Fault component identification method and device of power distribution network, medium and equipment

Abstract

The invention discloses a fault component identification method and device of a power distribution network, a medium and equipment, and belongs to the field of fault identification, and the method comprises the steps: firstly obtaining topological structure data, load node data and source load historical operation data of the power distribution network; based on topological structure data, four types of fault influence incidence matrixes are constructed, and influence types of branch faults on load nodes are determined. And constructing an island recovery matrix by combining the distributed power supply capacity and the load node data, and identifying load nodes capable of independently operating. And generating an operation scene set through clustering by using source load historical operation data. And based on the fault influence incidence matrix and the island recovery matrix, calculating the sensitivity value of the preset parameter of each element to the reliability index. And finally, identifying the fault component according to the sensitivity numerical value. The method effectively solves the problem that the fault component of the power distribution network cannot be accurately and efficiently identified in the prior art.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of fault identification, and in particular to a method and device for identifying a fault component of a power distribution network, a medium and equipment. BACKGROUND

[0002] As an important part of the power system, the reliability and stability of the power distribution network are directly related to the user's power experience. Accurate identification of the fault component in the power distribution network is of great significance for quickly locating the fault position, timely repairing the fault, reducing the power outage time and improving the power supply reliability. However, the existing technology still has many deficiencies in this field, which is difficult to meet the operation and maintenance needs of modern complex power distribution networks.

[0003] Currently, the identification of fault components in the power distribution network mainly relies on sensitivity analysis, which identifies the fault components by evaluating the impact of parameter changes of each component on the system reliability index. There are mainly two types of traditional methods: finite difference method (perturbation method) and direct derivative method. The finite difference method observes the change of the reliability index by applying a small perturbation to the reliability influencing factors, and then judges the sensitivity of the influencing factors. This method is simple in principle and has strong universality, but its sensitivity analysis accuracy is affected by the accuracy of the fault parameter change, and the calculation amount is large and the efficiency is low. The direct derivative method calculates the partial derivative of the reliability index with respect to the fault parameter, and directly substitutes the fault parameter value into the calculation of the reliability index change rate, so as to compare the sensitivity of the influencing factors. This method has clear mathematical meaning and can obtain sensitivity analysis results in the whole domain, but its application is limited due to the lack of analytical expression form of the overall network reliability index.

[0004] In addition, the existing methods have the following deficiencies when dealing with complex power distribution networks, especially mesh-type power distribution networks containing distributed power sources and energy storage devices. First, the lack of analytical expression form of the overall network reliability index limits the direct derivative method in calculating the sensitivity of the reliability parameter, which further affects the accurate identification of the fault component. Second, most of the existing methods are based on a single load level for analysis, and do not fully consider the changes of distributed power output, energy storage device charging and discharging, and load demand in different time periods, resulting in inaccurate identification results. Finally, the existing methods have low calculation efficiency and are difficult to meet the real-time or near real-time fault identification requirements.

[0005] These deficiencies result in the inability of the existing technology to accurately and efficiently identify the fault components of the power distribution network. SUMMARY

[0006] The present application provides a method and device for identifying a fault component of a power distribution network, a medium and equipment to solve the problem of inaccurate and efficient identification of fault components of the power distribution network in the prior art.

[0007] In a first aspect, the application provides a method for identifying a fault element of a power distribution network, comprising:

[0008] obtaining topology data, load node data and source-load historical operation data of the power distribution network;

[0009] constructing four types of fault influence correlation matrices based on the topology data;

[0010] constructing an island restoration matrix based on a preset distributed power supply capacity and the load node data, the island restoration matrix being used to identify load nodes that can be independently operated under each fault scenario;

[0011] generating a set of operation scenarios through clustering according to the source-load historical operation data;

[0012] calculating the sensitivity values of preset parameters of each element of the power distribution network to preset reliability indicators through matrix operation and partial derivative calculation based on the fault influence correlation matrices and the island restoration matrix; the preset parameters include fault rate, repair time, switch operation time and island formation time, and the preset reliability indicators include system average outage duration, system average outage frequency and expected power supply shortage;

[0013] identifying the fault element of the power distribution network according to the sensitivity values.

[0014] The application quantifies the influence of branch faults on load nodes and the load nodes that can be islanded under each fault scenario by obtaining the topology data, load node data and source-load historical operation data of the power distribution network, and constructing four types of fault influence correlation matrices and an island restoration matrix. Further, the set of operation scenarios is generated through clustering using the source-load historical operation data, so that the analysis can cover a variety of actual operation conditions. On this basis, the sensitivity values of preset parameters of each element to preset reliability indicators can be efficiently calculated through matrix operation and partial derivative calculation, these parameters including fault rate, repair time, switch operation time and island formation time, and the reliability indicators covering system average outage duration, system average outage frequency and expected power supply shortage. Finally, according to these sensitivity values, the application realizes accurate identification of the fault element of the power distribution network. This process not only improves the accuracy of fault element identification, but also enhances the adaptability and practicality of the method by considering the influence of multi-period operation scenarios and distributed power supply, and effectively solves the problem that the prior art cannot accurately and efficiently identify the fault element of the power distribution network.

[0015] Further, the obtaining of the topology data, load node data and source-load historical operation data of the power distribution network is specific to:

[0016] The topology data includes connection relationships of nodes, branches, sectional switches and tie switches.

[0017] The load node data includes locations, power demands and user quantities of each load node.

[0018] The source-load historical operation data includes output records of distributed power sources, charging and discharging data of energy storage devices and historical power consumption load records of each load node.

[0019] The present application obtains topology data, load node data and source-load historical operation data of a distribution network, and can comprehensively master the operation state and characteristics of the distribution network. Specifically, the distribution network topology data covers connection relationships of nodes, branches, sectional switches and tie switches, provides a basis for constructing a fault influence correlation matrix and an island restoration matrix, and makes it possible to accurately analyze the influence of a fault on the network. The load node data includes locations, power demands and user quantities of each load node, which helps to evaluate the influence degree of a fault on user power supply reliability. The source-load historical operation data records output of distributed power sources, charging and discharging conditions of energy storage devices and historical power consumption load of each load node, which provides a basis for generating a running scenario set through clustering, and further makes the analysis consider the fault influence under different running conditions. Based on these detailed and comprehensive data, the present application can more accurately identify a fault element, improve the fault diagnosis capability of the distribution network and the reliability management level.

[0020] Further, four types of fault influence correlation matrices are constructed based on the topology data, specifically:

[0021] The first type of influence of the four types of fault influence correlation matrices is that a branch fault causes all power supply paths of a load node to be disconnected, and power supply can be restored only after the fault is repaired; the second type of influence is that a branch fault causes all power supply paths of a load node to be disconnected, and power supply can be restored by a main power source after fault isolation; the third type of influence is that a branch fault causes all power supply paths of a load node to be disconnected, and power supply can be restored by switching to a standby power source through a tie line after fault isolation; and the fourth type of influence is that a branch fault has no influence on a load node.

[0022] Each branch of the distribution network is traversed, and the influence type of each branch fault on each load node is determined according to the topology structure and the fault type.

[0023] Four types of fault influence correlation matrices are constructed according to the four types of influence, and the rows of the four types of fault influence correlation matrices correspond to branches, the columns correspond to load nodes, and the elements represent whether a branch fault causes a corresponding type of influence on a load node.

[0024] The application can quantitatively and classify the specific influence of branch faults on load nodes by constructing four types of fault influence correlation matrix based on the topology data of the power distribution network. Specifically, the first type of influence covers the case where all power supply paths of the load node are disconnected due to the branch fault, and the power supply can be restored only after the fault is repaired; the second type of influence involves the case where the load node can be restored by the main power supply after fault isolation; the third type of influence refers to the case where the load node is restored by the standby power supply through the tie line after the branch fault; and the fourth type is the case where the branch fault has no influence on the load node. By traversing each branch of the power distribution network and determining the influence type of each branch fault on each load node according to the topology structure and fault type, four types of fault influence correlation matrix are constructed. The rows of these matrices correspond to the branches, the columns correspond to the load nodes, and the elements explicitly indicate whether the branch fault has the corresponding type of influence on the load node. This process enables accurate classification and quantification of the influence of faults on the power distribution network, providing a solid data foundation for subsequent fault diagnosis and reliability assessment, significantly improving the accuracy and efficiency of fault component identification, and thus enhancing the overall reliability and operation and management level of the power distribution network.

[0025] Further, based on the preset distributed power supply capacity and the load node data, an island restoration matrix is constructed, specifically:

[0026] Based on the topology of the power distribution network, the fault point of each branch is taken as the starting point, and the depth-first search or breadth-first search method is used to identify the network area that forms island operation;

[0027] The total installed capacity and available power of the distributed power supply in the network area are analyzed, and the power demand in the load node data is combined to determine whether the distributed power supply has the ability to support the power supply of the load nodes in the network area, and the power supply capacity judgment result is obtained;

[0028] Based on the power supply capacity judgment result, an island restoration matrix is constructed; the rows of the island restoration matrix correspond to the fault scenarios, the columns correspond to the load nodes, and the elements of the island restoration matrix indicate whether the load node can be restored by island operation under the corresponding fault scenario.

[0029] The application can accurately identify the island area in the power distribution network after failure and evaluate the power supply recovery ability of the distributed power supply to the load node by constructing an island recovery matrix based on the preset distributed power supply capacity and load node data. Specifically, based on the power distribution network topology, the failure point of each branch is taken as the starting point, and the depth-first search or breadth-first search method is used to identify the network area forming island operation. On this basis, the total installed capacity and available power of the distributed power supply in the network area are analyzed, and the power demand in the load node data is combined to judge whether the distributed power supply has the ability to support the power supply of the load node in the network area. Based on the power supply capacity judgment result, an island recovery matrix is constructed, the rows of the matrix correspond to the failure scenarios, the columns correspond to the load nodes, and the elements clearly indicate whether the load node can be restored to power supply through island operation under the corresponding failure scenario. This process not only improves the efficiency of failure recovery, but also enhances the self-healing ability of the power distribution network under failure conditions, providing strong support for achieving fast and reliable failure recovery.

[0030] Further, the running scenario set is generated by clustering according to the source-load historical operation data, specifically:

[0031] The distributed power output, energy storage output and load demand are selected from the source-load historical operation data as clustering features;

[0032] Based on the clustering algorithm, the historical scenarios are selected from the source-load historical operation data as initial class centers;

[0033] According to the clustering features, the distance between each historical scenario and the initial class center is calculated, and each scenario is assigned to the nearest class;

[0034] In each class, the scenario with the smallest intra-class average distance is selected as the new class center until the class center no longer changes, and the clustering result of the current clustering number is obtained;

[0035] Adjust the clustering number and repeat the clustering process to obtain the clustering results corresponding to different clustering numbers;

[0036] Calculate the evaluation coefficients of each clustering result, select the clustering result with the largest evaluation coefficient, and simplify the scenarios in each class of the clustering result with the largest evaluation coefficient to the running scenario set.

[0037] The application can efficiently process and analyze the state of the distribution network under different operating conditions by generating a set of operating scenarios according to source load historical operation data. First, the distributed power output, energy storage output and load demand are selected as clustering features from the source load historical operation data, which can comprehensively reflect the operating state of the distribution network. Then, based on the clustering algorithm, the initial class center is selected from the historical operation data, and the distance between each historical scenario and the initial class center is calculated according to the clustering features, and the scenario is assigned to the nearest category. In each category, the scenario with the minimum average distance in the category is selected as the new class center, and the iteration is continuously performed until the class center is stable, and the clustering result of the current clustering number is obtained. By adjusting the clustering number and repeating the clustering process, the clustering results corresponding to different clustering numbers are obtained. Finally, the evaluation coefficient of each clustering result is calculated, the clustering result with the maximum evaluation coefficient is selected, and it is simplified as a set of operating scenarios. This process not only effectively reduces the data volume, but also retains the key information, provides more representative and accurate operating scenarios for subsequent fault element identification and reliability analysis, and improves the analysis efficiency and accuracy of the entire system.

[0038] Further, the sensitivity values of the preset parameters of each element of the distribution network to the preset reliability index in the set of operating scenarios are calculated by matrix operation and partial derivative calculation based on the fault influence correlation matrix and the island restoration matrix, specifically:

[0039] For each operating scenario in the set of operating scenarios, the analytical expressions of the system average interruption duration, the system average interruption frequency and the expected power supply shortage are constructed by combining the fault influence correlation matrix and the island restoration matrix, and the analytical expressions include the correlation between the fault influence correlation matrix elements, the island restoration matrix elements, the element preset parameters and the load node data;

[0040] For each operating scenario, the partial derivative of the analytical expression with respect to the preset parameters of each element is calculated to obtain the sensitivity values of each preset parameter to the preset reliability index under each scenario;

[0041] According to the occurrence probability of each operating scenario in the set of operating scenarios, the sensitivity values under each scenario are weighted and summed according to the occurrence probability to obtain the sensitivity values considering the timing differences of different scenarios in the set of operating scenarios.

[0042] The application can accurately analyze the sensitivity of preset parameters of each element of the power distribution network to the preset reliability index through matrix operation and partial derivative calculation based on the fault influence correlation matrix and the island recovery matrix. Specifically, for each operating scenario in the operating scenario set, the analytical expressions of the system average interruption duration, the system average interruption frequency and the expected energy not supplied are constructed by combining the fault influence correlation matrix and the island recovery matrix, which include the elements of the fault influence correlation matrix, the elements of the island recovery matrix, the element preset parameters and the load node data. Further, the partial derivatives of these analytical expressions with respect to the preset parameters of each element are calculated to obtain the sensitivity values of each preset parameter to the preset reliability index under each scenario. Finally, according to the occurrence probability of each operating scenario in the operating scenario set, the sensitivity values under each scenario are weighted and summed according to the occurrence probability to obtain the sensitivity values considering the differences in different scenarios in the operating scenario set. This process not only improves the accuracy of fault element identification, but also enhances the adaptability and practicality of the method by considering the influence of multi-period operating scenarios and distributed power sources, which provides a strong guarantee for the reliability and stability of the power distribution network.

[0043] Further, the fault element of the power distribution network is identified according to the sensitivity values, specifically:

[0044] The sensitivity values of each element are sorted according to their absolute values to form a sensitivity sorting list;

[0045] The element with the largest absolute value of the sensitivity value is selected from the sensitivity sorting list as a candidate fault element;

[0046] According to the matrix elements corresponding to the candidate fault element in the fault influence correlation matrix, the element preset parameters and the load node data, the actual sensitivity value of the preset reliability index when the candidate fault element fails is calculated by combining the connection relationship between the candidate fault element and the load node and the connection relationship between the candidate fault element and other branches in the topology structure of the power distribution network.

[0047] If the calculated actual sensitivity value is consistent with the sensitivity value of the candidate fault element, the candidate fault element is confirmed as the fault element.

[0048] The application can efficiently and accurately locate the fault element by identifying the fault element of the power distribution network based on the sensitivity value. Specifically, first, the sensitivity values of each element are sorted according to their absolute values to form a sensitivity sorting list. Then, the element with the largest absolute value of the sensitivity value is selected from the list as the candidate fault element. Further, in combination with the connection relationship in the topology structure of the power distribution network, the actual sensitivity value of the preset reliability index of the candidate fault element when the candidate fault element fails is calculated by using the matrix element corresponding to the candidate fault element in the fault influence correlation matrix, the element preset parameter and the load node data. If the calculated actual sensitivity value is consistent with the sensitivity value of the candidate fault element, the candidate fault element is confirmed as the fault element. This process not only improves the accuracy of fault element identification, but also enhances the reliability and practicality of the method by considering the topology structure and actual operation data, thereby providing strong support for the rapid fault recovery and reliability improvement of the power distribution network.

[0049] In a second aspect, the application provides a fault element identification device of a power distribution network. The fault element identification device of the power distribution network comprises:

[0050] An acquisition module is configured to acquire topology structure data, load node data and source-load historical operation data of the power distribution network.

[0051] A first construction module is configured to construct four types of fault influence correlation matrices based on the topology structure data.

[0052] A second construction module is configured to construct an island recovery matrix based on the preset distributed power supply capacity and the load node data, wherein the island recovery matrix is used to identify the load nodes that can be independently operated under each fault scenario.

[0053] A clustering module is configured to generate a set of operation scenarios by clustering according to the source-load historical operation data.

[0054] A calculation module is configured to calculate the sensitivity values of the preset parameters of each element of the power distribution network to the preset reliability index in the set of operation scenarios by matrix operation and partial derivative calculation based on the fault influence correlation matrices and the island recovery matrix, wherein the preset parameters include fault rate, repair time, switch operation time and island formation time, and the preset reliability index includes system average outage duration, system average outage frequency and expected power supply shortage.

[0055] An identification module is configured to identify the fault element of the power distribution network according to the sensitivity values.

[0056] The power distribution network fault element identification device provided by the application realizes efficient and accurate fault element identification through the cooperative work of multiple functional modules. The acquisition module first collects the topological structure data, load node data and source-load historical operation data of the power distribution network, providing basic information for subsequent analysis. The first construction module constructs four types of fault influence correlation matrices based on the topological structure data, clearly defining the influence type of branch faults on load nodes. The second construction module constructs an island restoration matrix in combination with the distributed power supply capacity and load node data, identifying load nodes that can operate independently under each fault scenario. The clustering module generates a set of operation scenarios using source-load historical operation data, covering a variety of actual operation conditions. The calculation module calculates the sensitivity values of each element's preset parameters on the preset reliability indicators, including key parameters such as failure rate and repair time, and important indicators such as system average outage duration, based on the fault influence correlation matrix and the island restoration matrix through matrix operation and partial derivative calculation. Finally, the identification module accurately identifies the fault elements according to the sensitivity values. This device not only improves the accuracy of fault element identification, but also enhances the adaptability and practicality of the method by considering the influence of multi-period operation scenarios and distributed power supplies, providing a strong guarantee for the reliability and stability of the power distribution network.

[0057] In a third aspect, the application provides a computer-readable storage medium comprising a stored computer program, wherein the computer program, when executed, controls a device in which the computer-readable storage medium is located to perform a power distribution network fault element identification method as described. The beneficial effects are the same as those of the power distribution network fault element identification method provided in the first aspect of the application.

[0058] In a fourth aspect, the application provides a terminal device comprising a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein the processor implements any of the power distribution network fault element identification methods described in the first aspect when executing the computer program. BRIEF DESCRIPTION OF DRAWINGS

[0059] Figure 1 FIG. 1 is a flowchart of an embodiment of the power distribution network fault element identification method provided by the application;

[0060] Figure 2 FIG. 3 is a structural schematic diagram of an embodiment of the power distribution system provided by the application;

[0061] Figure 3 FIG. 4 is a structural schematic diagram of an embodiment of the power distribution network fault element identification device provided by the application. DETAILED DESCRIPTION

[0062] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0063] Example 1

[0064] Please refer to Figure 1 In order to solve the problem that existing technologies cannot accurately and efficiently identify faulty components in distribution networks, this invention provides a method for identifying faulty components in distribution networks, including steps S01-S06.

[0065] S01: Obtain the topology data, load node data, and historical source-load operation data of the distribution network.

[0066] In a preferred embodiment of this invention, the acquisition of the distribution network topology data, load node data, and historical source-load operation data specifically includes:

[0067] When acquiring distribution network topology data, distribution network dispatch automation systems and geographic information systems (GIS) can be relied upon. From these systems, the connection relationships of nodes, branches, sectionalizing switches, and tie switches are extracted. Specifically, this includes the node number and type (e.g., power supply node, load node), the start and end nodes, length, and conductor type of branches, as well as the specific locations of sectionalizing switches and tie switches within the branches and their connection methods to the nodes. For example, in a distribution network with 10 nodes and 15 branches, each branch connects to different nodes at both ends, and some branches have sectionalizing switches installed, while some nodes are connected via tie switches. All this information needs to be accurately acquired and organized into structured data.

[0068] When acquiring load node data, it is necessary to combine the electricity consumption information collection system and user files. Collect the location information of each load node, such as specific geographical coordinates or transformer area names; calculate the power demand of each load node, including maximum load, minimum load, and average load data, for example, the maximum power demand of a residential load node is 50kW, and the average power demand is 20kW; simultaneously record the number of users at each load node, such as a load node serving 200 residential users.

[0069] The source load history operation data can be acquired through an energy management system (EMS) and a distributed power monitoring system. The output record of the distributed power is collected, such as the output data of a photovoltaic power station every 15 minutes, the real-time output of a wind power, the charging and discharging data of the energy storage device, including the charging power, the discharging power and the SOC (state of charge) change, and the historical power load curve of each load node, such as the power load data of each hour in the past year, so as to reflect the change law of the load.

[0070] S02: based on the topological structure data, a four-type failure influence incidence matrix is constructed.

[0071] As a preferred embodiment of the present embodiment, based on the topological structure data, the four-type failure influence incidence matrix is constructed, specifically:

[0072] Based on the acquired power distribution network topological structure data, four different failure influence types are first defined, which cover all possible influence of branch faults on load nodes. Specifically as follows:

[0073] The first type of influence: the branch fault causes all power supply paths of the load node to be disconnected, and the power supply can be restored only after the fault is repaired;

[0074] The second type of influence: the branch fault causes all power supply paths of the load node to be disconnected, but the power supply can be restored by the main power supply after the fault is isolated;

[0075] The third type of influence: the branch fault causes all power supply paths of the load node to be disconnected, but the power supply can be restored by the standby power supply through the tie line after the fault is isolated;

[0076] The fourth type of influence: the branch fault has no influence on the load node;

[0077] Next, by traversing each branch of the power distribution network, the specific influence type of each branch fault on each load node is determined in combination with the topological structure and the fault type. Based on this information, a four-type failure influence incidence matrix is constructed, wherein the rows of the matrix correspond to the branches in the power distribution network, the columns correspond to the load nodes, and the elements in the matrix explicitly indicate whether the fault of a certain branch causes one of the four influence types defined above on a certain load node under a certain fault condition;

[0078] More specifically, the four-type failure influence incidence matrix (FEIM) FEIM A, FEIM B, FEIM C and FEIM D are as follows: the elements A i,j , B i,j , C i,j and D in the four-type failure influence incidence matrixi,j The definition is as follows:

[0079]

[0080] by Figure 2 The simple power distribution system shown illustrates the process of constructing the topological correlation matrix. Figure 2 The diagram shows a simple radial power distribution system. The power supply node is the substation busbar. Load nodes LP1, LP2, and LP3 are powered by branches ①, ②, and ③, and branches ④, ⑤, and ⑥ equipped with fuses. Circuit breakers are installed at the busbar outlets, and tie switches are installed at the feeder ends. Section switches are normally closed, and tie switches are normally open. Assume that the substation busbar, circuit breakers, fuses, and switchgear have extremely high reliability and a very low probability of failure. Consider only the case of a single component failure.

[0081] Based on the definition of the correlation matrix of these four types of fault effects, we can... Figure 2 The correlation matrix of the four types of fault effects for the simple power distribution system shown is listed below:

[0082]

[0083] This process not only systematically quantifies the impact of faults on the distribution network, but also provides accurate data support for subsequent fault component identification and reliability assessment, enabling a more accurate assessment of the potential impact of each component's fault on the distribution network's reliability indicators, thereby achieving precise identification of fault components.

[0084] S03: Based on the preset distributed power supply capacity and the load node data, construct an island recovery matrix, which is used to identify load nodes that can operate independently under each fault scenario.

[0085] In a preferred embodiment of this invention, an islanding recovery matrix is ​​constructed based on a preset distributed power supply capacity and the load node data. This islanding recovery matrix is ​​used to identify load nodes that can operate independently under various fault scenarios. Specifically:

[0086] First, determine the preset capacity of each distributed power source in the distribution network, including the rated power and maximum output active power of photovoltaic, wind power, energy storage, etc. Simultaneously, retrieve key information from the load node data, such as the real-time power demand of each load node and the network topology connections in its area.

[0087] Next, for each possible fault scenario (i.e. the case of each branch fault), the fault point of the branch is taken as the starting point, and a depth-first search method is used to traverse the power distribution network topology to determine the network area isolated due to the fault. For example, when a fault occurs in a feeder branch, the location of the sectionalizing switch between the fault point and the upstream power supply, the downstream load node and the adjacent tie switch can be determined by searching, thereby delineating the range of possible island formation;

[0088] Then, the total available capacity of all distributed power sources in the isolated area is calculated (device losses and reserved margins need to be deducted), and compared with the total power demand of all load nodes in the area. If the total capacity of the distributed power sources is greater than or equal to the total power demand, it is determined that the area has island operation conditions, and each load node in the area can be included in the range of independent operation; if the capacity is insufficient, important load nodes are further screened, and their power supply demand is prioritized, until the distributed power source capacity matches the power demand of the screened load nodes;

[0089] Finally, an island recovery matrix is constructed according to the above determination results. The rows of the matrix correspond to different fault scenarios (i.e. each branch fault), the columns correspond to all load nodes, and the matrix elements are represented by "yes" or "no": when a branch fails, if the distributed power sources in the area of a load node can support its independent operation, the corresponding position in the matrix is marked as "yes", otherwise it is marked as "no". For example, if branch A fails, the distributed power sources in the areas of load nodes 1 and 2 can meet the demand, but the capacity of the area of load node 3 is insufficient, then in the row corresponding to branch A in the matrix, nodes 1 and 2 are marked as "yes", and node 3 is marked as "no". The island recovery matrix is as follows:

[0090] Island Recovery Matrix IRM N×N (Island Recovery Matrix):

[0091]

[0092] Through the matrix, the independent operation capability of the load nodes under each fault scenario can be intuitively presented, providing a key basis for subsequent reliability index calculation.

[0093] S04: According to the historical operation data of the source and load, a set of operation scenarios is generated by clustering.

[0094] As a preferred embodiment of the present embodiment, the set of operation scenarios is generated by clustering according to the historical operation data of the source and load, specifically:

[0095] Selecting the annual historical data of a power distribution network in North China as the basis, a typical operation scenario set construction method for reliability evaluation of the power distribution system is established based on the K-Medoids clustering algorithm;

[0096] The clustering process is to select the Euclidean distance (2-norm) to depict the distance between different classes. The silhouette coefficient s is used to measure the effectiveness of the number of clusters K:

[0097]

[0098] where m represents the total number of sample points, a(i) represents the average distance of sample x i to each sample point in the i-th cluster, and b(i) represents the average distance of sample x i to the nearest cluster sample point. The number of clusters K is in the range of

[0099] (1) Count the number of running scenarios to be clustered m, and initially set the number of clusters K = 2;

[0100] (2) Randomly select K scenarios as initial cluster centers (center samples) from them;

[0101] (3) According to the distributed power output, energy storage output, and load demand of each scenario, calculate the Euclidean distance between each scenario and the current center sample;

[0102] (4) According to the "shortest distance matching" principle, assign the remaining m-K scenarios to the nearest class;

[0103] (5) In each class, reselect the sample with the smallest "intra-class average distance" as the new center, that is, find the sample point that minimizes the sum of distances to the remaining samples in the class as the representative; that is:

[0104]

[0105] where x i represents the sample point in the i-th cluster, represents the center point of the i-th cluster in the j-th iteration process.

[0106] (6) Repeat the distance calculation and center point updating process until all cluster centers no longer change;

[0107] (7) Record the silhouette coefficient s of each round of clustering, and determine whether the maximum number of clusters is reached. If not, set K = K + 1 and recluster;

[0108] (8) Compare the silhouette coefficients corresponding to each K value, select the clustering result that maximizes s, and take it as the final typical scenario division scheme.

[0109] Thus, the m running scenarios can be reduced to K typical running scenarios;

[0110] S05: based on the fault impact correlation matrix and the island recovery matrix, calculating the sensitivity of the preset parameters of each element of the power distribution network in the operating scenario set to the preset reliability index by matrix operation and partial derivative calculation; the preset parameters include failure rate, repair time, switch operation time and island formation time, and the preset reliability index includes system average interruption duration, system average interruption frequency and expected power supply shortage.

[0111] As a preferred embodiment of the present embodiment, the sensitivity of the preset parameters of each element of the power distribution network in the operating scenario set to the preset reliability index is calculated based on the fault impact correlation matrix and the island recovery matrix by matrix operation and partial derivative calculation, specifically:

[0112] The introduction of the fault impact correlation matrix FEIM realizes explicit analytical calculation of the reliability index. For time-invariant element failure parameters, the sensitivity of each parameter can be compared based on the analytical calculation formula using partial derivative operation. Taking the device failure rate λ i as an example, the SAIDI-λ i sensitivity calculation method is introduced under a single load level.

[0113]

[0114] In the formula, λ i represents the device failure rate of the i-th element, n j represents the number of users of the j-th load node, column(A i ), column(B i ) and column(C i ) represent the i-th column of the fault impact correlation matrix FEIM A, FEIM B and FEIM C, respectively, row(A i ), row(B i ) and row(C i ) represent the i-th row of the fault impact correlation matrix FEIM A, FEIM B and FEIM C, respectively.

[0115] By repeating the above partial derivative calculation process, the sensitivity of the reliability level affected by each failure parameter under a single load level can be obtained, as shown in the following table:

[0116] Table 1 Device failure rate λ i sensitivity analysis

[0117]

[0118] SAIFI-λ iSensitivity is only related to the sum of the elements in the i-th row of the fault impact correlation matrix (i.e., the number of faults caused by the i-th component) and the sum of the products of the number of users at the load node, SAIDI-λ i The sensitivity is only related to the sum of the products of the load outage time caused by the i-th element and the number of users at the load node, EENS-λ i The sensitivity is only related to the sum of the products of the load outage time caused by the i-th component and the power demand of the load node. SAIFI, SAIDI, and EENS represent the system average number of outages, the system average outage time, and the expected power shortage, respectively, where n represents a row vector consisting of the numbered users of the N load nodes. cons This indicates the total number of users in the power distribution system.

[0119] Table 2 Equipment Fault Repair Time (μ) i Sensitivity analysis

[0120]

[0121] It can be seen that SAIDI-μ i The sensitivity is only related to the sum of the products of the number of Class a outages caused by the i-th element and the number of users at the load node, EENS-μ i The sensitivity is only related to the sum of the products of the number of Class A outages caused by the i-th component and the power demand of the load node. This is because the outage time for Class B and Class C faults is equal to the sectionalizing switch operation time t. sw Interchange switch operation time t op Unaffected by equipment failure repair time μ i Influence.

[0122] Table 3. Operation time t of the sectional switch sw Sensitivity analysis

[0123]

[0124] It can be seen that SAIDI-t sw Sensitivity is related to the sum of the products of the number of Class B outages at each node in the entire system and the number of users at each load node, EENS-t sw The sensitivity is related to the sum of the products of the number of Class B outages at each node in the entire system and the power demand of the load nodes. This is because the outage time for Class A and Class C faults is equal to the fault repair time μ. i Interchange switch operation time t op Unaffected by the sectionalizing switch operation time t sw Influence.

[0125] Table 4. Operation time t of the interconnection switch op Sensitivity analysis

[0126]

[0127] It can be seen that SAIDI-t op The sensitivity is related to the sum of the product of the number of c-type outage and the number of users at each node in the whole system, EENS-t op The sensitivity is related to the sum of the product of the number of c-type outage and the power demand at each node in the whole system. This is because the outage time of a-type and b-type fault is the fault repair time μ i , the operation time of sectionalizing switch t sw , which is not affected by the operation time of tie switch t op ;

[0128] Further, the probability distribution of each typical operating scenario is as follows:

[0129]

[0130] In the formula, P k represents the probability of the occurrence of the kth typical scenario, t k represents the duration represented by the kth typical scenario, T represents the statistical year of the system to be evaluated, m k represents the number of scenarios in the kth cluster in the clustering result, and m represents the total number of operating scenarios before reduction.

[0131] Still taking the SAIDI-λ i sensitivity analysis as an example, the time-invariant element fault parameter sensitivity analysis method considering the time sequence difference of scenarios is introduced:

[0132]

[0133] In the formula, P k represents the probability of the occurrence of the kth scenario, SAIDI(k) represents the system average outage time of the kth scenario, IRM(k) represents the island recovery matrix of the kth scenario, and the remaining parameters have the same meaning as Figure 2 The row vector represents the annual average outage number of N nodes, the row vector represents the annual average outage time of N nodes, the row vector represents the power shortage of N nodes, and the row vector P 1×K =[P1, P2, …, P K ] represents the probability of the occurrence of K scenarios, respectively represent the outage time and power shortage of the jth node under the ith scenario;

[0134] The above partial derivative calculation process is repeated to obtain the sensitivity of the reliability index considering the source and load multi-period time sequence scenarios to each time-invariant element fault parameter, as shown in the following table:

[0135] Table 5 Equipment failure rate λ i Sensitivity analysis

[0136]

[0137] SAIFI-λ i Sensitivity is only related to the sum of the product of the number of outages caused by the ith element and the number of users at the load nodes, SAIDI-λ i Sensitivity is only related to the sum of the product of the outage time caused by the ith element and the number of users at the load nodes, EENS-λ i Sensitivity is only related to the sum of the product of the outage time caused by the ith element and the power demand at the load nodes.

[0138] Table 6 Equipment failure repair time μ i Sensitivity analysis

[0139]

[0140] SAIDI-μ under each scenario i Sensitivity and EENS-μ i Sensitivity decreases because a part of the load originally affected by the a-type fault can be restored by the island, and the outage time of this part of the load nodes decreases to the island formation time t isld The increment of SAIDI and EENS brought by this part of the load nodes is irrelevant to the equipment failure repair time μ i .

[0141] Table 7 Section switch operation time t sw Sensitivity analysis

[0142]

[0143] SAIDI-t under each scenario sw Sensitivity results are consistent because the transfer process of the b-type fault node is irrelevant to whether there is a source-load power supply island. EENS-t under each scenario sw Sensitivity changes because the load demand of the load node is different under different timing scenarios;

[0144] Table 8 Tie switch operation time t op Sensitivity analysis

[0145]

[0146] SAIDI-t under each scenario op Sensitivity results are consistent because the transfer process of the c-type fault node is irrelevant to whether there is a source-load power supply island. EENS-t under each scenarioop The sensitivity appears to change because the load demand of the load node is different in different time sequence scenarios.

[0147] Table 9 Island formation time t isld Sensitivity analysis

[0148]

[0149] SAIDI-t in each scenario isld The sensitivity is related to the sum of the product of the number of times each node in the entire system is restored to power by an island and the number of users of the load node, EENS-t isld The sensitivity is related to the sum of the product of the number of times each node in the entire system is restored to power by an island and the power demand of the load node. This is because the outage time of a class a, b, and c fault is the fault repair time μ i , the tie switch operation time t op , and the section switch operation time t sw , which is not affected by the island formation time t isld .

[0150] The time-invariant element fault parameter sensitivity analysis method considering the source-load multi-period time sequence scenario retains the time sequence difference of different scenarios, can analytically calculate the sensitivity of each time-invariant element fault parameter, evaluate the influence of each element of the distribution system on reliability, and realize accurate identification of the fault element.

[0151] Taking a typical scenario of "high load-low wind and light" in the set of operation scenarios as an example (the probability of this scenario is 15%), four types of fault influence correlation matrices (FEIM-A to FEIM-D) and an island restoration matrix (IRM) are first called. The element in the 3rd row and the 5th column of FEIM-A is "yes", indicating that the fault of branch 3 will cause the load node 5 to be in the "only restored after fault repair" state; the element in the 3rd row and the 5th column of IRM is "no", indicating that node 5 cannot be islanded under this fault scenario.

[0152] Combined with the above matrix elements and the load node data (the power demand of node 5 is 200 kW, and the number of users is 50), an analytical expression of the system average interruption duration index (SAIDI) is constructed: SAIDI = Σ(FEIM element × (1-IRM element) × repair time × number of users) / total number of users. By substituting the repair time of branch 3 of 8 hours, the contribution of branch 3 to SAIDI under this scenario is (1 × 1 × 8 × 50) / 500 = 0.8 hours / household.

[0153] Sensitivities of the analytical expressions with respect to the pre-set parameters are calculated: when calculating the sensitivity of the repair time of branch 3 with respect to SAIDI, the result of the partial derivative is (FEIM element x (1-IRM element) x number of users) / total number of users = (1 x 1 x 50) / 500 = 0.1 (h / house) / h, that is, the SAIDI increases by 0.1 h / house for each 1 h increase in the repair time. Similarly, the sensitivity values of the branch 3 failure rate, switch operation time and other parameters with respect to SAIFI and EENS in this scenario can be calculated.

[0154] All scenarios in the scenario set (such as "low load-high wind", "fluctuating load-energy storage charging and discharging", etc.) are traversed, and the above matrix operation and partial derivative calculation are repeated to obtain the sensitivity values in each scenario. Finally, the sensitivity values are summed according to the scenario probability: for example, the total sensitivity of the repair time of branch 3 with respect to SAIDI = (0.1 x 15%) + (0.08 x 20%) + … + (0.12 x 10%) = 0.09 (h / house) / h, that is, the final sensitivity result considering the difference between scenarios is obtained.

[0155] S06: According to the sensitivity values, a faulty element of the power distribution network is identified.

[0156] As a preferred embodiment of the present embodiment, the faulty element of the power distribution network is identified according to the sensitivity values, specifically:

[0157] In a certain 10 kV power distribution network, the sensitivity values of each element with respect to the system average interruption duration index (SAIDI) are calculated by the foregoing steps, and a sensitivity ranking list is formed after sorting by absolute value. Among them, the sensitivity value of the failure rate of branch 7 with respect to SAIDI is 0.35 (times / year) / h, ranking first, far exceeding other elements (such as 0.12 of branch 3 and 0.08 of switch 5), so branch 7 is listed as a candidate faulty element.

[0158] Combined with the topology of the power distribution network, it can be known that branch 7 is connected to three important load nodes (nodes 12, 13 and 14) and is directly associated with tie switch 8. According to the fault influence association matrix, when branch 7 fails, nodes 12 and 13 belong to the "can be transferred after fault isolation" type (second type of influence), and node 14 belongs to the "only restored after fault repair" type (first type of influence); the island restoration matrix shows that the distributed power capacity in this area is insufficient to support island operation. Based on these matrix elements, the actual sensitivity value is calculated by substituting the actual failure rate (0.2 times / year) of branch 7, the repair time (6 hours) and the number of users (120 households) of node 14, which is 0.34 (times / year) / h, with an error of less than 3% compared with the previous sensitivity value (0.35), meeting the consistency determination standard.

[0159] Further verification found that the recent operation data of branch 7 shows that its failure rate has increased by 40% compared with the historical average, and the power failure record of node 14 is highly consistent with the failure period of branch 7. In summary, the branch 7 is finally confirmed as the faulty element, and needs to be arranged for maintenance in priority.

[0160] In summary, the present application quantifies the influence of branch failure on load nodes and the load nodes that can be islanded under each failure scenario by obtaining the topological structure data of the power distribution network, load node data, and source-load historical operation data, and constructing four types of fault influence correlation matrices and island restoration matrices. Further, the source-load historical operation data is used to generate a set of operation scenarios through clustering, enabling the analysis to cover a variety of actual operating conditions. Based on this, the present application can efficiently calculate the sensitivity values of the preset parameters of each element on the preset reliability indicators through matrix operations and partial derivative calculations. These parameters include failure rate, repair time, switch operation time, and island formation time, while the reliability indicators cover system average outage duration, system average outage frequency, and expected power supply shortage. Finally, based on these sensitivity values, the present application achieves accurate identification of the faulty elements of the power distribution network. This process not only improves the accuracy of fault element identification, but also enhances the adaptability and practicality of the method by considering multiple time period operation scenarios and the influence of distributed power sources. The present application effectively solves the problem of accurately and efficiently identifying the faulty elements of the power distribution network in the prior art.

[0161] Embodiment Two

[0162] Please refer to Figure 3 The power distribution network fault element identification device provided in the embodiments of the present application.

[0163] In this embodiment, the power distribution network fault element identification device includes an acquisition module 10, a first construction module 20, a second construction module 30, a clustering module 40, a calculation module 50, and an identification module 60.

[0164] The acquisition module 10 is used to acquire the topological structure data of the power distribution network, load node data, and source-load historical operation data.

[0165] The first construction module 20 is used to construct four types of fault influence correlation matrices based on the topological structure data.

[0166] The second construction module 30 is used to construct an island restoration matrix based on the preset distributed power source capacity and the load node data, the island restoration matrix being used to identify the load nodes that can be independently operated under each failure scenario.

[0167] The clustering module 40 is used to generate a set of operation scenarios through clustering according to the source-load historical operation data.

[0168] The computing module 50 is configured to calculate the sensitivity values of the preset parameters of each element of the power distribution network in the set of operation scenarios to the preset reliability indexes based on the fault influence correlation matrix and the island restoration matrix through matrix operation and partial derivative calculation; the preset parameters include fault rate, repair time, switch operation time and island formation time, and the preset reliability indexes include system average interruption duration, system average interruption frequency and expected energy not supplied;

[0169] The identifying module 60 is configured to identify the fault elements of the power distribution network according to the sensitivity values.

[0170] For the convenience and brevity of description, the device embodiments of the present application include all the implementation manners of the above-mentioned power distribution network fault element identification method embodiments, which will not be described herein.

[0171] Embodiment three:

[0172] The embodiment of the present application provides a computer readable storage medium, which comprises a stored computer program, wherein the computer readable storage medium controls the device where the computer readable storage medium is located to execute the power distribution network fault element identification method when the computer program runs.

[0173] The power distribution network fault element identification method can be stored in a computer readable storage medium if it is realized in the form of a software function unit and used as an independent product. Based on this understanding, all or part of the processes in the above-mentioned embodiment methods can also be completed by a computer program instructing related hardware, and the computer program can be stored in a computer readable storage medium. The computer program can implement the steps of each method embodiment when executed by a processor. The computer program includes computer program code, which can be in the form of source code, object code, executable files or some intermediate forms, etc. The computer readable medium can include any entity or device, recording medium, U disk, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electric carrier signal, telecommunication signal and software distribution medium, etc. that can carry the computer program code.

[0174] Embodiment four

[0175] The embodiment provides a terminal device, which comprises a processor, a memory and a computer program stored in the memory and configured to be executed by the processor, and the processor realizes any one of the power distribution network fault element identification methods according to the embodiment one when executing the computer program.

[0176] The above-described specific embodiments further illustrate the objects, technical solutions, and beneficial effects of the present application. It should be understood that the above-described specific embodiments are merely for the purpose of illustrating the present application and are not intended to limit the protection scope of the present application. It is particularly pointed out that any modification, equivalent replacement, improvement, etc. made by those skilled in the art within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A method for identifying a faulty element of a power distribution network, characterized in that, The method comprises the following steps: Obtain the topology data, load node data and source-load historical operation data of the power distribution network; Based on the topology data, construct four types of fault influence correlation matrices; Based on the preset distributed power capacity and the load node data, construct an island restoration matrix, which is used to identify the load nodes that can operate independently under each fault scenario; According to the source-load historical operation data, generate a set of operation scenarios through clustering; Based on the fault influence correlation matrix and the island restoration matrix, calculate the sensitivity values of the preset parameters of each element of the power distribution network in the set of operation scenarios through matrix operation and partial derivative calculation; the preset parameters include fault rate, repair time, switch operation time and island formation time, and the preset reliability indicators include system average outage duration, system average outage frequency and expected power supply shortage; According to the sensitivity values, identify the fault elements of the power distribution network.

2. The method of claim 1, wherein, The topology data, load node data and source-load historical operation data of the power distribution network are obtained in detail as follows: The topology data includes the connection relationship of nodes, branches, sectional switches and tie switches; The load node data includes the location, power demand and user quantity of each load node; The source-load historical operation data includes the output record of the distributed power source, the charge and discharge data of the energy storage device and the historical power consumption load record of each load node.

3. The method of claim 1, wherein, Based on the topology data, four types of fault influence correlation matrices are constructed in detail as follows: The first type of influence of the four types of fault influence correlation matrices is that the branch fault causes all power supply paths of the load node to be disconnected, and only the power supply can be restored after the fault is repaired; the second type of influence is that the branch fault causes all power supply paths of the load node to be disconnected, and the power supply can be restored by the main power source after the fault is isolated; the third type of influence is that the branch fault causes all power supply paths of the load node to be disconnected, and the power supply can be restored by the standby power source through the tie line after the fault is isolated; the fourth type of influence is that the branch fault has no influence on the load node; Each branch of the power distribution network is traversed, and the influence type of each branch fault on each load node is determined according to the topology structure and the fault type; According to the four types of influence, four types of fault influence correlation matrices are constructed, and the rows of the four types of fault influence correlation matrices correspond to the branches, the columns correspond to the load nodes, and the elements represent whether the branch fault causes the corresponding type of influence on the load node. Based on the preset distributed power capacity and the load node data, the island restoration matrix is constructed in detail as follows: Based on the topology structure of the power distribution network, the fault point of each branch is taken as the starting point, and the depth-first search or breadth-first search method is used to identify the network area that forms an island operation; 4. The method of claim 1, wherein, Analyze the total installed capacity and available power of the distributed power source in the network area, and combine the power demand in the load node data to determine whether the distributed power source has the ability to support the power supply of the load nodes in the network area, to obtain a power supply capacity judgment result; Based on the power supply capacity judgment result, construct an island restoration matrix; ​ ​ The row of the island recovery matrix corresponds to a fault scenario, the column corresponds to a load node, and the island recovery matrix element represents whether the load node can be restored to power supply through island operation under the corresponding fault scenario.

5. The method of claim 1, wherein, The operation scenario set is generated by clustering according to source load historical operation data, specifically: Distributed power output, energy storage output and load demand are selected from source load historical operation data as clustering features; Based on the clustering algorithm, historical scenarios are selected from the source load historical operation data as initial class centers; According to the clustering features, the distance between each historical scenario and the initial class center is calculated, and each scenario is assigned to the nearest class; In each class, the scenario with the minimum intra-class average distance is selected as the new class center until the class center no longer changes, and the clustering result of the current clustering number is obtained; Adjust the clustering number and repeat the clustering process to obtain the clustering results corresponding to different clustering numbers; Calculate the evaluation coefficient of each clustering result, select the clustering result with the maximum evaluation coefficient, and simplify each class in the clustering result with the maximum evaluation coefficient to an operation scenario set.

6. The method of claim 1, wherein, Based on the fault influence correlation matrix and the island recovery matrix, the sensitivity values of the preset parameters of each element of the distribution network in the operation scenario set to the preset reliability index are calculated by matrix operation and partial derivative calculation, specifically: For each operation scenario in the operation scenario set, the system average outage duration, the system average outage frequency and the expected power supply shortage are constructed by combining the fault influence correlation matrix and the island recovery matrix, and the analytical expression includes the correlation between the fault influence correlation matrix elements, the island recovery matrix elements, the element preset parameters and the load node data. For each operation scenario, the partial derivative of the analytical expression with respect to the preset parameters of each element is calculated to obtain the sensitivity value of each preset parameter to the preset reliability index under each scenario. According to the occurrence probability of each operation scenario in the operation scenario set, the sensitivity values under each scenario are weighted and summed according to the occurrence probability to obtain the sensitivity values considering the timing differences of different scenarios in the operation scenario set.

7. The method of claim 1, wherein, According to the sensitivity values, the fault elements of the distribution network are identified, specifically: Sort the sensitivity values of each element according to their absolute values to form a sensitivity sorting list; Select the element with the maximum absolute value of the sensitivity value from the sensitivity sorting list as the candidate fault element; According to the connection relationship between the candidate fault element and the load node and the connection relationship between the candidate fault element and other branches in the topology structure of the distribution network, the actual sensitivity value of the preset reliability index when the candidate fault element fails is calculated according to the matrix elements corresponding to the candidate fault element in the fault influence correlation matrix, the element preset parameters and the load node data. If the calculated actual sensitivity value is consistent with the sensitivity value of the candidate fault element, the candidate fault element is confirmed as the fault element.

8. A device for identifying a faulty element of a power distribution network, characterized in that It includes: An acquisition module is configured to acquire topology data of a distribution network, load node data and source load historical operation data; A first construction module is configured to construct four types of fault influence correlation matrices based on the topology data; a second constructing module, configured to construct an island restoration matrix based on a preset distributed power supply capacity and the load node data, the island restoration matrix being used to identify load nodes that can be independently operated under each fault scenario; a clustering module, configured to generate a set of operation scenarios through clustering according to source load historical operation data; a calculating module, configured to calculate sensitivity values of preset parameters of each element of the power distribution network in the set of operation scenarios to preset reliability indexes based on the fault influence correlation matrix and the island restoration matrix through matrix operation and partial derivative calculation; the preset parameters include a fault rate, a repair time, a switch operation time, and an island formation time, and the preset reliability indexes include a system average interruption duration, a system average interruption frequency, and an expected energy not supplied; an identifying module, configured to identify a fault element of the power distribution network according to the sensitivity values.

9. A computer-readable storage medium, characterized in that, The computer readable storage medium comprises a stored computer program, wherein the computer program controls a device in which the computer readable storage medium is located to perform the fault element identification method of the power distribution network according to any one of claims 1 to 7 when the computer program is executed.

10. A terminal device, comprising: The computer readable storage medium comprises a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, and the processor implements the fault element identification method of the power distribution network according to any one of claims 1 to 7 when the computer program is executed.

Images