Insulation board production line fault diagnosis method based on knowledge graph

By building a fault diagnosis knowledge graph for insulation board production line and using deep learning models for fault diagnosis, the problem of complexity of fault diagnosis in multi-device and multi-sensor environment is solved, and the diagnosis accuracy and accuracy of maintenance suggestions are improved under the conditions of missing signs. .

CN118917398APending Publication Date: 2024-11-08CHINA UNIV OF MINING & TECH

Patent Information

Application Number
CN202411045652.7
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-01
Publication Date
2024-11-08

AI Technical Summary

Technical Problem

In the insulation board production line, due to the multi-device, multi-sensor, and multi-source heterogeneity of fault data, the entity extraction is complex, and it is difficult to build a fault diagnosis knowledge graph, and the existing methods have low accuracy in diagnosis results when the fault signs are missing.

Method used

Using a fault diagnosis method based on knowledge graph, the fault diagnosis knowledge graph of the insulation board production line is constructed, and a strategy construction model that combines bottom-up and top-down, combined with the deep learning model BiLSTM+CRF for entity and relationship extraction, establishing A complete fault diagnosis knowledge graph, and introduce direct and indirect fault causes through the inference diagnosis process, giving sorted repair suggestions.

Benefits of technology

Improve the diagnostic accuracy rate under the conditions of missing signs, can effectively extract knowledge from equipment maintenance logs, expert knowledge bases and literature materials, analyze the causes of equipment failures and give maintenance suggestions, support query functions and input of new fault phenomena, and improve Troubleshooting knowledge graph.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118917398A_ABST
    Figure CN118917398A_ABST
Patent Text Reader

Abstract

The invention discloses an insulation board production line fault diagnosis method based on a knowledge graph. The insulation board production line fault diagnosis method comprises the steps of construction of an insulation board production line fault diagnosis knowledge graph and fault diagnosis reasoning based on the insulation board production line fault diagnosis knowledge graph. According to the invention, an indirect fault reason is introduced through the relationship among fault diagnosis reasoning, data measuring points, an equipment structure and a fault phenomenon, so that the problem of low accuracy of a traditional fault diagnosis method under the condition of symptom deficiency can be solved; knowledge extraction can be effectively carried out from an insulation board production line equipment maintenance log, an expert knowledge base, literature data and equipment technical documents, and when a user carries out insulation board production line fault diagnosis, production line equipment fault reasons are analyzed, and maintenance suggestions are given.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to a thermal insulation board production line fault diagnosis method, specifically to a thermal insulation board production line fault diagnosis method based on a knowledge graph, and belongs to the technical field of production line fault diagnosis. Background Art

[0002] Initially, the concept of knowledge graph was proposed for search engine services. However, with the development of technology, knowledge graph technology has been widely used in the medical field, e-commerce field, domain knowledge management and other fields.

[0003] The application of knowledge graph technology to the field of fault diagnosis can accurately and comprehensively describe fault diagnosis knowledge and its connections. Therefore, fault diagnosis based on knowledge graph has received widespread attention. However, there may be two problems in the application of knowledge graph technology to the field of fault diagnosis of insulation board production lines: on the one hand, due to the fact that insulation board production lines have multiple devices, multiple sensors, and multi-source heterogeneous fault data, entity extraction is complex, the construction of fault diagnosis knowledge graphs is difficult, and there is a lack of knowledge graph cases for fault diagnosis of insulation board production lines; on the other hand, the existing methods have the problem of low accuracy of diagnostic results when fault signs are missing. Summary of the invention

[0004] The purpose of the present invention is to provide a fault diagnosis method for an insulation board production line based on a knowledge graph to solve the problems raised in the above-mentioned background technology.

[0005] In order to solve the above technical problems, the present invention provides the following technical solutions: a fault diagnosis method for an insulation board production line based on a knowledge graph, the fault diagnosis method for an insulation board production line comprises the following steps:

[0006] S1. Construction of knowledge graph for fault diagnosis of insulation board production line

[0007] The construction of the fault diagnosis knowledge graph includes:

[0008] S11. Model construction of fault diagnosis knowledge graph for insulation board production line;

[0009] S12, performing knowledge ontology representation on the knowledge graph of fault diagnosis of insulation board production line according to the constructed model;

[0010] S13, extracting knowledge ontology from the insulation board production line fault diagnosis knowledge graph according to the represented knowledge ontology;

[0011] S14, constructing nodes and edges in the insulation board production line fault diagnosis knowledge graph based on the extracted knowledge ontology, and establishing a complete fault diagnosis knowledge graph;

[0012] S2. Fault diagnosis reasoning based on the insulation board production line fault diagnosis knowledge graph

[0013] The fault diagnosis reasoning includes:

[0014] S21, according to the complete fault knowledge graph obtained in step S14, input the fault phenomenon set F of the insulation board production line, and determine whether the fault phenomenon set F satisfies And Q = G ∪ R, we get the inference diagnosis set F d ;

[0015] S22, based on the inference diagnosis set F d Associated with the relation set R, the direct fault causes E are introduced by the relevant relations d and indirect fault causes E i , we get the fault cause set E0=E d ∪E i ;

[0016] S23, introducing a maintenance suggestion set A0 according to the fault cause set E0, and sorting the maintenance suggestions A0 in the order of maintenance suggestions introduced by direct causes first and maintenance suggestions introduced by indirect causes later;

[0017] S24. Output the fault causes and maintenance suggestions according to the sorted results.

[0018] As a further technical solution of the present invention: in step S11, the model construction of the insulation board production line fault diagnosis knowledge graph adopts a strategy construction model that combines bottom-up and top-down strategies.

[0019] As a further technical solution of the present invention: in step S12, the knowledge ontology of the insulation board production line fault diagnosis knowledge graph is represented according to the constructed model, specifically including:

[0020] Equipment structure, sensor point, fault phenomenon, fault cause and maintenance suggestion are taken as entity set G, and the relationship between these entities is taken as relationship set R;

[0021] Among them, the entity set is represented as:

[0022] G = {G1, G2, G3, G4, G5}

[0023] In the formula: G1—equipment structure entity; G2—sensor point entity; G3—fault phenomenon entity; G4—fault cause entity; G5—maintenance suggestion entity;

[0024] The relation set is represented as:

[0025] R={R1, R2, R3, R4, R5, R6}

[0026] In the formula: R1 is the relationship between the equipment structure and the cause of the fault; R2 is the relationship between the sensing point and the equipment structure; R3 is the relationship between the fault phenomenon and the cause of the fault; R4 is the relationship between the maintenance suggestion and the cause of the fault; R5 is the relationship between the sensing point and the fault phenomenon; R6 is the relationship between the equipment structure and the fault phenomenon.

[0027] As a further technical solution of the present invention: in step S13, the knowledge ontology is extracted from the insulation board production line fault diagnosis knowledge graph using the deep learning model BiLSTM+CRF entity extraction and relationship extraction; wherein BiLSTM represents a bidirectional long short-term memory network, and CRF represents a conditional random field.

[0028] As a further technical solution of the present invention: Step S14 specifically includes: in establishing a complete fault diagnosis knowledge graph, the entity set serves as the node in the fault diagnosis knowledge graph, and the relationship set between the entities serves as the edge in the fault diagnosis knowledge graph, wherein the edge weight represents the confidence of the entity relationship, and Neo4j is used to store the entities and relationships of the fault diagnosis knowledge graph.

[0029] As a further technical solution of the present invention: in step S21, the fault diagnosis reasoning process inputs the fault phenomenon set F that occurs to the complete fault knowledge graph obtained in step S14, and determines whether the fault phenomenon set F satisfies And Q = G ∪ R, we get the inference diagnosis set F d ;

[0030] For the case where the conditions are not met, remove the entity elements of the fault phenomenon set F that are not in Q to obtain F', and let F d =F'.

[0031] As a further technical solution of the present invention: in step S22, the inference diagnosis set F d Associated with the relation set R, the direct fault cause E is introduced by the relationship between the fault phenomenon and the fault cause d , the judgment condition is F d =Map(R5,R6);

[0032] If satisfied, the indirect fault cause E is introduced by the relationship between the sensing point and the fault phenomenon R5 and the relationship between the equipment structure and the fault phenomenon R6. i If not satisfied, then Finally, let the set of fault causes E0 = E d ∪E i .

[0033] As a further technical solution of the present invention: in step S24, the diagnosis result is output according to the sorted fault cause set E0 and the maintenance suggestion set A0, and the fault phenomenon entity is retrieved through the Neo4j database using cypher statements to obtain the fault cause and maintenance suggestion to realize fault diagnosis.

[0034] The beneficial effects of the present invention are:

[0035] 1) The present invention introduces indirect fault causes through fault diagnosis reasoning, data measurement points, the relationship between equipment structure and fault phenomena, which can solve the problem of low accuracy of traditional fault diagnosis methods under the condition of missing signs;

[0036] 2) The present invention can effectively extract knowledge from the insulation board production line equipment maintenance log, expert knowledge base, literature and equipment technical documents. When the user is diagnosing the insulation board production line fault, the user can analyze the cause of the production line equipment failure and give maintenance suggestions;

[0037] 3) The fault diagnosis knowledge graph of the present invention supports the query function. Users can input equipment fault phenomena and analyze and output the fault causes through the fault diagnosis reasoning process. It supports the introduction of new fault phenomena, fault causes, and maintenance suggestions. When a new fault occurs in the production line equipment, the technicians can input it, which is convenient for improving the constructed insulation board production line fault diagnosis knowledge graph. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1 It is a schematic diagram of the overall process of the present invention;

[0039] Figure 2 A schematic diagram of the knowledge graph entities and relationships of the present invention;

[0040] Figure 3 It is a schematic diagram of the knowledge graph fault diagnosis reasoning process of the present invention. DETAILED DESCRIPTION

[0041] In order to facilitate the problem to be solved, an embodiment of the present invention provides a plastic injection molding machine with a demoulding mechanism. The technical scheme in the embodiment of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiment of the present invention. Obviously, the described embodiment is only a part of the embodiment of the present invention, not all of the embodiments. Based on the embodiment of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0042] Examples, such as Figure 1 As shown, the present invention provides a thermal insulation board production line fault diagnosis method based on a knowledge graph, and the thermal insulation board production line fault diagnosis method comprises the following steps:

[0043] First: Construction of the insulation board production line fault diagnosis knowledge graph.

[0044] 1) The model construction of the insulation board production line fault diagnosis knowledge graph adopts a strategy construction mode that combines bottom-up and top-down;

[0045] 2) The knowledge ontology of the insulation board production line fault diagnosis knowledge graph is represented according to the constructed model, where the equipment structure, sensor points, fault phenomenon, fault cause and maintenance suggestion are taken as the entity set G, and the relationship between these entities is taken as the relationship set R;

[0046] like Figure 2 As shown, the entity set is represented as:

[0047] G = {G1, G2, G3, G4, G5}

[0048] In the formula: G1—equipment structure entity; G2—sensor point entity; G3—fault phenomenon entity; G4—fault cause entity; G5—maintenance suggestion entity;

[0049] The relation set is represented as:

[0050] R={R1, R2, R3, R4, R5, R6}

[0051] Where: R1—the relationship between the equipment structure and the cause of the fault; R2—the relationship between the sensing point and the equipment structure; R3—the relationship between the fault phenomenon and the cause of the fault; R4—the relationship between the maintenance suggestion and the cause of the fault; R5—the relationship between the sensing point and the fault phenomenon; R6—the relationship between the equipment structure and the fault phenomenon;

[0052] 3) Extract knowledge ontology from the insulation board production line fault diagnosis knowledge graph based on the represented knowledge ontology;

[0053] The knowledge ontology of the insulation board production line fault diagnosis knowledge graph is extracted using the deep learning model BiLSTM+CRF for entity extraction and relationship extraction; BiLSTM stands for bidirectional long short-term memory network, and CRF stands for conditional random field;

[0054] Among them, knowledge ontology extraction includes entity extraction, relationship extraction and attribute extraction. Knowledge ontology extraction is carried out from the equipment maintenance log of the insulation board production line, expert knowledge base, literature and equipment technical documents to construct triple structure data of entity-relationship-entity and entity-attribute-attribute value;

[0055] Entity extraction includes equipment structure entities, sensor point entities, fault phenomenon entities, fault cause entities, and maintenance suggestion entities; among them, equipment structure includes insulation board mixing equipment, compression molding equipment, cutting equipment, packaging machine, mixing kettle, crawler laminator and its subsystems, components, and parts; sensor points include temperature measurement points, pressure measurement points, flow measurement points, vibration measurement points, and voltage and current measurement points;

[0056] Relationship extraction includes the relationship between equipment structure and fault cause, the relationship between sensor sensing points and equipment structure, and the relationship between fault phenomenon and fault cause; the relationship between maintenance suggestions and fault cause, the relationship between sensor sensing points and fault phenomenon, and the relationship between equipment structure and fault phenomenon.

[0057] 4) Construct nodes and edges in the fault diagnosis knowledge graph of the insulation board production line based on the extracted knowledge ontology, and establish a complete fault diagnosis knowledge graph;

[0058] In terms of features, the entity set serves as the node in the fault diagnosis knowledge graph, the relationship set between entities serves as the edge in the fault diagnosis knowledge graph, the edge weight represents the confidence of the entity relationship, and Neo4j is used to store the entities and relationships of the fault diagnosis knowledge graph to establish a complete fault diagnosis knowledge graph.

[0059] Second: Fault diagnosis reasoning based on the knowledge graph of insulation board production line fault diagnosis, such as Figure 3 shown.

[0060] 1) Based on the complete fault diagnosis knowledge graph obtained in the "first" step, the fault phenomenon set F that occurs during the fault diagnosis reasoning process is input into it to determine whether the fault phenomenon set F satisfies And Q = G ∪ R, we get the inference diagnosis set F d ;

[0061] If the conditions are met, the inference diagnosis set F d =F;

[0062] If the condition is not met, remove the entity elements of the fault phenomenon set F that are not in Q to obtain F', and let F d =F';

[0063] 2) Based on the inference diagnosis set F d Associated with the relation set R, the direct fault cause E is introduced by the relation R3 between the fault phenomenon and the fault cause d , the judgment condition is F d =Map(R5,R6)

[0064] If satisfied, the indirect fault cause E is introduced by the relationship between the sensing point and the fault phenomenon R5 and the relationship between the equipment structure and the fault phenomenon R6. i ;

[0065] If not satisfied, Finally, let the set of fault causes E0 = E d ∪E i .

[0066] 3) Introduce the maintenance suggestion set A0 according to the fault cause set E0, and sort the maintenance suggestions A0 in the order of maintenance suggestions introduced by direct causes first and maintenance suggestions introduced by indirect causes later;

[0067] 4) Output the fault causes and maintenance suggestions according to the sorted ones;

[0068] The diagnostic results are output according to the sorted fault cause set E0 and maintenance suggestion set A0. The fault phenomenon entity is retrieved through the Neo4j database using cypher statements to obtain the fault cause and maintenance suggestion to realize fault diagnosis.

[0069] The insulation board production line fault diagnosis method constructs an insulation board production line fault diagnosis knowledge graph by establishing entities and their relationships. Through the reasoning diagnosis process, the equipment structure, sensor points, fault phenomenon, fault cause, and maintenance suggestions of the production line are linked to obtain direct and indirect fault causes. The influence of relevant factors on the fault phenomenon can be accurately judged, the fault cause is revealed, and maintenance suggestions are given, thereby reducing the impact of the fault on the normal production of the insulation board production line.

[0070] It will be apparent to those skilled in the art that the invention is not limited to the details of the exemplary embodiments described above and that the invention can be implemented in other specific forms without departing from the spirit or essential features of the invention. Therefore, the embodiments should be considered exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description, and it is intended that all variations falling within the meaning and scope of the equivalent elements of the claims be included in the invention. Any reference numeral in a claim should not be considered as limiting the claim to which it relates.

[0071] In addition, it should be understood that although the present specification is described according to implementation modes, not every implementation mode contains only one independent technical solution. This description of the specification is only for the sake of clarity. Those skilled in the art should regard the specification as a whole. The technical solutions in each embodiment may also be appropriately combined to form other implementation modes that can be understood by those skilled in the art.

Claims

1. A fault diagnosis method for an insulation board production line based on a knowledge graph, characterized in that: The insulation board production line fault diagnosis method comprises the following steps: S1. Construction of knowledge graph for fault diagnosis of insulation board production line The construction of the fault diagnosis knowledge graph includes: S11. Model construction of fault diagnosis knowledge graph for insulation board production line; S12, representing the knowledge ontology of the insulation board production line fault diagnosis knowledge graph according to the constructed model; S13, extracting knowledge ontology from the insulation board production line fault diagnosis knowledge graph according to the represented knowledge ontology; S14, constructing nodes and edges in the insulation board production line fault diagnosis knowledge graph based on the extracted knowledge ontology, and establishing a complete fault diagnosis knowledge graph; S2. Fault diagnosis reasoning based on the insulation board production line fault diagnosis knowledge graph The fault diagnosis reasoning includes: S21, according to the complete fault knowledge graph obtained in step S14, input the fault phenomenon set F of the insulation board production line, and determine whether the fault phenomenon set F satisfies And Q = G ∪ R, we get the inference diagnosis set F d ; S22. Based on the inference diagnosis set F d Associated with the relation set R, the direct fault causes E are introduced by the relevant relations d and indirect fault causes E i , we get the fault cause set E0=E d ∪E i ; S23, introducing a maintenance suggestion set A0 according to the fault cause set E0, and sorting the maintenance suggestions A0 in the order of maintenance suggestions introduced by direct causes first and maintenance suggestions introduced by indirect causes later; S24. Output the fault causes and maintenance suggestions according to the sorted results.

2. The insulation board production line fault diagnosis method according to claim 1 is characterized in that: In step S11, the model construction of the insulation board production line fault diagnosis knowledge graph adopts a strategy construction model that combines bottom-up and top-down strategies.

3. The insulation board production line fault diagnosis method according to claim 1, characterized in that: In step S12, the ontology of the insulation board production line fault diagnosis knowledge graph is represented according to the constructed mode, specifically including: Equipment structure, sensor point, fault phenomenon, fault cause and maintenance suggestion are taken as entity set G, and the relationship between these entities is taken as relationship set R; Among them, the entity set is represented as: G = {G1, G2, G3, G4, G5} In the formula: G1—equipment structure entity; G2—sensor point entity; G3—fault phenomenon entity; G4—fault cause entity; G5—maintenance suggestion entity; The relation set is represented as: R={R1, R2, R3, R4, R5, R6} In the formula: R1 is the relationship between the equipment structure and the cause of the fault; R2 is the relationship between the sensing point and the equipment structure; R3 is the relationship between the fault phenomenon and the cause of the fault; R4 is the relationship between the maintenance suggestion and the cause of the fault; R5 is the relationship between the sensing point and the fault phenomenon; R6 is the relationship between the equipment structure and the fault phenomenon.

4. The insulation board production line fault diagnosis method according to claim 3, characterized in that: In step S13, the knowledge extraction of the insulation board production line fault diagnosis knowledge graph adopts the deep learning model BiLSTM+CRF entity extraction and relationship extraction; wherein, BiLSTM represents the bidirectional long short-term memory network, and CRF represents the conditional random field.

5. The insulation board production line fault diagnosis method according to claim 4, characterized in that: Step S14 specifically includes: in establishing a complete fault diagnosis knowledge graph, the entity set serves as the node in the fault diagnosis knowledge graph, and the relationship set between entities serves as the edge in the fault diagnosis knowledge graph, wherein the edge weight represents the confidence of the entity relationship, and Neo4j is used to store the entities and relationships of the fault diagnosis knowledge graph.

6. The insulation board production line fault diagnosis method according to claim 4, characterized in that: In step S21, the fault diagnosis reasoning process inputs the fault phenomenon set F to the complete fault knowledge graph obtained in step S14 to determine whether the fault phenomenon set F satisfies And Q = G ∪ R, we get the inference diagnosis set F d ; For the case where the conditions are not met, remove the entity elements of the fault phenomenon set F that are not in Q to obtain F', and let F d =F'.

7. The insulation board production line fault diagnosis method according to claim 6, characterized in that: In step S22, the inference diagnosis set F d Associated with the relation set R, the direct fault cause E is introduced by the relation R3 between the fault phenomenon and the fault cause d , the judgment condition is F d =Map(R5,R6); If satisfied, the indirect fault cause E is introduced by the relationship between the sensing point and the fault phenomenon R5 and the relationship between the equipment structure and the fault phenomenon R6. i If not satisfied, then Finally, let the set of fault causes E0 = E d ∪E i .

8. The insulation board production line fault diagnosis method according to claim 7, characterized in that: In step S24, the diagnosis result is output according to the sorted fault cause set E0 and the maintenance suggestion set A0, and the fault phenomenon entity is searched through the Neo4j database using cypher statements to obtain the fault cause and maintenance suggestion to achieve fault diagnosis.

Citation Information

Patent Citations

  • Industrial equipment fault diagnosis method based on knowledge graph

    CN113723632A

  • Range hood equipment fault diagnosis method and system based on knowledge graph

    CN114971383A

  • Machine tool fault diagnosis method based on knowledge graph

    CN115374314A

  • Equipment fault diagnosis method and device based on knowledge graph driving

    CN117350377A

Cited By

  • Complex production line reliability analysis method based on combination of fault operation and maintenance knowledge graph and fault tree

    CN119903430A

  • Fault operation and maintenance knowledge graph combined with fault tree complex production line reliability analysis method

    CN119903430B

  • Process control station self-diagnosis system supporting user definition

    CN121325815A