Automatic relay protection checking method based on sequential logic

Abstract

The invention relates to the technical field of power system safety, in particular to an automatic checking method for a relay protection device based on sequential logic, which comprises the following steps of: acquiring power grid operation state data in real time, storing a state sequential sequence through an annular buffer area, calculating related characteristics, performing sequential logic verification on the state sequential sequence by adopting a bounded model, and checking the relay protection device based on the sequential logic. Generating a structured trigger signal; and matching a diagnosis rule base based on the trigger signal and the characteristic value, automatically obtaining a check parameter, carrying out check through threshold comparison and Spearman rank correlation coefficient analysis, outputting a check report, and updating the diagnosis rule base. The check response is improved to the second level, the diagnosis accuracy and efficiency are improved, and the resource consumption is reduced.

Description

Technical Field

[0001] This invention relates to the field of power system safety technology, and in particular to an automatic verification method for relay protection devices based on timing logic. Background Technology

[0002] In the safe and stable operation of power systems, relay protection devices play a crucial role as safety guardians, and their reliability directly affects the rapid isolation of grid faults and the safety of equipment. Therefore, regular verification of relay protection devices is a core aspect of power operation and maintenance.

[0003] Currently, the mainstream verification methods mainly rely on periodic maintenance and online monitoring technology based on fixed thresholds, as detailed below:

[0004] Regular maintenance mode: Regular maintenance is usually carried out on an annual or monthly cycle. This method has obvious response lag and cannot detect hidden faults or setpoint drift of protection devices between two maintenance cycles in a timely manner, which poses a hidden danger to power grid safety. In addition, regular maintenance requires indiscriminate inspection of all equipment, including a large number of normally operating equipment, resulting in a huge waste of human and material resources.

[0005] Online monitoring technology based on fixed thresholds: Although online monitoring technology can collect data in real time, its triggering mechanism mostly relies on the threshold exceeding the limit at a single time point (such as the current instantaneously exceeding the set value). The diagnostic strategy is fixed and relies on human experience to judge. It cannot effectively identify complex fault symptoms with time-series characteristics such as "current slowly rises and continues" or "voltage fluctuation decreases", resulting in a high false alarm and missed alarm rate. Summary of the Invention

[0006] Therefore, the technical problem to be solved by the present invention is to overcome the problem that the prior art cannot accurately and efficiently identify faults in relay protection devices.

[0007] To address the aforementioned technical problems, this invention provides an automatic verification method for relay protection devices based on timing logic, comprising: Real-time acquisition of power grid operation status data; A circular buffer is constructed to store the power grid operation status data as a state time sequence of a preset length on the circular buffer, and the feature values ​​of the power grid operation status data are calculated. Perform timing logic verification on the state timing sequence; if the timing logic verification result is true, generate a structured trigger signal; Based on the feature values ​​of the trigger signal and the power grid operating status data, a diagnostic rule base is matched to select a diagnostic strategy; Based on the diagnostic strategy, the parameters that the relay protection device needs to be verified are obtained, and a verification report is output. Update the diagnostic rule base based on the verification report.

[0008] Preferably, the method for acquiring real-time power grid operation status data is as follows: A communication connection is established with the relay protection device through the IEC 61850 protocol, and the power grid operation status data is obtained according to a predefined data point table using a cyclic reading method.

[0009] Preferably, the types of power grid operating status data include: current, voltage, frequency, and switching status.

[0010] Preferably, in the state time sequence, the data at each time point includes a timestamp, a data value, and a quality bit.

[0011] Preferably, the temporal logic verification of the state time sequence includes the following steps: Based on a preset linear time-series logic formula, the true or false value of the data at each time point in the state time-series is determined, and the true or false values ​​corresponding to all time points are combined in time sequence to obtain a finite state sequence. If the finite state sequence satisfies the triggering condition of the sequential logic verification, then the verification result of the sequential logic verification is true; otherwise, the verification result of the sequential logic verification is false.

[0012] Preferably, after performing temporal logic verification on the state-time sequence, an auxiliary judgment is made on the state-time sequence whose verification result is true. The method is as follows: The similarity between the state time series sequence and the historical fault template sequence is calculated using the dynamic time warping algorithm. If the similarity is greater than a preset threshold, the diagnostic priority of the abnormal event corresponding to the state time series sequence is increased.

[0013] Preferably, the structured trigger signal includes: The comprehensive confidence score is calculated based on the abnormal event type determined by the state time sequence type, the trigger timestamp, the data snapshot of the abnormal event triggering event, the verification result based on the time sequence logic verification, and the result of auxiliary judgment.

[0014] Preferably, a diagnostic strategy is selected by matching the feature values ​​of the trigger signal and the power grid operating status data in a diagnostic rule base. The method is as follows: Calculate the feature values ​​of the trigger signal and power grid operation status data and the rule confidence score of each diagnostic rule in the diagnostic rule base; the rule confidence score is calculated based on the matching degree between the trigger signal and the diagnostic rule and the rule confidence weight; If the rule confidence score of the trigger signal and the diagnostic rule is greater than the activation threshold corresponding to the diagnostic rule, then the diagnostic strategy corresponding to the diagnostic rule is selected.

[0015] Preferably, after obtaining and verifying the parameters that the relay protection device needs to be checked, the Spearman rank correlation coefficient is used to determine whether the changing trend of the parameters that need to be checked is consistent with the aging trend of the equipment. If yes, it is considered normal; if no, it is determined that there is an abnormal risk.

[0016] Preferably, the diagnostic rule base is updated based on the verification report, and the method includes: online learning and offline learning; The online learning process involves updating the rule confidence weight of the diagnostic rule in the diagnostic rule base using the success rate and false alarm rate of the diagnostic rule in the verification report. The offline learning process involves updating the diagnostic rules in the diagnostic rule base based on the verification report using a rule mining algorithm.

[0017] Compared with the prior art, the above-described technical solution of the present invention has the following advantages: The automatic verification method for relay protection devices based on timing logic described in this invention verifies the timing logic of the state timing sequence through bounded model checks. This method can accurately capture timing patterns containing fault information, such as "continuous over-limit" and "excessive rate of change," in the power grid operation status data. This is fundamentally different from the single threshold point judgment in the prior art. It ensures that the verification process is triggered only when a real anomaly occurs, overcoming the lag of periodic verification. It can shorten the verification response from "days" to "seconds," while avoiding frequent interference with normal equipment and reducing verification resource consumption.

[0018] Furthermore, this invention integrates the verification results based on time-series logic with the historical fault matching results based on the dynamic time warping algorithm through a multi-dimensional auxiliary judgment mechanism. The dynamic time warping algorithm can effectively overcome the stretching and distortion of time series on the time axis and quantitatively assess the morphological similarity between the current sequence and the historical fault template. This dual verification mechanism greatly reduces the risk of misjudgment caused by noise interference or the failure of a single judgment criterion, and greatly improves the diagnostic accuracy.

[0019] Furthermore, this invention matches trigger signals constructed based on abnormal state time sequences within a diagnostic rule base, and selects the optimal diagnostic strategy according to the matching degree and rule confidence weights. The entire process requires no manual intervention, automatically completing the entire flow from signal triggering to report generation, significantly improving verification efficiency. Moreover, in the parameter verification stage, this invention not only performs simple threshold comparisons but also introduces the Spearman rank correlation coefficient to analyze the consistency between the drift trend of the set value and the aging trend of the equipment, thereby enabling earlier detection of potential risks.

[0020] Finally, this invention constructs a complete data closed loop of "perception-decision-execution-learning," enabling the verification process to continuously optimize. It adjusts rule confidence weights in real time through online learning and conducts in-depth analysis of error cases through offline learning to uncover new association rules to optimize the diagnostic rule base. This invention can continuously learn from operational practices, adapt to changes in the power grid structure, and ultimately form a virtuous cycle of continuous performance improvement—something that existing static systems cannot match. Attached Figure Description

[0021] To make the content of this invention easier to understand, the invention will be further described in detail below with reference to specific embodiments and accompanying drawings, wherein: Figure 1 This is a flowchart of an automatic verification method for relay protection devices based on timing logic according to the present invention. Detailed Implementation

[0022] The present invention will be further described below with reference to the accompanying drawings and specific embodiments, so that those skilled in the art can better understand and implement the present invention. However, the embodiments described are not intended to limit the present invention.

[0023] Reference Figure 1 As shown, the present invention provides an automatic verification method for relay protection devices based on timing logic, comprising:

[0024] S1: System initialization.

[0025] The system initialization includes loading a preset protection device status monitoring rule base and a diagnostic rule base. These rule bases are stored in XML format and contain fields such as rule ID, trigger condition expression, associated diagnostic strategy ID, and confidence weight.

[0026] S2: Real-time acquisition of power grid operation status data.

[0027] This embodiment establishes a communication connection with the relay protection device through the IEC 61850 protocol and uses a cyclic reading method to obtain power grid operation status data according to a predefined point list.

[0028] It should be noted that the types of power grid operation status data include real-time monitoring parameters such as current, voltage, frequency, and switch status. The data point table is predefined by the administrator according to the actual configuration of the power grid.

[0029] S3: Construct a circular buffer to store the power grid operation status data as a state time sequence of a preset length on the circular buffer.

[0030] In this embodiment, a ring buffer is established as a time-series database to store the state time-series sequence of the most recent 300 time points (corresponding to 30 seconds of data, based on a 100ms sampling interval). The data at each time point includes a timestamp, data value, and quality bits. The design of the above time window size aims to balance the duration of typical power grid transient faults (usually a few seconds to tens of seconds) with system memory overhead.

[0031] Furthermore, this step calculates characteristic values ​​such as the first derivative (rate of change) and moving average of the stored raw data in real time, and stores the calculated multidimensional characteristic values ​​together with the raw data in a circular buffer.

[0032] The formulas for calculating the first derivative (D) and the moving average (MA) are as follows: ; ; in, The first derivative, i.e., the rate of change of current, is used to quantify the drasticness of the current change; t is the time point. and The values ​​are the current values ​​at time points t and t-1, respectively. The time interval is the sampling interval between two adjacent time points, which is set to 100ms in this embodiment; It is a moving average value used to smooth out instantaneous noise and reflect the overall trend of current change; The value of is the size of the sliding window, which is 5 in this embodiment, i.e., a sliding window of 5 × 100 ms = 500 ms; j is the summation index variable, specifically representing the backward movement of j time points from the current time point t (j = 0 to j = M−1). Let tj be the current value at time point tj.

[0033] The first derivative is used to characterize the rate of change of the power grid state, and the moving average is used to smooth data fluctuations, providing multi-dimensional feature inputs for subsequent time-series logic judgments.

[0034] S4: Perform timing logic verification on the state timing sequence; if the verification result of the timing logic verification is true, generate a structured trigger signal.

[0035] This invention is based on the principle of bounded model checking to perform temporal logic verification on state-time sequences. The steps include: S401: Based on a preset linear time-series logic formula, the true or false value of the data at each time point in the state time-series is determined, and the true or false values ​​corresponding to all time points are combined in time sequence to obtain a finite state sequence.

[0036] This step first parses the pre-defined linear sequential logic formula; the atomic propositions in the formula (such as...) The true or false value of ) is determined by the system querying the preset threshold (such as ) in the status monitoring rule base. ) and real-time data in the circular buffer The determination is made after comparison.

[0037] For example, at time point t, if the current value If the atomic proposition is true at time point t, then it is true; otherwise, it is false.

[0038] This step combines the truth values ​​of all atomic propositions at the N most recent time points in the circular buffer in temporal order to obtain a finite state sequence. Each state A snapshot of the state at the corresponding time point ti.

[0039] S402: If the finite state sequence satisfies the triggering condition of the sequential logic verification, then the verification result of the sequential logic verification is true; otherwise, the verification result of the sequential logic verification is false.

[0040] This step uses a bounded model checking algorithm to verify the preset triggering conditions.

[0041] For example, in the typical scenario of "current exceeding the limit for three consecutive sampling cycles", the formal expression for setting the trigger condition for its timing logic verification is as follows: , making ; in, It is an existential quantifier, indicating "at least one exists"; As a loop variable, in a finite sequence of states The starting position index for matching is N; N is the length of the finite state sequence, i.e., the length of the state time sequence in the circular buffer; t is the current latest time point. The current value at time point tk; The current protection setting is preset in the condition monitoring rule base; This is the logical AND operator, representing "and"; Traversing a finite sequence of states For all possible starting positions k, the trigger condition for executing the sequential logic verification is determined; if at least one k satisfies the condition, the verification result of the sequential logic verification is true; otherwise, the verification result of the sequential logic verification is false.

[0042] Preferably, after performing temporal logic verification on the state-time sequence, a multi-dimensional, high-confidence auxiliary judgment is performed on the state-time sequence whose verification result is true. The method is as follows: The similarity between the state time series sequence and the historical fault template sequence is calculated using the dynamic time warping algorithm. If the similarity is greater than a preset threshold, the diagnostic priority of the abnormal event corresponding to the state time series sequence is increased.

[0043] Specifically, a dynamic time warping algorithm is activated to calculate the similarity between the current computational state time series sequence and the historical fault template sequence. This dynamic time warping algorithm overcomes the stretching and distortion of the sequence along the time axis, matching the morphological similarity between the real-time monitored fault symptom sequence and the historical fault template sequence, thereby providing diagnostic auxiliary information with higher confidence.

[0044] If the similarity is greater than a preset threshold (i.e., a high match), a high-priority diagnostic event is immediately triggered. In this embodiment, the preset threshold is 80%. This value is a balance point achieved between the false positive rate and the false negative rate after extensive testing of the historical fault database. A threshold below 80% will significantly increase the false positive rate, increasing the operational burden; while a threshold above 85% will increase the false negative rate, failing to effectively warn of some potential faults.

[0045] If the similarity does not reach the threshold, a routine diagnostic event is triggered.

[0046] In one embodiment of the present invention, sequential logic verification and auxiliary judgment can be performed simultaneously. Finally, the results of sequential logic verification and auxiliary judgment, along with their respective confidence scores, are recorded to determine the diagnostic priority for subsequent iterative optimization of the diagnostic rule base.

[0047] The structured trigger signal includes: an anomaly event type determined according to the type of the state time sequence, a trigger timestamp, a data snapshot of the anomaly event triggering event, a verification result based on time-series logic verification, and a comprehensive confidence score calculated from the results of auxiliary judgment. Details are as follows:

[0048] `event_type` (Abnormal Event Type): Automatically determined based on the atomic proposition or sequential logic condition that triggers the abnormality. For example, if the verification involves continuous current exceeding the limit, the abnormal event type is "Continuous Current Abnormality". The generation method is to label events based on the content of matching atomic propositions (such as current, voltage, etc.) and the duration (such as three consecutive cycles) in the state sequence.

[0049] For example, when the detected current value exceeds for three consecutive cycles The system automatically sets event_type to "current continuity anomaly".

[0050] Timestamp (trigger timestamp): Records the specific moment the event was triggered using the current system time or data acquisition timestamp. It is generated by directly obtaining the system clock or data timestamp.

[0051] For example, when the trigger occurs, the system records the current time as "2023-10-01 12:00:00.000".

[0052] `data_snapshot`: Extracts a sequence of state data related to the trigger moment from the circular buffer. It is generated by copying the original data and calculated features (such as current value sequences) from the most recent N time points in the circular buffer.

[0053] For example, for current anomalies, `data_snapshot` might contain a sequence of current values ​​from the most recent 30 time points (3 seconds of data), such as... .

[0054] confidence_score (overall confidence score): Calculated based on the results of time-series logic verification and auxiliary judgment, primarily relying on: If the sequential logic verification is true and the similarity of the auxiliary judgment is high, the score is high (for example, 80% similarity in the auxiliary judgment corresponds to a confidence level of 0.8). If only the temporal logic verification is true, the score will be lower (e.g., a base confidence level of 0.5, which will be adjusted according to the specific matching degree).

[0055] Optionally, historical confidence data may also affect the overall confidence score.

[0056] For example, if the sequential logic verification is true and the similarity of the auxiliary judgment is 85%, the confidence score might be calculated as 0.85; if only the sequential logic is true, it might be 0.6. This score reflects the reliability of the multi-dimensional judgment.

[0057] This invention employs a bounded model checking method, which clarifies that verification is performed on a finite-length state time sequence S, resolving the contradiction between infinite paths and finite buffers, and conforming to the actual situation of real-time processing in industrial systems.

[0058] S5: Match the trigger signal and the feature values ​​of the power grid operation status data in the diagnostic rule base and select a diagnostic strategy.

[0059] Specifically, a diagnostic strategy is selected by matching the trigger signal in the diagnostic rule base, and the method is as follows: The Rete algorithm is used to calculate the rule confidence score between the trigger signal and each diagnostic rule in the diagnostic rule base; the rule confidence score is calculated based on the matching degree between the trigger signal and the diagnostic rule and the rule confidence weight. If the rule confidence score of the trigger signal and the diagnostic rule is greater than the activation threshold corresponding to the diagnostic rule, then the diagnostic strategy corresponding to the diagnostic rule is selected.

[0060] When multiple diagnostic rules are matched in the diagnostic rule base, the diagnostic strategies corresponding to the multiple diagnostic rules are activated to form a comprehensive diagnostic solution, and a diagnostic strategy identifier (such as "overload protection verification strategy ID001") and a confidence score (such as "0.85", i.e. 85% confidence) are generated.

[0061] The diagnostic strategy identifier is a unique identifier for matching rules in the diagnostic rule base, used to indicate the specific strategy that should be executed during the current verification.

[0062] The following are typical diagnostic strategies and their identifiers: Overload protection verification strategy: Overload protection verification strategy ID001; Low voltage protection verification strategy: Low voltage protection verification strategy ID002; Frequency protection verification strategy: Frequency protection verification strategy ID003.

[0063] It should be noted that the diagnostic rule base is constructed as a production rule system, containing "if-then" rules, for example: If the duration of the current anomaly is greater than 50ms and the rate of change is greater than 20% / s, then the overload protection verification strategy is matched; the value of 50ms is based on the typical operating time threshold of the relay protection device in the power system; the value of 20% / s is based on the experience of power grid operation that a current change rate exceeding 20% / s is usually an abnormal situation. If the voltage drop exceeds 10% and the recovery time is greater than 100ms, the "low voltage protection verification strategy" will be applied. The 10% value is based on the industry standard threshold for voltage drop in the power grid, and the 100ms value is based on the typical recovery time of the voltage drop.

[0064] If the frequency fluctuation exceeds 0.5Hz and the duration is greater than 200ms, the "frequency protection verification strategy" will be matched; the value of 0.5Hz is based on the industry standard threshold for power grid frequency fluctuation; the value of 200ms is based on the typical duration of power grid frequency fluctuation.

[0065] S6: Based on the diagnostic strategy, obtain the parameters that the relay protection device needs to be verified, perform verification, and output a verification report.

[0066] This embodiment calls the Get_Setting() function in the SDK provided by the relay protection device supplier to obtain the real-time set value of the parameter that needs to be checked, and automatically determines the standard set value range of the parameter that needs to be checked according to the diagnostic strategy, so as to verify whether the real-time set value is within the standard set value range.

[0067] Furthermore, in addition to comparing the standard setpoint range, this embodiment also verifies the consistency of the parameter trends, including: After obtaining and verifying the parameters that the relay protection device needs to be checked, the Spearman rank correlation coefficient is used to determine whether the changing trend of the parameters that need to be checked is consistent with the aging trend of the equipment. If yes, it is considered normal; if no, it is determined that there is an abnormal risk.

[0068] If a function call fails or a communication timeout occurs, the system will automatically retry 3 times. If it still fails, a "communication error" alarm will be generated, and the current verification process will be terminated to proceed to the next step.

[0069] It should be noted that when the Spearman rank correlation coefficient is greater than 0.8, the trend of the parameter change is consistent with the aging trend of the equipment and is considered normal; when the correlation coefficient is less than 0.5, it is determined that there is an abnormal risk.

[0070] The Spearman rank correlation coefficient threshold is set based on statistical analysis of a large amount of historical equipment aging data and setpoint drift data. The analysis shows that when the coefficient is greater than 0.8, the consistency (accuracy) between the diagnostic conclusion (considered normal) and the actual situation exceeds 95%. When the coefficient is less than 0.5, the detection rate of abnormal risks is significantly improved. Thresholds of 0.8 and 0.5 represent the best balance between ensuring high accuracy and maintaining high detection sensitivity.

[0071] S7: Update the diagnostic rule base based on the verification report.

[0072] If the verification result is normal, a "Verification Passed" report will be generated, which includes information such as verification time, verification parameters, and power grid status. If the verification result is abnormal, a "Verification Required" report will be generated, which includes information such as abnormal parameters, possible causes, and suggested verification methods.

[0073] It should be noted that the diagnostic report is automatically appended with a timestamp and system identifier to ensure traceability. The system automatically dispatches "required verification" reports to designated maintenance personnel and sets priorities.

[0074] Preferably, the diagnostic rule base is updated based on the verification report, and the method includes online learning and offline learning. For cases that are successfully verified, the confidence weight of the rule that triggers the diagnostic strategy is automatically adjusted; for cases with errors, they are stored in the error case base and used to provide data for subsequent rule mining.

[0075] The online learning process involves updating the rule confidence weight of the diagnostic rule in the diagnostic rule base using the success rate and false alarm rate of the diagnostic rule in the verification report.

[0076] The formula for adjusting the confidence weights of rules is: ;in, The learning rate (ranging from 0.1 to 3). The updated rule confidence weights, The confidence weights of the rules before the update.

[0077] The offline learning process involves updating the diagnostic rules in the diagnostic rule base based on the verification report using a rule mining algorithm.

[0078] Specifically, the offline learning process can be initiated periodically (e.g., quarterly). The specific method is as follows: First, use rule mining algorithms such as Apriori to analyze the error case library, mine the association rules of error features, and generate new candidate diagnostic rules or correction suggestions; second, these candidate diagnostic rules or correction suggestions can only be officially updated to the diagnostic rule library after being confirmed by the administrator; finally, the updated diagnostic rule library also needs to be cross-validated and an optimization report needs to be generated.

[0079] It should be noted that this step involves in-depth offline learning and system optimization based on the historical case library, ensuring the continuous optimization capability of the verification.

[0080] The automatic verification method for relay protection devices based on timing logic described in this invention verifies the timing logic of the state timing sequence through bounded model checks. This method can accurately capture timing patterns containing fault information, such as "continuous over-limit" and "excessive rate of change," in the power grid operation status data. This is fundamentally different from the single threshold point judgment in the prior art. It ensures that the verification process is triggered only when a real anomaly occurs, overcoming the lag of periodic verification. It can shorten the verification response from "days" to "seconds," while avoiding frequent interference with normal equipment and reducing verification resource consumption.

[0081] Furthermore, this invention integrates the verification results based on time-series logic with the historical fault matching results based on the dynamic time warping algorithm through a multi-dimensional auxiliary judgment mechanism. The dynamic time warping algorithm can effectively overcome the stretching and distortion of time series on the time axis and quantitatively assess the morphological similarity between the current sequence and the historical fault template. This dual verification mechanism greatly reduces the risk of misjudgment caused by noise interference or the failure of a single judgment criterion, and greatly improves the diagnostic accuracy.

[0082] Furthermore, this invention matches trigger signals constructed based on abnormal state time sequences within a diagnostic rule base, and selects the optimal diagnostic strategy according to the matching degree and rule confidence weights. The entire process requires no manual intervention, automatically completing the entire flow from signal triggering to report generation, significantly improving verification efficiency. Moreover, in the parameter verification stage, this invention not only performs simple threshold comparisons but also introduces the Spearman rank correlation coefficient to analyze the consistency between the drift trend of the set value and the aging trend of the equipment, thereby enabling earlier detection of potential risks.

[0083] Finally, this invention constructs a complete data closed loop of "perception-decision-execution-learning," enabling the verification process to continuously optimize. It adjusts rule confidence weights in real time through online learning and conducts in-depth analysis of error cases through offline learning to uncover new association rules to optimize the diagnostic rule base. This invention can continuously learn from operational practices, adapt to changes in the power grid structure, and ultimately form a virtuous cycle of continuous performance improvement—something that existing static systems cannot match.

[0084] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0085] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0086] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0087] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0088] Obviously, the above embodiments are merely illustrative examples for clear explanation and are not intended to limit the implementation. Those skilled in the art will recognize that other variations or modifications can be made based on the above description. It is neither necessary nor possible to exhaustively list all possible implementations here. However, obvious variations or modifications derived therefrom are still within the scope of protection of this invention.

Claims

1. An automatic verification method for relay protection devices based on timing logic, characterized in that, include: Real-time acquisition of power grid operation status data; A circular buffer is constructed to store the power grid operation status data as a state time sequence of a preset length on the circular buffer, and the feature values ​​of the power grid operation status data are calculated. Perform temporal logic verification on the state-time sequence; If the result of the sequential logic verification is true, a structured trigger signal is generated. Based on the feature values ​​of the trigger signal and the power grid operating status data, a diagnostic rule base is matched to select a diagnostic strategy; Based on the diagnostic strategy, the parameters that the relay protection device needs to be verified are obtained, and a verification report is output. Update the diagnostic rule base based on the verification report.

2. The automatic verification method for a relay protection device based on timing logic according to claim 1, characterized in that, The method for obtaining real-time power grid operation status data is as follows: A communication connection is established with the relay protection device through the IEC 61850 protocol, and the power grid operation status data is obtained according to a predefined data point table using a cyclic reading method.

3. The automatic verification method for a relay protection device based on timing logic according to any one of claims 1 or 2, characterized in that, The types of power grid operation status data include: current, voltage, frequency, and switch status.

4. The automatic verification method for a relay protection device based on timing logic according to claim 1, characterized in that, In the state time sequence, the data at each time point includes a timestamp, data value, and quality bits.

5. The automatic verification method for a relay protection device based on timing logic according to claim 1, characterized in that, The steps for performing time-series logic verification on state-series sequences include: Based on a preset linear time-series logic formula, the true or false value of the data at each time point in the state time-series is determined, and the true or false values ​​corresponding to all time points are combined in time sequence to obtain a finite state sequence. If the finite state sequence satisfies the triggering condition of the sequential logic verification, then the verification result of the sequential logic verification is true; otherwise, the verification result of the sequential logic verification is false.

6. The automatic verification method for a relay protection device based on timing logic according to claim 1, characterized in that, After performing temporal logic verification on the state-time sequence, an auxiliary judgment is made for the state-time sequence whose verification result is true. The method is as follows: The similarity between the state time series sequence and the historical fault template sequence is calculated using the dynamic time warping algorithm. If the similarity is greater than a preset threshold, the diagnostic priority of the abnormal event corresponding to the state time series sequence is increased.

7. The automatic verification method for a relay protection device based on timing logic according to claim 6, characterized in that, The structured trigger signal includes: The comprehensive confidence score is calculated based on the abnormal event type determined by the state time sequence type, the trigger timestamp, the data snapshot of the abnormal event triggering event, the verification result based on the time sequence logic verification, and the result of auxiliary judgment.

8. The automatic verification method for a relay protection device based on timing logic according to claim 1, characterized in that, Based on the feature values ​​of the trigger signal and the power grid operating status data, a diagnostic rule base is matched to select a diagnostic strategy. The method is as follows: Calculate the feature values ​​of the trigger signal and power grid operation status data and the rule confidence score of each diagnostic rule in the diagnostic rule base; the rule confidence score is calculated based on the matching degree between the trigger signal and the diagnostic rule and the rule confidence weight; If the rule confidence score of the trigger signal and the diagnostic rule is greater than the activation threshold corresponding to the diagnostic rule, then the diagnostic strategy corresponding to the diagnostic rule is selected.

9. The automatic verification method for a relay protection device based on timing logic according to claim 1, characterized in that, After obtaining and verifying the parameters that the relay protection device needs to be checked, the Spearman rank correlation coefficient is used to determine whether the changing trend of the parameters that need to be checked is consistent with the aging trend of the equipment. If yes, it is considered normal; if no, it is determined that there is an abnormal risk.

10. The automatic verification method for a relay protection device based on timing logic according to claim 1, characterized in that, The diagnostic rule base is updated based on the verification report, and the methods include online learning and offline learning. The online learning process involves updating the rule confidence weight of the diagnostic rule in the diagnostic rule base using the success rate and false alarm rate of the diagnostic rule in the verification report. The offline learning process involves updating the diagnostic rules in the diagnostic rule base based on the verification report using a rule mining algorithm.

Images