Power plant relay protection fault diagnosis method for centralized control center

Abstract

The invention is suitable for the technical field of power system diagnosis, and provides a centralized control center-oriented power plant relay protection fault diagnosis method, which comprises the following steps of: reconstructing a fault event sequence with sequential logic by receiving a relay protection action alarm signal and automatically acquiring associated fault data; the method comprises the following steps of: configuring and executing multi-dimensional analysis tasks including electrical quantity analysis, protection logic check and historical case matching, automatically judging action behaviors of the protection device according to cross validation of task results, intelligently comparing and storing diagnosis characteristics and conclusions with a historical case library, and finally automatically generating a diagnosis report. According to the invention, automation of the whole process from fault data collection and intelligent analysis to report generation is realized, and the response speed and processing efficiency of fault analysis are effectively improved; through the built-in standardized judgment logic, the dependence on artificial experience is reduced, and the consistency and objective accuracy of the diagnosis conclusion are guaranteed.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of power system diagnosis, in particular to a power plant relay protection fault diagnosis method for a centralized control center. BACKGROUND

[0002] In the safe and stable operation of the power system, the relay protection device is a key equipment for quickly isolating faults, and the correctness of its action behavior is crucial. With the advancement of intensive management of power generation enterprises, the centralized control center mode is increasingly popular, and it is necessary to centrally monitor and analyze faults of the relay protection equipment of multiple decentralized power plants under the control of the center. The current fault analysis process includes: when the centralized control center monitoring system receives the protection action alarm sent by the power station, the operation and maintenance personnel need to manually retrieve the corresponding fault recording file, event sequence record and protection device action message in turn; professional personnel need to combine their own experience to analyze the time comparison, logic analysis and waveform interpretation of these multi-source, heterogeneous and time-stamped data to infer the fault point and evaluate the protection action behavior, and finally form an analysis report.

[0003] However, the above method analysis process is time-consuming and laborious, and the entire process from data collection to report generation is slow in response, which is difficult to meet the needs of rapid review and decision-making after the power grid fails. Secondly, the accuracy and consistency of the analysis conclusion depend on the professional skill level of the on-duty personnel, and the analysis results of the same fault by different personnel or at different times may differ, lacking standardized judgment criteria. Thus, it is unable to adapt to the intelligent needs of the centralized control center to implement efficient, accurate and standardized remote technical supervision of subordinate power stations. SUMMARY

[0004] The present application provides a power plant relay protection fault diagnosis method for a centralized control center, which is used to solve the problem of failing to adapt to the intelligent needs of the centralized control center to implement efficient, accurate and standardized remote technical supervision of subordinate power stations.

[0005] The present application provides a power plant relay protection fault diagnosis method for a centralized control center, which includes: receiving a relay protection action alarm signal sent by a subordinate power plant, and collecting fault data associated with the relay protection action alarm signal; analyzing the relay protection action alarm signal and the fault data, and reconstructing a fault event sequence with time sequence logic according to the power station and the interval; configuring and executing a plurality of analysis tasks of different dimensions for the fault events in the fault event sequence, the analysis tasks including at least a first type of task for analyzing electrical quantities, a second type of task for checking the protection device action logic, and a third type of task for matching historical cases; Determine the action behavior of the protection device according to the results of various analysis tasks; wherein: If the result of the first type of task indicates that there is a fault feature in the protection area of the protection device, and the result of the second type of task indicates that the action of the protection device meets the preset condition, it is determined that the action of the protection device is correct; If the result of the first type of task indicates that there is a fault feature in the protection area of the protection device, but the result of the second type of task indicates that the action of the protection device does not meet the preset condition, it is determined that the action of the protection device is abnormal; If the result of the first type of task does not detect a fault feature in the protection area of the protection device, but the result of the second type of task indicates that the protection device has acted, it is determined that the protection device misoperates; Compare and store the diagnosed feature information, the determination conclusion and the result of the third type of task with the historical diagnosis case library; Based on the determination conclusion and the comparison result with the historical diagnosis case library, automatically generate a fault diagnosis analysis report.

[0006] Further, the fault data includes a fault recording file, an event sequence record and a protection device action message.

[0007] Further, the analysis of the relay protection action alarm signal and the fault data, the reconstruction of the fault event sequence with time sequence logic according to the substation and the interval, comprises: Respectively perform time alignment processing on the relay protection action alarm signal, the fault recording file, the event sequence record and the protection device action message; Extract the start time of the fault recording file, the switch position change time in the event sequence record and the logic action time in the protection device action message from the time-aligned data; According to the start time, the switch position change time and the logic action time, associate the relay protection action alarm signal, the switch position change event and the protection logic action event in the same substation and interval in time sequence to generate a fault event sequence.

[0008] Further, the result of the second type of task indicating that the action of the protection device meets the preset condition comprises: Obtain the actual action value of the protection device output by the second type of task and the corresponding action setting value; Obtain the action time of the protection device output by the second type of task, and the fault start time determined from the fault event sequence; If the actual action value is greater than or equal to the action fixed value, and the time difference between the action time and the fault starting time is within the preset protection action time window, it is determined that the action of the protection device meets the preset condition.

[0009] Further, the determining that the protection device acts correctly comprises: obtaining a historical case matched with the current fault from the third type of task output; If the final conclusion of the matched historical case is that the action is correct, the label of the current determination conclusion is marked as a high confidence level; If no historical case is matched, or the final conclusion of the matched historical case is not that the action is correct, the label of the current determination conclusion is marked as a regular confidence level.

[0010] Further, the determining that the protection device acts abnormally comprises: If the result of the first type of task indicates that there is a fault feature in the protection area of the protection device, it is further determined whether the result of the second type of task contains the action signal of the protection device; If the result of the second type of task does not contain any action signal of the protection device, it is determined that the protection device fails to act.

[0011] Further, the determining that the protection device acts abnormally further comprises: If the result of the second type of task contains the action signal of the protection device, the actual action value and the action fixed value corresponding to the action signal, and the action time and the fault starting time are obtained; If the actual action value is less than the action fixed value, it is determined that the protection device is logically abnormal due to not reaching the fixed value; If the actual action value is greater than or equal to the action fixed value, but the time difference between the action time and the fault starting time is not within the preset protection action time window, it is determined that the protection device is logically abnormal due to action time sequence error.

[0012] Further, the determining that the protection device misacts comprises: Based on the result of the first type of task, the position of the fault feature is determined; If the fault feature is not detected in the protection area of the protection device, and the fault feature is also not detected in the adjacent backup protection area, it is determined that the protection device is a fault-free misaction; If the fault feature is detected to be located outside the protection area of the protection device, but is detected in the adjacent backup protection area, it is determined that the protection device is a zone-out fault misaction.

[0013] Furthermore, the comparison of the diagnostic feature information, judgment conclusion, and the results of the third type of task with the historical diagnostic case database includes: Based on the fault type, faulty device identifier and fault location in the feature information, a search is performed in the historical diagnostic case database to obtain a set of candidate cases. Calculate the similarity between the feature information of each historical case in the candidate case set and the feature information of the current diagnosis; Based on the similarity and a preset similarity threshold, a matching result is determined from the candidate case set. The matching result includes an indication of a matched historical case or a historical case that was not matched.

[0014] Furthermore, storing the diagnostic feature information, judgment conclusion, and the results of the third type of task with the historical diagnostic case database includes: The storage mode of this diagnostic record is determined based on the matching results; If the storage mode is a new storage mode, the complete record of this diagnosis will be stored as a new case in the historical diagnosis case library. If the storage mode is an update storage mode, then update the call count and last call time information of the matched historical cases; If the storage mode is an associated storage mode, a case record associated with the matched historical case is created for this diagnosis and stored in the historical diagnosis case library.

[0015] As can be seen from the above technical solutions, the present invention has the following advantages: This invention receives relay protection alarm signals and automatically collects associated fault data, reconstructs a fault event sequence with temporal logic, and then configures and executes multi-dimensional analysis tasks, including electrical quantity analysis, protection logic verification, and historical case matching. Based on the cross-validation of the results of each task, it automatically determines the behavior of the protection device and intelligently compares and stores diagnostic features and conclusions with a historical case database, ultimately automatically generating a diagnostic report. This invention achieves full automation from fault data collection and intelligent analysis to report generation, effectively improving the response speed and processing efficiency of fault analysis. Through built-in standardized judgment logic, it reduces reliance on human experience, ensuring the consistency and objectivity of diagnostic conclusions. Simultaneously, by utilizing the continuous comparison and evolution of the historical case database, the system possesses self-learning and knowledge accumulation capabilities. Therefore, it provides the control center with an efficient, accurate, and continuously evolving intelligent remote technical supervision method. Attached Figure Description

[0016] Figure 1 This is a flowchart illustrating a power plant relay protection fault diagnosis method for centralized control centers according to the present invention. Figure 2 This is a flowchart illustrating the logic for judging the action behavior of the protection device based on the results of various analysis tasks in this invention. Detailed Implementation

[0017] The terms “first,” “second,” “third,” “fourth,” etc. (if present) in the specification and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms “comprising” and “corresponding to,” and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0018] Example 1 The implementation method in this embodiment can be implemented in a system, on a server, or on a terminal; no specific limitation is made. The method in this application will be described below from the perspective of system implementation. Please refer to... Figure 1 The method provided in this application includes the following steps: S1. Receive relay protection alarm signals sent by subordinate power plants and collect fault data associated with the relay protection alarm signals; In this embodiment, this step is performed by a fault diagnosis server deployed in the central control center. This server monitors and receives relay protection alarm signals actively sent by the main or substation information stations of each power plant in real time via the power dispatch data network or a dedicated communication channel. This relay protection alarm signal is a digital communication signal generated by the protection device when it detects a power system fault and initiates or completes tripping logic. It includes information such as the plant / station identifier, protection device identifier, action type, action phase, and approximate time. After capturing this alarm signal, the server automatically triggers a more detailed fault data retrieval process associated with it.

[0019] The fault data includes fault waveform files, event sequence records, and protection device action messages. Fault waveform files are instantaneous waveform data files of key electrical quantities recorded by fault waveform recording devices installed in the plant or by the internal waveform recording function of the protection devices, representing a period of time before and after the fault. Event sequence records are sequences of events with timestamps, such as switch changes, protection actions, and alarm signals, recorded by the plant monitoring system or protection information system, reflecting the timing of actions of various devices during the fault process. Protection device action messages are detailed action reports generated by the protection device itself, recording the startup, action, return logic, action timestamps, and comparison results with set values ​​for each protection element within the device.

[0020] S2. Analyze the relay protection alarm signals and fault data, and reconstruct them into a fault event sequence with time-series logic according to the substation and bay. This step integrates and correlates the multi-source data received in step S1 to reconstruct the complete dynamic process of fault occurrence, development, and remediation. This is achieved through the following steps: 1. Time alignment processing is performed on relay protection alarm signals, fault recording files, event sequence records, and protection device action messages respectively; 2. Extract the start time of the fault waveform file, the switch change time in the event sequence record, and the logic action time in the protection device action message from the time-aligned data; 3. Based on the start time, switch change time, and logic action time, associate the relay protection action alarm signals, switch change events, and protection logic action events within the same substation and bay in chronological order to generate a fault event sequence.

[0021] Specifically, a time-stamped interpolation correction method is used to unify all data to the same reference time in the central control center, eliminating errors caused by clock differences between different devices. The absolute fault initiation time is extracted from the header information of the fault recording file; the precise timestamps of events such as circuit breaker opening / closing and protection hard contact actions are extracted from the event sequence records; and the logical times of each protection element's initiation, action, and return are extracted from the protection device action messages. Finally, using substation identifiers and electrical bays as basic units, the system sorts and logically associates all events belonging to the same fault process according to their timestamp order, thereby generating a structured fault event sequence. This sequence not only records what happened in chronological order but also reveals why it happened through the contextual relationships between events, providing precise and orderly input for subsequent intelligent analysis.

[0022] S3. For fault events in the fault event sequence, configure and execute multiple analysis tasks of different dimensions. The analysis tasks include at least a first type of task for analyzing electrical quantities, a second type of task for verifying the action logic of protection devices, and a third type of task for matching historical cases. The first type of task (electrical quantity analysis) analyzes electrical waveform data based on fault waveform files associated with the fault event sequence, extracting and calculating fault characteristic quantities, including but not limited to the amplitude, phase, harmonic content, impedance, and power direction of fault current and voltage, to determine the fault type, location, and severity. The second type of task (protection logic verification) verifies the actual operating logic of the protection device item by item based on the fault event sequence and protection device action messages, combined with preset protection device settings and the primary system topology model, analyzing whether its start-up, action, and return conditions meet preset setting thresholds and timing requirements. The third type of task (historical case matching) searches and calculates similarity in a historical diagnostic case database based on the characteristic information of the current fault to find similar historical fault records that can be used for reference. These three types of tasks are analyzed in parallel from three different dimensions: electrical facts, protection logic, and empirical analogy, providing multi-perspective and mutually corroborating evidence for subsequent comprehensive judgment.

[0023] S4. Determine the action behavior of the protection device based on the results of various analysis tasks; among which: Based on the parallel multi-dimensional analysis results in S3, automatic cross-validation and comprehensive judgment are performed through built-in logical rules, thereby replacing manual experience-based judgment and achieving objective conclusions regarding whether the protection device operates correctly, abnormally, or malfunctions. Its core lies in establishing a deterministic mapping relationship between electrical facts (first-type task results) and protection logic (second-type task results), supplemented by historical experience (third-type task results) for confidence assessment. Please refer to [link / reference]. Figure 2 The judgment logic includes the following: S41. If the result of the first type of task indicates that there are fault characteristics within the protection zone of the protection device, and the result of the second type of task indicates that the action of the protection device meets the preset conditions, then the action of the protection device is determined to be correct. Here, the presence of fault characteristics within the protection zone of the protection device means that the analysis of the first type of task confirms that the characteristics of the fault electrical quantity and its geographical location fall within the theoretical protection range of the protection device, such as the generator differential protection zone, transformer differential protection zone, and line distance protection section I. "Action meeting preset conditions" means that the actual response behavior of the protection device simultaneously meets the two core requirements of the action setting and the action sequence. These constitute the two necessary conditions for determining that the action is correct. If either condition is not met, even if fault characteristics exist, it cannot be determined as a correct action.

[0024] In this embodiment, the result of the second type of task indicates that the action of the protection device meets preset conditions, including the following: 1. Obtain the actual action value and corresponding action set value of the protection device output by the second type of task; 2. Obtain the action time of the protection device output by the second type of task, and the fault initiation time determined from the fault event sequence; 3. If the actual action value is greater than or equal to the action set value, and the time difference between the action time and the fault start time is within the preset protection action time window, then the action of the protection device is determined to meet the preset conditions.

[0025] Specifically, the actual action value refers to the value of the electrical quantity measured by the protection device and used as the start criterion at the time of the fault; the corresponding action setpoint is the threshold value that is pre-set and input into the protection device and serves as its action threshold. The action time is the precise time when the protection element completes the logical judgment and issues the action command, as parsed from the protection device's action message by the second type of task; the fault start time is the absolute time of the initial occurrence of the fault, recorded by the fault waveform file, obtained from the fault event sequence. The preset protection action time window is a reasonable time range pre-set according to the protection principle, equipment performance, and system stability requirements. For example, for the main protection, this time window may be 20-50 milliseconds after the fault starts. The time difference between the action time and the fault start time is calculated. If the difference falls within this time window, the action sequence is considered reasonable. Only when the actual action value exceeds or equals the setpoint (criterion satisfied) and the time difference is within the reasonable window (action rapid and reasonable) is it finally determined to meet the preset conditions. If the actual action value is less than the setpoint, it is determined that the setpoint has not been reached; if the time difference exceeds the time window, it is determined that the action sequence is incorrect, both of which are cases of not meeting the preset conditions.

[0026] In this embodiment, determining that the protection device has operated correctly includes: 1. Obtain historical cases from the third type of task output that match the current fault; 2. If the final conclusion of the matched historical cases is that the action was correct, then the conclusion of this judgment will be marked as high confidence level; 3. If no historical case is matched, or if the final conclusion of the matched historical case is not that the action was correct, the conclusion of this judgment will be marked as the normal confidence level.

[0027] Specifically, the system first reads the matching results from the third type of task output. This result may be a specific historical case record or an indication of no match. If a historical case is matched, and the final diagnostic conclusion recorded in that historical case is also that the protective device operated correctly, the system considers this judgment to be supported by historical experience. Therefore, the confidence level of this conclusion is marked as high, and in the subsequently generated report, this conclusion can be highlighted or accompanied by an explanation that it is consistent with historical experience. If no historical case is matched, or the conclusion of the matched historical case is not that the operation was correct, the system considers this judgment to lack strong support from historical experience or to have potential contradictions. In this case, the confidence level is marked as normal, indicating that the conclusion is based on this data analysis, and it is recommended to review it in conjunction with other information.

[0028] S42. If the result of the first type of task indicates that there are fault characteristics in the protection zone of the protection device, but the result of the second type of task indicates that the action of the protection device does not meet the preset conditions, then the action of the protection device is determined to be abnormal. If the action here does not meet the preset conditions, it means that the protection device's response to the fault in the zone is biased. Based on the specific information provided by the second type of task, the system will further distinguish whether this abnormality is manifested as a complete failure to respond or a logical abnormality with a response logic error, thereby providing a more accurate diagnostic conclusion.

[0029] In this embodiment, determining that the protection device is malfunctioning includes: 1. If the result of the first type of task indicates that there are fault characteristics within the protection zone of the protection device, then it is further determined whether the result of the second type of task includes the action signal of the protection device; 2. If the result of the second type of task does not contain any action signal of the protection device, then the protection device is deemed to have failed to operate.

[0030] Specifically, an action signal refers to a specific status bit or event record in the protection device's action message that indicates that its protective element has taken action. After confirming the presence of fault characteristics within the zone, the system will retrieve the results of the second type of task. If no action record of any protective element related to the current fault is found in the message, it indicates that the protection device did not issue any trip or alarm command when it should have acted, and is therefore judged as protection failure to operate. This is one of the most serious abnormal situations and may result in the fault not being able to be quickly cleared.

[0031] In this embodiment, determining that the protection device is malfunctioning also includes: 1. If the result of the second type of task includes the action signal of the protection device, then obtain the actual action value and action set value corresponding to the action signal, as well as the corresponding action time and fault start time; 2. If the actual action value is less than the action set value, the protection device is determined to be malfunctioning due to failure to reach the set value; 3. If the actual action value is greater than or equal to the action set value, but the time difference between the action time and the fault start time is not within the preset protection action time window, the protection device is judged to be logically abnormal due to an incorrect action timing.

[0032] Specifically, if the result of the second type of task includes an action signal, it indicates that the protection device has responded, but its response needs to be verified. The system first extracts the specific numerical value and timing information associated with the action signal. If the actual action value is less than the action setpoint, it means that the protection device issued the action command without meeting its activation criteria, which violates the most basic protection logic. This may stem from internal calculation errors, sampling anomalies, or setpoint drift, and is therefore judged as a logic anomaly due to failure to reach the setpoint. If the actual action value is greater than or equal to the action setpoint, but the time difference between the action time and the fault activation time is not within the preset protection action time window, such as severely ahead of schedule (action before the fault occurs) or severely behind schedule (far exceeding the main protection's expected action time), it indicates that the protection device's logic judgment or output circuit may have timing disorder, component damage, or abnormal interference, and is therefore judged as a logic anomaly due to incorrect action timing. Both of these situations belong to protection device malfunction or functional disorder.

[0033] S43. If the result of the first type of task does not detect fault characteristics within the protection zone of the protection device, but the result of the second type of task indicates that the protection device has been activated, then the protection device is determined to have malfunctioned. Here, maloperation refers to the protection device issuing an operation command when it should not have operated. Further differentiation, based on the actual location of the faulty electrical quantity, between maloperation caused by the device itself and its selectivity (maloperation to faults outside the designated area) is crucial for locating the root cause of the fault and developing countermeasures.

[0034] In this embodiment, determining that the protection device has malfunctioned includes: 1. Based on the results of the first type of task, determine the location of the fault characteristics; 2. If the fault characteristics are not detected in the protection zone of the protection device, and the fault characteristics are not detected in the adjacent backup protection zone either, then the protection device is determined to be a faultless malfunction. 3. If the fault characteristics are detected outside the protection zone of the protection device, but are detected in the adjacent backup protection zone, the protection device is determined to be malfunctioning due to an external fault.

[0035] Specifically, the system first determines the electrical location of the fault point based on characteristic quantities such as impedance and power direction calculated from the first type of task. Adjacent backup protection zones refer to the protection range of devices electrically adjacent to this protection device and serving as its backup protection; for example, distance protection section I of a line is the main protection for this line, and its adjacent backup protection zones may include distance protection sections II and III of this line, or the protection range of the opposite / next-level line. If the fault characteristics are not detected in all relevant protection zones (including the main protection zone of this device and all adjacent backup protection zones), it indicates that no fault has actually occurred in the system, and the operation of the protection device is entirely due to its own component defects, secondary circuit interference, setting errors, or accidental contact, etc., and is therefore judged as a faultless maloperation. If the fault characteristics are clearly detected outside the protection zone of this protection device, but within its adjacent backup protection zone, it indicates that a fault does exist in the system, but this protection device has incorrectly responded to an external fault that is not within its protection range. This is due to improper protection setting coordination, failure of directional element criteria, or inaccurate judgment under complex fault conditions, etc., and is therefore judged as an external fault maloperation. This distinction helps maintenance personnel take targeted inspection and handling measures.

[0036] S5. Compare and store the diagnostic feature information, judgment conclusions, and results of the third type of task with the historical diagnostic case database; By systematically comparing the complete information of this diagnosis with the historical case database and intelligently deciding on storage strategies based on the comparison results, the system can not only use historical experience to assist current analysis, but also transform new diagnostic cases into structured knowledge, thereby continuously optimizing and enriching the case database and improving the accuracy and efficiency of subsequent diagnoses.

[0037] In this embodiment, the diagnostic feature information, judgment conclusion, and the results of the third type of task are compared with the historical diagnostic case database, including the following steps: 1. Based on the fault type, faulty equipment identifier, and fault location in the feature information, a search is performed in the historical diagnostic case database to obtain a set of candidate cases; 2. Calculate the similarity between the feature information of each historical case in the candidate case set and the feature information of the current diagnosis; 3. Based on similarity and a preset similarity threshold, determine the matching results from the candidate case set. The matching results include indicators of matched historical cases or no matched historical cases.

[0038] Specifically, the system first uses fault type, faulty device identifier, and fault location as key search fields to perform a query in the historical diagnostic case database, obtaining a preliminary set of relevant candidate cases. For each historical case in the set, the system extracts its more detailed feature information vector and calculates its similarity with the feature information vector of the current diagnosis using cosine similarity. The system presets a similarity threshold to quantify whether a valid match is formed. If the similarity calculation result of a historical case in the candidate set is greater than the threshold, the case with the highest similarity is selected as the matched historical case; if the similarity of all historical cases does not exceed the threshold, or the candidate set is empty, the matching result is no matched historical case.

[0039] In this embodiment, the diagnostic feature information, judgment conclusion, and results of the third type of task are stored with the historical diagnostic case database, including the following steps: 1. Determine the storage mode for this diagnostic record based on the matching results; 2. If the storage mode is new storage mode, the complete record of this diagnosis will be stored as a new case in the historical diagnosis case database; 2. If the storage mode is update storage mode, then update the call count and last call time information of the matched historical cases; 3. If the storage mode is associated storage mode, a case record associated with the matched historical case will be created for this diagnosis and stored in the historical diagnosis case library.

[0040] Specifically, the system first determines the storage mode based on the matching results obtained from the aforementioned comparison steps. If no historical case is matched, the storage mode is determined to be the new storage mode. The system integrates all data from this diagnosis, including feature information, fault event sequence, task results, and judgment conclusions, into a new case record, stores it in the case library, and assigns it an initial state. If a historical case is matched, and the judgment conclusion of this diagnosis is consistent with the final conclusion recorded in the matched historical case, the storage mode is determined to be the update storage mode. The system does not add new records to the case library but updates the metadata of the matched historical case, such as increasing its call count and refreshing its last call time, to reflect the reuse frequency and timeliness of the case experience. If a historical case is matched, but the judgment conclusion of this diagnosis is inconsistent with the final conclusion of the historical case, the storage mode is determined to be the associated storage mode. The system creates a new case record for this diagnosis, but at the same time marks the unique identifier of the associated historical case with conflicting conclusions in this record, and marks the status of both records as pending review or questionable conclusion, thus forming an associated conflict group in the knowledge base.

[0041] S6. Based on the judgment conclusion and the comparison results with the historical diagnostic case library, automatically generate a fault diagnosis analysis report.

[0042] After completing fault diagnosis and case library management, the system automatically calls a preset report generation template to structurally integrate the key information and conclusions generated in the preceding steps, outputting a complete fault diagnosis and analysis report. The report content includes at least: a fault event overview, a detailed description of the multi-dimensional analysis process, the final judgment conclusion, fault location results, historical case matching information, and diagnostic conclusions and handling recommendations. The report is presented in a format combining structured text with key waveform screenshots, time-series diagrams, and other visual elements. It can be directly displayed on the human-machine interface of the central control center monitoring system or exported as a standard document for archiving and distribution. This step achieves full automation from data analysis to result delivery, ensuring the report's standardization, timeliness, and professionalism, providing operators with clear and direct decision-making and action guidelines.

[0043] It is understood that those skilled in the art can combine various implementation methods in the above embodiments under the guidance of the above examples to obtain technical solutions with multiple implementation methods.

[0044] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for fault diagnosis of power plant relay protection in a centralized control center, characterized in that, include: Receive relay protection alarm signals sent by subordinate power plants and collect fault data associated with the relay protection alarm signals; The relay protection alarm signals and fault data are analyzed and reconstructed by substation and bay to form a fault event sequence with time-series logic; For the fault events in the fault event sequence, configure and execute multiple analysis tasks of different dimensions. The analysis tasks include at least a first type of task for analyzing electrical quantities, a second type of task for verifying the operating logic of the protection device, and a third type of task for matching historical cases. The action behavior of the protection device is determined based on the results of various analysis tasks; among which: If the result of the first type of task indicates that there are fault characteristics within the protected area of ​​the protection device, and the result of the second type of task indicates that the action of the protection device meets the preset conditions, then the action of the protection device is determined to be correct. If the result of the first type of task indicates that there are fault characteristics within the protection zone of the protection device, but the result of the second type of task indicates that the action of the protection device does not meet the preset conditions, then the action of the protection device is determined to be abnormal. If the result of the first type of task does not detect fault characteristics within the protection zone of the protection device, but the result of the second type of task indicates that the protection device has been activated, then the protection device is determined to have malfunctioned. The diagnostic feature information, judgment conclusions, and results of the third type of task are compared and stored with the historical diagnostic case database. Based on the judgment conclusion and the comparison results with the historical diagnostic case database, a fault diagnosis analysis report is automatically generated.

2. The power plant relay protection fault diagnosis method for centralized control centers according to claim 1, characterized in that, The fault data includes fault waveform files, event sequence records, and protection device action messages.

3. The power plant relay protection fault diagnosis method for centralized control centers according to claim 2, characterized in that, The process of analyzing the relay protection alarm signals and fault data, and reconstructing them by substation and bay to form a fault event sequence with time-series logic, includes: The relay protection alarm signals, fault recording files, event sequence records, and protection device action messages are all time-aligned. Extract the start time of the fault waveform file, the switch change time in the event sequence record, and the logical action time in the protection device action message from the time-aligned data. Based on the start time, switch change time, and logic action time, the relay protection alarm signals, switch change events, and protection logic action events within the same substation and bay are associated in chronological order to generate a fault event sequence.

4. The power plant relay protection fault diagnosis method for centralized control centers according to claim 1, characterized in that, The result of the second type of task indicates that the operation of the protection device meets preset conditions, including: Obtain the actual action value and corresponding action set value of the protection device output by the second type of task; The action time of the protection device output by the second type of task is obtained, as well as the fault initiation time determined from the fault event sequence; If the actual action value is greater than or equal to the action set value, and the time difference between the action time and the fault start time is within the preset protection action time window, then the action of the protection device is determined to meet the preset conditions.

5. The power plant relay protection fault diagnosis method for centralized control centers according to claim 4, characterized in that, The determination that the protection device is functioning correctly includes: Obtain historical cases that match the current fault from the output of the third type of task; If the final conclusion of the matched historical cases is that the action was correct, then the conclusion of this judgment will be marked as high confidence level; If no historical case is matched, or if the final conclusion of the matched historical case is not that the action was correct, the conclusion of this judgment will be marked as the normal confidence level.

6. The power plant relay protection fault diagnosis method for centralized control centers according to claim 1, characterized in that, The determination that the protection device is malfunctioning includes: If the result of the first type of task indicates that there are fault characteristics within the protected area of ​​the protection device, then it is further determined whether the result of the second type of task includes the action signal of the protection device; If the result of the second type of task does not contain any action signal of the protection device, then the protection device is determined to be inactive.

7. The power plant relay protection fault diagnosis method for centralized control centers according to claim 6, characterized in that, The determination of abnormal operation of the protection device also includes: If the result of the second type of task includes the action signal of the protection device, then obtain the actual action value and action set value corresponding to the action signal, as well as the corresponding action time and fault start time. If the actual action value is less than the action set value, the protection device is determined to be malfunctioning due to failure to reach the set value. If the actual action value is greater than or equal to the action set value, but the time difference between the action time and the fault start time is not within the preset protection action time window, then the protection device is determined to be logically abnormal due to an action timing error.

8. The power plant relay protection fault diagnosis method for centralized control centers according to claim 1, characterized in that, The determination of the malfunction of the protection device includes: Based on the results of the first type of task, determine the location of the fault characteristics; If the fault characteristics are not detected in the protection zone of the protection device, and the fault characteristics are not detected in the adjacent backup protection zone, then the protection device is determined to be a faultless malfunction. If the fault characteristics are detected outside the protection zone of the protection device, but are detected in the adjacent backup protection zone, then the protection device is determined to be a false trip due to an external fault.

9. The power plant relay protection fault diagnosis method for centralized control centers according to claim 1, characterized in that, The step of comparing the diagnostic feature information, judgment conclusion, and the results of the third type of task with the historical diagnostic case database includes: Based on the fault type, faulty device identifier and fault location in the feature information, a search is performed in the historical diagnostic case database to obtain a set of candidate cases. Calculate the similarity between the feature information of each historical case in the candidate case set and the feature information of the current diagnosis; Based on the similarity and a preset similarity threshold, a matching result is determined from the candidate case set. The matching result includes an indication of a matched historical case or a historical case that was not matched.

10. The power plant relay protection fault diagnosis method for centralized control centers according to claim 9, characterized in that, The process of storing the diagnostic feature information, judgment conclusion, and the results of the third type of task with the historical diagnostic case database includes: The storage mode of this diagnostic record is determined based on the matching results; If the storage mode is a new storage mode, the complete record of this diagnosis will be stored as a new case in the historical diagnosis case library. If the storage mode is an update storage mode, then update the call count and last call time information of the matched historical cases; If the storage mode is an associated storage mode, a case record associated with the matched historical case is created for this diagnosis and stored in the historical diagnosis case library.

Images