A method for locating faults in distribution network

Abstract

The invention relates to distribution network fault monitoring, and in particular to a distribution network fault locating method, which collects transient zero-sequence voltage and transient zero-sequence current of each section, and calculates the volt-ampere characteristic feature vector and zero-sequence power feature vector of each section; obtains the comprehensive feature vector of each section according to the volt-ampere characteristic feature vector and zero-sequence power feature vector of each section; adopts a fuzzy C-means clustering algorithm combined with an improved adaptive immune algorithm, performs clustering analysis on each section according to the comprehensive feature vector, and obtains the fault section; the technical solution provided by the invention can effectively overcome the defect of the prior art that it is difficult to accurately and quickly locate the fault of the distribution network.

Description

Technical Field

[0001] The invention relates to distribution network fault monitoring, and in particular to a distribution network fault locating method. Background Art

[0002] With the continuous development of society, power users have higher and higher requirements for power quality and power supply reliability. Among all power outages, power outages caused by distribution network faults account for more than 90%. Accurate and rapid fault location of the distribution network is conducive to the isolation and rapid maintenance of the fault in a small area, which is of great significance to the safe, stable and efficient operation of the distribution network.

[0003] At present, the distribution network fault location methods are mainly divided into impedance method, injection signal method and traveling wave method. The impedance method is greatly affected by the transition resistance of the fault point, the neutral point grounding method and the distributed capacitance, and the ranging accuracy is poor. For the multi-branch structure distribution network, the impedance method has problems such as pseudo fault points. The energy of the injected signal of the injection signal method is limited by the voltage transformer, and it cannot be applied to intermittent faults. The degree of automation is low, and it is necessary to add signal injection equipment, which is costly. The traveling wave method is not affected by factors such as the transition resistance of the fault point, system oscillation and current transformer saturation, and has been successfully applied in the transmission network fault location.

[0004] However, the current topology of the distribution network is relatively complex, with a large number of branches on the feeder, and overhead line-cable hybrid lines are very common, which brings great difficulties to the fault location of the distribution network. In addition, the length of the distribution line is relatively short, so the accuracy requirement for fault location is higher than that of the transmission line. Some solutions for traveling wave location of transmission network faults in the prior art cannot achieve accurate fault location for distribution networks with complex topology structures, and even the fault location fails. Summary of the invention

[0005] In view of the above-mentioned shortcomings of the prior art, the present invention provides a distribution network fault location method, which can effectively overcome the defect of the prior art that it is difficult to accurately and quickly locate faults in the distribution network.

[0006] To achieve the above objectives, the present invention is implemented through the following technical solutions:

[0007] A method for locating a distribution network fault comprises the following steps:

[0008] S1. Collect the transient zero-sequence voltage and transient zero-sequence current of each section, and calculate the volt-ampere characteristic characteristic vector and zero-sequence power characteristic vector of each section;

[0009] S2. Obtaining a comprehensive characteristic vector of each section according to the characteristic vector of the volt-ampere characteristic and the zero-sequence power characteristic vector of each section;

[0010] S3, using the fuzzy C-means clustering algorithm combined with the improved adaptive immune algorithm, clustering analysis is performed on each section according to the comprehensive feature vector to obtain the fault section;

[0011] Among them, in the fuzzy C-means clustering algorithm combined with the improved adaptive immune algorithm, the cluster center is used as the antibody, and the comprehensive feature vector of each segment is used as the antigen. The mutation rate is dynamically updated according to the affinity between the antibody and the antigen, and the number of iterations to accelerate convergence and protect excellent antibodies. At the same time, the vaccination probability is dynamically updated according to the affinity between the antibody and the antigen, and the number of iterations to maintain the stability and diversity of the population.

[0012] Preferably, in S1, transient zero-sequence voltage and transient zero-sequence current of each section are collected, and the volt-ampere characteristic feature vector and zero-sequence power feature vector of each section are calculated, including:

[0013] S11, collecting the transient zero-sequence voltage and transient zero-sequence current of each section, and calculating the transient zero-sequence voltage derivative and zero-sequence power value of each section;

[0014] S12, obtaining the volt-ampere characteristic characteristic vector of each section according to the first half-wave transient zero-sequence voltage derivative and transient zero-sequence current of each section;

[0015] S13. Obtain a zero-sequence power characteristic vector of each section according to the zero-sequence power value of each section.

[0016] Preferably, in S11, the transient zero-sequence voltage and the transient zero-sequence current of each section are collected, and the transient zero-sequence voltage derivative and the zero-sequence power value of each section are calculated, including:

[0017] The transient zero-sequence voltage derivative of each section is calculated using the following formula:

[0018] ;

[0019] in, U 0 is the transient zero-sequence voltage, for The transient zero-sequence voltage at time U 0( t )for t The transient zero-sequence voltage at time is the time interval;

[0020] The zero-sequence power value of each section is calculated using the following formula:

[0021] P 0= U 0* I 0;

[0022] in, P 0 is the zero-sequence power value, I0 is the transient zero-sequence current.

[0023] Preferably, in S12, the volt-ampere characteristic feature vector of each section is obtained according to the first half-wave transient zero-sequence voltage derivative and transient zero-sequence current of each section, including:

[0024] The first half-wave transient zero-sequence voltage derivative and transient zero-sequence current of each section are normalized, and a volt-ampere characteristic scatter plot is constructed. A clustering algorithm is used to extract the cluster center that characterizes the distribution of the scatter plot, so as to obtain the corresponding volt-ampere characteristic feature vector.

[0025] Preferably, in S13, according to the zero-sequence power value of each section, a zero-sequence power characteristic vector of each section is obtained, including:

[0026] The amplitude of the zero-sequence power value of each section is accumulated according to time to obtain an accumulated value, thereby obtaining the corresponding zero-sequence power characteristic vector.

[0027] Preferably, in S2, a comprehensive characteristic vector of each section is obtained according to the characteristic vector of the volt-ampere characteristic and the zero-sequence power characteristic vector of each section, including:

[0028] S21, normalizing the volt-ampere characteristic feature vector and the zero-sequence power feature vector of each section to eliminate the influence of different dimensions and numerical ranges on the clustering results;

[0029] S22, performing feature splicing and fusion on the normalized volt-ampere characteristic feature vector and zero-sequence power feature vector of each section, so as to obtain a comprehensive feature vector of each section;

[0030] Among them, the comprehensive feature vector of each segment is expressed as follows:

[0031] CFV i =[ VCE i , ZD i ];

[0032] In the above formula, CFV i For the i The comprehensive feature vector of the segments, VCE i For the i The eigenvector of the volt-ampere characteristic of each segment is: ZD i For the i The zero-sequence power characteristic vector of each section.

[0033] Preferably, in S3, a fuzzy C-means clustering algorithm combined with an improved adaptive immune algorithm is used to perform cluster analysis on each section according to the comprehensive feature vector to obtain the fault section, including:

[0034] S31, inputting the comprehensive feature vector of each section into the fuzzy C-means clustering algorithm combined with the improved adaptive immune algorithm, performing cluster analysis, and obtaining the cluster center;

[0035] S32, analyzing the clustering results, selecting the cluster that best matches the fault characteristics, and taking the segment included in the cluster as the fault segment;

[0036] S33, verifying the fault location result to ensure the accuracy and reliability of the fault location;

[0037] Among them, in the fuzzy C-means clustering algorithm combined with the improved adaptive immune algorithm, the cluster center is used as the antibody and the comprehensive feature vector of each segment is used as the antigen.

[0038] Preferably, in S31, the comprehensive feature vector of each segment is input into the fuzzy C-means clustering algorithm combined with the improved adaptive immune algorithm to perform cluster analysis to obtain the cluster center, including:

[0039] S311, randomly generating an initial population of a certain size in the search space, each antibody in the initial population represents a potential solution, and initializing the membership matrix of each antibody;

[0040] S312, cloning antibodies with higher affinity in the current population to generate a new group of antibodies;

[0041] S313, dynamically updating the mutation rate according to the affinity between each antibody and the antigen in the current population and the number of iterations, and performing a mutation operation on each antibody in the current population based on the updated mutation rate to introduce a new search space;

[0042] S314, dynamically updating the vaccination probability according to the affinity between each antibody and antigen in the current population and the number of iterations, and vaccinating each antibody in the current population based on the updated vaccination probability to guide the algorithm to search for a better search space;

[0043] S315, obtaining the affinity between each antibody in the current population and the antigen by calculating the distance between the antigen and each antibody in the current population, and updating the membership matrix of each antibody;

[0044] S316, sorting each antibody in the current population according to affinity, and eliminating antibodies with lower affinity;

[0045] S317, judging whether the iteration termination condition is satisfied, if not, returning to S312, otherwise, outputting each antibody in the current population as a cluster center.

[0046] Preferably, in S313, the mutation rate is dynamically updated according to the affinity between each antibody and the antigen in the current population and the number of iterations, and a mutation operation is performed on each antibody in the current population based on the updated mutation rate to introduce a new search space, including:

[0047] S3131. According to the principle that the higher the affinity of the antibody, the smaller the mutation rate should be to protect its excellent characteristics, and the lower the affinity of the antibody, the larger the mutation rate should be to increase the population diversity, and that a higher mutation rate is used in the early stage of the algorithm to increase the population diversity, and a lower mutation rate is used in the later stage of the algorithm to accelerate convergence and protect excellent antibodies, the mutation rate is dynamically updated using the following formula:

[0048] ;

[0049] in, Antibodies x j In the k The mutation rate at iterations, f min , f max , f avg are the minimum affinity, maximum affinity, and average affinity of all antibodies in the current population, respectively. f ( x j ) is an antibody x j The affinity of k is the current iteration number, K is the maximum number of iterations, a is a constant, and the affinity of the new antibody generated by cloning is the same as that of the original antibody;

[0050] S3132. Based on the updated mutation rate, a mutation operation including adjusting the position and changing the elements of the membership matrix is ​​performed on each antibody in the current population to introduce a new search space.

[0051] Preferably, in S314, the vaccination probability is dynamically updated according to the affinity between each antibody and the antigen in the current population and the number of iterations, and the vaccine is vaccinated to each antibody in the current population based on the updated vaccination probability to guide the algorithm to search for a better search space, including:

[0052] S3141. According to the principle that the probability of vaccination should be smaller for antibodies with higher affinity to protect their excellent characteristics, and the probability of vaccination should be greater for antibodies with lower affinity to increase their chances of receiving excellent genes, and that a higher vaccination probability should be used in the early stage of the algorithm to accelerate the evolution of the population, and a lower vaccination probability should be used in the later stage of the algorithm to maintain the stability and diversity of the population, the vaccination probability is dynamically updated using the following formula:

[0053] ;

[0054] in, Antibodies x j In the k The mutation rate at iterations, f max , f avg are the maximum affinity and average affinity of all antibodies in the current population, respectively. f ( x j ) is an antibody x j The affinity of k is the current iteration number, K is the maximum number of iterations, b , c are all constants, and the affinity of the new antibodies generated by cloning is the same as that of the original antibodies;

[0055] S3142. Extract the vaccine library based on the characteristic information or prior knowledge of the problem, and replace or modify some genes of each antibody in the current population based on the updated vaccination probability to guide the algorithm to search towards a better search space.

[0056] Compared with the prior art, the method for locating a distribution network fault provided by the present invention has the following beneficial effects:

[0057] 1) Collect the transient zero-sequence voltage and transient zero-sequence current of each section, and calculate the volt-ampere characteristic feature vector and zero-sequence power feature vector of each section. According to the volt-ampere characteristic feature vector and zero-sequence power feature vector of each section, obtain the comprehensive feature vector of each section, so that the comprehensive feature vector of each section can be effectively extracted based on the transient zero-sequence voltage and transient zero-sequence current of each section, providing data support for the subsequent distribution network fault location through cluster analysis;

[0058] 2) Combine the adaptive immune algorithm with the fuzzy C-means clustering algorithm, and improve the adaptive immune algorithm. Take the cluster center as the antibody and the comprehensive feature vector of each segment as the antigen. Dynamically update the mutation rate according to the affinity between the antibody and the antigen, as well as the number of iterations, so as to accelerate convergence and protect excellent antibodies. At the same time, dynamically update the vaccination probability according to the affinity between the antibody and the antigen, as well as the number of iterations, so as to maintain the stability and diversity of the population, so that the algorithm can converge quickly, and then realize the accurate and rapid fault location of the distribution network, which is conducive to the safe, stable and efficient operation of the distribution network. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention, and for ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0060] Figure 1 It is a schematic diagram of the process of the present invention;

[0061] Figure 2 It is a schematic diagram of the process of obtaining the comprehensive feature vector of each section in the present invention;

[0062] Figure 3 The figure is a schematic diagram of the process of obtaining the fault section by clustering analysis on the comprehensive feature vectors of each section in the present invention. DETAILED DESCRIPTION

[0063] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0064] A distribution network fault location method, such as Figure 1 and Figure 2 As shown, S1, collect the transient zero-sequence voltage and transient zero-sequence current of each section, and calculate the volt-ampere characteristic characteristic vector and zero-sequence power characteristic vector of each section, specifically including:

[0065] S11, collecting the transient zero-sequence voltage and transient zero-sequence current of each section, and calculating the transient zero-sequence voltage derivative and zero-sequence power value of each section;

[0066] S12, obtaining the volt-ampere characteristic characteristic vector of each section according to the first half-wave transient zero-sequence voltage derivative and transient zero-sequence current of each section;

[0067] S13. Obtain a zero-sequence power characteristic vector of each section according to the zero-sequence power value of each section.

[0068] 1) S11 collects the transient zero-sequence voltage and transient zero-sequence current of each section, and calculates the transient zero-sequence voltage derivative and zero-sequence power value of each section, including:

[0069] The transient zero-sequence voltage derivative of each section is calculated using the following formula:

[0070] ;

[0071] in, U 0 is the transient zero-sequence voltage, for The transient zero-sequence voltage at time U 0( t )for t The transient zero-sequence voltage at time is the time interval;

[0072] The zero-sequence power value of each section is calculated using the following formula:

[0073] P 0= U 0* I 0;

[0074] in, P 0 is the zero-sequence power value, I 0 is the transient zero-sequence current.

[0075] 2) In S12, the volt-ampere characteristic vector of each section is obtained according to the first half-wave transient zero-sequence voltage derivative and transient zero-sequence current of each section, including:

[0076] The first half-wave transient zero-sequence voltage derivative and transient zero-sequence current of each section are normalized, and a volt-ampere characteristic scatter plot is constructed. A clustering algorithm is used to extract the cluster center that characterizes the distribution of the scatter plot, so as to obtain the corresponding volt-ampere characteristic feature vector.

[0077] 3) In S13, the zero-sequence power characteristic vector of each section is obtained according to the zero-sequence power value of each section, including:

[0078] The amplitude of the zero-sequence power value of each section is accumulated according to time to obtain an accumulated value, thereby obtaining the corresponding zero-sequence power characteristic vector.

[0079] S2. According to the volt-ampere characteristic characteristic vector and the zero-sequence power characteristic vector of each section, a comprehensive characteristic vector of each section is obtained, which specifically includes:

[0080] S21, normalizing the volt-ampere characteristic feature vector and the zero-sequence power feature vector of each section to eliminate the influence of different dimensions and numerical ranges on the clustering results;

[0081] S22, performing feature splicing and fusion on the normalized volt-ampere characteristic feature vector and zero-sequence power feature vector of each section, so as to obtain a comprehensive feature vector of each section;

[0082] Among them, the comprehensive feature vector of each segment is expressed as follows:

[0083] CFV i =[ VCE i , ZD i ];

[0084] In the above formula, CFV i For the i The comprehensive feature vector of the segments, VCE i For the i The eigenvector of the volt-ampere characteristic of each segment is: ZD i For the i The zero-sequence power characteristic vector of each section.

[0085] The above technical scheme collects the transient zero-sequence voltage and transient zero-sequence current of each section, and calculates the volt-ampere characteristic feature vector and zero-sequence power feature vector of each section. According to the volt-ampere characteristic feature vector and zero-sequence power feature vector of each section, the comprehensive feature vector of each section is obtained, so that the comprehensive feature vector of each section can be effectively extracted based on the transient zero-sequence voltage and transient zero-sequence current of each section, providing data support for the subsequent distribution network fault location through cluster analysis.

[0086] like Figure 1 and Figure 3 As shown in S3, the fuzzy C-means clustering algorithm combined with the improved adaptive immune algorithm is used to perform cluster analysis on each section according to the comprehensive feature vector to obtain the fault section.

[0087] In the technical solution of the present application, the fuzzy C-means clustering algorithm is combined with the improved adaptive immune algorithm, the cluster center is used as the antibody, the comprehensive feature vector of each segment is used as the antigen, and the mutation rate is dynamically updated according to the affinity between the antibody and the antigen, as well as the number of iterations to accelerate convergence and protect excellent antibodies. At the same time, the vaccination probability is dynamically updated according to the affinity between the antibody and the antigen, as well as the number of iterations to maintain the stability and diversity of the population.

[0088] S3 uses the fuzzy C-means clustering algorithm combined with the improved adaptive immune algorithm to perform cluster analysis on each section based on the comprehensive feature vector to obtain the fault section, including:

[0089] S31, inputting the comprehensive feature vector of each section into the fuzzy C-means clustering algorithm combined with the improved adaptive immune algorithm, performing cluster analysis, and obtaining the cluster center;

[0090] S32, analyzing the clustering results, selecting the cluster that best matches the fault characteristics, and taking the segment included in the cluster as the fault segment;

[0091] S33, verifying the fault location result to ensure the accuracy and reliability of the fault location;

[0092] Among them, in the fuzzy C-means clustering algorithm combined with the improved adaptive immune algorithm, the cluster center is used as the antibody and the comprehensive feature vector of each segment is used as the antigen.

[0093] Specifically, in S31, the comprehensive feature vector of each segment is input into the fuzzy C-means clustering algorithm combined with the improved adaptive immune algorithm to perform cluster analysis to obtain the cluster center, including:

[0094] S311, randomly generating an initial population of a certain size in the search space, each antibody in the initial population represents a potential solution, and initializing the membership matrix of each antibody;

[0095] S312, cloning antibodies with higher affinity in the current population to generate a new group of antibodies;

[0096] S313, dynamically updating the mutation rate according to the affinity between each antibody and the antigen in the current population and the number of iterations, and performing a mutation operation on each antibody in the current population based on the updated mutation rate to introduce a new search space;

[0097] S314, dynamically updating the vaccination probability according to the affinity between each antibody and antigen in the current population and the number of iterations, and vaccinating each antibody in the current population based on the updated vaccination probability to guide the algorithm to search for a better search space;

[0098] S315, obtaining the affinity between each antibody in the current population and the antigen by calculating the distance between the antigen and each antibody in the current population, and updating the membership matrix of each antibody;

[0099] S316, sorting each antibody in the current population according to affinity, and eliminating antibodies with lower affinity;

[0100] S317, judging whether the iteration termination condition is satisfied, if not, returning to S312, otherwise, outputting each antibody in the current population as a cluster center.

[0101] 1) In S313, the mutation rate is dynamically updated according to the affinity between each antibody and antigen in the current population and the number of iterations, and each antibody in the current population is mutated based on the updated mutation rate to introduce a new search space, including:

[0102] S3131. According to the principle that the higher the affinity of the antibody, the smaller the mutation rate should be to protect its excellent characteristics, and the lower the affinity of the antibody, the larger the mutation rate should be to increase the population diversity, and that a higher mutation rate is used in the early stage of the algorithm to increase the population diversity, and a lower mutation rate is used in the later stage of the algorithm to accelerate convergence and protect excellent antibodies, the mutation rate is dynamically updated using the following formula:

[0103] ;

[0104] in, Antibodies x j In the k The mutation rate at iterations, f min , f max , f avg are the minimum affinity, maximum affinity, and average affinity of all antibodies in the current population, respectively. f ( x j ) is an antibody x j The affinity of k is the current iteration number, K is the maximum number of iterations, a is a constant, and the affinity of the new antibody generated by cloning is the same as that of the original antibody;

[0105] S3132. Based on the updated mutation rate, a mutation operation including adjusting the position and changing the elements of the membership matrix is ​​performed on each antibody in the current population to introduce a new search space.

[0106] 2) In S314, the vaccination probability is dynamically updated according to the affinity between each antibody and antigen in the current population and the number of iterations, and the vaccine is vaccinated to each antibody in the current population based on the updated vaccination probability to guide the algorithm to search for a better search space, including:

[0107] S3141. According to the principle that the probability of vaccination should be smaller for antibodies with higher affinity to protect their excellent characteristics, and the probability of vaccination should be greater for antibodies with lower affinity to increase their chances of receiving excellent genes, and that a higher vaccination probability should be used in the early stage of the algorithm to accelerate the evolution of the population, and a lower vaccination probability should be used in the later stage of the algorithm to maintain the stability and diversity of the population, the vaccination probability is dynamically updated using the following formula:

[0108] ;

[0109] in, Antibodies x j In the k The mutation rate at iterations, f max , f avg are the maximum affinity and average affinity of all antibodies in the current population, respectively. f ( x j ) is an antibody x j The affinity of k is the current iteration number, K is the maximum number of iterations, b , c are all constants, and the affinity of the new antibodies generated by cloning is the same as that of the original antibodies;

[0110] S3142. Extract the vaccine library based on the characteristic information or prior knowledge of the problem, and replace or modify some genes of each antibody in the current population based on the updated vaccination probability to guide the algorithm to search towards a better search space.

[0111] The above technical scheme combines the adaptive immune algorithm with the fuzzy C-means clustering algorithm, and improves the adaptive immune algorithm, taking the cluster center as the antibody and the comprehensive feature vector of each segment as the antigen, and dynamically updates the mutation rate according to the affinity between the antibody and the antigen, as well as the number of iterations, so as to accelerate convergence and protect excellent antibodies. At the same time, the vaccination probability is dynamically updated according to the affinity between the antibody and the antigen, as well as the number of iterations, so as to maintain the stability and diversity of the population, so that the algorithm can converge quickly, and then realize accurate and rapid fault location of the distribution network, which is conducive to the safe, stable and efficient operation of the distribution network.

[0112] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements will not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A distribution network fault location method, characterized in that: The following steps are involved: S1. Collect the transient zero-sequence voltage and transient zero-sequence current of each section, and calculate the volt-ampere characteristic characteristic vector and zero-sequence power characteristic vector of each section; S2. Obtaining a comprehensive characteristic vector of each section according to the characteristic vector of the volt-ampere characteristic and the zero-sequence power characteristic vector of each section; S3, using the fuzzy C-means clustering algorithm combined with the improved adaptive immune algorithm, clustering analysis is performed on each section according to the comprehensive feature vector to obtain the fault section; Among them, in the fuzzy C-means clustering algorithm combined with the improved adaptive immune algorithm, the cluster center is used as the antibody, and the comprehensive feature vector of each segment is used as the antigen. The mutation rate is dynamically updated according to the affinity between the antibody and the antigen, as well as the number of iterations, to accelerate convergence and protect excellent antibodies. At the same time, the vaccination probability is dynamically updated according to the affinity between the antibody and the antigen, as well as the number of iterations, to maintain the stability and diversity of the population; S3 uses the fuzzy C-means clustering algorithm combined with the improved adaptive immune algorithm to perform cluster analysis on each section based on the comprehensive feature vector to obtain the fault section, including: S31, inputting the comprehensive feature vector of each section into the fuzzy C-means clustering algorithm combined with the improved adaptive immune algorithm, performing cluster analysis, and obtaining the cluster center; S32, analyzing the clustering results, selecting the cluster that best matches the fault characteristics, and taking the segment included in the cluster as the fault segment; S33, verifying the fault location result to ensure the accuracy and reliability of the fault location; Among them, in the fuzzy C-means clustering algorithm combined with the improved adaptive immune algorithm, the cluster center is used as the antibody and the comprehensive feature vector of each segment is used as the antigen; In S31, the comprehensive feature vector of each segment is input into the fuzzy C-means clustering algorithm combined with the improved adaptive immune algorithm to perform cluster analysis and obtain the cluster center, including: S311, randomly generating an initial population of a certain size in the search space, each antibody in the initial population represents a potential solution, and initializing the membership matrix of each antibody; S312, cloning antibodies with higher affinity in the current population to generate a new group of antibodies; S313, dynamically updating the mutation rate according to the affinity between each antibody and the antigen in the current population and the number of iterations, and performing a mutation operation on each antibody in the current population based on the updated mutation rate to introduce a new search space; S314, dynamically updating the vaccination probability according to the affinity between each antibody and antigen in the current population and the number of iterations, and vaccinating each antibody in the current population based on the updated vaccination probability to guide the algorithm to search for a better search space; S315, obtaining the affinity between each antibody in the current population and the antigen by calculating the distance between the antigen and each antibody in the current population, and updating the membership matrix of each antibody; S316, sorting each antibody in the current population according to affinity, and eliminating antibodies with lower affinity; S317, judging whether the iteration termination condition is met, if not, returning to S312, otherwise outputting each antibody in the current population as a cluster center; In S313, the mutation rate is dynamically updated according to the affinity between each antibody and antigen in the current population and the number of iterations, and each antibody in the current population is mutated based on the updated mutation rate to introduce a new search space, including: S3131. According to the principle that the higher the affinity of the antibody, the smaller the mutation rate should be to protect its excellent characteristics, and the lower the affinity of the antibody, the larger the mutation rate should be to increase the population diversity, and that a higher mutation rate is used in the early stage of the algorithm to increase the population diversity, and a lower mutation rate is used in the later stage of the algorithm to accelerate convergence and protect excellent antibodies, the mutation rate is dynamically updated using the following formula: in, Antibody x j The mutation rate at the kth iteration, f min 、f max 、f avg are the minimum affinity, maximum affinity, and average affinity of all antibodies in the current population, respectively. j ) is antibody x j affinity, k is the current iteration number, K is the maximum iteration number, a is a constant, and the new antibody generated by cloning has the same affinity as the original antibody; S3132. Based on the updated mutation rate, a mutation operation including adjusting the position and changing the elements of the membership matrix is ​​performed on each antibody in the current population to introduce a new search space.

2. The method for locating a distribution network fault according to claim 1, characterized in that: S1 collects the transient zero-sequence voltage and transient zero-sequence current of each section, and calculates the volt-ampere characteristic vector and zero-sequence power characteristic vector of each section, including: S11, collecting the transient zero-sequence voltage and transient zero-sequence current of each section, and calculating the transient zero-sequence voltage derivative and zero-sequence power value of each section; S12, obtaining the volt-ampere characteristic characteristic vector of each section according to the first half-wave transient zero-sequence voltage derivative and transient zero-sequence current of each section; S13. Obtain a zero-sequence power characteristic vector of each section according to the zero-sequence power value of each section.

3. The method for locating a distribution network fault according to claim 2, characterized in that: S11 collects the transient zero-sequence voltage and transient zero-sequence current of each section, and calculates the transient zero-sequence voltage derivative and zero-sequence power value of each section, including: The transient zero-sequence voltage derivative of each section is calculated using the following formula: Among them, U0 is the transient zero-sequence voltage, U0(t+Δt) is the transient zero-sequence voltage at time t+Δt, U0(t) is the transient zero-sequence voltage at time t, and Δt is the time interval; The zero-sequence power value of each section is calculated using the following formula: P0=U0*I0; Among them, P0 is the zero-sequence power value and I0 is the transient zero-sequence current.

4. The method for locating a distribution network fault according to claim 3, characterized in that: In S12, the volt-ampere characteristic characteristic vector of each section is obtained according to the first half-wave transient zero-sequence voltage derivative and transient zero-sequence current of each section, including: The first half-wave transient zero-sequence voltage derivative and transient zero-sequence current of each section are normalized, and a volt-ampere characteristic scatter plot is constructed. A clustering algorithm is used to extract the cluster center that characterizes the distribution of the scatter plot, so as to obtain the corresponding volt-ampere characteristic feature vector.

5. The method for locating a distribution network fault according to claim 4, characterized in that: In S13, the zero-sequence power characteristic vector of each section is obtained according to the zero-sequence power value of each section, including: The amplitude of the zero-sequence power value of each section is accumulated according to time to obtain an accumulated value, thereby obtaining the corresponding zero-sequence power characteristic vector.

6. The method for locating a distribution network fault according to claim 2, characterized in that: In S2, the comprehensive characteristic vector of each section is obtained according to the characteristic vector of the volt-ampere characteristic and the zero-sequence power characteristic vector of each section, including: S21, normalizing the volt-ampere characteristic feature vector and the zero-sequence power feature vector of each section to eliminate the influence of different dimensions and numerical ranges on the clustering results; S22, performing feature splicing and fusion on the normalized volt-ampere characteristic feature vector and zero-sequence power feature vector of each section, so as to obtain a comprehensive feature vector of each section; Among them, the comprehensive feature vector of each segment is expressed as follows: CFV i =[VCE i ,ZPDE i ]; In the above formula, CFV i is the comprehensive feature vector of the ith segment, VCE i is the eigenvector of the volt-ampere characteristic of the i-th section, ZPDE i is the zero-sequence power eigenvector of the ith section.

7. The method for locating a distribution network fault according to claim 1, characterized in that: In S314, the vaccination probability is dynamically updated according to the affinity between each antibody and antigen in the current population and the number of iterations, and the vaccine is vaccinated to each antibody in the current population based on the updated vaccination probability to guide the algorithm to search for a better search space, including: S3141. According to the principle that the probability of vaccination should be smaller for antibodies with higher affinity to protect their excellent characteristics, and the probability of vaccination should be greater for antibodies with lower affinity to increase their chances of receiving excellent genes, and that a higher vaccination probability should be used in the early stage of the algorithm to accelerate the evolution of the population, and a lower vaccination probability should be used in the later stage of the algorithm to maintain the stability and diversity of the population, the vaccination probability is dynamically updated using the following formula: in, Antibody x j The mutation rate at the kth iteration, f max 、f avg are the maximum affinity and average affinity of all antibodies in the current population, respectively, f(x j ) is antibody x j affinity, k is the current iteration number, K is the maximum iteration number, b and c are constants, and the new antibody generated by cloning has the same affinity as the original antibody; S3142. Extract the vaccine library based on the characteristic information or prior knowledge of the problem, and replace or modify some genes of each antibody in the current population based on the updated vaccination probability to guide the algorithm to search towards a better search space.

Images