Power distribution network fault intelligent positioning and response system

Abstract

The invention relates to the technical field of fault diagnosis, in particular to a power distribution network fault intelligent positioning and response system, which comprises a data conversion module, a power distribution network fault intelligent positioning and response module, a power distribution network fault intelligent positioning and response module, a power distribution network fault intelligent positioning and response module and a power distribution network fault intelligent positioning and response module, the voltage fluctuation and the circuit breaker state are represented, and power distribution network graph data are generated. According to the invention, through converting the operation data of the power distribution network into the point cloud and utilizing a persistent coherence method, topology analysis is carried out on the power grid state, and through capturing and analyzing changes of rings and holes in graphic data, topology changes related to voltage fluctuation and load abnormity are identified, so that the accuracy of fault diagnosis and the perspectiveness of fault prediction are improved; fault points are monitored in real time, time sequence analysis is adopted to predict the time and place of potential faults, the fault response speed and precision are accelerated, and stable operation of the power system is enhanced through a data analysis and prediction mechanism.

Description

Technical Field

[0001] The present invention relates to the technical field of fault diagnosis, and in particular to a distribution network fault intelligent positioning and response system. Background Art

[0002] The field of fault diagnosis technology covers a range of methods and tools designed to detect, analyze and identify anomalies or failures in systems, equipment or components. These technologies are widely used in various industrial and technological systems, such as power networks, manufacturing machinery, communication facilities, etc. The core purpose is to improve the reliability and efficiency of the system, reduce unexpected downtime, and optimize maintenance and repair operations. Fault diagnosis usually relies on sensor data, historical maintenance records and intelligent algorithms to predict potential problems and intervene in advance, so that intervention can be made before the fault develops into a more serious problem. With the advancement of technology, fault diagnosis has evolved from traditional manual detection methods to more automated and intelligent solutions, including the integration of machine learning and artificial intelligence technologies.

[0003] Among them, the intelligent location and response system for distribution network faults refers to an intelligent fault handling technology specifically used in power systems. The main purpose of this system is to automatically detect the location of faults in the distribution network and respond quickly to minimize the impact of service interruptions. Through real-time data analysis and the application of historical fault data, the system can predict and identify fault types, automatically isolate affected areas, and redirect current to restore power supply. This not only improves the operating efficiency and reliability of the power grid, but also significantly reduces the economic losses and safety risks caused by faults.

[0004] Existing technologies mainly rely on basic sensor data and historical maintenance records, and their fault handling usually fails to reflect the dynamic changes of complex power grid status in real time. In particular, in dealing with atypical faults and predicting potential problems, they often show limitations due to the lack of high-dimensional data analysis. As a result, when faced with sudden voltage fluctuations or circuit breaker events, conventional technologies may not be able to accurately locate the source of the fault immediately, increasing the time for diagnosis and response. In addition, because the technology fails to predict the development of faults in a timely manner, the best time for intervention is often missed, which increases the risk of power grid operation and affects the continuity and safety of power supply. Summary of the invention

[0005] The purpose of the present invention is to solve the shortcomings of the prior art and to propose a distribution network fault intelligent positioning and response system.

[0006] In order to achieve the above object, the present invention adopts the following technical solution: A distribution network fault intelligent positioning and response system comprises:

[0007] The data conversion module collects the operation data of the distribution network and converts it into point clouds. Each point cloud represents an operation state of the distribution network. By calculating the fit between the operation states and building connections between the point clouds, the voltage fluctuation and circuit breaker state are represented, and the distribution network graphic data is generated.

[0008] The topology feature analysis module receives the distribution network graph data, uses the persistent coherence method to analyze the formation and dissipation of loops and holes in the graph, pays attention to the impact of load anomalies and voltage fluctuations on loops and holes, analyzes the changes of loops and holes, and generates topology feature dynamics;

[0009] The fault point identification module uses the topological feature dynamics to compare the changes of the rings and holes with the distribution network fault events, identifies the topological changes associated with the fault, and locates the fault point through the changes of the rings and holes;

[0010] The fault prediction module continuously monitors the state changes of the fault point, uses the change trend to detect abnormal points in the time series, analyzes the current data to predict the potential time and location of the fault, and generates a fault prediction result.

[0011] As a further solution of the present invention, the step of acquiring the distribution network graphic data is specifically as follows:

[0012] Collect voltage, current and circuit breaker status data from the distribution network monitoring system, organize the data format and generate raw operation data sets based on the timestamp and geographic information of the operation data;

[0013] Converting the original operation data set into a point cloud format, each point cloud including voltage fluctuation and circuit breaker status information, and mapping the time series data into a high-dimensional space using a coordinate transformation method to generate a point cloud data set;

[0014] Based on the point cloud data set, the fit between point clouds is calculated using the formula:

[0015]

[0016] Compare the multi-dimensional features between multiple point clouds and generate a set of fitting results;

[0017] Among them, a i and b i They represent the eigenvalues ​​of the two compared point clouds in the i-th dimension, and S represents the fitting scores of the two point clouds;

[0018] The fitness result set is used to construct a connection graph between point clouds. The connection graph represents the fitness between point clouds through edge weights. The connection priority is determined according to the weights to generate graphical data of the distribution network.

[0019] As a further solution of the present invention, the steps of analyzing the formation and dissipation of rings and holes in the graph are specifically as follows:

[0020] receiving the distribution network graphic data, converting the distribution network graphic data into a graphic format, including key voltage and load status information, and generating graphic data;

[0021] Applying a persistent coherence method to identify loops and holes in the graphical data, drawing current paths and potential disconnection points in the power grid, and generating preliminary identification results of loops and holes;

[0022] Analyze the preliminary identification results of the rings and voids, track the formation and dissipation process, and record the information of topological structure changes over time. The formula is:

[0023]

[0024] Generates ring and cavity variation analysis results;

[0025] Where C(t) represents the change of rings and holes at time t, λ i Indicates the sensitivity of the target ring or cavity to voltage fluctuations, V i (t) is the voltage fluctuation of the ith ring or hole.

[0026] As a further solution of the present invention, the step of dynamically acquiring the topological features is specifically as follows:

[0027] Receiving the ring and hole change data, the data records the change of the topological structure in the power grid, and generating the ring and hole change data;

[0028] Using multi-scale topological analysis technology to process the change data of the rings and voids, tracking the changes of each ring and void under differentiated time and voltage fluctuation conditions, and generating dynamic tracking results of the rings and voids;

[0029] Integrate the ring and hole dynamic tracking results, using the formula:

[0030]

[0031] Calculate the topological feature dynamics accumulated over generation time t;

[0032] Where T(t) represents the accumulated topological feature dynamics, ω k To represent the weight of the target loop or hole in the topology analysis, reflecting its importance in the grid stability, d k (t′) is the rate of change of the ring or hole at time t′, which represents the topological structure of the change, and α k is the adjustment factor, referring to the effect of voltage fluctuations on the sensitivity to rings or voids.

[0033] As a further solution of the present invention, the step of obtaining the fault point is specifically as follows:

[0034] Extracting loop and hole data in the topological feature dynamics, the data records the topological structure changes at key time points in the power grid, and generating loop and hole time series data;

[0035] Comparing and analyzing the ring and hole time series data with the fault event records, determining the correlation between the fault event and the target topology change through statistical correlation analysis, and generating fault-correlated topology change analysis results;

[0036] Using the fault-related topology change analysis results, perform topology change tracking and apply the formula:

[0037]

[0038] Locate the fault point;

[0039] Where P(x) represents the fault point prediction model, βi represents the importance weight of the i-th topology change indicator, which is used to weight the influence of the differentiation indicator, γi is the fault response sensitivity coefficient, which is used to adjust the weight to reflect the actual fault sensitivity, and x i It is the topology change monitored in real time.

[0040] As a further solution of the present invention, the steps of obtaining the fault prediction result are specifically as follows:

[0041] Deploy a sensor network to continuously monitor the state of the power grid near the fault point, record key parameters of current, voltage and load, form a real-time data stream, and generate real-time state monitoring data;

[0042] Based on the real-time status monitoring data, applying a time series analysis method to perform anomaly detection, identify data patterns that are estimated to indicate faults, and generate anomaly detection results;

[0043] According to the abnormal detection results, the prediction algorithm is used to analyze the time and location of the fault, using the formula:

[0044]

[0045] Generate fault prediction results;

[0046] Where F(t) represents the risk function of failure prediction, α j represents the influence of differentiated data indicators in the fault prediction model, β j is the time attenuation coefficient, which adjusts the sensitivity of the time span to the impact of the fault, d j are the data metrics identified from anomaly detection.

[0047] Compared with the prior art, the advantages and positive effects of the present invention are:

[0048] In the present invention, by converting the distribution network operation data into a point cloud and using the persistent coherence method, this technical solution can perform a more detailed topological analysis of the power grid state. This method captures and analyzes the changes in loops and holes in the graphic data, accurately identifies topological changes related to voltage fluctuations and load anomalies, and greatly improves the accuracy of fault diagnosis and the foresight of predicting faults. In addition, real-time monitoring of fault points and using time series analysis to predict the time and location of potential faults can significantly speed up the fault response speed and accuracy. This in-depth data analysis and prediction mechanism effectively enhances the stable operation of the power system and significantly reduces the impact caused by faults. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] Figure 1 is a system flow chart of the present invention;

[0050] Figure 2 A flow chart of the steps for obtaining the distribution network graphic data of the present invention;

[0051] Figure 3 A flow chart of the steps for analyzing the formation and dissipation of rings and voids in a graph according to the present invention;

[0052] Figure 4 A flowchart of the steps for obtaining the dynamic topological features of the present invention;

[0053] Figure 5 A flowchart of the steps for obtaining the fault point of the present invention;

[0054] Figure 6 This is a flow chart of the steps for obtaining the fault prediction results of the present invention. DETAILED DESCRIPTION

[0055] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0056] In the description of the present invention, it should be understood that the terms "length", "width", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside" and the like indicate positions or positional relationships based on the positions or positional relationships shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, in the description of the present invention, "multiple" means two or more, unless otherwise clearly and specifically defined.

[0057] Embodiment 1

[0058] See also Figure 1 , a distribution network fault intelligent positioning and response system includes:

[0059] The data conversion module collects the operation data of the distribution network and converts it into point clouds. Each point cloud represents an operation state of the distribution network. By calculating the fit between the operation states and building connections between the point clouds, the voltage fluctuation and circuit breaker state are represented, and the distribution network graphic data is generated.

[0060] The topology feature analysis module receives distribution network graph data, uses the persistent coherence method to analyze the formation and dissipation of loops and holes in the graph, pays attention to the impact of load anomalies and voltage fluctuations on loops and holes, analyzes the changes of loops and holes, and generates topology feature dynamics;

[0061] The fault point identification module uses topological features dynamically to compare the changes in rings and holes with distribution network fault events, identify topological changes associated with the fault, and locate the fault point through the changes in rings and holes;

[0062] The fault prediction module continuously monitors the state changes of the fault point, uses the change trend to detect abnormal points in the time series, analyzes the current data to predict the potential time and location of the fault, and generates a fault prediction result.

[0063] The distribution network graphic data specifically include voltage fluctuation graphics and circuit breaker status graphics. The topological feature dynamics include ring dynamics and void dynamics. The fault point identification results specifically include fault-related topological changes. The fault prediction results include fault time prediction and fault location prediction.

[0064] See also Figure 2 ,The specific steps for obtaining distribution network graphical data are as follows:

[0065] Collect voltage, current and circuit breaker status data from the distribution network monitoring system, organize the data format and generate raw operation data sets based on the timestamp and geographic information of the operation data;

[0066] In this step, the voltage, current and circuit breaker status data in the monitoring system are sorted according to timestamps and geographic information. For the voltage data V, according to the formula Calculate the average voltage of multiple data points;

[0067] Where V i represents the voltage of the ith data point, n represents the total number of data points, V avg Represents the calculated average voltage.

[0068] Detailed explanation of the formula and the process of formula calculation and derivation:

[0069] To calculate the average voltage in the distribution network, each voltage data point V i The total voltage is then divided by the total number of data points n to obtain the average value of the overall voltage V avg It not only provides a baseline for the data but also provides a reference for outlier detection. For example, if the collected data points are V1=120, V2=125, V3=118, and n=3, then the average voltage V avg The calculation is as follows:

[0070]

[0071] The results show the average voltage level of the distribution network during the time period, providing basic data for subsequent analysis.

[0072] The original operation data set is converted into a point cloud format, each point cloud includes voltage fluctuation and circuit breaker status information, and the time series data is mapped into a high-dimensional space using a coordinate transformation method to generate a point cloud data set;

[0073] In this step, the sorted data is converted into a point cloud format, involving data on voltage fluctuations and circuit breaker status, where the voltage fluctuation ΔV is calculated by the formula ΔV = max(V) - min(V), and the fluctuation characteristics in the point cloud are plotted;

[0074] Where max(V) and min(V) represent the maximum and minimum values ​​of the voltage data during the monitoring period, respectively, and ΔV represents the range of voltage fluctuation.

[0075] Detailed explanation of the formula and the process of formula calculation and derivation:

[0076] To characterize the voltage fluctuation in the point cloud, the difference between the maximum and minimum voltage values ​​collected from the power grid data is calculated to obtain the range of voltage fluctuation. This calculation helps capture the stability and fluctuation range of voltage in the point cloud representation. For example, if the voltage data collected in the same time period is V1=130, V2=120, and V3=135, the voltage fluctuation ΔV is calculated as follows:

[0077] ΔV=135-120=15volts

[0078] The results show the maximum voltage fluctuation range during the observation period, which is critical for analyzing voltage stability and device performance.

[0079] Based on the point cloud data set, the fit between point clouds is calculated using the formula:

[0080]

[0081] Compare the multi-dimensional features between multiple point clouds and generate a set of fitting results;

[0082] Among them, a i and b i They represent the eigenvalues ​​of the two compared point clouds in the i-th dimension, and S represents the fitting scores of the two point clouds;

[0083] formula:

[0084]

[0085] Detailed explanation of the formula and the process of formula calculation and derivation:

[0086] The formula is based on the improvement of cosine fit and evaluates the fit between differentiated point clouds in the point cloud dataset. The fit between point clouds is obtained by calculating the dot product of the eigenvalues ​​of each dimension and dividing it by the modulus of each eigenvector. For example, there are two point cloud eigenvalue vectors a=[1,2,3] and b=[2,1,2], then the fit S is calculated as follows:

[0087]

[0088] The results show that the two point clouds have a high degree of fit in the high-dimensional space, indicating that they represent similar operating states or grid conditions.

[0089] Using the fit result set, a connection graph between point clouds is constructed. The connection graph represents the fit between point clouds through the weights of the edges. The priority of the connection is determined according to the weights, and the graphical data of the distribution network is generated.

[0090] In this step, the connection graph between point clouds is constructed using the fit results S between point clouds. The weight w of the edge is determined by the formula w = f(S), where f(S) is a conversion function that converts the fit into a weight value to represent the strength of the connection;

[0091] In the formula, S represents the fit between point clouds, and w represents the weight of the edge in the constructed connection graph.

[0092] Detailed explanation of the formula and the process of formula calculation and derivation:

[0093] When constructing the point cloud connection graph, the fit S is converted into the edge weight w through a function f, which is used to represent the strength of the connection between point clouds. For example, f(S) = 10×S, and the fit S obtained from the above results is 0.89, then the weight w is calculated as follows:

[0094] w=10×0.89=8.9

[0095] The weight value w indicates the strength of the connection between the two point clouds. A high weight means that the two states are very similar, which is suitable for emphasizing this relationship in the graphical data representation.

[0096] See also Figure 3, the steps for analyzing the formation and dissipation of rings and holes in the graph are as follows:

[0097] receiving distribution network graphic data, converting the distribution network graphic data into a graphic format, including key voltage and load status information, and generating graphic data;

[0098] In receiving the graphical data of the distribution network, the data includes voltage and load status information, according to the formula

[0099] Calculate the sum of multiple node voltages, where F(g) represents the sum of voltages in the graph data and gi represents the voltage value of the i-th node.

[0100] Detailed explanation of the formula and the process of formula calculation and derivation:

[0101] In this step, the total voltage value of the entire distribution network is obtained by summing the voltage of each node, which helps to analyze the impact of voltage fluctuations on the network topology. Formula F(g) represents the accumulation of all node voltages. i Indicates the voltage value recorded by a node in the graph data. For example, there are 5 nodes in the distribution network, and the voltage values ​​are 10V, 20V, 30V, 40V and 50V respectively. The calculation process is:

[0102] F(g)=10+20+30+40+50=150V

[0103] The total 150V represents a voltage overview of the current distribution network status and provides basic data for analysis.

[0104] Apply the persistent coherence method to identify loops and holes in the graphic data, draw the current paths and potential disconnection points in the power grid, and generate preliminary identification results of loops and holes;

[0105] When applying the persistent homology method to identify loops and voids in graphic data, special attention is paid to the current paths and potential breakpoints in the data, and the structural characteristics of loops and voids are calculated according to the formula H(t)=∫sφ(t,x)dx, where H(t) represents the structural characteristics of loops and voids at time t, and φ(t,x) represents the current density function at position x and time t.

[0106] Detailed explanation of the formula and the process of formula calculation and derivation:

[0107] The formula H(t) is used to evaluate how current paths form loops and holes at differentiated time points and indicate potential weaknesses in the grid. In the formula, φ(t, x) is the current density function, the variable x represents the location in the network, and t represents the time variable. The distribution of current density at differentiated locations is set to φ(t, x) = 5t·e -x, and refer to S as a target area [0, 5] in the power grid, the integral calculation is:

[0108]

[0109] The integral is calculated by:

[0110]

[0111] The numerical calculation is done by inserting the t value. For example, if t = 1, then:

[0112] H(1)=5·(1-e -5 )≈4.966

[0113] This means that at time t = 1, the quantitative evaluation of the structural characteristics of loops and holes in the network is 4.966, which helps to understand the topological security status of the network.

[0114] Analyze the preliminary identification results of rings and voids, track the formation and dissipation process, and record the information of topological structure changes over time. The formula is:

[0115]

[0116] Generates ring and cavity variation analysis results;

[0117] Where C(t) represents the change of rings and holes at time t, λ i Indicates the sensitivity of the target ring or cavity to voltage fluctuations, V i (t) is the voltage fluctuation of the ith ring or hole.

[0118] formula:

[0119]

[0120] Detailed explanation of the formula and the process of formula calculation and derivation:

[0121] The formula is used to analyze the changes of the identified rings and holes over time. In the formula, C(t) represents the change of the rings and holes at time t, and λ i is the attenuation coefficient corresponding to the ith ring or hole, reflecting the rate of change over time, while V i (t) is the voltage fluctuation of the ith ring or hole at time t.

[0122] Three rings or holes are set, with attenuation coefficients of λ1=0.1, λ2=0.2, λ3=0.3, and the voltage fluctuations at the target time t are

[0123] V1(t)=100V,V2(t)=150V,V3(t)=200V. Then the formula is as follows:

[0124] C(t)=e -0.1t 100+e -0.2t 150+e -0.3t 200

[0125] For example, when t=1 is set:

[0126] C(1)=e- 0.1·1 100+e -0.2·1 150+e -0.3·1 200

[0127] C(1)=0.904·100+0.819·150+0.741·200

[0128] C(1)=90.4+122.85+148.2

[0129] C(1)=361.45

[0130] The value C(1) = 361.45 indicates that at time t = 1, the total voltage fluctuation impact of the rings and holes in the entire network is 361.45 V. The numerical value helps to assess the stability of the grid at a given time and its potential impact on grid operation.

[0131] See also Figure 4 ,The specific steps for obtaining dynamic topological features are:

[0132] Receive ring and hole change data, which records the changes in the topology of the power grid and generates ring and hole change data;

[0133] Among them, the change data of the receiving ring and the hole is calculated according to the formula P = N total -N effective , calculate the inefficiency of the data.

[0134] In the formula, P represents the ineffectiveness, N total Represents the total number of data points, N effective Represents the number of valid data points.

[0135] Detailed explanation of the formula and the process of formula calculation and derivation:

[0136] The formula is used to calculate the proportion of invalid data points in the data obtained from the persistent coherence method. Set the total number of data points N in a certain monitoring. total The number of valid data points is 1000. effective is 800 points, then the ineffectiveness P is calculated as follows:

[0137] P = 1000-800 = 200

[0138]

[0139] The results show that 20% of the data points in the current dataset are deemed invalid, indicating that this part of the data cannot be analyzed due to quality issues or monitoring errors.

[0140] Use multi-scale topological analysis technology to process the change data of rings and voids, track the changes of each ring and void under differentiated time and voltage fluctuation conditions, and generate dynamic tracking results of rings and voids;

[0141] Among them, the changes of tracking rings and voids under differentiated time and voltage fluctuation conditions are calculated according to the formula Calculate the average rate of change.

[0142] In the formula, R represents the average rate of change, V change represents the change value of rings and holes in each event, N events Represents the total number of events.

[0143] Detailed explanation of the formula and the process of formula calculation and derivation:

[0144] The formula is used to quantify the average change of rings and voids over time and voltage fluctuations. Set to record 50 events in a period of time, and the change value of rings and voids in each event V change The sum is 250, so the average rate of change R is calculated as follows:

[0145]

[0146] The results show that the average change value of rings and holes in each event is 5, which provides quantitative information on the sensitivity of power grid state changes and helps to evaluate the stability of the power grid and its ability to respond to voltage fluctuations.

[0147] Integrate the ring and hole dynamic tracking results, using the formula:

[0148]

[0149] Calculate the topological feature dynamics accumulated over generation time t;

[0150] Where T(t) represents the accumulated topological feature dynamics, ω k To represent the weight of the target loop or hole in the topology analysis, reflecting its importance in the grid stability, d k (t′) is the rate of change of the ring or hole at time t′, which represents the topological structure of the change, and α k is the adjustment factor, referring to the effect of voltage fluctuations on the sensitivity to rings or voids.

[0151] formula:

[0152]

[0153] Detailed explanation of the formula and the process of formula calculation and derivation:

[0154] The formula integrates multiple variables to calculate the cumulative changes of loops and holes in the power grid over time, and provides data support for analyzing the power grid status. Set the simulation scenario: a monitoring period t is 24 hours, there are 3 loops or holes (m=3), and the importance weights are ω1=0.5, ω2=0.3, ω3=0.2; within the period, the change rate of each loop or hole is d1(t′)=2, d2(t′)=1.5, d3(t′)=1, and the voltage fluctuation adjustment coefficient is α1=1.2, α2=1.0, α3=0.8. Substitute the parameters into the formula for calculation:

[0155]

[0156] The total amount of cumulative topological characteristic dynamics from t = 0 to t = 24 hours was calculated and integrated.

[0157] The results show that the cumulative changes of multiple rings and voids within 24 hours can help analyze the dynamic response capability of the power grid when facing continuous voltage fluctuations.

[0158] See also Figure 5 , the specific steps for obtaining the fault point are:

[0159] Extract ring and hole data in the topological feature dynamics, record the topological structure changes at key time points in the power grid, and generate ring and hole time series data;

[0160] During the analysis, the ring and hole time series data are transformed by the target technical formula and matched with the fault event records. Calculate feature change values.

[0161] In the formula, S(T) represents the transformed sequence characteristic value, t i Represents the i-th data point in the original time series.

[0162] Detailed explanation of the formula and the process of formula calculation and derivation:

[0163] The formula converts the original time series by accumulating the value of each data point, simplifying the comparative analysis. Set the time series data to T = [1, 2, 3, 4, 5], then the calculation process is:

[0164]

[0165] The result S(T)=15 represents the cumulative value of the entire sequence under a certain target treatment.

[0166] Compare and analyze the ring and hole time series data with the fault event records, determine the correlation between the fault event and the target topology change through statistical correlation analysis, and generate fault-related topology change analysis results;

[0167] When determining the correlation between the ring and hole time series data and the fault events, a calculation method is used to evaluate the correlation between the two. The correlation C is calculated according to the formula Calculate, where x i represents the analysis result of fault-related topology changes, y i Represents fault data.

[0168] Detailed explanation of the formula and the process of formula calculation and derivation:

[0169] The formula is the Pearson correlation coefficient, which is used to measure the linear relationship between two data sets. Set x = [1, 2, 3, 4] and y = [2, 3, 2, 5], where and is the average value of each data. The calculation process is as follows:

[0170]

[0171] The calculation result will be the C value, which indicates the strength of the correlation between the two sets of data and is used to determine the correlation between topology changes and faults.

[0172] Use the fault-correlated topology change analysis results to perform topology change tracking and apply the formula:

[0173]

[0174] Locate the fault point;

[0175] Where P(x) represents the fault point prediction model, βi represents the importance weight of the i-th topology change indicator, which is used to weight the influence of the differentiation indicator, γi is the fault response sensitivity coefficient, which is used to adjust the weight to reflect the actual fault sensitivity, and x i It is the topology change monitored in real time.

[0176] formula:

[0177]

[0178] Detailed explanation of the formula and the process of formula calculation and derivation:

[0179] The formula simplifies the prediction of fault points, where each variable has a meaning. For example, setting the weight parameter β i is 1.5, the fault sensitivity coefficient γ i is 2, and the topological change x i If is 4, the calculation process is as follows:

[0180] P(x)=1.5·2·4=12

[0181] The result P(x)=12 represents the index value of the predicted fault point under the given weight and sensitivity. This value helps to understand the possibility and specific location of the fault under the conditions.

[0182] See also Figure 6 , the specific steps for obtaining fault prediction results are:

[0183] Deploy a sensor network to continuously monitor the grid status near the fault point, record key parameters of current, voltage and load, form a real-time data stream, and generate real-time status monitoring data;

[0184] In this step, a sensor network is deployed to continuously monitor the state of the power grid near the fault point, including key parameters of current, voltage, and load. Through the sensor data, a real-time data stream is formed and real-time status monitoring data is generated. This process involves real-time calculation and update of the data stream, which is expressed by the following reference formula:

[0185] R(t)==∫(I(t)·V(t)·L(t))dt

[0186] Where R(t) represents the accumulated status monitoring data, I(t), V(t), and L(t) represent the current, voltage, and load that change with time, respectively.

[0187] Detailed explanation of the formula and the process of formula calculation and derivation:

[0188] R(t) is calculated by integrating the data of current I(t), voltage V(t), and load L(t) collected by the real-time monitoring sensor. In this process, the value of each parameter is obtained through real-time data monitoring, updated and included in the cumulative value. For example, at a certain target time, I(t) = 10A, V(t) =

[0189] 220V, L(t) = 0.8, then the contribution value of the instantaneous state monitoring data at this moment is:

[0190] R(t)=10A×220V×0.8=1760W

[0191] This value indicates the state of the power grid monitored at this moment, reflecting the comprehensive situation of real-time current, voltage and load, and evaluating and warning the health status of the power grid.

[0192] Based on real-time condition monitoring data, time series analysis methods are applied to perform anomaly detection, identify data patterns that are expected to indicate failures, and generate anomaly detection results;

[0193] In this step, based on the real-time status monitoring data, time series analysis is performed to identify anomalies. This is a key analysis process involving the recognition of data patterns and the detection of anomalies. The reference formula is as follows:

[0194]

[0195] Where A(t) represents the anomaly detection value at time t, x i (t) represents the monitoring data point at time t, μ(t) is the average value of the data points, and n is the number of data points.

[0196] Detailed explanation of the formula and the process of formula calculation and derivation:

[0197] By calculating the sum of the squares of the differences between multiple data points and their average values, the degree of abnormality of the data at any given moment can be quantitatively evaluated. i (t) has three values: 10, 12 and 14, with an average value of μ(t) = 12. The anomaly detection value is calculated as follows:

[0198] A(t)=(10-12) 2 +(12-12) 2 +(14-12) 2

[0199] A(t)=4+0+4=8

[0200] The value A(t)=8 describes the degree of abnormality of the monitoring data at the time instant. A higher A(t) value indicates that the data deviates greatly from the average value, indicating the early signs of failure.

[0201] According to the abnormal detection results, the prediction algorithm is used to analyze the time and location of the fault, using the formula:

[0202]

[0203] Generate fault prediction results;

[0204] Where F(t) represents the risk function of failure prediction, α j represents the influence of differentiated data indicators in the fault prediction model, β j is the time attenuation coefficient, which adjusts the sensitivity of the time span to the impact of the fault, d j are the data metrics identified from anomaly detection.

[0205] formula:

[0206]

[0207] formula:

[0208]

[0209] Detailed explanation of the formula and the process of formula calculation and derivation:

[0210] F(t) represents the risk function of failure prediction, which is used to calculate the failure risk at the target time t. Each coefficient α j , β j and d j These are all parameters obtained based on actual data:

[0211] α j Represents the weights of multiple differentiated data indicators in the fault prediction model.

[0212] β j is the time attenuation coefficient, which is used to adjust the sensitivity of the time span to the impact of faults.

[0213] d j is the target data metric identified from anomaly detection.

[0214] For example, there are three data indicators, and the parameters are: α1 = 0.5, β1 = 0.1, d1 = 100; α2 = 0.3, β2 = 0.05, d2 = 150; α3 = 0.2, β3 = 0.2, d3 = 200. At t = 10, the calculation is as follows:

[0215] F(10)=0.5·e -0.1·10 100+0.3 e -0.05·10 150+0.2 e -0.2·10 200

[0216] Perform the calculation:

[0217] F(10)=0.5·e -1 100+0.3 e -0.5 150+0.2 e -2 200

[0218] F(10)=0.5·0.3679·100+0.3·0.6065·150+0.2·0.1353·200

[0219] F(10)=18.395+27.2925+5.412=51.0995

[0220] The value indicates the predicted failure risk at time t=10.

[0221] The above are only preferred embodiments of the present invention and are not intended to limit the present invention in other forms. Any technician familiar with the profession may use the technical contents disclosed above to change or modify them into equivalent embodiments with equivalent changes and apply them to other fields. However, any simple modification, equivalent change and modification made to the above embodiments based on the technical essence of the present invention without departing from the technical solution of the present invention still falls within the protection scope of the technical solution of the present invention.

Claims

1. A distribution network fault intelligent location and response system, characterized in that: The system comprises: The data conversion module collects the operation data of the distribution network and converts it into point clouds. Each point cloud represents an operation state of the distribution network. By calculating the fit between the operation states and building connections between the point clouds, the voltage fluctuation and circuit breaker state are represented, and the distribution network graphic data is generated. The topology feature analysis module receives the distribution network graph data, uses the persistent coherence method to analyze the formation and dissipation of loops and holes in the graph, pays attention to the impact of load anomalies and voltage fluctuations on loops and holes, analyzes the changes of loops and holes, and generates topology feature dynamics; The fault point identification module uses the topological feature dynamics to compare the changes of the rings and holes with the distribution network fault events, identifies the topological changes associated with the fault, and locates the fault point through the changes of the rings and holes; The fault prediction module continuously monitors the state changes of the fault point, uses the change trend to detect abnormal points in the time series, analyzes the current data to predict the potential time and location of the fault, and generates a fault prediction result.

2. The distribution network fault intelligent location and response system according to claim 1, characterized in that: The steps for acquiring the distribution network graphical data are specifically as follows: Collect voltage, current and circuit breaker status data from the distribution network monitoring system, organize the data format and generate raw operation data sets based on the timestamp and geographic information of the operation data; Converting the original operation data set into a point cloud format, each point cloud including voltage fluctuation and circuit breaker status information, and mapping the time series data into a high-dimensional space using a coordinate transformation method to generate a point cloud data set; Based on the point cloud data set, the fit between point clouds is calculated using the formula: Compare the multi-dimensional features between multiple point clouds and generate a set of fitting results; Among them, a i and b i They represent the eigenvalues ​​of the two compared point clouds in the i-th dimension, and S represents the fitting scores of the two point clouds; The fitness result set is used to construct a connection graph between point clouds. The connection graph represents the fitness between point clouds through edge weights. The connection priority is determined according to the weights to generate graphical data of the distribution network.

3. The distribution network fault intelligent location and response system according to claim 2, characterized in that: The steps of forming and dissipating the rings and holes in the analysis graph are specifically as follows: receiving the distribution network graphic data, converting the distribution network graphic data into a graphic format, including key voltage and load status information, and generating graphic data; Applying a persistent coherence method to identify loops and holes in the graphical data, drawing current paths and potential disconnection points in the power grid, and generating preliminary identification results of loops and holes; Analyze the preliminary identification results of the rings and voids, track the formation and dissipation process, and record the information of topological structure changes over time. The formula is: Generates ring and cavity variation analysis results; Where C(t) represents the change of rings and holes at time t, λ i Indicates the sensitivity of the target ring or cavity to voltage fluctuations, V i (t) is the voltage fluctuation of the ith ring or hole.

4. The distribution network fault intelligent location and response system according to claim 3, characterized in that: The steps of dynamically acquiring the topological features are specifically as follows: Receiving the ring and hole change data, the data records the change of the topological structure in the power grid, and generating the ring and hole change data; Using multi-scale topological analysis technology to process the change data of the rings and voids, tracking the changes of each ring and void under differentiated time and voltage fluctuation conditions, and generating dynamic tracking results of the rings and voids; Integrate the ring and hole dynamic tracking results, using the formula: Calculate the topological feature dynamics accumulated over generation time t; Where T(t) represents the accumulated topological feature dynamics, ω k To represent the weight of the target loop or hole in the topology analysis, reflecting its importance in the grid stability, d k (t′) is the rate of change of the ring or hole at time t′, which represents the topological structure of the change, α k is the adjustment factor, referring to the effect of voltage fluctuations on the sensitivity to rings or voids.

5. The distribution network fault intelligent location and response system according to claim 4, characterized in that: The steps for obtaining the fault point are specifically as follows: Extracting loop and hole data in the topological feature dynamics, the data records the topological structure changes at key time points in the power grid, and generating loop and hole time series data; Comparing and analyzing the ring and hole time series data with the fault event records, determining the correlation between the fault event and the target topology change through statistical correlation analysis, and generating fault-correlated topology change analysis results; Using the above fault-related topology change analysis results, perform topology change tracking and apply the formula: Locate the fault point; Among them, P(x) represents the fault point prediction model, β i represents the importance weight of the i-th topological change index, which is used to weight the influence of the differentiation index, γ i is the fault response sensitivity coefficient, which is used to adjust the weight to reflect the actual fault sensitivity, x i It is the topology change monitored in real time.

6. The distribution network fault intelligent location and response system according to claim 5, characterized in that: The steps for obtaining the fault prediction result are specifically as follows: Deploy a sensor network to continuously monitor the state of the power grid near the fault point, record key parameters of current, voltage and load, form a real-time data stream, and generate real-time state monitoring data; Based on the real-time status monitoring data, applying a time series analysis method to perform anomaly detection, identify data patterns that are estimated to indicate faults, and generate anomaly detection results; According to the abnormal detection results, the prediction algorithm is used to analyze the time and location of the fault, using the formula: Generate fault prediction results; Where F(t) represents the risk function of failure prediction, α j represents the influence of differentiated data indicators in the fault prediction model, β j is the time attenuation coefficient, which adjusts the sensitivity of the time span to the impact of the fault, d j are the data metrics identified from anomaly detection.

Images