Self-adaptive reclosing control method and system for distribution automation feeder terminal

Abstract

The invention discloses a self-adaptive reclosing control method and system for a distribution automation feeder terminal, and relates to the technical field of intelligent power distribution, and the method comprises the steps: carrying out the real-time monitoring of a distribution network line, obtaining a line monitoring sequence, synchronously obtaining a terminal state sequence, carrying out the abnormal interference correction of a terminal, and obtaining a line operation feature sequence; building a reclosing locking detection space, and inputting the line operation characteristic sequence into the reclosing locking detection space to obtain a reclosing locking coefficient; if the reclosing locking coefficient is smaller than the reclosing locking threshold value, multi-dimensional fault prediction is carried out on the distribution network line, a line fault feature graph is generated, protection tripping probability prediction is carried out on the distribution network line, and the line protection tripping probability is determined; and performing self-adaptive reclosing control by combining the line protection tripping probability and the feeder terminal. The technical problems that reclosing locking judgment is inaccurate and reclosing control lacks adaptivity in the prior art are solved, and the technical effect of improving reclosing action decision accuracy and reliability is achieved.

Description

Technical Field

[0001] This invention relates to the field of intelligent power distribution technology, and more specifically to an adaptive reclosing control method and system for power distribution automation feeder terminals. Background Technology

[0002] Whether a distribution network line performs reclosing after a fault has a significant impact on power supply reliability. In actual operation, feeder terminals need to determine whether reclosing is feasible based on line status, fault characteristics, and various blocking conditions. Due to the complex operating environment of lines, monitoring parameters often include noise, abnormal interference, or sudden changes in state, making it difficult to comprehensively identify blocking factors and easily leading to false or missed blocking. Furthermore, traditional reclosing strategies often rely on fixed logic or single-feature triggers, making it difficult to dynamically adjust the control method according to the actual operating characteristics of the line. This often results in a lack of adaptability in reclosing actions, limiting both stability and safety. Summary of the Invention

[0003] This application provides an adaptive reclosing control method and system for feeder terminals in power distribution automation, which addresses the technical problems of inaccurate reclosing interlocking determination and lack of adaptability in existing technologies.

[0004] In view of the above problems, this application provides an adaptive reclosing control method and system for feeder terminals in distribution automation.

[0005] The first aspect of this application provides an adaptive reclosing control method for a feeder terminal block in a distribution automation system, the method comprising: Real-time monitoring of the distribution network lines by the feeder terminal is performed to obtain the line monitoring sequence, and the terminal status sequence of the feeder terminal is acquired simultaneously. Terminal anomaly interference correction is performed on the line monitoring sequence based on the terminal status sequence to obtain the line operation characteristic sequence. Reliable mining of the operation characteristics of the distribution network lines is performed based on the multi-dimensional factors of reclosing blocking to build a reclosing blocking detection space. The line operation characteristic sequence is input into the reclosing blocking detection space to obtain the reclosing blocking coefficient. If the reclosing blocking coefficient is less than the reclosing blocking threshold, multi-dimensional fault prediction is performed on the distribution network lines based on the line operation characteristic sequence to generate a line fault feature map. The protection trip probability is predicted for the distribution network lines based on the line fault feature map to determine the line protection trip probability. An adaptive reclosing mechanism is introduced, combining the line protection trip probability and the feeder terminal for adaptive reclosing control.

[0006] A second aspect of this application provides an adaptive reclosing control system for a distribution automation feeder terminal, the system comprising: The system includes the following modules: a real-time monitoring module for monitoring distribution network lines in real time based on feeder terminals, obtaining line monitoring sequences, and synchronously acquiring terminal status sequences of the feeder terminals; an interference correction module for performing terminal anomaly interference correction on the line monitoring sequences based on the terminal status sequences, obtaining line operation characteristic sequences; a mining module for performing reliable mining of operation characteristics of the distribution network lines based on multi-dimensional factors of reclosing blocking, building a reclosing blocking detection space, and inputting the line operation characteristic sequences into the reclosing blocking detection space to obtain reclosing blocking coefficients; a fault prediction module for performing multi-dimensional fault prediction on the distribution network lines based on the line operation characteristic sequences if the reclosing blocking coefficients are less than the reclosing blocking threshold, generating a line fault feature map; a probability prediction module for predicting the protection trip probability of the distribution network lines based on the line fault feature map, determining the line protection trip probability; and a control module for introducing an adaptive reclosing mechanism, combining the line protection trip probability and the feeder terminals for adaptive reclosing control.

[0007] One or more technical solutions provided in this application have at least the following technical effects or advantages: This application performs real-time monitoring of distribution network lines using feeder terminals to obtain line monitoring sequences and simultaneously acquires the terminal status sequences of the feeder terminals. Based on the terminal status sequences, it performs terminal anomaly interference correction on the line monitoring sequences to obtain line operation characteristic sequences. It then performs reliable mining of the operation characteristics of the distribution network lines based on multi-dimensional factors of reclosing blocking, constructs a reclosing blocking detection space, and inputs the line operation characteristic sequences into the reclosing blocking detection space to obtain reclosing blocking coefficients. If the reclosing blocking coefficients are less than the reclosing blocking threshold, it performs multi-dimensional fault prediction on the distribution network lines based on the line operation characteristic sequences to generate a line fault feature map. Based on the line fault feature map, it predicts the protection trip probability of the distribution network lines to determine the line protection trip probability. Finally, it introduces an adaptive reclosing mechanism, combining the line protection trip probability and the feeder terminals for adaptive reclosing control. This invention addresses the technical problems of inaccurate reclosing interlock determination and lack of adaptability in existing technologies. By constructing a reclosing interlock detection space to obtain the interlock coefficient, and combining fault prediction and protection tripping probability, adaptive reclosing control is achieved, thereby improving the accuracy and reliability of reclosing action decision-making. Attached Figure Description

[0008] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0009] Figure 1 A schematic flowchart of an adaptive reclosing control method for a power distribution automation feeder terminal provided in an embodiment of this application; Figure 2 This is a schematic diagram of the adaptive reclosing control system for a power distribution automation feeder terminal provided in an embodiment of this application.

[0010] Explanation of reference numerals in the attached diagram: Real-time monitoring module 11, Interference correction module 12, Mining module 13, Fault prediction module 14, Probability prediction module 15, Control module 16. Detailed Implementation

[0011] This application provides an adaptive reclosing control method and system for distribution automation feeder terminals. It addresses the technical problems of inaccurate reclosing interlocking determination and lack of adaptability in existing technologies. By constructing a reclosing interlocking detection space to obtain the interlocking coefficient, and combining fault prediction and protection tripping probability, adaptive reclosing control is achieved, thereby improving the accuracy and reliability of reclosing action decision-making.

[0012] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0013] It should be noted that any variation of the terms "comprising" and "having" is intended to cover non-exclusive inclusion, for example, a process, method, system, product, or server that includes a series of steps or units is not necessarily limited to those steps or units that are explicitly listed, but may include other steps or modules that are not explicitly listed or that are inherent to such processes, methods, products, or devices.

[0014] Example 1, as Figure 1 As shown, this application provides an adaptive reclosing control method for feeder terminals in distribution automation, the method comprising: Step S100: Real-time monitoring of the distribution network lines is performed based on the feeder terminal to obtain the line monitoring sequence, and the terminal status sequence of the feeder terminal is obtained synchronously.

[0015] In this embodiment, the feeder terminal continuously collects operational parameters such as current, voltage, zero-sequence current, zero-sequence voltage, and frequency of the distribution network line, recording the changes of each electrical parameter point by point using a fixed sampling period. By performing basic amplitude calculations and waveform processing on the collected electrical signals, a line monitoring sequence reflecting the changes in the line's operating status over time is obtained.

[0016] Simultaneously, during real-time line monitoring, the feeder terminal records its own operational status, including remote or local location status, fault handling activation / deactivation status, protection activation / deactivation status, communication status, self-test status, and manual operation status. By recording these status changes chronologically, a terminal status sequence is formed. This terminal status sequence reflects the working conditions of the feeder terminal within the monitoring period.

[0017] Preferably, the feeder terminal is a feeder detection terminal with detection function.

[0018] Step S200: Perform terminal anomaly interference correction on the line monitoring sequence based on the terminal status sequence to obtain the line operation characteristic sequence.

[0019] In this embodiment, when correcting for terminal anomaly interference in the line monitoring sequence based on the terminal status sequence, the line monitoring sequence is first marked with status indicators for the corresponding time period based on the fault handling activation / deactivation status, protection activation / deactivation status, remote or local location status, communication status, self-test status, and manual operation status recorded in the terminal status sequence, in order to identify monitoring sections that may be affected by the feeder terminal's operating status. Subsequently, the monitoring data within the marked sections is corrected using anomaly detection methods. Specifically, missing data due to abnormal communication status or self-test anomalies is compensated for; voltage and current abrupt changes due to protection activation / deactivation or function switching are smoothed for amplitude; and invalid data due to non-load changes caused by manual operation is removed. After the above processing steps of status indicator, anomaly identification, and data correction, the abnormal interference caused by the feeder terminal's own operating status on the line monitoring sequence is removed, resulting in the line operating characteristic sequence.

[0020] Step S300: Based on the multi-dimensional factors of reclosing blocking, perform reliable mining of the operating characteristics of the distribution network line, build a reclosing blocking detection space, and input the line operating characteristic sequence into the reclosing blocking detection space to obtain the reclosing blocking coefficient.

[0021] In this embodiment, firstly, using interruption current blocking, second harmonic inrush current blocking, reverse power supply blocking, double-sided voltage blocking, and manual tripping blocking as multi-dimensional factors for reclosing blocking, event retrieval is performed on the historical operation records of the distribution network lines, forming interruption current blocking event clusters, second harmonic inrush current blocking event clusters, reverse power supply blocking event clusters, double-sided voltage blocking event clusters, and manual tripping blocking event clusters, respectively. Subsequently, based on each event cluster, reliable analysis of line operation characteristics is conducted to construct corresponding interruption current blocking detection spaces, second harmonic inrush current blocking detection spaces, reverse power supply blocking detection spaces, double-sided voltage blocking detection spaces, and manual tripping blocking detection spaces. Finally, these multiple blocking detection spaces are encapsulated and integrated to form the reclosing blocking detection space.

[0022] Next, the line operation characteristic sequence is input into the reclosing lockout detection space. In this process, the line operation characteristic sequence is input into the interruption current lockout detection space, the second harmonic inrush current lockout detection space, the reverse current blocking detection space, the double-sided voltage blocking detection space, and the manual tripping lockout detection space, respectively. By comparing the feature similarity with the lockout detection samples in each lockout detection space, the interruption current lockout detection coefficient, the second harmonic inrush current lockout detection coefficient, the reverse current blocking detection coefficient, the double-sided voltage blocking detection coefficient, and the manual tripping lockout detection coefficient are obtained. The interruption current lockout detection coefficient is determined by the maximum similarity coefficient among multiple interruption current lockout detection samples. Subsequently, the maximum value is selected from the above multiple lockout detection coefficients, and the maximum coefficient is used as the reclosing lockout coefficient.

[0023] Furthermore, the method provided in the application embodiments also includes: The reclosing interlocking multi-dimensional factors include interruption current interlocking, second harmonic inrush current interlocking, reverse power supply interlocking, dual-sided pressurized interlocking, and manual tripping interlocking.

[0024] In this embodiment, the reclosing blocking multi-dimensional factor consists of interruption current blocking, second harmonic inrush current blocking, reverse current blocking, double-sided voltage blocking, and manual tripping blocking, used to characterize the blocking characteristics of distribution network lines in reclosing condition determination from different dimensions. Specifically, interruption current blocking reflects whether the current during line operation reaches the interruption current range, thereby determining whether the fault current constitutes a blocking condition; second harmonic inrush current blocking identifies the second harmonic component in the current signal to distinguish between excitation inrush current and fault current characteristics; reverse current blocking determines whether there are reverse current characteristics during power restoration, thereby identifying whether reclosing conditions are not met; double-sided voltage blocking determines whether the line is in a double-sided energized condition based on the voltage status on both sides of the switch, to determine whether reclosing needs to be blocked; and manual tripping blocking identifies tripping behavior caused by manual operation to ensure that reclosing blocking is maintained after manual intervention.

[0025] Furthermore, the method provided in the application embodiment, which performs reliable mining of the operating characteristics of the distribution network line based on the multi-dimensional factors of reclosing interlocking to build a reclosing interlocking detection space, also includes: Based on the reclosing blocking multi-dimensional factors, historical events are retrieved for the distribution network lines to obtain event clusters for interruption current blocking, second harmonic inrush current blocking, reverse current blocking, double-sided voltage blocking, and manual tripping blocking. Based on the interruption current blocking event clusters, reliable analysis of line operation characteristics is performed to establish an interruption current blocking detection space. Based on the second harmonic inrush current blocking event clusters, reliable analysis of line operation characteristics is performed to establish a second harmonic inrush current blocking detection space. Based on the reverse current blocking event clusters, reliable analysis of line operation characteristics is performed to establish a second harmonic inrush current blocking detection space. A reliable analysis of operational characteristics is conducted to establish a reverse power-on blocking detection space; a reliable analysis of line operational characteristics is conducted based on the dual-sided pressurized blocking event cluster to establish a dual-sided pressurized blocking detection space; a reliable analysis of line operational characteristics is conducted based on the manual trip blocking event cluster to establish a manual trip blocking detection space; the interruption current blocking detection space, the second harmonic inrush current blocking detection space, the reverse power-on blocking detection space, the dual-sided pressurized blocking detection space, and the manual trip blocking detection space are encapsulated into the reclosing blocking detection space.

[0026] In this embodiment of the application, when retrieving historical events of distribution network lines based on the multi-dimensional factors of reclosing blocking, the historical operation records of the distribution network lines are filtered using an event feature matching method. The criteria corresponding to interruption current blocking, second harmonic inrush current blocking, reverse power supply blocking, double-sided voltage blocking, and manual tripping blocking are used as search conditions. Feature comparison is performed on historical current records, historical voltage records, and historical operation records. Events that meet the interruption current feature, second harmonic feature, reverse power supply feature, double-sided voltage feature, and manual tripping feature are respectively grouped to obtain interruption current blocking event clusters, second harmonic inrush current blocking event clusters, reverse power supply blocking event clusters, double-sided voltage blocking event clusters, and manual tripping blocking event clusters.

[0027] Next, when performing a reliable analysis of line operation characteristics based on the interruption current blocking event clusters, the following steps are taken: First, operational feature identification is performed on the interruption current blocking event clusters to extract multiple interruption current blocking line operation samples reflecting the interruption current blocking characteristics. Then, support evaluation is performed on these interruption current blocking line operation samples to form an operation sample support sequence. Based on the support evaluation results, a reliability evaluation is conducted to obtain an operation sample reliability evaluation sequence. Next, based on the operation sample reliability evaluation sequence and using the sample reliability evaluation threshold as a screening criterion, reliability optimization identification is performed on multiple interruption current blocking line operation samples, ultimately forming an interruption current blocking detection space containing multiple interruption current blocking detection samples.

[0028] Similarly, when performing reliable analysis of line operation characteristics based on the second harmonic inrush current blocking event cluster, multiple line operation samples reflecting the characteristics of second harmonic inrush current blocking are first identified from the cluster. Then, support evaluation is performed on these line operation samples to form corresponding operation sample support sequences, and a reliability evaluation is conducted to form corresponding operation sample reliability evaluation sequences. Based on the sample reliability evaluation threshold, reliability optimization identification is performed on multiple second harmonic inrush current blocking line operation samples, ultimately forming a second harmonic inrush current blocking detection space containing multiple second harmonic inrush current blocking detection samples.

[0029] When performing reliable analysis of line operation characteristics based on reverse power-in blocking event clusters, the process begins by identifying operation features within each cluster, extracting multiple line operation samples with reverse power-in blocking characteristics. Subsequently, support evaluation is performed on these sample operation data to generate corresponding support sequences. The support results are then reliably evaluated to obtain corresponding reliable evaluation sequences. Finally, based on the sample reliability evaluation threshold, reliable optimization is performed on the multiple reverse power-in blocking line operation samples to form a reverse power-in blocking detection space, which contains multiple reverse power-in blocking detection samples.

[0030] When performing reliable analysis of line operation characteristics based on double-sided pressurized blocking event clusters, the process begins by identifying operation features of the double-sided pressurized blocking event clusters, extracting multiple line operation samples representing the characteristics of double-sided pressurized blocking. Subsequently, support evaluation is performed on these line operation samples to generate a support sequence, and this sequence is then reliably evaluated to form a reliable evaluation sequence. Based on the sample reliability evaluation threshold, reliable optimization identification is performed on multiple double-sided pressurized blocking line operation samples, ultimately establishing a double-sided pressurized blocking detection space, which consists of multiple double-sided pressurized blocking detection samples.

[0031] When performing reliable analysis of line operation characteristics based on manual tripping and blocking event clusters, the process begins by identifying the operation features of each event cluster, extracting multiple line operation samples with manual tripping and blocking characteristics. Subsequently, support evaluation is performed on these manually tripping and blocking line operation samples to obtain an operation sample support sequence. Based on this, a reliability evaluation is conducted to obtain an operation sample reliability evaluation sequence. According to the sample reliability evaluation threshold, reliability optimization identification is performed on multiple manually tripping and blocking line operation samples to form a manual tripping and blocking detection space, which contains multiple manually tripping and blocking detection samples.

[0032] Finally, the interruption current blocking detection space, second harmonic inrush current blocking detection space, reverse power supply blocking detection space, dual-sided voltage blocking detection space, and manual trip blocking detection space are encapsulated and integrated. These spaces are then grouped in parallel according to their respective corresponding reclosing blocking multi-dimensional factors, organizing the five types of blocking feature samples in a unified structural form. This allows each type of detection space to function within the same judgment system. Through this encapsulation step, the interruption current blocking detection space, second harmonic inrush current blocking detection space, reverse power supply blocking detection space, dual-sided voltage blocking detection space, and manual trip blocking detection space are merged into a single reclosing blocking detection space.

[0033] Furthermore, in the method provided in the application embodiments, the process of performing reliable analysis of line operation characteristics based on the interruption current blocking event cluster to establish an interruption current blocking detection space also includes: The operation characteristics of the distribution network lines are identified based on the interruption current blocking event clusters to obtain multiple interruption current blocking line operation samples. Support is evaluated on these multiple interruption current blocking line operation samples based on the interruption current blocking event clusters to obtain an operation sample support sequence. A credibility evaluation is performed on these multiple interruption current blocking line operation samples based on the operation sample support sequence to obtain an operation sample credibility evaluation sequence. Based on the sample credibility evaluation threshold, credibility optimization identification is performed on these multiple interruption current blocking line operation samples based on the operation sample credibility evaluation sequence to generate the interruption current blocking detection space, which includes multiple interruption current blocking detection samples.

[0034] In this embodiment of the application, when performing operation feature identification of distribution network lines based on interruption current blocking event clusters, a feature identification method based on threshold comparison is used to analyze the historical current records in the interruption current blocking event clusters. The current amplitude threshold, duration characteristics, and current change rate corresponding to the interruption current blocking characteristics are used as identification conditions. The current time series data in the event clusters are compared one by one, and the record segments that meet the characteristics of interruption current surge and continuity are selected to obtain multiple interruption current blocking line operation samples.

[0035] Next, when evaluating the support of multiple circuit operation samples with interruption current blocking, a statistical evaluation method based on occurrence frequency is used. The number of times each circuit operation sample with interruption current blocking occurs in the interruption current blocking event cluster is counted. By calculating the occurrence frequency of each operation sample, the support of the operation sample is obtained, and the support of multiple operation samples is arranged into an operation sample support sequence according to the sample order.

[0036] Subsequently, when conducting a credibility evaluation of multiple operating samples of interrupted current blocking lines, a credibility assessment method based on normalized ratio calculation was used. The support sequence of the operating samples was summed, and the formula for calculating the credibility evaluation coefficients of multiple operating samples was obtained by dividing the support of each operating sample by the sum of the elements of the support sequence of the operating sample. These credibility evaluation coefficients of the operating samples constituted the credibility evaluation sequence of the operating samples, so that the credibility of each operating sample of interrupted current blocking line in the overall event cluster can be characterized in a unified quantitative way.

[0037] Finally, when performing reliable optimization identification of multiple circuit operation samples with interruption current blocking based on the sample reliability evaluation threshold, a feature optimization method based on threshold screening is used to compare the reliability evaluation sequence of the operation samples item by item. The circuit operation samples with interruption current blocking corresponding to the reliability evaluation coefficient of the operation samples that are greater than or equal to the sample reliability evaluation threshold are identified as interruption current blocking detection samples. These interruption current blocking detection samples are added to the interruption current blocking detection sample set to generate the interruption current blocking detection space, so that the interruption current blocking detection space contains multiple interruption current blocking detection samples that have undergone reliable screening.

[0038] Furthermore, in the method provided in the application embodiment, inputting the line operation characteristic sequence into the reclosing blockade detection space to obtain the reclosing blockade coefficient further includes: The line operation characteristic sequence is input into the interruption current blocking detection space to obtain the interruption current blocking detection coefficient, which is the maximum similarity coefficient between the line operation characteristic sequence and multiple interruption current blocking detection samples in the interruption current blocking detection space. The line operation characteristic sequence is input into the second harmonic inrush current blocking detection space to obtain the second harmonic inrush current blocking detection coefficient. The line operation characteristic sequence is input into the reverse power supply blocking detection space to obtain the reverse power supply blocking detection coefficient. The line operation characteristic sequence is input into the double-sided voltage blocking detection space to obtain the double-sided voltage blocking detection coefficient. The line operation characteristic sequence is input into the manual trip blocking detection space to obtain the manual trip blocking detection coefficient. The maximum values ​​of the interruption current blocking detection coefficient, the second harmonic inrush current blocking detection coefficient, the reverse power supply blocking detection coefficient, the double-sided voltage blocking detection coefficient, and the manual trip blocking detection coefficient are filtered to generate the reclosing blocking coefficient.

[0039] In this embodiment of the application, when the line operation feature sequence is input into the interruption current blocking detection space, a method based on cosine similarity calculation is used to compare the line operation feature sequence with multiple interruption current blocking detection samples in the interruption current blocking detection space one by one. Multiple similarity coefficients are generated by performing cosine similarity calculation on the feature vectors, and the largest similarity coefficient is selected from the multiple similarity coefficients and used as the interruption current blocking detection coefficient.

[0040] Similarly, when the line operation feature sequence is input into the second harmonic inrush current blocking detection space, a feature comparison method based on cosine similarity calculation is used to perform harmonic feature comparison between the line operation feature sequence and multiple second harmonic inrush current blocking detection samples in the second harmonic inrush current blocking detection space. Multiple similarity coefficients are obtained by calculating the cosine similarity of the second harmonic feature vectors, and the largest similarity coefficient is selected as the second harmonic inrush current blocking detection coefficient.

[0041] When the line operation feature sequence is input into the reverse power-on blocking detection space, a method based on cosine similarity calculation is used to compare the power-on direction features of the line operation feature sequence with multiple reverse power-on blocking detection samples in the reverse power-on blocking detection space. By performing cosine similarity calculation on the feature vector composed of voltage recovery direction, phase change and power-on time features, multiple similarity coefficients are obtained, and the largest similarity coefficient is selected as the reverse power-on blocking detection coefficient.

[0042] When the line operation feature sequence is input into the double-sided voltage blocking detection space, a method based on cosine similarity calculation is used to perform voltage feature comparison between the line operation feature sequence and multiple double-sided voltage blocking detection samples in the double-sided voltage blocking detection space. By performing cosine similarity calculation on the feature vector composed of double-sided voltage amplitude, double-sided voltage phase difference and voltage stability, multiple similarity coefficients are obtained, and the largest similarity coefficient is selected as the double-sided voltage blocking detection coefficient.

[0043] When the line operation feature sequence is input into the manual tripping interlock detection space, a method based on cosine similarity calculation is used to compare the electrical quantity change features of the line operation feature sequence and multiple manual tripping interlock detection samples in the manual tripping interlock detection space. By performing cosine similarity calculation on the feature vector composed of current changes, voltage changes and transient features before and after the tripping, multiple similarity coefficients are obtained, and the largest similarity coefficient is selected as the manual tripping interlock detection coefficient.

[0044] After obtaining the detection coefficients for interruption current blocking, second harmonic inrush current blocking, reverse power supply blocking, double-sided pressurized blocking, and manual tripping blocking, a numerical screening method based on the maximum value is used to compare each of the above five types of detection coefficients, and the one with the largest value among the five types of detection coefficients is selected as the reclosing blocking coefficient.

[0045] Step S400: If the reclosing blocking coefficient is less than the reclosing blocking threshold, perform multi-dimensional fault prediction on the distribution network line according to the line operation characteristic sequence and generate a line fault feature map.

[0046] In this embodiment, the reclosing blocking coefficient is compared with a preset reclosing blocking threshold. When the reclosing blocking coefficient is less than the reclosing blocking threshold, multi-dimensional fault prediction is performed on the distribution network line based on the line operation characteristic sequence. In this process, short-circuit fault prediction, ground fault prediction, and other fault prediction are first performed based on the line operation characteristic sequence, thereby obtaining the line short-circuit fault prediction results, line ground fault prediction results, and line other fault prediction results. Subsequently, a graph neural network is used to structure and organize the three types of fault prediction results. By integrating and representing the correlation between the short-circuit fault prediction results, ground fault prediction results, and other fault prediction results, a line fault feature map is formed.

[0047] Furthermore, in the method provided in the application embodiments, the method for performing multi-dimensional fault prediction on the distribution network line based on the line operation characteristic sequence to generate a line fault feature map further includes: Short-circuit fault prediction is performed on the distribution network line based on the line operation characteristic sequence to obtain line short-circuit fault prediction results; ground fault prediction is performed on the distribution network line based on the line operation characteristic sequence to obtain line ground fault prediction results; other fault prediction is performed on the distribution network line based on the line operation characteristic sequence to obtain line other fault prediction results; the line short-circuit fault prediction results, the line ground fault prediction results, and the line other fault prediction results are processed using a graph neural network to obtain the line fault feature map.

[0048] In this embodiment, when predicting short-circuit faults in distribution network lines based on line operation characteristic sequences, a pre-trained short-circuit fault identification neural network model is used. During the training phase, this model uses historical line operation characteristic sequences as input data and historically labeled short-circuit fault prediction results as output data. Through multiple rounds of iterative learning on the time-dimensional change patterns of the input sequences, the model internally forms discrimination parameters related to short-circuit conditions. By inputting the line operation characteristic sequence formed in the current period into the short-circuit fault identification neural network model, the model performs calculations on the input sequence based on the parameters formed during the training phase to obtain the line short-circuit fault prediction result.

[0049] When predicting ground faults in distribution network lines based on line operation characteristic sequences, a pre-trained ground fault identification neural network model is used. During the training phase, this neural network model uses historical line operation characteristic sequences as input data and ground fault prediction results marked in historical operation records as output data. Through multiple iterative learning processes on the grounding patterns presented in the sequences, the model develops an internal parameter structure related to grounding conditions. By inputting the line operation characteristic sequences into the ground fault identification neural network model, the model analyzes the input sequences based on the parameters formed during the training phase and outputs the line ground fault prediction results.

[0050] When predicting other faults in distribution network lines based on line operation characteristic sequences, a pre-trained neural network model for identifying other faults is used. During the training phase, this neural network model uses historical line operation characteristic sequences as input data and previously labeled line fault prediction results from historical records as output data. By learning the changing patterns corresponding to atypical operating states in the sequence, the model can develop internal parameters to distinguish fault types other than short-circuit and grounding faults. By inputting the line operation characteristic sequences into the other fault identification neural network model, the model processes the input sequences based on the parameters developed during training and outputs the predicted results for other line faults.

[0051] Finally, the graph neural network (Graph Neural Network) is used to organize the prediction results of line short-circuit faults, line grounding faults, and other line faults. During the training phase, the Graph Neural Network uses the three types of historical fault prediction results as input data and the line fault feature map as output data. It iteratively learns the internal parameters by studying the combination patterns of different fault categories in terms of temporal, correlational, and distributional relationships. By inputting the prediction results of line short-circuit faults, line grounding faults, and other line faults into the Graph Neural Network, it organizes the three types of prediction results and outputs the line fault feature map.

[0052] Furthermore, the method provided in the application embodiments also includes: If the reclosing interlocking coefficient is greater than or equal to the reclosing interlocking threshold, a reclosing interlocking command is generated.

[0053] In this embodiment, when the reclosing lockout coefficient is greater than or equal to the reclosing lockout threshold, a prohibition command for reclosing action is generated. This command is used as a reclosing lockout command to require the control system to keep the reclosing in a locked state under the current operating condition, so as to prevent the switch from performing a closing operation.

[0054] Step S500: Based on the line fault characteristic map, predict the protection trip probability of the distribution network line and determine the line protection trip probability.

[0055] In this embodiment, when predicting the protection tripping probability of a distribution network line based on the line fault feature map, the historical protection tripping records of the distribution network line are first retrieved to form a line fault feature sample set and a protection tripping probability sample set. Then, a Support Vector Regression (SVR) model and a Gated Recurrent Unit (GRU) model are trained under supervision using the line fault feature sample set and the protection tripping probability sample set, respectively, to obtain a first protection tripping probability prediction model and a second protection tripping probability prediction model. Next, the output results of the two models are used as a fusion object. A probability prediction output fusion model is constructed through training, and the first and second protection tripping probability prediction models are used as the first protection tripping probability prediction node, while the probability prediction output fusion model is used as the second protection tripping probability prediction node, forming a protection tripping probability prediction channel. Finally, the line fault feature map is input into the protection tripping probability prediction channel to obtain the line protection tripping probability.

[0056] Furthermore, in the method provided in the application embodiment, predicting the protection tripping probability of the distribution network line based on the line fault feature map and determining the line protection tripping probability further includes: Based on the protection tripping records retrieved from the distribution network lines, a line fault feature sample set and a protection tripping probability sample set were obtained. The SVR model was then trained under supervision using the line fault feature sample set and the protection tripping probability sample set to obtain a first protection tripping probability prediction model. The GRU model was then trained under supervision using the line fault feature sample set and the protection tripping probability sample set to obtain a second protection tripping probability prediction model. The first and second protection tripping probability prediction models were then fused to obtain a probability prediction output fusion model. A protection tripping probability prediction channel was established using the first and second protection tripping probability prediction models as the first protection tripping probability prediction node and the probability prediction output fusion model as the second protection tripping probability prediction node. The line fault feature map was then input into the protection tripping probability prediction channel to obtain the line protection tripping probability.

[0057] In this embodiment of the application, when retrieving protection trip records of distribution network lines, the historical operation files of the distribution network lines are read one by one, and data sequences that can represent line fault conditions are extracted from the files and organized into a line fault feature sample set in chronological order. At the same time, corresponding trip information, including whether a trip has occurred or the trip probability value, is extracted from the same file and organized into a protection trip probability sample set, so that the two types of sample sets maintain a corresponding relationship in time.

[0058] Next, when supervising the training of the SVR model based on the line fault feature sample set and the protection trip probability sample set, the line fault feature sample set is input into the SVR model as training input, and the protection trip probability sample set is input into the SVR model as training target value. Through multiple rounds of training, the internal parameters of the model are continuously adjusted so that the SVR model can generate prediction results consistent with the training target value based on the input data, thus forming the first model for protection trip probability prediction.

[0059] When supervising the training of the GRU model based on the line fault feature sample set and the protection trip probability sample set, the line fault feature sample set is input into the GRU model in chronological order, and the protection trip probability sample set is used as the corresponding sequence target value input into the model. This enables the model to establish the temporal correlation between the input sequence and the target sequence during the training process. Through multiple rounds of training, the internal control structure is continuously adjusted so that the GRU model can output a stable trip probability estimate based on the input sequence, thus obtaining the second model for protection trip probability prediction.

[0060] Subsequently, during the output fusion training based on the first and second protection trip probability prediction models, the trip probability prediction values ​​obtained from the two models under the same input conditions are simultaneously input into the fusion structure. Through multiple rounds of training, the internal parameters of the fusion structure are adjusted so that the fusion structure can integrate the prediction results of the two models and generate a more balanced output, thus obtaining the probability prediction output fusion model.

[0061] Then, when constructing the protection trip probability prediction channel, the first protection trip probability prediction model and the second protection trip probability prediction model are used as the first protection trip probability prediction node, and the probability prediction output fusion model is used as the second protection trip probability prediction node. The nodes are arranged in order so that the input data can first pass through the first node to obtain the initial prediction, and then pass through the second node to complete the comprehensive prediction, thus forming a continuous trip probability prediction process.

[0062] Finally, when inputting the line fault feature map into the protection trip probability prediction channel, the line fault feature map is sequentially input into the first node and the second node of the protection trip probability prediction. The first node outputs the initial trip probability prediction result according to its structure, and the second node performs comprehensive processing on the initial trip probability prediction result and generates the final line protection trip probability.

[0063] Step S600: Introduce an adaptive reclosing mechanism, combining the line protection tripping probability and the feeder terminal to perform adaptive reclosing control.

[0064] Furthermore, in the method provided in the application embodiments, the adaptive reclosing mechanism includes: Determine whether the line protection trip probability is greater than or equal to the protection trip probability threshold; if the line protection trip probability is greater than or equal to the protection trip probability threshold, generate a reclosing trigger command and simultaneously generate trigger timing information; if the trigger timing information meets a first preset duration, execute reclosing control based on the reclosing trigger command and obtain the reclosing execution result from the feeder terminal; if the reclosing execution result indicates a trip within a second preset duration, generate a failure blocking command; if the reclosing execution result indicates no trip within the second preset duration, generate a reclosing logic reset command.

[0065] In this embodiment, an adaptive reclosing mechanism is introduced, which combines the line protection tripping probability and the feeder terminal for adaptive reclosing control.

[0066] When performing adaptive reclosing control based on the adaptive reclosing mechanism, the first step is to determine whether the line protection tripping probability is greater than or equal to a preset protection tripping probability threshold. When the line protection tripping probability is greater than or equal to the protection tripping probability threshold, a reclosing trigger command is generated, and a timer is started simultaneously with the generation of the reclosing trigger command. The trigger timing information is formed by recording the current time.

[0067] Next, it is determined whether the trigger timing information meets the first preset duration. A duration judgment method based on time accumulation comparison is used to accumulate the trigger timing information and compare the accumulated duration with the first preset duration. When the accumulated duration reaches or exceeds the first preset duration, it indicates that the reclosing trigger command has met the execution conditions, and reclosing control is executed. At this time, based on the reclosing execution result obtained from the feeder terminal, that is, using the reclosing trigger command as the action trigger source, the switch completes the closing action by executing the reclosing control operation. Simultaneously, the line operating status after closing is collected through the feeder terminal and used as the reclosing execution result.

[0068] Subsequently, when determining whether the reclosing execution result is a trip that occurred within the second preset time period, the reclosing execution result is continuously monitored based on the response judgment method of status detection within the time window, and the monitored operating status is compared with the second preset time period. When the monitoring result shows that a trip occurred within the second preset time period, a failure blocking instruction is generated based on the result, and the system enters the blocking state to prevent further execution of the reclosing action.

[0069] When the reclosing execution result indicates that no tripping has occurred within the second preset time period, the reclosing execution result is recorded based on the instruction generation method for determining the normal operating state. When the reclosing execution result indicates that the operation remains stable and no tripping has occurred within the second preset time period, a reclosing logic reset instruction is generated based on this state, so that the reclosing logic is restored to a state that can be executed again.

[0070] In summary, the embodiments of this application have at least the following technical effects: This application performs real-time monitoring of distribution network lines using feeder terminals to obtain line monitoring sequences and simultaneously acquires the terminal status sequences of the feeder terminals. Based on the terminal status sequences, it performs terminal anomaly interference correction on the line monitoring sequences to obtain line operation characteristic sequences. It then performs reliable mining of the operation characteristics of the distribution network lines based on multi-dimensional factors of reclosing blocking, constructs a reclosing blocking detection space, and inputs the line operation characteristic sequences into the reclosing blocking detection space to obtain reclosing blocking coefficients. If the reclosing blocking coefficients are less than the reclosing blocking threshold, it performs multi-dimensional fault prediction on the distribution network lines based on the line operation characteristic sequences to generate a line fault feature map. Based on the line fault feature map, it predicts the protection trip probability of the distribution network lines to determine the line protection trip probability. Finally, it introduces an adaptive reclosing mechanism, combining the line protection trip probability and the feeder terminals for adaptive reclosing control. This invention addresses the technical problems of inaccurate reclosing interlock determination and lack of adaptability in existing technologies. By constructing a reclosing interlock detection space to obtain the interlock coefficient, and combining fault prediction and protection tripping probability, adaptive reclosing control is achieved, thereby improving the accuracy and reliability of reclosing action decision-making.

[0071] Example 2 is based on the same inventive concept as the adaptive reclosing control method for distribution automation feeder terminals in the foregoing examples, such as... Figure 2 As shown, this application provides an adaptive reclosing control system for distribution automation feeder terminals. The system and method embodiments in this application are based on the same inventive concept. The system includes: The real-time monitoring module 11 is used to monitor the distribution network line in real time according to the feeder terminal, obtain the line monitoring sequence, and synchronously acquire the terminal status sequence of the feeder terminal; the interference correction module 12 is used to perform terminal anomaly interference correction on the line monitoring sequence according to the terminal status sequence, and obtain the line operation feature sequence; the mining module 13 is used to perform reliable mining of the operation features of the distribution network line according to the reclosing blocking multi-dimensional factors, build a reclosing blocking detection space, and input the line operation feature sequence into the reclosing blocking detection space to obtain the reclosing blocking coefficient; the fault prediction module 14 is used to perform multi-dimensional fault prediction on the distribution network line according to the line operation feature sequence if the reclosing blocking coefficient is less than the reclosing blocking threshold, and generate a line fault feature map; the probability prediction module 15 is used to predict the protection trip probability of the distribution network line according to the line fault feature map, and determine the line protection trip probability; the control module 16 is used to introduce an adaptive reclosing mechanism, and perform adaptive reclosing control by combining the line protection trip probability and the feeder terminal.

[0072] Furthermore, the system is also used to implement the following functions: Based on the reclosing blocking multi-dimensional factors, historical events are retrieved for the distribution network lines to obtain event clusters for interruption current blocking, second harmonic inrush current blocking, reverse current blocking, double-sided voltage blocking, and manual tripping blocking. Based on the interruption current blocking event clusters, reliable analysis of line operation characteristics is performed to establish an interruption current blocking detection space. Based on the second harmonic inrush current blocking event clusters, reliable analysis of line operation characteristics is performed to establish a second harmonic inrush current blocking detection space. Based on the reverse current blocking event clusters, reliable analysis of line operation characteristics is performed to establish a second harmonic inrush current blocking detection space. A reliable analysis of operational characteristics is conducted to establish a reverse power-on blocking detection space; a reliable analysis of line operational characteristics is conducted based on the dual-sided pressurized blocking event cluster to establish a dual-sided pressurized blocking detection space; a reliable analysis of line operational characteristics is conducted based on the manual trip blocking event cluster to establish a manual trip blocking detection space; the interruption current blocking detection space, the second harmonic inrush current blocking detection space, the reverse power-on blocking detection space, the dual-sided pressurized blocking detection space, and the manual trip blocking detection space are encapsulated into the reclosing blocking detection space.

[0073] Furthermore, the system is also used to implement the following functions: The operation characteristics of the distribution network lines are identified based on the interruption current blocking event clusters to obtain multiple interruption current blocking line operation samples. Support is evaluated on these multiple interruption current blocking line operation samples based on the interruption current blocking event clusters to obtain an operation sample support sequence. A credibility evaluation is performed on these multiple interruption current blocking line operation samples based on the operation sample support sequence to obtain an operation sample credibility evaluation sequence. Based on the sample credibility evaluation threshold, credibility optimization identification is performed on these multiple interruption current blocking line operation samples based on the operation sample credibility evaluation sequence to generate the interruption current blocking detection space, which includes multiple interruption current blocking detection samples.

[0074] Furthermore, the system is also used to implement the following functions: The line operation characteristic sequence is input into the interruption current blocking detection space to obtain the interruption current blocking detection coefficient, which is the maximum similarity coefficient between the line operation characteristic sequence and multiple interruption current blocking detection samples in the interruption current blocking detection space. The line operation characteristic sequence is input into the second harmonic inrush current blocking detection space to obtain the second harmonic inrush current blocking detection coefficient. The line operation characteristic sequence is input into the reverse power supply blocking detection space to obtain the reverse power supply blocking detection coefficient. The line operation characteristic sequence is input into the double-sided voltage blocking detection space to obtain the double-sided voltage blocking detection coefficient. The line operation characteristic sequence is input into the manual trip blocking detection space to obtain the manual trip blocking detection coefficient. The maximum values ​​of the interruption current blocking detection coefficient, the second harmonic inrush current blocking detection coefficient, the reverse power supply blocking detection coefficient, the double-sided voltage blocking detection coefficient, and the manual trip blocking detection coefficient are filtered to generate the reclosing blocking coefficient.

[0075] Furthermore, the system is also used to implement the following functions: Short-circuit fault prediction is performed on the distribution network line based on the line operation characteristic sequence to obtain line short-circuit fault prediction results; ground fault prediction is performed on the distribution network line based on the line operation characteristic sequence to obtain line ground fault prediction results; other fault prediction is performed on the distribution network line based on the line operation characteristic sequence to obtain line other fault prediction results; the line short-circuit fault prediction results, the line ground fault prediction results, and the line other fault prediction results are processed using a graph neural network to obtain the line fault feature map.

[0076] Furthermore, the system is also used to implement the following functions: Based on the protection tripping records retrieved from the distribution network lines, a line fault feature sample set and a protection tripping probability sample set were obtained. The SVR model was then trained under supervision using the line fault feature sample set and the protection tripping probability sample set to obtain a first protection tripping probability prediction model. The GRU model was then trained under supervision using the line fault feature sample set and the protection tripping probability sample set to obtain a second protection tripping probability prediction model. The first and second protection tripping probability prediction models were then fused to obtain a probability prediction output fusion model. A protection tripping probability prediction channel was established using the first and second protection tripping probability prediction models as the first protection tripping probability prediction node and the probability prediction output fusion model as the second protection tripping probability prediction node. The line fault feature map was then input into the protection tripping probability prediction channel to obtain the line protection tripping probability.

[0077] Furthermore, the system is also used to implement the following functions: Determine whether the line protection trip probability is greater than or equal to the protection trip probability threshold; if the line protection trip probability is greater than or equal to the protection trip probability threshold, generate a reclosing trigger command and simultaneously generate trigger timing information; if the trigger timing information meets a first preset duration, execute reclosing control based on the reclosing trigger command and obtain the reclosing execution result from the feeder terminal; if the reclosing execution result indicates a trip within a second preset duration, generate a failure blocking command; if the reclosing execution result indicates no trip within the second preset duration, generate a reclosing logic reset command.

[0078] Furthermore, the system is also used to implement the following functions: The reclosing interlocking multi-dimensional factors include interruption current interlocking, second harmonic inrush current interlocking, reverse power supply interlocking, dual-sided pressurized interlocking, and manual tripping interlocking.

[0079] Furthermore, the system is also used to implement the following functions: If the reclosing interlocking coefficient is greater than or equal to the reclosing interlocking threshold, a reclosing interlocking command is generated.

[0080] It should be noted that the order of the embodiments described above is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. Furthermore, the above description focuses on specific embodiments of this specification. The processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired results. In some implementations, multitasking and parallel processing are possible or may be advantageous.

[0081] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any modifications, equivalent changes, and alterations made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the scope of the present invention.

Claims

1. An adaptive reclosing control method for feeder terminals in distribution automation, characterized in that, The method includes: The feeder terminal performs real-time monitoring of the distribution network lines to obtain the line monitoring sequence and simultaneously acquires the terminal status sequence of the feeder terminal. Based on the terminal status sequence, the line monitoring sequence is corrected for terminal anomalies to obtain the line operation characteristic sequence. Based on the multi-dimensional factors of reclosing interlocking, the operating characteristics of the distribution network line are reliably mined, a reclosing interlocking detection space is built, and the operating characteristic sequence of the line is input into the reclosing interlocking detection space to obtain the reclosing interlocking coefficient. If the reclosing blocking coefficient is less than the reclosing blocking threshold, multi-dimensional fault prediction is performed on the distribution network line based on the line operation characteristic sequence to generate a line fault characteristic map. Based on the line fault feature map, the protection trip probability of the distribution network line is predicted to determine the line protection trip probability. An adaptive reclosing mechanism is introduced, which combines the line protection tripping probability with the feeder terminal to perform adaptive reclosing control.

2. The adaptive reclosing control method for distribution automation feeder terminals as described in claim 1, characterized in that, Based on the multi-dimensional factors of reclosing interlocking, the operating characteristics of the distribution network line are reliably mined to build a reclosing interlocking detection space, including: Based on the reclosing blocking multidimensional factor, historical events are retrieved for the distribution network line to obtain the interruption current blocking event cluster, the second harmonic inrush current blocking event cluster, the reverse power supply blocking event cluster, the double-sided pressure blocking event cluster, and the manual tripping blocking event cluster. Based on the interruption current blocking event cluster, reliable analysis of line operation characteristics is performed to establish an interruption current blocking detection space. Based on the second harmonic inrush current blocking event cluster, reliable analysis of line operation characteristics is performed to establish a second harmonic inrush current blocking detection space. Based on the reverse power blocking event cluster, reliable analysis of line operation characteristics is performed to establish a reverse power blocking detection space. Based on the aforementioned double-sided pressurized interlocking event cluster, a reliable analysis of the line operation characteristics is performed to establish a double-sided pressurized interlocking detection space. Based on the aforementioned manual tripping and blocking event clusters, reliable analysis of line operation characteristics is performed to establish a manual tripping and blocking detection space; The interruption current blocking detection space, the second harmonic inrush current blocking detection space, the reverse power supply blocking detection space, the dual-sided pressurized blocking detection space, and the manual trip blocking detection space are encapsulated into the reclosing blocking detection space.

3. The adaptive reclosing control method for distribution automation feeder terminals as described in claim 2, characterized in that, Based on the aforementioned interruption current blocking event cluster, a reliable analysis of line operation characteristics is performed to establish an interruption current blocking detection space, including: Based on the interruption current blocking event cluster, the operation characteristics identification of the distribution network line is performed to obtain multiple interruption current blocking line operation samples; Based on the interruption current blocking event cluster, the support of the multiple interruption current blocking line operation samples is evaluated to obtain the operation sample support sequence. Based on the support sequence of the operating samples, a credibility evaluation of the multiple interruption current blocking line operating samples is performed to obtain a credibility evaluation sequence of the operating samples. Based on the sample credibility evaluation threshold, the multiple interruption current blocking line operation samples are reliably optimized and identified according to the operation sample credibility evaluation sequence to generate the interruption current blocking detection space, which includes multiple interruption current blocking detection samples.

4. The adaptive reclosing control method for distribution automation feeder terminals as described in claim 1, characterized in that, The line operation characteristic sequence is input into the reclosing blockade detection space to obtain the reclosing blockade coefficients, including: The line operation feature sequence is input into the interruption current blocking detection space to obtain the interruption current blocking detection coefficient. The interruption current blocking detection coefficient is the maximum similarity coefficient between the line operation feature sequence and multiple interruption current blocking detection samples in the interruption current blocking detection space. The line operation characteristic sequence is input into the second harmonic inrush current blocking detection space to obtain the second harmonic inrush current blocking detection coefficient. The line operation characteristic sequence is input into the reverse power-on blocking detection space to obtain the reverse power-on blocking detection coefficient. The line operation characteristic sequence is input into the double-sided pressurized interlocking detection space to obtain the double-sided pressurized interlocking detection coefficient. The line operation characteristic sequence is input into the manual tripping interlock detection space to obtain the manual tripping interlock detection coefficient; The maximum values ​​of the interruption current blocking detection coefficient, the second harmonic inrush current blocking detection coefficient, the reverse power supply blocking detection coefficient, the dual-sided pressurized blocking detection coefficient, and the manual trip blocking detection coefficient are filtered to generate the reclosing blocking coefficient.

5. The adaptive reclosing control method for distribution automation feeder terminals as described in claim 1, characterized in that, Based on the line operation characteristic sequence, multi-dimensional fault prediction is performed on the distribution network lines to generate a line fault feature map, including: Short-circuit fault prediction is performed on the distribution network line based on the line operation characteristic sequence to obtain the line short-circuit fault prediction result; Based on the line operation characteristic sequence, ground fault prediction is performed on the distribution network line to obtain the line ground fault prediction result; Based on the line operation characteristic sequence, other fault predictions are performed on the distribution network lines to obtain other fault prediction results. Based on the graph neural network, the prediction results of the line short-circuit fault, the prediction results of the line ground fault, and the prediction results of other line faults are processed to obtain the line fault feature map.

6. The adaptive reclosing control method for feeder terminals in distribution automation as described in claim 1, characterized in that, Based on the line fault characteristic map, the probability of protection tripping of the distribution network line is predicted to determine the line protection tripping probability, including: Based on the protection trip records of the distribution network lines, a line fault feature sample set and a protection trip probability sample set are obtained. The SVR model is trained under supervision based on the line fault feature sample set and the protection trip probability sample set to obtain the first protection trip probability prediction model; The GRU model is trained under supervision based on the line fault feature sample set and the protection trip probability sample set to obtain a second protection trip probability prediction model. The output fusion training is performed based on the first protection trip probability prediction model and the second protection trip probability prediction model to obtain the probability prediction output fusion model; Using the first protection trip probability prediction model and the second protection trip probability prediction model as the first protection trip probability prediction node, and the probability prediction output fusion model as the second protection trip probability prediction node, a protection trip probability prediction channel is established. The line fault feature map is input into the protection trip probability prediction channel to obtain the line protection trip probability.

7. The adaptive reclosing control method for feeder terminals in distribution automation as described in claim 1, characterized in that, The adaptive reclosing mechanism includes: Determine whether the line protection trip probability is greater than or equal to the protection trip probability threshold; If the line protection trip probability is greater than or equal to the protection trip probability threshold, a reclosing trigger command is generated, and trigger timing information is generated simultaneously. If the trigger timing information meets the first preset duration, reclosing control is executed based on the reclosing trigger command, and the reclosing execution result is obtained according to the feeder terminal; If the reclosing execution result is a tripping within a second preset time period, a failure blocking instruction is generated. If the reclosing execution result is that no tripping occurs within the second preset time period, a reclosing logic reset instruction is generated.

8. The adaptive reclosing control method for distribution automation feeder terminals as described in claim 1, characterized in that, The reclosing interlocking multi-dimensional factors include interruption current interlocking, second harmonic inrush current interlocking, reverse power supply interlocking, dual-sided pressurized interlocking, and manual tripping interlocking.

9. The adaptive reclosing control method for feeder terminals in distribution automation as described in claim 1, characterized in that, If the reclosing interlocking coefficient is greater than or equal to the reclosing interlocking threshold, a reclosing interlocking command is generated.

10. An adaptive reclosing control system for feeder terminals in power distribution automation, characterized in that, The system is used to execute the adaptive reclosing control method for distribution automation feeder terminals as described in any one of claims 1-9, the system comprising: The real-time monitoring module is used to monitor the distribution network lines in real time based on the feeder terminals, obtain the line monitoring sequence, and synchronously acquire the terminal status sequence of the feeder terminals. Interference correction module is used to perform terminal anomaly interference correction on the line monitoring sequence based on the terminal state sequence to obtain the line operation characteristic sequence; The mining module is used to perform reliable mining of the operating characteristics of the distribution network line based on the multi-dimensional factors of reclosing blockade, build a reclosing blockade detection space, and input the line operating characteristic sequence into the reclosing blockade detection space to obtain the reclosing blockade coefficient. The fault prediction module is used to perform multi-dimensional fault prediction on the distribution network line based on the line operation characteristic sequence if the reclosing blocking coefficient is less than the reclosing blocking threshold, and generate a line fault feature map. The probability prediction module is used to predict the protection trip probability of the distribution network line based on the line fault feature map and determine the line protection trip probability. The control module is used to introduce an adaptive reclosing mechanism, which combines the line protection tripping probability and the feeder terminal to perform adaptive reclosing control.

Images