Method for predicting occurrence of tire vulcanization failure and tire manufacturing method

Abstract

To provide a prediction method capable of more highly accurately predicting the possibility of a vulcanization failure occurrence by using specification data of a green tire and sectional shape data of a molded surface of a used mold, and a tire manufacturing method using this prediction method.SOLUTION: A prediction method stores, in an arithmetic unit 2, a prediction model PM generated by machine-learning tire specification data D1 including a sectional shape of a green tire G, sectional shape data D2 of a molded surface 7 of a mold 6, vulcanization failure occurrence data D3 in a tire T manufactured by vulcanizing multiple green tires G different in tire specification data D1 by using the mold 6 as teacher data, and predicts the occurrence presence of a vulcanization failure in a target tire Ta manufactured by vulcanizing a green tire Ga by the mold 6 by assigning the tire specification data D1 of the green tire Ga to be predicted and the sectional shape data D2 of the molded surface 7 to the prediction model PM to perform an arithmetic operation.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present invention relates to a method for predicting the occurrence of vulcanization failures in tires and a tire manufacturing method, and more specifically, to a prediction method that can predict the possibility of vulcanization failures occurring with higher accuracy by using specification data of a green tire and cross-sectional shape data of the molding surface of a mold to be used, and a tire manufacturing method utilizing this prediction method. [Background technology]

[0002] Tires are manufactured by vulcanizing unvulcanized green tires in a vulcanization mold. Green tires are molded to preset specifications, but depending on the tire specifications, they may not match well with the mold used, causing vulcanization defects in the vulcanized tire. For example, vulcanized tires may have vulcanization defects such as localized excess or deficiency of rubber, excess or deficiency of vulcanization, and deformation.

[0003] In order to prevent such failures, it has been proposed to use the size of the gap between the predicted cross-sectional shape of a green tire and the molding surface of the mold used, calculated using cross-sectional shape data, as an index of vulcanization failures (see Patent Document 1). In the invention proposed in Patent Document 1, a large number of coordinate points are extracted at equal intervals in the tire radial direction on the outer surface of the tire of the predicted cross-sectional shape of the green tire, and the size of the gap between the outer surface of the tire and the molding surface at each coordinate point is calculated. A correlation coefficient between the size of this gap and the failure rate is calculated, and it is determined that failures are likely to occur in areas (positions) where the absolute value of this correlation coefficient is large.

[0004] In this invention, the parts (positions) where failures are likely to occur are identified, but the clear appropriate range of the size of the above-mentioned gap that can prevent the occurrence of failures is not identified. Therefore, when the green tire to be judged is vulcanized using a specified mold, it is difficult to accurately predict whether or not a vulcanization failure will occur, and there is room for improvement in predicting the possibility of the occurrence of a vulcanization failure with higher accuracy. [Prior art documents] [Patent documents]

[0005] [Patent Document 1] JP 2010-18010 A Summary of the Invention [Problem to be solved by the invention]

[0006] An object of the present invention is to provide a prediction method capable of predicting with higher accuracy the possibility of vulcanization failure occurring by using specification data of a green tire and cross-sectional shape data of the molding surface of a mold to be used, and a tire manufacturing method utilizing this prediction method. [Means for solving the problem]

[0007] In order to achieve the above object, the method for predicting the occurrence of curing defects in tires of the present invention uses as training data tire specification data including the cross-sectional shape of a green tire, cross-sectional shape data of the molding surface of a mold for vulcanizing the green tire, and occurrence data of curing defects that have occurred in tires manufactured by vulcanizing a large number of the green tires having different tire specification data using the mold, and generates a prediction model for predicting whether or not a curing defect will occur in a tire manufactured by vulcanization by machine learning using the training data, inputs the tire specification data of the green tire to be predicted into a calculation device in which this prediction model is stored, and predicts whether or not a curing defect will occur in a target tire manufactured by vulcanizing the green tire to be predicted using the mold by calculation processing using the prediction model based on this input tire specification data and the cross-sectional shape data of the molding surface. The tire specification data includes viscosity data of unvulcanized rubber that contacts the molding surface of the green tire. It is characterized by: Another tire vulcanization fault occurrence prediction method of the present invention uses, as training data, tire specification data including a cross-sectional shape of a green tire, cross-sectional shape data of a molding surface of a mold for vulcanizing the green tire, and occurrence data of vulcanization faults that have occurred in tires manufactured by vulcanizing a large number of the green tires having different tire specification data using the mold, to generate a prediction model that predicts whether or not a vulcanization fault will occur in a tire manufactured by vulcanization by performing machine learning using the training data, inputting tire specification data of the green tire to be predicted into a calculation device in which this prediction model is stored, and predicting whether or not a vulcanization fault will occur in a target tire manufactured by vulcanizing the green tire to be predicted using the mold through calculation processing by the calculation device using the prediction model based on the input tire specification data and the cross-sectional shape data of the molding surface, adding vulcanization condition data including at least the vulcanization temperature when vulcanizing the green tire as the training data, inputting the vulcanization condition data of the green tire to be predicted into the calculation device, and using the input vulcanization condition data also for predicting whether or not the vulcanization fault will occur in the target tire. Yet another tire vulcanization fault occurrence prediction method of the present invention uses tire specification data including the cross-sectional shape of a green tire, cross-sectional shape data of the molding surface of a mold for vulcanizing the green tire, and occurrence data of vulcanization faults that have occurred in tires manufactured by vulcanizing a large number of the green tires having different tire specification data using the mold as training data, thereby generating a prediction model that predicts whether or not a vulcanization fault will occur in a tire manufactured by vulcanization, and generates a prediction model that predicts whether or not a vulcanization fault will occur in a tire manufactured by vulcanization by inputting tire specification data of the green tire to be predicted into a calculation device in which this prediction model is stored, and predicts whether or not a vulcanization fault will occur in a target tire manufactured by vulcanizing the green tire to be predicted using the mold by calculation processing using the prediction model based on this input tire specification data and the cross-sectional shape data of the molding surface, and includes in the tire specification data the volume of unvulcanized rubber outside the carcass layer of the green tire.

[0008] The tire manufacturing method of the present invention is characterized in that, based on the results predicted by the tire vulcanization fault prediction method described above, the tire specification data of the green tire in which the vulcanization fault will not occur in the target tire is determined, a green tire corresponding to the determined tire specification data is molded, and the molded green tire is vulcanized using a mold having the same specifications as the mold.

[0009] According to the tire curing fault occurrence prediction method of the present invention, a prediction model is used that is generated by machine learning the tire specification data, the cross-sectional shape data of the molding surface, and the occurrence data of the curing fault as the teaching data. Therefore, by inputting the tire specification data of the green tire to be predicted and the cross-sectional shape data of the molding surface to the prediction model and performing calculation processing by the calculation device, it becomes possible to predict with high accuracy the occurrence or non-occurrence of a curing fault in a target tire manufactured by vulcanizing the green tire to be predicted using the mold.

[0010] According to the tire manufacturing method of the present invention, a green tire corresponding to the tire specification data of the green tire free of the vulcanization failure is molded based on the results predicted by the tire vulcanization failure prediction method, and the molded green tire is vulcanized using a mold of the same specifications as the mold. Therefore, it is possible to stably and efficiently manufacture tires free of vulcanization failure. [Brief description of the drawings]

[0011] [Figure 1] FIG. 1 is an explanatory diagram illustrating a prediction system used in the present invention. [Diagram 2] FIG. 2 is an explanatory diagram illustrating a right half of a green tire in cross section. [Diagram 3] FIG. 13 is an explanatory diagram illustrating the right half of the mold in cross section. [Figure 4] FIG. 2 is an explanatory diagram illustrating a cross-sectional view of the right half of a green tire in a state in which the tire is placed in a mold. [Diagram 5]FIG. 2 is an explanatory diagram illustrating an example of an appropriate range of the outer surface (profile) of the cross-sectional shape of a green tire predicted using a prediction model, together with the cross-sectional shape of the molding surface. [Figure 6] FIG. 6 is an explanatory diagram illustrating the appropriate range in FIG. 5 together with the profile of a green tire to be predicted. [Figure 7] FIG. 2 is an explanatory diagram illustrating a profile of a green tire formed based on the prediction results using a prediction model, together with the cross-sectional shape and appropriate range of the molding surface. [Figure 8] 8 is an explanatory diagram illustrating, in cross section, a vulcanizing device that vulcanizes the green tire of FIG. 7.

[0023] FIG. [Figure 9] 9 is an explanatory diagram illustrating a right half of the green tire in FIG. 8 in cross section. FIG. [Figure 10] FIG. 2 is an explanatory diagram illustrating a right half of a vulcanized tire in cross section. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0012] DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, a method for predicting the occurrence of a vulcanization fault in a tire and a method for manufacturing a tire according to the present invention will be specifically described with reference to the embodiments shown in the drawings.

[0013] The tire vulcanization fault occurrence prediction method of the present invention is carried out using a prediction system 1 exemplified in Fig. 1. This prediction system 1 has a calculation device 2, an input unit 3, a display unit 4, and further a measurement device 5. The calculation device 2, the input unit 3, and the display unit 4 are connected to each other so as to be able to communicate with each other by wire or wirelessly. Measurement data acquired by the measurement device 5 is input to the calculation device 2 directly or through the input unit 3.

[0014] Various data are input and stored in the arithmetic unit 2, and the arithmetic unit 2 performs arithmetic processing using these data. Various computers can be used as the arithmetic unit 2. Therefore, the arithmetic unit 2 has a memory in which various data are stored, and a CPU that performs arithmetic processing.

[0015] The input unit 3 is an input means for inputting various data to the arithmetic unit 2. As the input unit 3, a keyboard, a mouse, or various terminal devices can be used.

[0016] The display unit 4 is a display means for displaying various data input to the arithmetic device 2 and the results of arithmetic processing using these data (numerical values, tables, drawings, etc.). As the display unit 4, various monitors can be used.

[0017] The measuring device 5 is a means for measuring and acquiring necessary data. For example, tire specification data D1 including the cross-sectional shape of the green tire G and the like are acquired using the measuring device 5. Cross-sectional shape data D2 of the molding surface 7 of the mold 6 for vulcanizing the green tire G can also be acquired using the measuring device 5. In this invention, at least the cross-sectional shape of the green tire G is measured by the measuring device 5 to acquire the data.

[0018] As tire specification data D1, cross-sectional shape data of a green tire G as exemplified in FIG. 2 is used. In a typical green tire G, an inner liner layer m1 is disposed on the innermost circumference, a carcass layer m2 is disposed on its outer surface, and unvulcanized rubber R is disposed on the outer surface of the carcass layer m2. A belt layer m4 and the like are embedded in the unvulcanized rubber R of the tread portion. Since the cross-sectional shape of the green tire G does not generally change in the tire circumferential direction, cross-sectional shape data at any circumferential position of the green tire G can be used. Note that, although the right half of the green tire, tire, and mold are exemplified in the drawing, the left half has the same specifications as the right half.

[0019] Other examples of the tire specification data D1 include various characteristic data of each component of the green tire G. Specifically, examples include volume data (cross-sectional area data) of the unvulcanized rubber R that forms the green tire G, viscosity data of this unvulcanized rubber R, cross-sectional shape data (and arrangement data) and rigidity data of the carcass layer m2, and cross-sectional shape data (and arrangement data) and rigidity data of the bead portion m3. The tire specification data D1 to be adopted may be only one type of cross-sectional shape data of the green tire G, but one or more other types of tire specification data D1 may also be added. The two-dot chain line CL in the figure indicates the center in the tire width direction.

[0020] Specifically, the cross-sectional shape data D2 of the molding surface 7 of the mold 6 is cross-sectional shape data of the molding surface 7 as shown in Fig. 3. The mold 6 shown in Fig. 3 is a sectional type, and is therefore composed of a side mold 6a and a sector mold 6b. The side mold 6a mainly vulcanizes and molds the outside of the side portion and the outside of the bead portion m3 of the green tire G, and the sector mold 6b mainly vulcanizes and molds the tread portion.

[0021] Since the cross-sectional shape of the molding surface 7 generally changes in the tire circumferential direction, cross-sectional shape data of the molding surface 7 at a preset position in the tire circumferential direction is used. For example, cross-sectional shape data D2 of a representative molding surface 7 including the most characteristic grooves, or cross-sectional shape data D2 of a molding surface 7 from which grooves are omitted, can be used. If there is a portion that is known to be prone to vulcanization failure, cross-sectional shape data D2 of the molding surface 7 including the portion where that vulcanization failure is likely to occur (the portion where it is most likely to occur) can also be used.

[0022] Alternatively, it is possible to exclude the portions (sites) of the molding surface 7 whose cross-sectional shape changes in the tire circumferential direction from the prediction target, and to predict only the portions (at least one of the outer side of the side portion and the outer side of the bead portion) whose cross-sectional shape does not substantially change in the tire circumferential direction. In other words, the presence or absence of vulcanization failure may be predicted only for the outer portion of the side portion, only for the outer portion of the bead portion m3, or only for the outer portion of the side portion and the outer portion of the bead portion m3. In this case, it is sufficient to use the tire specification data D1 of the portion corresponding to the prediction target and the cross-sectional shape data D2 of the molding surface 7.

[0023] Examples of the measuring device 5 that measures and obtains the cross-sectional shape data of the green tire G include a known profile sensor, an image capturing device such as a digital camera, and an X-ray imaging device (CT scanner). When a profile sensor or an X-ray imaging device (CT scanner) is used, the cross-sectional shape data can be measured and obtained without cutting the green tire G into a cross section.

[0024] A known viscosity measuring device such as a Mooney viscometer can be exemplified as the measuring device 5 for measuring and acquiring the viscosity data of the unvulcanized rubber R. This viscosity data may be a measured value at room temperature, but considering that the unvulcanized rubber R is heated in the vulcanization process and its fluidity is improved, a measured value at a temperature at which the fluidity is highest can also be adopted. A known image taking device such as a digital camera and an X-ray taking device can be exemplified as the measuring device 5 for measuring and acquiring the volume data (cross-sectional area data) of the unvulcanized rubber R, the cross-sectional shape data (and arrangement data) of the carcass layer m2, and the cross-sectional shape data (and arrangement data) of the bead portion m3. A known bending tester can be exemplified as the measuring device 5 for measuring and acquiring the stiffness data of the carcass layer m2 and the bead portion m3.

[0025] The cross-sectional shape data D2 of the molding surface 7 can be obtained by measuring the mold 6 using a measuring device 5 such as a known profile sensor, an image capture device such as a digital camera, etc. However, since design data (CAD data, etc.) used in the manufacture of the mold 6 generally exists, data corresponding to the cross-sectional shape of the molding surface 7 can be extracted from this design data and used as the cross-sectional shape data of the molding surface 7 without measurement.

[0026] In the present invention, a prediction model PM generated by machine learning is used. This prediction model PM is a computer program for predicting the possibility of occurrence of a vulcanization failure (presence or absence of occurrence) in a tire T manufactured by vulcanizing a green tire G. An example of a procedure for generating this prediction model PM will be described.

[0027] First, a large number of green tires G having different tire specification data D1 are vulcanized using a predetermined mold 6 to manufacture a tire T as shown in Fig. 10. Then, the presence or absence of a vulcanization fault and the location of the vulcanization fault (location data) are confirmed for the manufactured tire T. Note that the predetermined mold 6 is not limited to one specific mold 6, and may be any mold 6 having the same specifications as the predetermined mold 6.

[0028] As illustrated in Fig. 4, a green tire G to be vulcanized is placed in a mold 6, and then the mold 6 is closed and pressure and heat are applied between a molding surface 7 of the mold 6 and an inflated vulcanization bladder 9. In Fig. 4, the green tire G in its initial shape (before being placed inside the mold 6) is depicted, and the cross-sectional shapes of the molding surface 7 and the vulcanization bladder 9 are depicted by dashed lines.

[0029] The presence or absence of a vulcanization failure and the location of the occurrence thereof confirmed in the manufactured tire T are adopted as vulcanization failure occurrence data D3. The vulcanization failure occurrence data D3 includes data in cases where a vulcanization failure has occurred and data in cases where a vulcanization failure has not occurred.

[0030] Regardless of the type or degree of the vulcanization failure, the data can be used as the vulcanization failure occurrence data D3 by distinguishing between cases where a vulcanization failure has occurred and cases where a vulcanization failure has not occurred. Alternatively, the type of vulcanization failure (excess or shortage of rubber, excess or shortage of vulcanization, deformation, etc.) and the degree of the vulcanization failure (small, medium, large) can be classified and confirmed. Then, a combination of these classifications can be used as the vulcanization failure occurrence data D3.

[0031] The vulcanization fault to be predicted can also be specified to one or more desired types of faults. In this case, the occurrence data D3 of the vulcanization fault is adopted as data indicating the occurrence of a vulcanization fault only when the specified type of vulcanization fault occurs, and is adopted as data indicating the absence of a vulcanization fault in other cases. Specifically, when the type of vulcanization fault to be predicted is specified to one type, rubber excess or shortage, if rubber excess or shortage occurs in the tire T, it is adopted as data indicating the occurrence of a vulcanization fault, and if there is no rubber excess or shortage even if there is over or under vulcanization, it is adopted as data indicating the absence of a vulcanization fault.

[0032] In this manner, data D3 on the occurrence of vulcanization failures is obtained. In addition, tire specification data D1 including a cross-sectional shape of a green tire G before the manufactured tire T is vulcanized, and data D2 on a cross-sectional shape of a molding surface 7 of a mold 6 in which the green tire G is vulcanized are prepared.

[0033] The prediction model PM is generated by machine learning the above-mentioned tire specification data D1, cross-sectional shape data D2 of the molding surface 7, and vulcanization fault occurrence data D3 as basic data (i.e., teacher data) for predicting the occurrence or non-occurrence of vulcanization faults. Specifically, the above-mentioned data D1, D2, and D3 are input to the calculation device 2, and machine learning is performed using these data D1, D2, and D3. That is, the presence or non-occurrence of vulcanization faults is linked to the relationship between the tire specifications of the green tire G and the cross-sectional shape of the molding surface 7, and the influence of the tire specifications of the green tire G and the cross-sectional shape of the molding surface 7 on the occurrence of vulcanization faults is analyzed and evaluated.

[0034] The incidence of vulcanization failures in the tire T is greatly affected by the size of the interval (gap) between the outer surface of the green tire G and the molding surface 7 of the mold 6, and the volume of the unvulcanized rubber R between the carcass layer m2 and the molding surface 7. Since the unvulcanized rubber R flows during vulcanization, the viscosity (fluidity) of the unvulcanized rubber R between the carcass layer m2 and the molding surface 7 also affects the incidence of vulcanization failures.

[0035] If the distance between the outer surface of the green tire G and the molding surface 7 of the mold 6 is too large, or if the volume of the unvulcanized rubber R between the carcass layer m2 and the molding surface 7 is too small, rubber defects are likely to occur, which may result in deformation of the tire. The portion with the rubber defects may become insufficiently vulcanized relative to the unvulcanized rubber R due to insufficient pressure or heat. If the distance between the outer surface of the green tire G and the molding surface 7 of the mold 6 is too small, or if the volume of the unvulcanized rubber R between the carcass layer m2 and the molding surface 7 is too large, excess rubber and tire deformation are likely to occur. The portion with the excess rubber may become over-vulcanized relative to the unvulcanized rubber R due to excessive pressure.

[0036] If the viscosity of the unvulcanized rubber R between the carcass layer m2 and the molding surface 7 is high, the flowability is low, and therefore deformation of the green tire G is unlikely to occur. If the viscosity is low, the flowability is high, and therefore deformation of the green tire G is likely to occur.

[0037] In addition, the cross-sectional shape (and arrangement) and rigidity of the carcass layer m2 affect the fluidity of the unvulcanized rubber R between the carcass layer m2 and the molding surface 7, and therefore affect the incidence of vulcanization failures. The cross-sectional shape (and arrangement) and rigidity of the bead portion m3 also affect the fluidity of the unvulcanized rubber R between the carcass layer m2 and the molding surface 7, and therefore affect the incidence of vulcanization failures. The above-mentioned factors each affect each other in a complex way, and thus the incidence of vulcanization failures changes.

[0038] Therefore, by using the data D1, D2, and D3 as teacher data, it becomes possible to generate a prediction model PM by having an artificial intelligence (AI) learn the differences and the degree of difference between the combinations of the data D1 and D2 when a vulcanization failure occurs and when a vulcanization failure does not occur, and the characteristics of each case. As a machine learning method, various well-known methods such as deep learning using a neural network can be exemplified. In deep learning, a network is used in which a multi-layer structure is formed by an input layer, a plurality of intermediate layers, and an output layer by a well-known method, and weights are set between nodes between each layer and linked. Then, tire specification data D1 and cross-sectional shape data D2 that affect the vulcanization failure occurrence data D3 are input from the input layer, and a predicted value of the vulcanization failure occurrence data D3 is output to the output layer. The calculated predicted value is compared with the actual measured value of the vulcanization failure occurrence data D3, and the weights of each are changed so as to reduce the error between the two. As a result, a prediction model PM with improved prediction accuracy is generated (constructed). The generated prediction model PM is stored in the calculation device 2.

[0039] Next, an example of a procedure for predicting whether or not a vulcanization failure will occur in a target tire Ta manufactured by vulcanizing a green tire Ga that is a prediction target in a predetermined mold 6 will be described.

[0040] The calculation device 2 receives input of tire specification data D1 of the green tire Ga to be predicted using the input unit 3. The calculation device 2 substitutes the input tire specification data D1 of the green tire Ga and cross-sectional shape data D2 of the molding surface 7 of the mold 6 into the prediction model PM, and performs a calculation process to execute the prediction model PM. By this calculation process, when the green tire Ga to be predicted is vulcanized using the mold 6 to manufacture a target tire Ta, it is determined whether or not a vulcanization failure will occur in the target tire Ta, and the prediction result is displayed on the display unit 4.

[0041] This prediction is based on the premise that the bead cores of the pair of bead portions m2 of the green tire Ga are fixed to the correct positions on the molding surface 7 and vulcanized. Therefore, the vulcanization failure occurrence data D3 also uses data obtained when the bead cores of the pair of bead portions m2 are fixed to the correct positions on the molding surface 7 and vulcanized.

[0042] The prediction results using the prediction model PM are displayed as a range between dashed lines for the appropriate range AR of the outer surface (profile) of the cross-sectional shape of the green tire Ga, as shown in Fig. 5 and Fig. 6. In Fig. 5, this appropriate range AR is shown together with the cross-sectional shape of the molding surface 7, and in Fig. 6, this appropriate range AR is shown together with the outer surface of the cross-sectional shape of the green tire Ga. The appropriate range AR is exaggerated to make it easier to see on the drawings.

[0043] Any range outside this appropriate range AR becomes an inappropriate range for the outer surface of the cross-sectional shape of the green tire Ga. Therefore, in the prediction result, if the entire outer surface of the cross-sectional shape of the green tire Ga falls within the appropriate range AR, it is predicted that no vulcanization failure will occur in the target tire Ta. On the other hand, if there is a part on the outer surface of the cross-sectional shape of the green tire Ga that falls within the inappropriate range, it is predicted that a vulcanization failure will occur in the target tire Ta in that part. Therefore, it is possible to predict whether or not a vulcanization failure will occur in the target tire Ta (green tire Ga) and the distribution of the occurrence positions. Note that the prediction result is not limited to the display in Figures 5 and 6, and it is sufficient if it is possible to grasp whether or not a vulcanization failure will occur in the target tire Ta and the appropriate range AR.

[0044] The prediction results are displayed in a planar coordinate system of the tire cross section, so it is possible to know at a glance at which position of the green tire Ga a vulcanization failure will occur. Therefore, it is easy to know which position of the green tire Ga is too large or too small relative to the mold 6 (molding surface 7). Furthermore, since it is possible to know the distance to the outer surface of the cross-sectional shape of the green tire Ga relative to the appropriate range AR and the distance to the molding surface 7 in the cross section, it becomes clear to what extent the cross-sectional shape of the green tire Ga should be changed in order to prevent the occurrence of a vulcanization failure. Therefore, it is very useful for obtaining a cross-sectional shape of the green tire Ga that can prevent the occurrence of a vulcanization failure.

[0045] This prediction method uses a prediction model PM that is generated by machine learning the above-mentioned tire specification data D1, the cross-sectional shape data D2 of the molding surface 7, and the vulcanization fault occurrence data D3 as teacher data. Therefore, by inputting the tire specification data D1 of the green tire Ga to be predicted and the cross-sectional shape data D2 of the molding surface 7 into the prediction model PM and performing calculation processing, it is possible to predict vulcanization faults that occur due to the complex interactions of various factors with each other with even greater accuracy.

[0046] Moreover, the cross-sectional shape data of green tires G, Ga employed as tire specification data D1 is obtained by measuring an actual green tire G using a measuring device 5. Moreover, since the molding surface 7 is manufactured by precision machining, even if design drawing data (such as CAD data) used in manufacturing the mold 6 is employed as the cross-sectional shape data D2 of the molding surface 7, the data substantially matches the measurement data obtained by measuring the actual molding surface 7 using the measuring device 5. The employment of such tire specification data D1 and cross-sectional shape data D2 of the molding surface 7 is advantageous in improving prediction accuracy.

[0047] The tire specification data D1 may include the volume of the unvulcanized rubber R outside the carcass layer m2 of the green tires G and Ga. This is advantageous for improving prediction accuracy.

[0048] In addition, the tire specification data D1 may include viscosity data of the unvulcanized rubber R that contacts the molding surface 7 of the green tires G, Ga. That is, including the viscosity data of the unvulcanized rubber R outside the carcass layer m2 is advantageous for improving prediction accuracy.

[0049] Furthermore, including the various data described above as the tire specification data D1 is advantageous in improving prediction accuracy. However, the more types of tire specification data D1 are included, the greater the burden of calculation processing becomes, and the improvement in prediction accuracy may not be commensurate with the increased burden. Therefore, in order to make predictions efficiently and accurately, it is advisable to include at least one type of cross-sectional shape data of the green tire G, which has the greatest influence, and add one or more other types of tire specification data D1 as necessary.

[0050] By using data classifying the type of vulcanization failure (excess or insufficient rubber, over or under vulcanization, deformation, etc.) and the degree of vulcanization failure (small, medium, large), it becomes possible to make clear predictions for each type of vulcanization failure or combination of types, as well as a clear prediction of the degree of vulcanization failure.

[0051] As training data, vulcanization condition data D4 including at least the vulcanization temperature when vulcanizing the green tire G can also be added. The vulcanization temperature affects the fluidity of the unvulcanized rubber R, and therefore affects the incidence of vulcanization failures. Therefore, including the vulcanization temperature in the training data is advantageous for improving the accuracy of predicting whether or not a vulcanization failure will occur in the tire Ta. Another example of the vulcanization condition data D4 is the vulcanization pressure applied to the green tire G by the vulcanization bladder 9. The vulcanization pressure also affects the fluidity of the unvulcanized rubber R, and therefore affects the incidence of vulcanization failures.

[0052] When the vulcanization condition data D4 is included in the teacher data, the vulcanization condition data D4 of the green tire Ga to be predicted is input to the calculation device 2, and the input vulcanization condition data D4 is also used to predict whether or not a vulcanization failure occurs in the target tire Ta. If the vulcanization condition data D4 is included in the teacher data, the burden of the calculation processing increases, and the improvement in prediction accuracy may not be commensurate with the increased burden. Therefore, it is advisable to add the vulcanization condition data D4 to the teacher data as necessary.

[0053] If it is assumed that the vulcanization conditions when using this mold 6 do not change much and vulcanization is performed under substantially the same vulcanization conditions, there is no need to consider the vulcanization condition data D4. In such a case, the data D3 on the occurrence of vulcanization failures under the given vulcanization conditions that are the assumption is used.

[0054] The tire manufacturing method of the present invention utilizes the curing failure occurrence prediction method of the present invention. An example of the procedure of the tire manufacturing method will be described.

[0055] In this tire manufacturing method, tire specification data of a green tire Ga in which no vulcanization failure occurs in the manufactured target tire Ta is grasped based on the results predicted using the prediction model PM. Then, the green tire Ga corresponding to the grasped tire specification data is molded by a known method. For example, as shown in FIG. 7, the entire outer surface of the cross-sectional shape of the green tire Ga is made to have a cross-sectional shape that falls within an appropriate range AR in which no vulcanization failure is predicted to occur. In this way, the green tire Ga is molded by reflecting the prediction results using the prediction model PM.

[0056] Next, as illustrated in Fig. 8 and Fig. 9, the molded green tire Ga is placed in a mold 6 having the same specifications as a predetermined mold 6 installed in a known vulcanizing device 8. Then, the green tire Ga is vulcanized in the closed mold 6 by a known method, thereby producing a tire Ta illustrated in Fig. 10.

[0057] In this tire manufacturing method, a green tire Ga corresponding to tire specification data D1 predicted with high accuracy as not causing a vulcanization failure is molded and vulcanized using a mold 6 having the same specifications as a predetermined mold 6. Therefore, it becomes possible to stably and efficiently manufacture a tire Ta that does not cause a vulcanization failure.

[0058] In a test case for improving a vulcanization failure caused by an excess or deficiency of rubber on the outer side of a tire Ta for a general passenger vehicle, a tire Ta was manufactured by molding and vulcanizing a green tire Ga so that the outer surface of the cross-sectional shape of the green tire Ga was within the appropriate range AR based on the results of the method for predicting the occurrence of vulcanization failure of the present invention. As a result, it was confirmed that vulcanization failure in the tire Ta could be prevented. [Explanation of symbols]

[0059] 1. Prediction System 2 Arithmetic unit 3. Input section 4 Display section 5. Measuring Equipment 6 Mold 6a Side mold 6b Sector Mall 7 Molding surface 8. Vulcanizing Equipment 9 Vulcanizing bladder PM Prediction Model G, Ga Green Tire m1 Inner liner layer m2 carcass layer m3 bead part m4 belt layer R Unvulcanized rubber Rc Vulcanized Rubber T, Ta tires (vulcanized tires)

Claims

1. A method for predicting the occurrence of vulcanization defects in tires, comprising: using as training data tire specification data including the cross-sectional shape of a green tire, cross-sectional shape data of the molding surface of a mold for vulcanizing the green tire, and occurrence data of vulcanization defects that have occurred in tires manufactured by vulcanizing a large number of the green tires having different tire specification data using the mold, and performing machine learning using the training data to generate a prediction model that predicts whether or not vulcanization defects will occur in tires manufactured by vulcanization; inputting tire specification data of the green tire to be predicted into a calculation device in which this prediction model is stored; and predicting whether or not vulcanization defects will occur in a target tire manufactured by vulcanizing the green tire to be predicted using the mold through calculation processing by the calculation device using the prediction model based on the input tire specification data and the cross-sectional shape data of the molding surface; and including in the tire specification data viscosity data of unvulcanized rubber that comes into contact with the molding surface of the green tire.

2. a tire vulcanization failure occurrence prediction method, comprising: using as training data tire specification data including a cross-sectional shape of a green tire, cross-sectional shape data of a molding surface of a mold for vulcanizing the green tire, and occurrence data of vulcanization failures that have occurred in tires manufactured by vulcanizing a large number of green tires having different tire specification data using the mold, and performing machine learning using the training data to generate a prediction model that predicts whether or not a vulcanization failure will occur in a tire manufactured by vulcanization; inputting tire specification data of the green tire to be predicted into a calculation device in which this prediction model is stored; predicting whether or not a vulcanization failure will occur in a target tire manufactured by vulcanizing the green tire to be predicted using the mold through calculation processing using the prediction model based on the input tire specification data and the cross-sectional shape data of the molding surface; adding vulcanization condition data including at least a vulcanization temperature when vulcanizing the green tire as the training data; inputting the vulcanization condition data of the green tire to be predicted into the calculation device; and using the input vulcanization condition data for the green tire to be predicted to also use the input vulcanization condition data for predicting whether or not the vulcanization failure will occur in the target tire.

3. A method for predicting the occurrence of vulcanization defects in tires manufactured by vulcanizing a large number of green tires having different tire specification data using tire specification data including a cross-sectional shape of a green tire, cross-sectional shape data of a molding surface of a mold for vulcanizing the green tire, and occurrence data of vulcanization defects that have occurred in tires manufactured by vulcanizing a large number of green tires having different tire specification data using the mold as training data, thereby generating a prediction model for predicting whether or not vulcanization defects will occur in tires manufactured by vulcanization, the tire specification data of the green tire to be predicted is input to a calculation device in which this prediction model is stored, and based on this input tire specification data and the cross-sectional shape data of the molding surface, the calculation device performs calculation processing using the prediction model to predict whether or not vulcanization defects will occur in a target tire manufactured by vulcanizing the green tire to be predicted using the mold, the tire specification data including the volume of unvulcanized rubber outside the carcass layer of the green tire.

4. A method for predicting the occurrence of vulcanization failures in a tire as described in claim 2 or 3, wherein the tire specification data includes viscosity data of unvulcanized rubber in contact with the molding surface of the green tire.

5. A tire manufacturing method comprising the steps of: determining tire specification data of a green tire in which the vulcanization fault will not occur in the target tire based on a result predicted by a tire vulcanization fault prediction method according to any one of claims 1 to 4; molding a green tire corresponding to the determined tire specification data; and vulcanizing the molded green tire using a mold having the same specifications as the mold.

Images