An operation and maintenance fault monitoring system modeling method based on a knowledge graph and an OPC UA protocol

Abstract

The application relates to a kind of operation and maintenance fault monitoring system modeling method based on knowledge graph and OPC UA protocol. First, the knowledge and previous processing rules in the operation and maintenance field are summarized, the previous operation and maintenance fault event knowledge of workshop is extracted, the fault data in production is sorted out, the corresponding knowledge base is established, and the concept model of workshop operation and maintenance fault entity and relationship is constructed. Then, the operation and maintenance fault model data source is acquired, the data layer is built, and then the ontology, entity, relationship and attribute are perfected from bottom to top, and the knowledge graph is constructed. Further, based on the OPC UA protocol, the information model is abstracted into the software system, and the operation and maintenance fault information model is instantiated. The whole step is simple and clear, can be applied to various discrete manufacturing production systems, solves the fault handling problem in operation and maintenance scheduling, can express the whole content of operation and maintenance fault monitoring at the same time, and does not affect the subsequent production, and has great application value in the operation and maintenance field.

Description

Technical Field

[0001] This invention relates to the field of operation and maintenance fault modeling in discrete manufacturing, specifically to a modeling method for operation and maintenance fault monitoring systems based on knowledge graphs and the OPC UA protocol. Background Technology

[0002] Traditional discrete manufacturing workshops suffer from heterogeneous production equipment systems, relatively closed production lines, and fragmented production process data. This makes information exchange particularly inconvenient, creating information silos and hindering the interaction between upper-level application systems and CNC equipment. Furthermore, the low efficiency, incomplete structure, and inconsistent formats of production line data acquisition make it difficult to promptly address equipment malfunctions, significantly impacting production processes and economic efficiency.

[0003] Traditional knowledge graph modeling only provides the ability to represent simple data, while OPC UA supports mechanisms to expose complex data semantics. The OPC UA protocol lacks a unified device model definition, and the model definition is not standardized for specific objects; knowledge graphs can precisely fill this gap. Previously, in the modeling process of knowledge graph-based operation and maintenance fault monitoring systems, effective data sampling and monitoring were often difficult to achieve, making integration with the OPC UA protocol challenging. Summary of the Invention

[0004] The purpose of this invention is to provide a modeling method for an operation and maintenance fault monitoring system based on knowledge graphs and the OPC UA protocol, in order to solve the problems mentioned in the background art. This invention divides the model into a conceptual layer and a data layer, mapping the underlying device data sources to corresponding variable nodes, enabling the fault monitoring system model built using the knowledge graph to achieve data storage and retrieval with the data center via the OPC UA protocol.

[0005] To achieve the above objectives, the present invention adopts the following technical solution:

[0006] This invention discloses a modeling method for an operation and maintenance fault monitoring system based on knowledge graphs and the OPC UA protocol, comprising the following steps:

[0007] Step 1: Extract past maintenance failure events from industrial maintenance documents. Based on the consistent and complete address space model and service model provided by the OPC UA protocol, and referring to the integration of DA, object types, reference types, complex data, and commands by the OPC UA protocol, determine the modeling scope from attribute elements to attributes, attribute sets, component sets, and then to production equipment.

[0008] Step 2: Based on the actual needs of the workshop, extract the information of the production line equipment ontology, and construct a conceptual model of workshop operation and maintenance fault entities and relationships through a seven-step method. On the basis of determining the scope of the modeling domain, list the important production line terms to construct a dictionary in the form of manual participation, and sort out the concepts, define the attributes and hierarchies of the classes, so as to realize the standardized description and formal expression of the attributes, concepts, and relationships of the discrete manufacturing workshop.

[0009] Step 3: Obtain the data source of the operation and maintenance fault model, extract specific fault instance information, clarify the definition and description of different types of faults, and build the data layer;

[0010] Step 4: Based on the conceptual model and data layer, refine the ontology, entities, relationships, and attributes from the bottom up to build a knowledge graph;

[0011] Step 5: Referring to the OPC UA protocol specification, abstract the information model into the software system based on XML class diagram modeling to realize the instantiation of the operation and maintenance fault information model;

[0012] Step 6: Determine the main parameters of the production line equipment failure, then determine the processing sequence based on past failure handling patterns and operation and maintenance knowledge, further identify the faulty equipment from the database based on instance information, obtain the processing sequence plan, and complete the selection of the priority faults to be handled.

[0013] Preferably, in step 1, the modeling scope is determined, including the system set, data sampling set, component set, attribute set, and relation set; the system set, component set, attribute set, and relation set together clarify the schema layer of the model; the data sampling set clarifies the data sampling equipment channels and update frequency.

[0014] Preferably, in step 2, based on the actual needs of the workshop, information on the production line equipment is extracted, and a conceptual model of workshop operation and maintenance fault entities and relationships is constructed. The detailed implementation steps are as follows:

[0015] 1) Review industry standards and relevant fault reports and maintenance records to determine the production process of the discrete manufacturing workshop production line, and determine the concepts, modes and equipment attributes of each system in the production line;

[0016] 2) Adopt a bottom-up approach to build a conceptual model, clarify the entities, concepts, attributes, relationships, and characteristics of the production line system and equipment, ensure the standardized description and formal expression of the model, and complete the underlying foundation of the knowledge graph.

[0017] Preferably, in step 3, the data source for the operation and maintenance fault model is obtained, specific fault instance information is extracted, the definitions and descriptions of different types of faults are clarified, and a data layer is built. The implementation steps are as follows:

[0018] 1) For structured data, relational databases generated in the previous production process can be directly used for knowledge graph representation based on the schema layer that has been built above.

[0019] 2) For unstructured data, starting from the analysis of fault causes, fault types and requirements, and using the production line equipment sampling use case table and sampling use case diagram as the data source, the data in the use case diagram and use case table are further processed through industry-standard semantic rules, dictionaries and related templates. Extraction is performed from three dimensions: entity, attribute and relationship. The entity dimension is extracted through specific things on the industrial production line, the data attributes and object attributes are extracted through the attribute information of equipment and data, the relationship dimension is extracted based on the relationship between entities, and the entities are linked based on the relationship between the system and the equipment.

[0020] Preferably, in step 4, based on the conceptual model and data layer, the ontology, entities, relationships, and attributes are improved from the bottom up to construct a knowledge graph. The detailed operation is as follows:

[0021] 1) In the knowledge graph pattern layer, system concepts, equipment entities, fault phenomena, fault handling operations, and fault scenarios are used as nodes, and the relationships between equipment and the system and between different types of faults and their own attributes are used as edges; nodes of different colors represent different entities of maintenance equipment and faults, and edges of different colors represent different attributes and relationships of faulty equipment, operations, and phenomena.

[0022] 2) In the knowledge graph data layer, equipment fault attribute values, fault concepts, and fault instances are used as nodes, and the relationships between equipment and systems and between different types of fault attribute values ​​are used as edges; nodes of different colors represent the attribute values ​​of different maintenance equipment faults, and edges of different colors represent different associations between faulty equipment and attribute values.

[0023] Preferably, the entity linking process first needs to screen out and remove erroneous parts, and then for entities with similar or identical names, the entities are extracted and grouped into a set, and then error correction and ambiguity removal are performed through entity linking.

[0024] Preferably, step 5, referring to the OPC UA protocol specification, abstracts the information model into the software system based on XML class diagram modeling, thereby instantiating the operation and maintenance fault information model. The implementation steps are as follows:

[0025] 1) Create a new project using the visual model generator Ua Modeler, name it, and choose a save path;

[0026] 2) Select the code generation type; Ua Modeler already includes all standard type nodes built into OPC UA. Create a new object type node in the model by selecting Types > ObjectTypes > BaseObjectType under InformationModel; Ua Modeler also allows users to extend their own types in specific fields. Add information types in the Information Model window. Object types must be added under the base object type BaseObjectType.

[0027] 3) Generate an XML file. Select "Export XML" to instantiate the operation and maintenance fault information model.

[0028] Preferably, when creating a new object type node in the model, you can choose to name it and add variables or methods.

[0029] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0030] (1) In response to the problems of information silos and difficulties in information interaction caused by the complex equipment, heterogeneous and multi-source production data, and relatively closed production lines in the discrete manufacturing industry, OPC UA has platform independence. It no longer relies on existing communication technologies, operating systems and programming languages, and is independent of various manufacturers around the world. It adopts a service-oriented architecture and achieves seamless transmission of information and data between sensors and between machines by providing services and describing the meaning of underlying devices.

[0031] (2) Compared with most other protocols currently widely used in industry, the OPC UA protocol has reliable data communication and security mechanisms, and defines various data and structural nodes of the device as objects to describe the device model, in order to support the complex data structure of the device. For production units, it improves the overall automation efficiency, further saves production costs, and obtains better operating benefits.

[0032] (3) Knowledge graphs express concepts, entities, attributes and their semantic relationships in the objective world through directed graphs, enabling interconnection between concepts and entities. They can provide systematic and deep-level structured domain knowledge, constructing human knowledge into a large semantic network that is computer-understandable, computable and reasonable, and providing the possibility of moving from "data intelligence" to "knowledge intelligence".

[0033] (4) By modeling the operation and maintenance fault monitoring system through knowledge graph and OPC UA protocol, it is possible to effectively realize the preventive maintenance and replacement of highly sensitive components, thereby avoiding the occurrence of fault downtime and meeting the needs of flexible deployment, agile production, frequent updates and high concurrency access of the production system. Attached Figure Description

[0034] Figure 1 This is a schematic diagram of the basic implementation framework of the present invention;

[0035] Figure 2 This is a schematic diagram of the production workshop framework based on the OPC UA protocol in this invention;

[0036] Figure 3 This is the overall architecture of the information model for the machine tool operation and maintenance monitoring system in this invention;

[0037] Figure 4 This invention is a model of cooling system information extracted from a knowledge graph in an impeller production line.

[0038] Figure 5 This is an instantiation information model of a cooling system sensor conforming to the OPC UA protocol in this invention. Detailed Implementation

[0039] The technical solution of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the protection scope of the present invention.

[0040] like Figures 1 to 5 As shown, this embodiment describes a modeling method for an operation and maintenance fault monitoring system based on knowledge graphs and the OPC UA protocol, combined with... Figure 1 It includes the following steps:

[0041] Step 1: Extract past maintenance failure events from industrial maintenance documents. Based on the consistent and complete address space model and service model provided by the OPC UA protocol, and referring to the integration of DA, object types, reference types, complex data, and commands by the OPC UA protocol, determine the modeling scope from attribute elements to attributes, attribute sets, component sets, and then to production equipment.

[0042] Step 2: Based on the actual needs of the workshop, extract the information of the production line equipment objects, and construct a conceptual model of workshop operation and maintenance fault entities and relationships through a seven-step method. On the basis of determining the scope of the modeling domain, list the important production line terms to construct a dictionary in the form of manual participation, and sort out the concepts, define the attributes and layers of the classes.

[0043] Step 3: Obtain the data source of the operation and maintenance fault model, extract specific fault instance information, clarify the definition and description of different types of faults, and build the data layer;

[0044] Step 4: Based on the conceptual model and data layer, refine the ontology, entities, relationships, and attributes from the bottom up to build a knowledge graph;

[0045] Step 5: Referring to the OPC UA protocol specification, abstract the information model into the software system based on XML class diagram modeling to realize the instantiation of the operation and maintenance fault information model;

[0046] Step 6: Determine the main parameters of the production line equipment failure, then determine the processing sequence based on past failure handling patterns and operation and maintenance knowledge, further identify the faulty equipment from the database based on instance information, obtain the processing sequence plan, and complete the selection of the priority faults to be handled.

[0047] In a preferred embodiment: Step 1, based on the characteristics of discrete manufacturing workshops and industry standards, deconstructs the digital structure of workshop production equipment and determines the modeling scope based on the OPC UA protocol. A schematic diagram of the production workshop framework based on the OPC UA protocol is shown below. Figure 2 The OPC UA client access control system's OPC UA server address space, starting from the entry node, exposes the target node through the semantics and direction of references, achieving the purpose of data collection and modification. OPC UA provides the possibility of more effectively displaying data semantics, not only providing data, but also representing that it is provided by a specific type of device, and allowing the exposure of the type hierarchy supported by that device.

[0048] In a preferred embodiment: In step 2, based on the actual needs of the workshop, information on the production line equipment is extracted, and a conceptual model of the overall architecture of workshop operation and maintenance fault entities and relationships is constructed, such as... Figure 3 The detailed implementation steps are as follows:

[0049] 1) Review industry standards and relevant maintenance reports and repair records to determine the production process of the discrete manufacturing workshop production line, and determine the concepts, modes and equipment attributes of each system in the production line;

[0050] 2) Adopt a bottom-up approach to build a conceptual model, clarify the entities, concepts, attributes, relationships, and characteristics of the production line system and equipment, ensure the standardized description and formal expression of the model, and complete the underlying foundation of the knowledge graph.

[0051] In a preferred embodiment: Step 3 involves obtaining the data source of the operation and maintenance fault model, extracting specific fault instance information, clarifying the definitions and descriptions of different types of faults, and building a data layer. The implementation steps are as follows:

[0052] 1) For structured data, statistical tables, data documents, relational databases, etc. generated in the previous production process can be directly used for knowledge graph representation based on the above-mentioned established schema layer.

[0053] 2) For unstructured data, starting from the causes of failures, failure types and requirements analysis, the sample use case table or use case diagram of production line equipment can be selected as the data source. The data in the use case diagram and use case table can be further processed through the semantic rules, dictionaries and related templates of industry standards, and extracted from the dimensions of entities, attributes, relationships, etc., and the entities can be linked based on the relationship between the system and the equipment.

[0054] In a preferred embodiment: In step 4, based on the conceptual model and data layer, the ontology, entities, relationships, and attributes are improved from the bottom up to construct a knowledge graph. Detailed operations are as follows:

[0055] 1) Construct a knowledge graph by using system concepts, equipment entities, fault phenomena, fault handling operations, and fault scenarios as nodes, and the relationships between equipment and the system, as well as different types of faults and their own attributes as edges.

[0056] 2) This knowledge graph implements knowledge representation modeling of the ontology, entities, relations, and attributes of the operation and maintenance fault system through the schema layer;

[0057] 3) This knowledge graph uses a data layer to store fragmented knowledge in discrete manufacturing workshop production line fault-related data in a structured form.

[0058] Information model of the cooling system in the impeller production line extracted from the knowledge graph, such as Figure 4 .

[0059] In a preferred embodiment: In step 2), entity linking first needs to screen out error parts, such as duplicate information, ambiguous information, and incorrect information. Then, for entities with similar or identical names, the above entities are extracted and formed into a set N = {"water chiller", "shaft", "water tank", ...}. After that, error correction and ambiguity removal are performed through entity linking.

[0060] In a preferred embodiment: Step 5, referring to the OPC UA protocol specification, uses UAModeler to abstract the information model into the software system based on XML class diagram modeling, thereby instantiating the operation and maintenance fault information model. The implementation steps are as follows:

[0061] 1) Use the UA Modeler software to create a new project, name it, and choose a save path;

[0062] 2) Select the code type to generate. Since we are using UA Modeler with the ultimate goal of generating XML files, the code type can be selected arbitrarily.

[0063] 3) Create a new object type node in the model. Under InformationModel, select Types > ObjectTypes > BaseObjectType to create it. You can name it and add variables or methods to it.

[0064] 4) UAModeler provides a graphical design interface for address spaces, enabling graphical design of information models. Users can quickly add nodes and references between nodes in the graphical user interface to rapidly implement complex models. Furthermore, the created information model can automatically generate corresponding XML files. After establishing the graphically structured information model, UAModeler automatically generates the corresponding XML description file and imports it into the address space of the OPC UA server. Sensor type instantiation is as follows... Figure 5 The sensor type describes all sensors used in manufacturing the Internet of Things (IoT). In addition to common variables such as serial number and model number, the sensor type should also include two variable nodes: reading and engineering units. The definition of sensor types is shown in Table 1.

[0065] Table 1: Definition of Sensor Types

[0066]

[0067] In a preferred embodiment: Step 6 determines the main parameters of the production line equipment fault, then determines the processing sequence based on past fault handling patterns and maintenance knowledge, and further identifies the faulty equipment from the database based on instance information. The implementation steps are as follows:

[0068] 1) The PLC connects to the OPC server via Siemens TCP / IP Ethernet, matches fault alert I / O and intermediate variables on the PLC, establishes tags and groups them.

[0069] 2) Import the node IDs of the established tags and corresponding equipment fault information into the data center. The data center manages the PLCs of the devices and connects to all information points, including tags, node IDs, addresses, descriptions, and regions, grouping them by equipment and production line. The OPC server scans the PLCs cyclically via subscription.

[0070] 3) Retrieve all node IDs in the data center and request the OPC server to subscribe. If a change occurs at a PLC monitoring point, the OPC server accepts the change by scanning and sends the corresponding node ID and value to the monitoring system terminal. The system will confirm whether this value change indicates a fault, retrieve the corresponding fault information from the data center, and display it on the terminal.

[0071] 4) The configuration and deployment of the equipment fault monitoring system enables comprehensive real-time monitoring of the corresponding workshop production line. In this example, the system can scan 973 monitoring points in real time at 100 millisecond intervals.

[0072] Equipment operation diagram of the equipment fault monitoring system (e.g.) Figure 4 As shown, the software intuitively reflects the operating status of the equipment and the location of the fault. It changes the previous monotonous text descriptions, presenting fault information with both text and graphics, facilitating quick and accurate judgment by operators and maintenance personnel. The fault information collected by the software generates a complete log file, recording the duration of the equipment fault in detail, providing a basis for statistics, tracing, and analysis of equipment fault causes.

[0073] By constructing a production line knowledge graph through data collection and fusion, and then instantiating the information model through the OPC UA protocol, fragmented knowledge in the production line fault-related data of discrete manufacturing workshops is stored in a structured form. This is of great significance for monitoring the production status of discrete manufacturing workshops, spare parts reserves, preventive maintenance, reducing the probability of failure, and improving the reliability of equipment operation.

[0074] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely preferred examples and are not intended to limit the invention. Various changes and modifications can be made to the present invention without departing from its novel spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of the present invention is defined by the appended claims and their equivalents.

Claims

1. A modeling method of an operation and maintenance fault monitoring system based on a knowledge graph and an OPC UA protocol, characterized in that, Comprise the following steps: Step 1: Extract the past operation and maintenance fault event knowledge of the workshop from the industrial overhaul document, determine the modeling range from attribute elements to attributes, attribute sets, component sets to production equipment based on the consistent and complete address space model and service model provided by the OPC UA protocol, and refer to the integration of DA, object type, reference type, complex data and command of the OPC UA protocol; Step 2: Extract the production line equipment ontology object information in combination with the actual requirements of the workshop, construct the concept model of workshop operation and maintenance fault entity and relationship through the seven-step method, list the important terms of the production line in the form of manual participation to construct the dictionary on the basis of determining the modeling field category, and comb the concept, define the attributes and layers of the class; Step 3: Obtain the operation and maintenance fault model data source, extract the specific instance information of the fault, and clarify the definition and description of different types of faults to build the data layer; Step 4: On the basis of the concept model and the data layer, perfect the ontology, entity, relationship and attribute from bottom to top, and construct the knowledge graph; the detailed operation is as follows: 1) In the mode layer of the knowledge graph, the system concept, equipment entity, fault phenomenon, fault handling operation and fault scene are taken as nodes, and the relationship between the equipment and the system and different types of faults and the attributes thereof are taken as edges; different colored nodes represent different ontology and entity of the operation and maintenance equipment and faults, and different colored edges represent different attributes and relationships of the fault equipment, operation and phenomenon; 2) In the data layer of the knowledge graph, the equipment fault attribute value, fault concept and fault instance are taken as nodes, and the relationship between the equipment and the system and different types of fault attribute values is taken as edges; different colored nodes represent different attribute values of the operation and maintenance equipment fault, and different colored edges represent different associations of the fault equipment and attribute values; Step 5: Refer to the OPC UA protocol specification, abstract the information model to the software system based on the XML class diagram modeling, and realize the instantiation of the operation and maintenance fault information model; Step 6: Determine the main parameters of the production line equipment fault, then determine the processing sequence based on the past handling fault rule and operation and maintenance knowledge, further determine the fault equipment based on the instance information from the database, obtain the processing sequence scheme, and complete the selection of the priority handling fault.

2. The knowledge graph and OPC UA protocol-based operation and maintenance fault monitoring system modeling method according to claim 1, characterized in that: The modeling range determined in the step 1 includes system set, data sampling set, component set, attribute set and relationship set; the system set, component set, attribute set and relationship set jointly illustrate the mode layer of the model; the data sampling set clearly illustrates the sampling equipment channel and update frequency of the data.

3. The knowledge graph and OPC UA protocol-based operation and maintenance fault monitoring system modeling method according to claim 1, characterized in that: In the step 2, the production line equipment ontology object information is extracted in combination with the actual requirements of the workshop, and the concept model of workshop operation and maintenance fault entity and relationship is constructed, and the detailed implementation steps are as follows: 1) Refer to the industry specifications and related fault reports and maintenance records to determine the production process of the discrete manufacturing workshop production line, and determine the system concept, mode and equipment attribute of the production line; 2) Establish the concept model in a bottom-up form, clarify the entity, concept, attribute, relationship and feature of the production line system and equipment, ensure the standardized description and formalized expression of the model, and complete the bottom foundation of the knowledge graph.

4. The knowledge graph and OPC UA protocol-based operation and maintenance fault monitoring system modeling method according to claim 1, characterized in that: The step 3 obtains the operation and maintenance fault model data source, extracts fault specific instance information, defines and explains different types of faults, and builds a data layer, and the implementation steps are as follows: 1) For structured data, the relational database generated in the past production process is directly used for knowledge graph expression based on the above built mode layer; 2) For unstructured data, taking fault reason, fault type and demand analysis as the starting point, taking the sampling case table and sampling case diagram of the production line equipment as the data source, further processing the data in the case diagram and case table through the semantic rules, dictionaries and related templates of the industry specifications, extracting from three dimensions of entity, attribute and relationship, extracting the entity dimension through the specific things on the industrial production line, extracting the data attribute and object attribute through the attribute information of the equipment and data, extracting the relationship dimension based on the relationship between entities, and linking the entities based on the relationship between the system and the equipment.

5. The knowledge graph and OPC UA protocol-based operation and maintenance fault monitoring system modeling method according to claim 4, characterized in that: Entity linking needs to first screen out the error part, and then for entities with similar or identical names, extract the above entities and form a set, and then correct errors and remove ambiguities through entity linking.

6. The knowledge graph and OPC UA protocol-based operation and maintenance fault monitoring system modeling method according to claim 1, characterized in that: The step 5 refers to the OPC UA protocol specification, abstracts the information model into the software system based on the XML class diagram modeling, and realizes the instantiation of the operation and maintenance fault information model, and the implementation steps are as follows: 1) Use the visual model generator Ua Modeler to create a new project, name it and select the save path; 2) Select the code generation type; Ua Modeler contains all standard type nodes built in OPC UA, create a new object type node in the model, select Types>ObjectTypes>BaseObjectType under InformationModel, and create it; Ua Modeler also allows users to extend their own types in specific fields, add information types under the Information Model window, and object types must be added under the basic object type BaseObjectType; 3) Generate an XML file, select Export XML, and the operation and maintenance fault information model instantiation is realized.

7. The knowledge graph and OPC UA protocol-based operation and maintenance fault monitoring system modeling method according to claim 6, characterized in that, When creating a new object type node in the model, select its name, add variables or methods.

Images