A cable production quality data management system based on IoT big data

Abstract

The present invention relates to the technical field of product quality management, and more specifically, to a cable production quality data management system based on Internet of Things big data, the system comprising: an acquisition module for acquiring a production quality sequence; a sample division module for sequentially screening target data points in the production quality sequence, taking the target data points screened each time and their neighboring data points as a cluster cluster until all data points are divided to obtain multiple cluster clusters; an anomaly detection module for taking the cluster center of each cluster cluster as the initial cluster center of the ISODATA algorithm, clustering the production quality sequence to obtain a clustering result, and managing the cable production quality based on the result of anomaly detection on the production quality sequence. The present invention can improve the accuracy of cable production quality management.

Description

Technical Field

[0001] The present invention relates to the technical field of product quality management, and more specifically, to a cable production quality data management system based on Internet of Things big data. Background Art

[0002] As a core component of modern power and information transmission, the production process of cables involves a variety of complex processes and materials. To ensure the stability of the production process and the quality of cable products, it is essential to monitor relevant data. With the development of Internet of Things technology, production data collected by sensors can be uploaded to the cloud for centralized management and analysis. Therefore, identifying abnormalities in production data is crucial to ensure the stability of cable production and product consistency.

[0003] In the related technology, for example, a patent application document with publication number CN117893103A discloses a cross-linked cable product quality management and control system based on a cable production line. The system includes: calculating parameter indicators of each link, and judging the parameter indicators of each link by setting thresholds. If the parameter indicators of the process ahead in the sequence do not meet the standards, they are directly returned for repair, and then a weighted calculation of the cable quality is performed using multiple parameter indicators to obtain the cable quality reliability.

[0004] However, the above scheme only considers the impact of single data on cable quality, but does not consider the impact of interaction between different data on cable production quality. This may reduce the accuracy of the reliability of the determined cable quality to a certain extent, thus making it impossible to accurately manage cable production quality. Summary of the invention

[0005] In order to solve the problem of low accuracy of cable production quality management due to the inability to accurately identify abnormal data in the cable production process, the present invention provides a cable production quality data management system based on Internet of Things big data. The system includes:

[0006] The acquisition module is used to collect data of various dimensions in the production process at each sampling moment to obtain the production quality sequence;

[0007] The sample partitioning module is used to sequentially screen the target data points in the production quality sequence, and take the target data point and its neighboring data points screened each time as a cluster until all data points are partitioned to obtain multiple clusters, wherein the target data point is the data point with the largest number of neighboring data points in the production quality sequence;

[0008] The anomaly detection module is used to use the cluster center of each cluster cluster as the initial cluster center of the ISODATA algorithm, cluster the production quality sequence to obtain a clustering result, and manage the cable production quality based on the result of anomaly detection on the production quality sequence based on the clustering result;

[0009] Among them, the neighboring data points are the data of each dimension in the production quality sequence, and the distance between them and the corresponding dimension data of the target data point is less than or equal to the neighboring distance of each dimension data; the neighboring distance is the product of the significance of the corresponding dimension data and the preset initial distance; the significance is the average difference between the corresponding dimension data and the average value of the dimension data.

[0010] The present invention can adaptively adjust the initial clustering center when the ISODATA algorithm clusters the production quality sequence based on the distribution characteristics of the data, so that the position of the initial clustering center conforms to the distribution characteristics of each dimensional data, which can improve the clustering effect. Therefore, based on the clustering results with higher accuracy, it can accurately judge whether the production quality sequence has an abnormality, and then based on the detection results, the cable production quality can be effectively managed.

[0011] According to a cable production quality data management system based on IoT big data provided by the present invention, the cable production quality is managed based on the result obtained by performing abnormal detection on the production quality sequence based on the clustering result, including:

[0012] The clusters in the clustering results are used as the target clusters, and the cumulative sum of the average distances between the data points in all the target clusters and the cluster centers of the corresponding target clusters is calculated. The normalized value of the product of the number of target clusters and the cumulative sum is used as the abnormality degree of the production quality sequence;

[0013] Get the preset threshold value, and when the abnormality level is greater than or equal to the preset threshold value, an alarm is issued to manage the cable production quality.

[0014] The present invention forms a quantifiable abnormality degree by limiting the range of the abnormality degree to 0-1, which can intuitively reflect the abnormal situation of the production quality sequence and reduce the amount of calculation when determining the abnormality detection result.

[0015] According to a cable production quality data management system based on IoT big data provided by the present invention, the abnormality degree of the production quality sequence satisfies the following relationship:

[0016] ;

[0017] In the formula, is the degree of abnormality in the production quality sequence; is the number of target clusters; The first in the production quality sequence The average distance between all data points in a target cluster and the cluster center of the target cluster; The natural constant An exponential function with base ; is the summation symbol.

[0018] According to a cable production quality data management system based on Internet of Things big data provided by the present invention, the significance satisfies the following relationship:

[0019] ;

[0020] In the formula, The first in the production quality sequence The significance of the data in these dimensions; The first in the production quality sequence The first dimension The value of the data; The first in the production quality sequence The average value of the dimension data; is the normalization function; for the summation symbol; is the absolute value symbol.

[0021] According to a cable production quality data management system based on Internet of Things big data provided by the present invention, a cluster center acquisition method includes:

[0022] The average value of all data points in any cluster in any dimension is taken as the value of the cluster center in any dimension;

[0023] Calculate the values ​​of the cluster center of any cluster in all dimensions and obtain the position of the cluster center of any cluster.

[0024] The location of the cluster center determined by the present invention can well reflect the distribution of data in the corresponding cluster.

[0025] According to a cable production quality data management system based on Internet of Things big data provided by the present invention, the system also includes:

[0026] A quantity threshold is preset. When the number of neighboring data points of any target data point is less than or equal to the quantity threshold, any target data point and its neighboring data points are divided into nearest neighbor clusters respectively.

[0027] The present invention can divide data with similar data distribution into a cluster, thereby ensuring the accuracy of the production quality sequence division.

[0028] According to a cable production quality data management system based on Internet of Things big data provided by the present invention, data of multiple dimensions in the production process are collected at each sampling moment, including:

[0029] At each sampling moment, readings of various sensors electrically connected to the cable extruder are collected to obtain data of various dimensions, and the dimensions correspond one to one with the data types collected by the sensors.

[0030] The present invention can improve the accuracy of the determined cable production quality to a certain extent by collecting various types of data as the basis for data analysis.

[0031] According to a cable production quality data management system based on Internet of Things big data provided by the present invention, a method for obtaining clustering results includes:

[0032] When the number of iterations for clustering the production quality sequence is greater than the preset number of iterations or the positions of all cluster centers tend to converge, the iteration is stopped and the clustering result is obtained.

[0033] The present invention has the following technical effects:

[0034] 1. The present invention uses the cluster center of each cluster determined by the sample division module as the initial cluster center of the ISODATA algorithm, so that the selection of the initial cluster center can be close to the actual distribution of data points in the production quality sequence. In the process of iterative clustering of the production quality sequence using the algorithm, it can better adapt to the influence of the interaction between different types of data, so that the clustering results can be obtained quickly and accurately.

[0035] 2. When the sample partitioning module of the present invention determines each cluster, the selection of neighbor data points depends on the neighbor distance of data of different dimensions, and the determination of the neighbor distance is based on the significance of the corresponding dimensional data, so that the determined neighbor data points can conform to the distribution of the corresponding dimensional data, so that data that are easy to cluster together can be divided into a cluster, ensuring the accuracy of the partitioning results.

[0036] 3. The clustering results obtained by the present invention have high accuracy, so that based on the clustering results with high accuracy, anomalies in the production quality sequence can be accurately detected, thereby achieving effective management of the cable production quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] 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:

[0038] Figure 1 It is a structural schematic diagram of a cable production quality data management system based on Internet of Things big data in an embodiment of the present invention. DETAILED DESCRIPTION

[0039] 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.

[0040] The specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0041] Reference Figure 1 , a cable production quality data management system 100 based on Internet of Things big data, including modules 110-130, specifically as follows:

[0042] The acquisition module 110 is used to collect data of multiple dimensions in the production process at each sampling moment to obtain a production quality sequence.

[0043] The production quality sequence refers to a sequence of data such as temperature, pressure, flow rate and speed collected at each sampling time during the production of the cable using the three-layer co-extrusion process. This embodiment does not specifically limit the length of the production quality sequence.

[0044] In an exemplary embodiment of the present invention, the collection of data in each dimension can be achieved by the following steps:

[0045] At each sampling moment, readings of various sensors electrically connected to the cable extruder are collected to obtain data of various dimensions, and the dimensions correspond one to one with the data types collected by the sensors.

[0046] The dimension refers to the type of data collected by the sensor. For example, when a temperature sensor is used to collect temperature data in the cable production process, temperature is used as a dimension.

[0047] Specifically, corresponding sensors, such as temperature sensors, pressure sensors, and speed sensors, can be arranged at appropriate locations of the cable extruder, and data such as temperature, pressure, and speed during the cable production process can be collected at a certain collection frequency, such as once per second, so as to obtain data of multiple dimensions at each sampling moment. This embodiment does not specifically limit the collection frequency.

[0048] The sample partitioning module 120 is used to sequentially screen the target data points in the production quality sequence, and treat the target data point and its neighboring data points screened each time as a cluster until all data points are partitioned to obtain multiple clusters, wherein the target data point is the data point with the largest number of neighboring data points in the production quality sequence.

[0049] It should be noted that when the number of neighboring data points of any data point is large, it means that the data point and its neighboring data points are more likely to be gathered in one set. Therefore, the data point and its neighboring data points can be divided into one set, thereby realizing the division of the production quality sequence.

[0050] It should be further explained that due to the different distribution of data of different dimensions, if the same distance standard is used to determine the neighboring data points of each dimension, the number of determined neighboring data points may be small, and the production quality sequence cannot be effectively divided. Therefore, when measuring the number of neighboring data points of each dimension, it is necessary to determine the neighboring distance in combination with the distribution of the corresponding dimension data, so as to ensure the accuracy of the determined neighboring data points of each dimension.

[0051] Specifically, the average difference between each dimension data in the production quality sequence and the average value of the dimension data can be calculated to obtain the significance of each dimension data, thereby measuring the distribution of the corresponding dimension data. For example, when the average difference between any dimension data and the average value of the dimension data is large, the significance of the dimension data is large, indicating that the distribution of the dimension data is wide and it is difficult to form a tight group. In this case, a larger neighbor distance needs to be set so that the set composed of any data point in the dimension data and its neighboring data points can reflect the true distribution of the dimension data.

[0052] Among them, the significance of any dimension data in the production quality sequence satisfies the following relationship:

[0053] ;

[0054] In the formula, The first in the production quality sequence The significance of the data in these dimensions; The first in the production quality sequence The first dimension The value of the data; The first in the production quality sequence The average value of the dimension data; is the normalization function; for the summation symbol; is the absolute value symbol.

[0055] Optionally, normalizing the average difference between each dimension data and the average value of the dimension data can limit the significance of each dimension data to 0-1, thereby reducing the amount of calculation.

[0056] Furthermore, after determining the significance of each dimension of data in the production quality sequence, the neighbor distance of each dimension of data may be calculated, thereby determining the number of neighboring data points of each data point in each dimension of data.

[0057] Specifically, the neighbor distance of any dimension of data in the production quality sequence satisfies the following relationship:

[0058] ;

[0059] In the formula, The first in the production quality sequence The nearest neighbor distance of the dimensional data; To preset the initial distance, in this embodiment ; The first in the production quality sequence The significance of the data in these dimensions; is the ceiling function.

[0060] Furthermore, the number of neighboring data points of each data point in the production quality sequence can be determined according to the neighboring distance of each dimension of data in the sequence, so as to filter the target data points in the sequence in turn and obtain multiple clusters. The specific process is:

[0061] First, any data point in the production quality sequence is taken as a reference data point, and the remaining data points are taken as data points to be confirmed. When the distance between the data of each dimension in any data point to be confirmed and the corresponding data of the reference data point is less than or equal to the neighbor distance of the corresponding dimension, the data point to be confirmed is taken as the neighbor data point of the reference data point.

[0062] Then, the number of neighboring data points of each data point in the sequence can be counted in turn. When the number of neighboring data points of any data point in the sequence is the largest, the data point is used as the target data point, and the target data point and its neighboring data points are removed from the sequence to obtain the first cluster.

[0063] Next, the screening process of the target data point is repeated to obtain a cluster consisting of the target data point obtained in each screening and its neighboring data points, until all data points are divided and multiple clusters are obtained.

[0064] In an exemplary embodiment of the present invention, the process of dividing the samples of the production quality sequence further includes:

[0065] A quantity threshold is preset. When the number of neighboring data points of any target data point is less than or equal to the quantity threshold, any target data point and its neighboring data points are divided into nearest neighbor clusters respectively.

[0066] Optionally, the quantity threshold can be set to 4. When the number of neighboring data points of any target data point is less than or equal to 4, the target data point and the neighboring data points of the target data point are respectively divided into the nearest neighbor clustering cluster. The nearest neighbor clustering cluster refers to the clustering cluster with the shortest distance to the corresponding data point.

[0067] Optionally, the distance involved in the present invention may be Euclidean distance or Manhattan distance. Of course, a suitable distance may be selected according to specific circumstances. This embodiment does not specifically limit the distance type used.

[0068] The anomaly detection module 130 is used to cluster the production quality sequence using the cluster center of each cluster cluster as the initial cluster center of the ISODATA algorithm to obtain clustering results, so as to manage the cable production quality based on the results of anomaly detection on the production quality sequence.

[0069] It should be noted that the ISODATA (Iterative Selforganizing Data Analysis Techniques Algorithm) algorithm is a clustering method based on self-organizing characteristics and iterative optimization. The algorithm optimizes the clustering results by randomly presetting the cluster center and continuously iterating until the position of the cluster center no longer changes or the preset number of iterations is reached. However, when the algorithm randomly presets the cluster center, it cannot take into account the distribution of the data, resulting in poor clustering effect. Therefore, the present invention improves the process of presetting the cluster center of the algorithm, so that the initial cluster center can be adaptively selected according to the data distribution, thereby ensuring the accuracy of the clustering results.

[0070] The present invention only improves the process of presetting the clustering center of the algorithm, and does not improve other steps of the algorithm during clustering.

[0071] In an exemplary embodiment of the present invention, the cluster center of each cluster determined by module 102 may be determined by the following steps:

[0072] The average value of all data points in any cluster in any dimension is taken as the value of the cluster center of the cluster in any dimension; the values ​​of the cluster center of any cluster in all dimensions are calculated to obtain the position of the cluster center of any cluster.

[0073] Furthermore, the cluster center of each cluster determined based on module 120 can be used as the initial cluster center of the ISODATA algorithm, and then the production quality sequence is clustered to obtain a clustering result. It should be noted that the process of clustering data using the algorithm is a prior art and will not be described in detail in this embodiment.

[0074] In an exemplary embodiment of the present invention, the clustering result can be determined by the following steps:

[0075] When the number of iterations for clustering the production quality sequence is greater than the preset number of iterations or the positions of all cluster centers tend to converge, the iteration is stopped and the clustering result is obtained.

[0076] Optionally, the preset number of iterations can be set to 100 according to empirical values, and when the number of iterations when clustering the production quality sequence using the algorithm is greater than 100, the iteration process is terminated to obtain a clustering result. This embodiment does not specifically limit the preset number of iterations.

[0077] Optionally, when the algorithm is used to iteratively cluster the production quality sequence, if the positions of the cluster centers of the cluster clusters tend to converge, the iterative process is stopped to obtain the clustering result of clustering the sequence.

[0078] Furthermore, after the clustering results are determined, anomaly detection can be performed on the production quality sequence based on the clustering results, so that the cable production quality can be managed according to the obtained detection results.

[0079] In an exemplary embodiment of the present invention, the management of cable production quality can be achieved through the following steps:

[0080] Step 1: Take the clusters in the clustering results as the target clusters, calculate the cumulative sum of the average distances between the data points in all target clusters and the cluster centers of the corresponding target clusters, and take the normalized value of the product of the number of target clusters and the cumulative sum as the abnormality degree of the production quality sequence;

[0081] Specifically, the degree of abnormality of the production quality sequence satisfies the following relationship:

[0082] ;

[0083] In the formula, is the degree of abnormality in the production quality sequence; is the number of target clusters; The first in the production quality sequence The average distance between all data points in a target cluster and the cluster center of the target cluster; The natural constant An exponential function with base ; is the summation symbol.

[0084] Optionally, when the number of target clusters in the clustering result is large, it means that there are many dissimilar data points in the production quality sequence, that is, the data difference in the production process is large, and the abnormality of the sequence is relatively large.

[0085] in, It reflects the cumulative sum of the average distances between the data points in all target clusters and the cluster center of the corresponding target cluster. The larger the value, the greater the data difference in each target cluster in the clustering result, and the corresponding abnormality of the sequence is relatively large.

[0086] Step 2: Obtain a preset threshold value. When the abnormality level is greater than or equal to the preset threshold value, an alarm is issued to manage the cable production quality.

[0087] Optionally, the preset threshold can be set to 0.7. When the abnormality degree of the production quality sequence is greater than 0.7, it means that there is an abnormality in the quality of the cables produced during the sampling period corresponding to the sequence. Then an alarm is issued to remind the operator to pay attention to the cable production quality problems during the period, so that corresponding adjustment measures can be taken in time to ensure the quality of the produced cables.

[0088] In the description of this specification, "plurality" or "several" means at least two, such as two, three or more, etc., unless otherwise clearly and specifically defined.

[0089] Although this specification has shown and described a number of embodiments of the present invention, it will be apparent to those skilled in the art that such embodiments are provided by way of example only. Those skilled in the art will conceive of many modifications, changes and alternatives without departing from the ideas and spirit of the present invention. It should be understood that in the practice of the present invention, various alternatives to the embodiments of the present invention described herein may be employed.

Claims

1. A cable production quality data management system based on Internet of Things big data, characterized in that: include: The acquisition module is used to collect data of various dimensions in the production process at each sampling moment to obtain the production quality sequence; A sample partitioning module is used to sequentially screen the target data points in the production quality sequence, and use the target data point and its neighboring data points screened each time as a cluster until all data points are partitioned to obtain multiple clusters, wherein the target data point is the data point with the largest number of neighboring data points in the production quality sequence; The abnormality detection module is used to use the cluster center of each cluster as the initial cluster center of the ISODATA algorithm, cluster the production quality sequence to obtain a clustering result, use the cluster in the clustering result as the target cluster, calculate the cumulative sum of the average distances between the data points in all target clusters and the cluster center of the corresponding target cluster, and use the normalized value of the product of the number of target clusters and the cumulative sum as the abnormality degree of the production quality sequence, satisfying the following relationship: ； In the formula, is the degree of abnormality in the production quality sequence; is the number of target clusters; The first in the production quality sequence The average distance between all data points in a target cluster and the cluster center of the target cluster; The natural constant An exponential function with base ; is a summation symbol; when the abnormality level is greater than or equal to a preset threshold, an alarm is issued to manage the cable production quality; The method for obtaining the cluster center includes: The average value of all data points in any cluster in any dimension is taken as the value of the cluster center of the cluster in any dimension; Calculate the values ​​of the cluster center of any cluster in all dimensions to obtain the position of the cluster center of any cluster; The neighboring data points are the data points whose distances from the dimensional data corresponding to the target data point to each dimensional data in the production quality sequence are less than or equal to the neighboring distances of each dimensional data; the neighboring distance is the product of the significance of the corresponding dimensional data and the preset initial distance; the significance is the average difference between the corresponding dimensional data and the average value of the dimensional data, satisfying the following relationship: ； In the formula, The first in the production quality sequence The significance of the data in these dimensions; is the number of the i-th dimension data in the production quality sequence; The first in the production quality sequence The first dimension The value of the data; The first in the production quality sequence The average value of the dimension data; is the normalization function; for the summation symbol; is the absolute value symbol.

2. According to claim 1, a cable production quality data management system based on Internet of Things big data is characterized in that: The system further comprises: A quantity threshold is preset, and when the number of neighboring data points of any target data point is less than or equal to the quantity threshold, the target data point and its neighboring data points are respectively divided into nearest neighbor clusters.

3. According to claim 1, a cable production quality data management system based on Internet of Things big data is characterized in that: The data of various dimensions in the production process are collected at each sampling moment, including: At each sampling moment, readings of a plurality of sensors electrically connected to the cable extruder are collected to obtain data of a plurality of dimensions, wherein the dimensions correspond one to one with the data types collected by the sensors.

4. A cable production quality data management system based on Internet of Things big data according to claim 1, characterized in that: The method for obtaining the clustering result includes: When the number of iterations for clustering the production quality sequence is greater than a preset number of iterations or the positions of all cluster centers tend to converge, the iteration is stopped to obtain a clustering result.

Images