Industrial equipment fault knowledge base construction method and device, and fault diagnosis method and system

Abstract

The application discloses a kind of industrial equipment failure knowledge base construction method, device and fault diagnosis method and system, the method steps include: obtaining the association between each component in the equipment to be measured and failure mode;Abstract the graph model of each failure mode to build and form failure mode node, abnormal fault event is regarded as node and is built in sequence to form fault tree, and each component in the equipment to be measured is regarded as node and builds basic graph model, each node in basic graph model, fault tree is connected with the corresponding failure mode node respectively, and is built to form fault knowledge system;Using graph database, the fault knowledge system is stored according to preset logic model, and the failure knowledge base of the equipment to be measured is obtained.The application realizes method simple, low in cost, data integrity is high, can fully reflect the mutual association between equipment configuration, internal structure relationship and failure mode, improves the efficiency and precision of fault diagnosis and the like advantages.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of industrial equipment fault diagnosis, and in particular to an industrial equipment fault knowledge base construction method and device, and a fault diagnosis method and system. BACKGROUND

[0002] For fault diagnosis of large-scale industrial equipment with complex structure, the current common methods are manual detection and troubleshooting or fault judgment based on simple operation data. For example, the maintenance of a railway locomotive mainly involves state and fault diagnosis based on the state and alarm information detected during online operation of the locomotive. Generally, the locomotive is temporarily connected after arriving at a section, the fault data of the locomotive on the same day and the same trip are downloaded, and the fault data of the locomotive are analyzed on site to locate the fault point and determine the solution by combining manual experience and manual detection. However, the above fault diagnosis method depends on the knowledge level and experience of the analysis personnel, and has low diagnosis efficiency and low intelligentization degree, and is prone to missed detection and false detection, and has problems such as untimely fault handling, inability to accurately find the fault location, and incomplete fault handling, making it difficult to timely and accurately locate the fault.

[0003] The intelligent fault diagnosis method can solve the problems of low fault positioning efficiency and low intelligentization degree in the traditional manual troubleshooting method. In the prior art, the intelligent fault diagnosis method generally involves collecting data of equipment in different fault states, constructing a fault diagnosis model based on the data in different fault states, and inputting the real-time fault data into the fault diagnosis model to realize intelligent fault diagnosis. However, the above intelligent fault diagnosis method requires a large amount of data in different fault states for model training, and for large-scale industrial equipment with complex structure, a large amount of model training data is required and the model training time is long. Meanwhile, the fault diagnosis model can only represent the relationship between the fault state and the fault data, and cannot fully utilize the information contained in the equipment configuration, internal structure relationship, and failure mode, resulting in low fault diagnosis accuracy.

[0004] If a fault knowledge base can be constructed using the information contained in the equipment configuration, internal structure relationship, and failure mode, the intelligent fault diagnosis of equipment failure can be realized using the fault knowledge base. In the prior art, the fault knowledge base is usually composed of only the operation data when the equipment fails, and cannot reflect the mutual relationship between the equipment configuration, internal structure relationship, and failure mode. Therefore, using such a fault knowledge base can only realize simple pattern matching of faults, and it is difficult to accurately diagnose the source of the fault, locate the fault position, and locate the components that may be affected after the fault occurs. SUMMARY

[0005] The technical problem to be solved by this invention is: in view of the technical problems existing in the prior art, this invention provides a method and device for constructing an industrial equipment fault knowledge base that is simple to implement, low in cost, and has high data integrity, and can fully reflect the interrelationship between equipment configuration, internal structural relationships and failure modes; as well as a fault diagnosis method and system that has high diagnostic accuracy and efficiency and can accurately diagnose the source of faults and locate the components that may be affected by the faults.

[0006] To solve the above-mentioned technical problems, the technical solution proposed by this invention is as follows:

[0007] A method for constructing an industrial equipment fault knowledge base, comprising the following steps:

[0008] Obtain the correlation between each component and the failure mode;

[0009] The graph model of each failure mode is abstracted to form failure mode nodes. Abnormal failure events are used as nodes and constructed into a fault tree in a hierarchical order from top to bottom. The components in the device under test are used as nodes and a basic graph model is constructed according to the structural relationship between the components. The nodes in the basic graph model and the fault tree are connected to the corresponding failure mode nodes according to the association relationship to construct the fault knowledge system of the device under test.

[0010] The fault knowledge system is stored using a graph database according to a preset logical model to obtain the fault knowledge base of the device under test.

[0011] Furthermore, there is a causal logical relationship between parent nodes and child nodes in the fault tree, where a fault event in the next layer is the cause of a fault event in the previous layer.

[0012] Furthermore, the logical model includes abnormal fault events, failure modes, component definition node labels, attributes, and the relationships between nodes in the fault knowledge system. The relationships between nodes include the relationships within the same type of node and the relationships between different types of nodes.

[0013] Furthermore, the node label includes the type of abnormal failure event, failure mode, component, and subsystem, and the attributes include abnormal failure event ID, failure mode ID, component ID, and name.

[0014] Furthermore, the relationships between the nodes include fault causes, belonging relationships, energy transfer relationships between components, fault and failure mode correlation relationships, and failure mode and component correlation relationships. The fault causes correspond to the relationship from parent node to child node in the fault tree, the belonging relationships correspond to the relationship from child component to total component, and the component type nodes include the energy transfer relationships between components and the subordinate relationship from child component to total component.

[0015] Further, the graph database is a Neo4j graph database.

[0016] An industrial equipment fault diagnosis method, comprising the steps of:

[0017] A fault knowledge base of the measured equipment is constructed according to the above-mentioned fault knowledge base construction method.

[0018] When an abnormal fault event is detected, the current detected abnormal fault event is input into the fault knowledge base, leaf nodes under the node corresponding to the current abnormal fault event in the fault tree are found out, all possible failure modes are determined according to all the found abnormal fault events and the corresponding relationship between the abnormal fault events and the failure modes, and all the components associated with the current abnormal fault event are judged according to all the determined failure modes and the relationship between the failure modes and the components.

[0019] An industrial equipment fault knowledge base construction device, comprising:

[0020] An acquisition module is configured to acquire fault domain knowledge of a measured equipment, wherein the fault domain knowledge comprises an association relationship between components and failure modes.

[0021] A knowledge system construction module is configured to abstract a graph model of each failure mode to form a failure mode node, to construct a fault tree by taking abnormal fault events as nodes and in a top-down order, and to construct a basic graph model by taking each component in the measured equipment as a node and according to a structural relationship between the components, and to connect each node in the basic graph model and the fault tree to a corresponding failure mode node according to the association relationship, so as to construct a fault knowledge system of the measured equipment.

[0022] A graph database storage module is configured to store the fault knowledge system according to a preset logic model by using a graph database, so as to obtain a fault knowledge base of the measured equipment.

[0023] An industrial equipment fault knowledge base construction device, comprising a processor and a memory, wherein the memory is configured to store a computer program, and the processor is configured to execute the computer program to perform the above-mentioned method.

[0024] An industrial equipment fault diagnosis system, comprising:

[0025] The above-mentioned fault knowledge base construction device;

[0026] The diagnostic device is used for inputting a current detected abnormal fault event into a fault knowledge base of the fault knowledge base construction device when an abnormal fault event is detected, finding out leaf nodes under a node corresponding to the current abnormal fault event from the fault tree, determining all possible failure modes according to all the abnormal fault events and the corresponding relationship between the abnormal fault events and the failure modes, and judging all components associated with the current abnormal fault event according to all the failure modes and the relationship between the failure modes and the components.

[0027] Compared with the prior art, the present application has the following advantages:

[0028] 1. The present application constructs a fault tree based on abnormal fault events as nodes, and the nodes of the fault tree are connected with failure mode nodes respectively to establish a connection between basic events and failure modes, and meanwhile, basic graph models are constructed by taking each component in the measured equipment as a node, and each node in the basic graph model is connected with a failure mode node respectively to form a fault knowledge system having an associated relationship network among abnormal fault events, fault trees, failure modes and components, so that the fault domain knowledge and the equipment mechanism structure can be compatible in the same knowledge system, thereby realizing the organic combination and unified representation of the fault domain knowledge and the equipment representation method, and meanwhile, the complex corresponding relationship between components and abnormal events can be realized, and the abnormal fault events caused by multiple factors or multiple components can be fully represented and covered, and the complex coupling cause-and-effect relationship of the fault problems across the system level can be realized.

[0029] 2. The present application uses a graph database as a physical model for realizing a knowledge base, and can completely represent the association among abnormal fault events, failure modes and components, and meanwhile, the associated failure modes and components can be quickly obtained based on abnormal fault events, and meanwhile, the graph retrieval performance and the expansion advantage of the graph database can be exerted to construct a fault domain knowledge base suitable for complex industrial equipment.

[0030] 3. The present application further determines a corresponding abnormal fault event when an abnormal state of the measured equipment triggers an alarm, inputs the abnormal fault event into the constructed fault knowledge base, and can quickly and accurately analyze the initial source of the fault and the associated components possibly affected by the abnormal fault event based on the association relationship among the abnormal fault events, the fault trees and the components. BRIEF DESCRIPTION OF DRAWINGS

[0031] Figure 1 is a schematic diagram of a principle flow of the present embodiment for realizing fault diagnosis based on an industrial equipment fault knowledge base.

[0032] Figure 2 is a schematic diagram of an implementation flow of the present embodiment for constructing an industrial equipment fault knowledge base.

[0033] Figure 3 is a schematic diagram of the principle of constructing a fault tree based on an abnormal failure event in the embodiment of the present application.

[0034] Figure 4 is a schematic diagram of the effect of constructing a model when the present application is applied to a lubrication system in a specific application embodiment.

[0035] Figure 5 is a schematic diagram of the nodes and relationships in a graph database in which the power subsystem is stored in a specific application embodiment of the present application.

[0036] Figure 6 is a schematic diagram of the implementation process of industrial equipment failure diagnosis in the embodiment of the present application. DETAILED DESCRIPTION

[0037] The present application is further described below in conjunction with the accompanying drawings and specific preferred embodiments, but the scope of protection of the present application is not limited thereby.

[0038] As shown in Figure 1 , the process of abnormal failure events and failure diagnosis analysis can form a complete loop. For example, in a large industrial equipment (such as a locomotive), if a module (referred to as a component) representing a subsystem or component in the equipment fails, it means that a failure mode related to the component affects the corresponding function. By calculating the indicators in the component health state evaluation system based on the monitored operation data, abnormal values can be found, and an abnormal event alarm is triggered, thereby starting the failure analysis and diagnosis link. Based on the constructed failure knowledge base, failure analysis, positioning, and prediction are performed.

[0039] As shown in Figure 2 , the steps of the industrial equipment failure knowledge base construction method in the embodiment of the present application include:

[0040] S01. Obtain the failure domain knowledge of the equipment under test, and the failure domain knowledge includes the association between each component and failure mode;

[0041] S02. Abstract the graph model construction of each failure mode to form a failure mode node, construct a fault tree from top to bottom according to the order of the abnormal failure event as a node, and construct a basic graph model according to the structure relationship between each component in the equipment under test, and connect each node in the basic graph model and the fault tree to the corresponding failure mode node according to the association relationship, to construct a failure knowledge system of the equipment under test;

[0042] S03. Store the failure knowledge system in a graph database according to a preset logical model to obtain a failure knowledge base of the equipment under test.

[0043] The embodiment takes the structured and hierarchical device configuration and internal structure as the core basis of fault knowledge data storage management. The fault domain knowledge is abstracted into the topology structure on the device system structure diagram and the independent graph structure associated with the system structure diagram. The fault tree is constructed based on the uniformly defined abnormal fault event. Each node in the fault tree corresponds to an abnormal fault event. The nodes of the fault tree are connected to the failure mode nodes, respectively, to establish the association between the basic events and the failure modes. Meanwhile, each component in the tested device is taken as a node to construct a basic graph model. Each node in the basic graph model is connected to the failure mode node, respectively, to establish the association between the components and the failure modes. The fault knowledge system with the association relationship network among the abnormal fault event, the fault tree, the failure mode and the components is formed. The fault domain knowledge and the device mechanism structure are compatible in the same knowledge system, so as to realize the organic combination and unified representation of the fault domain knowledge and the device representation method. Meanwhile, the complex corresponding relationship between the components and the abnormal events can be realized, the abnormal fault events caused by multiple factors or multiple components can be fully represented and covered, and the complex coupling cause-and-effect relationship of the fault problems across the system level can be realized.

[0044] The embodiment further combines the storage mode of the graph database, uses the graph database as the physical model of the knowledge base, can completely represent the association among the abnormal fault event, the failure mode and the components, can quickly obtain the associated failure mode and components based on the abnormal fault event, can exert the graph retrieval performance and expansibility advantage of the graph database, and can construct the fault domain knowledge base suitable for the complex industrial device.

[0045] The failure mode can be obtained by analyzing the component-failure mode analysis table and the component potential failure mode analysis table. Based on the analysis table, the association relationship between the failure mode and the potential failure mode and the components can be determined. The component-failure mode relationship table is the relatively clear and formal fault domain knowledge in the system. After the association between the failure mode and the components is determined by using the relationship table, the fault knowledge base is constructed in the above manner. When the abnormal fault event occurs, the specific components of the device can be connected through the association relationship of the abnormal fault event-fault tree-failure mode-component relationship, so as to realize the intelligent fault diagnosis and analysis. The potential failure mode analysis table can be the items that are relatively important or have relatively large risks.

[0046] Since one component can correspond to multiple failure modes, if the failure modes are taken as attributes on the nodes of the basic graph model, the convenience of query will be affected if one node attribute contains multiple failure modes. Therefore, in this embodiment, the failure modes are abstracted as independent nodes after the graph model is built, and then connected to the basic graph model through the correspondence between components and failure modes. In the specific application embodiment, the data relationship model of function-component-failure mode is generated through the component-failure mode relationship table and the potential failure mode table, and then the graph model of the failure mode is abstracted and connected to each component node and fault tree node in the basic graph model as an independent node.

[0047] In this embodiment, the graph model is used to represent the conceptual model of the device structure under test and the fault tree and failure mode. Since there is no direct correspondence between the nodes of the fault tree and the components, the fault tree is taken as a graph structure independent of the basic graph model. For the construction of the fault tree, this embodiment first establishes a health state evaluation index system of the device under test, and then defines abnormal fault events according to abnormal states, taking the abnormal fault events as the basic component events (i.e., nodes) of the fault tree, and integrating the fault tree after analyzing the system hierarchy from top to bottom, as shown in Figure 3 The fault tree has a causal logical relationship between the parent nodes and the child nodes, in which the next layer of fault events is the inducement of the previous layer of fault events. Since the abnormal events are defined to maintain consistency with the failure modes, the nodes of the fault tree can establish corresponding edges with the failure mode nodes.

[0048] Taking the lubrication system in the locomotive device as an example in the specific application embodiment, as shown in Figure 4 The left side is a component structure diagram of the lubrication system, and the basic graph structure is constructed according to the component structure diagram. The upper part on the right side is the constructed fault tree, and the lower part is the constructed failure mode node. Each failure mode node is connected to each node in the fault tree, and each failure mode node is also connected to each component in the lubrication system, i.e., there is a one-to-many relationship between the basic graph model representing the component structure relationship and the failure mode nodes, and there is a many-to-many relationship between the failure modes and the nodes of the fault tree. It should be noted that, Figure 4 The scattered distribution of the failure mode nodes in

[0049] The nodes in the logic model of the embodiment are respectively defined with labels and attributes of abnormal failure events, failure modes, components, and the relationships between the nodes in the failure knowledge system, including the relationships within the same type of nodes and the relationships between different types of nodes. The node labels specifically include the types of abnormal failure events, failure modes, components, and subsystems, and the attributes specifically include the IDs and names of abnormal failure events, failure modes, components, and subsystems. The relationships between the nodes include the relationships of failure causes, belonging, energy transmission between components, failure modes and components, and the relationships of failure modes and components. The relationship of failure causes corresponds to the relationship from a parent node to a child node in a fault tree, the relationship of belonging corresponds to the relationship from a child component to a total component, and the node of the component type includes the energy transmission relationship between components and the relationship of subordination from a child component to a total component. In a specific application example, the logic model is applied to a lubricating system in a locomotive device, and the configuration of the logic model is shown in Table 1.

[0050] The logic model of the failure knowledge base needs to define the types and attributes of the nodes and edges of the graph model. In the embodiment, the Neo4j graph database is used to store the failure knowledge system. In the Neo4j graph database, the types of nodes are defined using labels, and a node can be assigned multiple labels. Both nodes and relationships have attributes and can be queried by attribute values. Specifically, a label is defined for each of the failure events, failure modes, and system components to distinguish the three types of nodes. A label representing a subsystem (such as the lubricating system described above) is also defined, and the component nodes belonging to the subsystem have both the labels representing the component and the lubricating system. In the embodiment, the subsystem is used as a label rather than an attribute value of the node. On the one hand, this can avoid redundancy caused by storing subsystem information through node attributes. On the other hand, through multiple labels, the hierarchical relationship of the system structure can be reflected, and it is also convenient to quickly find the required nodes in different subsystems, such as finding a component under the condition of having determined the belonging subsystem. Through the subsystem label, the search can be performed within the node set having the label, and if the subsystem is used as an attribute value, the search needs to be performed in all nodes with the attribute value of the subsystem and meeting the search conditions.

[0051] Table 1: Logic model of the locomotive lubricating system

[0052]

[0053]

[0054] The embodiment specifically defines the corresponding Id of each type of node in the attribute, which can set a unique identifier for the node, forming a primary key similar to that in a relational database. Although the Neo4j graph database provides an automatically generated Id for each node, Neo4j will reuse the automatic Id of the deleted node. The embodiment defines the corresponding Id of each type of node, which can form a unique identifier for the node and avoid the reuse of the Id of the deleted node. The name in the attribute is a public attribute of all nodes and relationships. In addition to the above attribute information, other attribute information can also be configured according to actual needs. If there is information about fault measures, the attribute can be added to the node, so that the node can obtain the suggestions for dealing with the fault.

[0055] The relationship of the node in the logical model in the embodiment includes the relationship within the same type of node and the relationship between different types of nodes. The node of the component type has both the energy transmission relationship between components and the subordinate relationship of the sub-component to the total component. If the size of the energy transmitted between components needs to be recorded, an attribute representing the order of magnitude can be added to the flow relationship. If there is more detailed information in the fault tree, such as the degree of causal correlation between different fault events, an attribute representing the impact or probability can be added to the reason relationship. If there is probability information between the component and the failure mode, an attribute representing the possibility can also be added to the relationship between the two, so as to provide qualitative or quantitative analysis results when searching for faults.

[0056] After determining the logical model according to the above steps, the data abstracted in the conceptual model is further stored in the graph database in the manner defined by the logical model. As shown in Figure 5 The nodes and relationships of the power subsystem stored in the graph database are shown in the left. The fault event nodes on the left form a fault tree. The middle is the failure mode node. The right is the component node. The nodes of the fault, failure mode, and component form a mesh structure. With the help of the characteristics of the graph database engine, any reachable node or path can be easily found through the relationship between the nodes.

[0057] The embodiment forms a fault knowledge base by storing the fault domain knowledge in the graph database. The non-key attribute information of the fault domain knowledge, the equipment history model, and the monitoring data are all stored in the relational database, and are connected through the node identifier (such as the node Id and the fault signature). Based on the fault knowledge base, it can be applied to the fault diagnosis and analysis platform to combine the monitoring data of the equipment for fault diagnosis and analysis, and further to realize fault retrieval, analysis, prediction, display, and other knowledge base-based applications.

[0058] As shown in Figure 6 The steps of the industrial equipment fault diagnosis method of the embodiment include:

[0059] The fault knowledge base of the device under test is constructed according to the above fault knowledge base construction method;

[0060] When an abnormal fault event is detected, the current detected abnormal fault event is input into the fault knowledge base, leaf nodes under the node corresponding to the current abnormal fault event in the fault tree are found out, all possible failure modes are determined according to all the abnormal fault events found out and the corresponding relationship between the abnormal fault events and the failure modes, and all the components associated with the current abnormal fault event are judged according to all the failure modes determined and the relationship between the failure modes and the components.

[0061] After the device under test triggers an alarm in an abnormal state, the corresponding abnormal fault event is determined, the abnormal fault event is input into the above-constructed fault knowledge base, and the initial source of the fault and the associated components possibly affected can be quickly and accurately analyzed by using the fault tree constructed by the abnormal fault event, the failure modes and the association relationship between the components.

[0062] Since the fault tree in the fault knowledge base is composed of abnormal events in a top-down manner, the leaf nodes of the fault tree are the most basic fault events, and there is a causal logical relationship between the parent nodes and the child nodes of the fault tree, the bottom fault events are the inducement of the upper fault events, and the fault source finding problem is changed into a problem of finding the leaf nodes under a given node (input abnormal event / fault identifier) of the fault tree, that is, a graph traversal finding problem. After the abnormal fault event of the system is input into the fault knowledge base in this embodiment, the leaf nodes under the abnormal event node in the fault tree are found out according to the structure of the fault knowledge base, the nodes of the fault tree are associated with the failure modes through the representative basic events, the fault root of the abnormal event is found out through the fault tree, and then the components possibly affected are inferred in combination with the corresponding relationship between the failure modes and the components.

[0063] The industrial equipment fault knowledge base construction device of this embodiment comprises:

[0064] The acquisition module is configured to acquire the fault domain knowledge of the device under test, and the fault domain knowledge comprises the association relationship between the components and the failure modes;

[0065] The knowledge system construction module is configured to abstract a graph model of each failure mode to construct a failure mode node, take an abnormal fault event as a node and construct a fault tree in a top-down order according to the hierarchy, take each component in the device under test as a node and construct a basic graph model according to the structural relationship between the components, connect each node in the basic graph model and the fault tree with the corresponding failure mode node according to the association relationship, and construct a fault knowledge system of the device under test;

[0066] The graph database storage module is configured to store the fault knowledge system according to a preset logic model using a graph database, and obtain a fault knowledge base of the device under test.

[0067] The industrial equipment fault knowledge base construction device of the embodiment corresponds to the above-described industrial equipment fault knowledge base construction method, and thus will not be described again.

[0068] In another embodiment, the industrial equipment fault knowledge base construction of the present application can also include a processor and a memory, the memory being configured to store a computer program, and the processor being configured to execute the computer program to perform the above-described method.

[0069] The industrial equipment fault diagnosis system of the embodiment includes:

[0070] The fault knowledge base construction device described above;

[0071] The diagnosis device is configured to input the current detected abnormal fault event into the fault knowledge base of the fault knowledge base construction device, find out the leaf nodes under the node corresponding to the current abnormal fault event from the fault tree, determine all possible failure modes according to all the abnormal fault events found out and the corresponding relationship between the abnormal fault events and the failure modes, and determine all the components associated with the current abnormal fault event according to all the failure modes determined and the relationship between the failure modes and the components.

[0072] The industrial equipment fault diagnosis system of the embodiment corresponds to the above-described industrial equipment fault diagnosis method, and both have the same principles and effects.

[0073] The above is only a preferred embodiment of the present application, and does not limit the present application in any form. Although the present application has been disclosed as above with the preferred embodiment, it is not intended to limit the present application. Therefore, any simple modification, equivalent change and modification made to the above embodiment without departing from the technical solution of the present application, and according to the technical essence of the present application, should fall within the scope of protection of the technical solution of the present application.

Claims

1. An industrial equipment failure knowledge base construction method characterized by the steps of The method comprises the following steps: Obtain the fault domain knowledge of the device under test, and the fault domain knowledge comprises the association between components and failure modes; Abstract the fault domain knowledge into a topology structure on a device system structure diagram, and abstract the fault domain knowledge into an independent graph structure connected with the system structure diagram, abstract the graph model of each failure mode to form a failure mode node, construct an abnormal fault event as a node and form a fault tree in a top-down order according to the hierarchy, and construct a basic graph model by taking each component in the device under test as a node and according to the structural relationship between the components, and connect each node in the basic graph model and the fault tree with the corresponding failure mode node according to the association, each node in the fault tree corresponds to an abnormal fault event, and the nodes of the fault tree are connected with the failure mode nodes respectively to establish the association between the basic events and the failure modes, and construct a basic graph model by taking each component in the device under test as a node, and the nodes of the basic graph model are connected with the failure mode nodes respectively to establish the association between the components and the failure modes, and construct a fault knowledge system of the device under test with the association between abnormal fault events, fault trees, failure modes and components, the fault trees have a causal logical relationship between parent nodes and child nodes, wherein the next layer of fault events is the inducement of the previous layer of fault events, the abnormal events and the failure modes are defined to be consistent, and the nodes of the fault tree and the failure mode nodes establish corresponding edges; Store the fault knowledge system in a graph database according to a preset logical model to obtain a fault knowledge base of the device under test, and the fault knowledge base is used to connect to specific components of the device through the association between abnormal fault events, fault trees, failure modes and components when an abnormal fault event is detected, which comprises the following steps: finding leaf nodes under the current abnormal fault event node according to the fault tree, determining all possible failure modes according to all abnormal fault events and the corresponding relationship between the abnormal fault events and the failure modes, and judging all components associated with the current abnormal fault event according to all determined failure modes and the relationship between the failure modes and the components. The logical model defines node labels, attributes of abnormal fault events, failure modes and components, and the relationship between the nodes in the fault knowledge system, the relationship between the nodes comprises the relationship within the same type of nodes and the relationship between different types of nodes, the nodes of the component type have both the energy transmission relationship between components and the subordinate relationship from sub-components to total components, so that the nodes of the fault, the failure mode and the component form a mesh structure, and any reachable node or path can be found through the relationship between the nodes.

2. The industrial equipment failure knowledge base construction method according to claim 1, characterized by, The node labels comprise the types of abnormal fault events, failure modes, components and subsystems, and the attributes comprise abnormal fault event IDs, failure mode IDs, component IDs and names.

3. The industrial equipment failure knowledge base construction method according to claim 1, characterized by, The relationship between the nodes includes a fault cause, a belonging relationship, an energy transmission relationship between components, a fault and failure mode correlation relationship, and a correlation relationship between failure modes and components, the fault cause corresponds to a relationship from a parent node to a child node in a fault tree, the belonging relationship corresponds to a relationship from a child component to a total component, and the node of the component type includes an energy transmission relationship between components and a subordinate relationship from a child component to a total component.

4. The industrial equipment failure knowledge base construction method according to any one of claims 1 to 3, characterized by, The graph database is a Neo4j graph database.

5. An industrial equipment failure diagnosis method characterized by the steps of The method comprises the following steps: The fault knowledge base of the device under test is constructed by the method for constructing a fault knowledge base according to any one of claims 1-4. When an abnormal fault event is detected, the current detected abnormal fault event is input into the fault knowledge base, leaf nodes under the node of the current abnormal fault event in the fault tree are found out, all possible failure modes are determined according to all the found abnormal fault events and the corresponding relationship between the abnormal fault events and the failure modes, and all components associated with the current abnormal fault event are determined according to all the determined failure modes and the relationship between the failure modes and the components.

6. An industrial equipment failure knowledge base construction apparatus for use in the method of any one of claims 1 to 4, characterized by, The method comprises the following steps: The acquisition module is configured to acquire fault domain knowledge of the device under test, wherein the fault domain knowledge comprises a correlation relationship between components and failure modes. The knowledge system construction module is configured to abstract a graph model of each failure mode to construct a failure mode node, take an abnormal fault event as a node and construct a fault tree in a top-down order according to a hierarchy, take each component in the device under test as a node and construct a basic graph model according to a structural relationship between the components, and connect each node in the basic graph model and the fault tree to a corresponding failure mode node according to the correlation relationship to construct a fault knowledge system of the device under test. The graph database storage module is configured to store the fault knowledge system in a graph database according to a preset logical model to obtain a fault knowledge base of the device under test, wherein the fault knowledge base is used to find leaf nodes under a node of a current abnormal fault event according to a fault tree when an abnormal fault event is detected, determine all possible failure modes according to all the found abnormal fault events and the corresponding relationship between the abnormal fault events and the failure modes, and determine all components associated with the current abnormal fault event according to all the determined failure modes and the relationship between the failure modes and the components.

7. An industrial equipment failure knowledge base construction apparatus comprising a processor and a memory for storing a computer program, characterized by, The processor is configured to execute the computer program to execute the method according to any one of claims 1-5.

8. An industrial equipment failure diagnosis system characterized by comprising: The method comprises the following steps: The fault knowledge base construction device according to claim 6 or 7; The diagnostic device is used for inputting the current detected abnormal failure event into the failure knowledge base of the failure knowledge base construction device when an abnormal failure event is detected, finding out the leaf nodes under the node of the current abnormal failure event from the failure tree, determining all possible failure modes according to all the found abnormal failure events and the corresponding relationship between the abnormal failure events and the failure modes, and judging all the components associated with the current abnormal failure event according to all the determined failure modes and the relationship between the failure modes and the components.

Images