Data maintenance system of converter station

Abstract

The invention provides a data maintenance system of a converter station, and relates to the technical field of data management, and the system comprises a document reading and preprocessing module which is used for reading a document in at least one format, and distinguishing and identifying a document structure and document information; the equipment tree construction module is used for reading a converter station design drawing, identifying and extracting equipment entity information, region interval information and a logic relationship between equipment and a region, and constructing an equipment tree of the converter station based on the identified and extracted information; the equipment parameter maintenance module is used for identifying a converter station container and an equipment entity by reading a converter station document, and associating attribute parameters of the equipment entity with a pre-constructed equipment model; and the UI interface display module is used for visually displaying the equipment tree and carrying out equipment parameter maintenance and manual confirmation through manual operation. According to the scheme, the work complexity and workload of front-line personnel can be reduced, and the operation and maintenance efficiency is further improved.

Description

Technical Field

[0001] The present invention relates to the technical field of data management, and in particular to a data maintenance system for a converter station. Background Art

[0002] DC transmission technology is widely used in projects such as new energy transmission. With the development of new energy, DC technology has been widely used. Moreover, with the advancement of the construction of the power Internet of Things and the development of status monitoring technology, converter stations have generated huge amounts of monitoring data.

[0003] At present, due to the large number of DC devices and the emergence of new technologies and new equipment, in the process of solidifying basic data, there are problems such as large manual maintenance workload and lack of data quality assurance. In particular, the product manuals, operation manuals, topology, operating procedures and other materials provided by many equipment manufacturers have not yet formed structured electronic documents, making it difficult to use and maintain these dormant data. In addition, in daily work, it is often necessary to query relevant equipment parameters, find component nameplates, and count the number of switch openings and closings. The manuals and operating procedures contain these basic data. Frontline personnel need to obtain data for operation and maintenance by flipping through drawings and checking documents, which is cumbersome and labor-intensive, greatly reducing the work efficiency of frontline personnel. Summary of the invention

[0004] In view of this, in order to address the above shortcomings, it is necessary to propose a data maintenance system for the converter station to reduce the tediousness and workload of front-line personnel, thereby improving operation and maintenance efficiency.

[0005] The present invention provides a data maintenance system for a converter station, the system comprising: a document reading and preprocessing module, a device tree building module, a device parameter maintenance module and a UI interface display module;

[0006] The document reading and preprocessing module is used to read documents in at least one format, and distinguish and identify document structures and document information;

[0007] The device tree construction module is used to read the converter station design drawings, identify and extract device entity information, area interval information and logical relationships between devices and areas, and construct a device tree for the converter station based on the identified and extracted information;

[0008] The device parameter maintenance module is used to identify the converter station container and the device entity by reading the converter station document, and associate the attribute parameters of the device entity with the pre-built device model;

[0009] The UI interface display module is used to visually display the device tree and to perform manual operation to maintain and confirm device parameters.

[0010] Preferably, the document reading and preprocessing module includes: a document reading unit, a distinguishing mark unit and a word segmentation processing unit;

[0011] The document reading unit supports reading converter station documents in at least PDF and word formats;

[0012] The distinguishing and identifying unit is used to distinguish the document structure and document information of the read converter station document, identify the information in different formats contained in the document, distinguish the title hierarchy and paragraphs, and remove redundant fields; wherein the formats of the data in the document include: text format, image format and table format;

[0013] The word segmentation processing unit is used to extract key information from the processed converter station document information according to the keywords in the preset corpus.

[0014] Preferably, the device tree construction module comprises: an electrical main wiring document reading unit, a main wiring area positioning unit, a spacing area identification unit, a device identification unit and a device tree construction unit;

[0015] The electrical main wiring document reading unit is used to read the electrical main wiring document of the converter station; wherein the electrical main wiring document includes CAD and PDF documents of the electrical main wiring;

[0016] The main wiring area positioning unit is used to determine whether there is an electrical main wiring diagram in the electrical main wiring document; and when there is an electrical main wiring diagram in the electrical main wiring document, position the read content to the electrical main wiring diagram area;

[0017] The interval area identification unit is used to identify the area identification, interval identification and text identification of the converter station from the electrical main wiring diagram area, and form a corresponding relationship table after matching the area identification, interval identification and text identification;

[0018] The device identification unit is used to identify the devices within each interval one by one, and use the line connection between two devices as a noise shield;

[0019] The device tree construction unit is used to construct a device tree structure with a nested relationship according to the logical relationship between the identified areas, intervals and devices.

[0020] Preferably, the interval region identification unit comprises: a region identification subunit, an interval identification subunit and a region interval table subunit;

[0021] The area recognition subunit is used to identify the virtual frame range in the electrical main wiring diagram area and the text marks and symbols at the corners of the virtual frame to obtain an area recognition result; wherein the intervals or areas in the converter station electrical main wiring document are delineated by a dotted box;

[0022] The interval recognition subunit is used to recognize the internally nested virtual box range and text logo when the virtual box is nested, and obtain the interval recognition result;

[0023] The area interval table subunit is used to fill the identified area recognition results and interval recognition results into a pre-built model table to obtain an area interval table.

[0024] Preferably, the region recognition result includes recognition results of multiple levels of regions.

[0025] Preferably, the device identification unit comprises: a first identification subunit and a second identification subunit;

[0026] The first identification subunit is used to determine the device type by using the device tag when there is a device tag in the graph, and extract the device tag as the device ID;

[0027] The second identification subunit is used for, when there is no device label in the graphic, reading the device graphic in the area based on the Tesseract OCR engine, and determining the device type by matching the device graphic stored in the device graphic library; wherein the device graphic library stores graphics and device information of various devices.

[0028] Preferably, the device parameter maintenance module includes: a title device tree association unit, a device attribute list construction unit, and a body document reading unit;

[0029] The title-device tree association unit is used to read the preprocessed document directory and titles at all levels, and form a document structure tree according to the titles at all levels; and match the read document structure tree with the constructed device tree to establish an association relationship between the title and the device tree;

[0030] The device attribute list construction unit is used to pre-construct each device attribute list under the device tree structure;

[0031] The body document reading unit is used to read each body paragraph information under the title one by one when the title matches the device tree.

[0032] Preferably, the body document reading unit comprises: a text reading subunit, a table reading subunit and a picture reading subunit;

[0033] The text reading subunit is used to read the text sentences one by one, split the text sentences into several short sentences according to punctuation marks, and remove redundant words as structured information; and segment the content of the short sentences, and extract the container class and device class data related to the title to identify the Chinese name data of the device class;

[0034] The table reading subunit is used to extract text information in the table by column, and identify the attribute field and attribute value of each column through the BERT semantic recognition function; and enter the identified attribute field into the corresponding attribute in the device attribute list;

[0035] The image reading subunit is used to read the text information in the image through the Tesseract OCR engine, and push it to manual confirmation, so that the device technical parameters can be entered after manual confirmation.

[0036] Preferably, the text reading subunit is also used to identify device-related technical parameters through the BERT semantic recognition function when the text contains device configuration and attribute information, and enter the technical parameters into the corresponding attributes in the device attribute list.

[0037] Preferably, the system further comprises: a manual correction module for manually correcting the equipment tree and the maintained equipment parameters.

[0038] It can be seen from the above technical scheme that the data maintenance system of the converter station provided by the present invention includes a document reading and preprocessing module, a device tree construction module, a device parameter maintenance module and a UI interface display module. The document reading and preprocessing module can read documents of at least one format, and distinguish and identify document structures and document information; the device tree construction module can read the converter station design drawings, identify and extract device entity information, regional interval information and the logical relationship between devices and regions, and build the device tree of the converter station based on this information; the device parameter maintenance module can identify the converter station container and device entity by reading the converter station document, and associate the attribute parameters of the device entity with the pre-built device model; the UI interface display module can visualize the device tree, and can perform manual operation to maintain and manually confirm the device parameters. It can be seen that this scheme can make full use of the existing converter station design drawings and other document achievements, realize the intelligent maintenance of converter station data, and do not require front-line personnel to flip through drawings and check documents to obtain data for operation and maintenance, which greatly reduces the workload and the tediousness of work, and improves the efficiency of front-line personnel in operation and maintenance. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 A schematic diagram of a data maintenance system for a converter station provided in an embodiment of the present invention.

[0040] Figure 2 A schematic diagram of a document reading and preprocessing module provided in an embodiment of the present invention.

[0041] Figure 3 A schematic diagram of a device tree construction module provided by an embodiment of the present invention.

[0042] Figure 4 A schematic diagram of a spacing region identification unit provided in an embodiment of the present invention.

[0043] Figure 5 A schematic diagram of a device identification unit provided in an embodiment of the present invention.

[0044] Figure 6 A schematic diagram of a device parameter maintenance module provided in an embodiment of the present invention.

[0045] Figure 7 A schematic diagram of a text document reading unit provided in an embodiment of the present invention.

[0046] In the figure: document reading and preprocessing module 10, document reading unit 11, distinguishing identification unit 12, word segmentation processing unit 13, device tree construction module 20, electrical main wiring document reading unit 21, main wiring area positioning unit 22, interval area identification unit 23, area identification subunit 231, interval identification subunit 232, area interval table subunit 233, device identification unit 24, first identification subunit 241, second identification subunit 242, device tree construction unit 25, device parameter maintenance module 30, title device tree association unit 31, device attribute list construction unit 32, main document reading unit 33, text reading subunit 331, table reading subunit 332, picture reading subunit 333, UI interface display module 40, manual correction module 50. DETAILED DESCRIPTION

[0047] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0048] like Figure 1-7 As shown, the present invention provides a data maintenance system for a converter station, the system comprising: a document reading and preprocessing module 10, a device tree building module 20, a device parameter maintenance module 30 and a UI interface display module 40;

[0049] The document reading and preprocessing module 10 is used to read documents in at least one format, and distinguish and identify document structures and document information;

[0050] The device tree construction module 20 is used to read the converter station design drawings, identify and extract device entity information, area interval information and logical relationship between devices and areas, and construct a device tree of the converter station based on the identified and extracted information;

[0051] The equipment parameter maintenance module 30 is used to identify the converter station container and the equipment entity by reading the converter station document, and associate the attribute parameters of the equipment entity with the pre-built equipment model;

[0052] The UI interface display module 40 is used for visually displaying the device tree and for manual operation to perform device parameter maintenance and manual confirmation.

[0053] The document reading and preprocessing module 10 may specifically include a document reading unit 11, a distinguishing mark unit 12 and a word segmentation processing unit 13;

[0054] The document reading unit 11 supports reading converter station documents in formats such as PDF and word;

[0055] A distinguishing and identifying unit 12 is used to distinguish the document structure and document information of the read converter station document, identify information of different formats contained in the document such as text, image and table, distinguish the title hierarchy and paragraphs, and remove redundant tag fields;

[0056] The word segmentation processing unit 13 is used to extract key information from the processed converter station document information according to the keywords in the preset corpus.

[0057] It should be understood that the required keywords are stored in the corpus, and when reading a file, the file is processed according to the keywords in the corpus to lay the foundation for subsequent computer vision and semantic recognition.

[0058] In this embodiment, the document reading and preprocessing module 10 cleans and preprocesses CAD, word, PDF and other files, including unified character encoding, eliminating redundant and repeated content, removing special characters and HTML tags, processing spelling errors, performing word segmentation, identifying and removing stop words, segmentation, sentence segmentation, and converting text to lowercase, etc. All these processes are to ensure that a clean, structured and consistent data foundation is provided for subsequent extraction work.

[0059] The device tree construction module 20 is used to read the converter station design drawings in PDF format and other formats, identify the bay area information and device entity information through computer vision technology, and extract the logical relationship between the bay and the device, such as aggregation and association, so as to form the converter station AC field and DC field device tree. Specifically, the device tree construction module 20 may include: an electrical main wiring document reading unit 21, a main wiring area positioning unit 22, a bay area identification unit 23, a device identification unit 24 and a device tree construction unit 25;

[0060] The electrical main wiring document reading unit 21 is used to read the electrical main wiring document of the converter station; wherein the electrical main wiring document includes the CAD and PDF documents of the electrical main wiring;

[0061] The main wiring area positioning unit 22 is used to determine whether there is an electrical main wiring diagram in the electrical main wiring document; and when there is an electrical main wiring diagram in the electrical main wiring document, locate the read content to the electrical main wiring diagram area;

[0062] The interval area identification unit 23 is used to identify the area identification, interval identification and text identification of the converter station from the electrical main wiring diagram area, and form a corresponding relationship table after matching the area identification, interval identification and text identification;

[0063] A device identification unit 24, for identifying the devices within each interval range one by one, and using the line connection between two devices as a noise shield;

[0064] The device tree construction unit 25 is used to construct a device tree structure with a nested relationship according to the logical relationship between the identified areas, intervals and devices.

[0065] The interval region identification unit 23 may further include: a region identification subunit 231, an interval identification subunit 232 and a region interval table subunit 233;

[0066] The area identification subunit 231 is used to identify the virtual frame range in the electrical main wiring diagram area and the text marks and symbols at the corners of the virtual frame to obtain the area identification result; wherein, the interval or area in the converter station electrical main wiring document is delineated by a dotted box;

[0067] The interval recognition subunit 232 is used to recognize the internally nested virtual box range and text mark when the virtual box has a nested relationship, and obtain the interval recognition result;

[0068] The area interval table subunit 233 is used to fill the identified area recognition results and interval recognition results into a pre-built model table to obtain an area interval table.

[0069] In this embodiment, when the interval area is identified, since the interval or area in the converter station design drawing is delineated by a dotted frame, the interval area range can be extracted by identifying the box. Therefore, it can be considered to use the Tesseract OCR engine to identify the virtual box range in the design drawing, as well as the specific text logos and symbols at the corners of the virtual box, and use it as the area logo of the box. If there is a nested relationship between the dotted boxes, a dotted box with a smaller range is further obtained, which is intended to be the interval area. When performing interval identification, first establish a table of the first-level area, second-level area and interval model of the converter station, including information such as identification, name, coordinates, etc. Then match the extracted virtual box text logo with the converter station area or interval logo, and identify the data of the first-level area, second-level area and interval respectively, and fill in the first-level area, second-level area and interval model. As shown in Table 1 below, the corresponding relationship between the DC field interval logo and the interval of the converter station.

[0070] Table 1 DC field area interval correspondence table

[0071]

[0072]

[0073] The device identification unit 24 is used to select an area within an interval virtual frame one by one, identify the primary devices within the interval one by one, and use the line connection between the two devices as a noise shield, that is, to distinguish the two primary devices and isolate the information to ensure the accuracy of the reading device. Further, when performing device identification, different methods can be used to perform device identification according to whether there is a device label in the graphic. Specifically, the device identification unit 24 may include: a first identification subunit 241 and a second identification subunit 242;

[0074] The first identification subunit 241 is used to determine the device type by using the device tag when there is a device tag in the graph, and extract the device tag as the device ID;

[0075] The second identification subunit 242 is used for, when there is no device label in the graphic, reading the device graphic in the area based on the Tesseract OCR engine, and determining the device type by matching the device graphic stored in the device graphic library; wherein the device graphic library stores graphics and device information of various devices.

[0076] In this embodiment, if there is a device label in the graph, the device label is directly used to determine the device type. The label of the device in the document of this area is extracted as the ID of the device, the device name is determined, and the device type is determined by the meaning of the English characters. If there is no device label, the device type is determined by computer vision technology, the Tesseract OCR engine is used to read the device graphics in the area, the line connection between the devices is shielded, the device primitive shape is extracted, and the device type is matched one by one with the device type in the device graphics library to determine the device type. And the devices in this area belong to this interval, and the association relationship between the device and the interval is established in the device model. The corresponding relationship between the device text symbol label and the device type is shown in Table 2 below.

[0077] Table 2 Correspondence between device type and device label

[0078] Serial number Device Type Equipment text symbol Remark 1 Y / Y commutation transformer Ty 2 Y / △ commutation transformer Td 3 lightning arrester F 4 Capacitors C 5 Voltage transformer U 6 Current transformer T 7 Flow control valve V 8 breaker Q Unit digits 9 Disconnector Q Double digits 10 Wall bushing X 11 Grounding switch Q Double digits 12 Reactor L 13 filter Z 14 AC filter bank ACF

[0079] The device tree construction unit 25 is used to construct a device tree structure containing a nested relationship of "primary area-secondary area-interval-device" according to the identified logical relationships among the primary area, secondary area, interval and device.

[0080] The converter station operation procedures and other PDF and word documents contain rich equipment information, including equipment records, parameters, configurations and other data. However, these data are stored as offline, unstructured data and are not converted into structured information. If manual maintenance is used, it will consume a lot of manpower and material resources, and the data quality is difficult to guarantee. Therefore, it is considered to perform data maintenance through the equipment parameter maintenance module 30. Specifically, the equipment parameter maintenance module 30 may include: a title equipment tree association unit 31, an equipment attribute list construction unit 32, and a text document reading unit 33;

[0081] The title device tree association unit 31 is used to read the preprocessed document directory and the titles at all levels, and form a document structure tree according to the titles at all levels; and match the read document structure tree with the constructed device tree to establish an association relationship between the title and the device tree;

[0082] The device attribute list building unit 32 is used to pre-build each device attribute list under the device tree structure;

[0083] The text document reading unit 33 is used to read each text paragraph information under the title one by one when the title matches the device tree.

[0084] Further, the body document reading unit 33 may include: a text reading subunit 331, a table reading subunit 332 and a picture reading subunit 333;

[0085] The text reading subunit 331 is used to read the text sentences one by one, split the text sentences into several short sentences according to punctuation marks, and remove redundant words as structured information; and to segment the contents of the short sentences, and extract the container class and device class data related to the title, and then identify the Chinese name data of the device class; the text reading subunit 331 is also used to identify the device-related technical parameters through the BERT semantic recognition function when the text contains device configuration and attribute information, and enter the technical parameters into the corresponding attributes in the device attribute list.

[0086] The table reading subunit 332 is used to extract text information in the table by column, and identify the attribute field and attribute value of each column through the BERT semantic recognition function; and enter the identified attribute field into the corresponding attribute in the device attribute list;

[0087] The image reading subunit 333 is used to read the text information in the image through the Tesseract OCR engine and push it to manual confirmation so that the equipment technical parameters can be entered after manual confirmation.

[0088] In this embodiment, when the device parameter maintenance module 30 extracts document information, it can be mainly implemented through the following process:

[0089] (1) After the PDF and word documents are preprocessed, the directory and titles at all levels of the document are read, a document structure tree is formed based on the titles at all levels, and the titles at all levels are matched with the device tree to establish an association between the titles and the device tree, that is, the read document structure tree is matched to the established device tree.

[0090] (2) Under the device tree structure, a list of device attributes is pre-built.

[0091] (3) If the title and the device tree completely match, directly enter the text document where this title is located; read the text paragraph information under this title one by one. If it is plain text, read the text sentences one by one, split them into several short sentences according to punctuation marks, remove unnecessary text regular expressions, and use several short sentences as structured information; use Chinese word segmentation technology to segment the content of the short sentences, extract container class and device class data related to the title, and use BERT to identify the Chinese name data of the device class after segmentation; if there are entities and relationships in the device component class data that are identified in the short sentences, then form a "device-relationship-component" triple, and store the device, component and relationship separately. In addition, if the text includes device configuration and attribute information, the device and related technical parameters are identified through the BERT semantic recognition function, and the technical parameters are automatically entered into the relevant attributes;

[0092] (4) If it is a table, first extract the text information in the table by column, identify the information category described in each column, that is, the attribute field and attribute value through the BERT semantic recognition function, match the attribute field with the attribute field in the model, and automatically enter the attribute value into the relevant attribute after the match is successful. If the match cannot be made, push it to manual confirmation and correction;

[0093] (5) If it is a picture, the text information in the image is read through the Tesseract OCR engine and pushed to manual confirmation. After confirmation, the equipment technical parameters are entered.

[0094] In addition, the data maintenance system of the converter station provided by the present solution also includes a manual correction module 50, which is used to manually correct the equipment tree and the equipment parameters of automatic maintenance. For example, when the correspondence with the model fails or there is data that does not correspond, manual correction is performed. At the same time, the manual correction module 50 can also realize functions such as information editing, structuring, importing and exporting.

[0095] In summary, in the data maintenance system of the converter station provided by this solution, the structural information processing of unstructured documents based on artificial intelligence is proposed, which makes full use of existing document achievements such as design drawings, user manuals, and operating procedures, and uses computer vision and natural language processing technology to semantically identify and classify the document content, convert its internal text, table and image data into structured information, automatically build the association between equipment and components, intelligently maintain equipment and attribute parameters, improve the maintenance efficiency of basic data of converter station equipment, and reduce the workload of data governance. At the same time, it provides convenience for data query and data mining of front-line personnel, thereby improving the operation and maintenance efficiency of front-line personnel.

[0096] The present specification also provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed in a computer, the computer is caused to execute a method in any one of the embodiments in the specification.

[0097] The present specification also provides a computing device, including a memory and a processor, wherein executable codes are stored in the memory, and when the processor executes the executable codes, a method in any embodiment of the present specification is implemented.

[0098] The modules or units in the device of the embodiment of the present invention can be combined, divided and deleted according to actual needs. The above disclosure is only the preferred embodiment of the present invention, and of course it cannot be used to limit the scope of the rights of the present invention. Those skilled in the art can understand that all or part of the processes of the above embodiment are implemented, and the equivalent changes made according to the claims of the present invention still fall within the scope of the invention.

Claims

1. A data maintenance system for a converter station, characterized in that: The system includes: document reading and preprocessing module, device tree building module, device parameter maintenance module and UI interface display module; The document reading and preprocessing module is used to read documents in at least one format, and distinguish and identify document structures and document information; The device tree construction module is used to read the converter station design drawings, identify and extract device entity information, area interval information and logical relationships between devices and areas, and construct a device tree for the converter station based on the identified and extracted information; The device parameter maintenance module is used to identify the converter station container and the device entity by reading the converter station document, and associate the attribute parameters of the device entity with the pre-built device model; The UI interface display module is used to visually display the device tree and to perform manual operation to maintain and confirm device parameters.

2. The data maintenance system for a converter station according to claim 1, characterized in that: The document reading and preprocessing module includes: a document reading unit, a distinguishing mark unit and a word segmentation processing unit; The document reading unit supports reading converter station documents in at least PDF and word formats; The distinguishing and identifying unit is used to distinguish the document structure and document information of the read converter station document, identify the information in different formats contained in the document, distinguish the title hierarchy and paragraphs, and remove redundant fields; wherein the formats of the data in the document include: text format, image format and table format; The word segmentation processing unit is used to extract key information from the processed converter station document information according to the keywords in the preset corpus.

3. The data maintenance system for a converter station according to claim 1, characterized in that: The device tree construction module includes: an electrical main wiring document reading unit, a main wiring area positioning unit, a spacing area identification unit, a device identification unit and a device tree construction unit; The electrical main wiring document reading unit is used to read the electrical main wiring document of the converter station; wherein the electrical main wiring document includes CAD and PDF documents of the electrical main wiring; The main wiring area positioning unit is used to determine whether there is an electrical main wiring diagram in the electrical main wiring document; and when there is an electrical main wiring diagram in the electrical main wiring document, position the read content to the electrical main wiring diagram area; The interval area identification unit is used to identify the area identification, interval identification and text identification of the converter station from the electrical main wiring diagram area, and form a corresponding relationship table after matching the area identification, interval identification and text identification; The device identification unit is used to identify the devices within each interval one by one, and use the line connection between two devices as a noise shield; The device tree construction unit is used to construct a device tree structure with a nested relationship according to the logical relationship between the identified areas, intervals and devices.

4. The data maintenance system for a converter station according to claim 3, characterized in that: The interval region identification unit comprises: a region identification subunit, an interval identification subunit and a region interval table subunit; The area recognition subunit is used to identify the virtual frame range in the electrical main wiring diagram area and the text marks and symbols at the corners of the virtual frame to obtain an area recognition result; wherein the intervals or areas in the converter station electrical main wiring document are delineated by a dotted box; The interval recognition subunit is used to recognize the internally nested virtual box range and text logo when the virtual box is nested, and obtain the interval recognition result; The area interval table subunit is used to fill the identified area recognition results and interval recognition results into a pre-built model table to obtain an area interval table.

5. The data maintenance system for a converter station according to claim 4, characterized in that: The region recognition result includes recognition results of multiple levels of regions.

6. The data maintenance system for a converter station according to claim 3, characterized in that: The device identification unit comprises: a first identification subunit and a second identification subunit; The first identification subunit is used to determine the device type by using the device tag when there is a device tag in the graph, and extract the device tag as the device ID; The second identification subunit is used for, when there is no device label in the graphic, reading the device graphic in the area based on the Tesseract OCR engine, and determining the device type by matching the device graphic stored in the device graphic library; wherein the device graphic library stores graphics and device information of various devices.

7. The data maintenance system for a converter station according to claim 1, characterized in that: The device parameter maintenance module includes: a title device tree association unit, a device attribute list construction unit, and a text document reading unit; The title-device tree association unit is used to read the preprocessed document directory and titles at all levels, and form a document structure tree according to the titles at all levels; and match the read document structure tree with the constructed device tree to establish an association relationship between the title and the device tree; The device attribute list construction unit is used to pre-construct each device attribute list under the device tree structure; The body document reading unit is used to read each body paragraph information under the title one by one when the title matches the device tree.

8. The data maintenance system for a converter station according to claim 7, characterized in that: The body document reading unit includes: a text reading subunit, a table reading subunit and a picture reading subunit; The text reading subunit is used to read the text sentences one by one, split the text sentences into several short sentences according to punctuation marks, and remove redundant words as structured information; and segment the content of the short sentences, and extract the container class and device class data related to the title to identify the Chinese name data of the device class; The table reading subunit is used to extract text information in the table by column, and identify the attribute field and attribute value of each column through the BERT semantic recognition function; and enter the identified attribute field into the corresponding attribute in the device attribute list; The image reading subunit is used to read the text information in the image through the Tesseract OCR engine, and push it to manual confirmation, so that the device technical parameters can be entered after manual confirmation.

9. The data maintenance system for a converter station according to claim 8, characterized in that: The text reading subunit is also used to identify the device-related technical parameters through the BERT semantic recognition function when the text contains device configuration and attribute information, and enter the technical parameters into the corresponding attributes in the device attribute list.

10. The data maintenance system for a converter station according to any one of claims 1 to 9, characterized in that: The system also includes: a manual correction module, which is used to manually correct the equipment tree and the maintained equipment parameters.

Images