measuring device 8, and a tensile stress-strain curve (SS curve) is obtained as physical property measurement data D of a test sample (dumbbell-shaped test piece) 13 of the vulcanized rubber.

[0013] This physical property measuring device 8 has a gripping unit 9 that grips and pulls both ends of a test sample 13, an elongation detection unit 10 that measures the elongation of the test sample 13, a force detection unit 11 that detects the tensile force acting on the test sample 13, and a calculation unit 12 that calculates the detection data by the detection units 10 and 11 as physical property measurement data D. The physical property measurement data D is input from the calculation unit 12 to the anomaly detection system 1. The elongation detection unit 10 measures the distance between reference points (marked lines) 14 marked on the test sample 13. For example, a plurality of digital camera devices with different shooting ranges are used for the elongation detection unit 10.

[0014] The physical property measuring device 8 is not limited to a tensile tester, and devices for measuring various known physical properties of vulcanized rubber can be used. Examples of the physical property measuring device 8 include a viscoelasticity tester for vulcanized rubber, an abrasion tester, and a friction tester. The physical property measurement data D is a series of time-continuous data that is continuously acquired for a predetermined time (e.g., several seconds to several tens of seconds) for one test sample 13 during one measurement by the physical property measuring device 8, such as the SS curve, the elastic modulus, the loss tangent, and the friction coefficient.

[0015] The anomaly detection system 1 has a calculation device 2 and an output unit 7. A known computer can be used as the calculation device 2, and has a calculation processing unit (CPU) 3, a main storage unit (memory) 4, an auxiliary storage unit (e.g., HDD) 5, and an input unit (keyboard, mouse) 6. The calculation device 2 performs various calculation processes using various stored data and various input data. Physical property measurement data D acquired by a physical property measuring device 8 is input and stored in the auxiliary storage unit 5.

[0016] A display, a printer, etc. are used as the output unit 7. The output unit 7 outputs the results of the arithmetic processing by the arithmetic unit 2 and the data stored and input in the arithmetic unit 2.

[0017] The present invention does not detect abnormalities caused by the individual test sample 13, but detects abnormalities caused by the measurement when acquiring the physical property measurement data D of the test sample 13 using the physical property testing device 8. Examples of abnormalities caused by this measurement include a defect where the test sample 13 is not securely gripped by the gripping unit 9 and the test sample 13 shifts when pulled, a defect where the reference points 14 are stretched and become unclear when the test sample 13 is pulled, and a defect where the distance between the reference points 14 cannot be measured accurately when switching from one elongation detection unit (camera device) 10 to the other elongation detection unit (camera device) as the distance between the reference points 14 increases, and the like.

[0018] Next, an example of the procedure of an embodiment of a method for detecting an anomaly in physical property measurement data of vulcanized rubber using the anomaly detection system 1 described above will be described.

[0019] As shown in the procedure illustrated in FIG. 2, a conforming product sample 13A having a quality suitable for measuring predetermined physical properties of vulcanized rubber is used as the test sample 13. The physical property measurement data D of this conforming product sample 13A is obtained by the physical property measuring device 8 (S110). The conforming product sample 13A means a sample that is free of morphological defects as the test sample 13 and free of abnormalities in its physical properties. The fact that the test sample 13 is free of morphological defects and free of abnormalities in its physical properties can be understood by the experience of a person skilled in the art (appearance, texture, etc.). For this reason, test samples 13 that are considered to be morphologically defective or have abnormalities in their physical properties are excluded in advance, and only the conforming product sample 13A is used to obtain the physical property measurement data D.

[0020] Next, the calculation device 2 judges whether or not there is an abnormality in the physical property measurement data D based on a comparison between the standard S of the data of the predetermined physical properties of the vulcanized rubber test sample 13 and the acquired physical property measurement data D of the conforming product sample 13A (S120). This standard S is set individually depending on the type of vulcanized rubber, etc. For example, an allowable range Ar for the standard S is set, and if the magnitude of the difference between the physical property measurement data D and the standard S exceeds the allowable range Ar, it is judged that there is an abnormality in the physical property measurement data D, and if the magnitude of this difference is within the allowable range Ar, it is judged that there is no abnormality in the physical property measurement data D (normal). The details of the standard S and the allowable range Ar will be described later.

[0021] Next, the result of the judgment made by the calculation device 2 is output to the output unit 7 (S130). The output unit 7 outputs the result of the judgment made by the calculation device 2, indicating that there is an abnormality or no abnormality in the physical property measurement data D of the conforming product sample 13A, so that the person in charge can know from this output that there was an abnormality caused by the measurement when acquiring the physical property measurement data D.

[0022] The contents of each step (S110 to S130) will be described in detail.

[0023] In step (S110), the physical property measurement data D of the conforming product sample 13A is measured and obtained by the physical property measurement device 8. For vulcanized rubber of the same specifications (e.g., vulcanized rubber from the same production batch) manufactured under the same compounding and manufacturing conditions, typically three or more conforming product samples 13A manufactured using the same vulcanized rubber are used to determine whether or not there is an abnormality in the physical property measurement data D of the conforming product samples 13A. Alternatively, it is also possible to use one conforming product sample 13A to determine whether or not there is an abnormality in the physical property measurement data D. The physical property measurement data D acquired by the physical property measurement device 8 is input to the calculation device 2 and stored in the auxiliary storage unit 5.

[0024] In step (S120), a standard S for data of a predetermined physical property of the vulcanized rubber test sample 13 and an allowable range Ar for this standard S are set. This standard S is data close to optimal for which the physical property measurement data D of the test sample 13 acquired by the physical property measuring device 8 can be determined to be normal. The allowable range Ar is a slight allowance for values ​​centered on this standard S, and if the physical property measurement data D of the test sample 13 is within the allowable range Ar for the standard S, it can be determined that the physical property measurement data D of the test sample 13 is normal. In other words, if the physical property measurement data D of the test sample 13 exceeds the allowable range Ar for the standard S, it can be determined that the physical property measurement data D of the test sample 13 is abnormal.

[0025] The narrower the allowable range Ar, the more likely it is that the physical property measurement data D is judged to have an abnormality, and the wider the allowable range, the less likely it is that the physical property measurement data D is judged to have an abnormality. This allowable range Ar may be set appropriately for each specification of vulcanized rubber based on accumulated past data. A person skilled in the art can appropriately set the allowable range Ar by referring to experience gained in the process of manufacturing and developing rubber materials and rubber products, various publicly known documents, a large amount of accumulated experimental and test data, accumulated computer simulation results, and the like.

[0026] This standard S can be set by various methods. For example, the standard S can be set based on the physical property measurement data Dg of the conforming product sample 13A obtained in the past when there is no abnormality. Alternatively, the standard S can be set based on the physical property measurement data D of the conforming product sample 13A obtained by the physical property measuring device 8 when determining the presence or absence of an abnormality, without using past data.

[0027] When setting based on the physical property measurement data D of the conforming product sample 13A obtained when determining the presence or absence of anomalies without using past data, specifically, a plurality of physical property measurement data D are compared with each other. In this method described with reference to Fig. 3, one of the measured physical property measurement data D is set as a reference S, and it is determined whether or not the difference between the physical property measurement data D serving as the reference S and the other physical property measurement data D exceeds the allowable range Ar. The distance between the data D (known Euclidean distance or Manhattan distance) can be used as the difference between the respective data D.

[0028] The physical property measurement data D (D1, D2, D3) shown in Fig. 3 are tensile stress-strain curves (SS curves). These physical property measurement data D were obtained from three conforming product samples 13A manufactured using the same production batch of vulcanized rubber. The physical property measurement data D2 shown by the dashed line in Fig. 3 and the physical property measurement data D3 shown by the dashed line are the standard S data.

[0029] In this method, it is determined whether the difference between the physical property measurement data D2, D3 serving as the reference S and the physical property measurement data D1 shown by the solid line exceeds the allowable range Ar. Therefore, as the data distance, a representative value of the distance between the stresses at the same strain position between the physical property measurement data D is used. The representative value of the distance for all combinations of the physical property measurement data D1, D2, D3 (D1 and D2, D2 and D3, D3 and D1) is calculated using the following formula (1). In the following formula (1), D(y1, y2) is the representative value of the distance, the strain positions are from 1st to Nth, and y1i, y2i are the stresses at the position of the i-th strain.

[0030]

number

[0031] In Fig. 3, the representative value of the data distance between the physical property measurement data D1 and D2 is 5 [MPa], the representative value of the data distance between the physical property measurement data D1 and D3 is 6 [MPa], and the representative value of the data distance between the physical property measurement data D2 and D3 is 1 [MPa]. If the allowable range Ar is set to ±4 [MPa], the representative value of the data distance between the physical property measurement data D1 and D2 exceeds the allowable range Ar, and the representative value of the data distance between the physical property measurement data D1 and D3 also exceeds the allowable range Ar. In other words, the physical property measurement data D1 is determined to exceed the allowable range Ar with respect to the approximation data D2 and D3, and the physical property measurement data D1 is determined to have an abnormality due to measurement.

[0032] In this method, the difference from the standard S is not limited to the distance between the data. For example, the difference from the standard S may be a representative value of the differential value of each physical property measurement data D. The differential value in the tensile stress-strain curve represents the slope of the physical property measurement data D, and the representative value of the slope differs between physical property measurement data D with an abnormality and physical property measurement data D without an abnormality.

[0033] The second method, which will be described with reference to FIG. 4, compares the measured physical property measurement data D with predicted data Da, which is obtained on the assumption that the measured physical property measurement data D is free from abnormalities due to measurement. That is, this is a method for analyzing the difference between the predicted value and the measured value. In this method, the calculation device 2 predicts the predicted data Da, sets the predicted predicted data Da as a reference S, and judges whether the difference (inter-data distance) between the predicted data Da and the measured physical property measurement data D exceeds the allowable range Ar. Note that the first method requires three or more physical property measurement data D to compare the physical property measurement data D, but the second method can be performed with only one physical property measurement data D.

[0034] The physical property measurement data D (D4, D5, D6) illustrated in Fig. 4 are acquired similarly to the physical property measurement data D1, D2, D3 illustrated in Fig. 3. In Fig. 4, the solid line indicates the physical property measurement data D4, the dashed line indicates the physical property measurement data D5, and the dashed dotted line indicates the physical property measurement data D6.

[0035] 5 and 6 show a part of the physical property measurement data D4 and D5. Since the physical property measurement data D4 and D5 are point cloud data, in FIG. 5, each circle drawn with a thin solid line indicates the physical property measurement data D4, and in FIG. 6, each circle drawn with a thin dashed line indicates the physical property measurement data D5. Also, in FIG. 5, each circle drawn with a thick solid line indicates the predicted data Da4, and the approximation line of the predicted data Da4 is drawn with a thick solid line. In FIG. 6, each circle drawn with a thick dashed line indicates the predicted data Da5, and the approximation line of the predicted data Da5 is drawn with a thick dashed line.

[0036] In this method, first, the interval area and the initial position are determined. The interval area and the initial position are determined by outputting an input screen of the interval area and the initial position to the output unit 7 of the calculation device 2, and inputting the interval area and the initial position to the input screen by the input unit 6. The interval area is determined by selecting any n consecutive points from the points of the data. In Figs. 5 and 6, three consecutive points are selected as the interval area. The initial position indicates the position at which prediction of the predicted data Da4 and Da5 starts, and is determined by selecting any strain position. In Figs. 5 and 6, the strain position a is selected as the initial position. In a range where the strain (stress) is very small, the data is often unstable (changing irregularly), so a position where the data is stable at a glance is selected as the initial position.

[0037] Next, the predicted data Da4, Da5 are predicted. The calculation device 2 executes data processing to predict the predicted data Da4, Da5 by performing data fitting using the n points extracted by determining the section region and the initial position. The data fitting can be a known straight line fitting or curve fitting. The straight line or curve obtained by the data fitting is the predicted data Da4, Da5. The predicted data Da4, Da5 are data predicted on the assumption that there is no abnormality due to the measurement in the physical property measurement data D4, D5, and can be regarded as data that quantifies the previously grasped standard S.

[0038] Next, the prediction data Da4, Da5 are compared with the corresponding physical property measurement data D4, D5. The calculation device 2 executes data processing to calculate the difference between the prediction data Da4, Da5 and the physical property measurement data D4, D5 up to m points ahead from the end point of the section area (the point m points ahead from the end point toward the left in Figs. 5 and 6). Next, the calculation device 2 executes data processing to obtain the degree of anomaly based on the calculated difference. m can be set arbitrarily. The degree of anomaly is positively correlated with the difference, and indicates the degree of difference between the prediction data Da4, Da5 and the physical property measurement data D4, D5. The higher the degree of anomaly, the larger the difference between the prediction data Da4, Da5 and the physical property measurement data D4, D5, and the lower the degree of anomaly, the smaller the difference between the prediction data Da4, Da5 and the physical property measurement data D4, D5. For the degree of anomaly, a representative value of the distance, such as the kNN distance between a group of n points and a group of m points, can also be adopted.

[0039] Next, the section area is moved one point ahead. The calculation device 2 executes data processing to move the current section area one point ahead. The number of points can be arbitrarily set within a range in which data fitting is performed continuously without interruption, for example, between n and m.

[0040] The above process of acquiring the degree of abnormality by data fitting is repeated. Next, the position where the acquired degree of abnormality exceeds the allowable range Ar is regarded as a change point. The calculation device 2 executes data processing to determine whether the acquired degree of abnormality is within the allowable range Ar, and determines whether a change point exists. Therefore, in this method, the determination of whether a change point exists in the physical property measurement data D is performed as a determination of whether the physical property measurement data D exceeds the allowable range with respect to the predicted data Da, that is, a determination of whether the physical property measurement data D exceeds the allowable range Ar with respect to the standard S. This method may also compare a representative value of the degree of abnormality with the allowable range Ar to determine whether the physical property measurement data D exceeds the allowable range Ar with respect to the standard S.

[0041] 5, the predicted data Da4 and the physical property measurement data D4 do not deviate significantly, and the acquired abnormality degrees are all within the allowable range Ar. As a result, since there is no change point in the physical property measurement data D4, it is determined that there is no abnormality due to measurement in the physical property measurement data D4.

[0042] In FIG. 6, there is a point where the predicted data Da5 and the physical property measurement data D5 significantly deviate from each other, and the degree of anomaly obtained at that point exceeds the allowable range Ar. This point is the change point. Therefore, since the physical property measurement data D5 has a change point, it is determined that the physical property measurement data D5 has an anomaly caused by the measurement. Similarly, since the physical property measurement data D6 also has a change point, it is determined that the physical property measurement data D6 also has an anomaly caused by the measurement.

[0043] This method can identify the time point when the abnormality occurred in addition to the presence or absence of an abnormality caused by the measurement of the physical property measurement data D. For example, it can identify that an abnormality occurred during measurement at the change point of the physical property measurement data D5 in FIG.

[0044] When setting the standard S based on the physical property measurement data D of the previously obtained conforming product sample 13A when no abnormalities are found, specifically, the physical property measurement data to be the standard S is calculated and used, for example, by simply averaging the physical property measurement data Dg of vulcanized rubber of the same specifications as the conforming product sample 13A when no abnormalities are found that have been accumulated up to now, and this data is then used.

[0045] Alternatively, previously acquired physical property measurement data Dg of the conforming product sample 13A when there is no abnormality and previously acquired physical property measurement data Db of the conforming product sample 13A when there is an abnormality are machine-learned as learning data to create an anomaly determination model M that determines whether or not there is an abnormality in the physical property measurement data D of the conforming product sample 13A. By inputting the physical property measurement data D of the conforming product sample 13A into this created anomaly determination model M, it is also possible to determine whether or not there is an abnormality in the physical property measurement data D of the conforming product sample 13A.

[0046] In more detail, an abnormality determination model M (classification model M1 or prediction model M2) constructed by machine learning using a data set 15 prepared in a preparation stage before step (S110) is used. The abnormality determination model M is stored in the auxiliary storage unit 5 of the calculation device 2. The data set 15 used in the classification model M1 is an accumulation of physical property measurement data D (Dg, Db) of a large number of conforming product samples 13A of the same vulcanized rubber previously acquired by the physical property measurement device 8. Each physical property measurement data D of the data set 15 is classified as physical property measurement data Dg when there is no abnormality due to the measurement, and is classified as physical property measurement data Db when there is an abnormality due to the measurement. The classification model M1 is a type of computer program that classifies the input physical property measurement data D into data when there is no abnormality due to the measurement or data when there is an abnormality due to the measurement. The classification model M1 is constructed by supervised machine learning using the physical property measurement data D of each conforming product sample in the data set 15 as an explanatory variable and the classification of normal data or abnormal data as a target variable. Examples of machine learning for constructing the classification model M1 and the prediction model M2 include various well-known supervised machine learning methods such as multiple regression analysis, random forest, K-nearest neighbor algorithm, and deep neural network (DNN).

[0047] The method using the prediction model M2 inputs the physical property measurement data D of the conforming product sample 13A to be judged, predicts and calculates predicted data Da on the assumption that the physical property measurement data D does not have anomalies due to measurement, and sets the predicted data Da as the standard S. Then, it is judged whether the predicted data Da exceeds the allowable range Ar. As with the second method described above, the judgment of whether the predicted data Da exceeds the allowable range Ar is made by judging whether the difference between the predicted data Da and the measured physical property measurement data D exceeds the allowable range Ar.

[0048] The data processing in step (S120) is not limited to one of the above-mentioned methods, and multiple methods can be used. For example, both the first method and the second method described above may be performed. In the first method described above, the representative values ​​of the data distances of the physical property measurement data D4, D5, and D6 shown in FIG. 4 are approximately equal, and all of the data D4, D5, and D6 are within the allowable range Ar, so it is determined that there is no abnormality due to the measurement. On the other hand, in the second method, there is no change point in any of the physical property measurement data D1, D2, and D3 shown in FIG. 3, so it is determined that there is no abnormality due to the measurement. Therefore, by performing both the first method and the second method, it is possible to detect various abnormalities due to the measurement without overlooking them, which is advantageous for improving the reliability of the physical property test.

[0049] In step (S130), the arithmetic device 2 executes data processing to output the judgment result of the above-mentioned step (S120) to the output unit 7. Specifically, the arithmetic device 2 executes data processing to output the presence or absence of an abnormality in the measured physical property measurement data D to the output unit 7 as the judgment result.

[0050] As described above, according to the various embodiments described above, the conforming product sample 13A is used as the test sample 13 to obtain its physical property measurement data D, so that a test sample 13 that is not conforming to the measurement is not used. That is, when determining whether or not there is an abnormality in the physical property measurement data D, an abnormality caused by the individual test sample 13 (13A) used is eliminated. In this way, it is possible to assume that there is no abnormality caused by the individual test sample 13 (13A) used, so that it is possible to grasp whether or not there is an abnormality caused by the measurement of the physical property measurement data D of the test sample 13 (13A). As a result, this is advantageous in ensuring the reliability of the physical property test by this physical property measuring device 8.

[0051] Then, based on a comparison between the standard S set for data on predetermined physical properties of vulcanized rubber and the physical property measurement data D of the conforming product sample 13A, the presence or absence of an abnormality in the physical property measurement data D of the conforming product sample 13A is judged, so that the physical property measurement data D can be applied even if it is physical property measurement data of vulcanized rubber obtained by various physical property tests. And, there is no need to make special modifications to the physical property measurement device 8, and it is sufficient to use the physical property measurement data D of the conforming product sample 13A obtained using the existing physical property measurement device 8. Therefore, each of the above-mentioned embodiments can easily detect an abnormality caused by the measurement of the physical property measurement data D of the vulcanized rubber test sample 13 (13A), and has high versatility.

[0052] Even if the physical property measuring devices 8 obtain the same predetermined physical properties of the vulcanized rubber, there are individual differences in the measurements of each physical property measuring device 8. Therefore, it is possible to specify the physical property measuring devices 8 that obtain the physical property measurement data D and the data of the set standard S, and classify each data for each physical property measuring device 8. By matching the physical property measuring device 8 that obtains the physical property measurement data D with the physical property measuring device 8 that obtains the data of the set standard S, it is advantageous to accurately detect the presence or absence of anomalies due to the measurement of the physical property measurement data D. Also, by classifying the data for each physical property measuring device 8 in this way, the individual differences in the measurements of each physical property measuring device 8 can be clearly understood.

[0053] The present invention is not limited to a specific embodiment, and various modifications and variations are possible within the scope of the gist of the present invention. [Explanation of symbols]

[0054] 1 Anomaly Detection System 2 Arithmetic unit 3. Processing Unit 4 Main memory 5 Auxiliary storage 6 Input section 7 Output section 8. Physical property measuring equipment 9 Gripping part 10. Extension detector 11 Force detection unit 12 Arithmetic section 13. Test Samples 13A compliant sample 14 Gauge points 15 Dataset D Physical property measurement data S standard

Claims

1. 1. A method for detecting anomalies in physical property measurement data of a vulcanized rubber, the method comprising the steps of: detecting anomalies in physical property measurement data of a test sample of vulcanized rubber, the physical property measurement data being continuously acquired for a predetermined period of time by a physical property measuring device during a single measurement of the vulcanized rubber; Setting standards for data of the predetermined physical properties of the vulcanized rubber; A method for detecting abnormalities in physical property measurement data of vulcanized rubber, comprising the steps of: acquiring the physical property measurement data using a conforming product sample having quality suitable for measuring the specified physical properties of the vulcanized rubber as the test sample; and determining whether or not there is an abnormality in the physical property measurement data of the conforming product sample based on a comparison between the acquired physical property measurement data of the conforming product sample and the standard.

2. 2. The method for detecting an anomaly in physical property measurement data of vulcanized rubber according to claim 1, wherein the standard is set based on the physical property measurement data of the conforming product sample.

3. 3. The method for detecting an anomaly in physical property measurement data of vulcanized rubber according to claim 2, wherein the criterion is set based on the physical property measurement data of three or more of the conforming product samples.

4. 2. The method for detecting anomalies in physical property measurement data of vulcanized rubber according to claim 1, wherein the standard is set based on the physical property measurement data of the conforming product sample previously obtained when no anomalies were found.

5. A method for detecting abnormalities in physical property measurement data of vulcanized rubber as described in claim 4, wherein an abnormality judgment model is created to determine whether or not there is an abnormality in the physical property measurement data of the conforming sample by machine learning using as learning data the physical property measurement data of the conforming sample obtained in the past when there is no abnormality and the physical property measurement data of the conforming sample obtained in the past when there is an abnormality, and the presence or absence of an abnormality in the physical property measurement data of the conforming sample is judged by inputting the physical property measurement data of the conforming sample into this created abnormality judgment model.

6. 1. A system for detecting anomalies in physical property measurement data of vulcanized rubber, the system detecting anomalies in physical property measurement data of a test sample of vulcanized rubber, the physical property measurement data being continuously acquired for a predetermined period of time by a physical property measuring device during a single measurement of the vulcanized rubber, The system for detecting anomalies in physical property measurement data of vulcanized rubber comprises: a calculation device into which standards for the specified physical properties of the vulcanized rubber and an acceptable range for these standards are input; and the physical property measurement data obtained by the physical property measuring device using a conforming product sample having quality suitable for measurement of the specified physical properties of the vulcanized rubber as the test sample is input; the calculation device compares the input physical property measurement data of the conforming product sample with the standard value to determine whether the difference between the two exceeds the acceptable range; and an output unit that outputs the result of the determination made by the calculation device.

Images