Relay protection configuration information intelligent identification system and method based on natural language technology

Abstract

The invention discloses a relay protection configuration information intelligent identification system and method based on a natural language technology. The system comprises a data acquisition module, a processing unit, a feature extraction module, a configuration information matching unit and a result output module. The data acquisition module is responsible for receiving and analyzing text information sent by a relay protection device from a power system; the processing unit cleans, standardizes and formats the received text information so as to improve the accuracy of subsequent analysis; the feature extraction module extracts key features from the preprocessed text by using a natural language processing technology; the configuration information matching unit compares the extracted features with a predefined relay protection configuration information base so as to identify specific configuration information corresponding to the text information; the result output module matches the result according to the configuration information; the text information sent by the relay protection device is quickly and accurately analyzed and processed, the working efficiency and accuracy of operation and maintenance personnel are improved, and safe and stable operation of a power system is guaranteed.

Description

Technical Field

[0001] The present invention belongs to the field of intelligent identification of relay protection, and more specifically, it particularly relates to an intelligent identification system of relay protection configuration information based on natural language technology. Meanwhile, the present invention also relates to an intelligent identification method of relay protection configuration information based on natural language technology. Background Art

[0002] In the power system, relay protection devices are important equipment to ensure the safe and stable operation of the power grid. They monitor the operating status of the power system and take corresponding protection measures when faults or abnormalities are detected to avoid the spread of faults and damage to equipment. However, with the continuous expansion and increase in complexity of power systems, the amount of text information generated by relay protection devices has also increased dramatically, bringing huge challenges to operation and maintenance personnel.

[0003] The traditional way of processing relay protection information mainly relies on manual interpretation and analysis, which has many shortcomings. First, manual processing of large amounts of complex text information is inefficient and difficult to meet real-time requirements. In an emergency, every second may affect the safety and stability of the power grid, and manual analysis takes a lot of time and cannot achieve immediate response. Secondly, the influence of human factors in the information processing process cannot be ignored. Different operation and maintenance personnel may have deviations in their interpretation of the same information due to differences in experience, knowledge level and working status, which may lead to misjudgment or omission, and thus affect the accuracy of decision-making. Furthermore, in the face of massive amounts of information, manual screening and classification are very difficult, and it is easy to miss important information. Moreover, with the continuous growth of the amount of information, this method of relying on manpower has become increasingly unrealistic. In addition, the manual processing of information lacks systematicity and coherence, which is not conducive to the mining and utilization of historical data, and it is difficult to discover potential laws and trends from it to provide a basis for preventive maintenance; therefore, we propose a system and method for intelligent identification of relay protection configuration information based on natural language technology. Summary of the invention

[0004] The purpose of the present invention is to solve the shortcomings existing in the prior art, and to propose a relay protection configuration information intelligent identification system and method based on natural language technology, including a data acquisition module, a processing unit, a feature extraction module, a configuration information matching unit and a result output module; to achieve fast and accurate parsing and processing of text information issued by the relay protection device, improve the work efficiency and accuracy of operation and maintenance personnel, and ensure the safe and stable operation of the power system.

[0005] To achieve the above object, the present invention provides the following technical solutions:

[0006] The intelligent identification system of relay protection configuration information based on natural language technology includes:

[0007] A data acquisition module, used to receive and parse text information sent from the relay protection device of the power system;

[0008] A processing unit cleans, standardizes and formats the received text information to improve the accuracy of subsequent analysis;

[0009] The feature extraction module uses natural language processing technology to extract key features from the preprocessed text, including entity recognition, relationship extraction, and tendency analysis;

[0010] The configuration information matching unit compares the extracted features with the predefined relay protection configuration information library to identify the specific configuration information corresponding to the text information;

[0011] The result output module generates easy-to-understand reports or notifications based on the configuration information matching results to guide operation and maintenance personnel to perform corresponding operations.

[0012] The intelligent identification method of relay protection configuration based on natural language technology is applied to the above system and includes the following steps:

[0013] S1. Data acquisition: receiving and analyzing data information sent by relay protection devices of the power system;

[0014] S2, data processing, cleaning, standardizing and formatting the received text information to improve the accuracy of subsequent analysis;

[0015] S3, feature extraction, using natural language processing technology to extract key features from the preprocessed text, including entity recognition, relationship extraction and tendency analysis;

[0016] S4, configuration information matching, comparing the extracted features with the predefined relay protection configuration information library to identify the specific configuration information corresponding to the text information;

[0017] S5. Result output: Generate an easy-to-understand report or notification based on the configuration information matching results to guide the operation and maintenance personnel to perform corresponding operations.

[0018] Preferably, when processing the data in step S2, it also includes using an algorithm to detect whether there are abnormal values ​​in the processed data, and making adjustments and corrections in time. The expression is:

[0019] ;

[0020] In the formula, X is the observed value, μ is the mean, and σ is the standard deviation. ≥1.5, considered as an outlier

[0021] Preferably, step S3 specifically comprises:

[0022] S31. According to the requirements of the relay protection information model, define the entity types that need to be identified, including device name, fault type, and time;

[0023] S32. Determine the types of relationships that exist in the text, including causal relationships, inclusion relationships, and contrast relationships;

[0024] S33. Construct a dictionary containing positive and negative words to judge the tendency of the text, analyze each sentence or paragraph in the text, calculate its score, and judge whether its tendency is positive or negative;

[0025] S34. Based on the results of entity recognition, relationship extraction and tendency analysis, select key features related to the relay protection information model, and use the cross-validation method to evaluate the effect of feature extraction to ensure that the extracted features can effectively support configuration information matching.

[0026] Preferably, in step S31, the word sequence obtained after the text T is processed by word segmentation is W={ 1, 2, ..., n},in i represents the i-th word;

[0027] Define entity type set E = {equipment name, fault type, time};

[0028] For each word i , use named entity recognition technology to determine whether it belongs to a type in the entity type set E. If so, mark it as the corresponding entity type;

[0029] The set of entities extracted from the text T can be expressed as:

[0030] Entity collection = {(e j , type j)|e j ∈T, type j∈E};

[0031] In the formula, e j represents the jth entity extracted, and type type j represents the type to which the entity belongs;

[0032] In step S32, the relationship type existing in the text is determined, and the expression is:

[0033] R={(E i , E j )|E i , E j ∈ entity set, (E i , Ej )∈relationtype};

[0034] In the formula, R represents the set of relations in the text, E i and E j is the identified entity;

[0035] In step S33, the propensity score of the text is calculated as follows:

[0036] Propensity score = ;

[0037] In the formula, if the tendency score is greater than 0, the text tendency is positive; if the tendency score is less than 0, the text tendency is negative.

[0038] Preferably, step S4 is specifically:

[0039] S41, initializing the configuration information matching module, loading parameters and resources, and calling the feature extraction module to extract key features from the text;

[0040] S42, loading a predefined relay protection configuration information library, and for each model in the model library, calculating a similarity score between the model and the extracted feature;

[0041] S43. For each model in the model library, calculate its similarity score with the extracted features, verify whether the selected model is accurate, and output the final result.

[0042] Preferably, the similarity score between the extracted feature and the extracted feature is calculated as follows:

[0043] ;

[0044] In the formula, ‖A‖ represents the Euclidean norm of vector A, ‖B‖ represents the Euclidean norm of vector B, and the calculation formula is:

[0045] ;

[0046] ;

[0047] A and B are two n-dimensional feature vectors, where A=(a1, a2, ..., a n ), B=(b1, b2, ..., b n ).

[0048] Preferably, step S5 is specifically:

[0049] S51. Generate an easy-to-understand report or notification content based on the parsed matching results, including the model name, key features, possible problem description, and recommended operation steps;

[0050] S52. Send the report or notification to the relevant operation and maintenance personnel, ensure that the sent report or notification can reach the operation and maintenance personnel in time, and remind them to check and read it;

[0051] S53. Record the detailed information of the reports or notifications sent, and track the feedback and operations of the operation and maintenance personnel, so as to timely understand the progress and effect of problem solving.

[0052] Technical effects and advantages of the present invention: Compared with the prior art, the system and method for intelligent identification of relay protection configuration information based on natural language technology provided by the present invention have the following advantages:

[0053] Through the automated data acquisition, analysis and processing process, the need for manual intervention is reduced, and the efficiency of information processing is greatly improved. The feature extraction module uses natural language processing technology to quickly and accurately extract key features from text, providing strong support for subsequent analysis;

[0054] Automatically matching the predefined relay protection configuration information library and identifying the specific configuration information corresponding to the text information helps to quickly locate the cause of the fault. By providing possible problem descriptions and recommended operating procedures, the system provides valuable reference information for operation and maintenance personnel, helping them to solve problems more quickly. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] Figure 1 This is a block diagram of the intelligent identification system for relay protection configuration information based on natural language technology of the present invention;

[0056] Figure 2 This is a flow chart of the intelligent identification method for relay protection configuration based on natural language technology of the present invention. DETAILED DESCRIPTION

[0057] In order to make the purpose, technical scheme and advantages of the present invention clearer, the present invention is further described in detail below in conjunction with specific embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0058] The present invention provides a system and method for intelligent identification of relay protection configuration information based on natural language technology, such as Figure 1 As shown, the system includes: a data acquisition module for receiving and parsing text information sent from a relay protection device of a power system;

[0059] A processing unit cleans, standardizes and formats the received text information to improve the accuracy of subsequent analysis;

[0060] The feature extraction module uses natural language processing technology to extract key features from the preprocessed text, including entity recognition, relationship extraction, and tendency analysis;

[0061] The configuration information matching unit compares the extracted features with the predefined relay protection configuration information library to identify the specific configuration information corresponding to the text information;

[0062] The result output module generates easy-to-understand reports or notifications based on the configuration information matching results to guide operation and maintenance personnel to perform corresponding operations;

[0063] Among them, the data acquisition module is responsible for receiving and parsing the text information sent by the relay protection device of the power system; the processing unit cleans, standardizes and formats the received text information to improve the accuracy of subsequent analysis; the feature extraction module uses natural language processing technology to extract key features from the preprocessed text; the configuration information matching unit compares the extracted features with the predefined relay protection configuration information library to identify the specific configuration information corresponding to the text information; the result output module is based on the result of the configuration information matching; to achieve fast and accurate parsing and processing of the text information sent by the relay protection device, improve the work efficiency and accuracy of the operation and maintenance personnel, and ensure the safe and stable operation of the power system.

[0064] This embodiment also proposes a method for intelligent identification of relay protection configuration based on natural language technology, which is applied to the above-mentioned system, such as Figure 2 As shown, the following steps are included:

[0065] S1. Data acquisition: receiving and analyzing data information sent by relay protection devices of the power system;

[0066] S2, data processing, cleaning, standardizing and formatting the received text information to improve the accuracy of subsequent analysis; when processing the data in step S2, it also includes using an algorithm to detect whether there are abnormal values ​​in the processed data, and making timely adjustments and corrections. The expression is:

[0067] ;

[0068] In the formula, X is the observed value, μ is the mean, and σ is the standard deviation. ≥1.5, considered as outliers.

[0069] S3, feature extraction, using natural language processing technology to extract key features from the preprocessed text, including entity recognition, relationship extraction and tendency analysis; step S3 is specifically as follows:

[0070] S31. According to the requirements of the relay protection information model, define the entity types that need to be identified, including device name, fault type, and time;

[0071] S32. Determine the types of relationships that exist in the text, including causal relationships, inclusion relationships, and contrast relationships;

[0072] S33. Construct a dictionary containing positive and negative words to judge the tendency of the text, analyze each sentence or paragraph in the text, calculate its score, and judge whether its tendency is positive or negative;

[0073] S34. Based on the results of entity recognition, relationship extraction and tendency analysis, select key features related to the relay protection information model, and use the cross-validation method to evaluate the effect of feature extraction to ensure that the extracted features can effectively support configuration information matching;

[0074] It should be noted that in step S31, the word sequence obtained after the text T is processed by word segmentation is W={ 1, 2, ..., n},in i represents the i-th word;

[0075] Define entity type set E = {equipment name, fault type, time};

[0076] For each word i , use named entity recognition technology to determine whether it belongs to a type in the entity type set E. If so, mark it as the corresponding entity type;

[0077] The set of entities extracted from the text T can be expressed as:

[0078] Entity collection = {(e j , type j)|e j ∈T, type j∈E};

[0079] In the formula, e j represents the jth entity extracted, and type type j represents the type to which the entity belongs;

[0080] In step S32, the relationship type existing in the text is determined, and the expression is:

[0081] R={(E i , E j )|E i , E j ∈ entity set, (E i , E j )∈relationtype};

[0082] In the formula, R represents the set of relations in the text, E i and E j is the identified entity;

[0083] In step S33, the propensity score of the text is calculated as follows:

[0084] Propensity score = ;

[0085] In the formula, if the tendency score is greater than 0, the text tendency is positive; if the tendency score is less than 0, the text tendency is negative;

[0086] S4, configuration information matching, comparing the extracted features with the predefined relay protection configuration information library to identify the specific configuration information corresponding to the text information; step S4 is specifically:

[0087] S41, initializing the configuration information matching module, loading parameters and resources, and calling the feature extraction module to extract key features from the text;

[0088] S42, loading a predefined relay protection configuration information library, and for each model in the model library, calculating a similarity score between the model and the extracted feature;

[0089] S43, for each model in the model library, calculate its similarity score with the extracted features, verify whether the selected model is accurate, and output the final result;

[0090] Furthermore, the similarity score between the extracted features and the extracted features is calculated as follows:

[0091] ;

[0092] In the formula, ‖A‖ represents the Euclidean norm of vector A, ‖B‖ represents the Euclidean norm of vector B, and the calculation formula is:

[0093] ;

[0094] ;

[0095] A and B are two n-dimensional feature vectors, where A=(a1, a2, ..., a n ), B=(b1, b2, ..., b n ).

[0096] S5, result output, based on the configuration information matching result, generate an easy-to-understand report or notification to guide the operation and maintenance personnel to perform corresponding operations; step S5 is specifically:

[0097] S51. Generate an easy-to-understand report or notification content based on the parsed matching results, including the model name, key features, possible problem description, and recommended operation steps;

[0098] S52. Send the report or notification to the relevant operation and maintenance personnel, ensure that the sent report or notification can reach the operation and maintenance personnel in time, and remind them to check and read it;

[0099] S53. Record the detailed information of the reports or notifications sent, and track the feedback and operations of the operation and maintenance personnel, so as to timely understand the progress and effect of problem solving.

[0100] In summary, compared with the prior art, the present invention has the following advantages:

[0101] Through the automated data acquisition, analysis and processing process, the need for manual intervention is reduced, and the efficiency of information processing is greatly improved. The feature extraction module uses natural language processing technology to quickly and accurately extract key features from text, providing strong support for subsequent analysis;

[0102] Automatically matching the predefined relay protection configuration information library and identifying the specific configuration information corresponding to the text information helps to quickly locate the cause of the fault. By providing possible problem descriptions and recommended operating procedures, the system provides valuable reference information for operation and maintenance personnel, helping them to solve problems more quickly.

[0103] Finally, it should be noted that the above is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, it is still possible for those skilled in the art to modify the technical solutions described in the aforementioned embodiments, or to make equivalent substitutions for some of the technical features therein. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. The intelligent identification system of relay protection configuration information based on natural language technology is characterized by: include: A data acquisition module, used to receive and parse text information sent from the relay protection device of the power system; A processing unit cleans, standardizes and formats the received text information to improve the accuracy of subsequent analysis; The feature extraction module uses natural language processing technology to extract key features from the preprocessed text, including entity recognition, relationship extraction, and tendency analysis; The configuration information matching unit compares the extracted features with the predefined relay protection configuration information library to identify the specific configuration information corresponding to the text information; The result output module generates easy-to-understand reports or notifications based on the configuration information matching results to guide operation and maintenance personnel to perform corresponding operations.

2. A method for intelligent identification of relay protection configuration based on natural language technology, the method being applied to the system of claim 1, characterized in that: The steps include: S1. Data acquisition: receiving and analyzing data information sent by relay protection devices of the power system; S2, data processing, cleaning, standardizing and formatting the received text information to improve the accuracy of subsequent analysis; S3, feature extraction, using natural language processing technology to extract key features from the preprocessed text, including entity recognition, relationship extraction and tendency analysis; S4, configuration information matching, comparing the extracted features with the predefined relay protection configuration information library to identify the specific configuration information corresponding to the text information; S5. Result output: Generate an easy-to-understand report or notification based on the configuration information matching results to guide the operation and maintenance personnel to perform corresponding operations.

3. The method for intelligent identification of relay protection configuration based on natural language technology according to claim 2 is characterized in that: When processing the data in step S2, it also includes using an algorithm to detect whether there are abnormal values ​​in the processed data, and making adjustments and corrections in time. The expression is: ； In the formula, X is the observed value, μ is the mean, and σ is the standard deviation. ≥1.5, considered as outliers.

4. The method for intelligent identification of relay protection configuration based on natural language technology according to claim 2 is characterized in that: Step S3 is specifically as follows: S31. According to the requirements of the relay protection information model, define the entity types that need to be identified, including device name, fault type, and time; S32. Determine the types of relationships that exist in the text, including causal relationships, inclusion relationships, and contrast relationships; S33. Construct a dictionary containing positive and negative words to judge the tendency of the text, analyze each sentence or paragraph in the text, calculate its score, and judge whether its tendency is positive or negative; S34. Based on the results of entity recognition, relationship extraction and tendency analysis, select key features related to the relay protection information model, and use the cross-validation method to evaluate the effect of feature extraction to ensure that the extracted features can effectively support configuration information matching.

5. The method for intelligent identification of relay protection configuration based on natural language technology according to claim 4 is characterized in that: In step S31, the word sequence obtained after the text T is segmented is W={ 1, 2, ..., n },in i represents the i-th word; Define entity type set E = {equipment name, fault type, time}; For each word i , use named entity recognition technology to determine whether it belongs to a type in the entity type set E. If so, mark it as the corresponding entity type; The set of entities extracted from the text T can be expressed as: Entity collection = {(e j , type j)|e j ∈T, type j∈E}; In the formula, e j represents the jth entity extracted, and type type j represents the type to which the entity belongs; In step S32, the relationship type existing in the text is determined, and the expression is: R={(E i , E j )|E i , E j ∈ entity set, (E i , E j )∈relationtype}; In the formula, R represents the set of relations in the text, E i and E j is the identified entity; In step S33, the propensity score of the text is calculated as follows: Propensity score = ; In the formula, if the tendency score is greater than 0, the text tendency is positive; if the tendency score is less than 0, the text tendency is negative.

6. The method for intelligent identification of relay protection configuration based on natural language technology according to claim 2 is characterized in that: Step S4 is specifically as follows: S41, initializing the configuration information matching module, loading parameters and resources, and calling the feature extraction module to extract key features from the text; S42, loading a predefined relay protection configuration information library, and for each model in the model library, calculating a similarity score between the model and the extracted feature; S43. For each model in the model library, calculate its similarity score with the extracted features, verify whether the selected model is accurate, and output the final result.

7. The method for intelligent identification of relay protection configuration based on natural language technology according to claim 6 is characterized in that: The similarity score between the extracted features and the extracted features is calculated as follows: ； In the formula, ‖A‖ represents the Euclidean norm of vector A, ‖B‖ represents the Euclidean norm of vector B, and the calculation formula is: ； ； A and B are two n-dimensional feature vectors, where A=(a1, a2, ..., a n ), B=(b1, b2, ..., b n ).

8. The method for intelligent identification of relay protection configuration based on natural language technology according to claim 2 is characterized in that: Step S5 is specifically as follows: S51. Generate an easy-to-understand report or notification content based on the parsed matching results, including the model name, key features, possible problem description, and recommended operation steps; S52. Send the report or notification to the relevant operation and maintenance personnel, ensure that the sent report or notification can reach the operation and maintenance personnel in time, and remind them to check and read it; S53. Record the detailed information of the reports or notifications sent, and track the feedback and operations of the operation and maintenance personnel, so as to timely understand the progress and effect of problem solving.

Images