A smart factory data processing method and system

Abstract

This disclosure provides a smart factory data processing method and system. After extracting monitoring data clusters to be processed from production monitoring logs, a descriptive carrier is extracted from each data cluster to obtain a data cluster descriptive carrier. Then, the data cluster operating status and weight corresponding to the data clusters to be processed are determined using the data cluster descriptive carrier. Based on the data cluster operating status and weight, the monitoring data clusters to be processed are sequentially determined, and one or more target monitoring data clusters are selected. Next, the target monitoring data clusters are decomposed, and one or more representative data items are obtained from the decomposed monitoring data items based on their contribution. This disclosure can accurately select target monitoring data clusters from the monitoring data clusters to be processed, thereby detecting representative data items within the target monitoring data clusters, thus increasing the accuracy of representative data item detection.

Description

Technical Field

[0001] This disclosure relates to the field of data processing, and more specifically, to a smart factory data processing method and system. Background Technology

[0002] With the rapid development of information technology and automation technology, people have begun to realize the many problems existing in traditional factory operation models, such as opaque production processes, low efficiency, and difficult equipment maintenance. Therefore, the concept of smart factories has emerged, aiming to achieve comprehensive optimization and management of the production process through digitalization, networking, and intelligence. A smart factory refers to a factory that utilizes advanced information technology (such as the Internet of Things, big data analytics, and artificial intelligence) and automation technology to achieve digital, networked, and intelligent management of the entire production process. Data processing is of paramount importance in smart factories. By collecting and storing the large amounts of data generated during the production process, smart factories can achieve real-time monitoring and recording of production status, making the production process more transparent and controllable. Through data analysis and mining, smart factories can identify potential problems and areas for improvement in the production process, thereby increasing production efficiency and reducing costs. Furthermore, data processing can also help smart factories achieve predictive maintenance of equipment failures, detecting equipment anomalies in advance and carrying out repairs, avoiding production losses due to equipment failures.

[0003] In data processing at smart factories, identifying representative data items is crucial. These representative data items are information that decisively impacts the production process and product quality. By identifying and processing this data, refined monitoring and management of the production process can be achieved, thereby optimizing production efficiency and improving product quality. Representative data identification is the foundation and prerequisite for production process optimization, equipment status monitoring, quality control, and intelligent decision support. How to accurately identify representative data in monitoring data is a current key technical challenge. Summary of the Invention

[0004] In view of this, the embodiments of this application provide at least one smart factory data processing method and system.

[0005] According to one aspect of the present disclosure, a smart factory data processing method is provided, applied to a data processing device, the data processing device being communicatively connected to at least one data monitoring device, the method comprising:

[0006] Obtain production monitoring logs from one or more target factory production lines, and extract monitoring data clusters to be processed from the production monitoring logs, wherein the production monitoring logs are collected and sent through the at least one data monitoring device;

[0007] The description carrier is extracted from the monitoring data cluster to be processed to obtain the data cluster description carrier of the monitoring data cluster to be processed.

[0008] The data cluster description carrier is used to determine the data cluster running status and data cluster weight corresponding to the monitoring data cluster to be processed. The data cluster running status is used to describe the running status type of the production monitoring log, and the data cluster weight is used to describe the contribution of the monitoring data cluster to be processed in the data cluster running status.

[0009] The monitoring data clusters to be processed are determined sequentially based on the running status and weight of the data clusters, so as to select one or more target monitoring data clusters;

[0010] The target monitoring data cluster is decomposed into data items. One or more representative data items are obtained from the monitoring data items based on the data item contribution degree of the decomposed monitoring data items. The data item contribution degree is used to describe the contribution of the monitoring data item in the target monitoring data cluster.

[0011] According to an example of an embodiment of this disclosure, the step of sequentially determining the monitoring data clusters to be processed based on the data cluster operating status and data cluster weights to select one or more target monitoring data clusters includes:

[0012] The data clusters to be processed are classified according to their operating status to obtain a set of data clusters to be processed corresponding to each data cluster's operating status.

[0013] The monitoring data clusters to be processed in the set of monitoring data clusters to be processed are determined sequentially according to the data cluster weights.

[0014] By sequentially determining the information, negative attribute data cleaning is performed on the monitoring data clusters to be processed in the set of monitoring data clusters to be processed, to obtain one or more target monitoring data clusters.

[0015] According to an example of an embodiment of this disclosure, the step of performing negative attribute data cleaning on the monitoring data clusters in the set of monitoring data clusters to be processed by sequentially determining information to obtain one or more target monitoring data clusters includes:

[0016] By sequentially determining the information, select the monitoring data clusters to be processed from the set of monitoring data clusters to be processed whose sequential results are less than or equal to the preset sequential critical results, and obtain one or more monitoring data clusters with negative attributes to be processed.

[0017] The negative attribute monitoring data clusters are cleaned in the set of monitoring data clusters to be processed to obtain one or more target monitoring data clusters.

[0018] According to an example of an embodiment of this disclosure, the step of obtaining one or more representative data items from the monitoring data items based on the data item contribution rate of the data decomposed monitoring data items includes:

[0019] Determine the statistical number of data items in the target monitoring data cluster to obtain the first statistical number of data items;

[0020] The contribution of each monitoring data item after data decomposition is determined by counting the number of the first data items.

[0021] Based on the contribution of the data items and the operating status of the data cluster, one or more representative data items are selected from the monitored data items.

[0022] According to an example of an embodiment of this disclosure, determining the data item contribution of each monitoring data item after data decomposition by counting the first data items includes:

[0023] In the target monitoring data cluster, determine the statistical number of data items that cover the monitoring data items to obtain the second statistical number of data items;

[0024] The data item frequency and discrimination coefficient of each monitored data item are obtained by counting the first data item and the second data item.

[0025] The data item contribution is obtained by producting the frequency of the data item and the discrimination coefficient;

[0026] The step of selecting one or more representative data items from the monitored data items based on the contribution of the data items and the operating status of the data cluster includes:

[0027] Select the monitoring data cluster corresponding to the running status of each data cluster from the target monitoring data cluster, and determine the statistical number of data items in the monitoring data cluster to obtain the statistical number of the third data item;

[0028] Valid data items are obtained from the monitoring data items. The conditional support of each monitoring data item is obtained by counting the number of valid data items and the third data item. The conditional support is used to describe the support of the monitoring data item in the target monitoring data cluster corresponding to the running status of the specified data cluster.

[0029] Based on the contribution and conditional support of the data items, one or more representative data items are selected from the monitored data items.

[0030] According to an example of an embodiment of this disclosure, obtaining the conditional support of each monitoring data item through the statistical count of the third data item and the valid data items includes:

[0031] The distribution support of each monitoring data item is obtained by counting the number of the third data item, thus obtaining the distribution support of the data item;

[0032] The distribution support of the effective data items is obtained based on the effective data items, the number of first data items, and the number of third data items, thus obtaining the effective distribution support.

[0033] The conditional support of the monitored data item is obtained by product of the distribution support and the effective distribution support.

[0034] The step of selecting one or more representative data items from the monitored data items based on the contribution and conditional support of the data items includes:

[0035] The monitoring data items are cleaned using the conditional support to obtain cleaned monitoring data items.

[0036] The cleaned monitoring data items are determined sequentially based on their contribution.

[0037] By sequentially determining the information, one or more representative data items are selected from the monitoring data items after cleaning.

[0038] According to an example of an embodiment of this disclosure, determining the data cluster operating status and data cluster weight corresponding to the monitoring data cluster to be processed through the data cluster description carrier includes:

[0039] The trained data processing algorithm obtains the data cluster runtime state description carrier from the data cluster description carrier.

[0040] The data cluster operation status corresponding to the monitoring data cluster to be processed is determined by the data cluster operation status description carrier.

[0041] The trained data processing algorithm transforms the data cluster description carrier into a positive monitoring description carrier for the monitoring data cluster to be processed, thereby obtaining the data cluster weights.

[0042] According to an example of an embodiment of this disclosure, before the trained data processing algorithm obtains the data cluster running status description carrier and the monitoring positive description carrier from the data cluster description carrier, it further includes:

[0043] Obtain a set of monitoring data cluster templates for one or more target factory production line templates, wherein the set of monitoring data cluster templates includes one or more monitoring data cluster templates that carry the target factory production line type and the data cluster operating status;

[0044] The estimated data cluster operating status is obtained by using a preset data processing algorithm to predict the data cluster operating status of the monitoring data cluster template.

[0045] Using the monitoring data cluster template, the target factory production line type of the target factory production line template is estimated by the preset data processing algorithm to obtain the estimated target factory production line type.

[0046] The preset data processing algorithm is optimized by carrying the target factory production line type, carrying the data cluster operating status, estimating the data cluster operating status, and estimating the target factory production line type, so as to obtain a trained data processing algorithm.

[0047] According to an example of an embodiment of this disclosure, the step of estimating the target factory production line type of the target factory production line template using the monitoring data cluster template and the preset data processing algorithm to obtain the estimated target factory production line type includes:

[0048] The preset data processing algorithm is used to extract the description carrier from the monitoring data cluster template, and the extracted template data cluster description carrier is transformed into a template monitoring front description carrier.

[0049] By monitoring the front description carrier through the template, the template data cluster weight of the monitoring data cluster template is determined, and the monitoring eccentricity adjustment parameter of each monitoring data cluster template is obtained according to the template data cluster weight.

[0050] The template data cluster description carrier is fused according to the monitoring eccentricity adjustment parameters, and the target factory production line type of the target factory production line template is determined according to the fused template data cluster description carrier to obtain the estimated target factory production line type.

[0051] The step of optimizing the preset data processing algorithm by carrying the target factory production line type, carrying the data cluster operating status, estimating the data cluster operating status, and estimating the target factory production line type to obtain the trained data processing algorithm includes:

[0052] The target factory production line cost of the monitoring data cluster template is determined by carrying the target factory production line type and the estimated target factory production line type.

[0053] The monitoring cost of the monitoring data cluster template is determined based on the operating status of the carried data cluster and the estimated operating status of the data cluster.

[0054] Obtain the combined eccentricity adjustment parameter of the monitoring cost, and fuse the monitoring cost and the target factory production line cost according to the combined eccentricity adjustment parameter;

[0055] The preset data processing algorithm is optimized based on the cost after fusion to obtain the trained data processing algorithm.

[0056] The acquisition of a monitoring data cluster template set for one or more target factory production line templates includes:

[0057] Obtain an initial set of monitoring data cluster templates for one or more target factory production line templates;

[0058] If the number of monitoring data cluster templates in the initial monitoring data cluster template set is greater than the template reference number, a preset number of initial monitoring data cluster templates are obtained by sampling from the initial monitoring data cluster template set based on the monitoring acquisition timing of the monitoring data cluster templates, and the remaining monitoring data cluster templates are obtained.

[0059] The remaining sampling number of the monitoring data cluster template is determined by the template reference quantity and the set quantity.

[0060] The target monitoring data cluster template corresponding to the remaining sampling number is obtained by randomly sampling from the remaining monitoring data cluster template.

[0061] The initial monitoring data cluster template and the target monitoring data cluster template are merged to obtain the monitoring data cluster template set of the target factory production line template.

[0062] According to another aspect of the present disclosure, a data processing system is provided, including a data monitoring device and a data processing device communicatively connected to each other, the data processing device comprising:

[0063] One or more processors;

[0064] and one or more memories, wherein the memories store computer-readable code that, when run by the one or more processors, causes the one or more processors to perform the methods described above.

[0065] The beneficial effects included in this disclosure are at least as follows:

[0066] The smart factory data processing method and system provided in this disclosure, after acquiring production monitoring logs of one or more target factory production lines and extracting monitoring data clusters to be processed from the production monitoring logs, extracts a description carrier for the monitoring data clusters to be processed, obtaining a data cluster description carrier. Then, the data cluster operating status and data cluster weight corresponding to the monitoring data clusters to be processed are determined through the data cluster description carrier. Based on the data cluster operating status and data cluster weight, the monitoring data clusters to be processed are sequentially determined, and one or more target monitoring data clusters are selected. Then, the target monitoring data clusters are decomposed, and one or more representative data items are obtained from the monitoring data items based on the data item contribution rate of the decomposed monitoring data items. After acquiring the monitoring data clusters to be processed from the production monitoring logs, this disclosure embodiment can extract the data cluster operating status and data cluster weight corresponding to the monitoring data clusters to be processed through the data cluster description carrier. Using the data cluster operating status and data cluster weight as known information, the target monitoring data clusters can be accurately selected from the monitoring data clusters to be processed, thereby detecting representative data items in the target monitoring data clusters, thus increasing the accuracy of representative data item detection.

[0067] It should be understood that the above general description and the following detailed description are merely exemplary and explanatory, and are not intended to limit the technical solutions of this application. Attached Figure Description

[0068] The above and other objects, features, and advantages of the present disclosure will become more apparent from the more detailed description of the embodiments thereof in conjunction with the accompanying drawings. The accompanying drawings are provided to further understand the embodiments of the present disclosure and form part of the specification. They are used together with the embodiments of the present disclosure to explain the disclosure and do not constitute a limitation thereof. In the drawings, the same reference numerals generally represent the same parts or operations.

[0069] Figure 1 This is a schematic diagram illustrating the application scenarios provided in the embodiments of this application.

[0070] Figure 2 This is a schematic diagram illustrating the implementation process of a smart factory data processing method provided in an embodiment of this application.

[0071] Figure 3 This is a schematic diagram of the composition structure of a data processing device provided in an embodiment of this application.

[0072] Figure 4 This is a schematic diagram of the hardware entity of a data processing device provided in an embodiment of this application. Detailed Implementation

[0073] The technical solutions of the embodiments of this disclosure will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this disclosure. All other embodiments obtained by those skilled in the art based on the embodiments of this disclosure without creative effort are within the scope of protection of this disclosure.

[0074] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions of this application are further described in detail below with reference to the accompanying drawings and embodiments. The described embodiments should not be regarded as limitations on this application. All other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0075] In the following description, references to "some embodiments" refer to a subset of all possible embodiments. It is understood that "some embodiments" may be the same or different subsets of all possible embodiments and may be combined with each other without conflict. The terms "first / second / third" are used merely to distinguish similar objects and do not represent a specific ordering of objects. It is understood that "first / second / third" may be interchanged in a specific order or sequence where permitted, so that the embodiments of this application described herein can be implemented in an order other than that illustrated or described herein.

[0076] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains. The terminology used herein is for descriptive purposes only and is not intended to limit the scope of this application.

[0077] The smart factory data processing method provided in this application embodiment can be applied to, for example... Figure 1 In the application environment shown, the data monitoring device 102 communicates with the data processing device 104 via a network. A data storage system can store the data that the data processing device 104 needs to process. The data storage system can be integrated into the data processing device 104 or placed in the cloud or on other network servers. The device monitoring data can be stored in the local storage of the data monitoring device 102, or in the data storage system or cloud storage associated with the data processing device 104. When data processing is required, the data processing device 104 can retrieve the monitoring data from the local storage of the data monitoring device 102, the data storage system, or the cloud storage. The data monitoring device 102 can be, but is not limited to, various personal computers, laptops, tablets, IoT devices, etc. The data processing device 104 can be implemented using a standalone server or a server cluster consisting of multiple servers.

[0078] The smart factory data processing method provided in this application is applied to a data processing device and specifically includes the following operations:

[0079] Operation S101: Obtain production monitoring logs for one or more target factory production lines, and extract monitoring data clusters to be processed from the production monitoring logs.

[0080] The method provided in this disclosure is applied to an industrial internet platform, specifically to data processing devices in the PaaS layer (i.e., platform layer) of the industrial internet. The industrial internet platform includes, from bottom to top, the endpoint layer, edge layer, IaaS layer (infrastructure layer), PaaS layer (platform layer), and SaaS layer (application layer). The endpoint layer, also known as the device layer, comprises various IoT-based industrial devices in the smart factory production site, such as CNC machine tools, industrial sensors, and industrial robots. These devices span the entire product lifecycle, playing different roles in production, testing, and monitoring to monitor the production site. Based on IoT technology, the endpoint layer generates and aggregates a large amount of industrial data, including historical and real-time data, making it the underlying foundation of the industrial internet platform. The edge layer collects the industrial data generated by the endpoint layer and performs protocol parsing and edge processing on industrial data from different sources, ensuring compatibility with various industrial communication protocols. It converts and unifies the format of the collected data and then remotely transmits the relevant data to the industrial internet platform via wired or wireless methods (such as 5G, NB-IoT, etc.) through links such as fiber optics and Ethernet. The Infrastructure as a Service (IaaS) layer primarily provides cloud infrastructure, such as computing, network, and storage resources, supporting the overall operation of the industrial internet platform. As the connection layer between devices and platform applications, the IaaS layer provides complete underlying infrastructure services for the functionality of the PaaS layer and the application services of the SaaS layer. The Platform as a Service (PaaS) layer is the core of the entire industrial internet platform. Built with cloud computing technology, it not only receives and stores data but also provides a powerful computing environment for cloud processing or control of industrial data. Fundamentally, it builds a highly scalable support system on the IaaS platform, providing a solid foundation for the development of industrial applications or software. The Application as a Service (SaaS) layer is crucial to the industrial internet platform. It serves as the gateway for external services, directly connecting with users and demonstrating the ultimate application value of industrial data.

[0081] Production monitoring logs are collected and transmitted through at least one data monitoring device, which can be located at the edge layer. The logs record equipment monitoring data for the corresponding production line, such as production processing data, equipment status data, energy consumption data, and environmental data. This data can include continuous numerical data, such as rotational speed, temperature, and humidity, as well as discrete status data, such as equipment status. After acquiring production monitoring logs from one or more target factory production lines, a data cluster to be processed (which can be understood as extracting useful information; for example, data irrelevant to the data to be processed is discarded, and useful data is extracted to form the data cluster to be processed, or the dataset to be processed) is obtained. The method for extracting the data cluster to be processed is, for example, obtaining an initial data cluster from the production monitoring logs, selecting target data items from the initial data cluster, selecting replacement data items corresponding to the target data items from a pre-prepared set of data item replacements based on the data item type of the target data item, and replacing the target data items with replacement data items in the initial data cluster to obtain the data cluster to be processed. The initial monitoring data cluster can be obtained from the production monitoring log in any way. For example, it can be obtained based on the data type of the production monitoring log. If the data type is equipment-related data (such as operational data or external environment data), the corresponding data item is considered valid and is used as the initial monitoring data cluster. If it is not equipment-related data, the corresponding data item is considered invalid and is ignored. After obtaining the initial monitoring data cluster from the production monitoring log, target data items are selected from it. The selection method can be, for example, selecting data items of a preset data item type (such as discrete data items describing equipment status) from the initial monitoring data cluster to obtain the target data item, or selecting preset data items from the initial monitoring data cluster to obtain the target data item.

[0082] After selecting the data item type of the target data item, a replacement data item corresponding to the target data item is selected from a pre-prepared data item replacement set based on the data item type of the target data item. For example, when the data item type is device on / off status information, a first replacement data item is selected from the pre-prepared data item replacement set; when the data item type is device current gear information, a second replacement data item is selected from the pre-prepared data item replacement set. The replacement data item is a continuous numerical value. In this embodiment, transforming the target data item in the initial monitoring data cluster is to unify the data, converting invalid data items in the initial monitoring data cluster into unified data items, thereby eliminating the disturbance of monitoring data items in the monitoring data cluster to be processed by invalid data items and increasing the reliability of representative data item identification.

[0083] In operation S102, the description carrier of the monitoring data cluster to be processed is extracted to obtain the data cluster description carrier of the monitoring data cluster to be processed.

[0084] Specifically, the data cluster description carrier extraction operator of the trained data processing algorithm can extract multi-dimensional operational description carriers from the monitoring data cluster to be processed, obtaining multi-dimensional initial data cluster description carriers. The data cluster description carrier is a carrier representation carrying the feature information of the monitoring data cluster to be processed (such as feature information representing the operational status). It can be a description vector, description matrix, or description tensor, depending on the actual data dimension, which is not limited in this disclosure. The initial data cluster description carriers of various dimensions are then fused to obtain the data cluster description carrier of the monitoring data cluster to be processed. Optionally, the data cluster description carrier extraction operator can be a feature extraction operator such as a convolution operator or a residual operator.

[0085] Operation S103 determines the running status and weight of the data cluster corresponding to the monitoring data cluster to be processed through the data cluster description carrier.

[0086] Data cluster operational status can be used to describe the operational status type of production monitoring logs, such as normal, abnormal, overloaded, or idle states. Data cluster weights describe the contribution of the data cluster to be processed within its operational status; they can be understood as its importance to the operational status. This determines the contribution of the data cluster to be processed within the corresponding production monitoring logs and assesses whether the production monitoring logs corresponding to the data cluster to be processed are valid. The method for determining the operational status and weight of the data cluster to be processed through a data cluster description carrier is as follows: For example, a trained data processing algorithm obtains the data cluster operational status description carrier from the data cluster description carrier; the operational status of the data cluster to be processed is determined through the data cluster operational status description carrier; and the trained data processing algorithm transforms the data cluster description carrier into a positive monitoring description carrier for the data cluster to be processed, thereby obtaining the data cluster weights.

[0087] Specifically, the data cluster runtime description carrier can be obtained from the data cluster description carrier by a trained data processing algorithm. For example, the data cluster runtime description carrier can be obtained from the data cluster description carrier by an artificial neural network of a trained data processing algorithm, or by other neural networks.

[0088] After obtaining the data cluster operational status description carrier, the operational status of the data cluster to be monitored is determined using this carrier. One method for determining the operational status is to transform the data cluster operational status description carrier into a probability distribution of the operational status of the data cluster to be monitored using a classification operator (such as a fully connected operator). Based on this probability distribution, the probability of selecting the data cluster to be monitored from a preset range of operational statuses is then used to determine its operational status. The positive monitoring description carrier describes the characteristics of the data cluster to be monitored in the production monitoring logs. The data cluster weights are obtained by transforming the data cluster description carrier into a positive monitoring description carrier using a trained data processing algorithm. Another method is to transform the data cluster description carrier into a positive monitoring description carrier using a trained artificial neural network. The weights of the data cluster to be monitored are then determined using this positive monitoring description carrier.

[0089] Specifically, the trained data processing algorithm can be constructed based on specific needs. Before obtaining the data cluster operation status description carrier and the monitoring positive description carrier respectively in the data cluster description carrier through the trained data processing algorithm, the smart factory data processing method provided in this disclosure embodiment may further include the following operations: obtaining a set of monitoring data cluster templates for one or more target factory production line templates, the set of monitoring data cluster templates including one or more monitoring data cluster templates carrying the target factory production line type and carrying the data cluster operation status; estimating the data cluster operation status of the monitoring data cluster templates through a preset data processing algorithm to obtain the estimated data cluster operation status; estimating the target factory production line type of the target factory production line templates through the monitoring data cluster templates and the preset data processing algorithm to obtain the estimated target factory production line type; and optimizing the preset data processing algorithm by carrying the target factory production line type, carrying the data cluster operation status, the estimated data cluster operation status, and the estimated target factory production line type to obtain the trained data processing algorithm.

[0090] The above process describes the training of a data processing algorithm. In practice, this may include the following operations:

[0091] Operation S11: Obtain a set of monitoring data cluster templates for one or more target factory production line templates.

[0092] The monitoring data cluster templates are the sample data used for training, including one or more monitoring data cluster templates that carry the target factory production line type and the data cluster operating status. Obtaining the monitoring data cluster template set for one or more target factory production line templates includes: obtaining an initial monitoring data cluster template set for one or more target factory production line templates; if the number of monitoring data cluster templates in the initial monitoring data cluster template set is greater than the template reference number (which can be understood as a quantity threshold), sampling a preset number of initial monitoring data cluster templates from the initial monitoring data cluster template set based on the monitoring acquisition time sequence (i.e., the time order of acquisition or generation) to obtain the remaining monitoring data cluster templates; determining the remaining sampling number of monitoring data cluster templates based on the template reference number and the set number; randomly sampling the remaining monitoring data cluster templates to obtain the target monitoring data cluster template corresponding to the remaining sampling number; and fusing the initial monitoring data cluster templates and the target monitoring data cluster templates to obtain the monitoring data cluster template set for the target factory production line template. In the above process, the initial monitoring data cluster template set is obtained in a way that, for example, by obtaining a set of factory production lines in a smart factory, the candidate target factory production lines in the target factory production line set are filtered using preset target factory production line information to obtain one or more target factory production line templates. The production monitoring logs and operating status identification results of each target factory production line template are obtained. Based on the operating status identification results of the target factory production line template, the target factory production line type (operating status type) of the target factory production line template is determined. Monitoring data clusters and data cluster operating statuses are obtained from the production monitoring logs of the target factory production line templates. The target factory production line type and data cluster operating status are carried in the monitoring data clusters to obtain monitoring data cluster templates. The monitoring data cluster templates of each target factory production line template are then merged to obtain the initial monitoring data cluster template set for the target factory production line templates. The method of filtering candidate target factory production lines in the target factory production line set using preset target factory production line information, for example, by obtaining preset target factory production line information and then cleaning the candidate target factory production lines included in the preset target factory production line information (e.g., production lines with low attention requirements, such as cleaning production lines), results in one or more target factory production line templates.

[0093] Operation S12 uses a preset data processing algorithm to estimate the data cluster operating status of the monitoring data cluster template and obtain the estimated data cluster operating status.

[0094] For example, the process of estimating the data cluster operating status of the monitoring data cluster template by using a preset data processing algorithm and obtaining the estimated data cluster operating status is the same as the process of determining the data cluster operating status of the monitoring data cluster to be processed.

[0095] Operation S13 involves monitoring the data cluster template and using a preset data processing algorithm to estimate the target factory production line type of the target factory production line template, thereby obtaining the estimated target factory production line type.

[0096] In practical implementation, a pre-set data processing algorithm can be used to extract descriptive carriers from the monitoring data cluster templates. These extracted descriptive carriers are then transformed into template monitoring front descriptive carriers. The template data cluster weights of the monitoring data cluster templates are determined using these front descriptive carriers. Based on these weights, a monitoring eccentricity adjustment parameter (i.e., a parameter used to assign a corresponding weighting value; a larger eccentricity adjustment parameter results in a larger weighting value) is obtained for each monitoring data cluster template. The template data cluster descriptive carriers are then fused based on these fused parameters. Finally, the target factory production line type of the target factory production line template is determined based on the fused template data cluster descriptive carriers, thus obtaining the estimated target factory production line type. The method for obtaining the monitoring eccentricity adjustment parameter for each monitoring data cluster template based on its weights can be as follows: For example, the data cluster weight value is obtained from the data cluster weights of each monitoring data cluster template. These weights are then fused (e.g., added together), and the ratio of the original template data cluster weight value to the fused weight value is calculated to obtain the monitoring eccentricity adjustment parameter for each monitoring data cluster template. One method for determining the target factory production line type of the target factory production line template based on the description carrier of the fused template data cluster is, for example, mapping the description carrier of the fused template data cluster using an artificial neural network to obtain a numerical value, and then obtaining the estimated probability of the target factory production line template through a ReLU logic function to determine the estimated target factory production line type of the target factory production line template.

[0097] Operation S14 optimizes the preset data processing algorithm by carrying the target factory production line type, the data cluster operating status, the estimated data cluster operating status, and the estimated target factory production line type, thereby obtaining the trained data processing algorithm.

[0098] In practice, the target factory production line cost of the monitoring data cluster template is determined by carrying the target factory production line type and the estimated target factory production line type. Based on the carrying and estimated operating states of the data clusters, the monitoring cost of the monitoring data cluster template is determined. A merging eccentricity adjustment parameter for the monitoring cost is obtained, and the monitoring cost and the target factory production line cost are fused according to the merging eccentricity adjustment parameter. The preset data processing algorithm is then optimized based on the fused cost to obtain the trained data processing algorithm. For example, the method of determining the target factory production line cost of the monitoring data cluster template by carrying the target factory production line type and the estimated target factory production line type can be achieved by carrying the estimated probability of the target factory production line type and the estimated target factory production line type, obtaining the initial target factory production line cost of each monitoring data cluster template through a negative log-likelihood loss function, and then fusing the initial target factory production line costs to obtain the final target factory production line cost of the monitoring data cluster template. One method for determining the monitoring cost of a monitoring data cluster template is to carry the running status and estimated running status of the data cluster. For example, the initial monitoring cost of each monitoring data cluster template can be obtained by carrying the running status and estimated running status of the data cluster and using the negative log-likelihood loss function.

[0099] The initial monitoring costs are merged to obtain the target factory production line cost of the monitoring data cluster template. For example, the individual monitoring costs are summed to obtain the target factory production line cost.

[0100] After determining the target factory production line cost and monitoring cost, the monitoring cost and the target factory production line cost are merged based on the merged eccentricity adjustment parameter of the obtained monitoring cost. For example, the merged eccentricity adjustment parameter can be merged with the monitoring cost to obtain the merged monitoring cost. This merged monitoring cost is then merged with the target factory production line cost to obtain the merged cost.

[0101] L=L1+β·L2

[0102] Where L is the cost after fusion, L1 is the cost of the target factory production line, β is the fusion eccentricity adjustment parameter, and L2 is the monitoring cost.

[0103] The preset data processing algorithm is a multi-branch algorithm. The merging eccentricity adjustment parameter β is a hyperparameter, and its value is randomly set in the early stage of training. After training starts, the effect of the preset data processing algorithm on the prediction of each monitoring data cluster template begins to stabilize. By reducing the merging eccentricity adjustment parameter β, the prediction effect of the target factory production line cost and the target factory production line template begins to stabilize, and the training of the preset data processing algorithm is completed.

[0104] Operation S104 involves sequentially determining the monitoring data clusters to be processed based on the data cluster operating status and data cluster weight, in order to select one or more target monitoring data clusters.

[0105] In practice, the monitoring data clusters to be processed can be classified according to their operating status to obtain a set of monitoring data clusters to be processed corresponding to each operating status. Based on the data cluster weights, the monitoring data clusters to be processed in the set are determined sequentially. Using this sequential determination information, negative attribute data cleaning is performed on the monitoring data clusters to be processed, resulting in one or more target monitoring data clusters. Negative attribute data cleaning involves cleaning the monitoring data clusters to be processed that have negative attributes (i.e., data clusters with no processing value) to obtain effective monitoring data clusters to be processed (i.e., monitoring data clusters whose contribution is within a preset sequential determination interval). Specifically, in the ascending order of the sequential results (e.g., sequentially numbered sequences), the monitoring data clusters to be processed that have a contribution greater than a preset sequential threshold are considered effective monitoring data clusters to be processed. The corresponding negative attribute monitoring data clusters are those whose sequential results are less than or equal to a preset sequential determination threshold. The method of cleaning negative attribute data based on sequentially determined information is as follows: Selecting monitoring data clusters whose sequential results are less than or equal to a preset sequential threshold result from the monitoring data cluster set to be processed, thereby obtaining one or more negative attribute monitoring data clusters, and cleaning the negative attribute monitoring data clusters in the monitoring data cluster set to be processed to obtain one or more target monitoring data clusters.

[0106] Operation S105 decomposes the target monitoring data cluster into data, and obtains one or more representative data items from the monitoring data items based on the data item contribution of the decomposed monitoring data items.

[0107] In actual implementation, the target monitoring data cluster is decomposed. For example, it can be decomposed using a preset data decomposition control according to a preset decomposition method to obtain one or more monitoring data items corresponding to the target monitoring data cluster. For example, it can be decomposed according to a complete data item, such as:

[0108] 1. Equipment status data:

[0109] Equipment Name: Machining Center A

[0110] Runtime: 360 hours

[0111] Number of failures N: 2

[0112] 2. Production process data:

[0113] Temperature sensor data T: 25℃

[0114] Humidity sensor data W: 50%

[0115] Pressure sensor data P: 100 psi

[0116] Production speed S: 1000 pieces / hour

[0117] ...During disassembly, the data items were as follows: Name-A; Time-360; N-2; T-25℃; W-50%; P-100psi; S-1000...

[0118] After decomposing the target monitoring data cluster, one or more representative data items are detected from the monitoring data items based on the contribution of each decomposed data item. These representative data items are the key data that represents the overall data. The method for detecting representative data items is, for example, determining the total number of data items in the target monitoring data cluster (i.e., the total number of data items included), obtaining a first data item count K. Using the first data item count K, the contribution of each decomposed monitoring data item is determined. Based on the data item contribution and the data cluster's operating status, one or more representative data items are selected from the monitoring data items. Here, the data item contribution describes the contribution of a monitoring data item to the target monitoring data cluster. The method for determining the contribution of each decomposed monitoring data item using the first data item count is, for example, determining the total number of data items covering the monitoring data items in the target monitoring data cluster, obtaining a second data item count L. The data item frequency and discrimination coefficient of each monitoring data item are obtained using the first data item count K and the second data item count L. The data item frequency and discrimination coefficient are then productted to obtain the data item contribution of the monitoring data item. The method for obtaining the data frequency and discrimination coefficient of each monitoring data item by counting the first data item and the second data item is, for example, to obtain the ratio of the count of the first data item and the count of the second data item, and then to obtain the discrimination coefficient of each monitoring data item based on the ratio. The discrimination coefficient can reflect the importance of the corresponding monitoring data item, and it can be calculated using the following formula:

[0119]

[0120] Where Q is the discrimination coefficient of each monitoring data item, K is the statistical number of the first data item, and L is the statistical number of the second data item covering the monitoring data items.

[0121] In the target monitoring data cluster, determine the statistical number of data items covered by the target monitoring data cluster in the operation status of each data cluster, and obtain the data item frequency of each monitoring data item.

[0122] After obtaining the data item frequency and discrimination coefficient for each monitoring data item, the data item frequency and discrimination coefficient can be fused (e.g., productized) to obtain the data item contribution of each monitoring data item. After determining the data item contribution of each monitoring data item after data decomposition, one or more representative data items are selected from the monitoring data items based on the data item contribution and the data cluster operating status. The method for selecting representative data items is, for example, selecting the monitoring data cluster corresponding to each data cluster operating status from the target monitoring data cluster, determining the statistical number of data items in the monitoring data cluster, obtaining the statistical number of third data items, obtaining valid data items from the monitoring data items, obtaining the conditional support (e.g., conditional probability) of each monitoring data item through the statistical number of third data items and the valid data items, and selecting one or more representative data items from the monitoring data items based on the data item contribution and conditional support. Conditional support describes the support of a monitored data item within a target monitored data cluster corresponding to the operating state of a specified data cluster. It is obtained by using the count of third data items and valid data items to calculate the conditional support of each monitored data item. For example, the distributional support (e.g., probability distribution) of each monitored data item is obtained by using the count of third data items to calculate the distributional support of the data item. The distributional support of the valid data item is then obtained based on the count of valid data items, the count of first data items, and the count of third data items. The conditional support of the monitored data item is obtained by productting the distributional support of the data item and the valid distributional support. Specifically, the distributional support of each monitored data item is obtained by using the count of third data items to calculate the distributional support of the monitored data item. For example, the count of data items covering the monitored data item is determined within the monitored data cluster to calculate the count of fourth data items. The ratio of the count of fourth data items to the count of third data items is then used to obtain the distributional support of the monitored data item. This distributional support describes the support of any monitored data item within a monitored data cluster where it appears in the operating state of that data cluster. The method for obtaining the distribution support of valid data items based on the number of valid data items, the first data item count, and the third data item count is as follows: For example, in the target monitoring data cluster, determine the number of data items covering the valid data items to obtain the fifth data item count. Based on the fifth data item count and the first monitoring data, obtain the global effective distribution support of the monitoring data item. In the monitoring data cluster, determine the number of data items covering the valid data items to obtain the sixth data item count. Based on the sixth data item count and the third data item count, obtain the effective distribution support of the monitoring data item. Use the global effective distribution support and the monitoring effective distribution support as the effective distribution support of the monitoring data item.Specifically, the method for obtaining the global effective distribution support of a monitoring data item based on the statistical count of the fifth data item and the first monitoring data is, for example, to obtain the ratio of the statistical count of the fifth data item to the statistical count of the first data item, to obtain the initial global effective distribution support corresponding to each effective data item of the monitoring data item, and to product the initial global effective distribution support to obtain the global effective distribution support. The global effective distribution support is the support of a data item in a specific monitoring data item appearing in any data item in the target monitoring data cluster. Specifically, the method for obtaining the monitoring effective distribution support of a monitoring data item based on the statistical count of the sixth data item and the statistical count of the third data item is, for example, to obtain the ratio of the statistical count of the sixth data item to the statistical count of the third data item, to obtain the initial monitoring effective distribution support corresponding to each effective data item of the monitoring data item, and to product the initial monitoring effective distribution support to obtain the monitoring effective distribution support. The monitoring effective distribution support is used to describe the support of any monitoring data cluster appearing in any number of data items in the monitoring data item for the operating state of a specific data cluster. After obtaining the data item distribution support and the effective distribution support, the data item distribution support and the effective distribution support are producted. The fusion method is, for example, to product the data item distribution support and the monitoring effective distribution support to obtain the fused distribution support. The support ratio of the fused distribution support to the global effective distribution support is obtained to obtain the conditional support of the monitoring data item. For example, it can be implemented based on the following formula:

[0123]

[0124] Wherein, S represents conditional support, describing the support of a specific monitoring data item appearing in a monitoring data cluster under a specific data cluster operating state; S1 represents data item distribution support; S2 represents effective distribution support, describing the support of a data item appearing in any monitoring data cluster under the data cluster operating state; S1·S2 represents fused distribution support; and S3 represents global effective distribution support, describing the support of a valid data item appearing in any data item in the target monitoring data cluster. After obtaining the conditional support of each monitoring data item, one or more representative data items are selected from the monitoring data items based on the data item contribution and conditional support. The method of selecting representative data items is, for example, cleaning the monitoring data items using conditional support to obtain cleaned monitoring data items, sequentially determining the cleaned monitoring data items based on the data item contribution, and then selecting one or more representative data items from the cleaned monitoring data items based on the sequential determination information. The monitoring data items can be cleaned by using conditional support, for example, by selecting monitoring data items whose conditional support is greater than a preset probability, so as to obtain cleaned monitoring data items.

[0125] By sequentially determining information, one or more representative data items are selected from the cleaned monitoring data items. For example, the first n target monitoring data items are selected from the cleaned monitoring data items and the target monitoring data items are determined as representative data items. Alternatively, the first n monitoring data items are selected from the cleaned monitoring data items, the data item contribution of the candidate monitoring data items is weighted, and one or more representative data items are selected from the weighted candidate monitoring data items.

[0126] The smart factory data processing method and system provided in this disclosure, after acquiring production monitoring logs of one or more target factory production lines and extracting monitoring data clusters to be processed from the production monitoring logs, extracts a description carrier for the monitoring data clusters to be processed, obtaining a data cluster description carrier. Then, the data cluster operating status and data cluster weight corresponding to the monitoring data clusters to be processed are determined through the data cluster description carrier. Based on the data cluster operating status and data cluster weight, the monitoring data clusters to be processed are sequentially determined, and one or more target monitoring data clusters are selected. Then, the target monitoring data clusters are decomposed, and one or more representative data items are obtained from the monitoring data items based on the data item contribution rate of the decomposed monitoring data items. After acquiring the monitoring data clusters to be processed from the production monitoring logs, this disclosure embodiment can extract the data cluster operating status and data cluster weight corresponding to the monitoring data clusters to be processed through the data cluster description carrier. Using the data cluster operating status and data cluster weight as known information, the target monitoring data clusters can be accurately selected from the monitoring data clusters to be processed, thereby detecting representative data items in the target monitoring data clusters, thus increasing the accuracy of representative data item detection.

[0127] Based on the same inventive concept, this application also provides a data processing apparatus for implementing the smart factory data processing method described above. The solution provided by this apparatus is similar to the implementation scheme described in the above method; therefore, the specific limitations in one or more data processing apparatus embodiments provided below can be found in the limitations of the smart factory data processing method described above, and will not be repeated here.

[0128] In one embodiment, such as Figure 3 As shown, a data processing apparatus 300 is provided, comprising:

[0129] The data cluster extraction module 310 is used to acquire production monitoring logs of one or more target factory production lines and extract monitoring data clusters to be processed from the production monitoring logs. The production monitoring logs are collected and sent through the at least one data monitoring device.

[0130] The description carrier extraction module 320 is used to extract the description carrier from the monitoring data cluster to be processed, and obtain the data cluster description carrier of the monitoring data cluster to be processed.

[0131] The data attribute acquisition module 330 is used to determine the data cluster running status and data cluster weight corresponding to the monitoring data cluster to be processed through the data cluster description carrier. The data cluster running status is used to describe the running status type of the production monitoring log, and the data cluster weight is used to describe the contribution of the monitoring data cluster to be processed in the data cluster running status.

[0132] The target data determination module 340 is used to sequentially determine the monitoring data clusters to be processed according to the running status and weight of the data clusters, so as to select one or more target monitoring data clusters;

[0133] The representative data item determination module 350 is used to decompose the target monitoring data cluster into data and obtain one or more representative data items from the monitoring data items by the data item contribution degree of the decomposed monitoring data items. The data item contribution degree is used to describe the contribution degree of the monitoring data item in the target monitoring data cluster.

[0134] Each module in the aforementioned tag processing device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in the processor of the data processing device in hardware form or independent of it, or stored in the memory of the data processing device in software form, so that the processor can call and execute the operations corresponding to each module.

[0135] In one embodiment, a data processing device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 4 As shown, this data processing device includes a processor, memory, input / output (I / O) interfaces, and a communication interface. The processor, memory, and I / O interfaces are connected via a system bus, and the communication interface is also connected to the system bus via the I / O interfaces. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides the environment for the operating system and computer programs stored in the non-volatile storage media. The database stores data including monitoring data. The I / O interfaces are used for exchanging information between the processor and external devices. The communication interface is used for communicating with external terminals via a network. When the computer program is executed by the processor, it implements a smart factory data processing method.

[0136] Those skilled in the art will understand that Figure 4 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the data processing device to which the present application is applied. A specific data processing device may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0137] In one embodiment, a data processing device is also provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the operations described in the above method embodiments.

[0138] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon that, when executed by a processor, performs the operations described in the above method embodiments.

[0139] In one embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the operations described in the above method embodiments.

[0140] It should be noted that the object information (including but not limited to the object's device information, corresponding personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the object or fully authorized by all parties, and the collection, use and processing of related data must comply with the relevant laws, regulations and standards of the relevant countries and regions.

[0141] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.

[0142] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0143] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A smart factory data processing method, characterized in that, Applied to a data processing device, the data processing device being communicatively connected to at least one data monitoring device, the method includes: Obtain production monitoring logs from one or more target factory production lines, and extract monitoring data clusters to be processed from the production monitoring logs, wherein the production monitoring logs are collected and sent through the at least one data monitoring device; The description carrier is extracted from the monitoring data cluster to be processed to obtain the data cluster description carrier of the monitoring data cluster to be processed. The data cluster description carrier is used to determine the data cluster running status and data cluster weight corresponding to the monitoring data cluster to be processed. The data cluster running status is used to describe the running status type of the production monitoring log, and the data cluster weight is used to describe the contribution of the monitoring data cluster to be processed in the data cluster running status. The monitoring data clusters to be processed are determined sequentially based on the running status and weight of the data clusters, so as to select one or more target monitoring data clusters; The target monitoring data cluster is decomposed into data items. One or more representative data items are obtained from the monitoring data items based on the data item contribution degree of the decomposed monitoring data items. The data item contribution degree is used to describe the contribution of the monitoring data item in the target monitoring data cluster.

2. The smart factory data processing method according to claim 1, characterized in that, The step of sequentially determining the monitoring data clusters to be processed based on the data cluster operating status and data cluster weight, to select one or more target monitoring data clusters, includes: The data clusters to be processed are classified according to their operating status to obtain a set of data clusters to be processed corresponding to each data cluster's operating status. The monitoring data clusters to be processed in the set of monitoring data clusters to be processed are determined sequentially according to the data cluster weights. By sequentially determining the information, negative attribute data cleaning is performed on the monitoring data clusters to be processed in the set of monitoring data clusters to be processed, to obtain one or more target monitoring data clusters.

3. The smart factory data processing method according to claim 2, characterized in that, The step involves sequentially determining information to perform negative attribute data cleaning on the monitoring data clusters in the set of monitoring data clusters to be processed, thereby obtaining one or more target monitoring data clusters, including: By sequentially determining the information, select the monitoring data clusters to be processed from the set of monitoring data clusters to be processed whose sequential results are less than or equal to the preset sequential critical results, and obtain one or more monitoring data clusters with negative attributes to be processed. The negative attribute monitoring data clusters are cleaned in the set of monitoring data clusters to be processed to obtain one or more target monitoring data clusters.

4. The smart factory data processing method according to any one of claims 1 to 3, characterized in that, The contribution of the monitoring data items after data decomposition is used to obtain one or more representative data items from the monitoring data items, including: Determine the statistical number of data items in the target monitoring data cluster to obtain the first statistical number of data items; The contribution of each monitoring data item after data decomposition is determined by counting the number of the first data items. Based on the contribution of the data items and the operating status of the data cluster, one or more representative data items are selected from the monitored data items.

5. The smart factory data processing method according to claim 4, characterized in that, The step of determining the data contribution of each monitoring data item after data decomposition by counting the first data item includes: In the target monitoring data cluster, determine the statistical number of data items that cover the monitoring data items to obtain the second statistical number of data items; The data item frequency and discrimination coefficient of each monitored data item are obtained by counting the first data item and the second data item. The data item contribution is obtained by producting the frequency of the data item and the discrimination coefficient; The step of selecting one or more representative data items from the monitored data items based on the contribution of the data items and the operating status of the data cluster includes: Select the monitoring data cluster corresponding to the running status of each data cluster from the target monitoring data cluster, and determine the statistical number of data items in the monitoring data cluster to obtain the statistical number of the third data item; Valid data items are obtained from the monitoring data items. The conditional support of each monitoring data item is obtained by counting the number of valid data items and the third data item. The conditional support is used to describe the support of the monitoring data item in the target monitoring data cluster corresponding to the running status of the specified data cluster. Based on the contribution and conditional support of the data items, one or more representative data items are selected from the monitored data items.

6. The smart factory data processing method according to claim 5, characterized in that, The step of obtaining the conditional support of each monitoring data item by counting the number of the third data item and the number of valid data items includes: The distribution support of each monitoring data item is obtained by counting the number of the third data item, thus obtaining the distribution support of the data item. The distribution support of the effective data items is obtained based on the effective data items, the number of statistics of the first data items, and the number of statistics of the third data items, thus obtaining the effective distribution support; The conditional support of the monitored data item is obtained by product of the distribution support and the effective distribution support. The step of selecting one or more representative data items from the monitored data items based on the contribution and conditional support of the data items includes: The monitoring data items are cleaned using the conditional support to obtain cleaned monitoring data items. The cleaned monitoring data items are determined sequentially based on their contribution. By sequentially determining the information, one or more representative data items are selected from the monitoring data items after cleaning.

7. The smart factory data processing method according to any one of claims 1 to 3, characterized in that, The step of determining the data cluster operating status and data cluster weight corresponding to the monitoring data cluster to be processed through the data cluster description carrier includes: The trained data processing algorithm obtains the data cluster runtime state description carrier from the data cluster description carrier. The data cluster operation status corresponding to the monitoring data cluster to be processed is determined by the data cluster operation status description carrier. The trained data processing algorithm transforms the data cluster description carrier into a positive monitoring description carrier for the monitoring data cluster to be processed, thereby obtaining the data cluster weights.

8. The smart factory data processing method according to claim 7, characterized in that, Before the trained data processing algorithm obtains the data cluster running status description carrier and the monitoring positive description carrier from the data cluster description carrier, it further includes: Obtain a set of monitoring data cluster templates for one or more target factory production line templates, wherein the set of monitoring data cluster templates includes one or more monitoring data cluster templates that carry the target factory production line type and the data cluster operating status; The estimated data cluster operating status is obtained by using a preset data processing algorithm to predict the data cluster operating status of the monitoring data cluster template. Using the monitoring data cluster template, the target factory production line type of the target factory production line template is estimated by the preset data processing algorithm to obtain the estimated target factory production line type. The preset data processing algorithm is optimized by carrying the target factory production line type, carrying the data cluster operating status, estimating the data cluster operating status, and estimating the target factory production line type, so as to obtain a trained data processing algorithm.

9. The smart factory data processing method according to claim 8, characterized in that, The step of estimating the target factory production line type of the target factory production line template using the monitoring data cluster template and the preset data processing algorithm to obtain the estimated target factory production line type includes: The preset data processing algorithm is used to extract the description carrier from the monitoring data cluster template, and the extracted template data cluster description carrier is transformed into a template monitoring front description carrier. By monitoring the front description carrier through the template, the template data cluster weight of the monitoring data cluster template is determined, and the monitoring eccentricity adjustment parameter of each monitoring data cluster template is obtained according to the template data cluster weight. The template data cluster description carrier is fused according to the monitoring eccentricity adjustment parameters, and the target factory production line type of the target factory production line template is determined according to the fused template data cluster description carrier to obtain the estimated target factory production line type. The step of optimizing the preset data processing algorithm by carrying the target factory production line type, carrying the data cluster operating status, estimating the data cluster operating status, and estimating the target factory production line type to obtain the trained data processing algorithm includes: The target factory production line cost of the monitoring data cluster template is determined by carrying the target factory production line type and the estimated target factory production line type. The monitoring cost of the monitoring data cluster template is determined based on the operating status of the carried data cluster and the estimated operating status of the data cluster. Obtain the combined eccentricity adjustment parameter of the monitoring cost, and fuse the monitoring cost and the target factory production line cost according to the combined eccentricity adjustment parameter; The preset data processing algorithm is optimized based on the cost after fusion to obtain the trained data processing algorithm. The acquisition of a monitoring data cluster template set for one or more target factory production line templates includes: Obtain an initial set of monitoring data cluster templates for one or more target factory production line templates; If the number of monitoring data cluster templates in the initial monitoring data cluster template set is greater than the template reference number, a preset number of initial monitoring data cluster templates are obtained by sampling from the initial monitoring data cluster template set based on the monitoring acquisition timing of the monitoring data cluster templates, and the remaining monitoring data cluster templates are obtained. The remaining sampling number of the monitoring data cluster template is determined by the template reference quantity and the set quantity. The target monitoring data cluster template corresponding to the remaining sampling number is obtained by randomly sampling from the remaining monitoring data cluster template. The initial monitoring data cluster template and the target monitoring data cluster template are merged to obtain the monitoring data cluster template set of the target factory production line template.

10. A data processing system, comprising a data monitoring device and a data processing device interconnected in communication, characterized in that, The data processing device includes: One or more processors; and one or more memories, wherein the memories store computer-readable code that, when executed by the one or more processors, causes the one or more processors to perform the method as described in any one of claims 1 to 9.

Images