Equipment fault influence range deduction method and system based on event knowledge graph

Abstract

The invention relates to an equipment fault influence range deduction method and system based on an event knowledge graph, and relates to the problems of equipment operation fault association event knowledge graph and fault influence range deduction. Time information is added into an equipment fault knowledge graph, and an entity-centered equipment fault calendar type knowledge graph is constructed. Further constructing an event-centered equipment fault calendar type knowledge graph by mining a logical relationship between the event and the entity; the method comprises the following steps: acquiring a real-time alarm fault event of equipment, deducing a mutual influence relationship and a fault influence range among entities involved in a fault event set through the alarm fault event in combination with an event-centered equipment fault calendar type knowledge graph, and providing an overhaul reference scheme for overhaul, operation and maintenance personnel. Maintenance and operation personnel are helped to quickly analyze accident reasons and comprehensively master key information of fault processing, so that the maintenance efficiency is improved.

Description

Technical Field

[0001] The present invention relates to the field of equipment fault maintenance, and more specifically, to a method and system for deducing the impact range of equipment faults based on an event knowledge graph. Background Art

[0002] As an emerging knowledge system of structured data resources, knowledge graph has become a hot spot for research and development in various industries. In daily maintenance and operation, with the help of knowledge graph models, equipment fault discrimination reasoning is carried out, the impact on the power grid and communication network is automatically analyzed, and the corresponding safety measures are confirmed. However, since the traditional knowledge graph is static, the knowledge graph expressed by triples focuses on entities and the relationship between entities, lacks the description of events and time relationships, and cannot reflect the dynamic evolution of facts over time. In fact, the relationship between entities is time-sensitive, and the relationship between entities will change in different time periods. The entity itself also has time sequence, and there is logic inside the fault event. Therefore, the three dimensions of timeliness, moment, and sequence are added to the knowledge graph to construct an event-centered event-history knowledge graph. It changes over time based on the validity of facts, absorbs new facts in the process of dynamic changes, and dynamically updates entities according to the latest relationship and time. At the same time, the event-centered equipment fault event-history knowledge graph deduces the mutual influence relationship between entities through fault events, which is convenient for maintenance and operation personnel to grasp the latest key information on fault handling and the scope of fault impact, assist in decision-making, and improve efficiency. Summary of the invention

[0003] In response to the above problems, the present invention provides a method and system for deducing the impact scope of equipment failure based on event knowledge graph.

[0004] On the one hand, the present invention provides a method for deducing the impact range of equipment failure based on an event knowledge graph, comprising the following steps:

[0005] Obtain basic data, including equipment ledger data, equipment failure history data, equipment maintenance plans and operating procedures data;

[0006] Based on basic data and introducing time elements, through data preprocessing, entity recognition, relationship extraction, knowledge representation, knowledge reasoning and completion, three types of event-based knowledge graphs are constructed, namely, event-based equipment ledger graph, event-based equipment failure graph, and event-based equipment maintenance plan and operating procedure graph.

[0007] Based on three types of historical knowledge graphs and introducing event elements, through the process of event extraction, entity connection, relationship fusion, knowledge embedding, and knowledge reasoning, an event-centered historical knowledge graph is constructed.

[0008] Acquire real-time equipment alarm fault events, combine them with event-centric history knowledge graph, use recurrent neural network to deduce the mutual influence relationship and fault impact range between entities involved in the fault event set, and automatically generate maintenance plans.

[0009] As a preferred implementation, the step of integrating and constructing an event-centered history-based knowledge graph includes:

[0010] Introduce event elements into the history-based knowledge graph to detect whether the event elements of the equipment failure event exist in any entity of the history-based knowledge graph;

[0011] If it exists, create an "event-entity" relationship between the event and the corresponding entity, and then detect the time relationship between the time when the equipment failure event occurs and the entity;

[0012] If there is a time relationship, create an "event-entity-time" relationship between the event and the corresponding entity;

[0013] Construct an event-centered history-based knowledge graph, which consists of six elements, namely QG = {Q, S q ,M qv ,T q ,M qe ,E q}, where Q is the set of fault events in the graph, S q is a collection of entities; M qv The label values ​​of entities and attributes; T q is the set of directed edges in the graph; M qe is the set of relationship values ​​on the graph edges; E q A collection of graph time elements.

[0014] As a preferred implementation, the step of using a recurrent neural network to deduce the mutual influence relationship between entities involved in the fault event set and the fault impact range includes:

[0015] Construct the equipment fault impact range event set according to the alarm fault event, Q = {q0, q1, ..., q n ,…q m} where q0 represents an alarm fault event, q n represents the nth secondary fault event caused by the alarm fault event, and m∈N* represents the maximum number of secondary fault events caused.

[0016] The entities involved in the secondary fault event caused by the alarm fault event q0 are expressed as Will Use calculation to get the entity set involved in the secondary fault event The secondary fault event q caused by the alarm fault event q0 nThe set of entities involved Where n∈N* represents the maximum number of entities involved in the secondary fault events caused by the alarm fault event q0; the fault events and entities form a bipartite graph, which can be expressed as

[0017] Constructed by entity set The determination coefficient matrix of the elements in under the influence of any element in the event set Q is in Used to determine the fault event q n Medium Entity The results of being affected;

[0018] Project the bipartite graph formed by fault events and entities to the entity set In the example, the projected vertices are represented as Constitute a single-part diagram; set the judgment in the fault event q n Medium Entity The rules of the affected results are analyzed, and each entity is obtained to obtain the transformation rules between the entities involved in the secondary fault event caused by the alarm fault event q0;

[0019] In this way, the association matrix is ​​obtained, and the mutual influence relationship between the entities involved in the secondary fault event set caused by the alarm fault event q0 is obtained. Then, the recurrent neural network is used to deduce the fault impact range caused by the alarm fault event q0, and the maintenance plan is automatically generated according to the fault impact range.

[0020] As a preferred implementation, the configuration determines that in the fault event q n Medium Entity The specific steps of analyzing each entity to obtain the transformation rules between entities involved in the secondary fault event caused by the alarm fault event q0 are as follows:

[0021] Will be used to determine in the event of a fault n Medium Entity The results of the impact There are 4 levels:

[0022]

[0023] After the alarm fault event q0 occurs, the collection The entity in the middle triggers a secondary fault event q under the action of q0 n , if the secondary fault event q n Entity Is the secondary failure event q n The influencing factor formed is n Next entity The judgment coefficient or

[0024] If the entity Only affects the secondary fault event q n ,but

[0025] If the entity Affects the secondary fault event q n , and will be affected again under this secondary event, then

[0026] By analyzing each entity, we can obtain the transformation rules between entities involved in the secondary fault event caused by the alarm fault event q0, which are as follows:

[0027] IF And j≠i,Then ir ij =1;

[0028] Else if And j≠i,Then ir ij =1;

[0029] Else ij =0;

[0030] End

[0031] This gives the correlation matrix

[0032] On the other hand, the present invention also provides a system for deducing the impact range of equipment failures based on an event knowledge graph, comprising:

[0033] Data acquisition module, used to obtain basic data, including equipment ledger data, equipment failure history data, equipment maintenance plan and operating procedure data;

[0034] The knowledge graph construction module is based on basic data and introduces the time element. Through data preprocessing, entity recognition, relationship extraction, knowledge representation, knowledge reasoning and completion, it constructs three types of event-based knowledge graphs, namely, event-based equipment ledger graph, event-based equipment fault graph, and event-based equipment maintenance plan and operating procedure graph, which are entity-centered.

[0035] The knowledge graph fusion module is based on three types of historical knowledge graphs and introduces event elements. Through the process of event extraction, entity connection, relationship fusion, knowledge embedding, and knowledge reasoning, it integrates and constructs an event-centered historical knowledge graph.

[0036] The deduction module is used to obtain real-time alarm fault events of equipment, combine the event-centric history knowledge graph, use the recurrent neural network to deduce the mutual influence relationship and fault impact range between the entities involved in the fault event set, and automatically generate maintenance plans.

[0037] As a preferred implementation, the step of integrating and constructing an event-centered history-based knowledge graph includes:

[0038] Introduce event elements into the history-based knowledge graph to detect whether the event elements of the equipment failure event exist in any entity of the history-based knowledge graph;

[0039] If it exists, create an "event-entity" relationship between the event and the corresponding entity, and then detect the time relationship between the time when the equipment failure event occurs and the entity;

[0040] If there is a time relationship, create an "event-entity-time" relationship between the event and the corresponding entity;

[0041] Construct an event-centered history-based knowledge graph, which consists of six elements, namely QG = {Q, S q ,M qv ,T q ,M qe ,E q}, where Q is the set of fault events in the graph, S q is a collection of entities; M qv The label values ​​of entities and attributes; T q is the set of directed edges in the graph; M qe is the set of relationship values ​​on the graph edges; E q A collection of graph time elements.

[0042] As a preferred implementation, the step of using a recurrent neural network to deduce the mutual influence relationship between entities involved in the fault event set and the fault impact range includes:

[0043] Construct the equipment fault impact range event set according to the alarm fault event, Q = {q0, q1, ..., q n ,…q m} where q0 represents an alarm fault event, q n represents the nth secondary fault event caused by the alarm fault event, and m∈N* represents the maximum number of secondary fault events caused.

[0044] The entities involved in the secondary fault event caused by the alarm fault event q0 are expressed as Will Use calculation to get the entity set involved in the secondary fault event The secondary fault event q caused by the alarm fault event q0n The set of entities involved Where n∈N* represents the maximum number of entities involved in the secondary fault events caused by the alarm fault event q0; the fault events and entities form a bipartite graph, which can be expressed as

[0045] Constructed by entity set The determination coefficient matrix of the elements in under the influence of any element in the event set Q is in Used to determine the fault event q n Medium Entity The results of being affected;

[0046] Project the bipartite graph formed by fault events and entities to the entity set In the example, the projected vertices are represented as Constitute a single-part diagram; set the judgment in the fault event q n Medium Entity The rules of the affected results are analyzed, and each entity is obtained to obtain the transformation rules between the entities involved in the secondary fault event caused by the alarm fault event q0;

[0047] In this way, the association matrix is ​​obtained, and the mutual influence relationship between the entities involved in the secondary fault event set caused by the alarm fault event q0 is obtained. Then, the recurrent neural network is used to deduce the fault impact range caused by the alarm fault event q0, and the maintenance plan is automatically generated according to the fault impact range.

[0048] As a preferred implementation, the configuration determines that in the fault event q n Medium Entity The specific steps of analyzing each entity to obtain the transformation rules between entities involved in the secondary fault event caused by the alarm fault event q0 are as follows:

[0049] Will be used to determine in the event of a fault n Medium Entity The results of the impact There are 4 levels:

[0050]

[0051] After the alarm fault event q0 occurs, the collection The entity in the middle triggers a secondary fault event q under the action of q0 n , if the secondary fault event q n Entity Is the secondary failure event q n The influencing factor formed is n Next entity The judgment coefficient or

[0052] If the entity Only affects the secondary fault event q n ,but

[0053] If the entity Affects the secondary fault event q n , and will be affected again under this secondary event, then

[0054] By analyzing each entity, we can obtain the transformation rules between entities involved in the secondary fault event caused by the alarm fault event q0, which are as follows:

[0055] IF And j≠i,Then ir ij =1;

[0056] Else if And j≠i,Then ir ij =1;

[0057] Else ij =0;

[0058] End

[0059] This gives the correlation matrix

[0060] On the other hand, the present invention further proposes an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the method described in any embodiment of the present invention when executing the program.

[0061] In yet another aspect, the present invention further provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method described in any embodiment of the present invention.

[0062] Compared with the prior art, the present invention has the following advantages:

[0063] 1. The present invention adds the time element to the knowledge graph, adding three dimensions: time series, moment, and timeliness. In terms of time series, the entities in the knowledge graph have 1-to-many, many-to-1, and many-to-many connections, and can be strung together into an orderly history through time series facts. In terms of moment, the relationship between the state of the entity and the event when the equipment fails is established with the time point as the link. In terms of timeliness, the time facts are classified and processed, reflecting the activation time and end time of the entity.

[0064] 2. Compared with the traditional entity-centric knowledge graph, the present invention proposes an event-centric history-based knowledge graph, which can provide maintenance personnel with comprehensive information about past fault events and assist in decision-making;

[0065] 3. The deductive analysis of equipment failures through event-centered history-based knowledge graphs can not only predict the scope of fault impact, but also provide maintenance plans with more reference value.

[0066] Additional aspects and advantages of the present invention will be set forth in the following description, and some of them will be apparent from the description, or may be understood by practicing the present invention. In addition, the various aspects and advantages of the present invention may be realized and obtained by the method steps and combinations particularly pointed out in the appended claims. BRIEF DESCRIPTION OF THE DRAWINGS

[0067] Figure 1 A flow chart of a method for inferring the impact range of equipment failures based on an event knowledge graph provided in an embodiment of the present application;

[0068] Figure 2 A schematic diagram of an entity-centric event-based knowledge graph provided in an embodiment of the present application;

[0069] Figure 3 A schematic diagram of an event-centric history-based knowledge graph provided in an embodiment of the present application;

[0070] Figure 4 A schematic diagram of a fault event and entity bipartite graph provided in an embodiment of the present application;

[0071] Figure 5 A schematic diagram of a fault event entity single-part diagram provided in an embodiment of the present application. DETAILED DESCRIPTION

[0072] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0073] It should be understood that the step numbers used in this document are only for convenience of description and are not intended to limit the order in which the steps are executed.

[0074] It should be understood that the terms used in the present specification are only for the purpose of describing specific embodiments and are not intended to limit the present invention. As used in the present specification and the appended claims, unless the context clearly indicates otherwise, the singular forms "a", "an" and "the" are intended to include plural forms.

[0075] The terms “include” and “comprising” indicate the presence of described features, integers, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or combinations thereof.

[0076] The term "and / or" means and includes any and all possible combinations of one or more of the associated listed items.

[0077] Embodiment 1:

[0078] This embodiment provides a method for deducing the impact range of equipment failures based on event knowledge graphs, which can be applied to the field of fault diagnosis and impact range deduction of power grid backbone optical communication systems, and can also be used in other fields. The present invention takes the failure of a power grid backbone optical communication network as an example for explanation, as shown in the attached figure. Figure 1 The specific contents are as follows:

[0079] Step 01: Collect equipment inventory data of a certain power grid backbone optical communication network, equipment failure history data, equipment maintenance plan and operating procedure data.

[0080] Step 02: Based on the equipment inventory data, equipment failure history data, equipment maintenance plan and operating procedures of the power grid backbone optical communication system, the data is constructed and the time element is introduced. Through data preprocessing, entity recognition, relationship extraction, knowledge representation, knowledge reasoning and completion, etc. Figure 2 As shown, an entity-centered event-based equipment inventory map is constructed; an entity-centered event-based equipment fault map is constructed; and an entity-centered event-based equipment maintenance plan and operating procedure map is constructed.

[0081] Step 03: Based on the above historical knowledge graph and introducing event elements, through the processes of event extraction, entity connection, relationship fusion, knowledge embedding, knowledge reasoning, etc. Figure 3 As shown in the figure, an event-centered history-based knowledge graph is constructed; first, the event element of the equipment failure event is detected to see if it exists in the entity of the above history-based knowledge graph. If so, an "event-entity" relationship between the event and the corresponding entity is created. Then, the time relationship between the time when the equipment failure event occurred and the entity is detected. If a time relationship exists, an "event-entity-time" relationship between the event and the corresponding entity is created.

[0082] Step 04: Obtain real-time alarm fault events of equipment, combine them with event-centric history-based knowledge graph, use recurrent neural network to deduce the mutual influence relationship and fault impact range between entities involved in the fault event set, and automatically generate maintenance plans.

[0083] The spectrum consists of 6 elements, QG = {Q, S q ,M qv ,T q ,M qe ,E q}, where Q is the set of fault events in the graph; S q =S qe ∪S ql , S qe is the entity set in the graph, S ql is the set of attributes in the graph; M qv are the label values ​​of graph entities and attributes; T q is the set of edges in the graph. qe is the set of edge relationship values; E q is the set of graph time elements;

[0084] M qe There are three types of edges: event connection edges, entity connection edges, and attribute edges. qei Connect two points qi and q i , if s qi ∈S qe ,q i ∈Q, indicating m qei What is connected is events and entities, m qei is the event connection edge, m qei = {q i ,M qei ,s qi ,[t se ]};m qei Connect two points qi and qj , if s qi ∈S qe ,s qj ∈S qe , indicating m qei The connection is between two entities, m qei is the entity connection edge, m qei ={s qi ,M qei ,s qj ,[t1,t2]}; if s qi ∈S qe ,s qj ∈S ql , indicating m qeiWhat is connected is the entity and the attribute, m qei is the attribute connection edge, m qei ={s qi ,M qei ,s qj}.

[0085] E q There are three types of time elements, one is {t s ,t e}, t s <t e , t s is the starting time, t e is the termination time. One is {t s ,+∝}, indicating that from t s So far, {+∝,t e} indicates from the past to t e moment. Another is {t se} indicates a moment.

[0086] The network management system obtains real-time alarm fault events and constructs an event set of equipment fault impact range based on the alarm fault events, Q = {q0, q1, …, q n ,…q m} where q0 represents an alarm fault event, q n Represents the secondary fault event caused by the alarm fault event, and m∈N* represents the maximum number of secondary fault events caused.

[0087] The entities involved in the secondary fault event caused by the alarm fault event q0 are expressed as Will Use calculation to get the entity set involved in the secondary fault event The secondary fault event q caused by the alarm fault event q0 n The set of entities involved Where n∈N* represents the maximum number of entities involved in the secondary fault events caused by the alarm fault event q0. Figure 4 As shown, the fault event and the entity form a bipartite graph, which is expressed as

[0088] Construct the entity set involved in the secondary fault event caused by the alarm fault event q0 The determination coefficient matrix of the elements in the event set Q under the influence of any element is in Used to determine the fault event q n Medium Entity The affected results are divided into 4 levels:

[0089]

[0090] like Figure 5 As shown, the fault event entity bipartite graph is projected onto the entity set involved in the secondary fault event caused by the alarm fault event q0. The vertices in are represented as Constitute a single-part graph, after the alarm fault event q0 occurs, the collection The entity in the middle triggers a secondary fault event q under the action of q0 n , if the secondary fault event q n Entity Is the secondary failure event q n The influencing factor formed is n Next entity The judgment coefficient or

[0091] If the entity Only affects the secondary fault event q n ,Right now

[0092] If the entity Affects the secondary fault event q n , and will be affected again under this secondary event, that is,

[0093] By analyzing each entity step by step in this way, we can obtain the transformation rules between entities involved in the secondary fault event caused by the alarm fault event q0:

[0094] IF And j≠i,Then ir ij =1;

[0095] Else if And j≠i,Then ir ij =1;

[0096] Else ij =0;

[0097] End

[0098] From this we can get the correlation matrix The mutual influence relationship between the entities involved in the secondary fault event set caused by the alarm fault event q0 is obtained, and finally the fault impact range caused by the alarm fault event q0 is deduced using the recurrent neural network (RNN), and the maintenance plan is automatically generated according to the fault impact range.

[0099] Embodiment 2:

[0100] This embodiment provides a system for deducing the impact range of equipment failures based on an event knowledge graph, including:

[0101] The data acquisition module is used to acquire basic data, including equipment inventory data, equipment failure history data, equipment maintenance plan and operating procedure data; this module is used to implement the function of step 01 in Example 1, which will not be repeated here.

[0102] The knowledge graph construction module is based on basic data and introduces the time element. Through data preprocessing, entity recognition, relationship extraction, knowledge representation, knowledge reasoning and completion, it constructs three types of historical knowledge graphs, namely, entity-centered historical equipment inventory graph, historical equipment fault graph, and historical equipment maintenance plan and operating procedure graph. This module is used to implement the function of step 02 in Example 1, which will not be repeated here.

[0103] The knowledge graph fusion module is based on three types of historical knowledge graphs and introduces event elements. Through the process of event extraction, entity connection, relationship fusion, knowledge embedding, and knowledge reasoning, it integrates and constructs an event-centered historical knowledge graph. This module is used to implement the function of step 03 in Example 1, which will not be repeated here.

[0104] The deduction module is used to obtain real-time alarm fault events of equipment, combine the event-centered history-based knowledge graph, and use the recurrent neural network to deduce the mutual influence relationship and fault impact range between the entities involved in the fault event set, and automatically generate a maintenance plan; this module is used to implement the function of step 04 in Example 1, which will not be repeated here.

[0105] Embodiment three:

[0106] This embodiment proposes an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, the method for deducing the impact range of equipment failure based on the event knowledge graph as described in any embodiment of the present invention is implemented.

[0107] Embodiment 4:

[0108] This embodiment proposes a computer-readable storage medium on which a computer program is stored. When the program is executed by a processor, the method for deducing the impact range of equipment failure based on the event knowledge graph as described in any embodiment of the present invention is implemented.

[0109] In the embodiments of the present application, "at least one" refers to one or more, and "more than one" refers to two or more. "And / or" describes the association relationship of associated objects, indicating that three relationships may exist. For example, A and / or B can represent the existence of A alone, the existence of A and B at the same time, and the existence of B alone. Among them, A and B can be singular or plural. The character " / " generally indicates that the previous and next associated objects are in an "or" relationship. "At least one of the following" and similar expressions refer to any combination of these items, including any combination of single or plural items. For example, at least one of a, b and c can be represented by: a, b, c, a and b, a and c, b and c, or a and b and c, where a, b, c can be single or multiple.

[0110] Those of ordinary skill in the art will appreciate that the various units and algorithm steps described in the embodiments disclosed herein can be implemented in a combination of electronic hardware, computer software, and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0111] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0112] In several embodiments provided in the present application, if any function is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (Read-Only Memory; hereinafter referred to as: ROM), random access memory (Random Access Memory; hereinafter referred to as: RAM), disk or optical disk, and other media that can store program codes.

[0113] The above descriptions are merely embodiments of the present invention and are not intended to limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made using the contents of the present invention specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A method for deducing the impact range of equipment failure based on event knowledge graph, characterized in that: The following steps are involved: Obtain basic data, including equipment ledger data, equipment failure history data, equipment maintenance plans and operating procedures data; Based on basic data and introducing time elements, through data preprocessing, entity recognition, relationship extraction, knowledge representation, knowledge reasoning and completion, three types of event-based knowledge graphs are constructed, namely, event-based equipment ledger graph, event-based equipment failure graph, and event-based equipment maintenance plan and operating procedure graph. Based on three types of historical knowledge graphs and introducing event elements, through the process of event extraction, entity connection, relationship fusion, knowledge embedding, and knowledge reasoning, an event-centered historical knowledge graph is constructed. Acquire real-time equipment alarm fault events, combine them with event-centric history knowledge graph, use recurrent neural network to deduce the mutual influence relationship and fault impact range between entities involved in the fault event set, and automatically generate maintenance plans.

2. According to claim 1, a method for deducing the impact range of equipment failure based on event knowledge graph is characterized in that: The steps of integrating and constructing an event-centered history-based knowledge graph include: Introduce event elements into the history-based knowledge graph to detect whether the event elements of the equipment failure event exist in any entity of the history-based knowledge graph; If it exists, create an "event-entity" relationship between the event and the corresponding entity, and then detect the time relationship between the time when the equipment failure event occurs and the entity; If there is a time relationship, create an "event-entity-time" relationship between the event and the corresponding entity; Construct an event-centered history-based knowledge graph, which consists of six elements, namely QG = {Q, S q ,M qv ,T q ,M qe ,E q }, where Q is the set of fault events in the graph, S q is a collection of entities; M qv The label values ​​of entities and attributes; T q is the set of directed edges in the graph; M qe is the set of relationship values ​​on the edge of the graph; E q A collection of graph time elements.

3. According to claim 1, a method for deducing the impact range of equipment failure based on event knowledge graph is characterized in that: The step of using a recurrent neural network to deduce the mutual influence relationship between entities involved in the fault event set and the fault impact range includes: Construct the equipment fault impact range event set according to the alarm fault event, Q = {q0, q1, ..., q n ,…q m } where q0 represents an alarm fault event, q n represents the nth secondary fault event caused by the alarm fault event, and m∈N* represents the maximum number of secondary fault events caused. The entities involved in the secondary fault event caused by the alarm fault event q0 are expressed as Will Use calculation to get the entity set involved in the secondary fault event The secondary fault event q caused by the alarm fault event q0 n The set of entities involved Where n∈N* represents the maximum number of entities involved in the secondary fault events caused by the alarm fault event q0; the fault events and entities form a bipartite graph, which can be expressed as Constructed by entity set The determination coefficient matrix of the elements in under the influence of any element in the event set Q is in Used to determine the fault event q n Medium Entity The results of being affected; Project the bipartite graph formed by fault events and entities to the entity set In the example, the projected vertices are represented as Constitute a single-part diagram; set the judgment in the fault event q n Medium Entity The rules of the affected results are analyzed, and each entity is obtained to obtain the transformation rules between the entities involved in the secondary fault event caused by the alarm fault event q0; In this way, the association matrix is ​​obtained, and the mutual influence relationship between the entities involved in the secondary fault event set caused by the alarm fault event q0 is obtained. Then, the recurrent neural network is used to deduce the fault impact range caused by the alarm fault event q0, and the maintenance plan is automatically generated according to the fault impact range.

4. According to claim 3, a method for deducing the impact range of equipment failure based on event knowledge graph is characterized in that: The setting determines that in the event of a fault q n Medium Entity The specific steps of analyzing each entity to obtain the transformation rules between entities involved in the secondary fault event caused by the alarm fault event q0 are as follows: Will be used to determine in the event of a fault n Medium Entity The results of the impact There are 4 levels: After the alarm fault event q0 occurs, the collection The entity in the middle triggers a secondary fault event q under the action of q0 n , if the secondary fault event q n Entity Is the secondary failure event q n The influencing factor formed is n Next entity The judgment coefficient or If the entity Only affects the secondary fault event q n ,but If the entity Affects the secondary fault event q n , and will be affected again under this secondary event, then By analyzing each entity, we can obtain the transformation rules between entities involved in the secondary fault event caused by the alarm fault event q0, which are as follows: and j ≠ i, Then ir ij = 1; Else if and j ≠ i, Then ir ij = 1; Elsa is ij =0; End Thus we get the correlation matrix IR = [ir ij ] n×n .

5. A system for deducing the impact range of equipment failures based on event knowledge graph, characterized in that: include: Data acquisition module, used to obtain basic data, including equipment ledger data, equipment failure history data, equipment maintenance plan and operating procedure data; The knowledge graph construction module is based on basic data and introduces the time element. Through data preprocessing, entity recognition, relationship extraction, knowledge representation, knowledge reasoning and completion, it constructs three types of event-based knowledge graphs, namely, event-based equipment ledger graph, event-based equipment fault graph, and event-based equipment maintenance plan and operating procedure graph, which are entity-centered. The knowledge graph fusion module is based on three types of historical knowledge graphs and introduces event elements. Through the process of event extraction, entity connection, relationship fusion, knowledge embedding, and knowledge reasoning, it integrates and constructs an event-centered historical knowledge graph. The deduction module is used to obtain real-time alarm fault events of equipment, combine the event-centric history knowledge graph, use the recurrent neural network to deduce the mutual influence relationship and fault impact range between the entities involved in the fault event set, and automatically generate maintenance plans.

6. According to claim 5, a device failure impact range deduction system based on event knowledge graph is characterized in that: The steps of integrating and constructing an event-centered history-based knowledge graph include: Introduce event elements into the history-based knowledge graph to detect whether the event elements of the equipment failure event exist in any entity of the history-based knowledge graph; If it exists, create an "event-entity" relationship between the event and the corresponding entity, and then detect the time relationship between the time when the equipment failure event occurs and the entity; If there is a time relationship, create an "event-entity-time" relationship between the event and the corresponding entity; Construct an event-centered history-based knowledge graph, which consists of six elements, namely QG = {Q, S q ,M qv ,T q ,M qe ,E q }, where Q is the set of fault events in the graph, S q is a collection of entities; M qv The label values ​​of entities and attributes; T q is the set of directed edges in the graph; M qe is the set of relationship values ​​on the edge of the graph; E q A collection of graph time elements.

7. According to claim 5, a device failure impact range deduction system based on event knowledge graph is characterized in that: The step of using a recurrent neural network to deduce the mutual influence relationship between entities involved in the fault event set and the fault impact range includes: Construct the equipment fault impact range event set according to the alarm fault event, Q = {q0, q1, ..., q n ,…q m } where q0 represents an alarm fault event, q n represents the nth secondary fault event caused by the alarm fault event, and m∈N* represents the maximum number of secondary fault events caused. The entities involved in the secondary fault event caused by the alarm fault event q0 are expressed as Will Use calculation to get the entity set involved in the secondary fault event The secondary fault event q caused by the alarm fault event q0 n The set of entities involved Where n∈N* represents the maximum number of entities involved in the secondary fault events caused by the alarm fault event q0; the fault events and entities form a bipartite graph, which can be expressed as Constructed by entity set The determination coefficient matrix of the elements in under the influence of any element in the event set Q is in Used to determine the fault event q n Medium Entity The results of being affected; Project the bipartite graph formed by fault events and entities to the entity set In the example, the projected vertices are represented as Constitute a single-part diagram; set the judgment in the fault event q n Medium Entity The rules of the affected results are analyzed, and each entity is obtained to obtain the transformation rules between the entities involved in the secondary fault event caused by the alarm fault event q0; In this way, the association matrix is ​​obtained, and the mutual influence relationship between the entities involved in the secondary fault event set caused by the alarm fault event q0 is obtained. Then, the recurrent neural network is used to deduce the fault impact range caused by the alarm fault event q0, and the maintenance plan is automatically generated according to the fault impact range.

8. The device failure impact range deduction system based on event knowledge graph according to claim 7 is characterized in that: The setting determines that in the event of a fault q n Medium Entity The specific steps of analyzing each entity to obtain the transformation rules between entities involved in the secondary fault event caused by the alarm fault event q0 are as follows: Will be used to determine in the event of a fault n Medium Entity The results of the impact There are 4 levels: After the alarm fault event q0 occurs, the collection The entity in the middle triggers a secondary fault event q under the action of q0 n , if the secondary fault event q n Entity Is the secondary failure event q n The influencing factor formed is n Next entity The judgment coefficient or If the entity Only affects the secondary fault event q n ,but If the entity Affects the secondary fault event q n , and will be affected again under this secondary event, then By analyzing each entity, we can obtain the transformation rules between entities involved in the secondary fault event caused by the alarm fault event q0, which are as follows: IF and j ≠ i, Then ir ij = 1; Else if and j ≠ i, Then ir ij = 1; Elsa is ij =0; End Thus we get the correlation matrix IR = [ir ij ] n×n .

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the method for deducing the impact range of equipment failure based on the event knowledge graph as described in any one of claims 1 to 4 is implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, it implements the method for deducing the impact range of equipment failure based on the event knowledge graph as described in any one of claims 1 to 4.

Images