A tire quality defect automatic early warning and analysis rectification system

Abstract

The application relates to the technical field of tire manufacturing, in particular to a tire quality defect automatic early warning and analysis rectification system. The system solves the technical problem that the accuracy of a technical scheme defect root cause tracing and the foresight of early warning cannot meet actual production requirements. The system comprises the following modules: a data acquisition module used for acquiring process parameters, equipment state data and finished product detection data of multiple processes contained in tire production; a risk quantification module used for determining an adjustment factor based on the process parameters, the equipment state data and the finished product detection data; a dynamic threshold generation module used for generating a confidence threshold of an association rule based on the adjustment factor; and a rule screening module used for comparing the confidence of the association rule with the confidence threshold, screening out the association rule with a confidence not lower than the confidence threshold as a quality early warning rule. The application is used for a quality early warning scene of tire manufacturing.

Description

Technical Field

[0001] This invention relates to the field of tire manufacturing technology, and more specifically to an automatic early warning, analysis and rectification system for tire quality defects. Background Technology

[0002] Tire manufacturing is a typical process industry. Its production process encompasses multiple coupled physical and chemical changes, including material preparation, component molding, vulcanization, and finished product inspection. The process parameters of each process are strongly correlated and sequentially transferable. Even a small deviation in any process can be propagated along the production chain and amplified in subsequent processes, ultimately affecting the quality and safety performance of the finished tire. Currently, the industry's commonly used quality control methods rely primarily on manual inspection and post-production sampling. This involves monitoring parameters based on operator experience during production and identifying defects through sampling after production. This approach struggles to provide real-time perception of quality risks throughout the entire production process and cannot effectively warn of potential risks before defects occur. Due to the large number of processes and the long process chain involved in tire production, even with the introduction of data mining methods for quality analysis, existing technologies often fail to effectively handle the propagation and nonlinear amplification of deviations across processes and over long timeframes. This results in the accuracy of defect root cause tracing and the foresight of early warnings falling short of actual production needs. Summary of the Invention

[0003] To address the technical problems of existing technologies, which suffer from insufficient accuracy in tracing the root causes of defects and in predictive capabilities for early warning due to the large number of processes and long production chains involved in tire manufacturing, the present invention aims to provide an automatic early warning, analysis, and rectification system for tire quality defects. The specific technical solution adopted is as follows: In a first aspect, the present invention provides an automatic early warning, analysis, and rectification system for tire quality defects. The system includes: a data acquisition module for acquiring process parameters, equipment status data, and finished product inspection data from multiple processes involved in tire production; a risk quantification module for determining adjustment factors based on process parameters, equipment status data, and finished product inspection data; the adjustment factors reflect the cumulative and amplified effects of deviations during transmission between multiple processes; a dynamic threshold generation module for generating confidence thresholds for association rules based on the adjustment factors; the association rules characterize the causal relationship between abnormal process parameters and defect occurrence; the confidence threshold decreases as the adjustment factor increases; and a rule filtering module for comparing the confidence of association rules with the confidence thresholds and filtering out association rules with confidence levels not lower than the confidence thresholds as quality early warning rules.

[0004] In conjunction with the first aspect mentioned above, in one possible implementation, the risk quantification module includes: a deviation sensitivity analysis submodule, used to determine a deviation sensitivity index based on the process standard values, allowable deviation ranges, and average processing time of process parameters in multiple processes; the deviation sensitivity index is used to characterize the sensitivity of process parameters to deviations; the ratio of the deviation sensitivity index to the process standard value is positively correlated and negatively correlated with the average processing time.

[0005] In conjunction with the first aspect mentioned above, in one possible implementation, the risk quantification module further includes: a transmission attenuation analysis submodule, used to determine the deviation transmission attenuation coefficient based on the average sensitivity of the source process, the number of processes traversed from the source process to the target process, and a preset attenuation factor; the source process is the process from which the deviation originates; the target process is the process to which the deviation is transmitted; the deviation transmission attenuation coefficient is used to characterize the remaining strength of the deviation after it is transmitted from the source process to the target process; the deviation transmission attenuation coefficient is negatively correlated with the number of process intervals and decreases as the average sensitivity of the source process increases.

[0006] In conjunction with the first aspect mentioned above, in one possible implementation, the risk quantification module further includes: a nonlinear amplification analysis submodule, used to determine the nonlinear amplification potential energy based on the deviation of the real-time process parameters of the target process from the reference values ​​of the process parameters of the target process, the amplification index of the process parameters, and the coupling strength between processes; the reference values ​​of the process parameters are determined based on the median of the process parameters in continuous production batches with a defect rate lower than a preset threshold; the amplification index is used to characterize the influence weight of the deviation of the process parameters on the severity of defects; the coupling strength between processes is determined based on the probability that a deviation in any process will cause a related defect in another process; the nonlinear amplification potential energy is used to characterize the potential possibility that the deviation of the source process will be amplified under the process state of the target process.

[0007] In conjunction with the first aspect mentioned above, in one possible implementation, the amplification index is calibrated based on the correlation between the degree of deviation of historical process parameters and the severity of defects.

[0008] In conjunction with the first aspect mentioned above, in one possible implementation, the risk quantification module further includes: a path integral submodule, used to determine the cumulative amplified potential energy of the path based on the nonlinear amplified potential energy of each pair of adjacent processes on the deviation transmission path; the deviation transmission path is a sequence of processes from the source process through one or more intermediate processes to the target process.

[0009] In conjunction with the first aspect mentioned above, in one possible implementation, the risk quantification module further includes: an adjustment factor calculation submodule, used to determine the adjustment factor based on the path cumulative amplification potential energy and the preset defect severity weight; the defect severity weight is a preset weight value based on the warning level of the defect; the warning level of the defect is a preset level based on the severity of the defect.

[0010] In conjunction with the first aspect mentioned above, in one possible implementation, the dynamic threshold generation module is specifically used to: determine the relative score of the adjustment factor based on the numerical value of the adjustment factor among all candidate association rules; adjust the preset basic confidence threshold based on the relative score and the preset adjustment coefficient to determine the confidence threshold; the higher the relative score, the greater the downward adjustment of the confidence threshold compared to the basic confidence threshold.

[0011] In conjunction with the first aspect mentioned above, in one possible implementation, the system further includes: an early warning and rectification closed-loop module, used to generate early warning information based on quality early warning rules, and generate rectification work orders containing the source process, deviation parameters in the source process, and the transmission path of the deviation; the rectification work orders are used to indicate the adjustment of the deviation parameters of the source process; the rectification effect is verified based on the quality data after rectification, and the verified effective rectification plan is stored.

[0012] In conjunction with the first aspect mentioned above, in one possible implementation, the process parameters collected by the data acquisition module include: the thickness and tension uniformity of the cord fabric in the calendering process, the temperature and pressure of the rubber compound in the extrusion process, the cutting angle and length in the cutting process, the bonding pressure and bead positioning accuracy in the molding process, and the vulcanization temperature and pressure in the vulcanization process; the finished product inspection data collected by the data acquisition module includes: X-ray defect data and dynamic balance data in the finished product inspection process.

[0013] The present invention has the following beneficial effects: This invention collects process parameters, equipment status, and finished product inspection data from multiple tire production processes through a data acquisition module, providing a comprehensive and real-time data foundation for subsequent analysis. A risk quantification module determines adjustment factors reflecting the cumulative and amplified effects of deviation transmission across processes based on this data, effectively quantifying the deviation transmission patterns in complex process chains. A dynamic threshold generation module dynamically generates confidence thresholds for association rules based on the adjustment factors, and these thresholds decrease as the adjustment factors increase, resulting in lower entry barriers for high-risk rules. A rule filtering module uses dynamic thresholds to filter association rules, thereby capturing low-frequency, high-hazard rules that are easily missed by traditional fixed threshold methods, significantly improving the coverage and accuracy of quality warnings, and achieving a leap from passive post-event detection to proactive risk warning. This solves the technical problem that existing solutions, due to the large number of processes and long process chains involved in tire production, struggle to meet the accuracy of defect root cause tracing and the foresight of warnings to meet actual production needs. Attached Figure Description

[0014] To more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0015] Figure 1 A schematic diagram of the system architecture of an automatic early warning, analysis and rectification system for tire quality defects provided in an embodiment of the present invention; Figure 2 This invention provides an automatic early warning, analysis and rectification device for tire quality defects, as one embodiment of the present invention. Detailed Implementation

[0016] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of an automatic early warning and analysis rectification system for tire quality defects proposed according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0017] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0018] The specific solution of the automatic early warning, analysis and rectification system for tire quality defects provided by the present invention will be described in detail below with reference to the accompanying drawings.

[0019] Please see Figure 1 The diagram illustrates a system architecture of an automatic early warning, analysis, and rectification system 100 for tire quality defects according to an embodiment of the present invention. The system includes: The data acquisition module 101 is used to acquire process parameters, equipment status data and finished product inspection data of multiple processes involved in tire production.

[0020] In one possible implementation, the data acquisition module 101 is deployed at each core process node of the tire production line to collect process parameters, equipment status data, and finished product inspection data of multiple processes involved in tire production in real time. The data acquisition module 101 is connected to the sensing units, detection equipment, and control systems deployed at each process through an industrial network. It continuously acquires multi-dimensional raw data reflecting the production process status at a preset sampling period. For example, the preset sampling period can be set to 1 second. For process parameters that change slowly (such as vulcanization temperature), a sampling period of 10 seconds can be used, while for high dynamic signals (such as vibration), a sampling period of 100 milliseconds can be used. The specific configuration can be flexibly configured according to the physical characteristics of the parameters, providing a complete and time-aligned data foundation for subsequent risk quantification.

[0021] For example, the data acquisition module 101 collects key process parameters reflecting processing quality, operational data reflecting equipment health status, and final inspection results reflecting whether the finished product is qualified for each process. For each collected parameter, the data acquisition module 101 simultaneously records its process standard value, allowable deviation range, and average processing time of the process to which it belongs; for defects in the finished product inspection data, a corresponding defect severity weight is preset according to its severity. All collected data is timestamped and then aggregated to the industrial data platform for use by the risk quantification module 102.

[0022] The risk quantification module 102 is used to determine adjustment factors based on process parameters, equipment status data, and finished product testing data.

[0023] The adjustment factor reflects the cumulative and amplified effect of deviations during the transmission process between multiple processes.

[0024] In one possible implementation, the risk quantification module 102 receives process parameters, equipment status data, and finished product inspection data output by the data acquisition module 101 to determine an adjustment factor that comprehensively reflects the cumulative and amplified effects of deviations during transmission across multiple processes. Internally, this module employs a hierarchical quantification strategy. First, it calculates the deviation sensitivity index for each parameter in each process. Then, it calculates the deviation transmission attenuation coefficient by combining the temporal relationship between processes. Next, it introduces the real-time process status of the current process to calculate the nonlinear amplification potential energy. Finally, it integrates along the deviation transmission path to obtain the path-accumulated amplification potential energy, and fuses it with a preset defect severity weight to output a dimensionless adjustment factor.

[0025] The dynamic threshold generation module 103 is used to generate confidence thresholds for association rules based on adjustment factors. Among them, the association rule is used to characterize the causal relationship between abnormal process parameters and the occurrence of defects; the confidence threshold decreases as the adjustment factor increases.

[0026] In one possible implementation, the dynamic threshold generation module 103 receives the adjustment factor output by the risk quantification module 102 and uses it to generate the confidence threshold of the association rule. The dynamic threshold generation module 103 first obtains a preset basic confidence threshold (e.g., 0.6), then calculates the relative score of its adjustment factor among all candidate rules for each candidate association rule, and then, combined with a preset adjustment coefficient (e.g., 0.5), lowers the basic confidence threshold to generate an adaptive confidence threshold specific to that rule. The candidate association rule refers to the association rule to be compared for confidence threshold, that is, all rules mined from historical data by an association rule mining algorithm (such as the Apriori algorithm).

[0027] The rule filtering module 104 is used to compare the confidence level of association rules with the confidence level threshold, and filter out association rules with a confidence level not lower than the confidence level threshold as quality warning rules.

[0028] In one possible implementation, the rule filtering module 104 receives the adaptive confidence threshold output by the dynamic threshold generation module 103, and the original confidence of each rule calculated by an association rule mining algorithm (such as the Apriori algorithm), to filter out the final quality warning rules. This module compares the original confidence of each rule with its corresponding adaptive confidence threshold, retaining only those rules whose original confidence is not lower than the adaptive confidence threshold.

[0029] For example, suppose the original confidence level of a certain association rule "calendering fabric thickness deviation → vulcanization delamination" is 0.55, while the adaptive confidence level threshold for this rule, calculated by the dynamic threshold generation module 103, is 0.57. Since the original confidence level is lower than the threshold, the rule filtering module 104 filters it out. Another rule, "calendering thickness deviation → cutting angle deviation → forming misalignment → vulcanization delamination," has an original confidence level of 0.52, but because its adjustment factor is high (reflecting a strong nonlinear amplification effect), the adaptive confidence level threshold is lowered to 0.50. Therefore, the original confidence level of 0.52 is not lower than 0.50, and this rule is retained as a quality warning rule. For rules with low adjustment factors and original confidence levels lower than the base threshold, their adaptive confidence level threshold is close to the base threshold and they will still be filtered out.

[0030] The technical solution provided by the above embodiments can bring at least the following beneficial effects: This embodiment collects process parameters, equipment status, and finished product inspection data of multiple processes in tire production through the data acquisition module 101, providing a comprehensive and real-time data foundation for subsequent analysis; the risk quantification module 102 determines the adjustment factor reflecting the cumulative and amplified effect of deviation transmission across processes based on these data, realizing the effective quantification of the deviation transmission law in complex process chains; the dynamic threshold generation module 103 dynamically generates the confidence threshold of the association rule based on the adjustment factor, and the threshold decreases as the adjustment factor increases, so that high-risk rules obtain a lower entry threshold; the rule screening module 104 uses the dynamic threshold to screen the association rules, thereby capturing low-frequency, high-hazard rules that are easily missed by traditional fixed threshold methods, significantly improving the coverage and accuracy of quality early warning, and realizing the leap from passive post-event detection to proactive risk early warning. This solves the technical problem that existing technical solutions, due to the large number of processes and long process chains involved in tire production, cannot meet the actual production needs in terms of the accuracy of defect root cause tracing and the foresight of early warning.

[0031] In one possible implementation, the risk quantification module 102 includes a deviation sensitivity analysis submodule, which is used to determine the deviation sensitivity index based on the process standard values, allowable deviation ranges and average processing time of process parameters in multiple processes.

[0032] The deviation sensitivity index is used to characterize the sensitivity of process parameters to deviations; the ratio of the deviation sensitivity index to the process standard value is positively correlated and negatively correlated with the average processing time.

[0033] In one possible implementation, the deviation sensitivity analysis submodule receives the process standard values, allowable deviation ranges, and average processing time of key parameters for each process provided by the data acquisition module 101, and uses them to calculate the deviation sensitivity index of each parameter, providing a weighting basis for subsequent cross-process deviation transmission analysis.

[0034] For example, the deviation sensitivity index of key parameter j in process i. Satisfy the following formula 1: Formula 1; Where i is the process number (ranging from 1 to 6, representing calendering, extrusion, cutting, molding, vulcanization, and finished product inspection, respectively); The process standard value (nominal value, in millimeters, degrees, megapascals, etc., depending on the parameter type) of key parameter j in process i. The allowable deviation range of parameter j (the maximum allowable fluctuation range of the process, with the same dimensions as the standard value). The average processing time for process i (unit: seconds / piece, obtained by averaging the total time spent on the continuous production of 100 tires).

[0035] Calculate the process stringency ratio. The larger the ratio, the smaller the allowable deviation relative to the standard value, meaning the more stringent the process control requirements and the more sensitive the parameters are to deviations. The reciprocal of processing time represents the production frequency per unit time. Shorter processing times indicate higher process variation frequencies, requiring greater precision in capturing and controlling deviations. Multiplying the process stringency ratio by the frequency density yields a comprehensive deviation sensitivity index. This index is positively correlated with the process stringency ratio and negatively correlated with processing time.

[0036] Understandably, the allowable deviation range This refers to the allowable fluctuation range specified in the process documentation. For any parameter with practical production significance, this value is always positive (e.g., allowable deviation of fabric thickness ±0.05mm, i.e., 0.05mm>0), and cannot be zero; average processing time of the process. This was calculated by averaging the total time taken to produce 100 tires consecutively. In actual production, each step requires time. This holds true consistently. In conclusion, neither of the two denominators in the formula can be zero, and the calculation always has mathematical meaning.

[0037] For example, taking the fabric thickness parameter in the calendering process as an example, if the standard process value is set to 1.0 mm and the allowable deviation range is set to ±0.05 mm, then the process stringency ratio for this parameter is 1.0 divided by 0.05, which equals 20. If the average processing time for this process is 120 seconds, then the frequency density is 1 / 120 ≈ 0.0083. Multiplying the process stringency ratio by the frequency density yields a deviation sensitivity index of approximately 0.166. For the rubber compound temperature parameter in the extrusion process, if the standard process value is 100℃, the allowable deviation range is ±2℃, the process stringency ratio is 50, the average processing time is 60 seconds, the frequency density is 0.0167, and the deviation sensitivity index is approximately 0.835. It can be seen that the more stringent the process requirements (smaller the allowable deviation) or the shorter the processing time (faster the production cycle), the higher the deviation sensitivity index for that parameter, indicating that the parameter is more sensitive to deviations and should be given higher weight in subsequent analysis. The technical solution provided by the above embodiments can bring at least the following beneficial effects: In this embodiment, a deviation sensitivity analysis submodule is added to the risk quantification module 102. The deviation sensitivity index is calculated using the process standard value, allowable deviation range and average processing time of the process parameters. This index is positively correlated with the process accuracy requirements and negatively correlated with the process processing speed. Thus, it can objectively reflect the inherent sensitivity of different processes and different parameters to deviations, providing a scientific weight basis for subsequent cross-process deviation transmission analysis, avoiding the deviation caused by subjective weighting, and improving the adaptability and accuracy of the risk quantification model.

[0038] In one possible implementation, the risk quantification module 102 further includes a transmission attenuation analysis submodule, used to determine the deviation transmission attenuation coefficient based on the average sensitivity of the source process, the number of processes passed from the source process to the target process, and a preset attenuation factor.

[0039] Here, the source process is the process from which the deviation originates; the target process is the process to which the deviation is transmitted; the deviation transmission attenuation coefficient is used to characterize the residual strength of the deviation after it is transmitted from the source process to the target process; the deviation transmission attenuation coefficient is negatively correlated with the number of process intervals and decreases as the average sensitivity of the source process increases.

[0040] For example, taking the calendering process as the source process and the vulcanization process as the target process, the deviation sensitivity analysis submodule has calculated the deviation sensitivity index of all key parameters in the calendering process, and taken the arithmetic mean to obtain the average sensitivity of the calendering process. Simultaneously, the overall average of the average sensitivity of the six major processes is calculated to obtain the relative sensitivity of the calendering process (i.e., the ratio of its average sensitivity to the overall average). The number of processes from calendering to vulcanization is the interval between them in the process flow, and the preset attenuation factor is calibrated to a fixed constant using historical defect transmission data. Multiplying the attenuation factor, the number of process intervals, and the relative sensitivity of the source process yields a comprehensive influence factor. Then, taking the negative exponent of this influence factor with the natural constant e as the base yields the deviation transmission attenuation coefficient. If the average sensitivity of the source process is low, the relative sensitivity is also low, the overall impact factor is small, and the attenuation coefficient is close to 1, indicating slow deviation attenuation and high residual strength. If the source process is an extrusion process, its average sensitivity is high, the relative sensitivity is significantly increased, the overall impact factor increases accordingly, and the attenuation coefficient decreases significantly, indicating faster deviation attenuation and lower residual strength. This is because deviations generated by highly sensitive processes are more easily amplified or transformed by the process conditions of subsequent processes during transmission, resulting in less residual energy actually being transmitted to the target process. If the source process remains unchanged but the target process is moved forward, the number of process intervals decreases, the overall impact factor decreases accordingly, the attenuation coefficient increases, and the residual strength of the deviation is higher.

[0041] In one possible implementation, the calibration method for the preset attenuation factor α is as follows: Retrieve recorded deviation propagation cases from historical production data, such as "cutting angle deviation → molding misalignment," "cutting angle deviation → vulcanization wear," etc. Each case includes the source process number, target process number, average sensitivity of the source process, overall average sensitivity, and the actual observed defect occurrence frequency or deviation residual strength ratio. The actual residual strength ratio can be approximated by statistically analyzing the conditional probability of the target process generating related defects when a deviation occurs in the source process. Then, construct an objective function, such as minimizing the sum of squared errors between the model-predicted deviation propagation attenuation coefficient and the actual observed ratio, and solve for the preset attenuation factor α using a nonlinear least squares method.

[0042] For example, the deviation propagation attenuation coefficient between process i and target process k The following formula 2 is satisfied: Formula 2; in, This is a preset attenuation factor used to control the attenuation rate. It is calibrated using historical defect data and has a default value of 0.2. This refers to the number of process intervals, which is the number of processes that must be traversed from the source process to the target process. The average sensitivity of source process i is obtained by the DSI arithmetic mean of all key parameters in that process; This is the overall average of the sensitivity of the six major processes, used to normalize the sensitivity of the source process; This is a natural exponential function (base is Euler's number e≈2.71828), representing exponential decay. The exponential term... The comprehensive impact factor includes, among which It represents the relative sensitivity of the source process; the larger the ratio, the more sensitive the source process is and the slower the degradation. It is an exponential decay function. When the process interval is larger or the sensitivity of the source process is higher, the absolute value of the exponential term is larger and the DTA value is smaller, that is, the residual strength of the deviation is lower. It is a dimensionless attenuation factor that represents the proportion of residual strength after the deviation is transmitted from the source process to the target process. The closer the value is to 1, the smaller the attenuation and the greater the impact; the closer the value is to 0, the more severe the attenuation and the negligible impact.

[0043] The technical solution provided by the above embodiments can bring at least the following beneficial effects: In this embodiment, the transmission attenuation analysis submodule calculates the deviation transmission attenuation coefficient based on the average sensitivity of the source process, the number of process intervals and the preset attenuation factor. This coefficient is negatively correlated with the number of process intervals and decreases as the sensitivity of the source process increases. It accurately simulates the exponential attenuation law of deviation in long process flow, enabling the system to quantitatively evaluate the remaining strength of deviation after it is transmitted from the preceding process to the subsequent process. This provides a quantitative indicator for judging whether the deviation is worth tracking, thereby effectively filtering out irrelevant minor disturbances and focusing on the key deviation sources that truly have cross-process influence.

[0044] In one possible implementation, the risk quantification module 102 further includes a nonlinear amplification analysis submodule, used to determine the nonlinear amplification potential energy based on the deviation between the real-time process parameters of the target process and the reference values ​​of the process parameters of the target process, the amplification index of the process parameters, and the coupling strength between processes.

[0045] Among them, the process parameter reference value is determined based on the median of process parameters in continuous production batches with a defect rate lower than a preset threshold (e.g., 0.5); the amplification index is used to characterize the influence weight of the deviation of process parameters on the severity of defects; the coupling strength between processes is determined based on the probability that a deviation in one process will cause related defects in another process; and the nonlinear amplification potential energy is used to characterize the potential possibility that the deviation of the source process will be amplified under the process state of the target process.

[0046] In one possible implementation, the nonlinear amplification analysis submodule receives the deviation propagation attenuation coefficient output by the propagation attenuation analysis submodule, and combines it with the deviation of the real-time process parameters of the target process from the reference value, the amplification index of the process parameters, and the coupling strength between processes to determine the nonlinear amplification potential energy. This potential energy is used to characterize the potential likelihood that the deviation of the source process will be amplified under the current process state of the target process.

[0047] For example, taking the vulcanization process as the target process, the data acquisition module 101 collects the vulcanization temperature parameter in real time. Its reference value is the median of this parameter in the continuous production batch with the lowest defect rate (e.g., 155℃). If the current real-time vulcanization temperature is 158℃, the deviation is a positive deviation of 3℃; the historical fluctuation range near the reference value is statistically ±5℃, so the relative deviation is approximately 0.6. The amplification index of this parameter is obtained through historical data analysis. For example, the Spearman correlation coefficient between vulcanization temperature and delamination defects is high, and the normalized amplification index is set to 1.8. Simultaneously, the coupling strength from the source process (such as the cutting process) to the vulcanization process is obtained by statistically analyzing the probability of vulcanization causing related defects when the cutting process deviates, for example, 1.2. The deviation transmission attenuation coefficient is obtained from the transmission attenuation analysis submodule, for example, 0.5. The nonlinear amplification potential energy is then obtained by multiplying the amplification factor (which grows superlinearly with increasing deviation) by the attenuation coefficient and the coupling strength. When the deviation is small, the amplification factor is close to 1, and the potential energy is mainly determined by the attenuation coefficient and coupling strength. When the deviation is large, the amplification factor increases significantly, and the potential energy rises sharply, indicating that the current process conditions can easily amplify the preceding deviation into a serious defect. If the other target process is a molding process, and its key parameter is the bonding pressure, with a reference value of 0.5 MPa and a real-time value of 0.55 MPa, the deviation is small, and the amplification index of this parameter is low (e.g., 1.1). Then the nonlinear amplification potential energy is relatively small, indicating that even if there is a preceding deviation, the risk of it being amplified under the current process conditions is low.

[0048] For example, nonlinear amplification of potential energy The following formula 3 is satisfied: Formula 3; in, For process sequence number, For the target process number, Let t be the set of parameters for the target process k, and t be the time interval. This is the real-time monitoring value of parameter j at time t (with the same dimensions as parameter j). This is a historical reference value for parameter j (the median of the consecutive production batches with the lowest defect rate, in the same dimension). The historical maximum value of parameter j The historical minimum value of parameter j (with the same dimension); Let be the amplification exponent of parameter j (dimensionless, normalized to [0.5, 2.0] by Spearman correlation coefficient); The symbol for multiplication indicates that the product is taken over all parameters in set J; The attenuation coefficient is the deviation propagation factor. The coupling strength between processes; This is the ratio of the real-time deviation to the historical extreme value range (between 0 and 1). Adding 1 changes the range to 1-2, then... The power of the power is used to obtain the magnification factor of parameter j; Multiplying the amplification factors of all key parameters in the current process together represents the comprehensive amplification effect under multi-parameter coupling; It is a dimensionless risk potential energy index, representing the potential for deviations in the source process to be nonlinearly amplified by the process conditions of the target process under the current technological state. The larger the value, the higher the risk.

[0049] In one possible implementation, when the difference between historical extreme values ​​is 0 (i.e. When the parameter is considered to have no fluctuation and its relative deviation is zero, the corresponding amplification factor is... Take 1.

[0050] The technical solution provided by the above embodiments can bring at least the following beneficial effects: This embodiment calculates the nonlinear amplification potential energy by introducing the deviation degree between the real-time process parameters of the target process and the reference value, the parameter amplification index, and the coupling strength between processes. This potential energy is used to characterize the potential possibility that the deviation of the source process is amplified under the current process state. It breaks through the limitations of the traditional linear superposition model and can capture the nonlinear mutation risk triggered by the superposition of specific process conditions and deviations. This enables the system to have the ability to perceive the phenomenon of a sharp increase in defects caused by the deterioration of the process state in advance, greatly enhancing the sensitivity of the early warning and the accuracy of the root cause analysis.

[0051] In one possible implementation, the amplification index is calibrated based on correlation analysis between the deviation of historical process parameters and the severity of defects. Specifically, process parameter records before the defect occurred are retrieved from historical production data. For each key parameter, the Spearman rank correlation coefficient is calculated between its deviation (i.e., the absolute value of the difference between the real-time value and the reference value divided by the historical extreme value difference) and the corresponding severity level of the defect. This coefficient reflects the strength and direction of the monotonic correlation between the two. Since defect severity levels are usually three-level (e.g., general, severe, and urgent), the Spearman correlation coefficient can effectively handle this level of data. Then, the calculated correlation coefficient is linearly or non-linearly mapped from the original range (-1 to 1) to a preset weighted interval, such as [0.5, 2.0]. A correlation coefficient of 0 is mapped to 1 (no amplification effect), a higher positive correlation coefficient results in a larger mapped value (stronger amplification effect), and a negative correlation coefficient is mapped to a value less than 1 (indicating a negative correlation between the deviation and the severity of the defect, which is less common in actual production). The mapped value is the amplification index of the parameter.

[0052] For example, regarding the vulcanization temperature parameter in the vulcanization process, the relative degree of deviation of the vulcanization temperature from the reference value in batches before the occurrence of delamination defects in history is statistically analyzed, and Spearman correlation analysis is performed with the severity level of the defect (e.g., delamination leading to tire scrapping, classified as a level three emergency defect). The correlation coefficient is 0.85, which is then mapped to the interval [0.5, 2.0] and set to 1.8. For the bonding pressure parameter in the molding process, if the correlation coefficient between its deviation and the severity level of the bead misalignment defect is 0.4, the amplification index after mapping is approximately 1.2. If the deviation of a parameter has no significant correlation with the severity of the defect (correlation coefficient close to 0), the amplification index is set to 1, indicating that the deviation of the parameter does not exacerbate the severity of the defect. For individual parameters with negative correlation (e.g., deviations in some parameters actually alleviate the defect), the amplification index can be set between 0.5 and 0.8 to reduce its contribution to the nonlinear amplification potential energy.

[0053] The technical solution provided in the above embodiments can bring at least the following beneficial effects: In this embodiment, the amplification index is calibrated based on the correlation between the deviation of historical process parameters and the severity of defects, so that the index has a clear physical meaning and statistical support, avoiding the uncertainty caused by arbitrary human setting. By dynamically adjusting the amplification index through data-driven methods, the system can adapt to the actual risk characteristics of different tire specifications, different production lines, and different process stages, thereby ensuring that the calculation of nonlinear amplification potential energy always fits the real production process, improving the model's generalization ability and long-term effectiveness.

[0054] In one possible implementation, the risk quantification module 102 further includes a path integral submodule, used to determine the path cumulative amplification potential energy based on the nonlinear amplification potential energy of each pair of adjacent processes on the deviation transmission path.

[0055] The deviation propagation path is a sequence of processes that goes from the source process through one or more intermediate processes to the target process.

[0056] In one possible implementation, the path integral submodule receives the nonlinear amplification potential energy of each pair of adjacent processes along the deviation propagation path from the nonlinear amplification analysis submodule, and uses this to determine the cumulative amplification potential energy of the entire path. The deviation propagation path is defined as a sequence of processes starting from the source process, passing through one or more intermediate processes, and finally reaching the target process. This module integrates the nonlinear amplification potential energy between each pair of adjacent processes along the path to obtain the cumulative amplification potential energy of the path, which reflects the overall amplification effect of the entire propagation chain.

[0057] For example, taking the calendering process as the starting point, the process proceeds through extrusion, cutting, and forming, ultimately reaching the vulcanization process as the target process. The nonlinear amplification analysis submodule calculates the nonlinear amplification potential energy between four adjacent processes: calendering to extrusion, extrusion to cutting, cutting to forming, and forming to vulcanization, respectively denoted as 0.85, 0.92, 0.78, and 0.88. The path integral submodule integrates these segmented potential energies to obtain a combined result of approximately 0.85, 0.92, 0.78, and 0.88 for the cumulative amplification potential energy along the path. This combined result reflects the total amplification degree of the nonlinear amplification effect after superposition of each segment in the entire deviation transmission chain from calendering to vulcanization. If the nonlinear amplification potential energy of a certain segment is particularly high (e.g., the potential energy of the forming to vulcanization segment is 1.5), the cumulative amplification potential energy of the entire path will also increase significantly, indicating that the deviation on this path is easily amplified into a serious defect in subsequent processes. Conversely, if the potential energy in a certain segment is close to zero, the cumulative amplified potential energy of the entire path will also approach zero, indicating that even if there is a deviation at the source, it will be effectively suppressed during the transmission process in that segment.

[0058] The technical solution provided by the above embodiments can bring at least the following beneficial effects: This embodiment utilizes the nonlinear amplification potential energy of each pair of adjacent processes on the deviation transmission path to calculate the cumulative amplification potential energy of the path. The single-step amplification effect is integrated along the complete path from the source process through intermediate processes to the target process, thereby obtaining the overall amplification effect of the entire deviation transmission chain. This allows the system to assess the comprehensive risk of deviation transmission in multiple processes and long chains from a global perspective, avoiding one-sided conclusions caused by isolated analysis of each process, and providing a reliable integral basis for the determination of subsequent adjustment factors.

[0059] In one possible implementation, the risk quantification module 102 further includes an adjustment factor calculation submodule, used to determine the adjustment factor based on the path cumulative amplification potential energy and the preset defect severity weight.

[0060] Among them, the severity weight is a preset weight value based on the warning level of the defect; the warning level of the defect is a preset level based on the severity of the defect.

[0061] In one possible implementation, the adjustment factor calculation submodule receives the path cumulative amplification potential energy output by the path integral submodule and combines it with a preset defect severity weight to determine the comprehensive adjustment factor for each association rule.

[0062] For example, for a certain association rule, the path integral submodule calculates a path cumulative amplification potential of 0.42, while the defect type that the rule points to (such as delamination defect) has a preset severity weight of 3 (Level 1 emergency defect). The adjustment factor calculation submodule combines the path cumulative amplification potential with the defect severity weight to obtain an adjustment factor of approximately 1.26 for this rule. For another rule pointing to a general defect (severity weight of 1) and with a low path cumulative amplification potential (e.g., 0.15), its adjustment factor is only 0.15. The adjustment factor reflects both the cumulative amplification strength of the deviation transmission chain and the severity level of the defect itself: even if the path cumulative amplification potential is high, if the defect severity is low, the adjustment factor may still be at a moderate level; conversely, if the defect severity is high, even if the path cumulative amplification potential is moderate, the adjustment factor will be significantly increased, resulting in a higher risk ranking.

[0063] For example, adjustment factor Satisfy the following formula 4: Formula 4; in, This is an association rule in the form of "abnormal parameter set of process i → defect occurs in process k"; t is the time point number, and the total time period T (e.g., the last 30 production days). This is an indicator function, taking the value 1 when rule R is met at time t, and 0 otherwise. The conditions for its validity are: the parameters of the source process deviate from the optimal range by more than 10%, and a defect occurs in the target process within a reasonable time window. The accumulated amplified potential energy of the path at time t (from each segment of the path) (The product is dimensionless). The severity weight of the defect at time t (for example, level 1 defect has a weight of 3, level 2 defect has a weight of 2, level 3 defect has a weight of 1, dimensionless). These are parameter tuning coefficients, and their values ​​should be extremely small positive numbers (e.g., 0.01) to avoid a denominator of 0. For all times when the rules are true, calculate Summation, that is, only when the rule is true, the product of the path accumulation amplification potential energy and the defect severity weight at that moment is accumulated; The total number of times the rule is true; The average risk score of rule R is obtained, i.e., the adjustment factor. ; It is a dimensionless comprehensive risk score used to measure the average severity of an association rule after considering the bias amplification effect and the severity of the defect. The higher the value, the greater the risk of the defect propagation path corresponding to that rule.

[0064] The technical solution provided by the above embodiments can bring at least the following beneficial effects: This embodiment determines the final adjustment factor based on the cumulative amplification potential energy of the path and the preset defect severity weight, wherein the defect severity weight is preset according to the warning level of the defect, realizing the preferential protection of high-risk defect types. This adjustment factor integrates the cumulative amplification intensity of the deviation transmission path and the hazard level of the defect itself, enabling the system to prioritize defect patterns that have both strong amplification effects and may lead to serious consequences when screening association rules, thereby ensuring the reasonable allocation of warning resources and the efficient investment of rectification measures.

[0065] In one possible implementation, the dynamic threshold generation module 103 is specifically used to: determine the relative score of the adjustment factor based on the numerical value of the adjustment factor among all candidate association rules; and adjust the preset basic confidence threshold based on the relative score and the preset adjustment coefficient to determine the confidence threshold.

[0066] Among them, the higher the relative score, the greater the reduction in the confidence threshold compared to the base confidence threshold.

[0067] In one possible implementation, the dynamic threshold generation module 103 receives the adjustment factors of each association rule output by the risk quantification module 102, and uses them to generate a unique adaptive confidence threshold for each rule. This module first compares the adjustment factors of all candidate association rules numerically to determine the relative ranking of each rule's adjustment factor among all rules, and then obtains a relative score between 0 and 1 through normalization. The higher the relative score, the higher the risk importance of the rule. Then, the module obtains a preset base confidence threshold and a preset adjustment coefficient, and dynamically adjusts the base threshold according to the relative score: the higher the relative score, the greater the reduction in the confidence threshold compared to the base threshold, i.e., obtaining a lower entry threshold.

[0068] For example, confidence threshold The following formula 5 is satisfied: Formula 5; in, This represents a specific association rule, in the form of "abnormal parameters in the preceding process → defect occurrence in the subsequent process"; The confidence threshold for rule R is used to compare it with the rule's traditional confidence level to determine whether the rule is retained as a quality warning rule. The preset baseline confidence threshold is a fixed constant, such as 0.6, which represents the baseline screening standard without adaptive adjustment. This is a preset adjustment coefficient used to control the magnitude of threshold reduction. Its value is usually between 0 and 1. The larger λ is, the greater the threshold reduction of the high-risk rule. R is the normalized relative score of the adjustment factor C(R) of rule R among all candidate association rules. It is obtained by linearly normalizing the adjustment factors of all rules, with a value range of [0, 1]. The higher the value of r(R), the higher the risk importance of the rule. It is a dynamic, adaptive confidence threshold that automatically adjusts based on the risk importance of the rule: the higher the risk of the rule, the lower its threshold, making it easier to be screened as a quality warning rule; the lower the risk of the rule, the lower its threshold is close to the basic threshold, and the screening criteria remain strict.

[0069] The technical solution provided by the above embodiments can bring at least the following beneficial effects: This embodiment calculates the relative score of the adjustment factor among all candidate association rules, and dynamically lowers the basic confidence threshold based on the relative score and the preset adjustment coefficient. The higher the relative score, the greater the threshold reduction. This mechanism transforms the uniform and fixed confidence threshold in traditional algorithms into an adaptive threshold related to the importance of rule risk. This allows low-frequency but high-risk rules to be retained even if they cross the threshold, while high-frequency but low-risk rules are still strictly restricted. This achieves algorithm optimization from finding the most common rules to finding the most important rules, greatly improving the practical value of the mining results.

[0070] In one possible implementation, the tire quality defect automatic early warning and analysis rectification system further includes: an early warning and rectification closed-loop module, which is used to generate early warning information based on quality early warning rules, generate rectification work orders containing the source process, deviation parameters in the source process, and the transmission path of the deviation, verify the rectification effect based on the quality data after rectification, and store the verified effective rectification plan.

[0071] Among them, the rectification work order is used to indicate the adjustment of deviation parameters of the source process.

[0072] In one possible implementation, the early warning and rectification closed-loop module receives quality early warning rules output by the rule filtering module 104, uses these rules to generate early warning information and automatically dispatch rectification work orders, while simultaneously tracking the rectification effect and reusing effective solutions. The early warning and rectification closed-loop module first parses the source process, deviation parameters within the source process, and the deviation propagation path contained in the quality early warning rules, generating a readable early warning description and a structured rectification work order. The rectification work order clearly indicates the process requiring adjustment, the target parameters, and the expected adjustment direction. Subsequently, the module pushes the rectification work order to the production execution system or relevant personnel terminals, instructing targeted adjustments to the deviation parameters of the source process. After rectification is completed, the module automatically acquires quality data from subsequent batches, verifying the effectiveness of the rectification measures by comparing indicators such as the defect incidence rate and deviation sensitivity index before and after rectification. For verified effective rectification solutions, the module stores them in a solution library for automatic recommendation or reuse when similar defects occur in the future.

[0073] For example, suppose the rule filtering module 104 outputs a quality warning rule: "Calendar fabric thickness deviation → Cutting angle deviation → Forming deviation → Vulcanization delamination". Based on this, the warning and rectification closed-loop module generates a warning message: "A high risk of delamination defects in the vulcanization process is detected due to calender fabric thickness deviation being transmitted through cutting and forming." Simultaneously, a rectification work order is generated, specifying the calender process as the source process, the fabric thickness as the deviation parameter, and recommending that the thickness deviation be controlled within 50% of the allowable range. The work order is automatically dispatched to the calender process operators and process engineers. After rectification, the module automatically reads the quality data of the subsequent 30 batches, finding that the delamination defect incidence rate decreased from 2.3% before rectification to 0.4%, and the sensitivity index of fabric thickness deviation in the calender process decreased from 0.166 to 0.102. The module determines that the rectification is effective and stores the rectification plan (calender fabric thickness deviation controlled within ±0.02mm) in the plan library. If a similar warning rule appears again, the module can automatically recommend the solution for operators to refer to. If the defect rate does not improve significantly after rectification, the module will not store the solution or will mark it as an invalid solution.

[0074] The technical solution provided in the above embodiments can bring at least the following beneficial effects: This embodiment generates early warning information based on the selected quality early warning rules, and issues rectification work orders containing source processes, deviation parameters, and transmission paths to guide operators in adjusting the root cause deviations; at the same time, the system verifies the rectification effect based on the quality data after rectification, and stores and reuses the verified effective solutions. This closed-loop design enables each early warning and rectification to be accumulated into knowledge, driving the continuous improvement of quality control level, and realizing a complete quality management process from intelligent early warning to precise analysis and then to closed-loop rectification.

[0075] In one possible implementation, the process parameters collected by the data acquisition module 101 include: the thickness and tension uniformity of the cord fabric in the calendering process, the temperature and pressure of the rubber compound in the extrusion process, the cutting angle and length in the cutting process, the bonding pressure and bead positioning accuracy in the molding process, and the vulcanization temperature and pressure in the vulcanization process; the finished product inspection data collected by the data acquisition module 101 includes: X-ray defect data and dynamic balance data in the finished product inspection process.

[0076] The technical solution provided in the above embodiments can bring at least the following beneficial effects: This embodiment limits key process parameters (such as fabric thickness, rubber temperature, cutting angle, bonding pressure, vulcanization temperature, etc.) and finished product inspection data (X-ray defects, dynamic balance), and simultaneously collects equipment vibration, current, raw material batch, and environmental data. This multi-level, multi-dimensional data acquisition system provides rich and accurate input features for the risk quantification model, ensuring that subsequent deviation sensitivity analysis, transmission attenuation calculation, and nonlinear amplification evaluation are all based on real and complete production data, thereby guaranteeing the reliability and industrial applicability of the entire early warning system.

[0077] Please see Figure 2 The diagram illustrates a schematic of an automatic early warning and analysis rectification device 200 for tire quality defects according to an embodiment of the present invention. The device includes: a communication unit 201 and a processing unit 202. The communication unit 201 is used to acquire process parameters, equipment status data, and finished product inspection data of multiple processes involved in tire production. The processing unit 202 is used to determine an adjustment factor based on the process parameters, equipment status data, and finished product inspection data. The adjustment factor reflects the cumulative and amplified effect of deviations during transmission between multiple processes. Based on the adjustment factor, a confidence threshold for association rules is generated. The association rules are used to characterize the causal relationship between abnormal process parameters and defect occurrence. The confidence threshold decreases as the adjustment factor increases. The confidence of the association rules is compared with the confidence threshold, and association rules with a confidence level not lower than the confidence threshold are selected as quality early warning rules.

[0078] It should be noted that the order of the above embodiments of the present invention is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0079] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

Claims

1. An automatic early warning, analysis, and rectification system for tire quality defects, characterized in that, include: The data acquisition module is used to acquire process parameters, equipment status data, and finished product inspection data for multiple processes involved in tire production. The risk quantification module is used to determine adjustment factors based on the process parameters, equipment status data, and finished product inspection data; the adjustment factors reflect the cumulative and amplified effects of deviations during the transmission process between the multiple processes. A dynamic threshold generation module is used to generate a confidence threshold for an association rule based on the adjustment factor; the association rule is used to characterize the causal relationship between abnormal process parameters and defect occurrence; the confidence threshold decreases as the adjustment factor increases. The rule filtering module is used to compare the confidence level of the association rule with the confidence level threshold, and filter out the association rules with a confidence level not lower than the confidence level threshold as quality warning rules.

2. The automatic early warning, analysis and rectification system for tire quality defects according to claim 1, characterized in that, The risk quantification module includes: The deviation sensitivity analysis submodule is used to determine the deviation sensitivity index based on the process standard values, allowable deviation ranges, and average processing time of the process parameters in the multiple processes. The deviation sensitivity index is used to characterize the sensitivity of the process parameters to deviations. The deviation sensitivity index is positively correlated with the process standard value and negatively correlated with the average processing time.

3. The automatic early warning, analysis and rectification system for tire quality defects according to claim 2, characterized in that, The risk quantification module also includes: The propagation attenuation analysis submodule is used to determine the deviation propagation attenuation coefficient based on the average sensitivity of the source process, the number of processes from the source process to the target process, and a preset attenuation factor; the source process is the process from which the deviation originates; the target process is the process to which the deviation is propagated; the deviation propagation attenuation coefficient is used to characterize the remaining strength of the deviation after it is propagated from the source process to the target process; the deviation propagation attenuation coefficient is negatively correlated with the number of process intervals and decreases as the average sensitivity of the source process increases.

4. The automatic early warning, analysis and rectification system for tire quality defects according to claim 3, characterized in that, The risk quantification module also includes: The nonlinear amplification analysis submodule is used to determine the nonlinear amplification potential energy based on the deviation of the real-time process parameters of the target process from the reference values ​​of the process parameters of the target process, the amplification exponent of the process parameters, and the coupling strength between processes. The reference values ​​of the process parameters are determined based on the median of the process parameters in continuous production batches with a defect rate below a preset threshold. The amplification exponent is used to characterize the influence weight of the deviation of the process parameters on the severity of defects. The coupling strength between processes is determined based on the probability that a deviation in one process will cause a related defect in another process. The nonlinear amplification potential energy is used to characterize the potential possibility that the deviation of the source process will be amplified under the process state of the target process.

5. The automatic early warning, analysis and rectification system for tire quality defects according to claim 4, characterized in that, The amplification index is calibrated based on the correlation between the degree of deviation of historical process parameters and the severity of defects.

6. The automatic early warning, analysis and rectification system for tire quality defects according to claim 4, characterized in that, The risk quantification module also includes: The path integral submodule is used to determine the cumulative amplified potential energy of the path based on the nonlinear amplified potential energy of each pair of adjacent processes on the deviation transmission path; the deviation transmission path is a sequence of processes from the source process through one or more intermediate processes to the target process.

7. The automatic early warning, analysis and rectification system for tire quality defects according to claim 6, characterized in that, The risk quantification module also includes: The adjustment factor calculation submodule is used to determine the adjustment factor based on the path accumulated amplification potential energy and the preset defect severity weight; the defect severity weight is a preset weight value based on the warning level of the defect; the warning level of the defect is a preset level based on the severity of the defect.

8. The automatic early warning, analysis and rectification system for tire quality defects according to claim 1, characterized in that, The dynamic threshold generation module is specifically used for: The relative score of the adjustment factor is determined based on the numerical value of the adjustment factor among all candidate association rules. Based on the relative score and a preset adjustment coefficient, the preset base confidence threshold is adjusted to determine the confidence threshold; the higher the relative score, the greater the downward adjustment of the confidence threshold compared to the base confidence threshold.

9. The automatic early warning, analysis and rectification system for tire quality defects according to claim 6, characterized in that, The aforementioned automatic early warning, analysis, and rectification system for tire quality defects also includes: The early warning and rectification closed-loop module is used to generate early warning information based on the quality early warning rules, and generate a rectification work order containing the source process, the deviation parameters in the source process, and the transmission path of the deviation; the rectification work order is used to indicate the adjustment of the deviation parameters of the source process. The effectiveness of the rectification is verified based on the quality data after rectification, and the verified and effective rectification solutions are stored.

10. The automatic early warning, analysis and rectification system for tire quality defects according to claim 1, characterized in that, The process parameters collected by the data acquisition module include: the thickness and tension uniformity of the cord fabric in the calendering process, the temperature and pressure of the rubber compound in the extrusion process, the cutting angle and length in the cutting process, the bonding pressure and bead positioning accuracy in the molding process, and the vulcanization temperature and pressure in the vulcanization process; the finished product inspection data collected by the data acquisition module includes: X-ray defect data and dynamic balance data in the finished product inspection process.

Images