A distribution station monitoring method, device and storage medium

Abstract

The present invention discloses a method, device and storage medium for monitoring a power distribution station. The method includes obtaining the distribution of electromagnetic interference intensity and real-time status data of the line in the power distribution station, determining the sensor selection scheme in combination with the equipment working condition; calculating the fault propagation path through the arrival time difference of the fault traveling wave, the line model parameters and the wave speed difference, determining the position of the fault point and generating the positioning result; associating the sensor selection with the fault location, judging whether the fault point is in the layout area, and preliminarily judging the fault type; generating the candidate recovery path scheme based on the real-time status data and the fault location result; integrating the fault traveling wave characteristics and the line status parameters of the recovery path, establishing a comprehensive evaluation model, and determining the optimal recovery path scheme with the power supply recovery efficiency and the load loss rate as the optimization target. The method has the following effects: realizing the rapid and accurate positioning of the distribution network fault and the intelligent power supply recovery, and improving the reliability and efficiency of the distribution network operation.

Description

Technical Field

[0001] The present invention relates to the technical field of power system automation, and in particular to a distribution station monitoring method, device and storage medium. Background Art

[0002] At present, the automation system of distribution station faces the dual challenges of complex electromagnetic environment and changeable working conditions in actual operation. The selection and layout of sensors need to ensure the reliability of data collection under strong interference conditions. At the same time, the real-time transmission and processing of massive monitoring data puts forward strict requirements on communication protocols and computing architecture.

[0003] In one existing technology, distributed optical fiber sensing and anti-interference communication technology are widely used in environmental parameter collection, edge computing nodes and cloud computing hybrid architecture support data hierarchical processing, and the traveling wave feature extraction method based on wavelet transform and neural network has been applied in engineering. At the system architecture level, the IEC 61850 standard protocol is used to realize device interconnection, and heuristic methods such as genetic algorithms are used to optimize power supply restoration.

[0004] In the prior art, there is a problem that the positioning of distribution network faults is not accurate enough, resulting in low reliability and efficiency of distribution network operation. Summary of the invention

[0005] The present invention provides a distribution station monitoring method, device and storage medium to achieve rapid and accurate positioning of distribution network faults and intelligent power supply restoration, and to improve the reliability and efficiency of distribution network operation.

[0006] In a first aspect, in order to solve the above technical problems, the present invention provides a method for monitoring a power distribution station, comprising:

[0007] Obtain the electromagnetic interference intensity distribution data in the power distribution station and the real-time status data of each line;

[0008] Determine the sensor selection scheme according to the electromagnetic interference intensity distribution data and in combination with the equipment operating condition parameters;

[0009] Obtain the arrival time difference data of the fault traveling wave, combine the preset line model parameters and wave speed difference characteristics, calculate the fault traveling wave propagation path, determine the fault point location, and generate the fault location result;

[0010] Associating the sensor selection scheme with the fault location result, determining whether the fault point is located in the sensor deployment area, and obtaining a preliminary determination result of the fault type;

[0011] Generate a candidate recovery path solution based on the real-time status data and in combination with the fault location result;

[0012] The fault traveling wave propagation characteristic parameters in the fault location result and the line state parameters in the restoration path candidate solution are integrated to establish a comprehensive evaluation model, and the optimal restoration path solution is determined with power supply restoration efficiency and load loss rate as optimization targets.

[0013] In an optional implementation, determining a sensor selection scheme based on the electromagnetic interference intensity distribution data and in combination with equipment operating condition parameters includes:

[0014] According to the electromagnetic interference intensity distribution data, measuring the intensity value of each interference source and generating an electromagnetic interference intensity distribution diagram;

[0015] Associating the electromagnetic interference intensity distribution map with the equipment operation condition database to determine the interference intensity value range;

[0016] According to the interference intensity value range, determine the sensor range value and select a sensor model that meets the range requirements;

[0017] According to the characteristics of different interference sources, it is judged whether the sensitivity of the sensor model meets the requirements, so as to determine the sensor selection scheme.

[0018] In an optional implementation, the acquisition of arrival time difference data of the fault traveling wave, combining preset line model parameters and wave speed difference characteristics, calculating the fault traveling wave propagation path, determining the fault point location, and generating the fault location result includes:

[0019] Extracting the time difference characteristic value in the traveling wave number according to the arrival time difference data of the fault traveling wave;

[0020] According to the time difference characteristic value, extracting the line model characteristic from the line model parameter, matching it with the wave speed difference characteristic, and obtaining the wave speed difference characteristic;

[0021] According to the wave speed difference characteristics, the fault traveling wave propagation path is calculated to obtain path length data;

[0022] Determine the fault point location data according to the path length data and in combination with the line model parameters;

[0023] Generate a fault location result based on the fault point location data and output the location value information;

[0024] If there is a deviation in the fault location result, recalculate the wave speed difference characteristic and adjust the propagation path;

[0025] Based on the adjusted propagation path, the fault point location is re-determined to generate the final fault location result.

[0026] In an optional implementation, associating the sensor selection scheme with the fault location result to determine whether the fault point is located in the sensor deployment area to obtain a preliminary determination result of the fault type includes:

[0027] Obtain the fault location output by the fault location system, and perform spatial topological relationship calculations through the geographic information system in combination with the spatial information of the pre-established sensor group to determine the spatial inclusion relationship between the fault location and the sensor deployment area;

[0028] If it is determined that the fault point is included in the sensor deployment area, all sensors identified by the sensor group are retrieved from the database according to the sensor group associated with the sensor deployment area;

[0029] According to all the sensors, the operating parameter data of the sensors within a period of time before and after the fault occurs are retrieved to form time series data, and a time series length threshold is set;

[0030] According to the time series data, the correlation coefficients between the operating parameters of each sensor are calculated and a correlation coefficient matrix is ​​constructed, and the correlation coefficients lower than the time series length threshold are set to zero;

[0031] According to the correlation coefficient matrix, the sensors are divided into different clusters by a community discovery algorithm, each cluster represents a closely related sensor subset, so as to determine the central sensor of each sensor subset;

[0032] For the sensor layout area of ​​the fault point, if it is determined that there are multiple sensor clusters, the central sensor closest to the fault point is selected, and its operating parameters are extracted for feature engineering processing;

[0033] Input the data processed by feature engineering into the pre-trained fault diagnosis model to obtain the preliminary judgment result of the fault type;

[0034] The training process of the fault diagnosis model:

[0035] Acquiring historical monitoring data of the sensor cluster, including vibration waveform spectrum, infrared thermal image spectrum and partial discharge soundprint signal;

[0036] Performing multimodal feature extraction and normalization processing on the historical monitoring data to generate a standardized feature matrix;

[0037] The standardized feature matrix is ​​used as input data, an initial fault diagnosis model is constructed based on a support vector machine algorithm, and the initial fault diagnosis model is trained according to a fused feature tensor and a fault type label to obtain a fault diagnosis model.

[0038] In an optional implementation, generating a candidate recovery path solution according to the real-time status data and in combination with the fault location result includes:

[0039] According to the real-time status data, the real-time active power and real-time reactive power of each line are calculated, and the switch status information of each line is obtained, so as to determine the on-off status of the line;

[0040] The real-time active power is calculated by the following formula:

[0041]

[0042] The real-time reactive power is calculated by the following formula:

[0043]

[0044] According to the pre-established load classification rules and the load types connected to each line, each line is evaluated and scored for load importance;

[0045] The load importance evaluation score of each line is calculated using the following formula:

[0046]

[0047] According to the load importance evaluation score, the load importance level is divided into a plurality of different levels, and the importance of each line is judged, thereby determining the load importance level of each line;

[0048] The specific line section where the fault occurs is determined through the fault location result, and the accurate position of the faulty line is obtained in combination with the switch status information;

[0049] According to the on / off state and the load importance level, different types of load transfer strategies are selected. If the faulty line is a non-important load line, load transfer is performed according to the actual situation;

[0050] According to the topological structure of the faulty line and the remaining capacity information of the non-faulty lines, all feasible paths from the power supply point to the power supply area of ​​the faulty line are calculated by using the shortest path algorithm in graph theory;

[0051] Comprehensively score all the feasible paths, and sort them from high to low according to the scores to generate candidate recovery path solutions;

[0052] The following formula is used to comprehensively score all the feasible paths:

[0053]

[0054] in, Indicates real-time active power, represents the total number of samples, and Respectively represent The sampled values ​​of voltage and current in samples, represents the total number of load type factors, Indicates the real-time reactive power, It represents the phase difference between current and voltage. represents the load importance assessment score, Indicates The weight of each load type factor, Indicates that the line is The scores under each load type factor are , and represents the factors of each load type, and , Indicates the comprehensive rating. Indicates the line operation status score, Indicates the line length.

[0055] In an optional implementation, the method of calculating all feasible paths from the power supply point to the power supply area of ​​the faulty line based on the topological structure of the faulty line and the remaining capacity information of the non-faulty lines by using the shortest path algorithm in graph theory includes:

[0056] Initialize the parameters of the shortest path algorithm and set the initial distance array as the first distance feature;

[0057] Iterate according to the shortest path algorithm, obtain the second distance feature after the iteration, and record the number of iterations;

[0058] Determine whether the second distance feature is better than the current distance feature, if so, update the second distance feature to the current distance feature; if not, discard the second distance feature;

[0059] Continue with the next round of iteration, and when the number of iterations is greater than or equal to the preset number of iterations, output the current distance feature as the optimal distance feature;

[0060] The optimal distance feature is used as a feasible path.

[0061] In an optional implementation, the fault traveling wave propagation characteristic parameters in the fault location result and the line state parameters in the restoration path candidate solution are integrated to establish a comprehensive evaluation model, and the power supply restoration efficiency and load loss rate are used as optimization targets to determine the optimal restoration path solution, including:

[0062] Acquire the traveling wave propagation characteristic parameters in the fault location result and the line state parameters in the restoration path candidate solution;

[0063] Using data fusion technology, the propagation characteristic parameters and the line state parameters are fused at the feature level to generate a comprehensive feature vector;

[0064] A comprehensive evaluation model based on random forest takes the comprehensive feature vector as input and power supply restoration efficiency and load loss rate as output;

[0065] A particle swarm optimization algorithm is used to search for the optimal solution among the candidate restoration path solutions with maximizing power restoration efficiency and minimizing load loss rate as objective functions;

[0066] If the objective function reaches a preset threshold, an optimal recovery path solution is output.

[0067] In a second aspect, the present invention provides a power distribution station monitoring device, comprising:

[0068] Data acquisition module, used to obtain electromagnetic interference intensity distribution data in the power distribution station and real-time status data of each line;

[0069] A sensor selection module, used to determine a sensor selection scheme according to the electromagnetic interference intensity distribution data and in combination with equipment operating condition parameters;

[0070] The fault location result generation module is used to obtain the arrival time difference data of the fault traveling wave, calculate the propagation path of the fault traveling wave by combining the line model parameters and the wave speed difference characteristics, determine the fault point location, and generate the fault location result;

[0071] A fault type judgment module is used to associate the sensor selection scheme with the fault location result, judge whether the fault point is located in the sensor deployment area, and obtain a preliminary judgment result of the fault type;

[0072] A recovery path candidate solution generating module, used to generate a recovery path candidate solution according to the real-time status data and in combination with the fault location result;

[0073] The module for determining the optimal restoration path is used to integrate the fault wave propagation characteristic parameters in the fault location result and the line state parameters in the restoration path candidate solution, establish a comprehensive evaluation model, and determine the optimal restoration path solution with power supply restoration efficiency and load loss rate as optimization targets.

[0074] In a third aspect, the present invention further provides an electronic device, comprising a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein the processor implements any one of the above-mentioned distribution station monitoring methods when executing the computer program.

[0075] In a fourth aspect, the present invention further provides a computer-readable storage medium, which includes a stored computer program, wherein when the computer program is running, the device where the computer-readable storage medium is located is controlled to execute any one of the above-mentioned distribution station monitoring methods.

[0076] Compared with the prior art, the present invention has the following beneficial effects:

[0077] The present invention discloses a method for monitoring a distribution substation, including obtaining the distribution of electromagnetic interference intensity and real-time status data of the line in the distribution substation, determining the sensor selection scheme in combination with the equipment working condition; calculating the fault propagation path through the arrival time difference of the fault traveling wave, the line model parameters and the wave speed difference, determining the position of the fault point and generating the positioning result; associating the sensor selection with the fault location, judging whether the fault point is in the layout area, and preliminarily judging the fault type; generating the candidate recovery path scheme based on the real-time status data and the fault location result; integrating the fault traveling wave characteristics and the line status parameters of the recovery path, establishing a comprehensive evaluation model, taking the power supply recovery efficiency and the load loss rate as the optimization target, and determining the optimal recovery path scheme. The method has the following effects: realizing the rapid and accurate positioning of the distribution network fault and the intelligent power supply recovery, and improving the reliability and efficiency of the distribution network operation. BRIEF DESCRIPTION OF THE DRAWINGS

[0078] Figure 1 It is a schematic flow chart of a method for monitoring a power distribution station provided by a first embodiment of the present invention;

[0079] Figure 2 It is a schematic diagram of the structure of a distribution station monitoring device provided in the second embodiment of the present invention. DETAILED DESCRIPTION

[0080] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0081] Reference Figure 1 The first embodiment of the present invention provides a method for monitoring a power distribution station, comprising the following steps:

[0082] S11, obtaining electromagnetic interference intensity distribution data in the power distribution station and real-time status data of each line;

[0083] S12, determining a sensor selection scheme according to the electromagnetic interference intensity distribution data and in combination with equipment operating condition parameters;

[0084] S13, obtaining arrival time difference data of the fault traveling wave, combining preset line model parameters with wave speed difference characteristics, calculating the propagation path of the fault traveling wave, determining the fault point location, and generating a fault location result;

[0085] S14, associating the sensor selection scheme with the fault location result, determining whether the fault point is located in the sensor deployment area, and obtaining a preliminary determination result of the fault type;

[0086] S15, generating a candidate recovery path solution according to the real-time status data and in combination with the fault location result;

[0087] S16, integrating the fault traveling wave propagation characteristic parameters in the fault location result and the line state parameters in the candidate restoration path solutions, establishing a comprehensive evaluation model, taking power supply restoration efficiency and load loss rate as optimization targets, and determining the optimal restoration path solution.

[0088] In step S11, the electromagnetic interference intensity distribution data in the power distribution station and the real-time status data of each line are obtained.

[0089] It should be noted that the electromagnetic interference intensity distribution data in the power distribution station includes electromagnetic interference intensity values ​​at different locations and frequencies. These data can help analyze interference sources, evaluate the safety of the equipment operating environment, and optimize electromagnetic compatibility design. To obtain electromagnetic interference intensity distribution data, it is necessary to select appropriate equipment according to the frequency range of the measurement, such as the Hydelogie SAM60MK2 spectrum detection equipment and its supporting omnidirectional antenna, which can sense electromagnetic spectrum signals in the frequency range of 30MHz to 6.3GHz. Then, measurement points are arranged in the power distribution station according to a certain grid spacing, covering key areas such as equipment areas, cable channels, and control rooms; the electromagnetic field strength at different frequencies is recorded at each measurement point, including electric field strength (V / m) and magnetic field strength (A / m); the signal is converted to the frequency domain through spectrum analysis technology to analyze the interference intensity of different frequency components.

[0090] The real-time status data of each line in the distribution station includes electrical parameters (such as voltage, current, power factor, etc.), equipment temperature, switch status, and environmental parameters (such as temperature and humidity, harmful gas concentration, etc.). These data are collected through various sensors and smart meters deployed in the distribution station. For example, smart power meters are used to collect electrical parameters, and temperature sensors are used to monitor equipment temperature. The collected data is processed and transmitted through industrial Internet of Things gateways or edge computing devices, supporting multiple communication protocols (such as 4G, 5G, etc.), and finally uploaded to the background data center or cloud platform. On the monitoring platform, these data are displayed in real time in a visual form, supporting fault warning, data analysis and remote management, thereby realizing efficient monitoring and intelligent management of distribution stations.

[0091] In step S12, a sensor selection scheme is determined according to the electromagnetic interference intensity distribution data and in combination with equipment operating condition parameters.

[0092] According to the electromagnetic interference intensity distribution data, measuring the intensity value of each interference source and generating an electromagnetic interference intensity distribution diagram;

[0093] Associating the electromagnetic interference intensity distribution map with the equipment operation condition database to determine the interference intensity value range;

[0094] According to the interference intensity value range, determine the sensor range value and select a sensor model that meets the range requirements;

[0095] According to the characteristics of different interference sources, it is judged whether the sensitivity of the sensor model meets the requirements, so as to determine the sensor selection scheme.

[0096] It should be noted that the electromagnetic interference intensity distribution map is obtained through measurement or simulation calculation, and is used to visually display the electromagnetic interference intensity at different locations in space in a two-dimensional or three-dimensional visual graph. The steps to generate the electromagnetic interference intensity distribution map are as follows: First, determine the radiation power and position of the interference source, and calculate the intensity of the interference source through experimental measurement or according to the equipment specifications; then, arrange multiple measurement points in the area to be measured, and use electromagnetic field probes or near-field probe arrays to collect electromagnetic field intensity data at each point; then, import the collected data into data analysis software or simulation tools (such as MATLAB, ANSYS, etc.), and use electromagnetic field theory, such as Maxwell's equations, for numerical calculations; finally, generate a distribution map based on the calculation results, and represent electromagnetic interference areas of different intensities in the form of color or vector field.

[0097] To determine the sensor range value based on the interference intensity value range, it is necessary to first clarify the minimum and maximum values ​​of the interference intensity, which define the range that the sensor needs to measure. The sensor range value should at least cover this interference intensity range to ensure that the sensor can accurately detect the minimum and maximum interference intensities. At the same time, in order to improve measurement accuracy and avoid sensor overload, select a range value slightly larger than the interference intensity range. For example, if the interference intensity range is 0 to 100 units, the sensor range value can be set to 0 to 120 units, which can not only ensure the integrity of the measurement range, but also leave a certain margin for emergencies.

[0098] To determine whether the sensitivity of a sensor model meets the requirements based on the characteristics of different interference sources, the following steps can be taken: clarify the characteristics of the interference source, including parameters such as the intensity, frequency, and direction of the interference, as well as the minimum signal change that the sensor needs to detect; then consult the manual of the sensor model to obtain its theoretical value of sensitivity and compare it with actual needs; then build a test platform, simulate actual application scenarios, test the sensor using a known standard signal source, and record its output signal changes; calculate the sensitivity formula , to verify whether the actual sensitivity of the sensor meets the requirements; finally, consider the impact of environmental factors such as temperature, humidity, and electromagnetic interference on sensitivity to ensure that the sensor can still meet the requirements in the actual working environment. If the test results show that the sensitivity of the sensor is within the allowable error range, it can be considered to meet the requirements.

[0099] In step S13, the arrival time difference data of the fault traveling wave is obtained, and the propagation path of the fault traveling wave is calculated in combination with the preset line model parameters and wave speed difference characteristics, the fault point position is determined, and the fault location result is generated.

[0100] Extracting the time difference characteristic value in the traveling wave number according to the arrival time difference data of the fault traveling wave;

[0101] According to the time difference characteristic value, extracting the line model characteristic from the line model parameter, matching it with the wave speed difference characteristic, and obtaining the wave speed difference characteristic;

[0102] According to the wave speed difference characteristics, the fault traveling wave propagation path is calculated to obtain path length data;

[0103] Determine the fault point location data according to the path length data and in combination with the line model parameters;

[0104] Generate a fault location result based on the fault point location data and output the location value information;

[0105] If there is a deviation in the fault location result, recalculate the wave speed difference characteristic and adjust the propagation path;

[0106] Based on the adjusted propagation path, the fault point location is re-determined to generate the final fault location result.

[0107] It should be noted that the arrival time difference data of the fault traveling wave refers to the time difference between the traveling wave signal generated when a fault occurs and the protection measuring points at both ends of the line in the power system. These data can be obtained by installing high-precision traveling wave sensors or voltage and current transformers at both ends of the transmission line. When a fault occurs, the traveling wave signal will propagate along the line, and the sensor will record the timestamp of the arrival of the traveling wave. By calculating the time difference recorded by the sensors at both ends and combining the traveling wave propagation speed, the location of the fault point can be inferred. For example, the wavelet transform modulus maximum method can be used to accurately calibrate the arrival time of the traveling wave head, and then calculate the time difference data.

[0108] According to the time difference characteristic value, the process of extracting the line model characteristics from the line model parameters and matching them with the wave speed difference characteristics to obtain the wave speed difference characteristics can be achieved through the following steps: using the distributed parameters of the line, such as resistance, inductance, capacitance, etc., to establish a line model, converting the three-phase voltage and current signals into mode components through phase mode transformation, such as Karen Bell transformation, and extracting line mode and zero mode components. The wave speed of the line mode component is close to the speed of light and is relatively stable, while the wave speed of the zero mode component is relatively small and is greatly affected by the line parameters. Then, based on methods such as wavelet transform, the characteristics of the traveling wave signal, such as waveform mutation points and frequency component changes, are extracted. By analyzing the propagation characteristics of different mode components, their wave speed differences are calculated, and combined with the arrival time difference of the fault traveling wave signal, the wave speed difference characteristics are matched.

[0109] The method for calculating the fault wave propagation path and obtaining the path length data based on the wave velocity difference characteristics is as follows: extract the characteristics of the fault wave signal through signal processing methods such as wavelet transform, and determine the time difference between the wave reaching the protection measuring points at both ends of the line. Then, combined with the wave velocity difference characteristics of the line, such as the difference between the line mode and zero mode wave velocities, the time difference formula of the fault wave propagation is used to calculate the distance from the fault point to both ends of the line. For example, for the line mode wave, the time difference formula is ,in is the distance from the fault point to the beginning of the line, is the line length, is the propagation speed of the traveling wave. Through optimization algorithms such as particle swarm optimization, the wave speed difference and time error are further corrected, and finally the propagation path of the fault traveling wave and the accurate path length data are obtained.

[0110] In fault location, to determine whether there is a deviation in the location result, the following method can be used: compare the location result with the actual measurement value or known reference value. If the difference exceeds the allowable error range, it is considered that there is a deviation. For example, in a complex transmission network, if the error between the fault location result and the actual fault location exceeds the set threshold, such as the threshold is set to 100 meters, it is determined to be a deviation. In addition, it can also be judged by analyzing the difference in the arrival time of the traveling wave head: if the calculated result of the traveling wave arrival time is inconsistent with the actual measurement value, it also indicates that there is a deviation in the location result. Once the deviation is confirmed, the wave speed difference characteristics need to be recalculated and the propagation path needs to be adjusted to improve the positioning accuracy.

[0111] In step S14, the sensor selection scheme is associated with the fault location result to determine whether the fault point is located in the sensor deployment area, thereby obtaining a preliminary determination result of the fault type.

[0112] Obtain the fault location output by the fault location system, and perform spatial topological relationship calculations through the geographic information system in combination with the spatial information of the pre-established sensor group to determine the spatial inclusion relationship between the fault location and the sensor deployment area;

[0113] If it is determined that the fault point is included in the sensor deployment area, all sensors identified by the sensor group are retrieved from the database according to the sensor group associated with the sensor deployment area;

[0114] According to all the sensors, the operating parameter data of the sensors within a period of time before and after the fault occurs are retrieved to form time series data, and a time series length threshold is set;

[0115] According to the time series data, the correlation coefficients between the operating parameters of each sensor are calculated and a correlation coefficient matrix is ​​constructed, and the correlation coefficients lower than the time series length threshold are set to zero;

[0116] According to the correlation coefficient matrix, the sensors are divided into different clusters by a community discovery algorithm, each cluster represents a closely related sensor subset, so as to determine the central sensor of each sensor subset;

[0117] For the sensor layout area of ​​the fault point, if it is determined that there are multiple sensor clusters, the central sensor closest to the fault point is selected, and its operating parameters are extracted for feature engineering processing;

[0118] Input the data processed by feature engineering into the pre-trained fault diagnosis model to obtain the preliminary judgment result of the fault type;

[0119] The training process of the fault diagnosis model:

[0120] Acquiring historical monitoring data of the sensor cluster, including vibration waveform spectrum, infrared thermal image spectrum and partial discharge soundprint signal;

[0121] Performing multimodal feature extraction and normalization processing on the historical monitoring data to generate a standardized feature matrix;

[0122] The standardized feature matrix is ​​used as input data, an initial fault diagnosis model is constructed based on a support vector machine algorithm, and the initial fault diagnosis model is trained according to a fused feature tensor and a fault type label to obtain a fault diagnosis model.

[0123] It should be noted that spatial topological relationship operations are calculations and analyses based on the logical structural relationships between spatial objects, such as points, lines, and surfaces: such as adjacency, association, and inclusion. Its significance lies in that through topological relationships, the relative position relationship between spatial entities can be determined without relying on specific coordinates or distances. This operation is stable and does not change with map projections. It can efficiently support the query, analysis, and reconstruction of spatial data and geographic entities. For example, in a geographic information system (GIS), topological relationships can be used to analyze the impact of rivers on surrounding areas, determine the best path, or reconstruct geographic entities.

[0124] To implement the operation of retrieving all sensors in the sensor deployment area from the database based on whether the fault point is included in the sensor deployment area, first, it is necessary to clarify the scope of the sensor deployment area, associate the location information of each sensor with the area, and store it in the database. When it is determined that the fault point is located in a certain sensor deployment area, the detailed information of all sensors in the area can be retrieved by querying the database according to the sensor group identifier. The specific implementation can be achieved by writing SQL query statements, combining the sensor's geographical location information and regional boundary conditions, and quickly locating and retrieving the associated sensor group.

[0125] According to all the sensors, the operating parameter data of the sensors in a period of time before and after the time point of the fault is retrieved to form time series data, and the time series length threshold is set. This can be achieved by the following method: determine the specific time point of the fault, and then query the sensor operating parameter data related to the time point before and after the time point from the database. These data are indexed by timestamps and record the real-time operating status of the sensor. Then, these data are arranged in chronological order to form time series data. In order to ensure the effectiveness of the analysis, it is necessary to set a time series length threshold. For example, a suitable window size can be selected based on domain knowledge or data characteristics. The sliding window method can be used to divide the time series at a fixed time interval, such as 15 minutes as a step size, and the data in each window is used as an independent time series. Finally, these time series data are analyzed according to needs, such as anomaly detection or fault diagnosis, and the validity of the data is judged according to the set threshold.

[0126] Community discovery is a method for identifying closely connected subgroups or communities in a network. Its goal is to divide the nodes in the network into different communities, so that the connections between nodes in a community are denser, while the connections between communities are relatively sparse. Through community discovery, we can better understand the structure and function of the network and reveal the behavioral patterns of individuals in the network and their relationships.

[0127] In step S15, a candidate restoration path solution is generated according to the real-time status data and in combination with the fault location result.

[0128] According to the real-time status data, the real-time active power and real-time reactive power of each line are calculated, and the switch status information of each line is obtained, so as to determine the on-off status of the line;

[0129] The real-time active power is calculated by the following formula:

[0130]

[0131] The real-time reactive power is calculated by the following formula:

[0132]

[0133] According to the pre-established load classification rules and the load types connected to each line, each line is evaluated and scored for load importance;

[0134] The load importance evaluation score of each line is calculated using the following formula:

[0135]

[0136] According to the load importance evaluation score, the load importance level is divided into a plurality of different levels, and the importance of each line is judged, thereby determining the load importance level of each line;

[0137] The specific line section where the fault occurs is determined through the fault location result, and the accurate position of the faulty line is obtained in combination with the switch status information;

[0138] According to the on / off state and the load importance level, different types of load transfer strategies are selected. If the faulty line is a non-important load line, load transfer is performed according to the actual situation;

[0139] According to the topological structure of the faulty line and the remaining capacity information of the non-faulty lines, all feasible paths from the power supply point to the power supply area of ​​the faulty line are calculated by using the shortest path algorithm in graph theory;

[0140] Comprehensively score all the feasible paths, and sort them from high to low according to the scores to generate candidate recovery path solutions;

[0141] The following formula is used to comprehensively score all the feasible paths:

[0142]

[0143] in, Indicates real-time active power, represents the total number of samples, and Respectively represent The sampled values ​​of voltage and current in samples, represents the total number of load type factors, Indicates the real-time reactive power, It represents the phase difference between current and voltage. represents the load importance assessment score, Indicates The weight of each load type factor, Indicates that the line is The scores under each load type factor are , and represents the factors of each load type, and , Indicates the comprehensive rating. Indicates the line operation status score, Indicates the line length.

[0144] It should be noted that Indicates real-time active power, reflecting the energy actually transmitted in the power system; It represents the total number of samples and is used to determine the frequency and precision of data collection; and Respectively represent The sampling values ​​of voltage and current in each sample are the basic electrical parameters in the power system; Indicates real-time reactive power, reflecting the energy used to establish magnetic and electric fields in the power system; Indicates the phase difference between current and voltage and is used to evaluate the power factor of the power system; It represents the load importance assessment score, which is used to measure the importance of the line in the power system; Indicates The weight of each load type factor is used to reflect the impact of different load types on the line; Indicates that the line is The score under each load type factor is used to evaluate the performance of the line under a specific load type; , and Represents the factors of each load type, and their sum is 1, which is used to describe the proportion of different load types in the total load; Represents a comprehensive score, which is used to comprehensively evaluate the performance of the line; Indicates the line operation status score, which is used to evaluate the line operation status; Indicates line length, which is used to reflect the physical length of the line. These parameters together constitute the index system for line performance evaluation in power systems, which helps to fully understand and optimize the operating status of power systems.

[0145] According to the load importance assessment score, the load importance level is divided into multiple different levels, and the importance of each line is judged to determine the load importance level of each line. The implementation method is as follows: According to the load's requirements for power supply reliability and the degree of loss or impact caused by power outage, the load is divided into primary load, secondary load and tertiary load. Primary load includes loads that cause personal injury, major losses or major impacts due to power outage; secondary load includes loads that cause major losses or impacts due to power outage; and tertiary load is power load that does not belong to primary and secondary levels. For each line, combined with the power load characteristics in its power supply area, such as equipment type, site importance, etc., its load level is determined by evaluating its importance level score.

[0146] The specific line section where the fault occurred is determined by the fault location results, and the accurate location of the faulty line is obtained in combination with the switch status information. This can be achieved through the following steps: Use fault location technology, such as a traveling wave-based fault location method, to determine the approximate section of the fault. Combined with switch status information, such as the action status and current data of the automated switch, the fault range is further narrowed. For example, if the switch status in a certain section shows an abnormality, such as switch action or current mutation, it can be determined that the fault occurred in the line section where the switch is located. Finally, by analyzing the timing relationship of the switch status information, such as the sequence of switch actions and the propagation time difference of the fault traveling wave, the location of the fault point is accurately determined. This method combines fault location technology and switch status information, can quickly and accurately determine the fault location, and improve fault handling efficiency.

[0147] The topological structure of the fault line refers to the connection relationship and organizational form between the fault line and its connected equipment in the power system, such as switches, transformers, busbars, etc. By abstracting the equipment in the power system as nodes and the lines as edges to form a topological graph, the electrical connection relationship between the fault line and other equipment can be clearly represented. The topological structure can help quickly determine the specific section where the fault occurs, and isolate the fault area through switch status information to reduce the scope of power outage; by analyzing the topological structure, the fault recovery path can be reasonably arranged, the power supply recovery efficiency can be improved, and the impact of the fault on the system stability can be reduced, thereby improving the overall power supply reliability.

[0148] Furthermore, according to the topological structure of the faulty line, combined with the remaining capacity information of the non-faulty lines, all feasible paths from the power supply point to the power supply area of ​​the faulty line are calculated by the shortest path algorithm in graph theory, including:

[0149] Initialize the parameters of the shortest path algorithm and set the initial distance array as the first distance feature;

[0150] Iterate according to the shortest path algorithm, obtain the second distance feature after the iteration, and record the number of iterations;

[0151] Determine whether the second distance feature is better than the current distance feature, if so, update the second distance feature to the current distance feature; if not, discard the second distance feature;

[0152] Continue with the next round of iteration, and when the number of iterations is greater than or equal to the preset number of iterations, output the current distance feature as the optimal distance feature;

[0153] The optimal distance feature is used as a feasible path.

[0154] It should be noted that the shortest path algorithm can use the Dijkstra algorithm, which is a classic algorithm for solving the shortest path from a single source and is applicable to directed or undirected graphs with weights. It maintains a priority queue, gradually expands outward from the starting point, and dynamically updates the shortest path and distance to each vertex. The core idea of ​​the algorithm is: each time, select the vertex closest to the starting point from the unprocessed vertices, mark it as processed, and update the shortest path estimate of its adjacent vertices. This process is repeated until all vertices are processed or the target vertex is processed.

[0155] In step S16, the fault traveling wave propagation characteristic parameters in the fault location result and the line state parameters in the restoration path candidate solution are integrated to establish a comprehensive evaluation model, and the optimal restoration path solution is determined with power supply restoration efficiency and load loss rate as optimization targets.

[0156] Acquire the traveling wave propagation characteristic parameters in the fault location result and the line state parameters in the restoration path candidate solution;

[0157] Using data fusion technology, the propagation characteristic parameters and the line state parameters are fused at the feature level to generate a comprehensive characteristic vector;

[0158] A comprehensive evaluation model based on random forest takes the comprehensive feature vector as input and power supply restoration efficiency and load loss rate as output;

[0159] A particle swarm optimization algorithm is used to search for the optimal solution among the candidate restoration path solutions with maximizing power restoration efficiency and minimizing load loss rate as objective functions;

[0160] If the objective function reaches a preset threshold, an optimal recovery path solution is output.

[0161] It should be noted that to obtain the characteristic parameters of traveling wave propagation in the fault location results and the line state parameters in the candidate restoration path solutions, the following methods can be used: Use the simulation platform of fault traveling waves combined with the distributed parameter model of the transmission line to simulate the propagation process of fault traveling waves. By detecting the arrival time, amplitude change, reflection and refraction of the traveling waves, the characteristic parameters of traveling wave propagation, such as the time of fault occurrence, traveling wave propagation speed and reflection coefficient, are extracted; based on the operation data of the power system, such as node voltage, current, power, etc., combined with data mining and machine learning technology, characteristic parameters related to line status are extracted. For example, the data is reduced in dimension using methods such as principal component analysis (PCA) or wavelet transform to extract line state parameters, such as voltage change trend, current mutation, etc.

[0162] Feature-level fusion is a data fusion technology that fuses feature vectors from different data sources after feature extraction and before decision making to generate a comprehensive feature vector. Specific implementation methods include feature concatenation, feature weighting, feature selection, etc. Feature concatenation is to directly connect feature vectors from different sources to form a longer vector; feature weighting is to fuse different feature vectors together by assigning them different weights; feature selection is to select several feature vectors for fusion according to certain criteria. The purpose of these methods is to integrate the information of multiple features, and by comprehensively considering the advantages of different features, extract more discriminative and expressive features, thereby improving the performance of the model.

[0163] It is worth noting that the comprehensive evaluation model is an evaluation framework that integrates multiple evaluation indicators into an overall framework, which is used to conduct a comprehensive, systematic and objective evaluation of complex systems or multiple alternative solutions. In the above scenario, the comprehensive evaluation model obtains the characteristic parameters of traveling wave propagation in the fault location results and the line state parameters in the candidate recovery path solutions, and uses data fusion technology to fuse these parameters at the feature level to generate a comprehensive feature vector. Subsequently, the comprehensive evaluation model based on random forest takes the power supply restoration efficiency and load loss rate as the optimization target, takes the comprehensive feature vector as input, and searches for the optimal solution in the candidate solutions through the particle swarm optimization algorithm. When the objective function reaches the preset threshold, the optimal recovery path solution is output.

[0164] The comprehensive evaluation model based on random forest can be implemented through the following steps: collect data related to power supply restoration, including fault characteristics, line status, equipment parameters, etc., and construct a comprehensive feature vector as model input; at the same time, collect historical data on power supply restoration efficiency and load loss rate as model output. Then, divide the data set into training set and test set, use the random forest algorithm to train on the training set, and optimize the model performance by adjusting model parameters such as the number of trees and maximum depth. Finally, use the trained model to predict the test set, evaluate the accuracy and generalization ability of the model, and thus achieve a comprehensive evaluation of power supply restoration efficiency and load loss rate.

[0165] Among them, the preset threshold refers to the set standard used to determine whether the objective function has reached the optimal solution during the optimization process. Specifically, the preset threshold is a quantitative standard for maximizing the power restoration efficiency and minimizing the load loss rate, which is used to measure whether the optimization algorithm, such as the particle swarm optimization algorithm, has found a sufficiently good recovery path solution. For example, in terms of power restoration efficiency, the preset threshold can be an efficiency value close to 100%, such as 95%, indicating that the recovery path can quickly restore most of the power supply; and in terms of load loss rate, the preset threshold can be a lower percentage, such as 5%, indicating that the recovery path can minimize the impact of power outages on users. When the power restoration efficiency calculated by the optimization algorithm reaches or exceeds the efficiency threshold, and the load loss rate is lower than the loss rate threshold, it can be considered that the optimal recovery path solution has been found.

[0166] The particle swarm optimization algorithm is used to maximize the power restoration efficiency and minimize the load loss rate as the objective function. The implementation method of searching for the optimal solution in the candidate restoration path scheme is as follows: Define the objective function: Take the power restoration efficiency as the maximization target and the load loss rate as the minimization target, and construct a multi-objective optimization problem. Then initialize the particle swarm, each particle represents a restoration path, and its position and speed represent the path state and adjustment direction respectively. During the iteration process, calculate the fitness value of each particle, update the individual optimal solution and the global optimal solution, and adjust the particle state according to the speed and position update formula. If the objective function reaches the preset threshold (such as the fitness value meets the requirements or the number of iterations reaches the upper limit), the optimal restoration path solution is output.

[0167] The following describes the working process of the present invention using a common scenario as an example. Figure 2 , which is Figure 1 Schematic diagram of the working scenario of the method.

[0168] In a certain area of ​​a city's power distribution network, the distribution station is densely populated with equipment and complex electromagnetic interference. The operation and maintenance personnel first collect the electromagnetic interference intensity distribution data in the station and the real-time current and voltage status of each line through sensors. Combined with the equipment operating conditions (such as load rate and equipment aging degree), determine the appropriate sensor selection scheme, such as high-sensitivity electromagnetic sensors and intelligent current sensors. One day, the line suddenly failed, and the system captured the arrival time difference data of the fault traveling wave. Combined with the line model parameters and wave speed differences, the fault traveling wave propagation path is calculated, the fault point is accurately located at 100 meters of a branch line, and the fault location result is generated. By associating the sensor selection scheme, it is confirmed that the fault point is located in the area where the sensor has been deployed, and it is initially judged to be a short circuit fault. Based on the real-time status data of the line and the fault location results, the system generates multiple candidate recovery path schemes. Further integrating the fault traveling wave characteristics (such as the degree of waveform distortion) and the line status parameters of the recovery path (such as line impedance and load conditions), a comprehensive evaluation model is established. With the highest power supply recovery efficiency and the lowest load loss rate as the optimization goal, the optimal recovery path scheme is finally determined to quickly restore power supply and reduce power outage losses.

[0169] In summary, the present invention discloses a method for monitoring a distribution station, including obtaining the distribution of electromagnetic interference intensity and real-time status data of the line in the distribution station, determining the sensor selection scheme in combination with the equipment working conditions; calculating the fault propagation path through the arrival time difference of the fault traveling wave, the line model parameters and the wave speed difference, determining the location of the fault point and generating the positioning result; associating the sensor selection with the fault location, judging whether the fault point is in the layout area, and preliminarily judging the type of fault; generating a candidate recovery path scheme based on the real-time status data and the fault location result; integrating the fault traveling wave characteristics with the line status parameters of the recovery path, establishing a comprehensive evaluation model, and determining the optimal recovery path scheme with the power supply recovery efficiency and the load loss rate as the optimization target. This method has the following effects: realizing the rapid and accurate positioning of distribution network faults and intelligent power supply recovery, and improving the reliability and efficiency of distribution network operation.

[0170] Reference Figure 2 The second embodiment of the present invention provides a power distribution station monitoring device, comprising:

[0171] Data acquisition module, used to obtain electromagnetic interference intensity distribution data in the power distribution station and real-time status data of each line;

[0172] A sensor selection module, used to determine a sensor selection scheme according to the electromagnetic interference intensity distribution data and in combination with equipment operating condition parameters;

[0173] The fault location result generation module is used to obtain the arrival time difference data of the fault traveling wave, calculate the propagation path of the fault traveling wave, determine the fault point location, and generate the fault location result by combining the preset line model parameters and wave speed difference characteristics;

[0174] A fault type judgment module is used to associate the sensor selection scheme with the fault location result, judge whether the fault point is located in the sensor deployment area, and obtain a preliminary judgment result of the fault type;

[0175] A recovery path candidate solution generating module, used to generate a recovery path candidate solution according to the real-time status data and in combination with the fault location result;

[0176] The module for determining the optimal restoration path is used to integrate the fault wave propagation characteristic parameters in the fault location result and the line state parameters in the restoration path candidate solution, establish a comprehensive evaluation model, and determine the optimal restoration path solution with power supply restoration efficiency and load loss rate as optimization targets.

[0177] It should be noted that a distribution station monitoring device provided in an embodiment of the present invention is used to execute all process steps of a distribution station monitoring method in the above embodiment, and the working principles and beneficial effects of the two correspond one to one, so they will not be repeated.

[0178] The embodiment of the present invention further provides an electronic device. The electronic device includes: a processor, a memory, and a computer program stored in the memory and executable on the processor, such as a data processing terminal program. When the processor executes the computer program, the steps in the above-mentioned embodiments of the power distribution station monitoring method are implemented, such as Figure 1 Alternatively, when the processor executes the computer program, the functions of the modules / units in the above-mentioned device embodiments are implemented, such as the module for determining the optimal recovery path.

[0179] Exemplarily, the computer program may be divided into one or more modules / units, which are stored in the memory and executed by the processor to implement the present invention. The one or more modules / units may be a series of computer program instruction segments capable of implementing specific functions, which are used to describe the execution process of the computer program in the electronic device.

[0180] The electronic device may be a computing device such as a desktop computer, a notebook, a PDA, and a smart tablet. The electronic device may include, but is not limited to, a processor and a memory. Those skilled in the art will appreciate that the above components are merely examples of electronic devices and do not constitute a limitation on the electronic device. The electronic device may include more or fewer components than the above components, or may combine certain components, or different components. For example, the electronic device may also include input and output devices, network access devices, buses, etc.

[0181] The processor may be a central processing unit (CPU), other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor, etc. The processor is the control center of the electronic device, and uses various interfaces and lines to connect various parts of the entire electronic device.

[0182] The memory can be used to store the computer program and / or module, and the processor realizes various functions of the electronic device by running or executing the computer program and / or module stored in the memory, and calling the data stored in the memory. The memory can mainly include a program storage area and a data storage area, wherein the program storage area can store an operating system, an application required for at least one function (such as a sound playback function, an image playback function, etc.), etc.; the data storage area can store data created according to the use of the mobile phone (such as audio data, a phone book, etc.), etc. In addition, the memory can include a high-speed random access memory, and can also include a non-volatile memory, such as a hard disk, a memory, a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), at least one disk storage device, a flash memory device, or other volatile solid-state storage devices.

[0183] Wherein, if the module / unit integrated in the electronic device is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on such an understanding, the present invention implements all or part of the processes in the above-mentioned embodiment method, and can also be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium, and the computer program can implement the steps of the above-mentioned various method embodiments when executed by the processor. Wherein, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form. The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording medium, U disk, mobile hard disk, disk, optical disk, computer memory, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), electric carrier signal, telecommunication signal and software distribution medium. It should be noted that the content contained in the computer-readable medium can be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer-readable media do not include electric carrier signals and telecommunication signals.

[0184] It should be noted that the device embodiments described above are merely schematic, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. In addition, in the accompanying drawings of the device embodiments provided by the present invention, the connection relationship between the modules indicates that there is a communication connection between them, which may be specifically implemented as one or more communication buses or signal lines. A person of ordinary skill in the art may understand and implement it without paying any creative effort.

[0185] The specific embodiments described above further illustrate the purpose, technical solutions and beneficial effects of the present invention. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. It is particularly pointed out that for those skilled in the art, any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for monitoring a power distribution station, characterized in that: Executed by a computer, including: Obtain the electromagnetic interference intensity distribution data in the power distribution station and the real-time status data of each line; Determine the sensor selection scheme according to the electromagnetic interference intensity distribution data and in combination with the equipment operating condition parameters; Obtain the arrival time difference data of the fault traveling wave, combine the preset line model parameters and wave speed difference characteristics, calculate the fault traveling wave propagation path, determine the fault point location, and generate the fault location result; Associating the sensor selection scheme with the fault location result, determining whether the fault point is located in the sensor deployment area, and obtaining a preliminary determination result of the fault type; Generate a candidate recovery path solution based on the real-time status data and in combination with the fault location result; The fault traveling wave propagation characteristic parameters in the fault location result and the line state parameters in the restoration path candidate solution are integrated to establish a comprehensive evaluation model, and the optimal restoration path solution is determined with the power supply restoration efficiency and load loss rate as optimization targets; The step of associating the sensor selection scheme with the fault location result to determine whether the fault point is located in the sensor deployment area and obtaining a preliminary determination result of the fault type includes: Obtain the fault location output by the fault location system, and perform spatial topological relationship calculations through the geographic information system in combination with the spatial information of the pre-established sensor group to determine the spatial inclusion relationship between the fault location and the sensor deployment area; If it is determined that the fault point is included in the sensor deployment area, all sensors identified by the sensor group are retrieved from the database according to the sensor group associated with the sensor deployment area; According to all the sensors, the operating parameter data of the sensors within a period of time before and after the fault occurs are retrieved to form time series data, and a time series length threshold is set; According to the time series data, the correlation coefficients between the operating parameters of each sensor are calculated and a correlation coefficient matrix is ​​constructed, and the correlation coefficients lower than the time series length threshold are set to zero; According to the correlation coefficient matrix, the sensors are divided into different clusters by a community discovery algorithm, each cluster represents a closely related sensor subset, so as to determine the central sensor of each sensor subset; For the sensor layout area of ​​the fault point, if it is determined that there are multiple sensor clusters, the central sensor closest to the fault point is selected, and its operating parameters are extracted for feature engineering processing; The data processed by feature engineering is input into the pre-trained fault diagnosis model to obtain the preliminary judgment result of the fault type.

2. The method for monitoring a power distribution station according to claim 1, characterized in that: Determining the sensor selection scheme according to the electromagnetic interference intensity distribution data and in combination with the equipment operating condition parameters includes: According to the electromagnetic interference intensity distribution data, measuring the intensity value of each interference source and generating an electromagnetic interference intensity distribution diagram; Associating the electromagnetic interference intensity distribution map with the equipment operation condition database to determine the interference intensity value range; According to the interference intensity value range, determine the sensor range value and select a sensor model that meets the range requirements; According to the characteristics of different interference sources, it is judged whether the sensitivity of the sensor model meets the requirements, so as to determine the sensor selection scheme.

3. The method for monitoring a power distribution station according to claim 1, characterized in that: The method of obtaining the arrival time difference data of the fault traveling wave, combining the preset line model parameters with the wave speed difference characteristics, calculating the fault traveling wave propagation path, determining the fault point location, and generating the fault location result includes: Extracting the time difference characteristic value in the traveling wave number according to the arrival time difference data of the fault traveling wave; According to the time difference characteristic value, extracting the line model characteristic from the line model parameter, matching it with the wave speed difference characteristic, and obtaining the wave speed difference characteristic; According to the wave speed difference characteristics, the fault traveling wave propagation path is calculated to obtain path length data; Determine the fault point location data according to the path length data and in combination with the line model parameters; Generate a fault location result based on the fault point location data and output the location value information; If there is a deviation in the fault location result, recalculate the wave speed difference characteristic and adjust the propagation path; Based on the adjusted propagation path, the fault point location is re-determined to generate the final fault location result.

4. The method for monitoring a power distribution station according to claim 1, characterized in that: The training process of the fault diagnosis model includes: Acquiring historical monitoring data of the sensor cluster, including vibration waveform spectrum, infrared thermal image spectrum and partial discharge soundprint signal; Performing multimodal feature extraction and normalization processing on the historical monitoring data to generate a standardized feature matrix; The standardized feature matrix is ​​used as input data, an initial fault diagnosis model is constructed based on a support vector machine algorithm, and the initial fault diagnosis model is trained according to a fused feature tensor and a fault type label to obtain a fault diagnosis model.

5. The method for monitoring a power distribution station according to claim 1, characterized in that: Generating a candidate recovery path solution according to the real-time status data and in combination with the fault location result includes: According to the real-time status data, the real-time active power and real-time reactive power of each line are calculated, and the switch status information of each line is obtained, so as to determine the on-off status of the line; The real-time active power is calculated by the following formula: The real-time reactive power is calculated by the following formula: According to the pre-established load classification rules and the load types connected to each line, each line is evaluated and scored for load importance; The load importance evaluation score of each line is calculated using the following formula: According to the load importance evaluation score, the load importance level is divided into a plurality of different levels, and the importance of each line is judged, thereby determining the load importance level of each line; The specific line section where the fault occurs is determined through the fault location result, and the accurate position of the faulty line is obtained in combination with the switch status information; According to the on / off state and the load importance level, different types of load transfer strategies are selected. If the faulty line is a non-important load line, load transfer is performed according to the actual situation; According to the topological structure of the faulty line and the remaining capacity information of the non-faulty lines, all feasible paths from the power supply point to the power supply area of ​​the faulty line are calculated by using the shortest path algorithm in graph theory; Comprehensively score all the feasible paths, and sort them from high to low according to the scores to generate candidate recovery path solutions; The following formula is used to comprehensively score all the feasible paths: in, Indicates real-time active power, represents the total number of samples, and Respectively represent The sampled values ​​of voltage and current in samples, represents the total number of load type factors, Indicates the real-time reactive power, It represents the phase difference between current and voltage. represents the load importance assessment score, Indicates The weight of each load type factor, Indicates that the line is The scores under each load type factor are , and represents the factors of each load type, and , Indicates the comprehensive rating. Indicates the line operation status score, Indicates the line length.

6. The method for monitoring a power distribution station according to claim 5, characterized in that: The method of calculating all feasible paths from the power supply point to the power supply area of ​​the faulty line based on the topological structure of the faulty line and the remaining capacity information of the non-faulty line through the shortest path algorithm in graph theory includes: Initialize the parameters of the shortest path algorithm and set the initial distance array as the first distance feature; Iterate according to the shortest path algorithm, obtain the second distance feature after the iteration, and record the number of iterations; Determine whether the second distance feature is better than the current distance feature, if so, update the second distance feature to the current distance feature; if not, discard the second distance feature; Continue with the next round of iteration, and when the number of iterations is greater than or equal to the preset number of iterations, output the current distance feature as the optimal distance feature; The optimal distance feature is used as a feasible path.

7. The method for monitoring a power distribution station according to claim 1, characterized in that: The method of fusing the fault traveling wave propagation characteristic parameters in the fault location result and the line state parameters in the candidate restoration path solution to establish a comprehensive evaluation model, taking the power supply restoration efficiency and the load loss rate as the optimization target, and determining the optimal restoration path solution includes: Acquire the traveling wave propagation characteristic parameters in the fault location result and the line state parameters in the restoration path candidate solution; Using data fusion technology, the propagation characteristic parameters and the line state parameters are fused at the feature level to generate a comprehensive characteristic vector; A comprehensive evaluation model based on random forest takes the comprehensive feature vector as input and power supply restoration efficiency and load loss rate as output; A particle swarm optimization algorithm is used to search for the optimal solution among the candidate restoration path solutions with maximizing power restoration efficiency and minimizing load loss rate as objective functions; If the objective function reaches a preset threshold, an optimal recovery path solution is output.

8. A power distribution station monitoring device, characterized in that: include: Data acquisition module, used to obtain electromagnetic interference intensity distribution data in the power distribution station and real-time status data of each line; A sensor selection module, used to determine a sensor selection scheme according to the electromagnetic interference intensity distribution data and in combination with equipment operating condition parameters; The fault location result generation module is used to obtain the arrival time difference data of the fault traveling wave, calculate the propagation path of the fault traveling wave, determine the fault point location, and generate the fault location result by combining the preset line model parameters and wave speed difference characteristics; A fault type judgment module is used to associate the sensor selection scheme with the fault location result, judge whether the fault point is located in the sensor deployment area, and obtain a preliminary judgment result of the fault type; A recovery path candidate solution generating module, used to generate a recovery path candidate solution according to the real-time status data and in combination with the fault location result; An optimal restoration path determination module is used to integrate the fault wave propagation characteristic parameters in the fault location result and the line state parameters in the restoration path candidate solution, establish a comprehensive evaluation model, and determine the optimal restoration path solution with power supply restoration efficiency and load loss rate as optimization targets; The step of associating the sensor selection scheme with the fault location result to determine whether the fault point is located in the sensor deployment area and obtaining a preliminary determination result of the fault type includes: Obtain the fault location output by the fault location system, and perform spatial topological relationship calculations through the geographic information system in combination with the spatial information of the pre-established sensor group to determine the spatial inclusion relationship between the fault location and the sensor deployment area; If it is determined that the fault point is included in the sensor deployment area, all sensors identified by the sensor group are retrieved from the database according to the sensor group associated with the sensor deployment area; According to all the sensors, the operating parameter data of the sensors within a period of time before and after the fault occurs are retrieved to form time series data, and a time series length threshold is set; According to the time series data, the correlation coefficients between the operating parameters of each sensor are calculated and a correlation coefficient matrix is ​​constructed, and the correlation coefficients lower than the time series length threshold are set to zero; According to the correlation coefficient matrix, the sensors are divided into different clusters by a community discovery algorithm, each cluster represents a closely related sensor subset, so as to determine the central sensor of each sensor subset; For the sensor layout area of ​​the fault point, if it is determined that there are multiple sensor clusters, the central sensor closest to the fault point is selected, and its operating parameters are extracted for feature engineering processing; The data processed by feature engineering is input into the pre-trained fault diagnosis model to obtain the preliminary judgment result of the fault type.

9. An electronic device, characterized in that: The method comprises a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein the processor implements the distribution substation monitoring method according to any one of claims 1 to 7 when executing the computer program.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes a stored computer program, wherein when the computer program is executed, the device where the computer-readable storage medium is located is controlled to execute the distribution substation monitoring method according to any one of claims 1 to 7.

Images