A problem rolling unit discovery method based on man-machine object ternary data fusion

Abstract

This invention relates to a method for discovering problematic rolling units based on the fusion of human, machine, and material data. The method includes: constructing a hot rolling scheduling ontology based on the human, machine, and material data; loading hot rolling scheduling instances into the hot rolling scheduling ontology; obtaining roll scheduling information using SPARQL statements based on the hot rolling scheduling ontology; obtaining a scheduling plan using an ant colony algorithm based on the roll scheduling information; and discovering problematic rolling units based on the scheduling plan using inference rules generated according to SWRL rules. Compared with existing technologies, this invention can quickly identify affected rolling units when uncertain events occur, which is beneficial for rapid response to emergencies and facilitates refined management of hot rolling production scheduling in the steel industry.

Description

Technical Field

[0001] This invention relates to the field of data fusion technology, and in particular to a method for discovering problem rolling units based on the fusion of human, machine, and object data. Background Technology

[0002] Data fusion originated in the military field and was subsequently applied to commerce and industry, such as in industrial fault detection. At this stage, the fusion primarily focused on sensor data, mainly occurring in the physical space. With the development of the internet and information technology, the fusion of "hardware" and "software" data, as well as the fusion of human-machine-object data, has emerged. Numerous studies have been conducted on data fusion, gradually extending from single spaces to multiple spaces (social space, cyberspace). Regarding data in the physical space, based on the "data-information-knowledge-intelligence" principle, much research has been carried out, such as predictive maintenance of equipment, prediction of remaining equipment lifespan, and recommendation of production-related process parameters. The rapid development of the internet is accelerating the flow and sharing of knowledge among virtual network data, thus extending data knowledge management from the physical level or within enterprises to the external world, profoundly impacting the entire data knowledge management system. Humans possess strong abilities of understanding, perception, reasoning, and learning, enabling them to handle many complex problems and compensate for the limitations of physical sensors, fully considering semantic information such as entity relationships. In addition to data computation, data fusion should also be able to perform complex logical reasoning. Ontologies, with their ability to share and jointly understand domain knowledge between humans and computers, as well as between machines, provide the possibility of implementation. As a semantic data description and sharing method, ontology mainly includes concept classes, relations, axioms, instances, and functions. It can solve problems such as semantic redundancy and data heterogeneity, and plays an important role in knowledge-intensive contexts. It has been widely used to solve problems such as data organization, representation, and fusion.

[0003] In the hot rolling process of steel production, steel billets from continuous casting machines or slab warehouses are heated in furnaces to reach the required temperature before being transported to continuous rolling mills to produce finished or semi-finished products. Hot rolling scheduling primarily addresses the hot rolling stage, determining the processing sequence of rolling units after the rolling units are divided to meet constraints and production targets. Currently, factors commonly considered in hot rolling scheduling include production skip penalties, order delivery dates, and equipment maintenance. Specific methods are often used to solve hot rolling scheduling problems under specific constraints. However, in the hot rolling process of steel production, uneven temperature within the furnace and malfunctions such as linear expansion of the furnace beams can affect slab heating. Customers may also have new demands. In such cases, the original scheduling plan becomes inapplicable, and the resulting scheduling arrangement becomes unsuitable. Summary of the Invention

[0004] The purpose of this invention is to overcome the shortcomings of the existing technology by providing a method for discovering problematic rolling units based on the fusion of human, machine, and material data. This method can quickly identify affected rolling units when uncertain events occur, which is beneficial for rapid response to emergencies and helps to improve the refined management of hot rolling production scheduling in the steel industry.

[0005] The objective of this invention can be achieved through the following technical solutions:

[0006] A method for discovering problematic rolling units based on the fusion of human-machine-object three-dimensional data includes:

[0007] Construct a hot rolling scheduling ontology based on the three elements of human, machine, and material;

[0008] Load the hot rolling scheduling instance into the hot rolling scheduling body;

[0009] Based on the hot rolling scheduling entity, roll scheduling information is obtained through SPARQL statements;

[0010] Based on the roll scheduling information, a scheduling plan is obtained using the ant colony algorithm;

[0011] Based on the scheduling plan, the problematic rolling unit is identified by reasoning using inference rules generated according to SWRL rules.

[0012] Furthermore, the process of constructing the hot rolling scheduling ontology includes:

[0013] By constructing a hierarchical structure of scheduling-related concepts, formalizing the hot rolling production order, decomposing the production task, determining the rolling unit involved in the production order and its specification attributes, determining the task under uncertain conditions, and then determining the class and subclass in the ontology. The uncertain conditions include equipment failure and emergency orders.

[0014] Define attributes and constraints, including object attributes, data attributes, and annotation attributes.

[0015] Furthermore, the aforementioned human-machine-object three-dimensional data includes human data, machine data, and object data.

[0016] Furthermore, the data on the individuals mentioned includes domain expert experience, user needs, and evaluations.

[0017] Furthermore, the machine's data includes document tables, databases, and calculation results.

[0018] Furthermore, the data of the objects includes equipment operating status data, material data, and workshop environment data collected by sensors.

[0019] Furthermore, the specific process of loading the hot rolling scheduling instance into the hot rolling scheduling body includes:

[0020] The hot rolling scheduling instance can be loaded into the hot rolling scheduling ontology via OWLAPI, or the hot rolling scheduling instance can be built into the hot rolling scheduling ontology via Protégé and then loaded into the computer via OWLAPI.

[0021] Furthermore, the construction process of the hot rolling scheduling instance includes:

[0022] The order is broken down into work type, operation items under work type, and rolling production line under operation item. The work type only includes the hot rolling stage, the operation item only includes rolling, and the rolling production line is set to one.

[0023] An electronic device includes a memory and a processor, the memory storing a computer program, the processor being able to execute the problem rolling cell discovery method as described above by invoking the program instructions.

[0024] A computer-readable storage medium includes a computer program that can be executed by a processor to implement the problem rolling cell discovery method.

[0025] Compared with the prior art, the present invention has the following beneficial effects:

[0026] This invention presents a method for identifying problematic rolling units. It constructs a hot rolling scheduling ontology based on human, machine, and material (HMI) data, loads hot rolling scheduling instances into this ontology, and uses SPARQL statements to obtain roll scheduling information. Based on this information, it obtains a scheduling plan using an ant colony algorithm. Then, based on the scheduling plan, it uses inference rules generated according to SWRL rules to identify problematic rolling units. This invention addresses the current lack of data organization, representation, and fusion in enterprises. It analyzes the characteristics of HMI data in the manufacturing field, studies an ontology-based data fusion framework, constructs a scheduling-oriented ontology model, and performs scheduling-related HMI data fusion. When equipment malfunctions or emergency orders are added, the original scheduling plan may need to be changed, potentially affecting rolling units. By inferring and identifying affected problematic rolling units based on SWRL, it quickly identifies affected rolling units in the event of uncertain events, facilitating rapid response to emergencies and further development of scheduling plans, thus contributing to the refined management of hot rolling production in the steel industry. Attached Figure Description

[0027] Figure 1 For scheduling ontology class hierarchy diagram;

[0028] Figure 2 This is a diagram illustrating the data attributes.

[0029] Figure 3 A diagram illustrating object properties;

[0030] Figure 4 The following diagram is provided as an example.

[0031] Figure 5 This is a scheduling diagram for the rolling unit;

[0032] Figure 6 This is a diagram illustrating the reasoning results;

[0033] Figure 7 This is a schematic diagram showing the location of the affected rolling units in the rolling unit scheduling diagram;

[0034] Figure 8 A flowchart of the method for identifying problematic rolling units. Detailed Implementation

[0035] The present invention will now be described in detail with reference to the accompanying drawings and specific embodiments. These embodiments are based on the technical solution of the present invention and provide detailed implementation methods and specific operating procedures. However, the scope of protection of the present invention is not limited to the following embodiments.

[0036] Example 1

[0037] A method for discovering problem rolling units based on the fusion of human-machine-object three-dimensional data, such as Figure 8 This includes the following steps:

[0038] S1. Human-Machine-Object Three-Dimensional Data Analysis: The fusion of human-machine-object three-dimensional data should be approached from the perspective of diverse data structures and large scale, taking into account the fusion objectives when considering the model framework structure. Given the diverse data sources and structures, a method capable of representing multi-source heterogeneous data needs to be selected. The human-machine-object three-dimensional data structures differ, and semantic ambiguity exists. Ontologies, with their well-defined concepts, hierarchical structure, knowledge reuse, and reasoning capabilities, are suitable for human-machine-object three-dimensional data fusion.

[0039] S2. Ontology Construction Techniques: To construct an ontology related to hot rolling scheduling, it is necessary to analyze the concepts and relationships involved from a holistic perspective. This involves hierarchically constructing scheduling-related concepts to determine the classes and subclasses within the ontology. The conceptual relationships related to scheduling are then clarified to determine the object attributes within the ontology. Taking scheduling as the research object, the ontology is constructed by extracting its concepts and relationships.

[0040] S3. Instance Generation: After the conceptual model of scheduling is built, scheduling instances need to be loaded into the conceptual model according to the specific scheduling problem. This can be done through OWLAPI, or it can be built in Protégé and then loaded into the computer through OWLAPI.

[0041] S4. Scheduling Plan Generation: The ontology integrates scheduling-related human, machine, and object data, preserving original scheduling concepts and semantic information. This facilitates data querying and use. SPARQL statements are used to retrieve information such as all units requiring rolling and delivery dates. The ontology achieves human-machine consensus on data, making it easy to check roll types when changing different rolling units. Furthermore, roll information and rolling time for different specifications of rolling units are stored in the ontology library for future use. Ant colony optimization (ACO) is an optimization algorithm that simulates the foraging behavior of real ants. It can update probabilities to enhance the algorithm's global search capability and supports distributed computing. Therefore, this embodiment uses ACO for initial scheduling generation.

[0042] S5. Semantic Rule Writing: Semantic Web Rule Language (SWRL) was proposed by the World Wide Web Consortium (W3C) in 2004 to describe reasoning rules. As a language that represents rules semantically, SWRL's rule concept evolved from RuleML (Rule Markup Language) and was combined with the OWL ontology. SWRL rules are written based on human experience;

[0043] S6. Human-machine-object three-dimensional data fusion, problem rolling unit discovery: Finally, when an uncertain event occurs, the problem rolling unit discovery is triggered.

[0044] Step S1 specifically includes the following steps:

[0045] S11. Determine the three key data points of human-machine-thing (HMI), and integrate these data points. This mainly includes relevant manufacturing data, which, from the perspective of the entire product lifecycle, includes design, manufacturing, maintenance, and service stages. In the product design stage, the designer's ideas and the user's actual needs are transformed into detailed descriptions. In the manufacturing stage, the concept is processed into a product according to the design process. In the maintenance and service stage, the manufacturer provides remote operation and maintenance services for the product.

[0046] S12. Analysis of Human-Machine-Thing (HMT) Data Characteristics: The above analysis reveals that data sources and types are becoming increasingly diverse. Focusing solely on production system data as in the past would negate the significance of enterprise datafication. The development of information technology has led to a surge in internet data and the emergence of new technologies such as big data and cloud computing, incorporating data from the internet, social networks, and the Internet of Things to enhance perception and decision-making. From the perspective of HMT data, human data primarily originates from the social space, including domain expert experience, user needs, and evaluations, mainly expressed in textual form with rich semantic information. Compared to mining knowledge from big data, expert experience has a higher value density. Typically, user needs and expert experience do not overlap, nor do maintenance expert experience and production scheduling experience, resulting in weak correlations between data. Machine data primarily originates from cyberspace, including documents, tables, databases, and calculation results. Enterprise information systems contain a large amount of data in various formats, leading to its diversity, heterogeneity, and massive volume. Material data primarily originates from the physical space, including equipment operating status data collected by sensors, material data, and workshop environmental data. Sensors generate a large amount of data every day, and this data changes over time, mainly characterized by its massive volume and real-time nature.

[0047] S13. Advantages of using ontology for data fusion: Semantic support: The conceptual hierarchy of an ontology can preserve the original semantic information of the data. Standardized representation: Ontologies have standardized representation forms and strict definitions. Using ontology can standardize the representation of raw data, facilitating subsequent applications. Human-computer consensus: Using ontology for data fusion allows relevant personnel to easily obtain ontology content by viewing the ontology. Furthermore, computers can also read ontology content to obtain relevant information. Knowledge reuse: As a formal representation method, ontology can preserve data information. In addition, the ontology can be extended and updated to continuously improve it.

[0048] Step S2 specifically includes the following steps:

[0049] S21. Determine the application scope. This embodiment mainly integrates human, machine, and material data by constructing an ontology, and then applies it to hot rolling scheduling-related issues. This includes the formal representation of hot rolling production orders, the decomposition of production tasks, the rolling units involved in the production orders and their specifications, as well as some tasks under uncertain circumstances such as equipment failure or emergency orders.

[0050] S22. Considering reuse, this stage mainly involves studying existing related ontology and literature on related topics.

[0051] S23. Define classes and hierarchical structures. By constructing hierarchical structures for scheduling-related concepts, the classes and subclasses in the ontology are determined.

[0052] S24. Define attributes and constraints. In this step, identify the terms used to enforce relationships between classes, i.e., select terms that belong to object attributes. Object attributes are used to modify the relationships between classes or instances. Object attributes have their own domain and range.

[0053] S25. Create an instance. An instance is used to represent a specific element. For each instance, select the class to which it belongs so that object properties, data properties, and annotation properties can be bound.

[0054] Step S3 specifically includes the following steps:

[0055] S31. The ontology is instantiated. Hot rolling scheduling is a small part of the steel rolling process, requiring the decomposition of existing orders into tasks, also known as the hot rolling batch planning problem. Only after this step are processable hot rolling units obtained, and hot rolling scheduling mainly occurs at this stage. Therefore, the orders received by the enterprise will be divided into processing orders (ProcessingOrders), and then these orders will be divided into different jobs. In the hot rolling stage, jobs are divided into operations, which only include the rolling step. If the enterprise has multiple rolling production lines, there will be different processes. However, rolling units have already been grouped together by proximity during planning, so it is assumed that each operation has only one selectable process.

[0056] Step S4 specifically includes the following steps:

[0057] S41. Rolling Unit Data Acquisition: To manipulate RDF / OWL format data, the W3C designed a standard query language called SPARQL, considered a key technology of the Semantic Web. Using SPARQL requires understanding information about the entities in the query and the language syntax. SPARQL typically consists of a series of triples containing variables, comprising two parts: a SELECT operator to identify the variables to appear in the query and a WHERE clause, providing the graph schema that the subgraph storing the RDF data should match. It can also have additional operators for filtering and grouping results, and supports operations such as counting and aggregation.

[0058] S42. The original scheduling plan for production uses the ant colony algorithm, a heuristic algorithm that utilizes the positive feedback mechanism of ants releasing pheromones to find the optimal solution. Ants will move along the direction with higher pheromone concentrations. In scheduling, the Operation processing priority knowledge (OPPK) is learned and updated from the scheduling solutions generated in each iteration. Given the rolling sequence of the rolling units and the Machine selects priority knowledge (MSPK), the algorithm determines which rolling line to select for a particular rolling unit. In this embodiment, it is assumed that there is only one line. The algorithm is then continuously optimized to find the optimal solution.

[0059] Step S5 specifically includes:

[0060] To enhance the expressive power of the original model, appropriate rules need to be written. Based on this embodiment, the following rules are required:

[0061] Rule 1 (from which the rolling units required for the Operation can be obtained):

[0062]

[0063] Rule 2 (This rule determines the source of the workpieces required for the Operation):

[0064]

[0065] Rule 3 (from which the potentially affected operations can be deduced):

[0066]

[0067] Step S6 specifically includes the following steps:

[0068] S61. Human-machine-object three-data fusion: The original scheduling plan information obtained above and the written SWRL rules are integrated into the ontology model.

[0069] S62. Problem rolling unit discovery: Apply ontology reasoning capabilities to obtain the affected rolling units.

[0070] Specific examples:

[0071] Step S1 involves the analysis of three key data features of humans, machines, and objects to support the ontology construction in step S2. Therefore, the specific implementation begins with step S2.

[0072] Following the above construction process, construct the scheduling-related ontology class hierarchy in Protégé, as follows: Figure 1As shown, the data related to people in scheduling includes DispatchingRule and ProductionObject, while machine data mainly includes SalesOrder, Process, Job, and Operation. Material data includes EquipmentFailure, Part, and Product. By integrating these three data types (human, machine, and material) through an ontology model, scheduling-related data can be considered more comprehensively, giving the solution method better adaptability and the ability to handle uncertainties.

[0073] Continue building object properties and data properties in Protégé. Figure 2 These are data attributes, which are mostly numerical data used to represent information related to the rolling unit.

[0074] Figure 3 As an object attribute, it is used to represent the relationship between classes. Scheduling involves the three elements of human, machine, and object data, and the semantic information between the three elements of data is preserved by the ontology object attributes.

[0075] S3, Instance Generation

[0076] Based on the production scenario of hot rolling of steel, simulation experiments were conducted to verify the feasibility and effectiveness of the proposed method model. In the experiment, the company has three ProcessingOrders that need to be processed, each order is a Job, and each Job contains six RollingUnits that need to be rolled, requiring 18 Operations. Each Process has different rolls, as shown in Table 1. Each Job and Operation has a due date, and each Process also has a processing time. The specific rolls required for each rolling unit are shown in Table 1.

[0077] Table 1. Rolls required for specific rolling units.

[0078]

[0079] The first row of the table contains 18 Operations and 18 rolling units. The first column is the roll number, which represents the selectable rolls for each rolling unit. A value of 1 indicates that a rolling unit requires a particular roll; otherwise, it's 0. The instantiated result is as follows. Figure 4 As shown.

[0080] The specific rolling unit schedule is shown in Table 2:

[0081] Table 2: Timetable for specific rolling units

[0082]

[0083] The first row in the table is the specific rolling unit number, which is 18 rolling units. The second row is the job release time, i.e., the time when the order was obtained. In this example, it is assumed that the order was obtained from any day in the previous week, i.e., releasetime = random.randint(1,7). The third row is the rolling time required for the operation, which is assumed to be a number between 3 and 10, i.e., processingtime = random.randint(3,10). The last row is the due time, which is based on the total work content. Therefore, in this example, it is automatically generated by multiplying the processing time by six, i.e., duetime = a * random.randint(18,60), where 'a' is a leniency coefficient. Each due time is obtained by sequentially adding the order release time.

[0084] S4. Scheduling Plan Generation: Simulation experiments were conducted on the instantiated rolling units using the algorithm model described above. A total of 18 rolling units were instantiated, and experiments were performed using different numbers of rolling units. The experimental system used a Windows 10 operating system, 8GB of RAM, an Intel Core i5-5200U / 2.2GHz processor, and Java language with an IntelliJ IDE environment.

[0085] The scheduling order obtained by applying the ant colony algorithm is as follows: Figure 5 As shown.

[0086] S6. Integration of human-machine-object three-dimensional data for problem detection in rolling units.

[0087] The above rules and scheduling order are updated in the ontology instance. Taking the failure of Furnace3 as an example, with the occurrence time set to 100, ontology reasoning query can obtain the affected units, such as... Figure 6 As shown. The affected rolling units are as follows: Figure 7 As shown ( Figure 7 The darker parts (in the middle) could not be rolled on time due to a malfunction in the heating furnace.

[0088] Example 2

[0089] An electronic device includes a memory and a processor, the memory storing a computer program, the processor calling program instructions to execute the problem rolling cell discovery method as described in Embodiment 1.

[0090] Example 3

[0091] A computer-readable storage medium includes a computer program that can be executed by a processor to implement the problem rolling cell discovery method described in Embodiment 1.

[0092] Examples 1, 2, and 3 propose a method, electronic device, and medium for discovering problematic rolling units based on the fusion of human-machine-object (HMI) data. Addressing the current lack of data organization, representation, and fusion in enterprises, these examples analyze the characteristics of HMI data in the manufacturing field, study an ontology-based data fusion framework, construct a scheduling-oriented ontology model, and perform scheduling-related HMI data fusion. Then, when equipment malfunctions or emergency orders are added, the original scheduling plan may need to be changed, potentially affecting rolling units. The method uses SWRL (Survey-Survey-Related Riddles) to infer and discover affected problematic rolling units.

[0093] The preferred embodiments of the present invention have been described in detail above. It should be understood that those skilled in the art can make numerous modifications and variations based on the concept of the present invention without creative effort. Therefore, all technical solutions that can be obtained by those skilled in the art based on the concept of the present invention through logical analysis, reasoning, or limited experimentation on the basis of existing technology should be within the scope of protection defined by the claims.

Claims

1. A method for discovering problem rolling units based on the fusion of human-machine-thing ternary data, characterized in that Comprising: Constructing a hot rolling scheduling ontology based on the ternary data of human, machine, and thing; Loading the hot rolling scheduling instances into the hot rolling scheduling ontology; Obtaining the roll scheduling information according to the hot rolling scheduling ontology through SPARQL statements; Obtaining the scheduling plan through the ant colony algorithm according to the roll scheduling information; Reasoning according to the scheduling plan through the inference rules generated according to SWRL rules to discover the problem rolling units.

2. The method for discovering a problem rolling unit based on the fusion of human-machine-matter ternary data according to claim 1, wherein The process of constructing the hot rolling scheduling ontology includes: Through hierarchical construction of scheduling-related concepts, formal representation of hot rolling production orders, decomposition of production tasks, determination of the rolling units involved in the production orders and their specification attributes, determination of tasks under uncertain circumstances, and further determination of the classes and subclasses in the ontology. The uncertain circumstances include equipment failures and emergency orders; Defining properties and constraints. The properties include object properties, data properties, and annotation properties.

3. The method for discovering a problem rolling unit based on the fusion of human, machine, and object ternary data according to claim 1, wherein The ternary data of human, machine, and thing includes human data, machine data, and thing data.

4. The method for discovering a problem rolling unit based on the fusion of human-machine-thing ternary data according to claim 3, wherein The human data includes domain expert experience, user requirements, and evaluations.

5. A method for discovering a problem rolling unit based on the fusion of human-machine-matter ternary data according to claim 3, characterized in that, The machine data includes document tables, databases, and calculation results.

6. The method for discovering a problem rolling unit based on the fusion of human-machine-thing ternary data according to claim 3, wherein The thing data includes equipment operation status data collected by sensors, material data, and workshop environment data.

7. A method for discovering a problem rolling unit based on the fusion of human-machine-thing ternary data according to claim 1, characterized in that, The specific process of loading the hot rolling scheduling instances into the hot rolling scheduling ontology includes: Loading the hot rolling scheduling instances into the hot rolling scheduling ontology through OWLAPI, or constructing the hot rolling scheduling instances into the hot rolling scheduling ontology through Protégé and then loading them into the computer through OWLAPI.

8. A method for discovering a problem rolling unit based on the fusion of human-machine-thing ternary data according to claim 1, characterized in that, The construction process of the hot rolling scheduling instances includes: Decomposing the order into work types, operation items under the work types, and rolling production lines under the operation items. The work types only include the hot rolling stage, the operation items only include rolling, and the rolling production line is set to one.

9. An electronic device, characterized in that, Comprising a memory and a processor. The memory stores a computer program, and the processor can execute the problem rolling unit discovery method as described in any one of claims 1 to 8 by invoking the program instructions.

10. A computer-readable storage medium, characterized in that, Comprising a computer program that can be executed by a processor to implement the problem rolling unit discovery method as described in any one of claims 1 - 8.

Images