A power supply protection method and system for a metrology switch

Abstract

The present application relates to the field of measurement switch control, and particularly relates to a power supply protection method and system for a measurement switch, which comprises: collecting multi-dimensional electrical quantity data of each node in a power distribution network within a preset time period at a fixed frequency and preprocessing; improving a cost function in a path search algorithm; searching for an optimal path node sequence from a power supply point to a target load point in the power distribution network based on the improved cost function, and generating a control instruction of the measurement switch according to the optimal path node sequence to control closing or opening of the corresponding switch. The present application fuses three core elements of a node's own abnormality degree, inter-node propagation correlation degree and local risk trend of path extension, guides the search algorithm to avoid nodes with obvious abnormality, and actively cuts off electrical channels with high abnormal propagation risk, thereby greatly reducing the probability of causing chain faults or safety accidents in the fault recovery process.

Description

Technical Field

[0001] This invention relates to the field of measurement switch control. In particular, it relates to a power supply protection method and system for measurement switches. Background Technology

[0002] In modern power systems, measuring switches are critical control and monitoring devices, and their operating status directly affects the stability, reliability, and power supply quality of the power grid. By monitoring and analyzing the electrical quantity data collected by the measuring switches in real time, it is possible to assess the operating status of each node in the power grid, identify and locate potential faults, and then, by controlling the actions of the measuring switches, complete fault isolation and restore safe power supply paths.

[0003] However, existing fault location and power supply path search methods still have significant shortcomings. Traditional methods typically treat each node in the power grid as an independent unit, judging its abnormal state solely based on whether its own electrical quantity data exceeds a preset threshold. This isolated analysis mode completely ignores the dynamic characteristics and spatial correlation of faults or anomalies propagating between nodes through electrical connections. When searching for backup power supply paths, the algorithm may fail to identify nodes and lines already affected by anomalies or on high-risk propagation paths. This results in power restoration paths still passing through abnormal areas or potential fault propagation channels, easily triggering secondary faults, expanding the power outage area, and even causing equipment damage, seriously threatening power supply security and system stability. Summary of the Invention

[0004] To address the technical problem that existing measurement switch control methods rely solely on node electrical quantity data for abnormal state analysis, which may result in the power supply path still passing through abnormal areas or potential fault propagation channels, this invention provides solutions in the following aspects.

[0005] In the first aspect, a power supply protection method for a measuring switch includes:

[0006] Collect and preprocess multi-dimensional electrical quantity data of each node in the distribution network within a preset time period at a fixed frequency.

[0007] Improve the cost function in the path search algorithm;

[0008] Based on the improved cost function, the optimal path node sequence from the power source point to the target load point is searched in the distribution network, and control instructions for the measurement switch are generated according to the optimal path node sequence.

[0009] The improved cost function includes: for each node, dividing the collected electrical quantity data into multiple local windows based on a sliding window of a preset length, and calculating the degree of electrical quantity anomaly in each local window to obtain the electrical quantity anomaly degree sequence for each node; calculating the anomaly propagation correlation between two nodes based on the node's impedance matrix and the electrical quantity anomaly degree sequence; and improving the cost function of the path search algorithm by combining the average electrical quantity anomaly degree of the nodes and the anomaly propagation correlation between nodes.

[0010] Preferably, the cost function of the path search algorithm is improved by combining the average electrical quantity anomaly degree of the nodes and the anomaly propagation correlation degree between nodes. This improvement includes: during the path search process, any neighboring node reachable from the current node is taken as a candidate node; the node with the largest average electrical quantity anomaly degree sequence among all the neighboring nodes of the current node is found and taken as the first reference node; the node with the largest average electrical quantity anomaly degree sequence among all the neighboring nodes of the candidate node is found and taken as the second reference node; and the cost function is improved based on the anomaly propagation correlation degree between the current node and the first reference node, between the candidate node and the second reference node, between the current node and the candidate node, and the average electrical quantity anomaly degree sequence of the current node and the candidate node.

[0011] Preferably, the improved cost function includes: setting a zero-prevention constant, calculating the sum of the anomaly propagation correlation degree between the current node and the first reference node and the zero-prevention constant, and taking the ratio of the anomaly propagation correlation degree between the candidate node and the second reference node to the aforementioned sum as the first ratio; selecting the maximum value among the mean values ​​of the electrical quantity anomaly degree sequences of the current node and the candidate node, and multiplying the maximum value among the mean values, the first ratio, and the anomaly propagation correlation degree between the current node and the candidate node to obtain the improved cost function.

[0012] Preferably, the method for calculating the degree of electrical quantity anomaly in each local window includes: for each local window, calculating the weight of each dimension based on the mean of the electrical quantity data of each dimension in the local window and the previous local window; calculating the coefficient of variation of each dimension based on the standard deviation and mean of the electrical quantity data of each dimension in the local window; summing the product of the weight and the coefficient of variation of each dimension, and normalizing the resulting sum to obtain the degree of electrical quantity anomaly.

[0013] Preferably, the weight of each dimension is calculated based on the mean of electrical quantity data in each dimension within the local window and the previous local window. This includes: for a single dimension, calculating the ratio of the mean of electrical quantity data in that dimension within the local window to the mean of electrical quantity data in that dimension within the previous local window, and using this ratio as a second ratio; calculating the absolute value of the difference between the first and the second ratio, and then normalizing the obtained absolute value to obtain the weight of that dimension.

[0014] Preferably, calculating the coefficient of variation for each dimension based on the standard deviation and mean of the electrical quantity data for each dimension within the local window includes: for a single dimension, calculating the standard deviation and mean of the electrical quantity data for that dimension within the local window, and using the ratio of the standard deviation to the mean as the coefficient of variation for that dimension.

[0015] Preferably, the calculation of the anomaly propagation correlation between two nodes based on the impedance matrix of the nodes and the electrical quantity anomaly degree sequence includes: calculating the electrical distance between the two nodes based on the impedance matrix of the nodes, and performing negative correlation normalization on the electrical distance; calculating the absolute value of the Pearson correlation coefficient between the electrical quantity anomaly degree sequences of the two nodes; and multiplying the absolute value of the Pearson correlation coefficient by the normalized electrical distance to obtain the anomaly propagation correlation between the two nodes.

[0016] Preferably, the multi-dimensional electrical quantity data includes voltage, current, and power.

[0017] Secondly, a power supply protection system for a measuring switch includes a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the above-mentioned power supply protection method for the measuring switch is implemented.

[0018] The present invention has the following effects:

[0019] 1. This invention transforms multi-dimensional electrical quantity data into a sequence of abnormal electrical quantity levels at nodes through sliding window analysis, and defines the correlation degree of abnormal propagation by combining the node impedance matrix. This enables the path search algorithm to perceive the health status of each node in the power grid and the propagation risk of abnormal fluctuations between nodes in real time and quantitatively, providing an accurate data foundation for safety decision-making.

[0020] 2. This invention integrates three core elements: the degree of anomaly of a node itself, the degree of propagation correlation between nodes, and the local risk trend of path extension. It guides the search algorithm to avoid nodes with obvious anomalies, actively cuts off electrical channels with high anomaly propagation risk, and encourages paths to migrate from high-risk areas to low-risk areas. This forms a defense-in-depth system, ensuring that the optimal path node sequence obtained by the search is the power supply channel with the lowest comprehensive anomaly risk and the highest safety margin. It effectively solves the major safety hazard of traditional methods that may lead loads to dangerous areas due to ignoring anomaly propagation, and greatly reduces the probability of cascading failures or safety accidents during fault recovery. Attached Figure Description

[0021] Figure 1 This is a flowchart of steps S1-S3 in a power supply protection method for a measuring switch according to an embodiment of the present invention.

[0022] Figure 2 This is a schematic diagram of steps S20-S22 in a power supply protection method for a measuring switch according to an embodiment of the present invention. Detailed Implementation

[0023] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are some embodiments of the present invention, but not all embodiments.

[0024] The specific embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0025] Reference Figure 1 A power supply protection method for a measuring switch includes steps S1-S3, as detailed below:

[0026] S1: Collect electrical quantity data of each node in the distribution network within a preset time period at a fixed frequency and preprocess it.

[0027] A power supply node network model representing the distribution network topology is constructed. Based on this model, multi-dimensional electrical quantity data of each node are synchronously acquired at a fixed acquisition frequency (e.g., 10Hz) within a preset time period. The multi-dimensional electrical quantity data includes, but is not limited to, voltage, current, and power. The specific set of dimensions can be configured and optimized according to actual monitoring needs, equipment capabilities, and algorithm performance.

[0028] Both the preset time period and the acquisition frequency are configurable hyperparameters that can be adjusted according to the monitoring requirements and hardware performance of specific application scenarios. In this embodiment, the preset time period is set to one hour. This setting is consistent with the update cycle of the measurement switch control command, thereby enabling a complete analysis and control strategy update to be performed once per hour.

[0029] The electrical quantity data for each dimension collected are preprocessed to ensure the accuracy and reliability of subsequent analysis. First, timestamp deviations caused by minor communication or acquisition delays between electrical quantity data at each node and in each dimension are corrected to ensure that all data are on a unified time reference. Second, filtering algorithms (such as moving average filtering and wavelet denoising) are used to smooth the data and suppress high-frequency noise introduced by measurement errors or transient interference. Finally, missing points in the data stream are detected and located, and linear interpolation is used to reasonably estimate and fill in the missing values ​​to maintain the continuity of the data sequence.

[0030] S2: Improve the cost function in the path search algorithm.

[0031] Reference Figure 2 Step S2 includes steps S20-S22, as follows:

[0032] S20: For each node, the collected electrical quantity data is divided into multiple local windows based on a sliding window of preset length, and the degree of electrical quantity anomaly in each local window is calculated to obtain the electrical quantity anomaly sequence for each node.

[0033] To capture and analyze the short-term dynamic changes and time-series anomaly patterns of electrical quantity data within a preset time period, this method uses the sliding window technique to segment and analyze the preprocessed continuous data stream.

[0034] Setting a sliding window with a duration of 1 minute and a sliding step size of 30 seconds, this overlapping sliding method ensures continuous coverage of data change boundaries and avoids missing critical anomalies occurring at the window edge due to window truncation.

[0035] The one-hour-long multi-dimensional electrical quantity data is divided into multiple continuous, partially overlapping local windows by using a sliding window. Each local window contains the time series of electrical quantity data for each dimension.

[0036] For each local window, the weight of each dimension is calculated based on the mean of the electrical quantity data for each dimension within that local window and the previous local window. Specifically: for a single dimension, the ratio of the mean of the electrical quantity data for that dimension within the local window to the mean of the electrical quantity data for that dimension within the previous local window is calculated and used as the second ratio; the absolute value of the difference between 1 and the second ratio is calculated, and the obtained absolute value is normalized to obtain the weight of that dimension, thus obtaining the weight of each dimension. This is expressed by the following formula:

[0037]

[0038] In the formula, Represents a computing node In the When the electrical quantity anomaly degree of the local window is the first Weights of electrical quantity data in each dimension; Represents a node The The first local window The mean of electrical quantity data in the dimension; Represents a node The The first local window The mean of electrical quantity data in the dimension; This represents the normalization function, which makes the sum of the weights of all electrical quantity data dimensions equal to 1.

[0039] When the value is close to 0, it indicates that the electrical quantity data in this dimension remains stable within two consecutive local windows, without significant abrupt changes, and belongs to a normal and stable operating state. A value significantly greater than 0 indicates a drastic jump in the electrical quantity data for that dimension between two consecutive local windows. Such large fluctuations on a short timescale are typical characteristic signals of abnormal events such as line faults, load switching, and intermittent grounding. Therefore, The larger the value, the more drastic the data mutation, and the higher the probability of anomalies in the electrical quantity data of the corresponding dimension.

[0040] It should be noted that, due to the nodes The first local window has no previous local window, so nodes collected in the previous preset time period (i.e., the previous hour) can be used. The last local window of the multi-dimensional electrical quantity data is used as the previous local window of the first local window in this analysis process.

[0041] The coefficient of variation for each dimension is calculated based on the standard deviation and mean of the electrical quantity data for each dimension within the local window. Specifically: for a single dimension, the standard deviation and mean of the electrical quantity data for that dimension are calculated within the local window, and the ratio of the standard deviation to the mean is used as the coefficient of variation for that dimension. This is expressed by the following formula:

[0042]

[0043] In the formula, Represents a computing node In the When the electrical quantity anomaly degree of the local window is the first Coefficient of variation of dimension; Represents a node The The first local window Standard deviation of electrical quantity data in the dimension; Represents a node The The first local window The mean of electrical quantity data in the dimension.

[0044] Reflects the nodes The The first local window The higher the coefficient of variation of the electrical quantity data in a given dimension, the more uneven the change in the electrical quantity data in that dimension within that local window, and the greater the possibility of anomalies.

[0045] The calculation method for the degree of electrical quantity anomaly in each local window includes: summing the products of the weight and coefficient of variation for each dimension, and then normalizing the sum to obtain the degree of electrical quantity anomaly. The specific formula is as follows:

[0046]

[0047] In the formula, Represents a node The The degree of electrical anomaly in a local window; Represents a computing node In the When the electrical quantity anomaly degree of the local window is the first Weights of electrical quantity data in each dimension; Represents a computing node In the When the electrical quantity anomaly degree of the local window is the first Coefficient of variation of dimension; Represents a node The total number of local windows.

[0048] After obtaining the electrical quantity anomaly level of each local window, the electrical quantity anomaly level of each local window constitutes the electrical quantity anomaly level sequence of the corresponding node.

[0049] Traditional methods typically rely on single electrical quantities or fixed thresholds for judgment, which are insufficient to comprehensively reflect complex faults. This invention overcomes the limitations of single-dimensional criteria by weightedly fusing abnormal information from multiple dimensions such as voltage, current, and power (represented by the coefficient of variation), and constructs a quantitative index that can comprehensively reflect the overall abnormal state of a node. This significantly improves the comprehensiveness and reliability of anomaly detection.

[0050] S21: Calculate the correlation degree of anomaly propagation between two nodes based on the impedance matrix of the nodes and the sequence of electrical quantity anomalies.

[0051] Traditional fault location methods typically treat nodes in the power grid as independent units, judging anomalies solely based on whether the node's electrical quantity data exceeds limits. This completely ignores the inherent electrical connections and coupling relationships between nodes, making it impossible to assess which related nodes a fault's abnormal fluctuations will propagate to, and the intensity of the propagation risk. It lacks a quantitative description of the fault propagation path. During fault isolation or power restoration, inaccurate assessment of the fault's impact range may lead to incomplete isolation or restoration paths that mistakenly pass through areas with hidden risks, thus affecting power supply safety and reliability.

[0052] To overcome the above-mentioned shortcomings, this invention proposes to introduce abnormal propagation correlation degree as a quantitative indicator of risk between nodes.

[0053] The method for calculating the anomaly propagation correlation degree includes: calculating the electrical distance between two nodes based on the impedance matrix of the nodes, and performing negative correlation normalization on the electrical distance; calculating the absolute value of the Pearson correlation coefficient between the anomaly degree sequences of the electrical quantities of the two nodes; and multiplying the absolute value of the Pearson correlation coefficient by the normalized electrical distance to obtain the anomaly propagation correlation degree between the two nodes. The specific formula is as follows:

[0054]

[0055] In the formula, Represents a node With nodes The degree of correlation in abnormal propagation between them; Indicates node-based and nodes The electrical distance between two nodes is obtained by calculating the impedance matrix. The specific calculation process is existing technology and will not be described in detail here. Represented by natural constant An exponential function with base 1; Represents a node The sequence of electrical quantity anomalies; Represents a node The sequence of electrical quantity anomalies; Represents a node Electrical quantity anomaly sequence and node The Pearson correlation coefficient between the electrical quantity anomaly sequences.

[0056] The electrical coupling strength between nodes was quantified. When two nodes are electrically closely connected (strongly coupled), their self-impedances are similar and their mutual impedances are close to their self-impedances, resulting in a small calculated electrical distance. Conversely, when the electrical connection between two nodes is weak, the electrical distance is larger.

[0057] electrical distance The smaller, The closer to 1, the stronger the node. Abnormal fluctuations are more likely to propagate to nodes Electrical distance The larger, A value closer to 0 indicates that the abnormal propagation is more difficult.

[0058] This reflects the correlation between the abnormal fluctuation patterns of the two nodes in the time dimension. The closer it is to 1, the stronger the synchronization of the electrical quantity anomaly sequence between the two nodes, and the greater the temporal correlation of the abnormal fluctuations. The closer it is to 0, the more independent the fluctuation patterns of the two are.

[0059] By fusing electrical distance, which characterizes spatial coupling, and Pearson correlation coefficient, which characterizes temporal evolution correlation, the anomaly propagation correlation degree defined in this invention can comprehensively and quantitatively identify potential anomaly propagation relationships between nodes. This anomaly propagation correlation degree provides a core quantitative basis for subsequent fault propagation path analysis, enabling the path search algorithm not only to proactively avoid nodes with obvious anomalies but also to effectively block high-risk (i.e., high-risk) nodes. This identifies potential abnormal propagation paths, thereby significantly improving the reliability of power supply path selection and the overall security of the power supply system.

[0060] S22: The cost function of the path search algorithm is improved by combining the average degree of electrical anomaly of the nodes and the correlation of anomaly propagation between nodes.

[0061] During the path search process, any neighboring node reachable from the current node is selected as a candidate node. The node with the largest mean of the electrical quantity anomaly sequence among all neighboring nodes of the current node is then selected as the first reference node. Similarly, the node with the largest mean of the electrical quantity anomaly sequence among all neighboring nodes of the candidate node is selected as the second reference node. Based on step S21, the anomaly propagation correlation between the current node and the first reference node, the anomaly propagation correlation between the candidate node and the second reference node, and the anomaly propagation correlation between the current node and the candidate nodes are calculated.

[0062] Set a zero-prevention constant, calculate the sum of the anomaly propagation correlation degree between the current node and the first reference node and the zero-prevention constant, and take the ratio of the anomaly propagation correlation degree between the candidate node and the second reference node to the aforementioned sum as the first ratio; select the maximum value among the mean values ​​of the electrical quantity anomaly degree sequences of the current node and the candidate node, and multiply the maximum value among the mean values, the first ratio, and the anomaly propagation correlation degree between the current node and the candidate node to obtain the improved cost function.

[0063] The improved cost function is expressed by the following formula:

[0064]

[0065] In the formula, Represents a node Connect to adjacent nodes The cost of time; Represents a node The mean of the electrical quantity anomaly sequence; Represents a node The mean of the electrical quantity anomaly sequence; This means that the maximum mean value is selected for cost calculation. The higher the degree of anomaly, the greater the cost, which can avoid introducing abnormal nodes during the path search process to the greatest extent. Represents a node With nodes The degree of correlation in abnormal propagation between them; Indicates from node The node with the largest mean of the electrical quantity anomaly sequence among all reachable neighboring nodes; Represents a node With nodes The degree of correlation in abnormal propagation between them; Indicates from node The node with the largest mean of the electrical quantity anomaly sequence among all reachable neighboring nodes; Represents a node With nodes The degree of correlation in abnormal propagation between them; This represents a very small zero-prevention constant to avoid the denominator being zero.

[0066] Used to quantify changes in proximity to abnormal nodes during path search. The smaller the value, the more likely it is to originate from the node. Move to node At this time, the path is moving away from the abnormal node, the risk of abnormal propagation is reduced, so the connection cost should be reduced accordingly; The larger the value, the closer the path is to an abnormal node, the higher the risk, and the greater the connection cost should be.

[0067] The improved cost function proposed in this scheme introduces three dynamic factors: the real-time electrical anomaly level of nodes, the correlation of anomaly propagation between nodes, and the local risk change trend in the path extension direction. This enables the algorithm to perceive the health status and risk distribution of the power grid in real time. The optimal path searched is essentially the power supply channel with the lowest comprehensive anomaly risk and the highest safety margin, avoiding the major safety hazards that traditional methods may cause the load to be directed to abnormal areas or fault propagation paths.

[0068] S3: Based on the improved cost function, search for the optimal path node sequence from the power source point to the target load point in the distribution network, and generate control commands for the measuring switches according to the optimal path node sequence to control the closing or opening of the corresponding switches.

[0069] Based on the improved cost function, using the classic... The heuristic search algorithm (A-Star Search Algorithm) finds the optimal power supply path, i.e., the optimal path node sequence, in the distribution network topology that minimizes the total cost from the power source point to the target load point. Based on the optimal path node sequence, the system automatically generates corresponding measurement switch control commands to close all series-connected switches on the optimal power supply path to construct a complete power supply path; simultaneously, it disconnects other switches unrelated to the optimal power supply path to achieve electrical isolation from non-power supply areas.

[0070] This solution enables the measurement switch control system to make the risk-optimal power supply decision autonomously based on real-time big data. In the event of a fault, the system can automatically and quickly generate a reliable and safe power supply path, effectively shortening the power outage time and preventing the fault from escalating.

[0071] This application also discloses a power supply protection system for a measuring switch. The system includes a processor and a memory. The memory stores computer program instructions. When the computer program instructions are executed by the processor, the power supply protection method for the measuring switch according to the above embodiments of the present invention is implemented.

[0072] The system also includes other components well known to those skilled in the art, such as communication buses and communication interfaces, the settings and functions of which are known in the art and will not be described in detail here.

[0073] It should be noted that those skilled in the art can make various modifications and improvements without departing from the inventive concept, and these all fall within the scope of protection of this invention. Therefore, the scope of protection of this patent should be determined by the appended claims.

Claims

1. A power supply protection method for a measuring switch, characterized in that, include: Collect and preprocess multi-dimensional electrical quantity data of each node in the distribution network within a preset time period at a fixed frequency. Improve the cost function in the path search algorithm; Based on the improved cost function, the optimal path node sequence from the power source point to the target load point is searched in the distribution network, and control instructions for the measurement switch are generated according to the optimal path node sequence. The improved cost function method includes: for each node, the collected electrical quantity data is divided into multiple local windows based on a sliding window of a preset length, and the degree of electrical quantity anomaly in each local window is calculated to obtain the electrical quantity anomaly degree sequence for each node; the anomaly propagation correlation degree between two nodes is calculated based on the impedance matrix of the node and the electrical quantity anomaly degree sequence. During the path search process, any neighboring node reachable from the current node is taken as a candidate node. The node with the largest mean of the electrical quantity anomaly degree sequence among all the neighboring nodes of the current node is found and used as the first reference node. The node with the largest mean of the electrical quantity anomaly degree sequence among all the neighboring nodes of the candidate node is found and used as the second reference node. Set a zero-prevention constant, calculate the sum of the anomaly propagation correlation degree between the current node and the first reference node and the zero-prevention constant, and take the ratio of the anomaly propagation correlation degree between the candidate node and the second reference node to the aforementioned sum as the first ratio; select the maximum value among the mean values ​​of the electrical quantity anomaly degree sequences of the current node and the candidate node, and multiply the maximum value among the mean values, the first ratio, and the anomaly propagation correlation degree between the current node and the candidate node to obtain the improved cost function.

2. The power supply protection method for a measuring switch according to claim 1, characterized in that, The calculation method for the degree of electrical quantity anomaly in each local window includes: for each local window, calculating the weight of each dimension based on the mean of the electrical quantity data of each dimension in the local window and the previous local window; calculating the coefficient of variation of each dimension based on the standard deviation and mean of the electrical quantity data of each dimension in the local window; summing the product of the weight and the coefficient of variation of each dimension, and normalizing the resulting sum to obtain the degree of electrical quantity anomaly.

3. The power supply protection method for a measuring switch according to claim 2, characterized in that, Based on the mean of electrical quantity data for each dimension within the local window and the previous local window, the weight of each dimension is calculated as follows: For a single dimension, the ratio of the mean of electrical quantity data for that dimension within the local window to the mean of electrical quantity data for that dimension within the previous local window is calculated and used as the second ratio; the absolute value of the difference between 1 and the second ratio is calculated, and the obtained absolute value is normalized to obtain the weight of that dimension.

4. The power supply protection method for a measuring switch according to claim 2, characterized in that, The coefficient of variation for each dimension is calculated based on the standard deviation and mean of the electrical quantity data for each dimension within the local window. This includes: for a single dimension, calculating the standard deviation and mean of the electrical quantity data for that dimension within the local window, and using the ratio of the standard deviation to the mean as the coefficient of variation for that dimension.

5. The power supply protection method for a measuring switch according to claim 1, characterized in that, The calculation of the anomaly propagation correlation between two nodes based on the node impedance matrix and the electrical quantity anomaly sequence includes: calculating the electrical distance between the two nodes based on the node impedance matrix and performing negative correlation normalization on the electrical distance; calculating the absolute value of the Pearson correlation coefficient between the electrical quantity anomaly sequences of the two nodes; and multiplying the absolute value of the Pearson correlation coefficient by the normalized electrical distance to obtain the anomaly propagation correlation between the two nodes.

6. The power supply protection method for a measuring switch according to claim 1, characterized in that, The multi-dimensional electrical quantity data includes voltage, current, and power.

7. A power supply protection system for a measuring switch, characterized in that, include: A processor and a memory, the memory storing computer program instructions that, when executed by the processor, implement the power supply protection method for a measuring switch according to any one of claims 1-6.

Images