Relay protection information correct verification system and method based on multi-scene adaptability technology

Abstract

The invention discloses a relay protection information correct verification system and method based on a multi-scene adaptability technology, and relates to the technical field of relay protection. An information type classification processing module; the protection information mapping module is associated with an expert knowledge verification model and is used for carrying out association mapping on the classified structured protection information and verification standard entries based on an intelligent matching algorithm, and mapping comprises semantic similarity matching of general information entries and expert knowledge weighted matching of personalized information entries; the intelligent verification module is associated with a verification strategy model; a fault-tolerant processing module; through cooperation of the configuration information acquisition module, the information type classification processing module, the protection information mapping module, the intelligent verification module and the fault-tolerant processing module, association mapping of relay protection information items to verification standard items can be realized, and the mapping success rate is improved in combination with an expert knowledge base.

Description

Technical Field

[0001] This invention belongs to the field of relay protection technology, specifically relating to a relay protection information correctness verification system and method based on multi-scenario adaptive technology. Background Technology

[0002] As the power system develops towards deep integration of control and regulation, remote operation of relay protection has become a core technical means to improve the efficiency of power grid operation and maintenance and reduce labor costs. Remote operation of relay protection includes remotely enabling and disabling switch plates, switching setting zones, modifying setting values, and resetting signals. In order to improve the accuracy of remote operation, a relay protection information verification system is needed to achieve precise control of the actions of relay protection devices.

[0003] Currently, relay protection information verification systems typically focus on electrical quantity monitoring, logic judgment, and action verification, combined with specific judgment rules for different protection types, to verify the correctness of relay protection device operation. Among these, protection information mapping is the core component of the relay protection information verification system. Protection information mapping establishes a unique association between the original information output by the relay protection device and the verification standard items preset by the master station, ensuring that the verification system can accurately identify device information, compare with the verification standards, and ultimately determine the correctness of the protection device's operating status.

[0004] However, because the existing mapping mechanism uses the same character-complete matching or simple keyword matching for all information entries, it does not take into account the universality and individual differences of the protected information entries. This leads to matching problems between the two types of entries, resulting in poor mapping adaptability. For example, if the main station's verification standard entry is the running fixed value area code, while a certain manufacturer's device outputs the current fixed value area code, the system cannot recognize that the two are the same information, determines that the information is missing, and requires manual intervention to confirm, which increases the operation and maintenance costs. Summary of the Invention

[0005] The purpose of this invention is to provide a relay protection information correct verification system and method based on multi-scenario adaptive technology, which can realize the association mapping between relay protection information items and verification standard items, and improve the mapping success rate by combining expert knowledge base, so as to solve the problems mentioned in the background technology.

[0006] To achieve the above objectives, the present invention adopts the following technical solution:

[0007] The relay protection information correct verification system based on multi-scenario adaptive technology includes: a configuration information acquisition module, which collects configuration information of relay protection devices of different models and equipment types, as well as relay protection information for calling the protection information substation;

[0008] The information type classification and processing module preprocesses and analyzes configuration information and relay protection information, first classifying them by device model and then by equipment type. At the same time, it transforms unstructured information into standardized structured information and removes invalid and redundant data.

[0009] The protection information mapping module is associated with an expert knowledge verification model. Based on the intelligent matching algorithm, it associates and maps the classified structured protection information with the verification standard items. The mapping includes semantic similarity matching of general information items and expert knowledge weighted matching of personalized information items.

[0010] The intelligent verification module, which is associated with the verification strategy model, performs intelligent verification of the mapped structured protection information entries based on the expert knowledge verification model and the verification strategy model. The multi-scenario adaptation includes different equipment model adaptation scenarios and different relay protection operation scenarios.

[0011] The fault-tolerant processing module corrects protocol conversion deviations and data transmission transient inconsistencies based on the general expert knowledge verification model for relay protection information.

[0012] Preferably, the configuration information acquisition module includes a security information substation summoning unit and a collection unit, both of which are connected to the information type classification and processing module;

[0013] The acquisition unit communicates with the physical interface of the relay protection device through a multi-interface industrial-grade acquisition gateway. The multi-interface industrial-grade acquisition gateway has a built-in dual power supply redundancy module and a lightning protection module. The configuration information includes relay protection device parameters, setting area configuration and soft pressure plate definition.

[0014] The protection information substation summoning unit summons the relay protection information of the protection information substation through the DL / T860 protocol or the substation 103 protocol. The relay protection information includes real-time remote signaling, telemetry and setting files.

[0015] Preferably, the information type classification processing module includes a classification rule engine unit, a structured conversion unit, and a redundancy filtering module connected in sequence. The classification rule engine unit stores a mapping table between device model and equipment type. The structured conversion unit has a built-in mapping table between unstructured description and standard fields. The redundancy filtering unit has preset invalid data judgment rules.

[0016] Preferably, the information type classification processing module performs the following preprocessing analysis process for configuration information and relay protection information:

[0017] A1. Receive configuration information from the module in JSON format, parse key fields, including device model, device type, parameter name and parameter value;

[0018] A2. Call the classification rule engine unit to execute primary and secondary classification. Primary classification groups the data by model and removes data without model identification. Secondary classification is based on model grouping and further subdivided by device function.

[0019] A3. Standardize and convert unstructured data customized by manufacturers into structured information;

[0020] A4. Call the redundancy filtering unit to delete invalid data that is irrelevant to protection verification, duplicate data, and data with incorrect formatting;

[0021] A5. The processed structured data is forwarded in batches to the protection information mapping module.

[0022] Preferably, the expert knowledge verification model includes an expert knowledge base, which includes a basic rule base and a dynamic update base. The basic rule base includes equipment model adaptation rules, parameter format rules, and semantic similarity thresholds. The dynamic update base includes historical mapping results and weight coefficients updated by machine learning. Basic information common to all models and types of relay protection devices is general information entries, while exclusive information strongly related to the relay protection device model and equipment type is personalized information entries.

[0023] The preferred semantic similarity matching process for general information items is as follows:

[0024] B1. Based on the expert knowledge base, classify the structured data output by the information type classification processing module and select the data belonging to the general information items;

[0025] B2. Convert the general entry name to be matched and the main site's verification standard entry name into 100-dimensional word vectors using the semantic analysis module;

[0026] B3. Calculate the cosine value of the angle between the general entry name to be matched and the main site's verification standard entry name using the cosine similarity algorithm to quantify their semantic similarity.

[0027] B4. Compare the calculated semantic similarity value with the preset semantic similarity threshold of the expert knowledge base. If the calculated semantic similarity value is greater than or equal to the preset semantic similarity threshold of the expert knowledge base, the match is considered successful; otherwise, the match is considered questionable and manual review is triggered.

[0028] The preferred expert knowledge-weighted matching process for personalized information entries is as follows:

[0029] C1. Based on the expert knowledge base, classify the structured data output by the information type classification processing module and select the data belonging to the personalized information items;

[0030] C2. Scores are calculated for personalized information entries from three dimensions: device model matching, parameter similarity, and scene adaptation.

[0031] C3. Calculate the total score using a weighted summation algorithm with preset weight coefficients;

[0032] C4. The total score is compared with the preset scoring threshold of the expert knowledge base to determine the matching result. When the total score is greater than or equal to 90 points, the matching is considered successful. When the total score is between 60 and 90 points, the matching is considered questionable and manual review is triggered. When the total score is below 60 points, the matching is considered unsuccessful and a warning message for information completion is issued.

[0033] The preferred process for intelligent verification in different device model adaptation scenarios is as follows:

[0034] D1. Extract the device model from the structured data of the configuration information acquisition module, and confirm the device model and exclusive characteristics through the model-characteristic mapping table in the expert knowledge base;

[0035] D2. Based on the equipment model, retrieve the corresponding model-specific verification template from the inspection strategy model library. The model-specific verification template includes parameter format conversion rules, model-specific thresholds, and signal association rules.

[0036] D3. To address the inconsistency in parameter formats caused by model differences, perform a unified format conversion to ensure consistent verification standards.

[0037] D4. Based on the current operating scenario, perform the adaptation verification, which includes parameter value verification, signal status verification, and protocol response verification.

[0038] D5. If all verification items are met, the verification passes. If the deviation is caused by model compatibility issues, the verification is deemed acceptable after adaptation. If the problem is not related to model compatibility, the verification fails and is reported to the main station.

[0039] Preferably, the fault tolerance processing module performs the following fault tolerance correction process:

[0040] E1. Monitor the intermediate data of the intelligent verification module in real time and identify the abnormal types of the intermediate data. The abnormal types include reduction conversion deviation abnormality and data transmission transient inconsistency abnormality.

[0041] E2. Invoke the corresponding fault tolerance strategy according to the exception type, that is, call the deviation correction strategy for the reduction conversion deviation exception, and call the retry acquisition strategy for the data transmission transient inconsistency exception.

[0042] E3. Perform fault tolerance correction based on the invoked fault tolerance strategy:

[0043] When correcting deviations, the absolute deviation rate must be calculated first. If the absolute deviation rate is less than or equal to the expert knowledge base threshold, it is determined to be fault-tolerant and the converted data is corrected to the original data. Otherwise, it is determined to be fault-intolerant and reported to the main station.

[0044] When retrying data acquisition, the function signal acquisition should be automatically triggered 3 times with a time interval of 10 seconds. If the function signal changes position after the retry, it is determined that the transient inconsistency has been resolved and the verification has passed. If the function signal still does not change position, it is determined that the equipment is faulty and the main station is notified.

[0045] E4. Provide real-time feedback on the fault tolerance correction results.

[0046] Based on the relay protection information correctness verification system based on multi-scenario adaptive technology described above, this invention also provides a relay protection information correctness verification method based on multi-scenario adaptive technology, including the following steps:

[0047] S1. Collect the configuration information of the relay protection device, and simultaneously summon the real-time protection information and setting file of the protection information substation;

[0048] S2. The collected information is first classified by device model, then further subdivided by equipment type. Unstructured information is converted into standardized structured information. At the same time, time setpoints and trip matrices are standardized and invalid redundant data is removed.

[0049] S3. Perform intelligent matching on the categorized structured information by combining the expert knowledge base: calculate semantic similarity to generate matching confidence for general information items, and perform weighted matching on personalized information items based on device-specific parameters to obtain mapped information items;

[0050] S4. Perform scenario-based verification on mapped information items based on expert knowledge verification model and verification strategy model: switch the corresponding verification logic according to the current operation scenario, and at the same time, use the fault tolerance processing module to correct the reduction conversion deviation and transient inconsistency.

[0051] S5. If the verification passes, generate a verification pass report; if the verification fails, link the expert knowledge base to output solution suggestions and record the exception information to the security audit log.

[0052] The relay protection information correctness verification system and method based on multi-scenario adaptability technology proposed in this invention have the following advantages compared with the prior art:

[0053] 1. This invention utilizes the coordinated operation of a configuration information acquisition module, an information type classification and processing module, a protection information mapping module, an intelligent verification module, and a fault-tolerant processing module. First, the configuration information acquisition module collects configuration information from different devices and obtains real-time remote signaling and setting files. Then, the information classification and processing module further subdivides the information according to device model and equipment type, converting unstructured data into standardized structured data. Simultaneously, it processes time settings and tripping matrices according to the requirements of the Southern Power Grid, eliminating redundancy. Finally, the protection information mapping module, combined with an expert knowledge base, generates confidence scores for general entries using semantic similarity, and weights personalized entries based on the equipment model, parameter similarity, and scenario adaptation weights preset in the expert knowledge base. The intelligent verification module then performs scenario-based verification on the mapped information entries, with the expert knowledge base dynamically updating weights and rules, significantly improving the mapping success rate.

[0054] 2. The intelligent verification module of the present invention is based on scenario-based verification logic, which accurately calls verification rules for different operation scenarios, satisfying the dual confirmation requirements of China Southern Power Grid, avoiding misjudgment by a single verification logic, reducing the misjudgment rate, and ensuring the security of remote operation.

[0055] 3. The configuration information acquisition module of the present invention collects device parameters, setting area configurations and soft pressure plate definitions of relay protection devices of different models and types, so as to achieve full coverage collection of configuration information of multiple manufacturers and multiple types of devices, avoid the subsequent verification gap due to missing information, and provide complete and accurate basic data support for verification.

[0056] 4. The fault-tolerant processing module of the present invention provides fault tolerance for protocol conversion deviations according to the expert database threshold, and sets a retry mechanism for transmission transient inconsistencies, thereby reducing verification interruptions caused by non-equipment failures, improving system stability, and reducing the frequency of manual intervention. Attached Figure Description

[0057] Figure 1 A system block diagram according to an embodiment of the present invention is shown;

[0058] Figure 2 A flowchart illustrating the preprocessing analysis process for configuration information and relay protection information according to an embodiment of the present invention is shown.

[0059] Figure 3 A flowchart illustrating the semantic similarity matching process for general information entries according to an embodiment of the present invention is shown.

[0060] Figure 4 A flowchart illustrating the expert knowledge-weighted matching process for personalized information entries according to an embodiment of the present invention is shown.

[0061] Figure 5 A flowchart of a method according to an embodiment of the present invention is shown. Detailed Implementation

[0062] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. The specific embodiments described herein are merely 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 those skilled in the art without creative effort are within the scope of protection of the present invention.

[0063] This invention provides, for example Figure 1-4 The relay protection information correctness verification system based on multi-scenario adaptive technology shown includes a configuration information acquisition module, an information type classification and processing module, a protection information mapping module associated with an expert knowledge verification model, an intelligent verification module associated with a verification strategy model, and a fault-tolerant processing module.

[0064] The configuration information acquisition module is used to collect configuration information of relay protection devices of different models and equipment types, as well as relay protection information for calling the protection information substation;

[0065] The configuration information acquisition module includes a security information substation summoning unit and a collection unit, both of which are connected to the information type classification and processing module.

[0066] The acquisition unit communicates with the physical interface of the relay protection device through a multi-interface industrial-grade acquisition gateway. The multi-interface industrial-grade acquisition gateway has a built-in dual power supply redundancy module and a lightning protection module. The configuration information includes relay protection device parameters, setting area configuration and soft pressure plate definition.

[0067] The protection information substation summoning unit summons the relay protection information of the protection information substation through the DL / T860 protocol or the substation 103 protocol. The relay protection information includes real-time remote signaling, telemetry and setting files.

[0068] The Baoxin substation summoning unit executes the following summoning process:

[0069] 1) Establish a TCP or UDP connection between the main station's information protection module and the information protection substation, and authenticate via IP whitelist. For TCP connection, the default port number is 102, and DL / T860 protocol is used for communication; for UDP connection, substation 103 protocol is used for communication; IP whitelist authentication only allows access from the main station's IP.

[0070] 2) When real-time remote signaling is invoked, DL / T860 issues GetDataValues ​​service, specifies the stVal attribute of SPCC logical node, and substation 103 issues general classification read service, ASDU type 0AH, function type FEH, information sequence number F8H;

[0071] When real-time telemetry is invoked, DL / T860 sends the GetDataValues ​​service, specifying the ActSG (fixed value area code) of SGCB and the Vol (voltage) of MMXU.

[0072] When the fixed value file is invoked, DL / T860 issues a FileTransfer service to read the file stored in the security information substation via the FTP protocol;

[0073] 3) After the data is returned by the security information substation, format verification and integrity verification are performed. The format verification is used to check whether the APDU structure of the DL / T860MMS message contains type identifier + length + data, and whether the type identifier + transmission reason of the 103 protocol ASDU is valid.

[0074] Integrity verification is used to calculate the CRC32 value of the fixed-value file and compare it with the CRC value reported by the security substation. If they match, the file is accepted; otherwise, it is retransmitted. The CRC32 value calculation formula is as follows:

[0075] 2,

[0076] in, The data in the set value file to be tested. As the initial value, To generate polynomials, For the i-th byte of data d, For XOR operation, Shift the i-th byte left by 8i bits. 2 represents the final result shifted left by 32 bits;

[0077] The CRC32 value is used to verify whether the fixed value file has been lost or tampered with during transmission. If the calculation result is consistent with the CRC value returned by the security substation, the data is intact.

[0078] 4) Encapsulate the verified real-time remote signaling, telemetry, and setting files into standardized data frames and forward them to the information type classification and processing module; real-time remote signaling such as disabling the reclosing soft pressure plate = exit, telemetry such as setting area number = 3;

[0079] The information type classification and processing module is used to preprocess and analyze configuration information and relay protection information. It first classifies information by device model, then by equipment type, and at the same time converts unstructured information into standardized structured information and removes invalid and redundant data.

[0080] The information type classification processing module includes a classification rule engine unit, a structured conversion unit, and a redundancy filtering module connected in sequence. The classification rule engine unit stores a mapping table between device model and equipment type. The structured conversion unit has a built-in mapping table between unstructured descriptions and standard fields. The redundancy filtering unit has preset invalid data judgment rules.

[0081] The information type classification and processing module performs the following preprocessing and analysis process for configuration information and relay protection information:

[0082] A1. Receive configuration information from the module in JSON format, parse key fields, including device model, device type, parameter name and parameter value;

[0083] A2. Call the classification rule engine unit to execute primary and secondary classification. Primary classification groups the data by model and removes data without model identification. Secondary classification is based on model grouping and further subdivided by device function.

[0084] A3. Standardize and convert unstructured data customized by manufacturers into structured information;

[0085] The standardized structured information transformation method uses the TF-IDF keyword extraction algorithm to extract parameters from unstructured data. The formula expression of the TF-IDF keyword extraction algorithm is as follows:

[0086] TF-IDF(w,D)=TF(w,D)*IDF(w,D),

[0087] ,

[0088] ,

[0089] Where TF-IDF(w,D) represents the keyword score, and TF(w,D) represents the frequency of keyword w in the unstructured data D. This represents the number of times the keyword w appears in the unstructured data D. Let w be the total number of occurrences of all words in the unstructured data D, IDF(w,D) be the scarcity of keyword w in all protection information data, and N be the total number of protection information data in the expert knowledge base. This represents the number of data items containing the keyword 'w'. Used to avoid betweenness numbers with a denominator of 0; the higher the keyword score, the more important the word is to the parameter data, and only words with a score greater than 0.5 are retained as keywords for standardized structured information conversion;

[0090] During the standardized structured information conversion, the keywords extracted by the TF-IDF keyword extraction algorithm are compared with a preset mapping table and converted into structured fields. The preset mapping table comes from the model and equipment type mapping rules stored in the expert knowledge base.

[0091] A4. Call the redundancy filtering unit to delete invalid data that is irrelevant to protection verification, duplicate data, and data with incorrect formatting;

[0092] A5. The processed structured data is forwarded in batches to the protection information mapping module.

[0093] The protection information mapping module is used to associate and map the classified structured protection information with the verification standard entries based on the intelligent matching algorithm. The mapping includes semantic similarity matching of general information entries and expert knowledge weighted matching of personalized information entries.

[0094] The expert knowledge verification model includes an expert knowledge base, which comprises a basic rule base and a dynamic update base. The basic rule base includes equipment model adaptation rules, parameter format rules, and semantic similarity thresholds. The dynamic update base includes historical mapping results and weight coefficients updated by machine learning. Basic information common to all models and types of relay protection devices is the general information entry, while exclusive information strongly related to the relay protection device model and equipment type is the personalized information entry.

[0095] The semantic similarity matching process for general information items is as follows:

[0096] B1. Based on the expert knowledge base, classify the structured data output by the information type classification processing module and select the data belonging to the general information items;

[0097] B2. The semantic analysis module converts the general entry names to be matched and the main site's verification standard entry names into 100-dimensional word vectors respectively; the vector dimension is an industry-standard optimized value, balancing computational efficiency and semantic representation accuracy.

[0098] The semantic analysis module integrates the Word2Vec word vector model, which includes a pre-trained lexicon for the power protection field. The pre-trained lexicon for the power protection field includes power-specific terms such as setting zone number, soft pressure plate, and remote signaling, as well as commonly used modifiers such as current, running, put into operation, and exited. It has been trained and optimized through 100,000 relay protection documents.

[0099] B3. Calculate the cosine value of the angle between the general entry name to be matched and the main site's verification standard entry name using the cosine similarity algorithm to quantify their semantic similarity.

[0100] The formula for the cosine similarity algorithm is:

[0101] ,

[0102] in, Here, A represents the semantic similarity value, and B represents the word vector of the general entry to be matched, while B represents the word vector of the standard entry verified by the main site. Let A be the dot product of vectors A and B. Let A be the magnitude of vector A. Let B be the magnitude of vector B, and k be the dimension of the word vector. Let i be the value of the i-th dimension of vector A. Let be the value of the i-th dimension of vector B, 1 k;

[0103] B4. Compare the calculated semantic similarity value with the preset semantic similarity threshold of the expert knowledge base. If the calculated semantic similarity value is greater than or equal to the preset semantic similarity threshold of the expert knowledge base, the match is considered successful; otherwise, the match is considered questionable and manual review is triggered.

[0104] The expert knowledge-weighted matching process for personalized information items is as follows:

[0105] C1. Based on the expert knowledge base, classify the structured data output by the information type classification processing module and select the data belonging to the personalized information items;

[0106] C2. Scores are calculated for personalized information entries from three dimensions: device model matching, parameter similarity, and scene adaptation.

[0107] The device model matching score is calculated as follows: it verifies whether the device model of the personalized information item to be matched is within the model range that the main site standard item is compatible with. If it is completely compatible, the device model matching score is 100 points; if it is a similar match, the device model matching score is 80 points; if it does not match, the device model matching score is 0 points.

[0108] The parameter similarity score is calculated as follows: The consistency between the format and value of the parameter to be matched and the standard parameter of the main site is verified. If the format and value of the parameter to be matched are completely consistent with the standard parameter of the main site, the parameter similarity score is 100 points; if the format of the parameter to be matched is consistent with the standard parameter of the main site, but the numerical deviation is less than 5%, the parameter similarity score is 80 points; if the format of the parameter to be matched is inconsistent with the standard parameter of the main site, the parameter similarity score is 0 points.

[0109] The scenario adaptation score is calculated as follows: the correlation between the current relay protection operation scenario and the item is required to be verified. If the current relay protection operation scenario is a mandatory verification item for the item, the scenario adaptation score is 100 points. If the current relay protection operation scenario is a non-mandatory verification item for the item, the scenario adaptation score is 60 points. If the current relay protection operation scenario is an irrelevant item for the item, the scenario adaptation score is 0 points.

[0110] C3. Calculate the total score using a weighted summation algorithm with preset weight coefficients;

[0111] The total score is calculated as follows: S = w1*S1 + w2*S2 + w3*S3.

[0112] Where S is the total matching score of the personalized information item, w1 is the device model matching weight coefficient, which is usually 40%, w2 is the parameter similarity weight coefficient, which is usually 30%, w3 is the scene adaptation weight coefficient, which is usually 30%, S1 is the device model matching score, S2 is the parameter similarity score, and S3 is the scene adaptation score.

[0113] C4. The total score is compared with the preset scoring threshold of the expert knowledge base to determine the matching result. When the total score is greater than or equal to 90 points, the matching is considered successful. When the total score is between 60 and 90 points, the matching is considered questionable and manual review is triggered. When the total score is below 60 points, the matching is considered unsuccessful and a warning message for information completion is issued.

[0114] The intelligent verification module is used to perform intelligent verification of the mapped structured protection information entries based on the expert knowledge verification model and the verification strategy model. The multi-scenario adaptation includes different equipment model adaptation scenarios and different relay protection operation scenarios.

[0115] The differences between different equipment models lie in parameter formats, signal descriptions, and protocol details, such as trip matrix format and timing accuracy. The intelligent verification process for different equipment models in various scenarios is as follows:

[0116] D1. Extract the device model from the structured data of the configuration information acquisition module, and confirm the device model and exclusive characteristics through the model-characteristic mapping table in the expert knowledge base;

[0117] D2. Based on the equipment model, retrieve the corresponding model-specific verification template from the inspection strategy model library. The model-specific verification template includes parameter format conversion rules, model-specific thresholds, and signal association rules.

[0118] D3. To address the inconsistency in parameter formats caused by model differences, perform a unified format conversion to ensure consistent verification standards.

[0119] For example, trip matrix format adaptation: CSC-103 outputs trip matrix binary 101101 → calls binary to hexadecimal algorithm to convert to hexadecimal 2D; PCS-931 directly outputs hexadecimal 2D → no conversion required; after both are unified into hexadecimal format, they are compared with the main station standard value 2D.

[0120] Time setting accuracy adaptation: CSC-103 output overcurrent stage I delay 0.05s (integer 50ms) → converted to floating point 0.05s; PCS-931 output overcurrent stage I delay 0.05s (direct floating point) → no conversion required; after unifying to floating point + second dimension, compare with the standard value of 0.05s;

[0121] D4. Based on the current operating scenario, perform the adaptation verification, which includes parameter value verification, signal status verification, and protocol response verification.

[0122] During parameter value verification, the fixed value comparison error algorithm is called to calculate the error between the converted actual value and the standard value, such as the trip matrix 2D and the standard 2D error being 0%;

[0123] The formula for the fixed-value comparison error algorithm is:

[0124] ,

[0125] in, To adjust the error rate of the setpoint comparison after adaptation, These are the actual values ​​of the device parameters after format adaptation. The standard values ​​preset by the main site;

[0126] Signal status verification: Check the model-specific abnormal signals, such as whether SPCC10.stVal of PCS-931 is 0, which indicates no abnormality;

[0127] Protocol response verification: Verify whether the device returns the correct response according to the exclusive protocol, such as PCS-931 returning Opera_rsp through DL / T860MMS service, and CSC-103 returning the execution result response through 103 protocol;

[0128] D5. If all verification items are met, the verification passes. If the deviation is due to model compatibility issues, the verification is deemed acceptable after adaptation. If the problem is not due to model compatibility issues, the verification fails and is reported to the main station.

[0129] Relay protection operations include remote activation / deactivation of soft pressure plates, remote switching of setting zones, remote modification of setting values, and remote reset signals. The intelligent verification methods for different relay protection operation scenarios are as follows:

[0130] When performing intelligent verification in the scenario of remote deployment / retraction of soft pressure plate, the system receives the command to deploy / retract the soft pressure plate from the main station, identifies the scenario as soft pressure plate deployment / retraction, calls the dual confirmation strategy of remote signaling status and function signal, collects the soft pressure plate status and corresponding function status, and determines whether the soft pressure plate status changes according to the operation target and whether the corresponding function status changes synchronously. If both the soft pressure plate status and the corresponding function status change and the response is qualified, the verification passes; otherwise, the verification fails.

[0131] During intelligent verification in the scenario of remote switching of the setpoint area, the system receives the command from the main station to switch to the target setpoint area, identifies the scenario as a setpoint area switch, calls the dual confirmation strategy of setpoint area number and setpoint content, collects the setpoint file of the currently running setpoint area number and the target area, determines whether the setpoint area number has changed according to the target, and compares the actual setpoint of the target area with the standard value of the main station. If the setpoint area number change is qualified and the setpoint content comparison error is less than 0.1%, the verification is deemed to have passed; otherwise, the verification fails.

[0132] When performing intelligent verification in a remote setting modification scenario, the system receives a command from the production station to modify the setting values ​​in the non-operational or preparatory areas. Recognizing the scenario as setting modification, it invokes a dual-confirmation strategy involving setting value comparison in the editing area and the integrity of the setting value file. The main station first issues a setting for the editing area number (target area), then issues a setting modification value instruction to set and modify the editing area. It then retrieves the actual setting values ​​in the editing area and the setting value file for that area, comparing the actual setting values ​​with the issued values ​​and verifying the file for tampering using CRC32. If the editing area comparison is successful and the file's CRC value matches that of the main station, the verification passes; otherwise, the verification fails.

[0133] During intelligent verification in remote reset signal operation scenarios, the system receives a signal command from the master station reset device, identifies the scenario as signal reset, and invokes the command response and signal reset confirmation strategy. The master station issues a reset command through the DL / T860 Oper_req service, collects the response returned by the acquisition unit, and collects the target signal status, such as the alarm signal Sig_Alarm.stVal. It determines whether the signal changes from 1=alarm to 0=normal, with a response time ≤5s. If the command response is successful and the signal reset is qualified, the verification passes; otherwise, the verification fails.

[0134] The fault-tolerant processing module is used to perform fault-tolerant correction of protocol conversion deviations and data transmission transient inconsistencies based on the general expert knowledge verification model of relay protection information.

[0135] The fault tolerance processing module performs the following fault tolerance correction process:

[0136] E1. Monitor the intermediate data of the intelligent verification module in real time and identify the abnormal types of the intermediate data. The abnormal types include protocol conversion deviation abnormality and data transmission transient inconsistency abnormality. Protocol conversion deviation is such as the 105ms time setting value sent by DL / T860 becoming 104ms after conversion by DL / T634.5104. Data transmission transient inconsistency is such as the soft pressure plate remote signaling status has changed, but the function signal has not changed within 5 seconds.

[0137] E2. Invoke the corresponding fault tolerance strategy according to the exception type, that is, call the deviation correction strategy for the reduction conversion deviation exception, and call the retry acquisition strategy for the data transmission transient inconsistency exception.

[0138] E3. Perform fault tolerance correction based on the invoked fault tolerance strategy:

[0139] When correcting deviations, the absolute deviation rate must be calculated first. If the absolute deviation rate is less than or equal to the expert knowledge base threshold, it is determined to be fault-tolerant and the converted data is corrected to the original data. Otherwise, it is determined to be fault-intolerant and reported to the main station.

[0140] When retrying data acquisition, the function signal acquisition should be automatically triggered 3 times with a time interval of 10 seconds. If the function signal changes position after the retry, it is determined that the transient inconsistency has been resolved and the verification has passed. If the function signal still does not change position, it is determined that the equipment is faulty and the main station is notified.

[0141] E4. Provide real-time feedback on the fault tolerance correction results, such as "deviation correction completed, verification passed" or "retry failed, equipment failure".

[0142] During verification, the configuration information acquisition module first collects configuration information of different devices and obtains real-time remote signaling and setting files. Then, the information classification and processing module subdivides the information according to device model and equipment type, converting unstructured data into standardized structured data. Simultaneously, it processes time setting values ​​and tripping matrices according to the requirements of China Southern Power Grid, eliminating redundancy. Finally, the protection information mapping module combines the expert knowledge base. General entries generate confidence scores using semantic similarity, while personalized entries are weighted and matched according to the equipment model, parameter similarity, and scenario adaptation weights preset by the expert database. The intelligent verification module then performs scenario-based verification on the mapped information entries. The expert database dynamically updates weights and rules, significantly improving the mapping success rate.

[0143] Based on the relay protection information correctness verification system based on multi-scenario adaptive technology described above, this invention also provides a relay protection information correctness verification method based on multi-scenario adaptive technology, such as... Figure 5 As shown, it includes the following steps:

[0144] S1. Collect the configuration information of the relay protection device, and simultaneously summon the real-time protection information and setting file of the protection information substation;

[0145] S2. The collected information is first classified by device model, then further subdivided by equipment type. Unstructured information is converted into standardized structured information. At the same time, time setpoints and trip matrices are standardized and invalid redundant data is removed.

[0146] S3. Perform intelligent matching on the categorized structured information by combining the expert knowledge base: calculate semantic similarity to generate matching confidence for general information items, and perform weighted matching on personalized information items based on device-specific parameters to obtain mapped information items;

[0147] S4. Perform scenario-based verification on mapped information items based on expert knowledge verification model and verification strategy model: switch the corresponding verification logic according to the current operation scenario, and at the same time, use the fault tolerance processing module to correct the reduction conversion deviation and transient inconsistency.

[0148] S5. If the verification passes, generate a verification pass report; if the verification fails, link the expert knowledge base to output solution suggestions and record the exception information to the security audit log.

[0149] Finally, it should be noted that the above description 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 foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A relay protection information correctness verification system based on multi-scenario adaptive technology, characterized in that: include: The configuration information acquisition module collects configuration information of relay protection devices of different models and equipment types, as well as relay protection information from the call protection substation; The information type classification and processing module preprocesses and analyzes configuration information and relay protection information, first classifying them by device model and then by equipment type. At the same time, it transforms unstructured information into standardized structured information and removes invalid and redundant data. The protection information mapping module is associated with an expert knowledge verification model. Based on the intelligent matching algorithm, it associates and maps the classified structured protection information with the verification standard items. The mapping includes semantic similarity matching of general information items and expert knowledge weighted matching of personalized information items. The intelligent verification module, which is associated with the verification strategy model, performs intelligent verification of the mapped structured protection information entries based on the expert knowledge verification model and the verification strategy model. The multi-scenario adaptation includes different equipment model adaptation scenarios and different relay protection operation scenarios. The fault-tolerant processing module corrects protocol conversion deviations and data transmission transient inconsistencies based on the general expert knowledge verification model for relay protection information.

2. The relay protection information correctness verification system based on multi-scenario adaptability technology according to claim 1, characterized in that: The configuration information acquisition module includes a security information substation summoning unit and a collection unit, both of which are connected to the information type classification and processing module. The acquisition unit communicates with the physical interface of the relay protection device through a multi-interface industrial-grade acquisition gateway. The multi-interface industrial-grade acquisition gateway has a built-in dual power supply redundancy module and a lightning protection module. The configuration information includes relay protection device parameters, setting area configuration and soft pressure plate definition. The protection information substation summoning unit summons the relay protection information of the protection information substation through the DL / T860 protocol or the substation 103 protocol. The relay protection information includes real-time remote signaling, telemetry and setting files.

3. The relay protection information correctness verification system based on multi-scenario adaptability technology according to claim 2, characterized in that: The information type classification processing module includes a classification rule engine unit, a structured conversion unit, and a redundancy filtering module connected in sequence. The classification rule engine unit stores a mapping table between device model and equipment type. The structured conversion unit has a built-in mapping table between unstructured descriptions and standard fields. The redundancy filtering unit has preset invalid data judgment rules.

4. The relay protection information correctness verification system based on multi-scenario adaptability technology according to claim 3, characterized in that: The information type classification and processing module performs the following preprocessing and analysis process for configuration information and relay protection information: A1. Receive configuration information from the module in JSON format, parse key fields, including device model, device type, parameter name and parameter value; A2. Call the classification rule engine unit to execute primary and secondary classification. Primary classification groups the data by model and removes data without model identification. Secondary classification is based on model grouping and further subdivided by device function. A3. Standardize and convert unstructured data customized by manufacturers into structured information; A4. Call the redundancy filtering unit to delete invalid data that is irrelevant to protection verification, duplicate data, and data with incorrect formatting; A5. The processed structured data is forwarded in batches to the protection information mapping module.

5. The relay protection information correctness verification system based on multi-scenario adaptability technology according to claim 4, characterized in that: The expert knowledge verification model includes an expert knowledge base, which comprises a basic rule base and a dynamic update base. The basic rule base includes equipment model adaptation rules, parameter format rules, and semantic similarity thresholds. The dynamic update base includes historical mapping results and weight coefficients updated by machine learning. Basic information common to all models and types of relay protection devices is the general information entry, while exclusive information strongly related to the relay protection device model and equipment type is the personalized information entry.

6. The relay protection information correctness verification system based on multi-scenario adaptability technology according to claim 5, characterized in that: The semantic similarity matching process for general information items is as follows: B1. Based on the expert knowledge base, classify the structured data output by the information type classification processing module and select the data belonging to the general information items; B2. Convert the general entry name to be matched and the main site's verification standard entry name into 100-dimensional word vectors using the semantic analysis module; B3. Calculate the cosine value of the angle between the general entry name to be matched and the main site's verification standard entry name using the cosine similarity algorithm to quantify their semantic similarity. B4. Compare the calculated semantic similarity value with the preset semantic similarity threshold of the expert knowledge base. If the calculated semantic similarity value is greater than or equal to the preset semantic similarity threshold of the expert knowledge base, the match is considered successful; otherwise, the match is considered questionable and manual review is triggered.

7. The relay protection information correctness verification system based on multi-scenario adaptability technology according to claim 6, characterized in that: The expert knowledge-weighted matching process for personalized information items is as follows: C1. Based on the expert knowledge base, classify the structured data output by the information type classification processing module and select the data belonging to the personalized information items; C2. Scores are calculated for personalized information entries from three dimensions: device model matching, parameter similarity, and scene adaptation. C3. Calculate the total score using a weighted summation algorithm with preset weight coefficients; C4. The total score is compared with the preset scoring threshold of the expert knowledge base to determine the matching result. When the total score is greater than or equal to 90 points, the matching is considered successful. When the total score is between 60 and 90 points, the matching is considered questionable and manual review is triggered. When the total score is below 60 points, the matching is considered unsuccessful and a warning message for information completion is issued.

8. The relay protection information correctness verification system based on multi-scenario adaptability technology according to claim 7, characterized in that: The process for intelligent verification in different device model adaptation scenarios is as follows: D1. Extract the device model from the structured data of the configuration information acquisition module, and confirm the device model and exclusive characteristics through the model-characteristic mapping table in the expert knowledge base; D2. Based on the equipment model, retrieve the corresponding model-specific verification template from the inspection strategy model library. The model-specific verification template includes parameter format conversion rules, model-specific thresholds, and signal association rules. D3. To address the inconsistency in parameter formats caused by model differences, perform a unified format conversion to ensure consistent verification standards. D4. Based on the current operating scenario, perform the adaptation verification, which includes parameter value verification, signal status verification, and protocol response verification. D5. If all verification items are met, the verification passes. If the deviation is caused by model compatibility issues, the verification is deemed acceptable after adaptation. If the problem is not related to model compatibility, the verification fails and is reported to the main station.

9. The relay protection information correctness verification system based on multi-scenario adaptability technology according to claim 8, characterized in that: The fault tolerance processing module performs the following fault tolerance correction process: E1. Monitor the intermediate data of the intelligent verification module in real time and identify the abnormal types of the intermediate data. The abnormal types include reduction conversion deviation abnormality and data transmission transient inconsistency abnormality. E2. Invoke the corresponding fault tolerance strategy according to the exception type, that is, call the deviation correction strategy for the reduction conversion deviation exception, and call the retry acquisition strategy for the data transmission transient inconsistency exception. E3. Perform fault tolerance correction based on the invoked fault tolerance strategy: When correcting deviations, the absolute deviation rate must be calculated first. If the absolute deviation rate is less than or equal to the expert knowledge base threshold, it is determined to be fault-tolerant and the converted data is corrected to the original data. Otherwise, it is determined to be fault-intolerant and reported to the main station. When retrying data acquisition, the function signal acquisition should be automatically triggered 3 times with a time interval of 10 seconds. If the function signal changes position after the retry, it is determined that the transient inconsistency has been resolved and the verification has passed. If the function signal still does not change position, it is determined that the equipment is faulty and the main station is notified. E4. Provide real-time feedback on the fault tolerance correction results.

10. A method for verifying the correctness of relay protection information based on multi-scenario adaptive technology, and a system for verifying the correctness of relay protection information based on multi-scenario adaptive technology as described in any one of claims 1-9, characterized in that: Includes the following steps: S1. Collect the configuration information of the relay protection device, and simultaneously summon the real-time protection information and setting file of the protection information substation; S2. The collected information is first classified by device model, then further subdivided by equipment type. Unstructured information is converted into standardized structured information. At the same time, time setpoints and trip matrices are standardized and invalid redundant data is removed. S3. Combine the expert knowledge base to perform intelligent matching on the classified structured information: calculate the semantic similarity of general information items to generate matching confidence, and perform weighted matching on personalized information items based on device-specific parameters to obtain mapped information items; S4. Perform scenario-based verification on mapped information items based on expert knowledge verification model and verification strategy model: switch the corresponding verification logic according to the current operation scenario, and at the same time, use the fault tolerance processing module to correct the reduction conversion deviation and transient inconsistency. S5. If the verification passes, a verification pass report will be generated; If the verification fails, the system will output a solution suggestion from the expert knowledge base and record the exception information in the security audit log.

Images