CN113220903B - Power accident visual analysis system and method based on knowledge graph - Google Patents

Abstract

The application discloses a power accident visual analysis system and method based on a knowledge graph, and the system and method comprises the following steps that firstly, a knowledge graph layer processes text data of a power accident event and constructs a graph database; the interface interaction layer acquires a data request and sends the data request to the data access layer; then, the data access layer processes the graph database according to the request and sends a processing result to the service logic layer, and the service logic layer respectively performs mode switching, node and relation editing and event process analysis according to the request and the processing result to determine a main responsible person and electric power equipment of the event; and finally, the interface interaction layer visually displays the processing result of the business logic layer. The method and the device solve the problem that people are difficult to find out the main responsible person or the main responsible power equipment when the power accident case happens because the condition of the power accident case is recorded in a traditional text mode and is not beneficial to analyzing the case, and make the visual content easier for people to understand.

Description

A system and method for visual analysis of power accidents based on knowledge graph

technical field

The present application relates to the field of electric power technology, in particular to a system and method for visual analysis of power accidents based on knowledge graphs.

Background technique

Every year, there are many power accident cases in our country, which will bring a lot of human and economic losses to the power grid company and the country. In order to reduce these losses and reduce the occurrence of power accident cases, it is necessary to conduct a detailed analysis of each power accident case And trace the source, find out the source of the accident and the main person responsible for the accident, so as to provide experience and lessons for future electric power safety production, promote the safe development of electric power production, and at the same time be able to alert electric workers to work carefully.

At present, in the field of electric power technology, the situation of electric accident cases is still recorded in traditional text, and text data is formed and analyzed. Since this method is not conducive to the correlation analysis and causal relationship mining of electric accident cases, it is difficult to Help people to accurately find out the main responsible person or the main responsible power equipment for the occurrence of electric accidents.

Therefore, this paper proposes a visual analysis system and method for electric accidents based on knowledge graphs, which is used to solve the problem that the traditional text recording of electric accident cases is not conducive to the analysis of the relevance of electric accident cases and the mining of causality. Therefore, it is difficult to help people accurately find out the main responsible person or the main responsible power equipment for the occurrence of electric accident cases.

Contents of the invention

In order to solve the problem of recording power accident cases in traditional text, it is not conducive to the correlation analysis and causal relationship mining of the cases, so it is difficult to help people accurately find out the main responsible person or the main responsible power equipment for the electric accident case. Problem, this application discloses a system and method for visual analysis of power accidents based on knowledge graphs through the following embodiments.

The first aspect of the present application discloses a visual analysis system for power accidents based on a knowledge map, including: a knowledge map layer, a data access layer, a business logic layer, and an interface interaction layer that are sequentially connected to each other;

The knowledge map layer is used to obtain text data of electric accident events, perform first preprocessing on the text data, and construct a graph database according to the result of the first preprocessing, and the first preprocessing includes data Acquisition, knowledge extraction, and knowledge update, the graph database includes a knowledge map, and the knowledge map includes nodes and relationships, and the nodes include operators related to the power accident event, events related to the power accident event, and For electrical equipment related to the electrical accident event, the relationship is the relationship between any of the nodes and other nodes;

The interface interaction layer is used to obtain a user's data request, and send the data request to the data access layer through the business logic layer;

The data access layer is configured to perform a second preprocessing on the nodes and the relationships in the graph database according to the data request, and send the result of the second preprocessing to the business logic layer Feedback, the second preprocessing includes adding, deleting, modifying, querying;

The business logic layer includes:

A mode switching module, configured to switch between the view mode of the person in charge and the view mode of electric equipment according to the data request and the result of the second preprocessing, and the view mode of the person in charge is used to display each of the tasks The operations of personnel in the power accident event and the links between the operations, the power equipment view mode is used to display the problems generated by each of the power equipment in the power accident event and the problems and the connection between the electrical accident events;

A node editing module, configured to perform node editing on the result of the second preprocessing according to the data request, and send the node editing result to the graph database through the data access layer for storage, and the node editing Including adding nodes, deleting nodes and updating node attributes;

A relationship editing module, configured to perform relationship editing on the result of the second preprocessing according to the data request, and send the relationship editing result to the graph database through the data access layer for storage, and the relationship editing Including adding relationship, deleting relationship and updating relationship attributes;

An auxiliary analysis module, configured to analyze the process of the power accident event according to the data request and the second preprocessing result, and obtain the influence of each node on the power accident event, wherein , the node with the greatest influence is the main responsible person or the main responsible power equipment for the power accident event;

The interface interaction layer is also used to obtain the processing result of the business logic layer, and to visually display the processing result according to the data request, the processing result includes the switching result of the mode switching module, the The node editing result of the node editing module, the relationship editing result of the relationship editing module, or the analysis result of the auxiliary analysis module.

Optionally, a server is also included;

The server is connected to the data access layer, and is used for backing up the result of the second preprocessing to realize data sharing.

Optionally, when the interface interaction layer visually displays the analysis results of the auxiliary analysis module according to the data request, the nodes with greater influence have a larger visualization radius.

The second aspect of the present application discloses a knowledge map-based power accident visualization analysis method, which is applied to the knowledge map-based power accident analysis method described in the first aspect of the present application In the visual analysis system, including:

Acquiring the text data of the power accident event, performing a first preprocessing on the text data, and constructing a graph database according to the result of the first preprocessing, the first preprocessing includes data collection, knowledge extraction and knowledge updating, The graph database includes a knowledge graph, the knowledge graph includes nodes and relationships, and the nodes include operators related to the power accident event, events related to the power accident event, and For electric equipment, the relationship is the relationship between any of the nodes and other nodes;

Obtain the user's data request, and send the data request to the data access layer through the business logic layer;

According to the data request, perform a second preprocessing on the nodes and the relationships in the graph database, and feed back the result of the second preprocessing to the business logic layer, the second preprocessing Including adding, deleting, modifying, querying;

According to the data request and the result of the second preprocessing, switch between the person in charge view mode and the electric equipment view mode, and obtain the switching result, and the person in charge view mode is used to display each of the operators The operation in the electric accident event and the connection between the operations, the electric equipment view mode is used to display the problems of each electric equipment in the electric accident event and the relationship between the problems and the The connection between the above-mentioned electrical accident events;

According to the data request, node editing is performed on the result of the second preprocessing, and the node editing result is sent to the graph database of the knowledge map layer through the data access layer for storage, and the node editing includes adding Node, delete node and update node attributes;

According to the data request, perform relation editing on the result of the second preprocessing, and send the relation editing result to the graph database of the knowledge map layer through the data access layer for storage, the relation editing Including adding relationship, deleting relationship and updating relationship attributes;

According to the data request and the result of the second preprocessing, analyze the process of the power accident event, and obtain the influence of each node on the power accident event, the node with the greatest influence That is, the main responsible person or the main responsible electrical equipment for the said electrical accident;

According to the data request, the switching result, the node editing result, the relationship editing result or the analysis result of the electric accident event process are visually displayed through the interface interaction layer.

Optionally, the knowledge graph layer is constructed by means of crowdsourcing construction of the knowledge platform.

Optionally, performing a second preprocessing on the nodes and the relationships in the graph database according to the data request, and feeding back the result of the second preprocessing to the business logic layer After that, also include:

The result of the second preprocessing is backed up to realize data sharing.

Optionally, the interface development of the interface interaction layer is realized by JavaScript programming language.

Optionally, according to the data request, when visually displaying the analysis results of the process of the electric accident event, nodes with greater influence have a larger visualization radius.

Optionally, the influence of any node on the power accident event is obtained by the following formula:

Among them, x represents any node, PR(x) represents the page ranking value of any node x, the larger the PR(x), the greater the influence of any node x, and β is between 0 and 1 The damping coefficient of , represents the random probability from any node x to the next node, Ai(i=1, 2, 3,...,n) represents the i-th node pointing to any node x, C(Ai) is the number of nodes pointing outward from node Ai, and E(x) is the decay factor, representing a certain vector corresponding to the set of nodes.

This application discloses a system and method for visual analysis of power accidents based on knowledge graphs, including: a knowledge graph layer, a data access layer, a business logic layer, and an interface interaction layer that are sequentially connected to each other; The text data is first preprocessed, and a graph database is built according to the processing results; the interface interaction layer obtains the user's data request and sends it to the data access layer; then, the data access layer performs the second preprocessing on the graph database according to the data request, And the processing result is fed back to the business logic layer, and the business logic layer includes: a mode switching module, a node editing module, a relationship editing module and an auxiliary analysis module, which are respectively used to switch modes according to user requests and the results of the second preprocessing , edit nodes and relationships, and analyze the event process to determine the main responsible person or the main responsible power equipment for the power accident event; finally, the interface interaction layer visually displays the processing results of the business logic layer according to user requests .

Using the above-mentioned device for testing the bearing capacity of briquetting blocks in photovoltaic power plants solves the problem that it is difficult to help people accurately because it is not conducive to the analysis of the relevance of the cases and the excavation of the causal relationship to record the power accident cases in the traditional text way. Find out the main responsible person or the main responsible power equipment problem in the occurrence of electric accident cases, and help the power grid company to analyze the causes of power accident cases in the past and the relationship and influencing factors among the entities in the event. Responsibility for the event is identified, while alerting business personnel, reducing the occurrence of power production accidents, and optimizing the traditional visualization technology by introducing knowledge graph technology, making the visualized content easier for people to understand.

Description of drawings

In order to illustrate the technical solution of the present application more clearly, the accompanying drawings that need to be used in the embodiments will be briefly introduced below. Obviously, for those of ordinary skill in the art, on the premise of not paying creative work, there are also Additional figures can be derived from these figures.

FIG. 1 is a schematic structural diagram of a knowledge map-based power accident visualization analysis system provided by an embodiment of the present application;

FIG. 2 is a schematic workflow diagram of a knowledge map-based power accident visualization analysis method provided by an embodiment of the present application;

FIG. 3 is a schematic diagram of a mode switching process of a knowledge graph-based power accident visualization analysis method provided by an embodiment of the present application;

FIG. 4 is a schematic diagram of the node editing process of a knowledge map-based power accident visualization analysis method provided by the embodiment of the present application;

FIG. 5 is a schematic diagram of a relationship editing process of a knowledge graph-based power accident visualization analysis method provided by an embodiment of the present application.

Detailed ways

In order to solve the problem of recording power accident cases in traditional text, it is not conducive to the correlation analysis and causal relationship mining of the cases, so it is difficult to help people accurately find out the main responsible person or the main responsible power equipment for the electric accident case. Problem, this application discloses a system and method for visual analysis of power accidents based on knowledge graphs through the following embodiments.

The first embodiment of the present application discloses a visual analysis system for power accidents based on knowledge graphs, referring to the structural diagram shown in Figure 1, which includes: a knowledge graph layer, a data access layer, a business logic layer, and an interface interaction layer that are sequentially connected to each other .

Specifically, the system of this application adopts a four-layer architecture, and the functions of each layer adopt modular interaction, which reduces the dependence between layers, which makes the system structure clear, the system scalability and independence are stronger, and it is convenient for developers to carry out Development and maintenance. The four-layer architecture is the interface interaction layer, business logic layer, data access layer, and knowledge graph layer from top to bottom.

The knowledge map layer is used to obtain text data of electric accident events, perform first preprocessing on the text data, and construct a graph database according to the result of the first preprocessing, and the first preprocessing includes data Acquisition, knowledge extraction, and knowledge update, the graph database includes a knowledge map, and the knowledge map includes nodes and relationships, and the nodes include operators related to the power accident event, events related to the power accident event, and For the electrical equipment related to the electrical accident event, the relationship is the relationship between any of the nodes and other nodes.

Specifically, the knowledge graph layer shown provides data storage and semantic query matching for the data access layer shown, and at the same time provides reasoning based on knowledge graphs to help clarify the context of events. The construction of the knowledge graph is carried out by means of knowledge platform crowdsourcing. Crowdsourcing is a new type of outsourcing model, which connects a group of loose task senders and task finishers to realize task contracting, matching, completion and Compared with traditional outsourcing, crowdsourcing performs better in terms of expenses, time and flexibility. The knowledge graph is composed of nodes (entities) and edges (relationships). Finally, the constructed knowledge graph is stored in the graph database. The graph database uses a high-performance Neo4j graph database, based on the reasoning of the knowledge graph, to help the Clearing the context of the incident can help people accurately find out the main responsible person or the main responsible power equipment problem in the electric accident case, and carry out accountability.

The interface interaction layer is used to obtain the user's data request, and send the data request to the data access layer through the business logic layer.

Specifically, the interface interaction layer directly interacts with the user, and its main function is to understand the user's operation, send the received data request to the business logic layer, and feed back the processing result to the interface interaction layer. The interaction layer does not involve logical judgment and other operations, and the main functions can be realized by visiting the website through a browser. Interface development mainly uses the most popular JavaScript programming language at present, and uses JavaScript programming language to manipulate the Document Object Model (DOM) to add interactive functions to hypertext markup (HTML) web pages. As a lightweight just-in-time compiled programming language, the JavaScript programming language has the characteristics of portability and cross-platform, and can run under most browsers. Use JQuery to optimize hypertext markup (HTML) document operations, event handling and animation design, JQuery allows asynchronous JavaScript and XML (AJAX) to load data in the background and display it on the global wide area network or World Wide Web On the (web) page, there is no need to reload the entire page, and using JQuery helps reduce the amount of code for agile development.

The interface interaction layer uses a front-end rendering method for rendering, adopts technologies such as selection interaction and filtering interaction, and simultaneously uses a double-layer canvas to optimize user experience. The visual channels used mainly include shape, color, direction, and size. There are many colors. In the visual editing page, as an example, red indicates unselected nodes, green indicates unselected relationships, and bright blue indicates nodes or relationships are selected. There are three types of shapes, namely straight lines and arrows Straight, circular. Among them, circles represent nodes, undirected straight lines represent general relations, and directed straight lines represent causal relations. In the node influence analysis interface, the radius of the node varies from 20px to 60px. The design of the interface interaction layer enables the present application to take into account the user experience when solving technical problems, further increasing the efficiency of solving technical problems.

The data access layer is configured to perform a second preprocessing on the nodes and the relationships in the graph database according to the data request, and send the result of the second preprocessing to the business logic layer Feedback, the second preprocessing includes adding, deleting, modifying, and querying.

Further, the server is also included.

The server is connected to the data access layer, and is used for backing up the result of the second preprocessing to realize data sharing.

Specifically, the data access layer is mainly used to connect the server and the graph database, and directly perform operations such as adding, deleting, modifying, and querying the graph database. The data access layer directly feeds back the processing results to the business logic layer, and also stores the calculation results of the business logic layer into the graph database.

The business logic layer includes:

A mode switching module, configured to switch between the view mode of the person in charge and the view mode of electric equipment according to the data request and the result of the second preprocessing, and the view mode of the person in charge is used to display each of the tasks The operations of personnel in the power accident event and the links between the operations, the power equipment view mode is used to display the problems generated by each of the power equipment in the power accident event and the problems and The link between the electrical accident events.

Specifically, the four modules of the business logic layer are independent of each other, and the person-in-charge view mode mainly shows the operation of each operator in the event and the relationship between them, so as to help the power grid company identify the person responsible for the event , the power equipment view mode mainly displays the problems of each power equipment in the event and the impact on the causality of the event, and this mode helps the business personnel of the power grid company to subsequently evaluate the quality of the power equipment. The design of this mode switching makes the view clearer, and the visual display is easier for the staff to understand. If various relationships are mixed in the same view, it will not have a good visual effect. Referring to the schematic diagram of the mode switching process shown in Figure 3, click the start button, the system starts to load the nodes and edges in the graph database of the knowledge map layer, and judges whether the current type is the power equipment view mode, and if it is the power equipment view mode, it will display One-way edge, if it is not in the power equipment view mode, judge whether you want to switch, if you don’t want to switch, directly jump to whether to display on the front end, if you want to switch, click switch, traverse all edges, and choose whether to Front-end display, if you choose to display at the front, then display at the front and end, if you do not choose to display at the front, hide at the front and end.

A node editing module, configured to perform node editing on the result of the second preprocessing according to the data request, and send the node editing result to the graph database through the data access layer for storage, and the node editing Including adding nodes, deleting nodes and updating node attributes.

Specifically, the node editing module mainly has three functions, that is, adding nodes, deleting nodes, and updating node attributes. A node represents an entity in an event (including people, events, and equipment, etc.). Since the retrospective analysis process of an event is a continuous update In the process, the information brought by the initial intelligence may not be very accurate, so it is necessary to modify the nodes and attributes. By adding nodes, deleting nodes, and updating nodes, the correctness of the knowledge map can be maintained, and the data can be continuously modified and updated. Maintaining the correctness and immediacy of the knowledge map also ensures that people can accurately find out the main responsible person or the main responsible power equipment problem for the electric accident case.

The functions of adding, deleting and updating nodes can be realized through the operation of Neo4j graph database. Refer to the flowchart of the node editing module shown in Figure 4, click the start button, enter the node name, and judge whether the node exists. If the node exists, update it. Node or delete a node, generate a delete update operation Cypher statement, and then perform corresponding operations in the Neo4j graph database, if the node does not exist, then add a node and generate a creation operation Cypher statement, and then perform corresponding operations in the Neo4j graph database, as described The corresponding operations include adding nodes, deleting nodes, and updating nodes. Finally, the front end returns the operation result and ends.

A relationship editing module, configured to perform relationship editing on the result of the second preprocessing according to the data request, and send the relationship editing result to the graph database through the data access layer for storage, and the relationship editing Including adding relationship, deleting relationship and updating relationship attributes.

Specifically, the relationship editing module mainly has three functions: adding relationships, deleting relationships, and updating relationship attributes. It is mainly used to edit the relationship between nodes. By using the neo4j graph database, you can directly edit the nodes in the graph database of the knowledge graph layer. The relationship is added, deleted, or the relationship attribute is updated to maintain the correctness and immediacy of the knowledge map. Referring to the flow diagram of the relationship editing module shown in Figure 5, click the start button, select the node, obtain the node and its relationship information, and create, delete, and update the relationship, wherein the created relationship is judged, if the created relationship If it is bidirectional, two edges will be created in the neo4j graph database for the selected node, and a Cypher statement for the corresponding operation will be generated to operate the neo4j graph database. If the created relationship is not bidirectional, it will be created in the neo4j graph database for the selected node One edge, generate the Cypher statement of the corresponding operation to operate the neo4j graph database; then, if there is an edge, the front end will display the directed edge and end, if the edge is measured, the front end will display the undirected edge and end, and the deleted relationship and the updated relationship will be checked The Cypher statement that completes the delete or update operation operates the neo4j graph database, and the front end displays the operation result and ends.

An auxiliary analysis module, configured to analyze the process of the power accident event according to the data request and the second preprocessing result, and obtain the influence of each node on the power accident event, wherein , the node with the greatest influence is the main responsible person or the main responsible power equipment of the power accident event.

Specifically, the auxiliary analysis module is mainly to help people analyze the entire event process of power accidents, and help people find links (nodes) with greater influence, so as to determine the main responsibility of the responsible person or equipment manufacturer. Business personnel can quickly search for the shortest path by selecting the input start node and end node, and taking the causality of adjacent nodes as the weight. At the same time, business personnel can also click the corresponding button to analyze the influence of the node, and visualize it according to the degree of influence. The node with greater influence has a larger visualization radius. The module uses the improved PageRank algorithm to calculate the size of influence, but if a node only has a relationship pointing to it, but no relationship pointed out from it, the PageRank algorithm is prone to precipitation. In order to avoid the precipitation of the PageRank value, in the original formula The decay factor is added for improvement, and the following formula is obtained, and the influence of any node on the power accident event is obtained by the following formula:

Among them, x represents any node, PR(x) represents the page ranking value of any node x, the larger the PR(x), the greater the influence of any node x, and β is between 0 and 1 The damping coefficient of , represents the random probability from any node x to the next node, Ai(i=1, 2, 3,...,n) represents the i-th node pointing to any node x, C(Ai) is the number of nodes pointing outward from node Ai, and E(x) is the decay factor, representing a certain vector corresponding to the set of nodes.

In order to enable the above algorithm to be applied visually, the system of this application will call the PageRank algorithm of the Neo4j graph database to return the PR values of different nodes, and perform differentiated presentations on the front end according to the PR values. When the user clicks the influence analysis button, the onPageRankClick event is triggered, a separate window pops up and the showPageRank function is called at the same time, the Cypher statement is run in the neo4j graph database, and then the PageRank algorithm is used for calculation, and the influence scores and values of all nodes are returned The larger the value, the greater the influence of the node, and then select the minimum and maximum values of the PR value of the node. If the two values are equal, all nodes are equal in size, indicating that the influence of all nodes is the same; if the two values are not equal, the difference between the PR values will be It is mapped to the interval with a radius of 20px to 60px, and the radius is set for each node according to the ratio. The greater the influence, the greater the node.

Among them, the key code for calling the PageRank algorithm is as follows:

String query="CALL algo.PageRank.stream(\n"+"\"MATCH(p)RETURN id(p)AS id\",\n"+"\"MATCH(p1)-[r]->( p2)RETURN id(p1)AS source,id(p2)AS target,r.weight as weight\",\n"+"{graph:\"Cypher\",weightProperty:\"weight\"})\n "+"YIELD node id,score\n"+

"RETURN algo.get Node By id(node id)AS node,score\n"+"ORDER BY scoreDESC\n"+"LIMIT10";

The interface interaction layer is also used to obtain the processing result of the business logic layer, and to visually display the processing result according to the data request, the processing result includes the switching result of the mode switching module, the The node editing result of the node editing module, the relationship editing result of the relationship editing module, or the analysis result of the auxiliary analysis module.

The present application discloses a system and method for visual analysis of power accidents based on knowledge graphs, including: a knowledge graph layer, a data access layer, a business logic layer, and an interface interaction layer that are sequentially connected to each other. First, the knowledge graph layer performs the first preprocessing on the text data of electric accident events, and builds a graph database according to the processing results. The interface interaction layer obtains the user's data request and sends it to the data access layer. Next, the data access layer performs second preprocessing on the graph database according to the data request, and feeds back the processing results to the business logic layer, which includes: a mode switching module, a node editing module, a relationship editing module and an auxiliary analysis module , which are respectively used for switching modes, editing nodes and relationships, and analyzing event processes according to user requests and the results of the second preprocessing, and determining the main responsible person or the main responsible electric equipment for the power accident event. Finally, the interface interaction layer visually displays the processing results of the business logic layer according to user requests.

Using the above-mentioned device for testing the bearing capacity of briquetting blocks in photovoltaic power plants solves the problem that it is difficult to help people accurately because it is not conducive to the analysis of the relevance of the cases and the excavation of the causal relationship to record the power accident cases in the traditional text way. Find out the main responsible person or the main responsible power equipment problem in the occurrence of electric accident cases, and help the power grid company to analyze the causes of power accident cases in the past and the relationship and influencing factors among the entities in the event. Responsibility for the event is identified, while alerting business personnel, reducing the occurrence of power production accidents, and optimizing the traditional visualization technology by introducing knowledge graph technology, making the visualized content easier for people to understand.

Further, when the interface interaction layer visually displays the analysis result of the auxiliary analysis module according to the data request, the node with greater influence has a larger visualization radius.

The following are method embodiments disclosed in this application, which are used to implement the above system embodiments. For details not disclosed in the method embodiments, please refer to the system embodiments.

The second embodiment of the present application discloses a knowledge graph-based power accident visualization analysis method, which is applied to the knowledge graph-based power accident analysis method disclosed in the first embodiment. In the accident visualization analysis system, see the work flow chart shown in Figure 2, including:

Step S1. Obtain the text data of the power accident event, perform first preprocessing on the text data, and construct a graph database according to the result of the first preprocessing, the first preprocessing includes data collection, knowledge extraction and Knowledge update, the graph database includes a knowledge graph, the knowledge graph includes nodes and relationships, the nodes include operators related to the power accident event, events related to the power accident event, and For event-related electrical equipment, the relationship is the relationship between any of the nodes and other nodes.

Further, the knowledge graph layer is constructed by means of crowdsourcing construction of the knowledge platform.

Specifically, step S1 is implemented in the knowledge graph layer.

Step S2, obtaining the user's data request, and sending the data request to the data access layer through the business logic layer.

Step S3. According to the data request, perform a second preprocessing on the nodes and the relationships in the graph database, and feed back the result of the second preprocessing to the business logic layer. The first The second preprocessing includes adding, deleting, modifying and querying.

Further, after step S3, it also includes:

The result of the second preprocessing is backed up to realize data sharing.

Specifically, the server backs up the result of the second preprocessing to realize data sharing.

Step S4. According to the data request and the result of the second preprocessing, switch between the view mode of the person in charge and the view mode of the electric equipment, and obtain the switching result. The view mode of the person in charge is used to display the The operations of the workers in the power accident event and the links between the operations, the power equipment view mode is used to display the problems of each of the power equipment in the power accident event and the The link between the problem and the electrical accident event in question.

Step S5, according to the data request, perform node editing on the result of the second preprocessing, and send the node editing result to the graph database of the knowledge map layer through the data access layer for storage, and the node Editing includes adding nodes, deleting nodes and updating node attributes.

Step S6. According to the data request, perform relational editing on the result of the second preprocessing, and send the relational editing result to the graph database of the knowledge map layer through the data access layer for storage. The above relation editing includes adding relation, deleting relation and updating relation property.

Step S7, according to the data request and the result of the second preprocessing, analyze the process of the electric accident event, and obtain the influence of each node on the electric accident event, the influence The largest node is the main responsible person or the main responsible power equipment of the power accident event.

Further, according to the data request, when the analysis result of the process of the electric accident event is visualized, the node with greater influence has a larger visualization radius.

Further, the influence of any node on the power accident event is obtained by the following formula:

Among them, x represents any node, PR(x) represents the page ranking value of any node x, the larger the PR(x), the greater the influence of any node x, and β is between 0 and 1 The damping coefficient of , represents the random probability from any node x to the next node, Ai(i=1, 2, 3,...,n) represents the i-th node pointing to any node x, C(Ai) is the number of nodes pointing outward from node Ai, and E(x) is the decay factor, representing a certain vector corresponding to the set of nodes.

Step S8, according to the data request, visually display the switching result, the node editing result, the relationship editing result or the analysis result of the electric accident event process through the interface interaction layer.

Further, the interface development of the interface interaction layer is realized by JavaScript programming language.

The present application discloses a system and method for visual analysis of power accidents based on knowledge graphs, including: a knowledge graph layer, a data access layer, a business logic layer, and an interface interaction layer that are sequentially connected to each other. First, the knowledge graph layer performs the first preprocessing on the text data of electric accident events, and builds a graph database according to the processing results. The interface interaction layer obtains the user's data request and sends it to the data access layer. Next, the data access layer performs second preprocessing on the graph database according to the data request, and feeds back the processing results to the business logic layer, which includes: a mode switching module, a node editing module, a relationship editing module and an auxiliary analysis module , which are respectively used for switching modes, editing nodes and relationships, and analyzing event processes according to user requests and the results of the second preprocessing, and determining the main responsible person or the main responsible electric equipment for the power accident event. Finally, the interface interaction layer visually displays the processing results of the business logic layer according to user requests.

Using the above-mentioned device for testing the bearing capacity of briquetting blocks in photovoltaic power plants solves the problem that it is difficult to help people accurately because it is not conducive to the analysis of the relevance of the cases and the excavation of the causal relationship to record the power accident cases in the traditional text way. Find out the main responsible person or the main responsible power equipment problem in the occurrence of electric accident cases, and help the power grid company to analyze the causes of power accident cases in the past and the relationship and influencing factors among the entities in the event. Responsibility for the event is identified, while alerting business personnel, reducing the occurrence of power production accidents, and optimizing the traditional visualization technology by introducing knowledge graph technology, making the visualized content easier for people to understand.

The present application has been described in detail above in conjunction with specific implementations and illustrative examples, but these descriptions should not be construed as limiting the present application. Those skilled in the art understand that without departing from the spirit and scope of the present application, various equivalent replacements, modifications or improvements can be made to the technical solutions and implementations of the present application, all of which fall within the scope of the present application. The scope of protection of the present application is subject to the appended claims.

Claims (9)

1. A visual analysis system for power accidents based on a knowledge map, characterized in that it includes: a knowledge map layer, a data access layer, a business logic layer and an interface interaction layer connected to each other in sequence; The knowledge map layer is used to obtain text data of electric accident events, perform first preprocessing on the text data, and construct a graph database according to the result of the first preprocessing, and the first preprocessing includes data Acquisition, knowledge extraction, and knowledge update, the graph database includes a knowledge map, and the knowledge map includes nodes and relationships, and the nodes include operators related to the power accident event, events related to the power accident event, and For electrical equipment related to the electrical accident event, the relationship is the relationship between any of the nodes and other nodes; The interface interaction layer is used to obtain a user's data request, and send the data request to the data access layer through the business logic layer; The data access layer is configured to perform a second preprocessing on the nodes and the relationships in the graph database according to the data request, and send the result of the second preprocessing to the business logic layer Feedback, the second preprocessing includes adding, deleting, modifying, querying; The business logic layer includes: A mode switching module, configured to switch between the view mode of the person in charge and the view mode of electric equipment according to the data request and the result of the second preprocessing, and the view mode of the person in charge is used to display each of the tasks The operations of personnel in the power accident event and the links between the operations, the power equipment view mode is used to display the problems generated by each of the power equipment in the power accident event and the problems and the link between the electrical accident events; A node editing module, configured to perform node editing on the result of the second preprocessing according to the data request, and send the node editing result to the graph database through the data access layer for storage, and the node editing Including adding nodes, deleting nodes and updating node attributes; A relationship editing module, configured to perform relationship editing on the result of the second preprocessing according to the data request, and send the relationship editing result to the graph database through the data access layer for storage, and the relationship editing Including adding relationship, deleting relationship and updating relationship attributes; An auxiliary analysis module, configured to analyze the process of the power accident event according to the data request and the second preprocessing result, and obtain the influence of each node on the power accident event, wherein , the node with the greatest influence is the main responsible person or the main responsible power equipment for the power accident event; The interface interaction layer is also used to obtain the processing result of the business logic layer, and to visually display the processing result according to the data request, the processing result includes the switching result of the mode switching module, the The node editing result of the node editing module, the relationship editing result of the relationship editing module, or the analysis result of the auxiliary analysis module. 2. A kind of power accident visual analysis system based on knowledge graph according to claim 1, is characterized in that, also comprises server; The server is connected to the data access layer, and is used for backing up the result of the second preprocessing to realize data sharing. 3. A knowledge map-based power accident visualization analysis system according to claim 2, wherein the interface interaction layer, according to the data request, when visually displaying the analysis results of the auxiliary analysis module , the node with greater influence has a larger visualization radius. 4. A method for visual analysis of power accidents based on knowledge graphs, characterized in that, the method for visual analysis of power accidents based on knowledge graphs is applied to any one of claims 1-3. In the power accident visualization analysis system, it includes: Acquiring the text data of the power accident event, performing a first preprocessing on the text data, and constructing a graph database according to the result of the first preprocessing, the first preprocessing includes data collection, knowledge extraction and knowledge update, The graph database includes a knowledge graph, the knowledge graph includes nodes and relationships, and the nodes include operators related to the power accident event, events related to the power accident event, and For electric equipment, the relationship is the relationship between any of the nodes and other nodes; Obtain the user's data request, and send the data request to the data access layer through the business logic layer; According to the data request, perform a second preprocessing on the nodes and the relationships in the graph database, and feed back the result of the second preprocessing to the business logic layer, the second preprocessing Including adding, deleting, modifying, querying; According to the data request and the result of the second preprocessing, switch between the person in charge view mode and the electric equipment view mode, and obtain the switching result, and the person in charge view mode is used to display each of the operators The operation in the electric accident event and the connection between the operations, the electric equipment view mode is used to display the problems of each electric equipment in the electric accident event and the relationship between the problems and the The connection between the above-mentioned electrical accident events; According to the data request, node editing is performed on the result of the second preprocessing, and the node editing result is sent to the graph database of the knowledge map layer through the data access layer for storage, and the node editing includes adding Node, delete node and update node properties; According to the data request, perform relation editing on the result of the second preprocessing, and send the relation editing result to the graph database of the knowledge map layer through the data access layer for storage, the relation editing Including adding relationship, deleting relationship and updating relationship attributes; According to the data request and the result of the second preprocessing, analyze the process of the power accident event, and obtain the influence of each node on the power accident event, the node with the greatest influence That is, the main responsible person or the main responsible electrical equipment for the said electrical accident; According to the data request, the switching result, the node editing result, the relationship editing result or the analysis result of the electric accident event process are visually displayed through the interface interaction layer. 5. A method for visual analysis of electric power accidents based on knowledge graph according to claim 4, characterized in that, the knowledge graph layer is constructed by means of crowdsourcing construction of a knowledge platform. 6. A method for visual analysis of power accidents based on knowledge graph according to claim 4, characterized in that, according to the data request, the nodes and the relationships in the graph database are performed for the second time Two preprocessing, and after feeding back the result of the second preprocessing to the business logic layer, it also includes: The result of the second preprocessing is backed up to realize data sharing. 7. A method for visual analysis of electric power accidents based on knowledge graph according to claim 4, characterized in that, the interface development of the interface interaction layer is realized by JavaScript programming language. 8. A method for visual analysis of power accidents based on knowledge graph according to claim 4, characterized in that, according to the data request, when the analysis results of the process of the power accident event are visualized, the more influential Larger nodes have a larger visualization radius. 9. A kind of knowledge map-based power accident visualization analysis method according to claim 4, characterized in that, The influence of any node on the power accident event is obtained by the following formula:

Among them, x represents any node, PR(x) represents the node influence value of any node x, the larger the PR(x), the greater the influence of any node x, and β is between 0 and 1 The damping coefficient of , represents the random probability from any node x to the next node, Ai represents the i-th node pointing to any node x, C(Ai) is the number of nodes pointing outward from node Ai, E(x) is Decay factor, which represents a certain vector corresponding to the set of nodes.

Images