A method for collecting production data of mesh belt layer of industrial and mining explosion-proof tire

Abstract

The invention relates to the technical field of data acquisition, and in particular to a method for acquiring process production data of a mesh belt layer of an industrial and mining explosion-proof tire. In the method, in a lamination process of a mesh belt layer of an industrial and mining explosion-proof tire, monitoring data sequences in N dimensions are respectively acquired; at least one suspected abnormal data is acquired according to the suspected abnormal degree of each data in the monitoring data sequence, and an adaptive weight of the suspected abnormal data is acquired; according to the adaptive weight of each suspected abnormal data, each suspected abnormal data is respectively subjected to weighted recursive average filtering to obtain filtered data, and the filtered data is used to replace the corresponding suspected abnormal data in the dimensional monitoring data sequence to obtain a final data acquisition sequence, and data acquisition is completed, and abnormal data caused by friction in the lamination process is filtered out, thereby ensuring that the process production data acquired in the lamination process are all valid data.

Description

Technical Field

[0001] The invention relates to the technical field of data collection, and in particular to a method for collecting process production data of a mesh belt layer of an industrial and mining explosion-proof tire. Background Art

[0002] The mesh belt process of industrial and mining explosion-proof tires is a manufacturing technology specifically used to improve tire wear resistance, puncture resistance and safety performance. This process is widely used in the manufacture of vehicle tires in harsh environments such as mines and construction sites. For industrial and mining explosion-proof tires, the lamination step of the belt layer is crucial. The lamination quality of the belt layer directly affects the overall performance and safety of the tire. Therefore, during the lamination process of the belt layer, it is necessary to collect pressure monitoring data and temperature process data at the lamination point to ensure the normal and stable lamination process. Since the lathe machines used in the lamination process are usually directly exposed to the workshop environment, it is easy to cause the collected temperature time series data and pressure time series data to have more fluctuations, such as noise data or mutation data. Therefore, it is necessary to filter the monitoring data collected during the lamination process, and judge whether the lamination process of the belt layer is normal by whether there are abnormal mutation points caused by lamination abnormalities in the filtered monitoring data.

[0003] In the prior art, a weighted recursive average filtering algorithm is usually used to filter the monitoring data collected during the lamination process. The basic idea of ​​the weighted recursive average filtering algorithm is to use the weighted average of the current sample value and the previous filtering result to calculate the new filtering result, thereby reducing the impact of random noise on the signal. It has strong applicability to real-time signal processing and is suitable for filtering real-time collected data during high-speed operation of the tire lamination device. However, during the lamination of the belt layer, friction may occur between the pressure roller and the rubber surface during the extrusion process, and the corresponding mutation data is not easy to be filtered out, resulting in abnormal data corresponding to non-lamination abnormalities in the filtered monitoring data, affecting the correct collection of valid data.

[0004] Therefore, how to improve the effect of filtering the monitoring data collected during the lamination process using the weighted recursive average filtering algorithm to ensure the effective collection of process production data of the mesh belt layer of industrial and mining explosion-proof tires has become an urgent problem to be solved. Summary of the invention

[0005] In view of this, an embodiment of the present invention provides a method for collecting process production data of the mesh belt layer of industrial and mining explosion-proof tires, so as to solve the problem of how to improve the effect of filtering the monitoring data collected during the lamination process using a weighted recursive average filtering algorithm, so as to ensure the effective collection of process production data of the mesh belt layer of industrial and mining explosion-proof tires.

[0006] In an embodiment of the present invention, a method for collecting process production data of a mesh belt layer of an industrial and mining explosion-proof tire is provided, and the method comprises the following steps:

[0007] During the lamination process of the mesh belt layer of the explosion-proof tire for industrial and mining, monitoring data sequences in N dimensions are collected respectively, and the N dimensions include lamination temperature and lamination pressure;

[0008] For a monitoring data sequence under any dimension, according to the data difference between each data in the monitoring data sequence and the previous data, respectively obtain the suspected abnormality degree of each data in the monitoring data sequence, and according to the suspected abnormality degree of each data in the monitoring data sequence, obtain at least one suspected abnormal data;

[0009] For any suspected abnormal data, according to the periodic correlation characteristics of the suspected abnormal data and the periodic correlation characteristics of the corresponding data at the same sampling time in other dimensions, the lamination abnormality characteristic degree of the suspected abnormal data is obtained, and according to the lamination abnormality characteristic degree of the suspected abnormal data, the adaptive weight of the suspected abnormal data is obtained;

[0010] According to the adaptive weight of each of the suspected abnormal data, weighted recursive average filtering is performed on each of the suspected abnormal data to obtain the corresponding filtered data. The filtered data is used to replace the corresponding suspected abnormal data in the monitoring data sequence under the dimension to obtain the final data collection sequence under the dimension, and the final data collection sequence under each dimension is obtained to complete the data collection.

[0011] Preferably, obtaining the suspected abnormality degree of each data in the monitoring data sequence respectively according to the data difference between each data in the monitoring data sequence and the previous data includes:

[0012] For any data in the monitoring data sequence, obtain the previous data of the data, and obtain the initial suspected abnormality degree of the data according to the data difference between the previous data and the data;

[0013] Obtain a preset number of historical data before the data, obtain the suspected abnormality degree of each of the historical data, obtain the average suspected abnormality degree, and in the preset number of historical data, respectively calculate the absolute value of the difference between each two adjacent historical data corresponding to the filtered data to obtain the average absolute value of the difference;

[0014] The maximum data is obtained in the monitoring data sequence, and the product of the reciprocal of the maximum data, the average suspected abnormality degree and the average absolute value of the difference is normalized to obtain a corresponding first normalized value, and the initial suspected abnormality degree of the data is normalized to obtain a corresponding second normalized value, and the sum of the first normalized value and the second normalized value is used as the suspected abnormality degree of the data.

[0015] Preferably, obtaining the initial suspected abnormality degree of the data according to the data difference between the previous data and the data includes:

[0016] The absolute value of the difference between the filtered data corresponding to the previous data and the data is calculated, the product between the data and the absolute value of the difference is obtained, and the ratio between the square of the maximum data and the product is used as the initial suspected abnormality degree of the data.

[0017] Preferably, obtaining at least one suspected abnormal data according to the suspected abnormality degree of each data in the monitoring data sequence includes:

[0018] If the suspected abnormality degree of any data in the monitoring data sequence is greater than or equal to a preset abnormality degree threshold, the data is determined to be suspected abnormal data.

[0019] Preferably, obtaining the lamination abnormality characteristic degree of the suspected abnormal data according to the periodic correlation characteristics of the suspected abnormal data and the periodic correlation characteristics of the corresponding data at the same sampling time in other dimensions includes:

[0020] In a lamination cycle of the mesh belt layer of the industrial and mining explosion-proof tire, a data fluctuation characteristic value and a phase of the suspected abnormal data in the corresponding lamination cycle are obtained, historical data corresponding to the same phase as the suspected abnormal data are obtained as first data in each historical lamination cycle, a data fluctuation characteristic value of each of the first data is obtained, and a variance between the data fluctuation characteristic values ​​of all the first data and the data fluctuation characteristic values ​​of the suspected abnormal data is calculated;

[0021] All the first data and the suspected abnormal data are combined into a first data set, data belonging to the same sampling time as the suspected abnormal data is obtained in the monitoring data sequence under other dimensions as target data, historical target data corresponding to the same phase as the target data is obtained in each historical lamination cycle as second data, and all the second data and the target data are combined into a second data set;

[0022] Calculating the absolute value of the difference between every two adjacent data in the first data set to form a first difference absolute value sequence, calculating the absolute value of the difference between every two adjacent data in the second data set to form a second difference absolute value sequence, and obtaining the Pearson correlation coefficient of the first difference absolute value sequence and the second difference absolute value sequence;

[0023] The variance is negatively mapped using an exponential function with a natural constant as the base to obtain a first mapping value, and the Pearson correlation coefficient is negatively mapped using an exponential function with a natural constant as the base to obtain a second mapping value, and the sum of the first mapping value and the second mapping value is used as the degree of lamination abnormality characteristics of the suspected abnormal data.

[0024] Preferably, in the lamination cycle of the mesh belt layer of the industrial and mining explosion-proof tire, obtaining the data fluctuation characteristic value of the suspected abnormal data in the corresponding lamination cycle includes:

[0025] The suspected abnormal data are normalized to obtain a corresponding normalized result, and the sum of the normalized result and the suspected abnormal degree of the suspected abnormal data is used as the data fluctuation characteristic value of the suspected abnormal data in the corresponding lamination cycle.

[0026] Preferably, the step of obtaining the adaptive weight of the suspected abnormal data according to the lamination abnormality characteristic degree of the suspected abnormal data includes:

[0027] The lamination abnormality characteristic degree of the suspected abnormal data at a preset multiple is used as the adaptive weight of the suspected abnormal data.

[0028] Compared with the prior art, the embodiments of the present invention have the following beneficial effects:

[0029] In the lamination process of the mesh belt layer of the industrial and mining explosion-proof tire, the present invention collects monitoring data sequences in N dimensions respectively, and the N dimensions include pressing temperature and pressing pressure; for the monitoring data sequence in any dimension, according to the data difference between each data in the monitoring data sequence and the previous data, respectively obtain the suspected abnormal degree of each data in the monitoring data sequence, and obtain at least one suspected abnormal data according to the suspected abnormal degree of each data in the monitoring data sequence; for any suspected abnormal data, according to the periodic correlation characteristics of the suspected abnormal data and the periodic correlation characteristics of the corresponding data at the same sampling time in other dimensions, obtain the lamination abnormality characteristic degree of the suspected abnormal data, and obtain the adaptive weight of the suspected abnormal data according to the lamination abnormality characteristic degree of the suspected abnormal data; according to the adaptive weight of each suspected abnormal data, weighted recursive average filtering is performed on each suspected abnormal data respectively, and filtered data is obtained accordingly, and the filtered data is used to replace the corresponding suspected abnormal data in the monitoring data sequence under the dimension to obtain the final data collection sequence under the dimension, and the final data collection sequence under each dimension is obtained to complete data collection. Among them, the suspected abnormal data in the monitoring data sequence under each dimension are screened, and the adaptive weight is obtained when the weighted recursive average filtering is performed on each suspected abnormal data according to the periodic correlation characteristics of each suspected abnormal data, so as to assign a relatively lower weight to the abnormal data caused by friction. When the monitoring data sequence is filtered by weighted recursive average filtering, the abnormal data caused by friction in the lamination process can be filtered out, thereby ensuring that the process production data collected in the lamination process are valid data, reducing the error in judging whether the lamination process of the belt layer is normal in the subsequent filtered monitoring data. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0031] Figure 1 This is a method flow chart of a method for collecting production data of a mesh belt layer of an industrial and mining explosion-proof tire provided in Embodiment 1 of the present invention;

[0032] Figure 2 It is a schematic diagram of a bonding drum provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0033] Embodiments of the present disclosure are described in detail below, and examples of the embodiments are shown in the accompanying drawings. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present disclosure, but should not be understood as limiting the present disclosure.

[0034] It should be noted that the terms "first", "second", etc. in the specification of the present disclosure and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present disclosure described herein can be implemented in an order other than those illustrated or described herein. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present disclosure. Instead, they are merely examples of devices and methods consistent with some aspects of the present disclosure.

[0035] In order to illustrate the technical solution of the present invention, specific embodiments are provided below for illustration.

[0036] See also Figure 1 , is a method flow chart of a method for collecting data of a mesh belt layer process for industrial and mining explosion-proof tires provided in the first embodiment of the present invention, such as Figure 1 As shown, the method may include:

[0037] Step S101 , during the lamination process of the mesh belt layer of the explosion-proof tire for industrial and mining, monitoring data sequences in N dimensions are respectively collected, where the N dimensions include lamination temperature and lamination pressure.

[0038] In the lamination process of the mesh belt layer of the industrial and mining explosion-proof tire, the following Figure 2 The laminating drum shown in the figure, after attaching a circle of carcass rubber layer to the drum surface, applies the steel wire belt layer and presses it on the rubber layer using a pressure roller. Before pressing, an infrared temperature sensor is arranged to measure the pressing temperature of the belt layer at the position of the pressing line where the laminating drum and the pressure roller meet where the mesh belt layer contacts the carcass; at the same time, a pressure sensor is installed on the inner wall of the laminating drum at the position of the pressing line where the laminating drum and the pressure roller meet to measure the pressing pressure of the belt layer. The collection frequency of the pressing temperature and the pressing pressure is the same, preferably, set to once every 10 seconds, so that the monitoring data sequence in two dimensions can be obtained, that is, the pressing temperature sequence and the pressing pressure sequence.

[0039] Step S102, for a monitoring data sequence under any dimension, obtain the suspected abnormality degree of each data in the monitoring data sequence according to the data difference between each data in the monitoring data sequence and its previous data, and obtain at least one suspected abnormal data according to the suspected abnormality degree of each data in the monitoring data sequence.

[0040] The purpose of the present invention is to filter each monitoring data sequence using a weighted recursive average filtering algorithm after collecting the monitoring data sequence in each dimension, and use the filtered monitoring data sequence as the final collected monitoring data to detect whether there is an abnormal mutation point caused by lamination abnormality, and then judge whether the lamination process of the belt layer is normal. However, during the lamination process of the belt layer, friction may be generated between the pressure roller and the rubber surface during the extrusion process, and the abnormal data caused by friction is not easily filtered out by the traditional weighted recursive average filtering algorithm. Therefore, in the embodiment of the present invention, after collecting the monitoring data sequence in each dimension, all suspected abnormal data in each monitoring data sequence are found by analyzing the characteristic differences between the data abnormality generated by friction and the lamination abnormality data. The suspected abnormal data includes abnormal data caused by lamination abnormality, noise data and abnormal data caused by friction.

[0041] Since the weighted recursive average filtering algorithm obtains the filtered result by weighted averaging the current sampling value and the previous sampling values ​​according to a certain weight coefficient, which also includes the weight of the current sampling value, the larger the current sampling value is and the greater the difference between the current sampling value and the historical sampling value, the more prominent the filtered current sampling value is in the overall filtered data, and the more likely the current sampling value is to become suspected abnormal data. Considering that the lamination process of the belt layer is in a stable operating state when it starts working, it is assumed that the first 29 monitoring data collected do not need to be filtered and are treated as normal data, and the monitoring data collected later are filtered in real time.

[0042] After obtaining the monitoring data sequence under each dimension, for the monitoring data sequence under any dimension, first, for each data in the monitoring data sequence, obtain the previous data of each data, for example, the previous data of the i-th data is recorded as the i-1th data, and then, for any data, according to the data difference between the data and the previous data, obtain the initial suspected abnormality degree of the data, wherein the method for obtaining the initial suspected abnormality degree of the data is:

[0043] The absolute value of the difference between the filtered data corresponding to the previous data and the data is calculated, the product between the data and the absolute value of the difference is obtained, and the ratio between the square of the maximum data and the product is used as the initial suspected abnormality degree of the data.

[0044] In one embodiment, since the weighted recursive average filtering algorithm filters each data in turn, for the previous data, the previous data of each data can obtain the filtered data. Taking the i-th data as an example, the calculation expression of the initial suspected abnormality degree of the i-th data is:

[0045]

[0046] in, Indicates the initial suspected abnormality degree of the i-th data, represents the value of the i-th data, represents the filtered data corresponding to the i-1th data, | | represents the absolute value sign, and S represents the maximum data in the monitoring data sequence to which the i-th data belongs.

[0047] It should be noted that the greater the difference between the filtered data corresponding to the i-th data and the i-1-th data, the greater the probability that the i-th data is suspected abnormal data, and the greater the corresponding initial suspected abnormality degree.

[0048] Similarly, the initial suspected abnormality degree of each data in the monitoring data sequence under this dimension can be obtained. However, considering that the historical data also contains more prominent data, the suspected abnormality degree of the current data (the i-th data in the monitoring data sequence under this dimension) will decrease accordingly. Therefore, it is necessary to further combine the suspected abnormality degree of the historical data before the current data with the distribution of the historical data, and use the weighted recursive average filtering algorithm to filter the current data. The larger the cumulative sum of the suspected abnormality degrees of the historical data required, the greater the difference between adjacent data between historical data, and the higher the suspected abnormality degree of the corresponding current data.

[0049] In the embodiment of the present invention, after obtaining the initial suspected abnormality degree of each data in the monitoring data sequence under the dimension, the initial suspected abnormality degree of each data is optimized according to the data difference between each data in the monitoring data sequence and the previous data, and the suspected abnormality degree of each data in the monitoring data sequence is obtained respectively. The specific method is as follows:

[0050] For any data in the monitoring data sequence, obtain the previous data of the data, and obtain the initial suspected abnormality degree of the data according to the data difference between the previous data and the data;

[0051] Obtain a preset number of historical data before the data, obtain the suspected abnormality degree of each of the historical data, obtain the average suspected abnormality degree, and in the preset number of historical data, respectively calculate the absolute value of the difference between each two adjacent historical data corresponding to the filtered data to obtain the average absolute value of the difference;

[0052] The maximum data is obtained in the monitoring data sequence, and the product of the reciprocal of the maximum data, the average suspected abnormality degree and the average absolute value of the difference is normalized to obtain a corresponding first normalized value, and the initial suspected abnormality degree of the data is normalized to obtain a corresponding second normalized value, and the sum of the first normalized value and the second normalized value is used as the suspected abnormality degree of the data.

[0053] In one embodiment, taking the i-th data as an example, referring to the above-mentioned method for obtaining the initial suspected abnormality level, the initial suspected abnormality level of the i-th data can be obtained. The weighted recursive average filtering uses a sliding window of fixed length. Each time new data is input, the data in the sliding window is weighted averaged according to a certain weight to obtain the filtered data to achieve the purpose of noise reduction. All data before the i-th data are taken as historical data. The historical data required for the traditional weighted recursive average filtering algorithm is recorded as That is, the data in the sliding window is Then, the suspected abnormality degree of each historical data is obtained. Since the historical data required for the traditional weighted recursive average filtering algorithm may include normal data in a stable operating state when the lamination process of the belt layer just starts working, these data do not have suspected abnormality degrees, and only the initial suspected abnormality degree obtained by the above-mentioned method for obtaining the initial suspected abnormality degree exists. Therefore, the initial suspected abnormality degree of the normal data is used as the suspected abnormality degree, and then according to the suspected abnormality degree of each historical data of the i-th data, the calculation expression of the suspected abnormality degree of the i-th data is obtained as follows:

[0054]

[0055] in, Indicates the suspected abnormality of the i-th data, norm() represents the normalization function, Indicates the initial suspected abnormality degree of the i-th data, Indicates the suspected abnormality of the j-th historical data before the i-th data. represents the amount of historical data required for the i-th data, Indicates that the j-th historical data before the i-th data corresponds to the filtered data, It indicates that the j-1th historical data before the ith data corresponds to the filtered data, and S indicates the maximum data in the monitoring data sequence to which the ith data belongs.

[0056] It should be noted that Represents the mean of the suspected abnormality of all historical data. The larger the mean, the greater the suspected abnormality of the i-th data. It represents the absolute value of the difference between the two adjacent historical data before the i-th data corresponding to the filtered data. The absolute values ​​of the differences between the corresponding filtered data of all adjacent historical data are accumulated. The larger the accumulated value of the absolute value of the difference, the more abnormal the distribution of the historical data before the i-th data is, the greater the probability that the i-th data is abnormal data, and the greater the suspected abnormality of the corresponding i-th data; Used as the denominator for normalization.

[0057] Similarly, the suspected abnormality degree of each data in the monitoring data sequence under this dimension is obtained. The greater the suspected abnormality degree, the greater the probability that the data is abnormal data caused by non-laminated abnormality. Therefore, according to the suspected abnormality degree of each data in the monitoring data sequence, at least one suspected abnormal data is obtained, which specifically includes:

[0058] If the suspected abnormality degree of any data in the monitoring data sequence is greater than or equal to a preset abnormality degree threshold, the data is determined to be suspected abnormal data.

[0059] In one embodiment, the abnormality threshold is set to 0.6. When the suspected abnormality degree of any data is greater than or equal to 0.6, it means that the data is abnormal data caused by non-lamination abnormality, and it is marked as suspected abnormal data, so that all suspected abnormal data in the monitoring data sequence can be obtained.

[0060] At this point, all suspected abnormal data in the monitoring data sequence under each dimension can be detected.

[0061] Step S103, for any suspected abnormal data, the degree of lamination abnormality characteristics of the suspected abnormal data is obtained according to the periodic correlation characteristics of the suspected abnormal data and the periodic correlation characteristics of the corresponding data at the same sampling time in other dimensions, and the adaptive weight of the suspected abnormal data is obtained according to the degree of lamination abnormality characteristics of the suspected abnormal data.

[0062] During the lamination process, as the laminating drum rotates, the mesh belt layer is gradually compacted and laminated to the rubber layer on the surface of the laminating drum. This process is periodic as the laminating drum rotates. If there is a lamination anomaly that causes the belt layer at a certain point to be damaged, broken, or not firmly laminated, the same problem will occur the next time it passes through that point, which is reflected in the collected temperature and pressure data, showing highly similar abnormal fluctuations at the same phase of the next cycle. The thickness of the rubber layer is uneven. The contact between the wire mesh and the rubber layer will be different in thinner or thicker areas. The deformation of the rubber layer after being embedded in the belt layer will cause changes in friction. During the lamination process, the friction at the place where the friction was too large last time may be reduced due to lamination the next time it passes through that point. Therefore, the pressure and temperature changes caused by friction are likely to fluctuate at the same point in the next cycle.

[0063] Based on the above analysis, in the embodiment of the present invention, the lamination cycle in the lamination process is first analyzed. Specifically, the lamination cycle is obtained by the lamination drum rotation speed and the circumference of the side edge of the lamination drum. The calculation expression of the lamination cycle is:

[0064]

[0065] Wherein, T is the lamination cycle, L is the circumference of the side edge of the laminating drum, and v is the rotation speed of the laminating drum.

[0066] In order to better highlight the similarity between lamination anomalies and cycle historical phases, and the variability between friction anomalies and cycle historical phases, in the lamination cycle of the mesh belt layer of industrial and mining explosion-proof tires, the data fluctuation characteristic value of each suspected abnormal data in the lamination cycle is obtained. Taking a suspected abnormal data in any dimension as an example, the specific method for obtaining the data fluctuation characteristic value in the lamination cycle is as follows:

[0067] The suspected abnormal data are normalized to obtain a corresponding normalized result, and the sum of the normalized result and the suspected abnormal degree of the suspected abnormal data is used as the data fluctuation characteristic value of the suspected abnormal data in the corresponding lamination cycle.

[0068] In one embodiment, the calculation expression of the data fluctuation characteristic value of any suspected abnormal data in the corresponding lamination cycle is:

[0069]

[0070] in, represents the data fluctuation characteristic value of the i-th suspected abnormal data in the corresponding lamination cycle, norm() represents the normalization function, represents the value of the i-th suspected abnormal data, Indicates the suspected abnormality degree of the i-th data.

[0071] Similarly, the data fluctuation characteristic value of each suspected abnormal data in the monitoring data sequence under any dimension in the corresponding lamination cycle can be obtained. Furthermore, for any suspected abnormal data in the monitoring data sequence under any dimension, the degree of lamination abnormality characteristics of the suspected abnormal data is obtained in combination with the fluctuation characteristics of the historical data belonging to the same phase. At the same time, in order to further improve the accuracy, considering that friction is an important source of heat, the friction force is proportional to the change in temperature, and the friction force is proportional to the normal pressure, so the temperature and pressure changes caused by friction are also positively correlated. For friction anomalies, if in the same phase of different cycles, the temperature and pressure change amplitudes between different phases should also be highly similar. Therefore, in the embodiment of the present invention, the multi-dimensional correlation characteristics of temperature and pressure are also combined to analyze the degree of lamination abnormality characteristics of the suspected abnormal data. The specific analysis process is as follows:

[0072] In a lamination cycle of the mesh belt layer of the industrial and mining explosion-proof tire, a data fluctuation characteristic value and a phase of the suspected abnormal data in the corresponding lamination cycle are obtained, historical data corresponding to the same phase as the suspected abnormal data are obtained as first data in each historical lamination cycle, a data fluctuation characteristic value of each of the first data is obtained, and a variance between the data fluctuation characteristic values ​​of all the first data and the data fluctuation characteristic values ​​of the suspected abnormal data is calculated;

[0073] All the first data and the suspected abnormal data are combined into a first data set, data belonging to the same sampling time as the suspected abnormal data is obtained in the monitoring data sequence under other dimensions as target data, historical target data corresponding to the same phase as the target data is obtained in each historical lamination cycle as second data, and all the second data and the target data are combined into a second data set;

[0074] Calculating the absolute value of the difference between every two adjacent data in the first data set to form a first difference absolute value sequence, calculating the absolute value of the difference between every two adjacent data in the second data set to form a second difference absolute value sequence, and obtaining the Pearson correlation coefficient of the first difference absolute value sequence and the second difference absolute value sequence;

[0075] The variance is negatively mapped using an exponential function with a natural constant as the base to obtain a first mapping value, and the Pearson correlation coefficient is negatively mapped using an exponential function with a natural constant as the base to obtain a second mapping value, and the sum of the first mapping value and the second mapping value is used as the degree of lamination abnormality characteristics of the suspected abnormal data.

[0076] In one embodiment, the lamination cycle of the mesh belt layer of the industrial and mining explosion-proof tire is known, and then the historical data contained in multiple historical lamination cycles can be obtained according to the lamination cycle. Since the phase between the lamination cycles is the same, that is, the sampling time, therefore, according to the data fluctuation characteristic value and phase of the suspected abnormal data in the lamination cycle to which it belongs, combined with the characteristics of the historical data of the same phase in each historical lamination cycle, and the data correlation between different dimensions, the lamination abnormality characteristic degree of any suspected abnormal data in the monitoring data sequence under any dimension is analyzed. Taking the i-th suspected abnormal data as an example, according to the phase and data type of the i-th suspected abnormal data in the lamination cycle to which it belongs, the historical data corresponding to the same phase in each historical lamination cycle is obtained as the first data, and the data fluctuation characteristic value of each first data is obtained, and the variance between the data fluctuation characteristic values ​​of all first data and the data fluctuation characteristic value of the i-th suspected abnormal data is calculated. ,in Represents the existing lamination cycle, which is used to characterize the i-th suspected abnormal data.

[0077] All the first data and the i-th suspected abnormal data are combined into a first data set, and the absolute value of the difference between every two adjacent data in the first data set is calculated to form a first difference absolute value sequence Similarly, obtain the i-th data in other dimensions as the target data, and according to the phase and data type of the target data in the lamination cycle to which it belongs, obtain the historical data corresponding to the same phase in each historical lamination cycle as the second data, and form all the second data and the target data into a second data set, and calculate the absolute value of the difference between each two adjacent data in the second data set to form a second difference absolute value sequence , and according to the data differences of different dimensions in the same phase, the dimension correlation characteristics are analyzed, and then the dimension correlation characteristics and period correlation characteristics are combined to obtain the lamination anomaly characteristic degree of the i-th suspected abnormal data. The calculation expression of the lamination anomaly characteristic degree is:

[0078]

[0079] in, It indicates the degree of lamination anomaly characteristics of the i-th suspected abnormal data, exp() indicates an exponential function with a natural constant as the base, It represents the data fluctuation characteristic value of the i-th suspected abnormal data in the corresponding lamination cycle, express The data belonging to the same phase in a lamination cycle is the set of data fluctuation characteristic values ​​in the lamination cycle to which it belongs, represents the variance, represents the second difference absolute value sequence, represents the absolute value sequence of the first difference, represents the Pearson correlation coefficient between two series, and 1 represents a constant.

[0080] It should be noted that The smaller the variance value of the ith suspected abnormal data, the more similar it is to the historical data of the same phase as the ith suspected abnormal data, indicating that the ith suspected abnormal data has a stronger characteristic of lamination anomaly, and the greater the degree of lamination anomaly characteristic of the ith suspected abnormal data; Characterizes the multidimensional correlation characteristics under the same phase. The larger the value, the more similar the data change amplitude of the same phase between dimensions is, the stronger the characteristics of friction anomaly are, and the weaker the corresponding characteristics of lamination anomaly are. The control result value range is in the positive value [0, 1], so, The larger the value of is, the smaller the degree of lamination anomaly characteristics corresponding to the i-th suspected abnormal data is.

[0081] At this point, the lamination anomaly characteristic degree of each suspected abnormal data in the monitoring data sequence under each dimension can be obtained, which is used for assigning adaptive weights when filtering each suspected abnormal data in the monitoring data sequence based on the lamination anomaly characteristic degree.

[0082] Step S104, according to the adaptive weight of each suspected abnormal data, perform weighted recursive average filtering on each suspected abnormal data, obtain the filtered data accordingly, use the filtered data to replace the corresponding suspected abnormal data in the monitoring data sequence under the dimension, obtain the final data collection sequence under the dimension, obtain the final data collection sequence under each dimension, and complete the data collection.

[0083] In an embodiment of the present invention, after obtaining the adaptive weight of each suspected abnormal data in the monitoring data sequence under each dimension, weighted recursive average filtering is performed on each suspected abnormal data according to the adaptive weight of each suspected abnormal data, and corresponding to the filtered data, a higher weight is assigned to the data with a high degree of lamination abnormality characteristics to ensure that the fitted data retains the actual abnormality; for data with a low degree of lamination abnormality characteristics, a lower weight is assigned to data with a high degree of friction abnormality to ensure that the data anomaly is fitted and filtered out, and then the filtered data is used to replace the corresponding suspected abnormal data in the monitoring data sequence under the dimension to obtain the final data collection sequence under the dimension.

[0084] Among them, according to the adaptive weight of each suspected abnormal data, weighted recursive average filtering is performed on each suspected abnormal data to obtain the corresponding filtered data, including: taking the lamination abnormality feature degree of the suspected abnormal data with a preset multiple as the adaptive weight of the suspected abnormal data.

[0085] In one embodiment, the calculation expression of the adaptive weight is:

[0086]

[0087] in, represents the adaptive weight of the i-th suspected abnormal data, It indicates the degree of lamination abnormality characteristics of the i-th suspected abnormal data, and 2 indicates the preset multiple.

[0088] It should be noted that The value range of the control weight is [0,1]. The stronger the lamination anomaly feature, the higher the weight obtained by the corresponding data. In the process of weighted recursive average filter fitting, the proportion of self-value is higher, and the actual anomaly can be better retained. Conversely, the weaker the lamination anomaly feature, the lower the proportion of self-value is in the process of weighted recursive average filter fitting, and it is easier to filter out irrelevant anomalies.

[0089] After obtaining the adaptive weight of each suspected abnormal data in the monitoring data sequence under each dimension, weighted recursive average filtering is performed according to its adaptive weight to obtain the filtered data, wherein the calculation expression of weighted recursive average filtering is:

[0090]

[0091] in, represents the filtered data, represents the adaptive weight of suspected abnormal data, Represents suspected abnormal data, that is, the value before filtering, Indicates the number of data in the corresponding sliding window when performing weighted recursive average filtering. Indicates the number of data other than suspected abnormal data.

[0092] It should be noted that weighted recursive average filtering belongs to the prior art and will not be described in detail here. The focus of the present invention is how to obtain the adaptive weights of suspected abnormal data in the monitoring data sequence for use in weighted recursive average filtering.

[0093] At this point, the weighted recursive average filtering process of the monitoring data sequence in each dimension is completed, and the corresponding filtered monitoring data sequence is the final collected data, which is used to monitor whether the lamination process of the belt layer is normal.

[0094] In summary, in the lamination process of the mesh belt layer of the industrial and mining explosion-proof tire, the embodiment of the present invention collects monitoring data sequences in N dimensions respectively, and the N dimensions include pressing temperature and pressing pressure; for the monitoring data sequence in any dimension, according to the data difference between each data in the monitoring data sequence and the previous data, the suspected abnormality degree of each data in the monitoring data sequence is obtained respectively, and at least one suspected abnormal data is obtained according to the suspected abnormality degree of each data in the monitoring data sequence; for any suspected abnormal data, according to the periodic correlation characteristics of the suspected abnormal data and the periodic correlation characteristics of the corresponding data at the same sampling time in other dimensions, the lamination abnormality characteristic degree of the suspected abnormal data is obtained, and according to the lamination abnormality characteristic degree of the suspected abnormal data, the adaptive weight of the suspected abnormal data is obtained; according to the adaptive weight of each of the suspected abnormal data, each of the suspected abnormal data is weighted recursive average filtered respectively, and the corresponding filtered data is obtained, and the filtered data is used to replace the corresponding suspected abnormal data in the monitoring data sequence under the dimension to obtain the final data collection sequence under the dimension, and the final data collection sequence under each dimension is obtained to complete data collection. Among them, the suspected abnormal data in the monitoring data sequence under each dimension are screened, and the adaptive weight is obtained when the weighted recursive average filtering is performed on each suspected abnormal data according to the periodic correlation characteristics of each suspected abnormal data, so as to assign a relatively lower weight to the abnormal data caused by friction. When the monitoring data sequence is filtered by weighted recursive average filtering, the abnormal data caused by friction in the lamination process can be filtered out, thereby ensuring that the process production data collected in the lamination process are valid data, reducing the error in judging whether the lamination process of the belt layer is normal in the subsequent filtered monitoring data.

[0095] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention, and should be included in the protection scope of the present invention.

Claims

1. A method for collecting process production data of mesh belt layer of industrial and mining explosion-proof tires, characterized in that: The method comprises: During the lamination process of the mesh belt layer of the explosion-proof tire for industrial and mining, monitoring data sequences in N dimensions are collected respectively, and the N dimensions include lamination temperature and lamination pressure; For a monitoring data sequence under any dimension, according to the data difference between each data in the monitoring data sequence and the previous data, respectively obtain the suspected abnormality degree of each data in the monitoring data sequence, and according to the suspected abnormality degree of each data in the monitoring data sequence, obtain at least one suspected abnormal data; For any suspected abnormal data, according to the periodic correlation characteristics of the suspected abnormal data and the periodic correlation characteristics of the corresponding data at the same sampling time in other dimensions, the lamination abnormality characteristic degree of the suspected abnormal data is obtained, and according to the lamination abnormality characteristic degree of the suspected abnormal data, the adaptive weight of the suspected abnormal data is obtained; According to the adaptive weight of each of the suspected abnormal data, weighted recursive average filtering is performed on each of the suspected abnormal data to obtain filtered data, and the filtered data is used to replace the corresponding suspected abnormal data in the monitoring data sequence under the dimension to obtain the final data collection sequence under the dimension, and the final data collection sequence under each dimension is obtained to complete data collection; The step of obtaining the lamination abnormality characteristic degree of the suspected abnormal data according to the periodic correlation characteristics of the suspected abnormal data and the periodic correlation characteristics of the corresponding data at the same sampling time in other dimensions includes: In a lamination cycle of the mesh belt layer of the industrial and mining explosion-proof tire, a data fluctuation characteristic value and a phase of the suspected abnormal data in the corresponding lamination cycle are obtained, historical data corresponding to the same phase as the suspected abnormal data are obtained as first data in each historical lamination cycle, a data fluctuation characteristic value of each of the first data is obtained, and a variance between the data fluctuation characteristic values ​​of all the first data and the data fluctuation characteristic values ​​of the suspected abnormal data is calculated; All the first data and the suspected abnormal data are combined into a first data set, data belonging to the same sampling time as the suspected abnormal data is obtained in the monitoring data sequence under other dimensions as target data, historical target data corresponding to the same phase as the target data is obtained in each historical lamination cycle as second data, and all the second data and the target data are combined into a second data set; Calculating the absolute value of the difference between every two adjacent data in the first data set to form a first difference absolute value sequence, calculating the absolute value of the difference between every two adjacent data in the second data set to form a second difference absolute value sequence, and obtaining the Pearson correlation coefficient of the first difference absolute value sequence and the second difference absolute value sequence; The variance is negatively mapped using an exponential function with a natural constant as the base to obtain a first mapping value, and the Pearson correlation coefficient is negatively mapped using an exponential function with a natural constant as the base to obtain a second mapping value, and the sum of the first mapping value and the second mapping value is used as the degree of lamination abnormality characteristics of the suspected abnormal data.

2. The method for collecting process data of mesh belt layer of industrial and mining explosion-proof tire according to claim 1 is characterized in that: The obtaining, according to the data difference between each data in the monitoring data sequence and the previous data, the suspected abnormality degree of each data in the monitoring data sequence respectively comprises: For any data in the monitoring data sequence, obtain the previous data of the data, and obtain the initial suspected abnormality degree of the data according to the data difference between the previous data and the data; Obtain a preset number of historical data before the data, obtain the suspected abnormality degree of each of the historical data, obtain the average suspected abnormality degree, and in the preset number of historical data, respectively calculate the absolute value of the difference between each two adjacent historical data corresponding to the filtered data to obtain the average absolute value of the difference; The maximum data is obtained in the monitoring data sequence, and the product of the reciprocal of the maximum data, the average suspected abnormality degree and the average absolute value of the difference is normalized to obtain a corresponding first normalized value, and the initial suspected abnormality degree of the data is normalized to obtain a corresponding second normalized value, and the sum of the first normalized value and the second normalized value is used as the suspected abnormality degree of the data.

3. The method for collecting process data of mesh belt layer of industrial and mining explosion-proof tire according to claim 2 is characterized in that: The step of obtaining the initial suspected abnormality degree of the data according to the data difference between the previous data and the data includes: The absolute value of the difference between the filtered data corresponding to the previous data and the data is calculated, the product between the data and the absolute value of the difference is obtained, and the ratio between the square of the maximum data and the product is used as the initial suspected abnormality degree of the data.

4. The method for collecting process production data of mesh belt layer of industrial and mining explosion-proof tire according to claim 1, characterized in that: The step of acquiring at least one suspected abnormal data according to the suspected abnormality degree of each data in the monitoring data sequence comprises: If the suspected abnormality degree of any data in the monitoring data sequence is greater than or equal to a preset abnormality degree threshold, the data is determined to be suspected abnormal data.

5. The method for collecting process production data of mesh belt layer of industrial and mining explosion-proof tire according to claim 1 is characterized in that: In the lamination cycle of the mesh belt layer of the industrial and mining explosion-proof tire, obtaining the data fluctuation characteristic value of the suspected abnormal data in the corresponding lamination cycle includes: The suspected abnormal data are normalized to obtain a corresponding normalized result, and the sum of the normalized result and the suspected abnormal degree of the suspected abnormal data is used as the data fluctuation characteristic value of the suspected abnormal data in the corresponding lamination cycle.

6. The method for collecting process data of mesh belt layer of industrial and mining explosion-proof tire according to claim 1, characterized in that: The step of obtaining the adaptive weight of the suspected abnormal data according to the lamination abnormality feature degree of the suspected abnormal data comprises: The lamination abnormality characteristic degree of the suspected abnormal data at a preset multiple is used as the adaptive weight of the suspected abnormal data.

Images