Method and system for analyzing fault data features of active distribution network

Abstract

A method and system for analyzing fault data features of an active distribution network. The method comprises: analyzing a voltage variation after a fault in an active distribution network and analyzing a single-phase ground fault (S1); acquiring distribution network fault data by means of analog simulation (S2); and establishing a fault location prediction model to analyze fault data features and realize fault localization (S3). In the method, a voltage variation after a fault in an active distribution network is analyzed, the voltage variation is equivalently represented using an injected current, a single-phase ground fault is analyzed, and the localization of a fault point is realized. By simulating a current variation during an actual fault injection process by means of an injected current function, a model can more closely reflect an actual operating state, thereby enhancing the authenticity of fault simulation and the accuracy of prediction. A complex relationship between a current variation and node impedance is deeply analyzed by means of a complex information filtering function, and more refined fault features are extracted, thereby optimizing the precision of fault localization, and enhancing the adaptability of the model to different fault scenarios.

Description

A method and system for analyzing fault data characteristics of active distribution network Technical Field

[0001] The present invention relates to the field of power distribution technology, and in particular to a method and system for analyzing fault data characteristics of an active power distribution network. Background Art

[0002] Active distribution networks represent a major advancement in power systems. By integrating distributed energy resources such as renewable energy, energy storage systems, and electric vehicles, they enable bidirectional energy flow and intelligent management. These networks not only improve the reliability and flexibility of power supply, but also promote energy efficiency and reduce environmental impact. The high level of automation and intelligence in active distribution networks enables real-time monitoring, analysis, and response to changes in various operating conditions, thereby optimizing grid operations and maintenance.

[0003] Fault data feature analysis plays a crucial role in active power distribution networks. By collecting and processing operational data such as voltage, current, and frequency, it identifies and predicts potential faults and anomalies. This analysis helps operators quickly locate faults, reduce power outages, and guide grid optimization and improvement measures. With technological advancements, the accuracy and efficiency of fault data analysis are continuously improving, providing strong support for ensuring stable and efficient grid operation.

[0004] Summary of the Invention

[0005] In view of the above-mentioned problems, the present invention is proposed.

[0006] Therefore, the technical problem solved by the present invention is that the existing fault data analysis method is inefficient and cannot quickly and accurately locate the fault point of the active power distribution network.

[0007] To solve the above technical problems, the present invention provides the following technical solutions: a method for analyzing the fault data characteristics of an active distribution network, comprising: analyzing the voltage changes and single-phase grounding faults of the active distribution network; obtaining distribution network fault data through simulation; establishing a fault location prediction model, analyzing the fault data characteristics and locating the fault.

[0008] As a preferred solution of the method for analyzing fault data characteristics of the active distribution network of the present invention, the voltage change includes, for the network before the active distribution network line fault, the voltage at the fault point f is expressed as:

[0009] In the node impedance matrix, each element represents the mutual impedance between the corresponding two nodes. When current is injected into any node s in the network, the voltage between the nodes fs is expressed as:

[0010] in, represents the voltage vector at point i before the fault; represents the voltage vector at point j before the fault; z if represents the impedance between point i and point f; represents the current vector flowing between i and f before the fault.

[0011] As a preferred solution of the method for analyzing the fault data characteristics of the active distribution network of the present invention, the voltage change also includes that when a fault occurs in the active distribution network line, each node will have a voltage drop of varying degrees, which can be equivalent to adding a voltage drop of magnitude at the fault point. The injected current is expressed as:

[0012] For different faults occurring at different locations, the voltage drop of the entire network can be expressed as:

[0013] in, represents the voltage change between nodes fs; Indicates the voltage drop of the entire network; Z indicates the equivalent value of the system impedance; Indicates the change in current at node s.

[0014] As a preferred embodiment of the method for analyzing fault data characteristics of an active power distribution network according to the present invention, the single-phase grounding fault analysis includes: when a single-phase grounding fault occurs, the distribution network undergoes a transient process from a pre-fault steady state to a post-fault steady state, and the current change caused by the fault can be expressed as:

[0015] Among them, I 1A I′ represents the effective value of the phase A current of line 1 in the steady state before the fault; 1A Indicates the effective value of phase A current of line 1 in steady state after the fault; ΔI 1A represents the change in phase A current of line 1; N0 represents the total number of lines; the feature ΔI related to the fault location is selected to locate the fault. Since different faults will produce different current distributions on different lines, each set of data will correspond to a unique fault location.

[0016] As a preferred embodiment of the method for analyzing fault data characteristics of an active distribution network according to the present invention, the simulation includes: during the sample generation process, setting the fault location distribution so that the fault is evenly distributed across all phases of all three-phase lines and randomly occurs at any position within the line; when setting the fault transition resistance, randomly selecting a value between 0 and 1400 Ω; due to the complex grid structure of the distribution network, the operating mode often changes, so the system impedance is re-valued every certain number of samples, with the value randomly selected between (3+4)j and (7+8)j Ω; to address the problem of random errors between simulation data and actual data, a certain margin is retained while simulating the actual grid operating data, and 30 dB white noise is added to the simulation data to participate in the offline training and online prediction processes of the model; in the actual distribution network, the connected load is always in a fluctuating state. To simulate the load changes, during the model simulation, the load is set to remain unchanged within each simulation time of 1 second. Before simulating each group of samples, each load is randomly selected within a range of 0.8 to 1.2 times to simulate the fluctuation changes of different loads.

[0017] As a preferred embodiment of the method for analyzing fault data characteristics of an active distribution network according to the present invention, the fault location prediction model includes introducing an attenuation function to simulate the natural attenuation process of voltage change over time after a fault occurs, based on the voltage change law before and after the fault in the distribution network, and establishing a fault location prediction model by predicting the voltage change value to locate the fault point, which is expressed as: I inj (t) = I max sin(2πht)e -γt

[0018] For single-phase grounding faults, the fault location information is extracted by comprehensively considering the change in current before and after the fault and the impedance value of each node, which is expressed as:

[0019] Among them, V f represents the predicted voltage change value of the fault section; N represents the total number of nodes in the distribution network; T represents the length of the observation time window; α represents the attenuation coefficient; t represents time; I inj (t) represents the injection current function; M represents the number of voltage change data points; ΔV ij represents the voltage change of node i before and after the fault occurs; λ represents the regularization parameter; Φ(ΔI i ,z i ) represents the information filtering function; I max represents the maximum amplitude of the injected current; h represents the frequency of the injected current; γ represents the attenuation coefficient; z i represents the impedance value of the i-th node; μ represents the adjustment parameter; ΔIavg It represents the average value of the current change of all nodes.

[0020] As a preferred solution of the method for analyzing fault data characteristics of an active power distribution network according to the present invention, the fault location prediction model further includes defining a loss function using a squared error with distance error attenuation, which is expressed as:

[0021] Among them, L represents the loss function; w i represents the weight of the i-th sample; represents the fault location predicted by the model for the i-th sample; y i represents the actual fault location of the i-th sample; β represents the adjustment parameter; d i represents the distance between the predicted fault location and the actual fault location; δ represents the adjustment parameter.

[0022] In a second aspect, the present invention also provides an analysis system for the fault data characteristics of an active distribution network, including an analysis module, which analyzes the voltage changes before and after the active distribution network line fault, and performs equivalentization by injecting current, analyzes single grounding faults, and locates single grounding fault points; a simulation module, which obtains fault sample data through simulation model training, considers fault location distribution, transition resistance setting, operation mode switching, noise interference and load fluctuation factors, so that the simulation results are close to the real data; a positioning module, which establishes a fault location prediction model according to the voltage change law before and after the fault in the distribution network, improves the prediction accuracy through the loss function, and realizes the positioning of the fault point.

[0023] In a third aspect, the present invention further provides a computing device, comprising: a memory and a processor;

[0024] The memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions. When the computer-executable instructions are executed by the processor, the steps of the method for analyzing fault data characteristics of the active power distribution network are implemented.

[0025] In a fourth aspect, the present invention further provides a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, implement the steps of the method for analyzing fault data characteristics of the active power distribution network.

[0026] Beneficial effects of the present invention: The method of the present invention analyzes the voltage changes after an active distribution network fault, equating the voltage changes with injected current, analyzing single-phase grounding faults and locating the fault point. By simulating the current changes during the actual fault injection process through the injection current function, the model can be closer to the actual operating state, enhancing the authenticity of the fault simulation and the accuracy of the prediction. The complex information filtering function deeply analyzes the complex relationship between current changes and node impedance, extracting more detailed fault characteristics, optimizing the level of detail in fault location, enhancing the model's adaptability to different fault scenarios, and improving the flexibility and robustness of the diagnostic process. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] To more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. Those skilled in the art can also derive other drawings based on these drawings without inventive effort. Among them:

[0028] FIG1 is an overall flow chart of a method for analyzing fault data characteristics of an active power distribution network provided by one embodiment of the present invention;

[0029] FIG2 is a schematic diagram of an equivalent current injection point after a fault in a method for analyzing fault data characteristics of an active power distribution network provided by an embodiment of the present invention;

[0030] FIG3 is an equivalent schematic diagram of current injection into nodes on both sides of a fault line after a fault, according to a method for analyzing fault data characteristics of an active power distribution network provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0031] To make the above-mentioned objects, features, and advantages of the present invention more clearly understood, the following detailed description of the specific embodiments of the present invention is given in conjunction with the accompanying drawings. It is obvious that the described embodiments are only part of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in this field without creative work should fall within the scope of protection of the present invention.

[0032] Example 1

[0033] 1-3 , an embodiment of the present invention provides a method for analyzing fault data characteristics of an active power distribution network, including:

[0034] S1: Perform voltage change and single-phase grounding fault analysis after active distribution network.

[0035] Furthermore, for a distribution network with N nodes, when a line fails, assuming the fault point is f and the nodes at both ends of the line are numbered i and j respectively. For the network before the fault, the voltage at the fault point can be expressed as:

[0036] In the node impedance matrix, each element represents the mutual impedance between the corresponding two nodes. When current is injected into any node s in the network, the voltage between the nodes fs is expressed as:

[0037] in, represents the voltage vector at point i before the fault; represents the voltage vector at point j before the fault; z if represents the impedance between point i and point f; represents the current vector flowing between i and f before the fault.

[0038] From the above formula, we can see that the mutual impedance between point f and any point s can be equivalent to the linear superposition of the mutual impedance between nodes is and the mutual impedance between nodes js.

[0039] Furthermore, when a fault occurs in the active distribution network line, each node will experience a voltage drop of varying degrees, which is equivalent to adding a voltage drop of As shown in Figure 2, the voltage change caused by the injected current can be expressed as:

[0040] That is, add a value of The injection current can be further equivalent to adding a current of magnitude of and The injected current is shown in Figure 3. Therefore, the above formula can be expanded into the following matrix form:

[0041] For different faults occurring at different locations, the voltage drop of the entire network can be expressed as:

[0042] in, represents the voltage change between nodes fs; Indicates the voltage drop of the entire network; Z indicates the equivalent value of the system impedance; Indicates the change in current at node s.

[0043] Therefore, different faults occurring at different locations correspond to different voltage drops, resulting in different voltage distributions across the entire network. This distribution can reflect the location information of the fault.

[0044] Furthermore, the distribution network undergoes a transient process from the pre-fault steady state to a new post-fault steady state. During this process, different faults occurring at different locations will produce different voltage dips at each node, resulting in different line current distributions. Assuming that the distribution network topology is known and the current RMS value of each line is available, then when a distribution network fault occurs, the current change caused by the fault can be expressed as:

[0045] Among them, I 1A I1′ represents the effective value of the phase A current of line 1 in the steady state before the fault; A Indicates the effective value of phase A current of line 1 in steady state after the fault; ΔI 1A represents the change in phase A current of line 1; N0 represents the total number of lines; the feature ΔI related to the fault location is selected to locate the fault. Since different faults will produce different current distributions on different lines, each set of data will correspond to a unique fault location.

[0046] S2: Obtain distribution network fault data through simulation.

[0047] Furthermore, distribution network failures are low-probability events relative to their operating time, and actual fault data is insufficient to support the data quality requirements of the training phase. However, existing power simulation software has developed rapidly and can already simulate field conditions to a large extent. Software such as Matlab / Simulink and PSCAD / EMTDC all have good simulation effects.

[0048] Therefore, we chose to use high-precision simulation software to simulate the operation of the distribution network. By superimposing noise and setting different fault types, we obtained a large amount of usable fault sample data to meet the data quantity and quality requirements of the training phase. For the application phase, the distribution terminals of the existing distribution network are equipped with current transformers. Through distribution automation and other systems, we can collect the effective value of the line current in the entire network, which meets the conditions for the online prediction application of the model.

[0049] Furthermore, Matlab / Simulink was used to build a model and simulate training samples. To ensure that the simulation results closely matched the real data, the sample generation process fully considered various factors, including fault location distribution, transition resistance setting, operating mode switching, noise interference, and load fluctuations, and these factors were reflected in the model simulation process.

[0050] To ensure a relatively uniform distribution of fault locations across all three-phase lines in the training data, faults were uniformly distributed across all phases of the three-phase lines, occurring randomly at any location within the lines. Furthermore, by testing the relationship between the number of training samples and training effectiveness, a number of training samples was selected that both ensured coverage of all lines and achieved sufficient training effectiveness. For example, the IEEE 123 node system has 118 lines, 59 of which are three-phase lines. Assuming 4000 sets of fault samples in the training data, at least 57 sets of fault samples were guaranteed for each three-phase line, ensuring relative sample coverage.

[0051] When setting the fault transition resistance, a random value between 0 and 1400 Ω is used. Research at Texas A&M University in the United States shows that when a high-resistance ground fault occurs in a distribution network, the transition resistance is generally below 100 Ω. However, during algorithm validation, this value is typically set to 500 to 1000 Ω. To test the model's performance under high-resistance ground faults and its ability to locate faults in different transition resistance ranges, the transition resistance is randomly set to a range of 0 to 1400 Ω.

[0052] Due to the complex structure of distribution networks, their operating modes often change. Typical models reflect these changes in system impedance and the switching of tie switches. To address this, the system impedance is reset every certain number of samples, with the value randomly selected between (3+4)j and (7+8)jΩ. Furthermore, before each sample is simulated, the tie switch state is reset, and a random operating mode is selected, ensuring that all lines maintain power.

[0053] To address the problem of random errors between simulation data and actual data, a certain margin is retained while simulating the actual power grid operation data, and 30dB white noise is added to the simulation data to participate in the offline training and online prediction processes of the model.

[0054] In the actual distribution network, the connected load is always in a fluctuating state. In order to simulate the change of load, during the model simulation, the load is set to remain unchanged within 1s of each simulation time. Before simulating each group of samples, each load is randomly selected within the range of 0.8 to 1.2 times to simulate the fluctuation of different loads.

[0055] S3: Establish a fault location prediction model, perform fault data feature analysis and achieve fault location.

[0056] Furthermore, by analyzing the voltage changes before and after the fault, the equivalent of the injected current, and the single-phase grounding fault analysis data, an attenuation function is introduced to simulate the natural attenuation process of the voltage changing with time after the fault occurs. A fault location prediction model is established to locate the fault point by predicting the voltage change value, which is expressed as: I inj (t) = I max sin(2πht)e -γt

[0057] For single-phase grounding faults, the current change before and after the fault and the impedance value of each node are comprehensively considered to extract the fault location information. At the same time, the logarithmic function is used to amplify the impact of small current changes, while the impact of large current changes is relatively reduced. On this basis, the difference between the current change and the average current change is smoothed to establish an information filtering function, which is expressed as:

[0058] Among them, V f represents the predicted voltage change value of the fault section; N represents the total number of nodes in the distribution network; T represents the length of the observation time window; α represents the attenuation coefficient; t represents time; I inj (t) represents the injection current function; M represents the number of voltage change data points; ΔV ij represents the voltage change of node i before and after the fault occurs; λ represents the regularization parameter; Φ(ΔI i ,z i ) represents the information filtering function; I max represents the maximum amplitude of the injected current; h represents the frequency of the injected current; γ represents the attenuation coefficient; z i represents the impedance value of the i-th node; μ represents the adjustment parameter; ΔI avg It represents the average value of the current change of all nodes.

[0059] It should be noted that by introducing the exponential decay function e -αt To simulate the natural attenuation process of current or voltage changing with time after a fault occurs, taking into account the decreasing effect of time factors on the fault, so that the model can reflect the dynamic changes in the actual physical process.

[0060] In fault location in active distribution networks, equivalence of injected current is not absolutely necessary. In real power systems, when a fault occurs, the system response will change due to the injected fault current. By incorporating the injected current function into the fault location formula, this response can be more accurately simulated, thereby improving the accuracy of fault location.

[0061] Furthermore, in order to optimize the fault location prediction model and improve the accuracy of the prediction results, the square error with distance error attenuation is used to define the loss function, which is expressed as:

[0062] Among them, L represents the loss function; w i represents the weight of the i-th sample; represents the fault location predicted by the model for the i-th sample; y i represents the actual fault location of the i-th sample; β represents the adjustment parameter; d i represents the distance between the predicted fault location and the actual fault location; δ represents the adjustment parameter.

[0063] This embodiment also provides an analysis system for fault data characteristics of an active distribution network, including an analysis module, which analyzes voltage changes before and after a fault in an active distribution network line, performs equivalent analysis by injecting current, analyzes single-phase grounding faults, and locates the single-phase grounding fault point; a simulation module, which obtains fault sample data through simulation model training, takes into account fault location distribution, transition resistance setting, operation mode switching, noise interference, and load fluctuation factors, so that the simulation results are close to real data; and a positioning module, which establishes a fault location prediction model based on the voltage change law before and after the fault in the distribution network, improves the prediction accuracy through a loss function, and realizes the positioning of the fault point.

[0064] This embodiment also provides a computing device, including a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to implement the method for analyzing fault data characteristics of an active power distribution network as proposed in the above embodiment.

[0065] This embodiment further provides a storage medium having a computer program stored thereon. When the program is executed by a processor, the method for analyzing fault data characteristics of an active power distribution network proposed in the above embodiment is implemented.

[0066] The storage medium proposed in this embodiment and the method for analyzing fault data characteristics of an active distribution network proposed in the above embodiment belong to the same inventive concept. For technical details not fully described in this embodiment, please refer to the above embodiment, and this embodiment has the same beneficial effects as the above embodiment.

[0067] Through the above description of the implementation methods, those skilled in the art can clearly understand that the present invention can be implemented with the help of software and necessary general-purpose hardware, and of course it can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as a computer's floppy disk, read-only memory (ROM), random access memory (RAM), flash memory (FLASH), hard disk or optical disk, etc., including a number of instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute the methods of various embodiments of the present invention.

[0068] Example 2

[0069] The following is an embodiment of the present invention, which provides a method for analyzing fault data characteristics of an active power distribution network. In order to verify the beneficial effects of the present invention, a scientific demonstration is carried out through simulation experiments.

[0070] First, we set up different fault scenarios, including various fault types, locations, and response conditions, to ensure comprehensiveness and representativeness of the test. Using high-precision simulation software Matlab / Simulink, we simulated the operation of the distribution network and the occurrence of faults. We also conducted multiple tests for each fault scenario to ensure the reliability of the results.

[0071] During the experiment, attention was paid to key indicators such as fault identification time, fault location accuracy, the number of test scenarios, average fault response time, and coverage of fault types. The fault location method of the present invention comprehensively utilizes the voltage change information before and after the fault, the equivalent processing of the injected current, and the single-phase grounding fault analysis technology. Through in-depth analysis of these data, a model that can accurately predict the fault location is established. In addition, in order to improve the adaptability and robustness of the model, the present invention considers various factors such as fault location distribution, transition resistance setting, operation mode switching, noise interference and load fluctuation during the simulation process. Three experimental data are selected for display. The experimental data are shown in Table 1, where the prior art A is based on the impedance calculation between the fault point and the measurement point for fault location; the prior art B is to use the traveling wave signal generated by the fault to locate the fault.

[0072] Table 1 Experimental data table

[0073] Several comparative experiments were conducted and the comprehensive data were statistically analyzed. The results are shown in Table 2.

[0074] Table 2 Comparison data table

[0075] The calculation of the fault location accuracy takes into account the influence of positioning accuracy. The traveling wave positioning of the existing technology B performs better than the method of the present invention in terms of positioning accuracy. However, traveling wave positioning is easily affected by line characteristics and environmental factors, and performs poorly in complex active distribution network environments. In addition, compared with the present invention, it takes longer to synchronize high-precision time measurements and requires a higher system configuration.

[0076] The method of the present invention outperforms existing technologies A and B in key performance indicators such as fault identification time, fault location accuracy, average fault response time, and fault type coverage, demonstrating the significant improvement of the present invention in improving fault location accuracy and system adaptability.

[0077] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A method for analyzing fault data characteristics of an active power distribution network, characterized in that: include: Conduct voltage change and single-phase grounding fault analysis after active distribution network; Obtain distribution network fault data through simulation; Establish a fault location prediction model, perform fault data feature analysis and achieve fault location.

2. The method for analyzing fault data characteristics of an active power distribution network according to claim 1, wherein: The voltage change includes, for the network before the active distribution network line fault, the voltage at the fault point f is expressed as: In the node impedance matrix, each element represents the mutual impedance between the corresponding two nodes. When current is injected into any node s in the network, the voltage between the nodes fs is expressed as: in, represents the voltage vector at point i before the fault; represents the voltage vector at point j before the fault; z if represents the impedance between point i and point f; represents the current vector flowing between i and f before the fault.

3. The method for analyzing fault data characteristics of an active power distribution network according to claim 2, wherein: The voltage change also includes that when the active distribution network line fails, each node will experience a voltage drop of varying degrees, which is equivalent to adding a voltage drop of magnitude at the fault point. The injected current is expressed as: For different faults occurring at different locations, the voltage drop of the entire network can be expressed as: in, represents the voltage change between nodes fs; Indicates the voltage drop of the entire network; Z indicates the equivalent value of the system impedance; Indicates the change in current at node s.

4. The method for analyzing fault data characteristics of an active power distribution network according to claim 3, wherein: The single-phase grounding fault analysis includes that when a single-phase grounding fault occurs, the distribution network goes through a transient process from the steady state before the fault to a new steady state after the fault. The current change caused by the fault can be expressed as: Among them, I 1A I′ represents the effective value of the phase A current of line 1 in the steady state before the fault; 1A Indicates the effective value of phase A current of line 1 in steady state after the fault; ΔI 1A Indicates the change in phase A current of line 1; N0 indicates the total number of lines; The feature ΔI related to the fault location is selected to locate the fault. Since different faults will produce different current distributions on different lines, each set of data will correspond to a unique fault location.

5. The method for analyzing fault data characteristics of an active power distribution network according to claim 4, wherein: The simulation includes, during the sample generation process, setting the fault location distribution to be uniformly distributed across all phases of all three-phase lines, and randomly occurring at any position in the line; When setting the fault transition resistance, a random value is selected between 0 and 1400Ω; Due to the complex grid structure, the operation mode of the distribution network often changes. Therefore, the system impedance is reset every certain number of samples, and the value is randomly selected between (3+4)j and (7+8)jΩ. To address the random error between simulation data and actual data, a certain margin is retained while simulating actual grid operation data. 30dB white noise is added to the simulation data to participate in the offline training and online prediction processes of the model. In the actual distribution network, the connected load is always in a fluctuating state. In order to simulate the change of load, during the model simulation, the load is set to remain unchanged within 1s of each simulation time. Before simulating each group of samples, each load is randomly selected within the range of 0.8 to 1.2 times to simulate the fluctuation of different loads.

6. The method for analyzing fault data characteristics of an active power distribution network according to claim 5, wherein: The fault location prediction model includes introducing an attenuation function to simulate the natural attenuation process of voltage change over time after a fault occurs based on the voltage change law before and after the fault in the distribution network, and establishing a fault location prediction model by predicting the voltage change value to locate the fault point, which is expressed as: I inj (t) = I max sin(2πht)e -γt For single-phase grounding faults, the change in current before and after the fault and the impedance of each node are comprehensively considered. Value, extract the fault location information, expressed as: Among them, V f represents the predicted voltage change value of the fault section; N represents the total number of nodes in the distribution network; T represents the length of the observation time window; α represents the attenuation coefficient; t represents time; I inj (t) represents the injection current function; M represents the number of voltage change data points; ΔV ij represents the voltage change of node i before and after the fault occurs; λ represents the regularization parameter; Φ(ΔI i ,z i ) represents the information filtering function; I max represents the maximum amplitude of the injected current; h represents the frequency of the injected current; γ represents the attenuation coefficient; z i represents the impedance value of the i-th node; μ represents the adjustment parameter; ΔI avg It represents the average value of the current change of all nodes.

7. The method for analyzing fault data characteristics of an active power distribution network according to claim 6, wherein: The fault location prediction model also includes defining a loss function using a squared error with distance error attenuation, which is expressed as: Among them, L represents the loss function; w i represents the weight of the i-th sample; represents the fault location predicted by the model for the i-th sample; y i represents the actual fault location of the i-th sample; β represents the adjustment parameter; d i represents the distance between the predicted fault location and the actual fault location; δ represents the adjustment parameter.

8. A system for analyzing fault data characteristics of an active power distribution network using the method according to any one of claims 1 to 7, characterized in that: include, The analysis module analyzes the voltage changes before and after the active distribution network line fault, and performs equivalent analysis by injecting current to analyze single grounding faults and locate the single grounding fault point; The simulation module obtains fault sample data through simulation model training, taking into account the fault location distribution, transition resistance setting, operation mode switching, noise interference and load fluctuation factors, so that the simulation results are close to the real data; The positioning module establishes a fault location prediction model based on the voltage change law before and after the fault in the distribution network, improves the prediction accuracy through the loss function, and realizes the positioning of the fault point.

9. A computing device comprising: memory and processor; The memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions. When the computer-executable instructions are executed by the processor, any one of claims 1 to 7 is implemented. The steps of a method.

10. A computer-readable storage medium storing computer-executable instructions, wherein the computer-executable instructions, when executed by a processor, implement the steps of the method according to any one of claims 1 to 7.

Images