A method for constructing a universal discrete manufacturing industry full-process supervision component library

Abstract

The present invention discloses a method for constructing a universal discrete manufacturing industry full-process supervision component library, and the steps include: 1) constructing a data access module for performing dynamic Bayesian preprocessing on the accessed data to obtain standardized data that conforms to the database storage structure, and then assigning identifiers to the standardized data based on relational reasoning and storing them in the database; 2) constructing a BOM list construction module for matching any two data with different identifiers in the database, and setting corresponding BOM list output rules for the required material quantity and price for materials of the same type according to the identifier matching; and when new standardized data is stored in the database, obtaining the most similar type and the corresponding BOM list output rules to construct the BOM list; 3) constructing a quality supervision module for detecting the BOM list generated in the production process of the production line, and storing the detection results in the database or synchronously pushing them to a user monitoring platform.

Description

Technical Field

[0001] The present invention belongs to the field of computer software technology and relates to a method for constructing a universal discrete manufacturing industry full-process supervision component library. Background Art

[0002] Summarizing the development status of my country's discrete manufacturing industry, when realizing digital transformation, it faces bottlenecks in the fit and reuse of the overall production line process component library, such as data access of multiple equipment model protocols, equipment scheduling, and product quality assurance.

[0003] The digital transformation of discrete manufacturing, represented by electronic product manufacturing, faces the common pain points of difficulty in accessing all factors of resources, as well as the application bottlenecks of product customization, such as difficulty in integrating rules and knowledge and supporting flexible production, and the application bottlenecks of large-scale collaborative manufacturing, such as difficulty in scheduling and low communication efficiency. The digital transformation of important people's livelihood discrete manufacturing, represented by textile manufacturing, faces the common pain points of information fragmentation in various links, as well as the application bottlenecks of product quality assurance, such as low intelligence and difficulty in locating quality problems, and the application bottlenecks of old equipment maintenance, such as the widespread existence of data islands and untimely fault response.

[0004] Academics and industry have made positive progress in multi-sensor data fusion analysis, identification resolution, sensor interface standards, etc., but they still face challenges such as incompatible access protocols, difficult data identification management, and suboptimal edge computing. The algorithm modules for actual production lines are highly specific and have relatively simple functions. Their deployment has high requirements for the local environment and cannot support the overall production process. The highly integrated internal structure makes the reuse rate of most current component libraries low.

[0005] Therefore, it is necessary to provide a reusable and highly universal component library that provides common technical support for the discrete manufacturing industry, covering the entire process from production data access to BOM construction to quality supervision. Summary of the invention

[0006] In view of the problems existing in the prior art, the purpose of the present invention is to provide a method for constructing a universal discrete manufacturing industry full-process supervision component library. The present invention can solve the pain points of data integration difficulties and low model reuse rate that are urgently needed in the current discrete manufacturing industry, thereby promoting the production of the discrete manufacturing industry and achieving the purpose of accelerating the digital transformation process of the industry.

[0007] The present invention includes designing a universal industrial full-process supervision component library from metadata collection to finished product quality inspection, accessing and integrating multi-protocol data resources, scheduling production lines and equipment according to inventory optimization algorithms, equipment management algorithms, and production planning algorithms, and finally performing quality inspection on finished products according to target detection algorithms and quality inspection algorithms. The purpose is to realize a decoupled component library that is highly adapted to the digital monitoring and management of the entire production process of the discrete manufacturing industry through the present invention.

[0008] The technical solution of the present invention is:

[0009] A method for constructing a universal discrete manufacturing industry full-process supervision component library, the steps of which include:

[0010] 1) Constructing a data access module for realizing multivariate data access; the data access module performs dynamic Bayesian preprocessing based on local reasoning on the accessed data to obtain standardized data that conforms to the database storage structure, and then assigns identifiers to the standardized data based on relational reasoning and stores them in the database;

[0011] 2) Constructing a BOM list construction module, which is used to match any two data with different identifiers in the database based on a multi-mode matching method, and set the corresponding type of BOM list output rules for the required material quantity and price of materials of the same type according to the identifier; and when new standardized data is stored in the database, clustering the new standardized data with each type of data, so as to obtain the type closest to the new standardized data and the corresponding BOM list output rules, and then use the screened BOM list output rules to construct the BOM list corresponding to the new standardized data;

[0012] 3) Construct a quality control module to detect the BOM sheets generated during the production process of the production line, and store the test results in the database or push them to the user monitoring platform simultaneously.

[0013] Furthermore, each module is integrated in a highly decoupled mode, each module is individually encapsulated and an API interface for parameter transmission is opened; the modules include the data access module, the BOM single construction module and the quality control module.

[0014] Furthermore, the modules are connected through API interface calls, and the relevant algorithms involving discrete manufacturing industry processes are separately encapsulated in Docker, deployed on the cloud, and called through API interfaces.

[0015] Furthermore, the business knowledge deviation correction technology based on progressive adversarial matching is used to detect the types and location information of quality defects in the BOM list, and the group training joint transfer framework based on federated contrastive learning is used to detect the types of faults in the BOM list.

[0016] Furthermore, the database is deployed on the client and joins the MA identification system through identification registration. The database serves as a third-level node of the MA identification system, enabling cross-regional, cross-industry, and cross-enterprise information query and sharing.

[0017] A server, characterized in that it comprises a memory and a processor, wherein the memory stores a computer program, the computer program is configured to be executed by the processor, and the computer program comprises instructions for executing each step in the above method.

[0018] A computer-readable storage medium stores a computer program, wherein the computer program implements the steps of the above method when executed by a processor.

[0019] The main advantages of the present invention are:

[0020] a) It is highly universal and implements key processes common to the discrete manufacturing industry, including multi-protocol data access, BOM single construction, and quality supervision;

[0021] b) High flexibility. The outer layer of the present invention is encapsulated using Docker, which can be deployed in the cloud and called through an open interface, thus reducing the requirements for the local operating environment and the technical difficulty of use;

[0022] c) With low coupling, each algorithm module in the present invention is relatively independent and interconnected through API calls. Users can replace or flexibly call the relevant algorithm modules in the present invention to make the overall component library more in line with actual needs;

[0023] The advantages of the present invention are as follows:

[0024] By using the universal discrete manufacturing industry full-process supervision component library construction method provided by the present invention, the fit and reuse rate of the component library with the actual production line can be improved; the algorithm modules involved in the present invention can be replaced according to the actual use requirements of the user, making the overall architecture more flexible and reducing the program operation coupling and secondary development difficulty; the present invention is encapsulated in docker, and the user can call it through the docker interface, reducing the deployment technology cost and simplifying the calling form.

[0025] This patent can not only improve the efficiency of the whole process supervision of the discrete manufacturing industry, but also ensure the universality of the component library. Compared with other methods, this patent has the characteristics of wide coverage of data access protocols, flexible decoupling and high reuse rate. It is a component library construction method suitable for the whole process supervision of the discrete manufacturing industry. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1 This is the architecture diagram of the universal industrial full-process supervision component library.

[0027] Figure 2 It is a business process diagram for the universal industrial full-process supervision component library. DETAILED DESCRIPTION

[0028] The present invention is further described in detail below in conjunction with the accompanying drawings. The examples given are only used to explain the present invention but not to limit the scope of the present invention.

[0029] The key points of the method of this patent mainly include: by combining and encapsulating highly common algorithm modules covering important processes in the discrete manufacturing industry, the pain points of low universality, low reuse rate and one-sided process supervision of the existing component library are solved. Through the calling method of the open API interface encapsulated by docker, each algorithm module is integrated but independent of each other in the decoupled component library of the present invention, reducing the degree of coupling, improving the flexible adaptability of the overall architecture, and reducing the technical difficulty of use, deployment, maintenance and secondary development.

[0030] 1. A universal supervision component library for the entire process of discrete manufacturing industry is proposed. The component library is composed of three process modules according to the common business logic of discrete manufacturing industry: data access module, BOM single construction module, and quality supervision module. Each process module is relatively independent and encapsulated with docker, and the connection is customized according to the actual business needs of the user through the parameterized API. The architecture diagram of the present invention is as follows: Figure 1 , its business logic is as follows Figure 2 , the detailed business logic is as follows:

[0031] Data access module

[0032] By integrating NB-IoT, CAN, Modbus, OPC-UA and other heterogeneous industrial protocol adaptation, the multi-data access function is realized, the accessed data is dynamically preprocessed by Bayesian, and the output is standardized data that conforms to the database storage structure. Finally, the standardized data is assigned an identifier based on relational reasoning and registered and coded and stored in the database. The database is connected to the MA identification system as a third-level node.

[0033] BOM Single Building Block

[0034] The relevant data is matched in turn based on the multi-mode matching method to find materials of the same type according to the identification, and the BOM output rules are set for the required material quantity, price, etc.

[0035] When new standardized data is stored in the database, the rules are clustered using embedded representation through an association mapping method based on the new standardized data and the data in each cluster, so as to obtain the cluster closest to the new standardized data and filter the available rules, and then use the filtered available rules to construct the BOM corresponding to the new standardized data.

[0036] This module matches the same type of materials according to the identification and sets the corresponding type of BOM output rules for the required material quantity and price; when new standardized data is stored in the database, the new standardized data is clustered with each type of data to obtain the type closest to the new standardized data and the corresponding BOM output rules, and then the filtered BOM output rules are used to construct the BOM corresponding to the new standardized data.

[0037] Quality control module

[0038] In the quality monitoring module, there are mainly defect detection and fault detection. The cloth image data generated by the camera during the production process can output the type and location information of the cloth quality defects through the Yolo object recognition algorithm; the machine and cloth data generated by the loom can output the type of machine fault through the random forest algorithm. Finally, the above results are stored in the database or pushed to the user monitoring platform synchronously.

[0039] 2. The present invention adopts a highly decoupled mode of integration, and each algorithm module in the architecture is individually encapsulated and opens an API interface for parameter transmission. Users can pass parameters through the outer calling code, customize the algorithm module to be called or the calling order, and can also replace the algorithm module with a custom algorithm module. The various modules in the present invention are connected through API interface calls, and the relevant algorithms involving the discrete manufacturing industry process can be individually encapsulated in docker, deployed on the cloud, and called through the API interface, so that users can achieve convenient deployment. The user database can be deployed locally and join the MA identification system through identification registration. At the same time, the local database serves as a third-level node of the MA identification system, so that data can be queried and shared across regions, industries, and enterprises, providing data support for users to subsequently explore deep functions.

[0040] Although the specific embodiments of the present invention are disclosed for the purpose of illustration, the purpose is to help understand the content of the present invention and implement it accordingly, those skilled in the art will understand that various substitutions, changes and modifications are possible without departing from the spirit and scope of the present invention and the appended claims. Therefore, the present invention should not be limited to the content disclosed in the best embodiment, and the scope of protection claimed by the present invention shall be subject to the scope defined in the claims.

Claims

1. A method for constructing a universal discrete manufacturing industry full-process supervision component library, the steps of which include: 1) Construct a data access module to realize multi-data access; The data access module performs dynamic Bayesian preprocessing based on local reasoning on the accessed data to obtain standardized data that conforms to the database storage structure, and then assigns identifiers to the standardized data based on relational reasoning and stores them in the database; 2) Constructing a BOM list construction module, which is used to match any two data with different identifiers in the database based on a multi-mode matching method, and set the corresponding type of BOM list output rules for the required material quantity and price of materials of the same type according to the identifier; and when new standardized data is stored in the database, clustering the new standardized data with each type of data, so as to obtain the type closest to the new standardized data and the corresponding BOM list output rules, and then use the screened BOM list output rules to construct the BOM list corresponding to the new standardized data; 3) Construct a quality control module to detect the BOM sheets generated during the production process of the production line, and store the test results in the database or push them to the user monitoring platform simultaneously; Among them, a high decoupling mode is adopted to integrate the modules, each module is separately encapsulated and an API interface for parameter transmission is opened; the modules are connected through API interface calls, and the relevant algorithms involving the discrete manufacturing industry process are separately encapsulated in docker, deployed on the cloud, and called through the API interface; the modules include the data access module, the BOM single construction module and the quality supervision module; the database is deployed on the client and joins the MA identification system through identification registration. The database serves as a third-level node of the MA identification system, enabling data to be queried and shared across regions, industries and enterprises.

2. The method according to claim 1, characterized in that The business knowledge deviation correction technology based on progressive adversarial matching is used to detect the types and location information of quality defects in the BOM list. The group training joint transfer framework based on federated contrastive learning is used to detect the types of faults in the BOM list.

3. A server, characterized in that: The method comprises a memory and a processor, wherein the memory stores a computer program, the computer program is configured to be executed by the processor, and the computer program comprises instructions for executing each step in the method according to any one of claims 1 to 2.

4. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 2 are implemented.

Images