A control ontology labeling system based on control cloud micro-scene

Abstract

The present invention discloses a control ontology label system based on a control cloud micro-scene, aiming to solve the technical problems of low efficiency and insufficient accuracy in the retrieval of control ontology label data of the power grid system in the prior art. It includes a data parsing module, a micro-scene construction module, a label query module, a label display module, a label construction module, and a control ontology label library, wherein the micro-scene construction module is used to call micro-scene slices according to the data field or domain vocabulary output by the data parsing module to generate a control cloud micro-scene; the label construction module is used to organize and analyze the control ontology data in the power grid, mine and learn, and generate control ontology labels; the label query module is used to query the control ontology label corresponding to the keywords or questions to be retrieved from the control ontology label library under the drive of the control cloud micro-scene. The present invention can query label data quickly, accurately and efficiently, meet the needs of control business, and improve the level of power grid management.

Description

Technical Field

[0001] The invention relates to a control ontology label system based on a control cloud micro-scene, belonging to the technical field of electric power automation. Background Art

[0002] With the continuous deepening of my country's ultra-high voltage interconnected power grid construction, the scale of the power grid has developed rapidly, the power grid structure has become increasingly complex, the national grid has basically been formed, and the construction of the control cloud has been completed, which has put forward higher requirements for the integrated operation level of the large power grid. It is necessary to further strengthen the support capabilities of the dispatching and control system, continuously improve the level of instant sharing of control information, improve data processing, application computing and service capabilities, realize on-demand access to application services in a wide area, and at the same time build a control entity label system to further improve the level of power grid management.

[0003] When building a control ontology labeling system, it is necessary to obtain and process various specific data. A large amount of model, operation, and management data is collected in control production. At present, there is a lack of unified data asset management methods, and there are problems such as being unable to find, inaccurate, and low retrieval efficiency; traditional data acquisition and use methods are single, and there are certain difficulties in quickly locating, acquiring, and deeply mining data, and it is necessary to improve data retrieval capabilities and value mining capabilities; with the continuous development of control business needs, the demand for application system response speed is also accelerating, and it is also necessary to combine the dispatching business scenario to integrate all-round data to assist business personnel in decision-making. Summary of the invention

[0004] In order to solve the problems of low efficiency and insufficient accuracy in retrieval of control ontology tag data of power grid systems in the prior art, the present invention proposes a control ontology tag system based on a control cloud micro-scene, which uniformly manages grid control data through tags, and uses the control cloud micro-scene to quickly, accurately and efficiently query tag data, meet the control business needs, and improve the level of grid management.

[0005] In order to solve the above technical problems, the present invention adopts the following technical means:

[0006] The present invention proposes a control ontology labeling system based on a control cloud micro-scene, comprising:

[0007] A data parsing module is used to parse the keywords or questions entered into the system into data fields or domain words through a search engine;

[0008] The micro-scenario construction module is used to call the micro-scenario slices according to the data fields or domain vocabulary output by the data parsing module to generate the control cloud micro-scenario;

[0009] The label construction module is used to organize, analyze, mine and learn the control ontology data in the power grid and generate control ontology labels;

[0010] The regulatory ontology tag library is used to store the regulatory ontology tags generated by the tag building module;

[0011] A tag query module is used to query the control ontology tag corresponding to the keyword or question to be retrieved from the control ontology tag library under the drive of the control cloud micro-scenario;

[0012] The label display module is used to display the regulation ontology labels queried by the data query module.

[0013] Furthermore, the construction method of micro scene slices is as follows:

[0014] Slice and classify the power grid scenarios according to the power grid level, research time, statistical period, management responsibility, equipment status, system status and business direction, and obtain multiple micro-scenario slices under each slice category;

[0015] Perform statistical analysis on micro-scenario slices based on historical power grid scenarios to obtain the correlation or mapping relationship between different micro-scenario slices;

[0016] A slice value of each micro-scene slice is generated based on the association degree or mapping relationship between different micro-scene slices.

[0017] Furthermore, the regulatory ontology labels are divided into four dimensions, namely attributes, events, indicators and inferences.

[0018] Furthermore, the method for generating labels of four dimensions is:

[0019] According to the steady-state or static data in the control cloud platform, the attribute label of the control entity is obtained;

[0020] Convert the power grid operation events corresponding to the control entity into SQL statements to obtain the event labels of the control entity;

[0021] Use big data mining and machine learning algorithms to process the historical operation data of the control entity to obtain the future state of the control entity, which is recorded as the inferred label of the control entity;

[0022] According to the grid operation index assessment rules, the real-time operation data of the control entity is analyzed to obtain the index label of the control entity.

[0023] Furthermore, the storage form of the control ontology tag library is: a two-dimensional table is generated using the object encoding of the control ontology and the tag value of the control ontology tag, and driven by the needs of the power grid micro-scene, different tag values ​​are dynamically generated for different control cloud micro-scenes.

[0024] Furthermore, the system also includes a label updating module, which is used to update the label value of the regulation ontology label according to the retrieval accuracy entered into the system by the user.

[0025] The following advantages can be obtained by using the above technical means:

[0026] The present invention proposes a control ontology labeling system based on the control cloud micro-scene, which converts the keywords or questions to be detected into the control cloud micro-scene. The control cloud micro-scene has richer scene information and is more easily recognized by the system language, which can improve the efficiency of label data retrieval. In the process of micro-scene construction, the present invention quickly locates and calls more micro-scene slices through slice values, which can ensure that the generated control cloud micro-scene is more accurate and complete, which is conducive to subsequent label retrieval. The present invention generates control ontology labels from four dimensions, which can not only make the most convenient and fast classification of power grid data, but also ensure the integrity and pertinence of power grid data labeling, which is conducive to quickly locating label data through micro-scenes.

[0027] The system of the present invention can collect a large amount of data in the regulation and control production, and perform unified data management through tags, and use the regulation and control cloud micro-scene to perform fast, accurate and efficient tag data query to meet the regulation and control business needs and improve the level of power grid management. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] Figure 1 It is a structural schematic diagram of a control ontology label system based on a control cloud micro-scene of the present invention;

[0029] Figure 2 It is a schematic diagram of the operation of the system of the present invention;

[0030] Figure 3 A schematic diagram of regulating cloud micro-scene construction and micro-scene slice classification in an embodiment of the present invention;

[0031] Figure 4 Schematic diagram of the construction process of the regulation ontology tag library in an embodiment of the present invention. DETAILED DESCRIPTION

[0032] The technical solution of the present invention is further described below in conjunction with the accompanying drawings:

[0033] The present invention proposes a control ontology labeling system based on the control cloud micro-scene, such as Figure 1 , 2 As shown, it mainly includes data parsing module, micro-scene construction module, label query module, label display module, label construction module, regulation ontology label library and label update module.

[0034] The data analysis module can obtain the keywords or questions to be searched that are input by the user into the system, perform semantic understanding and vocabulary analysis on the keywords or questions to be searched through the search engine, and parse the keywords or questions to be searched into one or more data fields or domain words, such as main network, steady state, primary equipment, etc.

[0035] The micro-scenario construction module can call micro-scenario slices according to the data fields or domain vocabulary output by the data parsing module to generate a control cloud micro-scenario that is adapted to the query content.

[0036] The present invention establishes micro-scenario slices of the power grid for various attributes, states and directions that need to be considered in the operation of power grid regulation, and then reassembles micro-scenario instances according to the events and instance data that actually occur in the power grid, thereby improving the granularity and flexibility of work handling in power grid regulation; Figure 3 As shown, the construction method of micro scene slices is as follows:

[0037] (1) Slice and classify the grid scenarios according to the grid level, research time, statistical period, management responsibility, equipment status, system status and business direction, and obtain multiple micro-scenario slices under each slice category; each slice category contains multiple specific micro-scenario slices. For example, the slice category grid level (Level) contains three micro-scenario slices, namely the main grid slice, distribution grid slice and microgrid slice.

[0038] (2) Perform statistical analysis on micro-scenario slices based on historical power grid scenarios to obtain the correlation or mapping relationship between different micro-scenario slices.

[0039] (3) Generate a slice value for each micro-scene slice based on the correlation or mapping relationship between different micro-scene slices.

[0040] The micro-scene slice combination layer can associate the parsed data fields or domain vocabulary with the divided micro-scene slices, and quickly locate and call more micro-scene slices through slice values, providing more data options for related micro-scenes, making the final generated control cloud micro-scene more accurate and complete.

[0041] According to the query content input by the user into the system, various types of control cloud micro-scenes are obtained through different combinations of micro-scene slices, so as to achieve the minimum granularity combination for dynamic scenes such as different business directions, management responsibilities, and equipment status.

[0042] In an embodiment of the present invention, when performing a steady-state analysis scenario based on the normal operation of equipment in a monthly period of a provincial main network, the following control cloud micro-scenario definition can be given through micro-scenario slicing:

[0043] Scene analyse ≡Level a →Duty c →Duration c →EState a →SSate a (1)

[0044] Among them, Sceneanalyse Represents the control cloud micro scene, according to Figure 2 , Level a Indicates that the power grid level is the main grid, Duty c Indicates that the management responsibility is provincial, Duration c Indicates that the statistical period is monthly operation statistics, EState a Indicates that the device is in normal operation. a Indicates that the system state is steady state.

[0045] Compared with the keywords or questions to be searched entered by the user, the control cloud micro-scene has richer scene information, which can be recognized by the system language, guide the subsequent retrieval and construction of power grid control ontology tags, and improve the efficiency of tag data retrieval.

[0046] The label construction module is used to generate specific regulatory ontology labels based on regulatory ontology related data, such as Figure 4 As shown, a large amount of raw data in the regulation and control production is obtained through operations such as data extraction, data conversion, data integration, data interface, and data encapsulation. These regulation and control data are extracted, sorted, analyzed, mined, and learned to define different label values.

[0047] In the present invention, the control entity business mainly includes the control of primary equipment, secondary equipment, automation equipment, material equipment, network structure, operation mode, control plan, maintenance plan, etc. The relevant data of the control entity mainly consists of relevant databases and data files.

[0048] The labels for the regulation ontology are based on the two-level labeling system of the headquarters and provincial (municipal) companies, and are built on a two-level data platform. They are an accurate evaluation method for the ontology status of the regulation ontology throughout its life cycle. Taking into account the business facts of the regulation ontology and the construction specifications of the regulation cloud, four labeling dimensions are constructed, namely attributes, events, indicators and inferences.

[0049] The attribute label is consistent with the model data type in the control cloud platform. It relies on the database, that is, it has fields and mainly stores steady-state or static data, such as voltage level, capacity, etc. The event label is consistent with the power grid planning data, operation data, maintenance data, etc. in the control cloud platform. It is mainly power grid operation data. The power grid operation events can be sorted according to the time dimension, and then the power grid operation events corresponding to the control entity are converted into SQL statements to obtain the event labels of the control entity, such as switch status, normal operation, etc. The inference label is consistent with the real-time cloud data in the control cloud platform. The encapsulated power grid historical operation data is iteratively learned through big data mining algorithms and machine learning technologies to obtain the future state data of the control entity to assist in future trend judgment. The future state data of the control entity is the inference label. Among them, commonly used algorithms include SVM, GBDT, KNN, KMEANS, etc. The indicator label mainly depends on the normalized operation indicator assessment rules of the power grid. The real-time operation data is compared, analyzed, and calculated with the indicator data to build a hierarchical evaluation tree. The corresponding evaluation results are generated for each control entity, which are recorded as indicator labels.

[0050] The system of the present invention generates control ontology labels from four dimensions, which can not only make the most convenient and rapid classification of power grid data, but also ensure the integrity and pertinence of power grid data labeling, which is conducive to quickly locating label data through micro-scenes.

[0051] The present invention can construct a power grid control ontology label system through four major capability layers: label sharing, label security, label service, and label statistics on the basis of the control ontology label library.

[0052] The control ontology tag library is used to store the control ontology tags generated by the tag construction module. The specific storage form is: a two-dimensional table is generated using the object coding of the control ontology and the tag value of the control ontology tag, and driven by the needs of the power grid micro-scene, tag values ​​that fit the scene characteristics are dynamically generated for different control cloud micro-scenes.

[0053] The tag query module is used to query the control ontology tag corresponding to the keyword or question to be retrieved from the control ontology tag library under the control cloud micro-scene drive, search for the tag value in the control ontology tag library through the control cloud micro-scene drive, and find the corresponding control ontology object code and control ontology tag.

[0054] The label display module is used to display the control ontology label data queried by the label query module.

[0055] After the label construction module completes the label construction, the label update module can also use the test data to evaluate the control ontology label accordingly, and decide whether to use the control ontology label based on the evaluation results. Unused labels will not be stored in the label library and need to be rebuilt. In addition, the label update module is also used to update the label value of the control ontology label according to the retrieval accuracy entered by the user into the system. After the label display module displays the label data, the user can provide feedback on the retrieval accuracy on the system page. The label update module adjusts the label content according to the feedback mechanism of the label system to improve the subsequent retrieval accuracy and complete the closed loop of the entire label library construction.

[0056] The system of the present invention can collect a large amount of data in the regulation and control production, and perform unified data management through tags, and use the regulation and control cloud micro-scene to perform fast, accurate and efficient tag data query to meet the regulation and control business needs and improve the level of power grid management.

[0057] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the technical principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A control ontology labeling system based on a control cloud micro-scene, characterized in that: include: A data parsing module is used to parse the keywords or questions entered into the system into data fields or domain words through a search engine; The micro-scenario construction module is used to call the micro-scenario slices according to the data fields or domain vocabulary output by the data parsing module to generate the control cloud micro-scenario; The label construction module is used to organize, analyze, mine and learn the control ontology data in the power grid and generate control ontology labels; The regulatory ontology tag library is used to store the regulatory ontology tags generated by the tag building module; A tag query module is used to query the control ontology tag corresponding to the keyword or question to be retrieved from the control ontology tag library under the drive of the control cloud micro-scenario; The label display module is used to display the control ontology labels queried by the data query module; The construction method of micro scene slice is as follows: Slice and classify the power grid scenarios according to the power grid level, research time, statistical period, management responsibility, equipment status, system status and business direction, and obtain multiple micro-scenario slices under each slice category; Perform statistical analysis on micro-scenario slices based on historical power grid scenarios to obtain the correlation or mapping relationship between different micro-scenario slices; A slice value of each micro-scene slice is generated based on the association degree or mapping relationship between different micro-scene slices.

2. According to the control ontology labeling system based on the control cloud micro-scene according to claim 1, it is characterized in that: The regulatory ontology labels are divided into four dimensions: attributes, events, indicators, and inferences.

3. According to claim 2, a control ontology labeling system based on a control cloud micro-scene is characterized in that: The method for generating labels of 4 dimensions is: According to the steady-state or static data in the control cloud platform, the attribute label of the control entity is obtained; Convert the power grid operation events corresponding to the control entity into SQL statements to obtain the event labels of the control entity; Use big data mining and machine learning algorithms to process the historical operation data of the control entity to obtain the future state of the control entity, which is recorded as the inferred label of the control entity; According to the grid operation index assessment rules, the real-time operation data of the control entity is analyzed to obtain the index label of the control entity.

4. According to claim 1, a control ontology labeling system based on a control cloud micro-scene is characterized in that: The storage form of the control ontology tag library is: a two-dimensional table is generated using the object encoding of the control ontology and the tag value of the control ontology tag, and driven by the needs of the power grid micro-scene, different tag values ​​are dynamically generated for different control cloud micro-scenes.

5. According to claim 1, a control ontology labeling system based on a control cloud micro-scene is characterized in that: The system further comprises a label updating module, which is used to update the label value of the regulation ontology label according to the retrieval accuracy entered into the system by the user.

Images