Direct current power distribution network fault section positioning method and system based on multi-source information fusion

Abstract

The invention relates to the technical field of distribution network fault positioning, in particular to a DC distribution network fault section positioning method and system based on multi-source information fusion, and the method comprises the steps: analyzing the nonlinear attenuation trend of the high-frequency current amplitude of each suspected fault element through employing a normal distribution model, so as to obtain an electrical quantity fault degree; correcting the node state of the relay protection Bayesian network model according to the local switching value action information, and calculating the switching value fault degree through reverse reasoning; taking the electrical quantity fault degree and the switching quantity fault degree as independent evidence bodies, and calculating a projection distance weight and an information entropy weight of a credible interval by using an improved Doster-Schelant evidence theory recognition framework which introduces a complete set and a subset, so as to carry out weighted fusion on all the evidence bodies, and obtaining a fusion fault probability of each suspected fault element; and positioning a DC power distribution network fault section according to the fusion fault probability. According to the method, the high-frequency transient electrical quantity characteristics and the switching value action information are fused, so that the fault section of the direct-current power distribution network can be quickly and accurately positioned.

Description

Technical Field

[0001] The present invention relates to the technical field of distribution network fault location, and in particular, to a method and system for locating a fault section of a DC distribution network based on multi-source information fusion. Background Art

[0002] With the development of DC distribution networks towards diversified and multi-level network structures, the increasing access of distributed photovoltaics, energy storage, and flexible loads has significantly increased the complexity of system faults. When a short-circuit fault occurs in a DC distribution system, the fault current rises rapidly, and the fault mechanism is complex and variable, posing a serious threat to the safe and stable operation of the system. To achieve accurate fault location and isolation, current fault section location methods for DC distribution networks mainly rely on three types of fault information-based methods: those based on electrical quantity fault information, those based on switch quantity fault information, and those based on multi-source fault information fusion. However, all of them have obvious limitations.

[0003] The fault location method based on electrical quantity fault information can be divided into single-end and multi-terminal measurement fault location. For example, the difference in the transient voltage derivatives of the positive and negative poles is used to identify faults. Although this fault location method has strong anti-interference ability, it has high requirements for sampling synchronization and the communication system. In practical applications, positioning failures are likely to occur due to synchronization errors; the method based on switch quantity fault information (such as relay protection action signals) realizes positioning by constructing fault diagnosis models such as expert systems or Bayesian networks. However, when signal missing or false reporting occurs in the Supervisory Control and Data Acquisition (SCADA) system, the reliability of switch quantity information drops sharply, resulting in insufficient discrimination between faulty components and other components, and even completely incorrect positioning; the fault location method based on multi-source fault information fusion combines the advantages of electrical quantity and switch quantity information. However, existing research still has the problem that the fused information sources are mostly homologous derivative features, with low difference between information sources and a cumbersome fusion calculation process. At the same time, the fault location method based on multi-source fault information fusion lacks an effective correction mechanism for incorrect switch quantity information, resulting in a fuzzy fault probability distribution. In addition, existing information fusion algorithms considering power grid fault location are prone to problems such as information loss or fusion failure in special scenarios, affecting the accuracy and universality of the positioning results and making it difficult to adapt to complex and variable fault scenarios.

[0004] In summary, there are many deficiencies in existing DC distribution network fault location technologies. These problems seriously restrict the accuracy and reliability of fault location and are difficult to meet the increasingly complex DC distribution network fault location requirements. Therefore, how to improve the accuracy and reliability of DC distribution network fault section location remains an urgent problem to be solved currently. Summary of the Invention

[0005] To solve the above technical problems, the present invention provides a method and system for locating fault sections in a DC distribution network based on multi-source information fusion.

[0006] In a first aspect, the present invention provides a method for locating fault sections in a DC distribution network based on multi-source information fusion. The method includes the following steps: According to the high-frequency transient voltage signal generated instantaneously when a bipolar short-circuit fault occurs in the DC distribution network, reconstruct a high-frequency fault current sparse vector; Use a normal distribution model to quantitatively analyze the non-linear attenuation trend of the high-frequency current amplitudes of each suspected fault component in the high-frequency fault current sparse vector to obtain estimated high-frequency current amplitudes; Perform cubic normalization on the estimated high-frequency current amplitudes of each suspected fault component to obtain the electrical quantity fault degree of each suspected fault component; Modify the node state of the relay protection Bayesian network model according to the pre-identified local switch quantity action information, and calculate the switch quantity fault degree of each suspected fault component through reverse inference; Take the electrical quantity fault degree and the switch quantity fault degree as independent evidence bodies, and use the improved Dempster-Shafer evidence theory recognition framework introducing the universal set and subsets to calculate the projection distance weight from the perspective of evidence theory and the information entropy weight from the perspective of probability theory respectively; Fuse the projection distance weight of the credible interval and the information entropy weight of each evidence body through a combined weighting method, and use the fused weights to weighted-fuse all evidence bodies to obtain the fused fault probability of each suspected fault component; Locate the fault section of the DC distribution network according to the fused fault probability, and drive the circuit breaker to operate to isolate the fault.

[0007] In a further implementation, the step of reconstructing a high-frequency fault current sparse vector according to the high-frequency transient voltage signal generated instantaneously when a bipolar short-circuit fault occurs in the DC distribution network includes: At the moment when a bipolar short-circuit fault occurs in the DC distribution network, construct a node high-frequency impedance matrix according to the node self-impedance data and the mutual impedance data between nodes of the DC distribution network; Synchronously collect the high-frequency transient voltage signals at the fault measurement point within a preset time window before and after the occurrence of the bipolar short-circuit fault through a voltage measurement device; Equivalent the fault bus node to a high-frequency current source, set the current of non-bus fault nodes to zero, and establish a sparse node equation for bus faults according to the node high-frequency impedance matrix and the high-frequency transient voltage signal; Equivalent the fault line to a high-frequency current source of adjacent fault measurement points, set the current between non-adjacent fault measurement points to zero, and establish a sparse node equation for line faults according to the node high-frequency impedance matrix and the high-frequency transient voltage signal; Integrate the bus fault sparse nodal equations and the line fault sparse nodal equations into an underdetermined sparse nodal equation system; Transform the underdetermined sparse nodal equation system into a sparse signal reconstruction problem, and use the Bayesian compressive sensing reconstruction algorithm to iteratively solve to obtain the high-frequency fault current sparse vector.

[0008] In a further implementation, the bus fault sparse nodal equations are used to describe the relationship between the node voltages and currents in the DC distribution network during a bus fault; The line fault sparse nodal equations are used to describe the relationship between the node voltages and currents in the DC distribution network during a line fault.

[0009] In a further implementation, the step of using the normal distribution model to quantitatively analyze the non-linear attenuation trend of the high-frequency current amplitudes of each suspected fault component in the high-frequency fault current sparse vector to obtain the estimated high-frequency current amplitude values includes: According to the high-frequency current amplitudes of each suspected fault component in the high-frequency fault current sparse vector, construct multiple independent normal distribution models according to the adjacent relationship of the suspected fault components; Take the maximum value of the high-frequency current amplitudes in adjacent suspected fault components as the peak value of the corresponding normal distribution model, and solve the attenuation parameters of each group of normal distribution models through the maximum likelihood estimation criterion; Based on the attenuation parameters, use the normal distribution model to calculate the estimated high-frequency current amplitude values at the midpoint positions of the lines between each pair of adjacent suspected fault components.

[0010] In a further implementation, the attenuation parameters include the standard deviation for controlling the attenuation rate and the electrical distance difference between two adjacent suspected fault components under the normal distribution.

[0011] In a further implementation, the step of performing cubic normalization on the estimated high-frequency current amplitude values of each suspected fault component to obtain the electrical quantity fault degrees of each suspected fault component includes: Perform cubic operations on the estimated high-frequency current amplitude values of each suspected fault component respectively to obtain the cubic high-frequency current amplitudes of each suspected fault component; Perform normalization processing on the cubic high-frequency current amplitudes to obtain the electrical quantity fault degrees of each suspected fault component.

[0012] In a further implementation, the step of correcting the node states of the relay protection Bayesian network model according to the pre-identified local switch quantity action information and calculating the switch quantity fault degrees of each suspected fault component through reverse inference includes: Set three - level sampling time windows according to the sequential action logic of DC distribution network relay protection, and determine the protection level where the tripping circuit breaker is located by detecting high - frequency transient voltage signals within each level of sampling time window, and obtain the local circuit breaker action status of the protection level where the tripping circuit breaker is located; Identify local digital quantity action information according to the local circuit breaker action status, and use the local digital quantity action information to correct the digital quantity information uploaded by the data acquisition and monitoring system to obtain corrected digital quantity action information; According to the relay protection action logic of the DC distribution network, construct a Bayesian network model for the relay protection of the DC distribution network, and use the corrected digital quantity action information to update the corresponding relay protection node status and circuit breaker node status in the relay protection Bayesian network model to obtain a corrected relay protection Bayesian network model; Based on each suspected fault component determined from the perspective of electrical quantities, use the corrected relay protection Bayesian network model for inverse probability reasoning to calculate the posterior probability of failure of each suspected fault component; Perform cubic normalization on the posterior probability of failure of each suspected fault component to obtain the digital quantity fault degree of each suspected fault component.

[0013] In a further implementation, the three - level sampling time windows include a first - level sampling time window, a second - level sampling time window, and a third - level sampling time window; among them, the first - level sampling time window corresponds to the main protection action period, the second - level sampling time window corresponds to the near - backup protection action period, and the third - level sampling time window corresponds to the far - backup protection action period.

[0014] In a further implementation, the steps of using the improved Dempster - Shafer evidence theory recognition framework introducing the universal set subset to calculate the projection distance weight from the perspective of evidence theory and the information entropy weight from the perspective of probability theory include: Define the credible interval and the maximum uncertainty interval of each suspected fault component in the evidence body according to the improved Dempster - Shafer evidence theory recognition framework introducing the universal set subset, and represent the credible interval and the maximum uncertainty interval in the form of interval numbers; Based on the credible intervals and maximum uncertainty intervals of each suspected fault component represented in the form of interval numbers, calculate the projection distance between the credible interval and the maximum uncertainty interval of each suspected fault component in the evidence body; Use the projection distance as an index to measure the uncertainty of the evidence body to obtain the uncertainty measurement value of the evidence body; According to the uncertainty measurement value of the evidence body and the preset evidence adjustment coefficient, obtain the projection distance weight of the credible interval of the evidence body from the perspective of evidence theory; Distribute the probability assignments of the universal set subset in the evidence body evenly to all suspected fault components to obtain the component probability distribution, and calculate the information entropy of the component probability distribution; The information entropy weight of the evidence body from the perspective of probability theory is calculated based on the information entropy of the component probability distribution and the preset probability adjustment coefficient.

[0015] In a second aspect, the present invention provides a DC distribution network fault section location system based on multi-source information fusion. The system includes: A signal reconstruction module for reconstructing a high-frequency fault current sparse vector according to the high-frequency transient voltage signal generated instantaneously when a bipolar short-circuit fault occurs in the DC distribution network. An amplitude estimation module for quantitatively analyzing the non-linear attenuation trend of the high-frequency current amplitudes of each suspected fault component in the high-frequency fault current sparse vector by using a normal distribution model to obtain high-frequency current amplitude estimation values. An electrical quantity analysis module for performing cubic normalization on the high-frequency current amplitude estimation values of each suspected fault component to obtain the electrical quantity fault degrees of each suspected fault component. A switch quantity analysis module for correcting the node states of the relay protection Bayesian network model according to the pre-identified local switch quantity action information and calculating the switch quantity fault degrees of each suspected fault component through reverse reasoning. A double-weight analysis module for using the electrical quantity fault degree and the switch quantity fault degree as independent evidence bodies, and respectively calculating the credible interval projection distance weight from the perspective of evidence theory and the information entropy weight from the perspective of probability theory by using an improved Dempster-Shafer evidence theory identification framework introducing the universal set and subsets. A probability fusion module for fusing the credible interval projection distance weight and the information entropy weight of each evidence body by a combined weighting method, and using the fused weights to weighted-fuse all evidence bodies to obtain the fusion fault probabilities of each suspected fault component. A fault location module for locating the fault section of the DC distribution network according to the fusion fault probability and driving the circuit breaker to act to isolate the fault.

[0016] The present invention provides a method and system for locating fault sections in a DC distribution network based on multi-source information fusion. The method reconstructs a high-frequency fault current sparse vector according to the high-frequency transient voltage signal generated instantaneously when a bipolar short-circuit fault occurs in the DC distribution network; uses a normal distribution model to quantitatively analyze the non-linear attenuation trend of the high-frequency current amplitudes of each suspected fault component in the high-frequency fault current sparse vector to obtain high-frequency current amplitude estimated values; performs cubic normalization on the high-frequency current amplitude estimated values of each suspected fault component to obtain the electrical quantity fault degrees of each suspected fault component; corrects the node states of the relay protection Bayesian network model according to the pre-identified local switch quantity action information, and calculates the switch quantity fault degrees of each suspected fault component through reverse inference; uses the electrical quantity fault degrees and the switch quantity fault degrees as independent evidence bodies, and respectively calculates the credible interval projection distance weight from the perspective of evidence theory and the information entropy weight from the perspective of probability theory by using an improved Dempster-Shafer evidence theory identification framework introducing the universal set and subsets; fuses the credible interval projection distance weight and the information entropy weight of each evidence body through a combined weighting method, and uses the fused weights to weighted-fuse all evidence bodies to obtain the fused fault probabilities of each suspected fault component; locates the fault sections of the DC distribution network according to the fused fault probabilities, and drives the circuit breaker to operate to isolate the fault. Compared with the prior art, this method realizes the rapid and accurate location of the fault sections in the DC distribution network and drives the circuit breaker to operate to isolate the fault by fusing the high-frequency transient electrical quantity characteristics and the switch quantity action information, and effectively improves the fault tolerance ability of the fault location, ensuring the safe and stable operation of the power grid. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 is a schematic flow chart of a method for locating fault sections in a DC distribution network based on multi-source information fusion provided by an embodiment of the present invention; Figure 2 is a schematic diagram of the architecture of a multi-terminal flexible DC distribution network provided by an embodiment of the present invention; Figure 3 is a schematic diagram of the relay protection timing logic of a DC distribution network provided by an embodiment of the present invention; Figure 4 is a schematic diagram of the equivalent high-frequency current source when a bus fault occurs provided by an embodiment of the present invention; Figure 5 is a schematic diagram of the equivalent high-frequency current source when a line fault occurs provided by an embodiment of the present invention; Figure 6 is a schematic diagram of the non-linear attenuation characteristic of the high-frequency fault current amplitude provided by an embodiment of the present invention; Figure 7 is a schematic diagram of a normal distribution model under different peaks provided by an embodiment of the present invention; Figure 8 is a schematic diagram of the relay protection Bayesian network model of a DC distribution network provided by an embodiment of the present invention; Figure 9 It is a schematic diagram of the transient voltage change trend at the fault location provided by an embodiment of the present invention; Figure 10 It is a schematic diagram of the fault voltage spectrum analysis at the breaker operation location provided by an embodiment of the present invention; Figure 11 It is a schematic diagram of the detection result of the high-frequency voltage signal provided by an embodiment of the present invention; Figure 12 It is a schematic diagram of the three-level sampling time window based on the relay protection timing logic provided by an embodiment of the present invention; Figure 13 It is a schematic diagram of the overall process of the switch quantity fault degree evaluation provided by an embodiment of the present invention; Figure 14 It is a schematic diagram of the fault location process based on the improved Dempster-Shafer evidence theory identification framework provided by an embodiment of the present invention; Figure 15 It is a three-dimensional schematic diagram of the high-frequency fault current sparse vector reconstructed by using the Bayesian compressive sensing reconstruction algorithm provided by an embodiment of the present invention; Figure 16 It is a block diagram of the DC distribution network fault section location system based on multi-source information fusion provided by an embodiment of the present invention.

[0018] Explanation of the reference numerals: 101, signal reconstruction module; 102, amplitude estimation module; 103, electrical quantity analysis module; 104, switch quantity analysis module; 105, double-weight analysis module; 106, probability fusion module; 107, fault location module. Detailed implementation manners

[0019] The following specifically illustrates the implementation manners of the present invention in conjunction with the drawings. The given embodiments are only for illustrative purposes and should not be construed as limiting the present invention. The drawings are only for reference and illustration, and do not constitute a limitation on the protection scope of the present invention patent, because many changes can be made to the present invention without departing from the spirit and scope of the present invention.

[0020] Figure 1 It is a schematic diagram of the DC distribution network fault section location method based on multi-source information fusion provided by an embodiment of the present invention. The embodiment of the present invention provides a DC distribution network fault section location method based on multi-source information fusion. As Figure 1 shown, the method includes the following steps: S1. According to the high-frequency transient voltage signal generated instantaneously when a bipolar short-circuit fault occurs in the DC distribution network, reconstruct the high-frequency fault current sparse vector.

[0021] In some embodiments, the steps of reconstructing the high-frequency fault current sparse vector according to the high-frequency transient voltage signal generated instantaneously when a bipolar short-circuit fault occurs in the DC distribution network include: At the instant when a bipolar short-circuit fault occurs in the DC distribution network, a node high-frequency impedance matrix is constructed according to the node self-impedance data and the mutual-impedance data between nodes of the DC distribution network; The high-frequency transient voltage signals at the fault measurement points within a preset time window before and after the occurrence of the bipolar short-circuit fault are synchronously collected by a voltage measurement device; The fault bus node is equivalent to a high-frequency current source, the current of the non-bus fault nodes is set to zero, and a sparse node equation for the bus fault is established according to the node high-frequency impedance matrix and the high-frequency transient voltage signal; The fault line is equivalent to a high-frequency current source of adjacent fault measurement points, the current between non-adjacent fault measurement points is set to zero, and a sparse node equation for the line fault is established according to the node high-frequency impedance matrix and the high-frequency transient voltage signal; The sparse node equation for the bus fault and the sparse node equation for the line fault are integrated into an underdetermined sparse node equation set; The underdetermined sparse node equation set is transformed into a sparse signal reconstruction problem, and the high-frequency fault current sparse vector is obtained by iterative solution using the Bayesian compressive sensing reconstruction algorithm.

[0022] Specifically, in this embodiment, The multi-terminal flexible DC distribution network architecture is used as the research object. Figure 2 is a schematic diagram of the multi-terminal flexible DC distribution network architecture. The multi-terminal flexible DC distribution network uses modular multilevel converters (MMC) to achieve flexible interconnection between alternating current and direct current. Photovoltaic power sources and DC loads are connected to the distribution network through DC-DC converters (DC-DC) or inverters (Direct Current to Alternating Current, DC-AC). At the same time, to ensure the safe operation of the system, each line and bus port is equipped with a DC circuit breaker (DCCB). The DC circuit breaker is used to quickly isolate the fault line or bus when a fault occurs. Among them, the photovoltaic power source realizes maximum power output through the maximum power point tracking (MPPT) control strategy. The protection configuration of the DC distribution network can quickly isolate the fault components from the power system to prevent further damage to the distribution network. Among them, the fault components include buses and lines. The protection configuration of each protection area is shown in Table 1, and Table 1 is as follows: Table 1 This embodiment analyzes the protection of lines and buses in a DC distribution network. Taking the Figure 2 line as an example, the operation logic of its protection at both ends is described. Figure 2 The multi-terminal flexible DC distribution network architecture shown includes buses B1 to B17. On line , the left protection Lm and the right protection Rm are the main protections, only responsible for protecting their own lines; the left near-backup protection Lp and the right near-backup protection Rp are the near-backup protections. When the main protection fails to operate due to reasons, the near-backup protection will start and drive the DC circuit breakers and the DC circuit breaker to trip, so as to isolate the fault. The left remote-backup protection Ls and the right remote-backup protection Rs are the remote-backup protections. When both the main protection and the near-backup protection fail to operate, the remote-backup protection will act, usually driving the DC circuit breakers and the DC circuit breaker to trip, realizing the complete isolation of the faulty line. In this embodiment, m represents the main protection, p represents the near-backup protection, s represents the remote-backup protection, L represents the protection at the left end outlet of the line, and R represents the protection at the right end outlet of the line.

[0023] The bus protection is also configured with main protection and backup protection. When the bus main protection (such as the left main protection B3Lm of bus B3 and the right main protection B3Rm of bus B3) acts, it will drive the circuit breakers and the circuit breaker to trip; if the main protection fails to operate, the bus backup protection (the left backup protection B3Lp of bus B3 and the right backup protection B3Rp of bus B3) will act, driving the DC circuit breakers and the DC circuit breaker to trip, so as to isolate the faulty bus. Figure 3 is the schematic diagram of the relay protection timing logic of the DC distribution network provided by the embodiment of the present invention. The action timing of each protection and circuit breaker follows the Figure 3 shown relay protection timing logic of the DC distribution network. Among them, is the initial fault moment, that is, the moment when the fault occurs; , and are the action moments of the main protection, near-backup protection and remote-backup protection respectively. Among them, the action moment of the main protection is the moment when the main protection detects the fault and starts to trip; the action moment of the near-backup protection is the moment when the near-backup protection starts to trip when the main protection fails to operate; the action moment of the remote-backup protection is the moment when the remote-backup protection starts to trip when both the main protection and the near-backup protection fail to operate. , and are the operating times of the circuit breaker driven by the main protection, near backup protection, and far backup protection, respectively. Among them, the operating time of the circuit breaker driven by the main protection is the moment when the circuit breaker trips in response to the main protection signal; the operating time of the circuit breaker driven by the near backup protection is the moment when the circuit breaker trips in response to the near backup protection signal; the operating time of the circuit breaker driven by the far backup protection is the moment when the circuit breaker trips in response to the far backup protection signal, which usually involves the circuit breakers of adjacent lines. , and are the setting operating time limits of the main protection, near backup protection, and far backup protection compared with the fault time, respectively. Among them, the setting operating time limit of the main protection compared with the fault time is the time required for the main protection to detect the fault and initiate a trip. , and are the delays of the circuit breaker operation after each protection action, respectively.

[0024] When a bipolar short-circuit fault occurs in the DC distribution network, the voltage at the fault point changes instantaneously, and the voltage at the fault point shows a step characteristic. The fault transient signal contains full-frequency domain information. Especially in the high-frequency domain, the voltage transient waveform at the fault point contains high-frequency components, and the impedance characteristic of the converter device in the high-frequency domain shows linearity and is not affected by the converter control strategy. Therefore, the fault point can be regarded as an additional high-frequency voltage source of the fault, and its high-frequency current flows through the entire DC distribution network. The mathematical representation form of the node high-frequency impedance matrix is: In the formula, is the node high-frequency impedance matrix, and the dimension of the node high-frequency impedance matrix is N×N, representing the impedance characteristics between each node at the angular frequency w; is the node high-frequency admittance matrix, and the dimension of the node high-frequency admittance matrix is N×N. The node high-frequency admittance matrix is the inverse matrix of the node high-frequency impedance matrix; the superscript -1 is the matrix inversion operation symbol; w is the angular frequency, representing the frequency characteristics of the high-frequency signal; N is the number of nodes in the DC distribution network; is the element of the node high-frequency impedance matrix at the frequency w. When i = j, represents the self-impedance of node i; when , represents the mutual impedance between node i and j; the subscript N1 represents an N-dimensional column vector; the subscript NN represents an N-order square matrix.

[0025] Combined with the node high-frequency impedance matrix, in this embodiment, the node high-frequency voltage equation can be further derived, and the mathematical representation form of the node high-frequency voltage equation is: Wherein, is the node high-frequency voltage column vector, the dimension of the node high-frequency voltage column vector is N×1, and it includes the voltage phasors of each node at the frequency of w; is the node high-frequency current column vector, the dimension of the node high-frequency current column vector is N×1, and it includes the current phasors of each node at the frequency of w.

[0026] Short-circuit faults can be divided into bus faults and line faults according to the fault location. When a fault occurs at node i in the DC distribution network, the fault current at this node i can be equivalent to a high-frequency current source. Figure 4 is the schematic diagram of the equivalent high-frequency current source during bus faults provided by the embodiments of the present invention. At this time, the high-frequency fault voltage of the whole network is only generated by the current source of node i, and only the high-frequency injection current of node i in the injection current column vector is not zero, which is specifically expressed as: Wherein, is the equivalent high-frequency current source of node i at the frequency of w.

[0027] When a fault occurs on the line between node i and node j, the equivalent high-frequency current source is located between the two nodes and is no longer distributed on the nodes. For simplicity of analysis, in this embodiment, the high-frequency current source between the nodes is equivalent to the two adjacent nodes closest to it. Figure 5 is the schematic diagram of the equivalent high-frequency current source during line faults provided by the embodiments of the present invention. At this time, the high-frequency fault voltage of the whole network is generated by node i and node j, and only the high-frequency injection currents of node i and node j in the node injection current column vector are not zero, which is specifically expressed as: Wherein, is the high-frequency current source equivalent to node i during line faults at the frequency of w; is the high-frequency current source equivalent to node j during line faults, where the subscripts i and j represent the node numbers; in Figure 5 where, is the fault location coefficient, , which is used to characterize the relative position of the fault point in the line; is the line the high-frequency current source before equivalence; is the high-frequency equivalent impedance of the converter at node i; is the high-frequency equivalent impedance of the converter at node j; is the grading parameter of the per-unit threshold interval, which is used for breaker tripping judgment; is the high-frequency current amplitude after equivalent distribution during line faults; is the high-frequency current source after equivalent distribution when node i fails; is the high-frequency current source after equivalent distribution when node j fails.

[0028] As can be seen from the above formulas, when a bus or line fault occurs in a DC distribution network, the high-frequency fault current exhibits sparsity. Based on this, in this embodiment, voltage measurement devices are configured at S (S < N) nodes to extract the high-frequency fault voltage amplitude and the impedance values of the corresponding rows in the node impedance matrix, and then a sparse node equation for bus and line faults is constructed. The sparse node equation for bus faults is used to describe the relationship between the node voltages and currents in the DC distribution network during bus faults, and the sparse node equation for line faults is used to describe the relationship between the node voltages and currents in the DC distribution network during line faults. The mathematical representation of the sparse node equation for bus faults is as follows: The mathematical representation of the sparse node equation for line faults is as follows: In the formula, is the column vector of the node high-frequency voltage at the measurement point, is an S×1 vector that contains the high-frequency voltage information of S measurement points in the distribution network; is the impedance sub-matrix of the measurement node, which represents the S-row and N-column sub-matrix extracted from the node high-frequency impedance matrix and corresponds to the impedance relationship between S measurement nodes and all N nodes; S represents the number of voltage measurement nodes.

[0029] According to the relationship between the ranks and rows / columns of the sparse node equation for bus faults and the sparse node equation for line faults, since the ranks of the sparse node equation for bus faults and the sparse node equation for line faults are less than or equal to N, the sparse node equation for bus faults and the sparse node equation for line faults are both underdetermined equations and have infinitely many solutions. Therefore, in this embodiment, the Bayesian compressed sensing (BCS) reconstruction algorithm is used to solve the sparse node equation for bus faults and the sparse node equation for line faults to obtain the high-frequency current sparse vector .

[0030] S2. Use the normal distribution model to quantitatively analyze the non-linear attenuation trend of the high-frequency current amplitudes of each suspected fault component in the high-frequency fault current sparse vector to obtain the estimated high-frequency current amplitudes.

[0031] In some embodiments, the steps of using the normal distribution model to quantitatively analyze the non-linear attenuation trend of the high-frequency current amplitudes of each suspected fault component in the high-frequency fault current sparse vector to obtain the estimated high-frequency current amplitudes include: According to the high-frequency current amplitudes of each suspected fault component in the high-frequency fault current sparse vector, construct multiple independent normal distribution models according to the adjacent relationship of the suspected fault components; Take the maximum value of the high-frequency current amplitude among adjacent suspected faulty components as the peak value of the corresponding normal distribution model, and solve the attenuation parameter of each normal distribution model through the maximum likelihood estimation criterion; Based on the attenuation parameter, use the normal distribution model to calculate the estimated value of the high-frequency current amplitude at the midpoint of the line between each pair of adjacent suspected faulty components.

[0032] Specifically, in the actual operation of the DC distribution network, due to the influence of interference factors such as transition resistance and noise, it is difficult to distinguish line faults and bus faults only relying on the reconstructed high-frequency current sparse vector. Under ideal conditions, the high-frequency current sparse vector obtained by the reconstruction algorithm only contains the high-frequency current amplitude of the faulty component. However, in the actual operating environment, due to the influence of interference factors such as transition resistance and noise, the accuracy of the reconstruction result is often challenged, resulting in the possible mixing of the high-frequency current amplitude of non-faulty components in the high-frequency current sparse vector. This uncertainty makes it difficult to effectively distinguish line faults and bus faults based on the reconstruction result alone, and it is impossible to achieve accurate fault location. Figure 6 It is a schematic diagram of the non-linear attenuation characteristic of the high-frequency fault current amplitude. For example, take the suspected faulty bus Bi as an example. The high-frequency current amplitude at the suspected faulty bus Bi reaches the peak value, while the high-frequency current amplitudes at the adjacent buses Bg, Bh, Bj, and Bk of the suspected faulty bus Bi are the high-frequency current amplitudes 、the high-frequency current amplitude 、the high-frequency current amplitude and the high-frequency current amplitude , 、 、 and are all less than the peak value . Although in the reconstruction result, the high-frequency current amplitudes of some non-faulty buses present local maximum values, this is not sufficient to distinguish bus faults or line faults. In addition, the high-frequency current amplitude of the faulty component may also deviate due to interference factors and may even be lower than the amplitudes of some non-faulty components.

[0033] Based on the above analysis, although the electrical quantity fault information can initially delimit the fault range, it cannot achieve accurate fault section location. To solve this problem, this embodiment proposes a high-frequency current amplitude estimation method based on the normal distribution model. Take the high-frequency current amplitudes of each suspected faulty component in the reconstructed high-frequency current sparse vector as the fault probability characterization index from the electrical quantity perspective, thereby determining the electrical quantity fault degree, and using the information fusion method to achieve the comprehensive evaluation of the multi-source fault degree, so as to accurately locate the faulty component. According to the sparse node equation during bus and line faults, the reconstruction result It can only directly characterize the high-frequency current amplitude of the suspected faulty bus, while the high-frequency current amplitudes of each suspected faulty line present an implicit characteristic distribution. To ensure the completeness of subsequent fault location based on multi-source fault information fusion, in this embodiment, these implicit characteristics are explicitly solved to obtain the high-frequency current amplitudes of each suspected faulty line. However, since the high-frequency current amplitudes of each suspected faulty component present a non-linear attenuation characteristic, it is impossible to directly construct a linear equation to solve the high-frequency current amplitudes of each suspected faulty line. Therefore, in this embodiment, the normal distribution model in probability statistics theory is introduced, and the non-linear attenuation trend of the high-frequency current amplitude is quantitatively characterized by constructing a normal distribution model, and the explicit solution of the high-frequency current amplitude is realized. Considering that it is difficult to ensure that all high-frequency current amplitudes are distributed on the normal distribution model by constructing only one normal distribution model to characterize the high-frequency current amplitudes of all suspected faulty buses, which will introduce a large error and increase the uncertainty of fusion location. Therefore, in this embodiment, based on the high-frequency current amplitudes of the suspected faulty buses in the high-frequency sparse vector, multiple groups of normal distribution models are constructed respectively. The specific mathematical expression of the constructed normal distribution model is as follows: In the formula, is the normal distribution model constructed for the suspected faulty bus Bi; is the standard deviation of the normal distribution, which characterizes the amplitude attenuation rate; x is the random variable of the normal distribution, which represents the spatial position variable; e is the base of the exponential function.

[0034] Based on the constructed multiple groups of normal distribution models, in this embodiment, the high-frequency current amplitude sequence of the set of suspected faulty buses is used as the input quantity, the larger value of the high-frequency current amplitudes between two adjacent suspected faulty buses is used as the peak value of the constructed normal distribution model, and the attenuation parameters of the normal distribution model are solved respectively based on the maximum likelihood estimation criterion. The attenuation parameters include the standard deviation used to control the attenuation rate and the electrical distance difference between two adjacent suspected faulty components under the normal distribution. The specific calculation formulas of the attenuation parameters and are as follows: In the formula, is the standard deviation of the normal distribution model between the bus Bi and the bus Bj, which is solved by maximum likelihood estimation; is the high-frequency current amplitude at the bus Bi; is used to represent the distance difference between the bus Bi and the bus Bj; is the high-frequency current amplitude at the bus Bj.

[0035] This embodiment combines and The combined estimation result can be further used to solve the amplitude of the high-frequency current on the line. During the calculation process, in this embodiment, the middle position of the line is taken for estimation to obtain the estimated value of the high-frequency current amplitude at the midpoint position of the line between each pair of adjacent suspected fault components. The specific calculation formula for the estimated value of the high-frequency current amplitude is: In the formula, is the high-frequency current amplitude at the midpoint of the line ; is the equivalent distance between bus bar Bi and bus bar Bj under the normal distribution, which is used to reflect the electrical distance of the bus bar.

[0036] By repeating the above steps in this embodiment, a comprehensive estimation of the high-frequency current amplitude of each suspected fault line can be achieved. By quantifying the non-linear attenuation characteristics of the amplitude, the implicit line amplitude distribution is transformed into explicit parameters, so that the high-frequency current amplitude of the suspected fault component is used as an indicator for characterizing the fault probability from the perspective of electrical quantities, determining the electrical quantity fault degree, and then using the information fusion method to achieve the fusion of multi-source fault degrees, so as to accurately locate the fault component.

[0037] To verify the completeness of the proposed method at the theoretical level, this embodiment further analyzes whether the above calculation process satisfies the objective law of the normal distribution function. Figure 7 is a schematic diagram of the normal distribution model under different peak values provided by the embodiment of the present invention. In the constructed normal distribution model, there is a clear quantitative relationship between parameters. When the high-frequency current amplitude of bus bar Bi increases, due to the influence of impedance parameters and interference factors in the distribution network, may show related characteristics of increasing or remaining unchanged, showing a negative correlation, Figure 7 The specific manifestation of the quantitative relationship between the model parameters in is: when increases to , the high-frequency current amplitude of the adjacent bus bar Bj synchronously increases to , resulting in decreasing to , increases to . At the same time, it can be seen from the above attenuation parameter calculation formula that and show a significant negative correlation. Based on the above analysis, and are positively correlated, while with There is a negative correlation, and this proportional relationship satisfies the parameter constraint conditions of the normal distribution function. Moreover, when the parameters change, it is still possible to effectively solve the amplitudes of each suspected fault line, verifying the theoretical completeness of the evaluation method. Among them, is the estimated value of the fault high-frequency current amplitude after the increase of bus Bi, which is an adjusted value considering interference factors; is the estimated value of the fault high-frequency current amplitude after the increase of bus Bj, which is an adjusted value considering interference factors; is for the line is the estimated value of the fault high-frequency current amplitude after the increase of the midpoint of the line, which is an adjusted value considering interference factors; is the corrected line 's normal distribution model; is the standard deviation of the normal distribution model between bus Bi and bus Bj; is for the line 's normal distribution model, which is used to describe the attenuation characteristics of the amplitude along the line; is the corrected spatial coordinate considering interference factors, which reflects the fault position adjustment amount; is the normal distribution distance at bus Bi, which is used to locate the relative position of the fault point; is the normal distribution distance at bus Bj, which is used to locate the relative position of the fault point.

[0038] S3. Cube-normalize the estimated values of the high-frequency current amplitudes of each suspected fault component to obtain the electrical quantity fault degrees of each suspected fault component.

[0039] In some embodiments, the step of cube-normalizing the estimated values of the high-frequency current amplitudes of each suspected fault component to obtain the electrical quantity fault degrees of each suspected fault component includes: Perform cube operations on the estimated values of the high-frequency current amplitudes of each suspected fault component respectively to obtain the cubic amplitudes of the high-frequency currents of each suspected fault component; Normalize the cubic amplitudes of the high-frequency currents to obtain the electrical quantity fault degrees of each suspected fault component.

[0040] Specifically, to achieve accurate fault location based on multi-source fault information fusion, this embodiment needs to convert two types of quantities with different natures, namely electrical quantities and switch quantities, into a unified dimension for comprehensive analysis. This embodiment uses the high-frequency current amplitude of the component as the characterization index of the electrical quantity fault degree. To ensure the comparability of the electrical quantity fault degrees of different components, this embodiment normalizes the relevant data such as the high-frequency current amplitudes of each component, thereby obtaining the electrical quantity fault degrees corresponding to each suspected fault component. The calculation formula of the electrical quantity fault degree is: In the formula, is the electrical quantity fault degree of the th suspected faulty component; is the cubic amplitude of the high-frequency current of the th suspected faulty component, that is, the cube of the amplitude of the high-frequency current of the th suspected faulty component; is the cubic amplitude of the high-frequency current of bus Bi; is the cubic amplitude of the high-frequency current of line ; n is the total number of suspected faulty components, where ; is the index of the suspected faulty component.

[0041] In this embodiment, through normalization processing, the electrical quantity fault degrees of each component can accurately reflect their relative importance in the case of a fault, effectively ensuring the probabilistic completeness of the obtained electrical quantity fault degrees, and providing an accurate and reliable data basis for the subsequent fusion of multi-source fault information and fault location.

[0042] S4. Modify the node states of the relay protection Bayesian network model according to the pre-identified local switch quantity action information, and calculate the switch quantity fault degrees of each suspected faulty component through reverse inference.

[0043] In some embodiments, the steps of modifying the node states of the relay protection Bayesian network model according to the pre-identified local switch quantity action information and calculating the switch quantity fault degrees of each suspected faulty component through reverse inference include: Set three-level sampling time windows according to the sequential action logic of DC distribution network relay protection, and determine the protection level where the tripped circuit breaker is located by detecting the high-frequency transient voltage signals within each level of sampling time window, and obtain the local circuit breaker action state of the protection level where the tripped circuit breaker is located; Identify the local switch quantity action information according to the local circuit breaker action state, and use the local switch quantity action information to correct the switch quantity information uploaded by the data acquisition and monitoring system to obtain the switch quantity action correction information; Construct a relay protection Bayesian network model for the DC distribution network according to the relay protection action logic of the DC distribution network, and update the corresponding relay protection node states and circuit breaker node states in the relay protection Bayesian network model by using the switch quantity action correction information to obtain a relay protection Bayesian network correction model; Based on each suspected faulty component determined from the perspective of electrical quantities, perform reverse probability inference using the relay protection Bayesian network correction model to calculate the post-fault probability of each suspected faulty component; Perform cubic normalization on the post-fault probabilities of each suspected faulty component to obtain the switch quantity fault degrees of each suspected faulty component.

[0044] Specifically, when a component in the DC distribution network fails, the relay protection system will drive the circuit breaker to trip, thereby isolating the faulty area and ensuring the safe and stable operation of the power grid. Based on this action logic, a Bayesian network model for relay protection of the DC distribution network is constructed in this embodiment. Figure 8 FIG. Figure 8 is a schematic diagram of the Bayesian network model for relay protection of the DC distribution network provided by the embodiment of the present invention. Each network node in the Bayesian network model for relay protection is interconnected based on the action logic of relay protection, and its connection sequence is: component, main protection, near backup protection, proximal circuit breaker, far backup protection, and distal circuit breaker. Among them, the malfunction state of the network node is represented by 0, and the normal operation state is represented by 1. In this embodiment, the prior probabilities of faults of DC lines and buses are set in combination with the actual DC distribution network engineering situation. Among them, the prior probability of DC line fault is 0.1187, and the prior probability of DC bus fault is 0.1123.

[0045] The state of network nodes in the Bayesian network model for relay protection is usually determined by the digital quantity action information collected by systems such as SCADA. However, when systems such as SCADA collect digital quantity information, situations such as missed reports and false reports may occur, which will affect the accuracy of subsequent multi-source fault information fusion and location. To address this problem, in this embodiment, the local digital quantity action information is accurately identified, and the missed report and false report information is corrected to ensure the accuracy of the subsequent fusion information. Specifically, this embodiment will elaborate on the acquisition method of local digital quantity action information from two aspects: the identification of circuit breaker action state and the identification of protection device action state. In this embodiment, the accurate identification of the high-frequency voltage signal generated at the moment of circuit breaker tripping is used to achieve the precise identification of the circuit breaker action state. First, taking the two-stage change trend of the transient voltage at the fault location as the starting point, the generation mechanism of the high-frequency voltage signal (hereinafter referred to as the high-frequency voltage signal) during the circuit breaker tripping process is specifically analyzed. Figure 9 FIG. Figure 9 is a schematic diagram of the transient voltage change trend at the fault location provided by the embodiment of the present invention. is the moment when a fault occurs. is the moment when the protection operates. is the moment when the circuit breaker trips to isolate the fault. At , the voltage at the fault location drops sharply from the normal operating voltage value to the voltage value ; at , the voltage drops suddenly from the voltage value to 0. and The change trends of the voltage at the fault location at the moment both show the characteristics of an approximate step signal. The special mutation form of the step signal makes it have rich full-frequency domain information. Among them, the mathematical expression of the unit step signal is: In the formula, is the unit step function, which is used to characterize the voltage mutation characteristic; t is the time variable of the fault process.

[0046] The mathematical expression of the Fourier transform result of the unit step signal is: In the formula, is the Fourier transform of the step signal, which reflects the full-frequency domain characteristic; is the pi; is the Dirac function, which represents the DC component and the impulse when the angular frequency w is zero; j in jw represents the imaginary unit, and w in jw represents the angular frequency.

[0047] According to the above analysis, it can be known that at moment, a stable high-frequency voltage signal can be detected at the fault location. On this basis, this embodiment further analyzes whether a high-frequency voltage signal can be detected at moment. At moment, the spectral characteristics of the voltage at the circuit breaker tripping location are as shown in Figure 10 . It can be seen from Figure 10 that when the circuit breaker trips, the amplitude of the voltage between poles changes greatly in the low-frequency band, the spectral density is small, the nonlinearity is strong, and the stability is poor; while in the high-frequency band, the amplitude changes little, the spectral density is large, its envelope is approximately linear, and the stability is good. This indicates that a high-frequency voltage signal can also be detected at the moment when the circuit breaker trips. Figure 11 is the schematic diagram of the detection result of the high-frequency voltage signal within the 0- time period provided by the embodiment of the present invention.

[0048] Within the 0- time period, high-frequency voltage signals can be detected at both moment and moment. However, the peak value of the high-frequency voltage signal at moment is much larger than the signal peak value detected at moment. The reason is that compared with the bipolar short-circuit fault, the voltage fluctuation caused by the moment when the circuit breaker trips is smaller, so that the amplitude of the high-frequency voltage signal is also smaller. Based on the above analysis, this embodiment constructs a three-level sampling time window mechanism in combination with the timing logic of the relay protection. The three-level sampling time window mechanism sets sampling time windows during the protection periods of each level. Among them, the three-level sampling time window includes a first-level sampling time window, a second-level sampling time window, and a third-level sampling time window; the first-level sampling time window corresponds to the main protection action period, the second-level sampling time window corresponds to the near-backup protection action period, and the third-level sampling time window corresponds to the far-backup protection action period. This embodiment can determine the protection period where the tripped circuit breaker is located by detecting high-frequency voltage signals in different sampling time windows. Figure 12It is a schematic diagram of a three - level sampling time window based on the time - sequence logic of relay protection. If signals are detected only within the single sampling time window 1, sampling time window 2, and sampling time window 3, it indicates that the circuit breaker tripped only during each level of protection and there was no circuit breaker refusal to trip during the process of isolating the fault. On the contrary, if signals are detected within multiple sampling time windows, it indicates that there is a circuit breaker malfunction during the process of isolating the fault (in this embodiment, circuit breaker misoperation is not considered here). It should be noted that since the bus protection only has a main protection and a backup protection, when isolating a bus fault, if there is a circuit breaker malfunction, signals can be detected in sampling time window 1 and sampling time window 2, and no signal can be detected in sampling time window 3. Among them, in this embodiment, the sampling frequency is preferably set between 1 kHz and 1.5 kHz.

[0049] Based on the known protection period where the tripped circuit breaker is located, this embodiment further determines the line to which the tripped circuit breaker belongs. The high - frequency voltage amplitude signal generated at the moment of circuit breaker tripping is small. At the same time, due to the existence of the high - frequency boundary, compared with the line at the current fault level, the high - frequency voltage flowing through the lower - level line decreases. Therefore, this embodiment stipulates that the measuring point can detect high - frequency voltage signals at its own line, adjacent lines, and separated lines. For the convenience of analysis, this embodiment sets a three - level per - unit threshold criterion as the discriminant basis for the line to which the tripped circuit breaker belongs. At the same time, taking the maximum value of the high - frequency voltage amplitude as the reference value and normalizing it to 1, the mathematical representation form of the three - level per - unit threshold interval constructed in this embodiment is: In the formula, is the first - level threshold interval; is the second - level threshold interval; is the third - level threshold interval.

[0050] When the per - unit value of the high - frequency voltage amplitude is within the three different intervals of the first - level threshold interval α, the second - level threshold interval β, and the third - level threshold interval λ, it indicates that there is a circuit breaker tripping in the measuring - point line, adjacent lines, and separated lines. When it is within the first - level threshold interval α, it indicates that the measuring - point line trips; when it is within the second - level threshold interval β, it indicates that the adjacent line trips; when it is within the third - level threshold interval λ, it indicates that the separated line trips. In addition, if the per - unit value is lower than 0.1 p.u., it is regarded as an error caused by system fluctuations, and it can be determined that there is no circuit breaker tripping. Among them, p.u. is the unit of the per - unit value. Through the above - mentioned method, this embodiment can accurately identify the operating state of the circuit breaker and the line to which it belongs, providing reliable information support for subsequent fault location.

[0051] After identifying the operating state of the circuit breaker, this embodiment further analyzes the specific protection types that drive the circuit breaker to trip. In this embodiment, the same circuit breaker can be driven by different protections. Therefore, this embodiment first summarizes and organizes various protection drive types of the line to which the tripped circuit breaker belongs; secondly, combines the detection results of the sampling time window and the sequence logic of the relay protection to preliminarily screen each protection drive type after summarization and organization, so as to narrow the range of protection drive types; finally, combines the action logic of the relay protection to reverse-deduce the actually operating protection type. For the convenience of clear description, this embodiment takes identifying the relay protection types of the tripped circuit breakers on line and line as an example for detailed description. Suppose high-frequency voltage signals are detected on line and line respectively in sampling time window 1 and sampling time window 3. This embodiment first summarizes the protection drive types of each protection on line and line . The summary results of the protection drive types of the circuit breakers on line and line are shown in Table 2. It can be seen from Table 2 that for line , since the circuit breaker tripped during the main protection period, this indicates that it may be driven by the main protections of line , bus B2 and bus B3; for line , since the circuit breaker tripped during the remote backup protection period, it means that it may be driven by the remote backup protections of line and line at this time. At the same time, this embodiment considers that there is a situation where the circuit breaker fails to operate during the isolation of the fault, and no high-frequency voltage signal is detected in sampling time window 2. From this, it can be inferred that the tripping of the circuit breaker on line is not driven by the relevant protections of the bus, and then excludes the main protections of bus B2 and bus B3, and determines that the operating circuit breaker on line is driven by the main protection of line . In addition, if the tripped circuit breaker on line is driven by the remote backup protection of line , a high-frequency signal should be detected on line , but this situation does not occur in this scenario. Based on this, this embodiment excludes the remote backup protection of line , and can determine that the tripped circuit breaker on line is driven by the remote backup protection of line . Through the above analysis, it can be seen that the method proposed in this embodiment can effectively identify the operating state of the protection and combine the identification results of the circuit breaker and the protection operating state to form local digital input / output action information. Table 2 is as follows: Table 2 In this embodiment, for each suspected faulty component determined from the perspective of electrical quantities, a corresponding Bayesian network model for relay protection is constructed. Then, the local switch action information is compared with the switch action information uploaded by systems such as SCADA. Through this comparison process, the false alarm information in the switch action information uploaded by systems such as SCADA is corrected and the missing information is supplemented. Furthermore, the incorrect network node states in each Bayesian network model for relay protection are corrected to ensure that the states of all network nodes in the Bayesian network model for relay protection are accurate and reliable. On this basis, this embodiment combines the corrected correct component state information and performs reverse reasoning on the Bayesian network model for relay protection to obtain the posterior probability of failure of each suspected faulty component. Finally, the posterior probability of failure of each suspected faulty component is subjected to cubic normalization processing to obtain the switch fault degree. Figure 13 It is a schematic diagram of the overall process for evaluating the switch fault degree provided by an embodiment of the present invention. The calculation formula for the switch fault degree is: In the formula, is the switch fault degree of the th suspected faulty component; is the cube of the original posterior probability of failure of the th suspected faulty component obtained based on logical judgment; is the cube of the posterior probability of failure of bus Bi; is the cube of the posterior probability of failure of line

[0052] In this embodiment, the switch fault degree focuses more on the inference of the probability of failure based on the probability model, while the electrical quantity fault degree directly comes from the physical measurement of the electrical quantity. The electrical quantity fault degree reflects the electrical characteristics during the fault. The two characterize the fault information of the component from different perspectives and are respectively used for the calculation of the switch fault degree and the electrical quantity fault degree.

[0053] S5. Taking the electrical quantity fault degree and the switch fault degree as independent evidence bodies, use the improved Dempster-Shafer evidence theory identification framework introducing the universal set subset to calculate the weight of the projection distance of the credible interval from the perspective of evidence theory and the weight of the information entropy from the perspective of probability theory respectively.

[0054] In some embodiments, the steps of using the improved Dempster-Shafer evidence theory identification framework introducing the universal set subset to calculate the weight of the projection distance of the credible interval from the perspective of evidence theory and the weight of the information entropy from the perspective of probability theory respectively include: According to the improved Dempster-Shafer evidence theory that introduces the subsets of the universal set, define the credible interval and the maximum uncertainty interval of each suspected faulty component in the evidence body, and represent the credible interval and the maximum uncertainty interval in the form of interval numbers; Based on the credible interval and the maximum uncertainty interval of each suspected faulty component represented in the form of interval numbers, calculate the projection distance between the credible interval and the maximum uncertainty interval of each suspected faulty component in the evidence body; Use the projection distance as an index to measure the uncertainty of the evidence body, and obtain the uncertainty measurement value of the evidence body; According to the uncertainty measurement value of the evidence body and the preset evidence adjustment coefficient, obtain the weight of the projection distance of the credible interval of the evidence body from the perspective of evidence theory; Equalize the probability assignments of the subsets of the universal set in the evidence body to all suspected faulty components to obtain the component probability distribution, and calculate the information entropy of the component probability distribution; According to the information entropy of the component probability distribution and the preset probability adjustment coefficient, calculate and obtain the information entropy weight of the evidence body from the perspective of probability theory.

[0055] Specifically, the Dempster-Shafer evidence theory (DS evidence theory) can effectively handle the uncertainty between evidence bodies and can present evidence information in a more intuitive and comprehensive way. Due to this special advantage, this evidence theory has been widely used in the field of information fusion. In the evidence theory, the identification framework is a finite non-empty set composed of different elements. In this embodiment, the identification framework is defined as the set of all suspected faulty components in the DC distribution network. The specific expression of the identification framework is: In the formula, is the identification framework, that is, the set of all suspected faulty components; is an element in the identification framework, which represents the nth suspected faulty component (such as a bus or a line); n is the total number of suspected faulty components.

[0056] On the given identification framework, if there is a function m that satisfies the specified constraint conditions, then this function m is called the basic probability assignment function of the evidence theory. The mathematical representation form of the constraint conditions specified by the function m is: In the formula, is the empty set is the basic probability assignment of the empty set; is the basic probability assignment function; is a subset in the identification framework, and this subset can contain one or more suspected faulty components; is the subset is a proper subset of the identification framework; is an empty set, which does not contain any faulty components.

[0057] When is the case, is called a focal element of the identification framework, where is the confidence assignment for the suspected faulty component . The confidence interval of the basic probability assignment function m can be expressed as: In the formula, is the confidence interval of the subset based on the evidence body m under the identification framework ; is the belief function of the subset based on the evidence body m under the identification framework ; is the likelihood function of the subset based on the evidence body m under the identification framework ; is the i-th subset in the identification framework ; is the basic probability assignment function of the focal element subset; is the subset of the focal element.

[0058] The core of the DS evidence theory lies in the Dempster combination rule, which reflects the fusion process among various evidence bodies. The specific expression of the combination rule is: In the formula, are all evidence bodies; is the basic probability assignment of the evidence body to the subset ; is the basic probability assignment of the evidence body to the subset ; is the basic probability assignment of the evidence body to the subset ; k is the conflict factor between evidence bodies, which is used to measure the conflict degree between different evidence bodies. When k = 1, it indicates that the evidence bodies are mutually contradictory, and at this time, the combination rule cannot perform fusion calculation; represents the intersection of evidence bodies, which reflects the consistency degree of different evidences.

[0059] To avoid the failure of the combination rule due to an excessive conflict coefficient, in this embodiment, the traditional DS evidence theory is improved from the perspective of modifying the evidence body. On the basis of adding the universal set to the identification framework, the allocation of the fault degree weight is realized from two perspectives: the evidence theory and the probability theory, and the fused processing of the weighted fault degree is carried out. It should be noted that the conflict between evidence bodies mainly stems from the situation where there is no intersection among the focal elements in the evidence body. Based on this, in this embodiment, by supplementing the subsets in the identification framework and increasing the intersection of the focal elements in the evidence body, the conflict between evidence bodies is reduced, and at the same time, it is ensured that the final decision result is not affected. Since the universal set has an intersection with all focal elements, in this embodiment, the universal set is introduced into the identification framework so that each evidence body contains this universal set. At this time, the expression of the identification framework of the improved Dempster-Shafer evidence theory introducing the universal set subset is: In the formula, is the identification framework of the improved Dempster-Shafer evidence theory introducing the universal set subset, which contains all subsets of the original identification framework and the universal set itself; is the universal set, which represents the set of all possible faulty components in the identification framework.

[0060] In this embodiment, the electrical quantity fault degree and the switch quantity fault degree are used as the evidence body and the evidence body respectively, and the specific representation form is: In the formula, is the universal set coefficient of the electrical quantity evidence; is the basic probability assignment of the electrical quantity evidence to the universal set; is the basic probability assignment of the switch quantity evidence to the universal set; is the universal set coefficient of the switch quantity evidence.

[0061] From the perspective of the evidence theory, in this embodiment, a method for measuring the uncertainty of the evidence body is proposed. This method uses the projection distance between the confidence interval of the evidence body and the maximum uncertainty interval as an index to measure the uncertainty of the evidence body. First, in this embodiment, the confidence interval of the suspected faulty components in the evidence body is constructed according to the confidence interval of the basic probability assignment function m, and at the same time, the confidence interval and the maximum uncertainty interval are processed as interval numbers. The definition of the interval number is as follows: In the formula, is the confidence interval number; is the minimum value of the confidence interval; is the maximum value of the confidence interval; x is the central value of the number of confidence intervals; R is the radius of the number of confidence intervals, which is used to measure the uncertainty range of the interval.

[0062] Secondly, in this embodiment, the projection distance between the confidence interval numbers and the maximum uncertainty interval numbers of each suspected fault in the evidence body is calculated, and the calculation formula of the projection distance is as follows: In the formula, is the projection distance between the confidence interval number and the maximum uncertainty interval number; is the upper bound of the maximum uncertainty interval; is the lower bound of the maximum uncertainty interval; b is the maximum uncertainty interval; is the Euclidean norm.

[0063] Then, in this embodiment, the projection distance is used as an index to measure the uncertainty of the evidence body, and the uncertainty measurement value of the evidence body is obtained. The specific expression of the uncertainty measurement value is: In the formula, is the uncertainty magnitude of each evidence body. The larger its value, the greater the uncertainty of the evidence body; MUI is the maximum uncertainty interval, . On this basis, in this embodiment, the evidence body weights are allocated according to the uncertainty measurement value of the evidence body and the preset evidence adjustment coefficient, and the confidence interval projection distance weight of the evidence body from the perspective of evidence theory is obtained. The mathematical expression of the confidence interval projection distance weight is: In the formula, is the confidence interval projection distance weight of the evidence body , that is, the evidence theory weight; is the preset evidence adjustment coefficient; is the evidence body 's uncertainty measurement; is the evidence body; M is the total number of evidence bodies.

[0064] It can be seen from the mathematical expression of the confidence interval projection distance weight that is inversely proportional to , that is, the greater the uncertainty of the evidence body, the smaller the weight it occupies.

[0065] Next, this embodiment elaborates in detail the weight allocation from the perspective of probability theory. Probability theory can be regarded as a further generalization of evidence theory. By converting the evidence under evidence theory into probability, the corresponding probability is obtained, and then the information entropy of the probability is calculated to indirectly measure the uncertainty of the evidence body from the perspective of probability theory. Among them, the definition of information entropy is: In the formula, is the information entropy of the evidence body m; is the evidence body for the subset basic probability assignment.

[0066] In this embodiment, the weights of the evidence bodies are allocated according to the proportion of information entropy, and the information entropy weight allocation formula is: In the formula, is the information entropy weight, that is, the probability theory weight; is the preset probability adjustment coefficient; is the evidence body information entropy.

[0067] It can be seen from the information entropy weight allocation formula that the smaller the information entropy of the evidence body, the less uncertain information it carries, and the more weight it occupies.

[0068] S6. Combine the credibility interval projection distance weights and information entropy weights of each evidence body through the combined weighting method, and use the combined weights to weight and fuse all evidence bodies to obtain the combined fault probabilities of each suspected fault component.

[0069] To achieve the fusion of multi-source fault information, this embodiment proposes a method for combining the weights of evidence bodies. Specifically, after completing the weight allocation from the perspectives of evidence theory and probability theory to obtain the credibility interval projection distance weight and information entropy weight After that, this embodiment combines the credibility interval projection distance weights and information entropy weights of each evidence body through the combined weighting method to determine the final weight allocation of each evidence body, ensuring that the weights can comprehensively reflect the credibility and uncertainty of the evidence body, so as to obtain the combined weights. The combined weights not only reflect the relative importance of the evidence body in the evaluation of electrical quantity and switch quantity fault degrees, but also balance the weight contributions from two different theoretical perspectives. The mathematical expression form of the combined weights is: In the formula, is the evidence body combined weight.

[0070] In this embodiment, the evidence bodies are weighted and averaged according to the fused weights to obtain the fused fault probabilities of each suspected fault component. The weighted average process fully considers the final weights of each evidence body, ensuring that the contribution to the fusion result is proportional to the credibility of the evidence body. The fused fault probabilities obtained through weighted average can more accurately reflect the comprehensive fault probabilities of each suspected fault component. In this embodiment, decisions are made based on the fused fault probabilities of each suspected fault component, and the component with the maximum fused fault probability is the fault component. Through dual-weight optimization, this embodiment not only considers the conflict degree between evidences but also takes into account the certainty of probability distribution, significantly improving the reliability of fault location. Among them, the mathematical representation of the weighted average process for evidence bodies is as follows: In the formula, is the fused fault probability; is the evidence body for the subset basic probability assignment.

[0071] S7. Locate the fault section of the DC distribution network according to the fused fault probability, and drive the circuit breaker to operate to isolate the fault.

[0072] In this embodiment, the high-frequency fault voltage components in the 3 ms before and after the fault are extracted according to the line parameters of the DC distribution network, and then the extracted high-frequency fault voltage components are processed by the BSC reconstruction algorithm to obtain the corresponding results to determine the suspected fault components. At the same time, this embodiment calculates the electrical quantity fault degree by using the electrical quantity evaluation method, and based on the accurate identification of the high-frequency voltage signal, combined with the timing action logic of the relay protection, accurately identifies the action states of the circuit breaker and the protection. On this basis, the error information in systems such as SCADA is corrected, and reverse reasoning is performed on each Bayesian network after the component state information is corrected to calculate the switch quantity fault degree. At the same time, this embodiment improves the traditional DS evidence theory. By supplementing the subsets of the recognition framework, the weights from the perspectives of evidence theory and probability theory are calculated respectively, and the weighted evidence bodies are fused based on the Dempster rule. The component corresponding to the maximum fused fault probability is determined as the fault section, so as to achieve the accurate and rapid location of the fault section of the DC distribution network.

[0073] To verify the correctness and effectiveness of the proposed fault section location method in this embodiment, this embodiment builds an IEEE33-node DC distribution network simulation model based on PSCAD / EMTDC, and conducts research taking line faults and bus faults as examples. For the simulation analysis of line bipolar short-circuit faults, this embodiment sets the line to have a bipolar short-circuit fault, and the main protection of the line drives the circuit breaker to operate, and the remote backup protection drives the line The circuit breaker at the left end trips to isolate the fault, Figure 14 is a schematic diagram of the fault location process based on the improved Dempster-Shafer evidence theory identification framework provided by an embodiment of the present invention. The high-frequency current sparse vector reconstructed by the BCS reconstruction algorithm in this embodiment, Figure 15 is a three-dimensional schematic diagram of the high-frequency fault current sparse vector reconstructed by using the Bayesian compressive sensing reconstruction algorithm provided by an embodiment of the present invention. It can be seen from it that there are relatively large high-frequency current amplitudes in bus B13, bus B14, bus B15 and bus B16. Furthermore, it is judged that the suspected fault components are bus B13, bus B14, bus B15, bus B16, line line and line . The set fault line is in the fault area. After determining the fault range, the electrical quantity fault degree is obtained based on the above fault degree evaluation method, as shown in Table 3: Table 3 As can be seen from Table 3, the discrimination degree of each suspected fault component in the electrical quantity fault degree is relatively low. The fault degree of line is 22.91%, which is lower than the fault degree of the non-fault component B14 and is relatively close to the fault degree of the non-fault component . This indicates that if only a single electrical quantity fault information is used for fault location, mislocation will occur. To avoid this problem, in this embodiment, the electrical quantity fault information is adjusted from the perspective of switch quantities. In this embodiment, it is necessary to judge whether there are missing reports and false reports in the relay protection action timing information collected by systems such as SCADA, and assign the corrected switch quantity information to the Bayesian network. On the premise of knowing the suspected fault components, the protection driving methods of the circuit breakers on line and line are shown in Table 4: Table 4 Table 4 shows the specific protection driving methods of the circuit breakers on line and line . Since the circuit breaker trip on line occurs during the main protection period, the backup protection of the line and the bus can be excluded. After analysis, it can be seen that the tripped circuit breaker on line is driven by the main protection of line . Combining the analysis of Table 4, it can be known that the local switch quantity action information is that line , circuit breaker and circuit breaker act, circuit breaker refuses to act, and the action of circuit breaker is false information, and at the same time, there is a missing report of For the rejection information, the corrected digital quantity information is assigned to the Bayesian network. On this basis, the faulty component has a digital quantity fault degree of 83.95%, which is much higher than that of other non-faulty components. Compared with the electrical quantity fault degree, the digital quantity fault degree has a higher discrimination degree and stronger credibility for the faulty component. To ensure the objectivity of the positioning result and further improve the positioning accuracy, the electrical quantity fault degree and the digital quantity fault degree are fused. The fusion result is shown in Table 5: Table 5 It can be seen from the fusion result in Table 5 that the faulty component is the line , and its fusion positioning result is 94.70%. Compared with the single fault degree, the positioning accuracy is further improved, avoiding the problem of positioning errors, and verifying the accuracy of the fault section positioning method proposed in the present invention when the line fails.

[0074] An embodiment of the present invention provides a method for locating a fault section of a DC distribution network based on multi-source information fusion. The method reconstructs a high-frequency fault current sparse vector according to the high-frequency transient voltage signal generated instantaneously when a bipolar short-circuit fault occurs in the DC distribution network; uses a normal distribution model to quantitatively analyze the non-linear attenuation trend of the high-frequency current amplitude of each suspected faulty component in the high-frequency fault current sparse vector to obtain an estimated value of the high-frequency current amplitude; performs cubic normalization on the estimated values of the high-frequency current amplitudes of each suspected faulty component to obtain the electrical quantity fault degree of each suspected faulty component; corrects the node state of the relay protection Bayesian network model according to the pre-identified local digital quantity action information, and calculates the digital quantity fault degree of each suspected faulty component through reverse reasoning; takes the electrical quantity fault degree and the digital quantity fault degree as independent evidence bodies, and respectively calculates the weight of the credible interval projection distance from the perspective of evidence theory and the weight of information entropy from the perspective of probability theory by using an improved Dempster-Shafer evidence theory identification framework introducing the universal set and subsets; fuses the weight of the credible interval projection distance and the weight of information entropy of each evidence body through a combined weighting method, and uses the fused weights to weight and fuse all evidence bodies to obtain the fusion fault probability of each suspected faulty component; locates the fault section of the DC distribution network according to the fusion fault probability, and drives the circuit breaker to operate to isolate the fault. Compared with the prior art, this method realizes the fast and accurate positioning of the fault section of the DC distribution network and drives the circuit breaker to operate to isolate the fault by fusing the high-frequency transient electrical quantity characteristics and the digital quantity action information, and effectively improves the fault tolerance ability of fault positioning, ensuring the safe and stable operation of the power grid.

[0075] It should be noted that the magnitudes of the serial numbers of the above processes do not mean the order of execution. The order of execution of each process should be determined by its function and internal logic, and should not constitute any limitation to the implementation process of the embodiments of the present application.

[0076] In one embodiment, as Figure 16 shown, an embodiment of the present invention provides a DC distribution network fault section location system based on multi-source information fusion. The system includes: A signal reconstruction module 101, configured to reconstruct a high-frequency fault current sparse vector according to a high-frequency transient voltage signal generated instantaneously when a bipolar short-circuit fault occurs in the DC distribution network; An amplitude estimation module 102, configured to quantitatively analyze the non-linear attenuation trend of the high-frequency current amplitudes of each suspected fault component in the high-frequency fault current sparse vector by using a normal distribution model, and obtain high-frequency current amplitude estimation values; An electrical quantity analysis module 103, configured to perform cubic normalization on the high-frequency current amplitude estimation values of each suspected fault component to obtain the electrical quantity fault degrees of each suspected fault component; A switch quantity analysis module 104, configured to correct the node state of the relay protection Bayesian network model according to the pre-identified local switch quantity action information, and calculate the switch quantity fault degrees of each suspected fault component through reverse inference; A double-weight analysis module 105, configured to use the electrical quantity fault degree and the switch quantity fault degree as independent evidence bodies, and respectively calculate the projection distance weight of the confidence interval from the perspective of evidence theory and the information entropy weight from the perspective of probability theory by using an improved Dempster-Shafer evidence theory recognition framework introducing the universal set subset; A probability fusion module 106, configured to fuse the projection distance weight of the confidence interval and the information entropy weight of each evidence body by using a combined weighting method, and use the fused weights to weight and fuse all evidence bodies to obtain the fusion fault probabilities of each suspected fault component; A fault location module 107, configured to locate the fault section of the DC distribution network according to the fusion fault probability, and drive the circuit breaker to act to isolate the fault.

[0077] For the specific limitations on a DC distribution network fault section location system based on multi-source information fusion, reference can be made to the above limitations on a DC distribution network fault section location method based on multi-source information fusion, which will not be elaborated here. Those of ordinary skill in the art can realize that, in combination with the various modules and steps described in the embodiments disclosed in the present application, they can be implemented by hardware, software, or a combination of both. Whether these functions are executed in a hardware or software manner depends on the specific application and design constraints of the technical solution. Professional technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of the present application.

[0078] An embodiment of the present invention provides a DC distribution network fault section location system based on multi-source information fusion. The signal reconstruction module of the system reconstructs a high-frequency fault current sparse vector according to the high-frequency transient voltage signal generated instantaneously when a bipolar short-circuit fault occurs in the DC distribution network; the amplitude estimation module uses a normal distribution model to quantitatively analyze the non-linear attenuation trend of the high-frequency current amplitudes of each suspected fault component in the high-frequency fault current sparse vector to obtain high-frequency current amplitude estimation values; the electrical quantity analysis module performs cubic normalization on the high-frequency current amplitude estimation values of each suspected fault component to obtain the electrical quantity fault degrees of each suspected fault component; the switch quantity analysis module corrects the node states of the relay protection Bayesian network model according to the pre-identified local switch quantity action information, and calculates the switch quantity fault degrees of each suspected fault component through reverse reasoning; the double-weight analysis module uses the electrical quantity fault degrees and the switch quantity fault degrees as independent evidence bodies, and respectively calculates the credible interval projection distance weight from the perspective of evidence theory and the information entropy weight from the perspective of probability theory by using an improved Dempster-Shafer evidence theory identification framework introducing the universal set and subsets; the probability fusion module fuses the credible interval projection distance weight and the information entropy weight of each evidence body through a combined weighting method, and uses the fused weights to weight and fuse all evidence bodies to obtain the fusion fault probabilities of each suspected fault component; the fault location module locates the fault section of the DC distribution network according to the fusion fault probabilities and drives the circuit breaker to operate to isolate the fault. Compared with the prior art, the system realizes the fast and accurate location of the fault section of the DC distribution network and drives the circuit breaker to operate to isolate the fault by fusing the high-frequency transient electrical quantity characteristics and the switch quantity action information, and effectively improves the fault tolerance ability of the fault location, ensuring the safe and stable operation of the power grid.

[0079] The above embodiments only represent several preferred embodiments of the present application. The description is relatively specific and detailed, but it should not be construed as a limitation on the scope of the invention patent. It should be noted that for those of ordinary skill in the art of this technology, without departing from the technical principle of the present invention, several improvements and substitutions can still be made, and these improvements and substitutions should also be regarded as the protection scope of the present application. Therefore, the protection scope of the patent of the present application should be subject to the protection scope of the claimed rights.

Claims

1. A method for locating fault sections in a DC distribution network based on multi-source information fusion, characterized in that, Including the following steps: According to the high-frequency transient voltage signal generated instantaneously when a bipolar short-circuit fault occurs in the DC distribution network, reconstruct the sparse vector of high-frequency fault current; Use the normal distribution model to quantitatively analyze the non-linear attenuation trend of the high-frequency current amplitude of each suspected fault component in the sparse vector of high-frequency fault current, and obtain the estimated value of the high-frequency current amplitude; Perform cubic normalization on the estimated values of the high-frequency current amplitudes of each suspected fault component to obtain the electrical quantity fault degree of each suspected fault component; Modify the node state of the relay protection Bayesian network model according to the pre-identified local switch quantity action information, and calculate the switch quantity fault degree of each suspected fault component through reverse inference; Take the electrical quantity fault degree and the switch quantity fault degree as independent evidence bodies, and use the improved Dempster-Shafer evidence theory identification framework introducing the universal set subset to calculate the projection distance weight from the perspective of evidence theory and the information entropy weight from the perspective of probability theory respectively; Fuse the projection distance weight of the credible interval and the information entropy weight of each evidence body through the combined weighting method, and use the fused weight to weighted-fuse all evidence bodies to obtain the fused fault probability of each suspected fault component; Locate the fault section of the DC distribution network according to the fused fault probability, and drive the circuit breaker to act to isolate the fault.

2. The method for fault section location of a DC distribution network based on multi-source information fusion according to claim 1, characterized in that The step of reconstructing the sparse vector of high-frequency fault current according to the high-frequency transient voltage signal generated instantaneously when a bipolar short-circuit fault occurs in the DC distribution network includes: At the instant when a bipolar short-circuit fault occurs in the DC distribution network, construct a node high-frequency impedance matrix according to the node self-impedance data and the mutual impedance data between nodes of the DC distribution network; Synchronously collect the high-frequency transient voltage signals at the fault measurement points within a preset time window before and after the occurrence of the bipolar short-circuit fault through a voltage measurement device; Equivalent the fault bus node to a high-frequency current source, set the current of non-bus fault nodes to zero, and establish a sparse node equation for bus fault according to the node high-frequency impedance matrix and the high-frequency transient voltage signal; Equivalent the fault line to a high-frequency current source of adjacent fault measurement points, set the current between non-adjacent fault measurement points to zero, and establish a sparse node equation for line fault according to the node high-frequency impedance matrix and the high-frequency transient voltage signal; Integrate the sparse node equation for bus fault and the sparse node equation for line fault into an underdetermined sparse node equation set; Convert the underdetermined sparse node equation set into a sparse signal reconstruction problem, and use the Bayesian compressive sensing reconstruction algorithm to iteratively solve to obtain the sparse vector of high-frequency fault current.

3. The method for locating a fault section of a DC distribution network based on multi-source information fusion according to claim 2, wherein: The sparse node equation for bus fault is used to describe the relationship between the voltages and currents of each node in the DC distribution network during bus fault; The sparse node equation for line fault is used to describe the relationship between the voltages and currents of each node in the DC distribution network during line fault.

4. A method for locating a fault section in a DC distribution network based on multi-source information fusion according to claim 1, characterized in that, The step of using the normal distribution model to quantitatively analyze the non-linear attenuation trend of the high-frequency current amplitude of each suspected fault component in the sparse vector of high-frequency fault current and obtaining the estimated value of the high-frequency current amplitude includes: According to the high-frequency current amplitudes of each suspected fault component in the sparse vector of high-frequency fault current, construct multiple groups of independent normal distribution models according to the adjacent relationship of the suspected fault components; Take the maximum value of the high-frequency current amplitude among adjacent suspected faulty components as the peak value of the corresponding normal distribution model, and solve the attenuation parameter of each group of normal distribution models through the maximum likelihood estimation criterion; Based on the attenuation parameter, use the normal distribution model to calculate the estimated value of the high-frequency current amplitude at the midpoint of the line between each pair of adjacent suspected faulty components.

5. The method for locating a fault section in a DC distribution network based on multi-source information fusion according to claim 4, wherein: The attenuation parameter includes the standard deviation for controlling the attenuation rate and the electrical distance difference between two adjacent suspected faulty components under the normal distribution.

6. A fault section location method for a DC distribution network based on multi-source information fusion according to claim 1, characterized in that, The step of cubic normalization of the estimated values of the high-frequency current amplitudes of each suspected faulty component to obtain the electrical quantity fault degree of each suspected faulty component includes: Perform cubic operations on the estimated values of the high-frequency current amplitudes of each suspected faulty component respectively to obtain the cubic amplitudes of the high-frequency currents of each suspected faulty component; Normalize the cubic amplitudes of the high-frequency currents to obtain the electrical quantity fault degrees of each suspected faulty component.

7. A fault section location method for a DC distribution network based on multi-source information fusion according to claim 1, characterized in that, The step of correcting the node state of the relay protection Bayesian network model according to the pre-identified local switch quantity action information and calculating the switch quantity fault degree of each suspected faulty component through reverse inference includes: Set three-level sampling time windows according to the sequential action logic of DC distribution network relay protection, and determine the protection level of the tripped circuit breaker by detecting the high-frequency transient voltage signals within each level of sampling time window, and obtain the local circuit breaker action state of the protection level where the tripped circuit breaker is located; Identify the local switch quantity action information according to the local circuit breaker action state, and use the local switch quantity action information to correct the switch quantity information uploaded by the data acquisition and monitoring system to obtain the corrected switch quantity action information; According to the relay protection action logic of the DC distribution network, construct a relay protection Bayesian network model of the DC distribution network, and use the corrected switch quantity action information to update the corresponding relay protection node state and circuit breaker node state in the relay protection Bayesian network model to obtain a corrected relay protection Bayesian network model; Based on each suspected faulty component determined from the perspective of electrical quantities, use the corrected relay protection Bayesian network model for reverse probability inference to calculate the posterior probability of failure of each suspected faulty component; Perform cubic normalization on the posterior probabilities of failure of each suspected faulty component to obtain the switch quantity fault degrees of each suspected faulty component.

8. A fault section location method for a DC distribution network based on multi-source information fusion according to claim 7, characterized in that: The three-level sampling time windows include the first-level sampling time window, the second-level sampling time window, and the third-level sampling time window; among them, the first-level sampling time window corresponds to the main protection action period, the second-level sampling time window corresponds to the near-backup protection action period, and the third-level sampling time window corresponds to the far-backup protection action period.

9. The method for locating a fault section of a DC distribution network based on multi-source information fusion according to claim 1, wherein The step of using the improved Dempster-Shafer evidence theory recognition framework introducing the universal set and subsets to calculate the projection distance weight of the confidence interval from the perspective of evidence theory and the information entropy weight from the perspective of probability theory includes: Define the confidence interval and the maximum uncertainty interval of each suspected faulty component in the evidence body according to the improved Dempster-Shafer evidence theory recognition framework introducing the universal set and subsets, and represent the confidence interval and the maximum uncertainty interval in the form of interval numbers; Based on the credible intervals and maximum uncertainty intervals of each suspected faulty component represented in the form of interval numbers, calculate the projection distances between the credible intervals and maximum uncertainty intervals of each suspected faulty component in the evidence body; Use the projection distance as an index to measure the uncertainty of the evidence body, and obtain the uncertainty measurement value of the evidence body; According to the uncertainty measurement value of the evidence body and the preset evidence adjustment coefficient, obtain the weight of the projection distance of the credible interval of the evidence body from the perspective of evidence theory; Equalize the probability assignments of the universal set subsets in the evidence body to all suspected faulty components to obtain the component probability distribution, and calculate the information entropy of the component probability distribution; According to the information entropy of the component probability distribution and the preset probability adjustment coefficient, calculate and obtain the information entropy weight of the evidence body from the perspective of probability theory.

10. A DC distribution network fault section location system based on multi-source information fusion, characterized in that, The system includes: A signal reconstruction module for reconstructing a high-frequency fault current sparse vector according to the high-frequency transient voltage signal generated instantaneously when a bipolar short-circuit fault occurs in the DC distribution network; An amplitude estimation module for quantitatively analyzing the non-linear attenuation trend of the high-frequency current amplitudes of each suspected faulty component in the high-frequency fault current sparse vector by using a normal distribution model to obtain high-frequency current amplitude estimation values; An electrical quantity analysis module for performing cubic normalization on the high-frequency current amplitude estimation values of each suspected faulty component to obtain the electrical quantity fault degrees of each suspected faulty component; A switch quantity analysis module for correcting the node states of the relay protection Bayesian network model according to the pre-identified local switch quantity action information, and calculating the switch quantity fault degrees of each suspected faulty component through reverse reasoning; A double-weight analysis module for using the electrical quantity fault degrees and the switch quantity fault degrees as independent evidence bodies, and respectively calculating the weight of the projection distance of the credible interval from the perspective of evidence theory and the information entropy weight from the perspective of probability theory by using an improved Dempster-Shafer evidence theory recognition framework introducing universal set subsets; A probability fusion module for fusing the weight of the projection distance of the credible interval and the information entropy weight of each evidence body by a combined weighting method, and using the fused weights to weightedly fuse all evidence bodies to obtain the fused fault probabilities of each suspected faulty component; A fault location module for locating the fault section of the DC distribution network according to the fused fault probabilities and driving the circuit breaker to act to isolate the fault.

Images