Process parameter monitoring method for silica gel foaming pillow production

Abstract

The invention relates to the technical field of data processing, in particular to a technological parameter monitoring method for silica gel foaming pillow production, which comprises the following steps of: acquiring a pressure data sequence and a temperature data sequence in a foaming cavity, acquiring the numerical value normal degree of each pressure datum according to the time sequence neighborhood segment of each pressure datum, and calculating the numerical value normal degree of each pressure datum according to the numerical value normal degree; according to the difference condition between each piece of pressure data and the temperature data in the temperature data sequence in the foaming cavity at the corresponding sampling moment, the sampling accuracy degree of each piece of pressure data is obtained; in the process of clustering the pressure data, according to the sampling accuracy, obtaining a merging distance between each group of clusters in the clustering process, inputting the merging distance between each group of clusters into an agglomerated hierarchical clustering algorithm, and gradually merging the clusters on the pressure data in an iteration mode by the algorithm to obtain a hierarchical clustering tree after clustering is completed; according to the hierarchical clustering tree, abnormal data are recognized, and the accuracy of abnormal data recognition is improved.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and more specifically, to a process parameter monitoring method for producing silica gel foam pillows. Background Art

[0002] With the improvement of people's living standards, especially in terms of health and comfort, consumers have higher and higher requirements for pillow products. Silicone foam pillows are favored because of their good breathability, support and comfort. In the production process of silicone foam pillows, the pressure data in the foaming cavity is crucial to the quality of the final product. If there are abnormalities in the pressure data, it may cause problems such as uneven foaming and unqualified density, which will affect the comfort and service life of the silicone pillow. Therefore, by performing anomaly detection on the pressure data of the foaming process, it can ensure that the production process of silicone foam pillows is stable and controllable, thereby improving product quality and production efficiency. An existing method for anomaly detection of pressure data in the foaming process is an agglomerative hierarchical clustering algorithm. This method does not require a preset number of clusters and has strong flexibility. It can provide effective support in the anomaly detection of pressure data in the foaming process.

[0003] When the agglomerative hierarchical clustering algorithm processes the recursive aggregation of pressure data, the principle of full-link aggregation is to select the distance between the two pressure data with the farthest distance in the two clusters as the merging distance for aggregation; however, in the process of collecting pressure data, the electromagnetic waves generated by the electrical equipment (such as heaters, motors, etc.) in the foaming process will interfere with the sensor signal, resulting in noise data in the collected pressure data. If there is noise in the pressure data of the two clusters, the merging distance between the two clusters obtained during the aggregation of the two clusters is inaccurate, resulting in an inaccurate hierarchical clustering tree, which makes the abnormal results obtained when anomaly detection is performed based on the hierarchical clustering tree inaccurate. Summary of the invention

[0004] In order to solve the problem that noise data affects the reading of the sensor and distorts the pressure data, which in turn reduces the accuracy of the merged distance between two clusters during agglomerative hierarchical clustering and causes the entire hierarchical clustering tree to be inaccurate, the present invention proposes a process parameter monitoring method for silicone foam pillow production, which comprises the following steps: Collect the pressure data sequence and temperature data sequence in the foaming cavity; obtain the sampling accuracy of each pressure data: , Indicates the sampling accuracy of the rth pressure data; Indicates the normality of the value of the rth pressure data; Indicates the normality of the value of the rth temperature data in the temperature data sequence in the foaming cavity; Represents the value of the rth temperature data in the temperature data sequence in the foaming cavity; Represents the temperature reference value of the rth temperature data in the temperature data sequence in the foaming cavity; represents a hyperparameter; Indicates taking the absolute value; represents the linear normalization function; Get the merged distance between each group of clusters in the clustering process , Indicates the merge distance between the nth group of clusters; I indicates a preset quantity parameter; represents the mean of the sampling accuracy of the i-th target pressure data pair of the n-th cluster; represents the distance of the i-th target pressure data pair of the n-th cluster; The merged distance between each group of clusters is input into the agglomerative hierarchical clustering algorithm. The algorithm gradually merges clusters of the pressure data in an iterative manner to obtain a hierarchical clustering tree after clustering is completed; based on the hierarchical clustering tree, abnormal data is identified.

[0005] The innovation of the present invention lies in that when using an agglomerative hierarchical clustering algorithm, multiple pressure data pairs can be obtained from two clusters, and the distances between the multiple pressure data pairs are weighted according to the sampling accuracy of the pressure data, so as to give a lower weight to the noise in the aggregation process, reduce the influence of the noise data on the clustering results, improve the accuracy of the hierarchical clustering tree, and thus make the subsequent abnormality detection results more accurate; further, according to the difference between each pressure data and the temperature data in the temperature data sequence in the foaming cavity at the corresponding sampling moment, the normality of the numerical value of each pressure data is corrected, so that the sampling accuracy of each pressure data is more accurate, and the accuracy of subsequent weighting is improved.

[0006] Preferably, the normality of the value of the r-th pressure data includes: ； In the formula, Indicates the normality of the value of the rth pressure data; Represents the variance of all pressure data in the temporal neighborhood segment of the rth pressure data; Represents the number of all pressure data in the temporal neighborhood segment of the rth pressure data; represents the variance of all pressure data in the time series neighborhood segment of the nth pressure data in the time series neighborhood segment of the rth pressure data; represents the nth pressure data in the time series neighborhood segment of the rth pressure data; represents the n+1th pressure data in the time series neighborhood segment of the rth pressure data; represents a hyperparameter; Indicates taking the absolute value; represents an exponential function with a natural constant as base; represents the linear normalization function.

[0007] Can reflect whether the pressure data is noise.

[0008] Preferably, the acquisition of the temporal neighborhood segment of the pressure data includes: Preset a neighborhood parameter In the pressure data sequence in the foaming cavity, the data segment consisting of the k pressure data before the r-th pressure data and the k pressure data after the r-th pressure data is recorded as the time neighborhood segment of the r-th pressure data; if the number of pressure data before or after the r-th pressure data is less than k, only the existing pressure data is used, that is, the length of the time neighborhood segment is adjusted to the actually available number.

[0009] Preferably, obtaining the normality of the numerical value of the r-th temperature data includes: ； In the formula, Indicates the normality of the value of the r-th temperature data; Represents the variance of all temperature data in the time series neighborhood segment of the rth temperature data; Represents the number of all temperature data in the time series neighborhood segment of the rth temperature data; represents the variance of all temperature data in the time series neighborhood segment of the nth temperature data in the time series neighborhood segment of the rth temperature data; Represents the nth temperature data in the time series neighborhood segment of the rth temperature data; Represents the n+1th temperature data in the time series neighborhood segment of the rth temperature data; represents a hyperparameter; Indicates taking the absolute value; represents an exponential function with a natural constant as base; represents the linear normalization function.

[0010] Preferably, obtaining the temperature reference value of the rth temperature data includes: In the pressure data sequence in the foaming cavity, the sampling moments corresponding to all pressure data with the same value as the r-th pressure data are recorded as the target sampling moments; the average value of the temperature data at all target sampling moments in the temperature data sequence in the foaming cavity is recorded as the temperature reference value of the r-th temperature data.

[0011] Preferably, the acquisition of the i-th target pressure data pair of the n-th cluster includes: Any two clusters are recorded as a group cluster, recorded as the nth group cluster; a pressure data is obtained from each of the two clusters in the nth group cluster, and the two pressure data are used as a group of pressure data pairs to obtain all the pressure data pairs of the nth group cluster; A quantity parameter I is preset; if the number of pressure data pairs in the nth cluster is greater than , sort the pressure data pairs of the nth cluster according to the order of distances of the pressure data pairs of the nth cluster from large to small, and obtain the pressure data pair sequence of the nth cluster, and record the first I pressure data pairs in the pressure data pair sequence of the nth cluster as the target pressure data pairs of the nth cluster.

[0012] This facilitates subsequent analysis of the merged distance between each group of clusters based on the target pressure data pair.

[0013] Preferably, the The distance of the i-th target pressure data pair of the cluster is obtained by: The first The absolute value of the difference between the data in each target pressure data pair of the cluster is recorded as The distance between each target pressure data pair of the cluster.

[0014] This facilitates the subsequent acquisition of the merged distance between each group of clusters.

[0015] Preferably, the collecting of the pressure data sequence and the temperature data sequence in the foaming cavity includes: The pressure data and temperature data in the silicone foaming cavity are collected every 1 second using the pressure sensor and the temperature sensor for one hour. The two sequences composed of the pressure data and the temperature data in the foaming cavity at each sampling moment are recorded as the pressure data sequence in the foaming cavity and the temperature data sequence in the foaming cavity.

[0016] Preferably, the identifying abnormal data according to the hierarchical clustering clusters includes: Preset a merge distance threshold T; prune the hierarchical clustering tree according to the merge distance threshold T to obtain all pruned clusters; if the number of pressure data in any cluster after pruned is not greater than 2, all pressure data in the cluster are recorded as abnormal data; The specific process of pruning is as follows: if the merge distance of two clusters to be merged in the hierarchical clustering tree is greater than the merge distance threshold T, then the merge operation between the two clusters is pruned to obtain two independent clusters; if the merge distance of two clusters to be merged in the hierarchical clustering cluster is less than or equal to the merge distance threshold T, then the merge operation of the two clusters is not pruned, and the two clusters are still merged into one independent cluster.

[0017] Improved the accuracy of abnormal data identification.

[0018] The present invention has the following beneficial effects: the purpose of the present invention is to obtain multiple pressure data pairs from two clusters when using an agglomerative hierarchical clustering algorithm, and to weight the distances between the multiple pressure data pairs according to the sampling accuracy of the pressure data, so as to give a lower weight to noise in the aggregation process, reduce the influence of noise data on the clustering results, improve the accuracy of the hierarchical clustering tree, and thus make the subsequent abnormality detection results more accurate; further, according to the difference between each pressure data and the temperature data in the temperature data sequence in the foaming cavity at the corresponding sampling moment, the numerical normality of each pressure data is corrected, so that the sampling accuracy of each pressure data is more accurate, and the accuracy of subsequent weighting is improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] By reading the following detailed description with reference to the accompanying drawings, the above and other objects, features and advantages of the exemplary embodiments of the present invention will become readily understood. In the accompanying drawings, several embodiments of the present invention are shown in an exemplary and non-limiting manner, and the same or corresponding reference numerals represent the same or corresponding parts, wherein: Figure 1 The present invention is a flowchart of a method for monitoring process parameters of silicone foam pillows according to an embodiment of the present invention. DETAILED DESCRIPTION

[0020] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0021] The specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0022] See also Figure 1 , which shows a flowchart of a process parameter monitoring method for producing a silicone foam pillow provided by an embodiment of the present invention, the method comprising the following steps: S001. Collecting a pressure data sequence and a temperature data sequence in a foaming cavity.

[0023] In the embodiment of the present invention, every 1 second is a sampling moment, and the pressure data and temperature data in the silicone foaming cavity are collected by the pressure sensor and the temperature sensor each time for a total of one hour; the two sequences consisting of the pressure data and the temperature data in the foaming cavity at each sampling moment are recorded as the pressure data sequence in the foaming cavity and the temperature data sequence in the foaming cavity.

[0024] S002. Obtain a time series neighborhood segment of each pressure data, and based on the time series neighborhood segment, obtain a normality of a value of each pressure data.

[0025] It should be noted that, the closer the data value between each pressure data in the pressure data sequence in the foaming cavity and other pressure data in its time series neighborhood is, the more normal the value of the pressure data is; and the more stable the value change in the data segment composed of each pressure data and other pressure data in its time series neighborhood is, the closer the value between the pressure data and the pressure data in its time series neighborhood is, the greater the numerical credibility of the pressure data is, and the greater the degree of normality of its numerical value is.

[0026] In the embodiment of the present invention, according to the fluctuation difference between each pressure data point and the surrounding pressure data points in the pressure data sequence in the foaming cavity, the specific method for obtaining the normality of the value of each pressure data point is as follows: Preset a neighborhood parameter k , in the pressure data sequence in the foaming cavity, the data segment consisting of the k pressure data before the r-th pressure data and the k pressure data after the r-th pressure data is recorded as the temporal neighborhood segment of the r-th pressure data; When the number of pressure data before or after the r-th pressure data is less than k, only the existing data is used, that is, the length of the time series neighborhood segment is adjusted to the actually available number.

[0027] The specific formula for the normality of the value of the rth pressure data is: ; In the formula, Indicates the normality of the value of the rth pressure data; Represents the variance of all pressure data in the temporal neighborhood segment of the rth pressure data; Represents the number of all pressure data in the temporal neighborhood segment of the rth pressure data; represents the variance of all pressure data in the time series neighborhood segment of the nth pressure data in the time series neighborhood segment of the rth pressure data; represents the nth pressure data in the time series neighborhood segment of the rth pressure data; represents the n+1th pressure data in the time series neighborhood segment of the rth pressure data; Represents a hyperparameter, in order to prevent the denominator from being 0; Indicates taking the absolute value; represents an exponential function with a natural constant as base; represents the linear normalization function.

[0028] It should be noted that The smaller it is, the smaller the difference between the fluctuation of all pressure data in the time series neighborhood segment of the r-th pressure data and the fluctuation of pressure data in the time series neighborhood segment of other pressure data in its time series neighborhood segment is, that is, the more similar it is, the greater the normality of the value of the r-th pressure data is; It represents the difference between adjacent pressure data in the time series neighborhood segment of the r-th pressure data. The smaller the value is, the more stable the numerical change between the r-th pressure data point and the pressure data in its time series neighborhood is, and the greater the normality of the numerical value of the r-th pressure data is.

[0029] S003. Obtain the sampling accuracy of each pressure data according to the difference between each pressure data and the temperature data in the temperature data sequence in the foaming cavity at the corresponding sampling moment; in the process of clustering the pressure data, obtain the merging distance between each group of clusters in the clustering process according to the sampling accuracy.

[0030] It should be noted that since the normality of the values ​​cannot distinguish between the pressure data whose values ​​of the noise data are too large or too small, in the scenario survey, the temperature and pressure in the production of silicone foam pillows are closely related; at higher temperatures, the foaming speed is faster, more gas is released, and the pressure usually increases; at lower temperatures, the foaming speed is slower, the gas release is less, and the pressure is also lower; therefore, the above characteristics are not possessed by noise data; in this step, the numerical change characteristics of the pressure data and the temperature data will be analyzed, and the sampling accuracy of each pressure data point will be obtained in combination with the normality of the values ​​of the pressure data. Furthermore, in the specific analysis, the closer the connection between the pressure data and the temperature data, the less likely it is that the pressure data point belongs to the noise data point, and the greater its sampling accuracy.

[0031] In the embodiment of the present invention, according to the difference between each pressure data and the temperature data in the temperature data sequence in the foaming chamber at the corresponding sampling moment, the specific method for obtaining the sampling accuracy of each pressure data is: Based on the same principle as obtaining the numerical normality of the rth pressure data, the numerical normality of each temperature data in the temperature data sequence in the foaming cavity is obtained, wherein the numerical normality of each temperature data satisfies the relationship:

[0032] In the formula, Indicates the normality of the value of the r-th temperature data; Represents the variance of all temperature data in the time series neighborhood segment of the rth temperature data; Represents the number of all temperature data in the time series neighborhood segment of the rth temperature data; represents the variance of all temperature data in the time series neighborhood segment of the nth temperature data in the time series neighborhood segment of the rth temperature data; Represents the nth temperature data in the time series neighborhood segment of the rth temperature data; Represents the n+1th temperature data in the time series neighborhood segment of the rth temperature data; represents a hyperparameter; Indicates taking the absolute value; represents an exponential function with a natural constant as base; represents the linear normalization function.

[0033] It should be noted that the method for determining the time series neighborhood segment of the rth temperature data is the same as the method for determining the time series neighborhood segment of the rth pressure data.

[0034] In the embodiment of the present invention, in the pressure data sequence in the foaming cavity, the sampling moments corresponding to all the pressure data with the same value as the r-th pressure data are recorded as the target sampling moments; the average value of the temperature data at all the target sampling moments in the temperature data sequence in the foaming cavity is recorded as the temperature reference value of the r-th temperature data; The specific formula for obtaining the sampling accuracy of the rth pressure data is: ; In the formula, Indicates the sampling accuracy of the rth pressure data; Indicates the normality of the value of the rth pressure data; Indicates the normality of the value of the rth temperature data in the temperature data sequence in the foaming cavity; Represents the value of the rth temperature data in the temperature data sequence in the foaming cavity; Represents the temperature reference value of the rth temperature data in the temperature data sequence in the foaming cavity; Represents a hyperparameter, in order to prevent the denominator from being 0; Indicates taking the absolute value; represents the linear normalization function.

[0035] It should be noted that The larger it is, the less likely the r-th pressure data is to be affected by noise based on the numerical performance analysis of the pressure data in its time series neighborhood, and the greater its sampling accuracy; The smaller it is, the more consistent the numerical performance of the temperature data in the temperature data sequence in the foaming cavity at the sampling moment corresponding to the r-th pressure data is in their respective time series neighborhood data segments, which means that the connection between the temperature data in the temperature data sequence in the foaming cavity at the sampling moment corresponding to the r-th pressure data is closer, which means that the possibility that the r-th pressure data belongs to noise data is smaller, and its sampling accuracy is greater.

[0036] It should be further explained that the change of temperature data may lag behind that of pressure data, so the closeness of the relationship quantified based on the normality of the two values ​​may not be actually closer; therefore, we continue to analyze the difference between the temperature data at the current sampling time and its corresponding temperature reference value; that is, in the formula The smaller it is, the smaller the difference between the temperature data in the temperature data sequence in the foaming cavity at the sampling moment corresponding to the r-th pressure data and the temperature reference value is, and the greater the possibility that the r-th pressure data belongs to real data rather than noise data. It can also be explained that the more consistent the numerical performance of the temperature data in the temperature data sequence in the foaming cavity at the sampling moment corresponding to the r-th pressure data in their respective time series neighborhood data segments, the greater the credibility, so the sampling accuracy of the r-th pressure data will be greater.

[0037] It should be noted that in the process of clustering pressure data using the agglomerative hierarchical clustering algorithm, each pressure data point is first regarded as a separate cluster, and then the merged distance between each two clusters (the absolute value of the difference between the pressure data of the two clusters) is calculated to merge the clusters, thereby completing the clustering of all pressure data and obtaining a hierarchical clustering tree after clustering. In order to improve the accuracy of pressure data clustering, it is necessary to use the sampling accuracy of the pressure data to correct it when calculating the merged distance between the two clusters in the clustering process. It is known that the greater the sampling accuracy between the pressure data of the two clusters, the more accurate the absolute value of the difference between the pressure data of the two clusters, and the more accurate the calculated merged distance of the two clusters. Therefore, the merged distance between each group of clusters is obtained to complete the optimization of the agglomerative hierarchical clustering algorithm.

[0038] In the embodiment of the present invention, any two clusters are recorded as a group cluster; a pressure data is obtained from each of the two clusters in the group cluster, and the two pressure data are used as a group of pressure data pairs to obtain all pressure data pairs of each group cluster; the absolute value of the difference between the data in each pressure data pair of each group cluster is recorded as the distance of each pressure data pair of each group cluster; Preset a quantity parameter I , if the number of pressure data pairs in the nth cluster is less than or equal to 1, then the maximum value of the distances of all pressure data pairs in the nth cluster is recorded as the merged distance between the nth clusters; If the number of the pressure data pairs of the nth cluster is greater than 1, the pressure data pairs of the nth cluster are sorted according to the order of the distances of the pressure data pairs of the nth cluster from large to small to obtain a pressure data pair sequence of the nth cluster, and the first 1 pressure data pairs in the pressure data pair sequence of the nth cluster are recorded as target pressure data pairs of the nth cluster; According to the target pressure data pair of the nth cluster, the calculation formula for obtaining the merged distance between the nth clusters is: ; In the formula, Indicates the merge distance between the nth group of clusters; I indicates a preset quantity parameter; represents the mean of the sampling accuracy of the i-th target pressure data pair of the n-th cluster; Represents the distance of the i-th target pressure data pair of the n-th cluster.

[0039] S004. Input the merged distance between each group of clusters into the agglomerative hierarchical clustering algorithm. The algorithm gradually merges clusters of the pressure data in an iterative manner to obtain a hierarchical clustering tree after clustering is completed; and identify abnormal data based on the hierarchical clustering tree.

[0040] In the embodiment of the present invention, the preset merge distance threshold T=1.1, in other embodiments, the implementer may preset the value of T according to the specific implementation situation; The merge distance between each group of clusters is input into the agglomerative hierarchical clustering algorithm. The algorithm gradually merges clusters of the pressure data in an iterative manner to obtain a hierarchical clustering tree after clustering. In each iteration, the cluster with the smallest merge distance is selected for merging, and the merge distance matrix is ​​updated.

[0041] It should be noted that the hierarchical clustering tree reflects the step-by-step merging process of clusters.

[0042] The hierarchical clustering tree is pruned according to the merge distance threshold T to obtain all the pruned clusters; the specific process of pruning is as follows: if the merge distance of the two clusters to be merged in the hierarchical clustering tree is greater than the merge distance threshold T, then the merge operation of the two clusters is pruned to obtain two independent clusters; if the merge distance of the two clusters to be merged in the hierarchical clustering tree is less than or equal to the merge distance threshold T, then the merge operation of the two clusters is not pruned, and the two clusters are still merged into one independent cluster; If the number of pressure data in any cluster after pruning is not greater than 2, all pressure data in the cluster are recorded as abnormal data.

[0043] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A process parameter monitoring method for the production of silica gel foam pillows, characterized in that: include: Collecting pressure data sequence and temperature data sequence in the foaming cavity; Get the sampling accuracy of each pressure data: , Indicates the sampling accuracy of the rth pressure data; Indicates the normality of the value of the rth pressure data; Indicates the normality of the value of the rth temperature data in the temperature data sequence in the foaming cavity; Represents the value of the rth temperature data in the temperature data sequence in the foaming cavity; Represents the temperature reference value of the rth temperature data in the temperature data sequence in the foaming cavity; represents a hyperparameter; Indicates taking the absolute value; represents the linear normalization function; Get the merged distance between each group of clusters in the clustering process , Indicates the merge distance between the nth group of clusters; I indicates a preset quantity parameter; represents the mean of the sampling accuracy of the i-th target pressure data pair of the n-th cluster; represents the distance of the i-th target pressure data pair of the n-th cluster; The merged distance between each group of clusters is input into the agglomerative hierarchical clustering algorithm. The algorithm gradually merges clusters of the pressure data in an iterative manner to obtain a hierarchical clustering tree after clustering is completed; based on the hierarchical clustering tree, abnormal data is identified.

2. A process parameter monitoring method for silica gel foam pillow production according to claim 1, characterized in that: The normality of the value of the r-th pressure data includes: ； In the formula, Indicates the normality of the value of the rth pressure data; Represents the variance of all pressure data in the time series neighborhood segment of the rth pressure data; Represents the number of all pressure data in the temporal neighborhood segment of the rth pressure data; represents the variance of all pressure data in the time series neighborhood segment of the nth pressure data in the time series neighborhood segment of the rth pressure data; represents the nth pressure data in the time series neighborhood segment of the rth pressure data; represents the n+1th pressure data in the time series neighborhood segment of the rth pressure data; represents a hyperparameter; Indicates taking the absolute value; represents an exponential function with a natural constant as base; represents the linear normalization function.

3. A process parameter monitoring method for silica gel foam pillow production according to claim 2, characterized in that: The acquisition of the temporal neighborhood segment of the pressure data includes: Preset a neighborhood parameter In the pressure data sequence in the foaming cavity, the data segment consisting of the k pressure data before the r-th pressure data and the k pressure data after the r-th pressure data is recorded as the time neighborhood segment of the r-th pressure data; if the number of pressure data before or after the r-th pressure data is less than k, only the existing pressure data is used, that is, the length of the time neighborhood segment is adjusted to the actually available number.

4. A process parameter monitoring method for silica gel foam pillow production according to claim 1, characterized in that: The acquisition of the normality of the numerical value of the r-th temperature data includes: ； In the formula, Indicates the normality of the value of the r-th temperature data; Represents the variance of all temperature data in the time series neighborhood segment of the rth temperature data; Represents the number of all temperature data in the time series neighborhood segment of the rth temperature data; represents the variance of all temperature data in the time series neighborhood segment of the nth temperature data in the time series neighborhood segment of the rth temperature data; Represents the nth temperature data in the time series neighborhood segment of the rth temperature data; Represents the n+1th temperature data in the time series neighborhood segment of the rth temperature data; represents a hyperparameter; Indicates taking the absolute value; represents an exponential function with a natural constant as base; represents the linear normalization function.

5. The process parameter monitoring method for the production of silicone foam pillows according to claim 1, characterized in that: The acquisition of the temperature reference value of the rth temperature data includes: In the pressure data sequence in the foaming cavity, the sampling moments corresponding to all pressure data with the same value as the r-th pressure data are recorded as the target sampling moments; the average value of the temperature data at all target sampling moments in the temperature data sequence in the foaming cavity is recorded as the temperature reference value of the r-th temperature data.

6. A process parameter monitoring method for silica gel foam pillow production according to claim 1, characterized in that: The acquisition of the i-th target pressure data pair of the n-th cluster includes: Any two clusters are recorded as a group cluster, recorded as the nth group cluster; a pressure data is obtained from each of the two clusters in the nth group cluster, and the two pressure data are used as a group of pressure data pairs to obtain all the pressure data pairs of the nth group cluster; A quantity parameter I is preset; if the number of pressure data pairs in the nth cluster is greater than , sort the pressure data pairs of the nth cluster according to the order of their distances from large to small, and obtain the pressure data pair sequence of the nth cluster, and record the first I pressure data pairs in the pressure data pair sequence of the nth cluster as the target pressure data pairs of the nth cluster.

7. The process parameter monitoring method for producing silicone foam pillows according to claim 1, characterized in that: The said The distance of the i-th target pressure data pair of the cluster is obtained by: The first The absolute value of the difference between the data in each target pressure data pair of the cluster is recorded as The distance between each target pressure data pair of the cluster.

8. The process parameter monitoring method for silica gel foam pillow production according to claim 1 is characterized in that: The collecting of the pressure data sequence and the temperature data sequence in the foaming cavity comprises: The pressure data and temperature data in the silicone foaming cavity are collected every 1 second using the pressure sensor and the temperature sensor for one hour. The two sequences composed of the pressure data and the temperature data in the foaming cavity at each sampling moment are recorded as the pressure data sequence in the foaming cavity and the temperature data sequence in the foaming cavity.

9. A process parameter monitoring method for silica gel foam pillow production according to claim 1, characterized in that: The step of identifying abnormal data according to the hierarchical clustering tree includes: Preset a merge distance threshold T; prune the hierarchical clustering tree according to the merge distance threshold T to obtain all pruned clusters; if the number of pressure data in any cluster after pruned is not greater than 2, all pressure data in the cluster are recorded as abnormal data; The specific process of pruning is as follows: if the merge distance of two clusters to be merged in the hierarchical clustering tree is greater than the merge distance threshold T, then the merge operation between the two clusters is pruned to obtain two independent clusters; if the merge distance of two clusters to be merged in the hierarchical clustering cluster is less than or equal to the merge distance threshold T, then the merge operation of the two clusters is not pruned, and the two clusters are still merged into one independent cluster.

Images